New glucocerebrosidase inhibitors by exploration of chemical diversity of N -substituted aminocyclitols using click chemistry and in situ screening

Article abstract of DOI:10.1021/jm101204u

A library of aminocyclitols derived from CuAAC reaction between N-propargylaminocyclitol 4 and a series of azides [1′25] is described and tested against GCase. Azides have been chosen from a large collection of potential candidates that has been filtered

Full text of DOI:10.1021/jm101204u

ARTICLE
pubs.acs.org/jmc
New Glucocerebrosidase Inhibitors by Exploration of Chemical
Diversity of N-Substituted Aminocyclitols Using Click Chemistry
and in Situ Screening
Lucía Díaz,† Josefina Casas,† Jordi Bujons,§ Amadeu Llebaria,† and Antonio Delgado*,†,‡
†Spanish National Research Council (Consejo Superior de Investigaciones Científicas, CSIC), Research Unit on Bioactive Molecules
(RUBAM), Departament de Química Biomꢀedica, Institut de Química Avanc-ada de Catalunya (IQAC), Jordi Girona 18-26, 08034
Barcelona, Spain
‡Universitat de Barcelona (UB), Facultat de Farmꢀacia, Unitat de Química Farmacꢀeutica (Unitat Associada al CSIC), Avenida Joan XXIII,
s/n, 08028 Barcelona, Spain
§Spanish National Research Council (Consejo Superior de Investigaciones Científicas, CSIC), Departament de Química Biolꢀogica i
Modelitzaciꢁo Molecular, Institut de Química Avanc-ada de Catalunya (IQAC), Jordi Girona 18-26, 08034 Barcelona, Spain
S Supporting Information
b
ABSTRACT: A library of aminocyclitols derived from CuAAC
reaction between N-propargylaminocyclitol 4 and a series of
azides [1-25] is described and tested against GCase. Azides
have been chosen from a large collection of potential candidates
that has been filtered according to physical and reactivity
constraints. A synthetic methodology has been optimized in
order to avoid the use of protecting groups on the aminocyclitol
scaffold. Because the reaction can be carried out in an aqueous
system, the resulting library members can be screened in situ
with minimal manipulation. From the preliminary GCase
inhibition data, the most potent library members have been individually resynthesized for further biological screening and
complete characterization. Some of the library members have shown biochemical data (IC50, Ki, and stabilization ratio) similar or
superior to those reported for NNDNJ. Docking studies have been used to postulate ligand-enzyme interactions to account for the
experimental results.
’ INTRODUCTION
and the potential to cross the blood-brain barrier of a small-
molecule approach, such as substrate reduction. This is especially
important to target CNS variants of the disease for which enzyme
replacement approaches are not suitable.10
Gaucher disease is one of the most prevalent lysosomal storage
disorders characterized by the accumulation of the sphingolipid
glucosylceramide (GC) in the lysosomes.1 The disease is caused
by the deficient activity elicited by several mutated forms of the
enzyme glucocerebrosidase (GCase, EC 3.2.1.45), a β-glucosi-
dase that hydrolyzes GC into glucose and ceramide (Cer).2
Cellular levels of the mutated, misfolded enzyme are abnormally
low due to its ER-associated degradation (ERAD) mediated by
the proteasome instead of proper folding and trafficking to the
lysosome.3,4 Several therapeutic strategies focused on Gaucher
disease have been developed over the last years. Classical
approaches rely on enzyme replacement and substrate reduction
therapies.5 However, new and promising approaches are starting
to emerge in order to circumvent some of the drawbacks
associated to the above strategies. Among them, the use of
pharmacological chaperones, small-molecule competitive inhibi-
tors of the target enzyme that raise the folded population of the
mutant enzyme at subinhibitory concentration by increasing the
stability of the enzyme-ligand complex,6-9 has become an
active field of research. This strategy combines the specificity
of an enzyme replacement approach with the oral bioavailability
Over recent years, we have worked actively on the development of
new aminocyclitols with potential applicability as pharmacological
chaperones of GCase. In particular, some symmetric N-alkyl sub-
stituted scyllo-aminoinositols (compounds 28, 29, and 30), closely
related to the iminocyclitol family of GCase inhibitors such as N-
nonyldeoxynojirimycin (NNDNJ)11 (Figure 1), behaved as efficient
competitive inhibitors of recombinant GCase12 and were able to
stabilize the enzyme against thermal denaturation.13 Moreover, some
of them were able to increase the activity of GCase in stably
transfected cell lines and in patient fibroblasts showing different
enzyme mutations.14
However, attempts to explore the chemical diversity of the
above aminocyclitols by the systematic variation of the amino
group functionalization afforded modest results.15 In addition, an
inherent limitation of our reported approach to N-substituted
Received: September 15, 2010
Published: March 03, 2011
r
2011 American Chemical Society
2069
dx.doi.org/10.1021/jm101204u J. Med. Chem. 2011, 54, 2069–2079
|

Journal of Medicinal Chemistry
ARTICLE
aminocyclitols lay on the required deprotection (by massive O-
benzyl removal) of each library member at the last synthetic step,
which rendered the process cumbersome. In this context, the
Huisgen Cu(I)-catalyzed alkyne-azide cycloaddition (CuAAC,
a paradigm of click chemistry) has become a widely used strategy
for chemical space exploration in drug design16,17 Because this
Cu-catalyzed process is extremely fast at room temperature and
compatible with water or water-rich solvent systems, the use of
protecting groups in most synthetic transformations is not
required. Interestingly, as the above process can be carried out
in water-containing solvent mixtures, it is possible to rapidly
assemble combinatorial libraries for direct in situ screening
without the need to carry out tedious isolation and purification
protocols.18-20 Moreover, the process can be performed, if
required, in a 96-well microtiter plate format for higher flexibility.
On the basis of these premises, we started a program aimed at the
exploration of the chemical diversity in N-substituted aminocy-
clitols based on the CuAAC click reaction. Preliminary results
with a set of aliphatic azides and a series of N-(ω-alky-
nyl)alkylaminocyclitols, differing on the length of the alkynyl
spacer, unravelled the importance of the correct placement of the
1,2,3-triazole moiety along the aliphatic chain in the resulting
systems.21 The above work set the stage for the development of
the protocol here reported, which allowed the construction of a small
demonstration library of N-substituted aminocyclitols by CuAAC
reaction between propargylamino derivative 4 (Scheme 1) and a
series of azides [1-25] (Figure 2). Interestingly, the process does
not require the use of protecting groups on the aminocyclitol
scaffold. Moreover, because the reaction can be carried out in an
aqueous system, the resulting library members can be screened in situ
with minimal manipulation. From the preliminary GCase inhibition
data, the most potent library members were individually resynthe-
sized for further biological screening and complete characterization.
Finally, docking studies have been used to postulate ligand-enzyme
interactions to account for the experimental results.
’ LIBRARY DESIGN
Reactants were selected from commercial databases based on
diversity criteria as well as synthetic suitability and economic
cost. Thus, starting from a collection of 528 commercial alkyl
alcohols and bromides, after applying an initial filter to remove
duplicates (compounds that would give the same product after
conversion to the corresponding azide) as well as highly ex-
pensive compounds and/or those with unfavorable properties
(for example, unsuitable molecular weight, incompatibility of
functional groups with biological matrices or the synthetic
sequence), we ended up with a collection of 343 potential
reactants. Diversity selection was carried out by enumerating
the virtual library of aminocyclitols, calculating a set of standard
molecular descriptors, and selecting a diverse subset using an
algorithm based on the calculation of the Euclidean distance among
compounds in normalized descriptor space and selection of those
which are farther apart (see Experimental Section and Supporting
Information). To this end, a subset of 25 diverse reactants22 was
selected, which were converted to the corresponding azides
[1]-[25] (Figure 2) following standard protocols.
’ LIBRARY SYNTHESIS
The aminocyclitols described in this study were obtained by
the CuAAC of N-(2-propynyl)aminocyclitol 4 with azides [1-
25] in H2O-THF as solvent system (Scheme 1).
Aminocyclitol 4 was obtained by a straightforward protocol
from tetra-O-benzyl conduritol B epoxide (2), as shown in
Scheme 1. Thus, the regio- and diastereoselective opening of
epoxide 2 with propargylamine, following our previously re-
ported protocol,23 afforded aminocyclitol 3.21,24 Removal of the
O-benzyl groups was carried out with BCl3 at -78 °C similar to
or superior than for 3 h, followed by MeOH quench at the same
temperature. Low temperature was crucial in this step to obtain
the required aminocyclitol 4 in high yields (higher than 90%) and
enough purity to be used in the CuAAC click reaction without
further manipulation. Initial deprotection attempts following our
previously reported conditions (typically, addition BCl3 acid
at -78 °C and further stirring at rt for 24 h)12 led to complex
Figure 1. Reported scyllo-aminocyclitols 28-30 as potential pharma-
cological chaperones and NNDNJ.
Figure 2. Azides used for CuAAC with aminocyclitol 4.
