Click chemistry approach to new N-substituted aminocyclitols as potential pharmacological chaperones for gaucher disease

Article abstract of DOI:10.1021/jm100198t

New N-alkylaminocyclitols bearing a 1,2,3-triazole system at different positions of the alkyl chain have been prepared as potential GCase pharmacological chaperones using click chemistry approaches. Among them, compounds 1d and 1e, with the shorter spacer

Full text of DOI:10.1021/jm100198t

5248 J. Med. Chem. 2010, 53, 5248–5255
DOI: 10.1021/jm100198t
Click Chemistry Approach to New N-Substituted Aminocyclitols as Potential Pharmacological
Chaperones for Gaucher Disease†
Lucıa Dıaz,‡ Jordi Bujons, Josefina Casas,§ Amadeu Llebaria,§ and Antonio Delgado*,‡,§
‡
ꢀ
Facultat de Farmacia, Unitat de Quımica Farmaceutica (Unitat Associada al CSIC), Universitat de Barcelona, Avda. Joan XXIII, s/n,
08028 Barcelona, Spain, Research Unit on Bioactive Molecules (RUBAM), Departament de Quımica Biomedica, Institut de Quımica Avanc-ada de
´
ꢀ
§
´
ꢀ
´
´
Catalunya (IQAC-CSIC, Spanish National Research Council), Jordi Girona 18-26, 08034 Barcelona, Spain, and Department de Quımica
Biologica i Modelitzacio Molecular, Institut de Quımica Avanc-ada de Catalunya (IQAC-CSIC, Spanish National Research Council), Jordi Girona
ꢀ
ꢁ
´
18-26, 08034 Barcelona, Spain
Received February 15, 2010
New N-alkylaminocyclitols bearing a 1,2,3-triazole system at different positions of the alkyl chain have been
prepared as potential GCase pharmacological chaperones using click chemistry approaches. Among them,
compounds 1d and 1e, with the shorter spacer (n = 1) between the alkyltriazolyl system and the amino-
cyclitol core, were the most active ones as GCase inhibitors, revealing a determinant effect of the location of
the triazole ring on the activity. Furthermore, SAR data and computational docking models indicate a
correlation between lipophilicity and enzyme inhibition and suggest “extended” and “bent” potential
binding modes for the compounds. In the “bent” mode, the most active compounds could establish a
hydrogen-bond interaction between the triazole moiety and enzyme residue Q284. Such an interaction
would be precluded in compounds with a longer spacer between the triazole and the aminocyclitol core.
Introduction
Gaucher disease is one of the most prevalent lysosomal
storage disorders characterized by the accumulation of the
sphingolipid glucosylceramide in the lysosomes. The disease
is caused by the deficient activity elicited by several mutated
forms of the enzyme glucocerebrosidase (GCasea), the
β-glucosidase that hydrolyzes glucosylceramide into glucose
and ceramide.1 Cellular levels of the mutated, misfolded
enzyme are abnormally low because of its premature degrada-
tion by specific cytosolic endoproteases in the endoplasmic
reticulum. Several therapeutic strategies for Gaucher disease
have been developed over the past years.2 Among them, the
use of pharmacological chaperones, competitive inhibitors of
the target enzyme thatassistthe properfoldingofthe defective
proteinatsubinhibitory concentrations,3-5 is an active field of
research.6 In this context, several iminosugars and amino-
cyclitols have been reported in the literature (Figure 1).
Figure 1. Aminocyclitols reported in this study (1-3), together
with representative iminosugars and aminocyclitols reported as
To date, a number of GCase structures in its native state7
GCase inhibitors.
with different degreesof glycosylation,8-10 under different pH
conditions,11,12 or as a complex with different inhibitors13-16
and closed conformations (see below). Differences on the sur-
face topology of GCase close to the entrance to the active site
have been observed in structures of the enzyme complexed with
different inhibitors. Thus, while in N-nonyldeoxynojirimycin
(NNDNJ) and N-butyldeoxynojirimycin (NBDNJ)-GCase
complexes the N-alkyl chains are stabilized by accommodation
into the hydrophobic valley between L3 and L4, an additional
hydrophobic groove between L1 and L3 is observed