2070
dx.doi.org/10.1021/jm101204u |J. Med. Chem. 2011, 54, 2069–2079

Journal of Medicinal Chemistry
ARTICLE
Scheme 1a
a Reagents and conditions: (a) propargylamine, LiClO4, CH3CN; (b) BCl3, CH2Cl2 (-78 °C); (c) RN3 ([1-25], see Figure 2), CuSO4, sodium
ascorbate, H2O/THF (1:1).
reaction mixtures, probably due to the formation of byproduct
arising from insertion of the C-C triple bond into the B-Cl
bond of the Lewis acid reagent.25 Thus, a strict control of the
reaction temperature and time was required in order to obtain
aminocyclitol 4 efficiently.
Azides [1-25] were obtained either from commercially
available alcohols (by reaction of the corresponding mesylates
with NaN3) or bromides (by direct azide displacement), as
described in the Experimental Section.
To set up a protocol amenable to in situ screening, CuAAC
were carried out at small scale (typically 10 mg of 4) and 1.2
equiv/mol of the required azide in 1 mL of a (1:1) H2O/THF
mixture for 24 h at rt to ensure total consumption of the starting
alkyne. The identity of the resulting aminocyclitols 1[1-25] was
confirmed by UPLC-MS analysis of the crude reaction mixtures.
These were directly screened as GCase inhibitors and IC50 values
were determined, as shown in Table 1.
Aminocyclitols with IC50 below 3.5 μM were resynthesized
independently for complete characterization and further enzy-
matic studies (see below). In this case, CuAAC reactions with the
required azides were carried out from tetra-O-benzyl aminocy-
clitol 3, followed by massive O-benzyl removal by hydrogenolysis
over Pd/C catalyst under acidic conditions (Scheme 2).26 This
approach was not suitable for the synthesis of the dibenzyloxy
derivative 1[1], which was obtained by reductive alkylation of
unprotected scyllo-aminocyclitol 612 with aldehyde 27, following
our previously reported protocol.27 The required aldehyde 27
was obtained, in turn, by Swern oxidation of alcohol 26, arising
from CuAAC reaction of propargyl alcohol with azide [1]
(Scheme 2).
Table 1. IC50 Values (pH = 5.2) of Library Members by in
Situ Screening from Crude CuAAC Reaction Mixtures
aminocyclitola
IC50 (μM)
1[1]
1[2]
0.10
237
1[3]
66.5
1.90
10.3
12.2
3.41
2.30
35.9
30.9
17.3
114
1[4]
1[5]
1[6]
1[7]
1[8]
1[9]
1[10]
1[11]
1[12]
1[13]
1[14]
1[15]
1[16]
1[17]
1[18]
1[19]
1[20]
1[21]
1[22]
1[23]
1[24]
1[25]
10.0
4.32
134
2.25
0.15
0.06
466
181
44.5
453
325
5.80
2.19
a See Scheme 1 and Figure 2 for structural details. Compounds with IC50
below 3.5 μM were individually synthesized for further studies (see Table 2).
’ RESULTS AND DISCUSSION
The crude reaction mixtures were tested as inhibitors of
recombinant GCase (Imiglucerase, from Genzyme) at the opti-
mum pH for enzyme activity (pH = 5.2), as described in the
Experimental Section. Independent blank experiments with the
CuAAC reagents (CuSO4 and sodium ascorbate) and each of the
reaction partners (azides [1-25] and aminocyclitol 4) were also
carried out to rule out unspecific effects (See Supporting
Information). Results in Table 1 show the corresponding IC50
values for the 25 library members. These results were used as
criteria for preliminary screening and compound selection.
In agreement with our previous results,21 compounds bearing
a long aliphatic chain at the N-triazolyl moiety (1[17] and 1[18])
showed low IC50 values as GCase inhibitors (below 1 μM).
Chain length seemed crucial for enzyme inhibition, as evidenced
by the around 10 times higher IC50 value found for the n-octyl
derivative 1[16]. In addition, the N-alkyl substituent on the
triazolyl moiety seems quite sensitive to the presence of polar
groups. Thus, both the presence of oxygen or sulfur atoms along
the aliphatic chain (compounds 1[19]-1[23]), as well as the
introduction of a terminal hydroxyl group, proved detrimental
for enzyme inhibition in comparison with the corresponding
hydrocarbon counterparts (see compound 1[15] and compare
the pairs 1[16] vs 1[24] and 1[18] vs 1[25]). Among analogues
2071
dx.doi.org/10.1021/jm101204u |J. Med. Chem. 2011, 54, 2069–2079

Journal of Medicinal Chemistry
ARTICLE
Scheme 2a
possessing a short aliphatic chain connecting the triazole moiety
with a densely heteroatom-substituted terminal aromatic or
cycloaliphatic group (see Computational Studies).
Compounds with IC50 values below 3.5 μM were individually
synthesized, as depicted in Scheme 2. To our delight, IC50 values
(determined at the enzyme optimal pH = 5.2) were consistent with
those determined from the crude CuAAC reaction mixtures,
confirming the reliability of this approach as a tool for preliminary
screening of GCase inhibitors. Interestingly, IC50 values at pH = 7.4
(Table 2) were similar to, or even better (as for 1[8]) to those found
at pH = 5.2, which indicates the ability of these compounds to
interact with the target enzyme at the neutral pH of the ER.
Inhibition constant (Ki) and GCase stabilization after in vitro
thermal denaturation were also determined for the aminocycli-
tols shown in Table 2. This property is generally used as an in
vitro model to evaluate the potential of the tested compounds as
pharmacological chaperones.11 For comparative purposes, the
stabilization ratio (SR) has been used as a measure of GCase
stabilization. This parameter is defined as the ratio of the relative
enzymatic activities (with vs without inhibitor) at a given
inhibitor concentration and incubation time. The relative enzy-
matic activity is obtained, in turn, by comparing the enzyme
activity at 48 °C for a given aminocyclitol concentration and
incubation time with that of a control assay at 37 °C (a
nondenaturalizing temperature) under otherwise identical con-
ditions. Relative activities for selected aminocyclitols (those with
Ki values around or lower than 1.0 μM) at different concentra-
tions and incubation times are plotted in Figure 3, while SR
values at 100 μM aminocyclitol concentration and 60 min
incubation time are recorded in Table 2. For the matter of
comparison, NNDNJ and 3013 (Figure 1) were also tested under
the same experimental conditions.
a Reagents and conditions: (a) H2, Pd/C, MeOH, HCl; (b) propargyl
alcohol, CuAAC; (c) Swern oxidation; (d) supported borohydride (see
text).
bearing a cycloaliphatic or an aromatic substituent, only 1[1],
1[4], 1[7], and 1[8] showed IC50 values below 3.5 μM, which
was set as the threshold for further studies. The high inhibitory
activity of 1[1] (bearing a bulky aromatic 3,5-dibenzyloxybenzyl
group) is remarkable, because this compound turned out to be
the most potent of the library, not only as GCase inhibitor (Ki =
0.05 μM, Table 2) but also as in vitro pharmacological chaperone
by its ability to stabilize GCase after thermal denaturation (see
below). A still remarkable activity profile was found for the
related analogue 1[4] (Ki = 0.77 μM, Table 2), whose 3,3-
diphenylpropyl group adopts a spatial disposition when bound to
GCase close to that observed for 1[1] (see Supporting In-
formation). A similar activity profile was found for 1[8] (R =
4-cyclohexylbutyl), whereas IC50 for the adamantylmethyl deri-
vative 1[7] was only slightly below the established threshold.
As in the aliphatic series, compound 1[14], a lower homo-
logue of 1[8], showed a reduced inhibitory activity (IC50 around
4 μM) and was thus disregarded. Library members arising from
CuAAC between azides [3], [5], [6], [9], [10], [11], and [13],
and alkyne 4 (Scheme 1) showed IC50 values between 10 and 70
μM and were also disregarded, as well as those arising from azides
[2], [12], and [15], with IC50 values above 100 μM. From a
structural standpoint, this set of compounds is characterized by
Inspection of Figure 3 reveals that the 3,3-diphenylpropyl and
the adamantylmethyl derivatives (compounds 1[4] and 1[7],
respectively) behave as weak thermal stabilizers at high concen-
tration (150 μM), while they act mainly as enzyme inhibitors at
lower concentrations. According to our previous findings, the
lipophilicity of the N-alkyltriazolyl moiety seems crucial in this
kind of compound, as evidenced by comparison of the activity
profiles of the N-octyl (1[16]), the N-decyl (1[17]),21 and the
N-undecyl (1[18]) derivatives as GCase and GCS inhibitors
(Table 2). In agreement with this observation, the increase of two
methylene units (moving from 1[16] to 1[18]) led to a
Table 2. GCase Inhibitory Activity, Thermal Stabilization, and GCS Inhibition of Selected Compounds
aminocyclitol
IC50 (μM)a
IC50 (μM)b
Ki (μM)c
stabilization ratio (SR)d
% inhibition GCasee
% inhibition GCS f
1[1]
0.05
1.5
0.06
1.8
0.05
0.77
1.05
0.29
1.84
0.09
0.05
0.89
0.30 h
0.30 i
31.2
0.5
93
73
63
30
20
90
97
28
ndl
nd
9
1[4]
18
1[7]
2.2
1.9
0.7
32
1[8]
1.5
0.30
1.1
7.5
8
1[16]g
1[17]g
1[18]
1[25]
30
1.0
1.0
0
0
0.10
0.12
1.8
0.09
0.10
1.3
21.0
30.5
6.9
29
28
1.8 h
(nd)
4.9
nd
30j 40k
NNDNJ
1.30
0.30
9.2
a GCase inhibition at pH = 5.2. b GCase inhibition at pH = 7.4. c GCase competitive inhibitors (pH = 5.2). d Relative enzymatic activities (inhibitor vs
control) at 100 μM inhibitor and 60 min incubation time. For values at different inhibitor concentration, see Figure 3 and Supporting Information. e In
intact fibroblasts at 50 μM aminocyclitol. f Determined in cell lysates using CerNBD as substrate at 250 μM inhibitor, see Supporting Information for
details. g Data from ref 21. h See ref 13. i See ref 32. j 10 μM inhibitor. k 50 μM inhibitor. l not determined.
2072
dx.doi.org/10.1021/jm101204u |J. Med. Chem. 2011, 54, 2069–2079

Journal of Medicinal Chemistry
ARTICLE
Figure 3. Relative enzymatic activity after thermal denaturation (48 °C)
for: 20 min (white bar), 40 min (gray bar), and 60 min (black bar) at the
indicated aminocyclitol concentrations (μM), compared to the correspond-
ing assay at 37 °C. C: control without inhibitor.
Figure 4. GCase inhibitory activities from Table 1, expressed
as -log(IC50), vs predicted clogP values (see Supporting Information)
for aminocyclitols 1[1]-[1[25].
both GCase structures, establishing hydrogen bond interactions
mainly with residues D127, W179, E235, W381, and N396, and
most of them exhibit an additional hydrogen bond between the
protonated secondary amino group and residue E235. Poses
obtained from docking to the 2V3E structure show the triazole-
attached substituent occupying the valley between loops L1-L2
and L3-L4, while those from docking to 2NSX show that the
substituent of the triazole is oriented toward or inside the narrow
groove between loops L1 and L3. These differences are more
noticeable for the compounds with linear chains, ie. 1[16]-
1[25], than for those with smaller or more rigid substituents on
the triazole ring, such as 1[1]. Thus, the docked poses obtained
for compound 1[1] from the two GCase targets (Figure 5C,D)
suggest that the two aromatic “arms” of the 3,5-dibenzyloxyben-
zyl substituent can establish hydrophobic interactions with the
side chains of residues in loops L1 (F347, W348) and L3 (W312,
L314, and F316). Similar arrangements were observed for the
less active compounds 1[4], 1[7], and 1[8] (see Supporting
Information), while the long and linear chains of the also active
1[16], 1[17], 1[18], and 1[25] were located in the valleys
between loops L3 and L4 or L1 and L3, depending on which
target structure was used (see Supporting Information and ref
21). Although the docking results for the rest of the less, or much
less, active compounds suggest that these could adopt bound
conformations which are similar to those of the active com-
pounds, most of them exhibit a lower hydrophobicity (clogP <
1.0) that probably reduces their capacity to efficiently interact
with the above-mentioned hydrophobic residues on loops L1 and
L3, as well as others (L240, L241, L286, L287) located near the
entrance to the catalytic site. Concerning the role of the triazole
ring on the binding to GCase, the present docking results do not
provide a clear answer. However, as it was previously observed,21
the docked poses obtained for some of the compounds with
structure 2NSX (i.e. 1[7], see Figure S2D of Supporting In-
formation) show the triazole ring in an orientation in which one
of the nitrogen atoms forms a hydrogen bond with residue Q284,
while other poses suggest that it could establish a stacking
interaction with Y313 (i.e. 1[1], Figure 5F). Alternatively, from
docking with 2V3E, a hydrogen bond with Y244 can be inferred
(i.e. compounds 1[8] and 1[18], see Figure S2E,G of Supporting
Information). Hence, the present results obtained from a larger
variety of compounds, suggest that the contribution of the
triazole ring, being located just outside of the tight cyclitol
significant SR enhancement (Table 2 and Figure 3). On the other
hand, introduction of a polar substituent in the aliphatic chain, as
in the N-(10-hydroxydecyl)triazolyl derivative 1[25], led to a
substantial decrease of both the relative enzyme activity and the
SR in comparison with the nonhydroxylated counterpart 1[17].
Very interestingly, the N-[3,5-bis(benzyolxy)benzyl] derivative
1[1] turned out to be the best compound of the library according
to its relative enzymatic activity and enzyme thermal stabiliza-
tion. In addition, it is a weak GCS inhibitor, a strong in vivo
GCase inhibitor (intact human wt fibroblasts) (Table 2), non-
cytotoxic (MTT test),28 and totally devoid of activity against a
panel of commercial glycosidases.29
’ COMPUTATIONAL STUDIES
We have previously reported that the GCase inhibition potency of a
related series of aminocyclitols showed a clear dependence on their
hydrophobicity, expressed as clogP.21 A similar correlation is also
observed for aminocyclitols 1[1]-1[25] (Figure 4), supporting our
previous hypothesis about the importance of hydrophobicity for GCase
inhibitory activity. However, other factors like the relative position of
the triazole ring with respect to the aminocyclitol moiety appeared to
influence the potency of these compounds because analogues differing
only on the length of the alkyl chain linker between the amino and
triazole groups showed large differences in activity, greater than 100-
fold in some cases, despite their very similar clogP values.
Docking studies with our previous collection of compounds
allowed us to propose two possible binding modes into the active
center of GCase, which arise from two slightly different struc-
tures for the protein (PDB codes 2V3E and 2NSX).21 These two
structures differ on their surface topology close to the entrance to
the active site. Thus, on 2V3E the access to a hydrophobic groove
between loops 1 and 3 (L1 and L3) is blocked while on 2NSX it is
open (Figure 5). Docking studies with the above two GCase
structures have now been extended to aminocyclitols 1[1]-
1[25] to determine their potential binding modes.
Parts A and B of Figure 5 show the best docked poses obtained
for compounds 1[1]-1[25] using GCase structures 2V3E and
2NSX, respectively, as targets. As determined for the previous
collection of analogue aminocyclitols, most of the compounds in
this work showed different binding modes depending on which
structure was used as docking target. Hence, all the compounds
could place their cyclitol moiety inside the active site cavity of
2073
dx.doi.org/10.1021/jm101204u |J. Med. Chem. 2011, 54, 2069–2079

Journal of Medicinal Chemistry
ARTICLE
Figure 5. Best poses obtained for aminocyclitols 1[1]-1[25] docked against GCase structures 2V3E (A) and 2NSX (B). L1-L4 denote the position of the
four loops that constitute the entrance to the GCase active site.21 In (B), the black arrow shows the accessible groove between loops L1 and L3. Detail of the best
poses obtained for compound 1[1] docked against 2V3E (C) and 2NSX (D), and corresponding interaction diagrams (E and F, respectively) that show strong
hydrogen bond interactions between the cyclitol hydroxyl groups and the side chains of residues D127, W179, E235, W381, and N396, as well as between the
protonated amino group and residue E235. The positive charge on this group is also stabilized by a π-cation interaction with the aromatic ring of residue Y313.
binding cavity, might depend on the nature of the substituent and
its capacity to interact with one or another region on the enzyme
surface. Studies to further elucidate the role of the triazole moiety
in this kind of aminocyclitols are currently underway and will be
reported in due course.
of N-substituted aminocyclitols as GCase inhibitors. The highly
aqueous content of the reaction media makes possible the direct
screening of the products by simple dilution of the crude reaction
mixtures in the adequate buffer. The process is suitable for its
implementation into a microtiter plate format and, hence,
amenable to combinatorial protocols. In this work, a small
demonstration library of 25 compounds has been produced in
order to test the potential of this approach. Maximum diversity
among the library members has been secured by application of a
proper selection algorithm from an initial collection of 343
compounds. Library members with IC50 values below 3.5 μM
’ CONCLUSION
A click chemistry approach based on the CuAAC reaction
between N-propargyl aminocyclitol 4 and a series of azides has
been optimized for the synthesis and in situ screening of a variety
2074
dx.doi.org/10.1021/jm101204u |J. Med. Chem. 2011, 54, 2069–2079

Journal of Medicinal Chemistry
ARTICLE
were individually synthesized and tested as GCase inhibitors and
also for their ability to induce enzyme thermal stabilization, an in
vitro indication of their potential as pharmacological chaperones.
Activity values of individually synthesized compounds were
consistent with those determined from the crude CuAAC
reaction mixtures, confirming the reliability of this approach as
a tool for preliminary screening of GCerase inhibitors. Note-
worthy, the highly lipophilic compounds 1[1], with two aromatic
“arms”, and 1[18], with a linear aliphatic side chain, showed the
highest enzyme stabilization ratios and are promising candidates
for further development. In summary, this approach has shown
to be very effective and simple for the rapid identification of
selective, potent GCase inhibitors within the aminocyclitol
family of compounds.
and 100 MHz (for 13C), respectively, unless otherwise indicated.
Chemical shifts (δ) are reported in ppm relative to the solvent
(CDCl3) signal. In cases where triazole carbon atoms in compounds
1[x] were not apparent, they were indirectly assigned by a combination
of 2D experiments (gHSQC and gHMBC) by means of long-range
scalar coupling through two or three chemical bonds of the CH2 groups
adjacent to the triazole ring. ESI/HRMS spectra were recorded on a
Waters LCT Premier mass spectrometer. Parallel syntheses of azides
were carried out in a 12-position Carousel reaction station under inert
atmosphere at low temperature, when required. Azide purifications were
carried out in a 12-position parallel purification system. Parallel click
chemistry reactions were carried out in 5 mL screw cap vials, provided
with a small magnetic bar, and placed in 24 position reaction blocks.
Parallel evaporations were carried out in a centrifugal evaporator.
For complete characterization of azides [1]-[25], see Supporting
Information. Purity of compounds used for enzyme assays was g95%,
as judged by HPLC analysis (Gemini C18 column, 4.6 mm ꢀ 250 mm, 5
μm) under the following chromatographic conditions: mobile phase A,
water containing 0.1% v/v trifluoroacetic acid; mobile phase B, ACN
containing 0.1% v/v trifluoroacetic acid; flow rate of 1.0 mL/min;
detection, SATIN-ELS (evaporative light scattering) (λ = 254 nm);
gradient elution, 0 min, from 80% A/20% B to 0% A/100% B over next
20 min. Each run was followed by a 3 min wash with 80% ACN,
20% water.
’ EXPERIMENTAL SECTION
Library Design. Electronic collections of alkyl alcohols and bro-
mides (R-OH and R-Br) were obtained as sdf formatted files from the
web pages of commercial brands. The initial collection of reactants
contained 528 compounds, which were processed in steps with the
program MOE (version 2008.10, Chemical Computing Group, Mon-
treal) by (1) converting them into 3D structures, (2) clipping the leaving
groups (OH or Br) and keeping the R-groups, (3) removing duplicate
R-groups, and (4) filtering undesired R-groups according to a predefined
set of rules (see Supporting Information). An additional filtering step
was manually carried out to remove compounds unsuitable from the
synthetic point of view or with too high price. In this way, a library of 343
R-groups (reactants) was obtained which was used to enumerate the
virtual library of aminocyclitols. The structures of the compounds in this
aminocyclitol library were energy minimized using the implemented
MMFF94x force field, a modified version of the MMFF94s force
field.30,31 The library was characterized by calculating a set of 47 standard
molecular descriptors including different physicochemical properties, as
well as spatial, topological, and information content indices (see
Supporting Information). Principal component analysis (PCA) showed
that the first three most important components accounted for 73% of the
variance, while up to nine components were required to account for
more than 90% of the variance (see Supporting Information). The
Diverse Subset module implemented in MOE, which relies on the
calculation of the Euclidean distance among compounds in normalized
descriptor space, was used to carry a diverse selection of 30 compounds
from the virtual library of aminocyclitols. However, synthetic problems
with some of them forced us to modify the original selection to the final
collection of reactants that was used to obtain azides [1]-[25]. The
diversity of this final subset was assessed by cell-based methods, both in
terms of space coverage (number of cells covered by selected subset/
total number of cells) and population coverage (total number of
compounds included in represented cells/total number of molecules
in library). Hence, binning the diversity space into 27 clusters (closest
possible number of bins to the number of selected compounds) that
arise from subdividing the space into three equiprobable regions for each
of the three most important principal components, the 25 aminocyclitols
resulting from the selected subset of reactants, representing only a 7.3%
of the possible products, achieved a 52% space coverage (14 clusters
occupied) and 53% population coverage (183 compounds represented).
Docking. Computational docking of compounds 1[1]-1[25]
against GCase structures 2V3E and 2NSX was performed as previously
described.21
(1R,2S,3r,4R,5S,6s)-6-(Prop-2-ynylamino)cyclohexane-1,2,3,4,5-pentaol
(4). A solution of aminocyclitol 321 (100 mg, 0.17 mmol) in CH2Cl2
(20 mL) is cooled to -78 °C under Ar and treated with 1 M BCl3
solution in heptane (1.8 mL, equivalent to 1.8 mmol). After stirring for 3
h at -78 °C, the reaction was quenched by addition of MeOH (4 mL),
previously cooled to -78 °C, and allowed to warm to rt with stirring.
Solvents were removed under reduced pressure in a centrifugal eva-
porator. EtOAc (3 mL) was next added to the residue, and the resulting
mixture was sonicated until precipitation of a white solid. This was
collected by filtration, thoroughly washed with additional EtOAc (3 ꢀ
2 mL), and dried to afford the final aminocyclitol 4 HCl (35 mg, 93%
3
yield). 1H NMR (500 MHz, CD3OD), δ (ppm) = 4.20 (d, 2H, J = 2.4
Hz) 3.56 (dd, 2H, J = 9.1 Hz, J = 10.4 Hz) 3.29 (t, 2H, J = 9.2 Hz) 3.23
(m, 2H) 3.15 (t, 1H, J = 10.7 Hz). 13C NMR (100 MHz, CD3OD), δ
(ppm) = 79.05, 76.62, 76.55, 75.18, 75.13, 71.00, 70.50, 62.41, 36.94. IR:
2128 cm-1 (CtC). HRMS: calcd for C9H16NO5 (M þ Hþ), 218.1028;
found, 218.1045
General Method for Parallel Click Chemistry and in Situ Screening.
A solution of alkyne 4 (10 mg, 0.046 mmol) and azide [1]-[25] (0.055
mmol) in a 1:1 mixture of H2O/THF (1 mL) in a 5 mL screw cap vial
was treated with a catalytic amount of CuSO4 5H2O (around 250 μg)
3
followed by sodium ascorbate (5 μmol, around 1 mg). After stirring for
24 h at rt, an aliquot was analyzed (UPLC-HRMS) to confirm click
adduct formation (see Supporting Information). The crude reaction
mixtures containing compounds 1[1-25] were directly used as mother
solutions (46 mM) for in vitro IC50 calculation of GCase inhibition, as
described above (for results, see Table 1).
Synthesis of Individual Aminocyclitols. (1RS,2SR,3RS,4RS,5SR,6RS)-
2,3,4,5-Tetrakis(benzyloxy)-6-(prop-2-ynylamino)cyclohexanol (3). A solu-
tion of the starting epoxide 2 (500 mg, 0.96 mmol) in CH3CN (10 mL)
was slowly added dropwise under argon at rt over previously dried LiClO4 (2.47
g, 23.2 mmol). A solution of 9.6 mmol (10 equiv/mol) of the required amine in
CH3CN (2 mL) was next added and the reaction mixture was stirred at 80 °C
under Ar. After 18 h, the reaction mixture was cooled to rt, quenched with H2O
(10 mL), extracted with CH2Cl2 (3 ꢀ 20 mL), and dried over anhydrous
Mg2SO4. Filtration and evaporation afforded crude aminocyclitols (3:1 mixture
of diastereomers), which were purified by filtration through a plug of silica on
elution with hexanes/EtOAc (6:4) to afford aminoalcohol 5 (435 mg, 79%
yield). 1H NMR: 7.36 (m, 20H), 4.95 (m, 5H), 4.82 (dd, 1H, J = 3.5 Hz, J =
11.1 Hz), 4.76 (d, 1H, J = 11.2 Hz), 3.62 (m, 2H), 3.46 (m, 3H), 2.71 (dt, 1H,
Chemistry. Solvents were distilled prior to use and dried by
1
standard methods. FT-IR spectra are reported in cm-1. H and 13C
NMR spectra were obtained in CDCl3 solutions at 500 MHz (for 1H)
2075
dx.doi.org/10.1021/jm101204u |J. Med. Chem. 2011, 54, 2069–2079

Journal of Medicinal Chemistry
ARTICLE
J= 1.8 Hz, J=9.9Hz),2.28(d,1H,J=2.1Hz).13CNMR:127.6-128.5, 84.52,
83.67, 83.04, 81.91, 75.24-75.78, 73.13, 71.51, 60.08, 37.53. HRMS: calcd for
C37H40NO5 (M þ Hþ), 578.2906; found, 578.2929.
J = 2.0 Hz) 5.44 (s, 2H) 5.02 (s, 4H) 4.79 (s, 2H). 13C NMR: 160.54,
136.51, 128.79, 128.28, 127.68, 107.37, 102.26, 70.30. HRMS: calcd for
C24H23N3O3Na (M þ Hþ), 424.1637; found, 424.1627.
b. Swern Oxidation of 26: Synthesis of 1-[3,5-Bis(benzyloxy)benzyl]-
1H-1,2,3-triazole-4-carbaldehyde (27). A solution of alcohol 26 (100
mg, 0.25 mmol) in EtOAc (10 mL) was treated with IBX (203 mg, 0.75
mmol). After stirring at reflux temperature for 4 h, the reaction mixture
was allowed to cool to rt and filtered over a pad of Celite, which was
thoroughly washed with EtOAc. The combined organic phases were
evaporated to dryness to give an oil, which, after hexane washings,
Click Chemistry Reaction of ω-Alkynyl Aminocyclitol 3
with Selected Azides. To a solution of 3 (50 mg, 0.23 mmol) and the
selected azide (0.28 mmol) in 1:1 H2O/THF (3 mL) in a screw cap glass
vial, a catalytic amount of CuSO4 5H2O and sodium ascorbate (0.023
3
mmol, 4.6 mg) were added. After stirring at rt for 10 min, the reaction
mixture was diluted with EtOAc (10 mL). The organic phase was co-
llected, dried over MgSO4, and worked up as usual to give a residue which
was purified by flash chromatography on elution with hexanes-EtOAc
(from 1:1 to 0:1) to afford the corresponding cycloaddition adducts 5.
(1RS,2SR,3RS,4RS,5SR,6RS)-2,3,4,5-Tetrakis(benzyloxy)-6-[[1-(3,3-diphen-
ylpropyl)-1H-1,2,3-triazol-4-yl]methylamino]cyclohexanol (5[4]). Ob-
tained in 76% yield. 1H NMR: 7.33 (m, 20H, 10H, 1H) 4.94 (m, 8H)
4.27 (s, 2H) 4.09 (m, AB system, 2H) 3.93 (t, 1H, J = 7.6 Hz) 3.67 (m,
5H) 2.68 (m, 2H, 1H). 13C NMR: 143.24, 138.71-138.36, 126.88-
128.84, 121.60, 80.72- 84.70, 73.01, 62.49, 48.25, 35.90. HRMS: calcd
for C52H55N4O5 (M þ Hþ), 815.4172; found, 815.4197.
1
afforded aldehyde 27 as a white solid in quantitative yield. H NMR:
10.14 (s, 1H) 7.99 (s, 1H) 7.37 (m, 10H) 6.64 (t, 1H, J = 2.2 Hz) 6.51 (d,
2H, J = 2.2 Hz) 5.49 (s, 2H) 5.02 (s, 4H). 13C NMR: 185.14, 160.68,
148.09, 136.34, 135.51, 128.79, 128.33, 127.64, 125.31, 107.59, 102.60,
70.34, 54.68.
c. Reductive Alkylation of Aminocyclitol 6 with Aldehyde 27:
Synthesis of (1R,2S,3r,4R,5S,6s)-6-[(1-(3,5-Bis(benzyloxy)benzyl)-1H-
1,2,3-triazol-4-yl)methylamino]cyclohexane-1,2,3,4,5-pentaol (1[1]).
To a screw cap glass vial, a solution of aminocyclitol 612 (15 mg, 84
μmol) in DMSO (3.5 mL), followed by AcOH (70 μL) and 22 mg (55
μmol) of aldehyde 27 at rt, was added. After shaking for 1 h, supported
cyanoborohydride (150 mg, equivalent to 300 μmol) was added
portionwise and the mixture was shaken for additional 12 h, followed
by addition of Wang resin (15 mg, equivalent to 45 μmol) and further
shaking of the resulting slurry for 12 h. Resins were filtrated and
thoroughly washed with MeOH/H2O (9/1 mixture). The combined
filtrates were evaporated to dryness and the residue taken up in 9/1
MeOH/H2O (3 mL), treated with carbonate resin (500 mg, equivalent
to 1.75 mmol) and Wang resin (15 mg, equivalent to 45 μmol), and
stirred for 4 h. Resins were filtrated, washed with MeOH/H2O (9/1
mixture), and evaporated to dryness to give a residue, which was purified
by flash chromatography (CH2Cl2/MeOH/NH4OH from 94:5:1 to
(1RS,2SR,3RS,4RS,5SR,6RS )-2,3,4,5-Tetrakis(benzyloxy)-6-[[1-adamantyl-
1H-1,2,3-triazol-4-yl]methylamino]cyclohexanol (5[7]). Obtained in 92%
yield. 1H NMR: 7.34 (m, 20H, 1H) 4.90 (m, 8H) 4.05 (m, 2H, 2H) 3.57 (m,
5H) 2.66 (m, 1H) 1.99 (m, 3H) 1.64 (m, 6H,) 1.49 (m, 6H). 13C NMR:
138.76-138.37, 128.63-127.76, 121.32, 84.82, 84.20, 83.11, 81.88, 75.99,
75.59, 73.05, 62.33, 40.32, 36.60, 28.16. HRMS: calcd for C48H57N4O5 (M þ
Hþ), 769.4330; found, 769.4352.
(1RS,2SR,3RS,4RS,5SR,6RS )-2,3,4,5-Tetrakis(benzyloxy)-6-[[1-(4-cyclohex-
ylbutyl)-1H-1,2,3-triazol-4-yl]methylamino]cyclohexanol (5[8]). Ob-
1
tained in 87% yield. H NMR: 7.41 (m, 20H, 1H) 4.93 (m, 8H) 4.31
(t, 2H, J = 6.7 Hz) 4.13 (m, AB system, 2H) 3.63 (m, 5H) 2.71 (m, 1H)
1.87 (m, 2H) 1.69 (m, 4H) 1.24 (m, 9H) 0.89 (m, 2H). 13C NMR:
146.52, 138.70-138.35, 127.61-128.74, 121.47, 84.71, 84.06, 83.05,
81.62, 75.90, 75.58, 72.92, 61.53, 50.41, 43.15, 37.48, 36.88, 33.37, 30.64,
26.71, 26.41, 23.87. HRMS: calcd for C47H59N4O5 (M þ Hþ),
759.44850; found, 759.4548.
1
83:17:1). Aminocyclitol 1[1] (12 mg, 25% yield) was obtained. H
NMR (500 MHz, CD3OD): 8.02 (s, 1H) 7.40 (m, 10H) 6.63 (m, 1H)
6.59 (d, 2H, J = 1.8 Hz) 5.54 (s, 2H) 5.04 (s, 4H) 4.42 (m, AB system,
2H) 3.48 (t, 2H, J = 9.6 Hz) 3.24 (m, 3H,) 2.88 (t, 1H, J = 10.5 Hz). 13
C
(1RS,2SR,3RS,4RS,5SR,6RS)-2,3,4,5-Tetrakis(benzyloxy)-6-[[1-undecyl-1H-
1,2,3-triazol-4-yl]methylamino]cyclohexanol (5[18]). Obtained in 83%
yield. 1H NMR: 7.38 (m, 20H, 1H) 4.94 (m, 8H) 4.30 (t, 2H, J = 7.2
Hz) 4.08 (m, AB system, 2H, J = 13.6 Hz) 3.57 (m, 5H) 2.67 (m, 1H)
1.87 (m, 2H) 1.21 (m, 16H) 0.90 (t, 3H, J = 6.8 Hz) 13C NMR: 147.20,
138.87-138.54, 128.77-127.91, 121.32, 84.91, 84.29, 83.26, 82.03,
76.11, 76.08, 75.71, 73.21, 61.72, 50.54, 43.28, 32.14, 30.54, 29.81, 29.80,
29.66, 29.56, 29.28, 26.76, 22.93, 14.40. HRMS: calcd for C48H63N4O5
(M þ Hþ), 775.4798; found, 775.4825.
NMR (100 MHz, CD3OD): 161.79, 151.35 (Cq triazole, 2D assign-
ment), 138.70, 138.35 (CH triazole, 2D assigment), 129.54, 128.98,
128.64, 125.93, 108.43, 102.93, 76.69, 75.31, 71.98, 71.12, 62.93, 54.96,
42.31. HRMS: calcd for C30H35N4O7 (M þ Hþ), 563.2506; found,
563.2511.
Debenzylation of Aminocyclitols 5 by Catalytic Hydroge-
nation. In a typical reaction, each of the above aminoalcohols (0.2 mmol)
was dissolved in a mixture of THF (5 mL), concentrated HCl (few drops),
and 50 mg of 10% Pd/C in a glass pressure flask. The flask was repeatedly
filled and evacuated with hydrogen and vigorously stirred at room
temperature for 24 h under H2 (1.5 atm). The reaction mixture was next
filtered through a plug of Celite, which was washed with 3 ꢀ 2 mL THF/
MeOH (1:1). The combined filtrates and washings were concentrated to
give aminocyclitols 1 as the corresponding hydrochloride salts.
(1RS,2SR,3RS,4RS,5SR,6RS)-2,3,4,5-Tetrakis(benzyloxy)-6- [[1-(10-hydro-
xydecyl)-1H-1,2,3-triazol-4-yl]methylamino]cyclohexanol (5[25]). Ob-
1
tained in 85% yield. H NMR: 7.38 (m, 20H, 1H) 4.91 (m, 8H) 4.30
(t, 2H, J = 7.2 Hz) 4.08 (m, AB system, 2H, J = 14.1 Hz) 3.59 (m, 7H)
2.67 (m, 1H) 1.87 (m, 2H) 1.56 (m, 2H) 1.21 (m, 12H). 13C NMR:
147.06, 138.40-138.73, 127.77-128.64, 121.21, 84.77, 84.15, 83.12,
81.90, 75.98, 75.94, 75.88, 73.06, 63.01, 61.62, 50.36, 43.18, 32.81, 30.32,
29.42, 29.36, 29.29, 28.95, 26.46, 25.78. HRMS: calcd for C47H61N4O6
(M þ Hþ), 777.4591; found, 777.4620.
(1R,2S,3r,4R,5S,6s)-6-[[1-(3,3-Diphenylpropyl)-1H -1,2,3-triazol-4-yl]meth-
ylamino]cyclohexane-1,2,3,4,5-pentaol (1[4]). Obtained from 5[4] in
1
quantitative yield. H NMR (500 MHz, CD3OD): 8.07 (s, 1H) 7.29
Synthesis of Aminocyclitol 1[1]. a. Huisgen Cycloaddition
between Azide [1] and Propargyl Alcohol: Synthesis of [1-[3,5-Bis-
(benzyloxy)benzyl]-1H-1,2,3-triazol-4-yl]methanol (26). To a solution
of propargyl alcohol (65 mg, 1 mmol) and azide [1] (400 mg, 1.16
mmol) in 1:1 H2O/THF (4 mL) in a screw cap glass vial, a catalytic
(d, 8H, J = 4.3 Hz) 7.19 (m, 2H) 4.57 (s, 2H) 4.38 (m, AB system, 2H)
3.94 (t, 1H, J = 7.9 Hz) 3.58 (t, 2H, J = 9.7 Hz) 3.23 (m, 3H) 3.08 (t, 1H,
J = 10.7 Hz) 2.71 (m, 2H). 13C NMR (100 MHz, CD3OD): 151.52 (Cq
triazole, 2D assignment), 144.97, 140.61 (CH triazole, 2D assigment),
139.2, 129.73, 128.84, 127.65, 76.58, 75.16, 70.67, 62.58, 50.16, 41.51,
36.80 (one carbon around 50, obscured by the solvent signal). HRMS:
calcd for C24H31N4O5 (M þ Hþ), 455.2294; found, 455.2306.
amount of CuSO4 5H2O and sodium ascorbate (20 mg, 0.1 mmol)
3
were added. After stirring at rt for 24 h, the reaction mixture was diluted
with EtOAc (10 mL). The organic phase was collected, dried over
MgSO4, and evaporated to give alcohol 26 in quantitative yield as an
orange solid, which was used in the next step without further purifica-
tion. 1H NMR: 7.41 (m, 10H, 1H) 6.61 (t, 1H, J = 2.1 Hz) 6.50 (d, 2H,
(1R,2S,3r,4R,5S,6s)-6-[(1-Adamantyl-1H-1,2,3-triazol-4-yl)methyl-
amino]cyclohexane-1,2,3,4,5-pentaol (1[7]). Obtained from 5[7] in
quantitative yield. 1H NMR (500 MHz, CD3OD): 8.10 (s, 1H) 4.62 (s,
2H) 4.15 (m, AB system, 2H) 3.61 (t, 2H, J = 9.5 Hz) 3.28 (m, 3H) 3.09
2076
dx.doi.org/10.1021/jm101204u |J. Med. Chem. 2011, 54, 2069–2079

Journal of Medicinal Chemistry
ARTICLE
(t, 1H, J = 10.6 Hz) 2.01 (m, 3H) 1.58 (m, 6H) 1.71 (m, 6H). 13C NMR
(100 MHz, CD3OD): 152.52 (Cq triazole, 2D assignment), 138.41 (CH
triazole, 2D assigment), 76.63, 75.14, 70.67, 62.52, 41.15, 37.62, 29.61.
HRMS: calcd for C20H33N4O5 (M þ Hþ), 409.2452; found, 409.2451.
(1R,2S,3r,4R,5S,6s)-6-[[1-(4-Cyclohexylbutyl)-1H-1,2,3-triazol-4-yl]methyl-
amino]cyclohexane-1,2,3,4,5-pentaol (1[8]). Obtained from 5[8] in
quantitative yield. 1H NMR (500 MHz, CD3OD): 8.26 (s, 1H) 4.63 (s,
2H) 4.47 (m, AB system, 2H) 3.64 (m, 2H) 3.25 (m,3H) 3.12 (t, 1H, J =
10.1 Hz) 1.91 (s, 2H) 1.68 (m,4H) 1.25 (m, 9H) 0.90 (m, 2H).
13C NMR (100 MHz, CD3OD): 139.57, 127.62, 76.49, 75.02, 70.60,
62.66, 51.81, 41.5, 38.74, 37.91, 34.40, 31.54, 27.69, 27.40, 24.71.
HRMS: calcd for C19H35N4O5 (M þ Hþ), 399.2607; found, 399.2599.
(1R,2S,3r,4R,5S,6s)-6-[[1-Undecyl-1H-1,2,3-triazol-4-yl ]methylamino]-
cyclohexane-1,2,3,4,5-pentaol (1[18]). Obtained from 5[18] in quan-
titative yield. 1H NMR (500 MHz, CD3OD): 8,18 (s, 1H) 4,59 (s, 2H)
4,45 (m, AB system, 2H) 3.61 (t, 2H, J = 9.3 Hz) 3.28 (m, 3H) 3.10 (t,
1H, J = 10.2 Hz) 1.93 (m, 2H) 1.28 (m, 16H) 0.90 (t, 3H, J = 6.8 Hz).
13C NMR (100 MHz, CD3OD): 152.43 (Cq triazole, 2D assignment),
139.44 (CH triazole, 2D assigment), 76.54, 75.09, 70.59, 62.65, 51.66,
41.56, 33.01, 31.33, 30.67, 30.65, 30.55, 30.41, 30.11, 27.46, 23.69, 14.41.
HRMS: calcd for C20H39N4O5 (M þ Hþ), 415.2920; found, 415.2931.
(1R,2S,3r,4R,5S,6s)-6-[[1-(10-Hydroxydecyl)-1H-1,2,3-triazol-4-yl]methyl-
amino]cyclohexane-1,2,3,4,5-pentaol(1[25]). Obtained from 5[25] in
performed as follows: substrate (100 μL, 5 mM 4-methylumbelliferyl β-
D-glucoside) in0.1Macetatebuffer(pH=4.0) wasaddedtoeachwellupto
a total volume of 260 μL; incubation was for 1 h at 37 °C. Enzyme reaction
was stopped with 2 mL of 100 mM glycine-NaOH buffer (pH =
10.6) and fluorescence measured at 355 nm (excitation) and 460 nm
(emission). For each experiment, untreated (no compound added) and
treated cells were plated in quadruplicate. Data are reported in Table 2.
Thermal Stabilization Assay. Following a modification of a
reported method,11 GCase aliquots (48 μL, 2 mg/mL) were incubated at
pH7.4at48°Cwith0(2μLH2O, control) or 2 μL test compound to reach
a final concentration of 10, 25, 50, 100, or 150 μM. Subsequently, 150 μL of
0.1 M acetate-phosphate buffer (pH 5.0) and 100 μL of substrate (4 mM
4-methyllumbelliferyl β-D-glucoside) in McIlvaine buffer (pH 5.2) were
added at different times (20, 40, 60 min) and incubated for 10 min at 37 °C.
Then, 300 μL of glycine buffer (200 mM, pH 10.6) were added and
liberated 4-methylumbelliferone was measured. Enzyme activity was re-
ported relative to that of the enzyme at 37 °C. Data are reported in Table 2,
Figure 3, and Supporting Information.
’ ASSOCIATED CONTENT
S
Supporting Information. Experimental protocols for the
b
parallel synthesis and characterization of azides 1-25, UPLC-
HRMS of adducts 1[1-25], evaluation of click chemistry
reagents in GCase assay, full SR data for compounds shown in
Figure 3, GCS activity and cytotoxicity data, MOE filter defini-
tions, molecular descriptors, diversity coverage of the virtual
library, pIC50 values derived from Table 1, and predicted clogP
values, Best poses for aminocyclitols 1[4], 1[7], 1[8], 1[18], and
1[25] against GCase structures 2V3E and 2NSX and NMR
spectra of the synthesized compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
1
quantitative yield. H NMR (500 MHz, CD3OD): 8.17 (s,1 H) 4.60
(s, 2H) 4.46 (m, AB system, 2H) 3.60 (t, 2H) 3.55 (t, 2H, J = 6.4 Hz)
3.27 (m, 3H) 3.10 (t, 1H, J = 10.4 Hz) 1.94 (m, 2H) 1.53 (m, 2H) 1.33
(m, 12H). 13C NMR (100 MHz, CD3OD): 153.25 (Cq triazole, 2D
assignment), 138.23 (CH triazole, 2D assigment), 76.57, 75.12, 70.61,
62.95, 62.64, 61.52, 51.62, 41.53, 33.61, 31.31, 30.59, 30.51, 30.08, 27.44,
26.90. HRMS: calcd for C19H37N4O6 (M þ Hþ), 417.2713; found,
417.2718.
Cells and Cultures. Fibroblasts were from Eucellbank and grown
in Eagle’s minimal essential medium (MEM) with 15% inactivated fetal
calf serum at 37 °C in a humidified, 5% CO2 incubator. Cells used were
between the 14th and the 30th passage.
’ AUTHOR INFORMATION
Enzyme Assays. 4-Methylumbelliferyl-β-D-glucopyranoside was
obtained from Sigma. GlucCerase (Imiglucerase) was kindly provided
by Genzyme Corporation (USA). HPLC measurements were per-
formed using a Waters 2690 Alliance system coupled to a Waters
2475 fluorescence detector (Milford, MA). Empower Software
(Waters Corporation) was utilized for data acquisition and processing.
GCase Activity. In vitro activity was determined with 4 mM 4-methy-
lumbelliferyl-β-D-glucopyranoside in McIlvaine buffer (pH = 5.2 or 7.4).
Enzyme solutions (25 μL from a stock solution containing 0.1 mg protein/
mL) in the presence of 0.2% (w/v) sodium taurocholate and 0.1% (v/v)
Triton X-100 in McIlvaine buffer (pH = 5.2 or 7.4) were incubated at 37 °C
without (control) or with inhibitor (variable concentrations) during 30 min
and, after addition of corresponding substrate solution (60 μL), incubations
were maintained at 37 °C for 10 min. Enzymatic reactions were stopped by
the addition of 150 μL of 100 mM glycine/NaOH buffer (pH 10.6). The
amount of 4-methylumbelliferone formed was determined with a fluorom-
eter at 355 nm (excitation) and 460 nm (emission).
Corresponding Author
*Fax: int þ34-932045904. E-mail: adelgado@cid.csic.es.
’ ACKNOWLEDGMENT
Partial financial support from the Ministerio de Ciencia e
Innovaciꢁon, Spain (project CTQ2008-01426/BQU) and Gen-
eralitat de Catalunya (Grant 2009SGR-1072) is acknowledged.
Lucía Díaz is grateful to CSIC for predoctoral research training
support within the JAE-Predoc program. The authors thank Dr.
Meritxell Egido-Gabas for technical assistance, Eva Dalmau for
HRMS analysis, and Genzyme Corporation for a generous
supply of Imiglucerase (Cerezyme). We also acknowledge the
Centre de Supercomputaciꢁo de Catalunya (CESCA) for allowing
the use of their software and hardware resources.
IC50 values were determined by plotting percent activity versus log
[I], using at least five different inhibitor concentrations. Type of
inhibition and Ki values for more active inhibitors were determined by
Lineweaver-Burk or Dixon plots of assays performed with different
concentrations of inhibitor and substrate. IC50 data from crude click
chemistry mixtures are reported in Table 1. Data from individually
synthesized compounds are reported in Table 2.
GCase Inhibition in Intact Fibroblasts. Fibroblasts were plated
into 24-well assay plates with DMEM containing aminocyclitol (50 μM,
final concentration) dissolved in DMSO and incubated as described
above for 24 h. Then, medium supplemented with aminocyclitols was
removed and cells were washed with PBS. The enzymatic assay was
’ ABBREVIATIONS USED
clogP, calculated logarithm of the partition coefficient; CNS,
central nervous system; CuAAC, copper-catalyzed alkyne azide
cycloaddition; GCase, β-glucocerebrosidase; GCS, glucosyl cer-
amide synthase.; CerNBD, 6-[(N-(7-nitrobenz-2-oxa-1,3-diazol-
4-yl)amino]hexanoylsphingosine
’ REFERENCES
(1) Grabowski, G. A. Phenotype, diagnosis, and treatment of
Gaucher’s disease. Lancet 2008, 372, 1263–1271.
2077
dx.doi.org/10.1021/jm101204u |J. Med. Chem. 2011, 54, 2069–2079

Journal of Medicinal Chemistry
ARTICLE
(2) Brady, R. O. Gaucher’s disease: past, present and future. Baillieres
Clin. Haematol. 1997, 10, 621–634.
Pharmacological Chaperones for Gaucher Disease. J. Med. Chem. 2010,
53, 5248–5255.
(3) Ron, I.; Horowitz, M. ER retention and degradation as the
molecular basis underlying Gaucher disease heterogeneity. Hum. Mol.
Genet. 2005, 14, 2387–2398.
(4) Schmitz, M.; Alfalah, M.; Aerts, J. M.; Naim, H. Y.; Zimmer, K. P.
Impaired trafficking of mutants of lysosomal glucocereb
rosidase in Gaucher’s disease. Int. J. Biochem. Cell Biol. 2005, 37,
2310–2320.
(5) Platt, F. M.; Lachmann, R. H. Treating lysosomal storage
disorders: current practice and future prospects. Biochim. Biophys. Acta,
Mol. Cell Res. 2009, 1793, 737–745.
(6) Morello, J.-P.; Bouvier, M.; Petaja-Repo, U. E.; Bichet, D. G.
Pharmacological chaperones: a new twist on receptor folding. Trends
Pharmacol. Sci. 2000, 21, 466–469.
(7) Fan, J.-Q. A contradictory treatment for lysosomal storage
disorders: inhibitors enhance mutant enzyme activity. Trends Pharmacol.
Sci. 2003, 24, 355–360.
(8) Butters, T. D. Pharmacotherapeutic strategies using small mo-
lecules for the treatment of glycolipid lysosomal storage disorders.
Expert Opin. Pharmacother. 2007, 8, 427–435.
(22) Some of the resulting fragments are also present in a variety of
iminocyclitol-derived modulators of glucosylceramide metabolism. Ada-
mantyl:(a) Wennekes, T.; van den Berg, R.; Boltje, T. J.; Donker-
Koopman, W. E.; Kuijper, B.; van der Marel, G. A.; Strijland, A.;
Verhagen, C. P.; Aerts, J.; Overkleeft, H. S. Synthesis and Evaluation
of Lipophilic Aza-C-glycosides as Inhibitors of Glucosylceramide Me-
tabolism. Eur. J. Org. Chem. 2010, 1258–1283. (b) Wennekes, T.; Meijer,
A. J.; Groen, A. K.; Boot, R. G.; Groener, J. E.; van Eijk, M.; Ottenhoff, R.;
Bijl, N.; Ghauharali, K.; Song, H.; O’Shea, T. J.; Liu, H. L.; Yew, N.;
Copeland, D.; van den Berg, R. J.; van der Marel, G. A.; Overkleeft, H. S.;
Aerts, J. M. Dual-Action Lipophilic Iminosugar Improves Glycemic
Control in Obese Rodents by Reduction of Visceral Glycosphingolipids
and Buffering of Carbohydrate Assimilation. J. Med. Chem. 2010,
53, 689–698. (c) Risseeuw, M. D. P.; van den Berg, R.; Donker-
Koopman, W. E.; van der Marel, G. A.; Aerts, J.; Overhand, M.;
Overkleeft, H. S. Synthesis and evaluation of D-gluco-pyranocyclopropyl
amines as potential glucosidase inhibitors. Bioorg. Med. Chem. Lett. 2009,
19, 6600–6603. (d) Wennekes, T.; van den Berg, R. J. B. H. N.; Bonger,
K. M.; Donker-Koopman, W. E.; Ghisaidoobe, A.; van der Marel, G. A.;
Strijland, A.; Aerts, J. M. F. G.; Overkleeft, H. S. Synthesis and evaluation
of dimeric lipophilic iminosugars as inhibitors of glucosylceramide
metabolism. Tetrahedron: Asymmetry 2009, 20, 836–846. (e) Bijl, N.;
Scheij, S.; Houten, S.; Boot, R. G.; Groen, A. K.; Aerts, J. M. The
glucosylceramide synthase inhibitor N-(5-adamantane-1-yl-methoxy-
pentyl)-deoxynojirimycin induces sterol regulatory element-binding
protein-regulated gene expression and cholesterol synthesis in HepG2
cells. J. Pharmacol. Exp. Ther. 2008, 326, 849–855. (f) Wennekes, T.;
Berg, R. J.; Donker, W.; Marel, G. A.; Strijland, A.; Aerts, J. M.;
Overkleeft, H. S. Development of Adamantan-1-yl-methoxy-Functiona-
lized 1-Deoxynojirimycin Derivatives as Selective Inhibitors of Gluco-
sylceramide Metabolism in Man. J. Org. Chem. 2007, 72, 1088–1097. (g)
Boot, R. G.; Verhoek, M.; Donker-Koopman, W.; Strijland, A.; van
Marle, J.; Overkleeft, H. S.; Wennekes, T.; Aerts, J. M. Identification of
the non-lysosomal glucosylceramidase as beta-glucosidase 2. J. Biol.
Chem. 2007, 282, 1305–1312. (h) Overkleeft, H. S.; Renkema, G. H.;
Neele, J.; Vianello, P.; Hung, I. O.; Strijland, A.; van der Burg, A. M.;
Koomen, G. J.; Pandit, U. K.; Aerts, J. M. Generation of specific
deoxynojirimycin-type inhibitors of the non-lysosomal glucosylcerami-
dase. J. Biol. Chem. 1998, 273, 26522–26527. Cyclohexylalkyl or phenyl-
alkyl: (i) Alonzi, D. S.; Dwek, R. A.; Butters, T. D. Improved cellular
inhibitors for glycoprotein processing alpha-glucosidases: biological
characterisation of alkyl- and arylalkyl-N-substituted deoxynojirimycins.
Tetrahedron: Asymmetry 2009, 20, 897–901. (j) Rawlings, A. J.; Lomas,
H.; Pilling, A. W.; Lee, M. J.; Alonzi, D. S.; Rountree, J. S.; Jenkinson,
S. F.; Fleet, G. W. J.; Dwek, R. A.; Jones, J. H.; Butters, T. D. Synthesis
and Biological Characterisation of Novel N-Alkyl-Deoxynojirimycin
alpha-Glucosidase Inhibitors. ChemBioChem 2009, 10, 1101–1105. N-
Alkoxyalkyl: (k) Mellor, H. R.; Platt, F. M.; Dwek, R. A.; Butters, T. D.
Membrane disruption and cytotoxicity of hydrophobic N-alkylated
imino sugars is independent of the inhibition of protein and lipid
glycosylation. Biochem. J. 2003, 374, 307–314. (l) Butters, T. D.; van
den Broek, L. A. G. M.; Fleet, G. W. J.; Krulle, T. M.; Wormald, M. R.;
Dwek, R. A.; Platt, F. M. Molecular requirements of imino sugars for the
selective control of N-linked glycosylation and glycosphingolipid bio-
synthesis. Tetrahedron: Asymmetry 2000, 11, 113–124. N-alkyl: (m)
Suzuki, Y.; Ogawa, S.; Sakakibara, Y. Chaperone therapy for neurono-
pathic lysosomal diseases: competitive inhibitors as chemical chaper-
ones for enhancement of mutant enzyme activities. Perspect. Med. Chem.
2009, 3, 7–19. (n) Delgado, A.; Casas, J.; Llebaria, A.; Abad, J. L.;
Fabrias, G. Chemical Tools to Investigate Sphingolipid Metabolism and
Functions. ChemMedChem 2007, 2, 580–606.
(9) Ringe, D.; Petsko, G. A. What are pharmacological chaperones
and why are they interesting? J. Biol. 2009, 8, 80.
(10) Barton, N. W.; Brady, R. O.; Dambrosia, J. M.; Di Bisceglie,
A. M.; Doppelt, S. H.; Hill, S. C.; Mankin, H. J.; Murray, G. J.; Parker,
R. I.; Argoff, C. E.; et al. Replacement therapy for inherited enzyme
deficiency—macrophage-targeted glucocerebrosidase for Gaucher’s dis-
ease. N. Engl. J. Med. 1991, 324, 1464–1470.
(11) Sawkar, A. R.; Cheng, W. C.; Beutler, E.; Wong, C. H.; Balch,
W. E.; Kelly, J. W. Chemical chaperones increase the cellular activity of
N370S beta-glucosidase: a therapeutic strategy for Gaucher disease.
Proc. Natl. Acad. Sci. U.S.A 2002, 99, 15428–15433.
(12) Egido-Gabas, M.; Serrano, P.; Casas, J.; Llebaria, A.; Delgado,
A. New aminocyclitols as modulators of glucosylceramide metabolism.
Org. Biomol. Chem. 2005, 3, 1195–1201.
(13) Egido-Gabas, M.; Canals, D.; Casas, J.; Llebaria, A.; Delgado, A.
Aminocyclitols as Pharmacological Chaperones for Glucocerebrosidase,
a Defective Enzyme in Gaucher Disease. ChemMedChem 2007,
2, 992–994.
(14) Sanchez-Olle, G.; Duque, J.; Egido-Gabas, M.; Casas, J.; Lluch,
M.; Chabas, A.; Grinberg, D.; Vilageliu, L. Promising results of the
chaperone effect caused by iminosugars and aminocyclitol derivatives on
mutant glucocerebrosidases causing Gaucher disease. Blood Cells, Mol.,
Dis. 2009, 42, 159–166.
(15) Serrano, P.; Casas, J.; Zucco, M.; Emeric, G.; Egido-Gabas, M.;
Llebaria, A.; Delgado, A. Combinatorial Approach to N-Substituted
Aminocyclitol Libraries by Solution-Phase Parallel Synthesis and Pre-
liminary Evaluation as Glucocerebrosidase Inhibitors. J. Comb. Chem
2007, 9, 43–52.
(16) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B. A
stepwise Huisgen cycloaddition process: copper(I)-catalyzed regiose-
lective “ligation” of azides and terminal alkynes. Angew. Chem., Int. Ed.
2002, 41, 2596–2599.
ꢁ
(17) Gil, M. V.; Arꢁevalo, M. J.; Lꢁopez, O. Click Chemistry—What’s
in a Name? Triazole Synthesis and Beyond. Synthesis 2007, 1589–1620.
(18) Brik, A.; Wu, C. Y.; Wong, C. H. Microtiter plate based
chemistry and in situ screening: a useful approach for rapid inhibitor
discovery. Org. Biomol. Chem. 2006, 4, 1446–1457.
(19) Wu, C. Y.; Chang, C. F.; Chen, J. S.; Wong, C. H.; Lin, C. H.
Rapid diversity-oriented synthesis in microtiter plates for in situ screen-
ing: discovery of potent and selective alpha-fucosidase inhibitors. Angew.
Chem., Int. Ed. 2003, 42, 4661–4664.
(20) Liang, F. S.; Brik, A.; Lin, Y. C.; Elder, J. H.; Wong, C. H.
Epoxide opening in water and screening in situ for rapid discovery of
enzyme inhibitors in microtiter plates. Bioorg. Med. Chem. 2006,
14, 1058–1062.
(23) Serrano, P.; Llebaria, A.; Delgado, A. Regio- and stereoselective
synthesis of aminoinositols and 1,2-diaminoinositols from conduritol B
epoxide. J. Org. Chem. 2005, 70, 7829–7840.
(21) Diaz, L.; Bujons, J.; Casas, J.; Llebaria, A.; Delgado, A. Click
(24) Exploratory attempts to carry out the nucleophilic opening of
Chemistry Approach to New N-Substituted Aminocyclitols as Potential
fully deprotected conduritol B epoxide with with the highly nucleophilic
2078
dx.doi.org/10.1021/jm101204u |J. Med. Chem. 2011, 54, 2069–2079

Journal of Medicinal Chemistry
ARTICLE
NaN3 were unsuccessful and unreacted material or complex reaction
mixtures were obtained in all cases.
(25) Meller, A.; Hirninger, F. J.; Noltemeyer, M.; Maringgele, W.
Umsetzungen von Metall- und Metalloidverbindungen mit mehrfunk-
tionellen Molek€ulen, XXXIII. Zur Reaktion von Propargylaminen bzw.
N-Organyl-N-(tri-methylsilyl)propargylaminen mit Bor-Halogen- bzw.
Bor-Schwefel-Verbindungen. Chem. Ber 1981, 114, 2519–2535.
(26) Gonzalez-Bulnes, P.; Casas, J.; Delgado, A.; Llebaria, A. Prac-
tical synthesis of (-)-1-amino-1-deoxy-myo-inositol from achiral pre-
cursors. Carbohydr. Res. 2007, 342, 1947–1952.
(27) Sisa, M.; Trapero, A.; Llebaria, A.; Delgado, A. Small-Scale One-
Pot Reductive Alkylation of Unprotected Aminocyclitols with Sup-
ported Reagents. Synthesis 2008, 3167–3170.
(28) Compounds shown in Table 2 were not cytotoxic at concen-
trations up to 300 μM . For details of the MTT test, see the Experimental
Section.
(29) Compounds shown in Table 2 were tested as inhibitors of the
following commercial glycosidases: R-glucosidase (yeast), R-glucosi-
dase (rice), R-galactosidase (coffee beans), and β-galactosidase (bovine
liver). All of them showed IC50 values above 100 μM, with the exception
of 1[25], which behaved as a weak β-galactosidase (bovine liver)
inhibitor (IC50 = 43 μM).
(30) Halgren, T. A. MMFF VII. Characterization of MMFF94,
MMFF94s, and other widely available force fields for conformational
energies and for intermolecular-interaction energies and geometries. J.
Comput. Chem. 1999, 20, 730–748.
(31) Halgren, T. A.; MMFF, V. I. MMFF94s option for energy
minimization studies. J. Comput. Chem. 1999, 20, 720–729.
(32) Compain, P.; Martin, O. R.; Boucheron, C.; Godin, G.; Yu, L.;
Ikeda, K.; Asano, N. ChemBioChem 2006, 7, 1356–1359.
2079
dx.doi.org/10.1021/jm101204u |J. Med. Chem. 2011, 54, 2069–2079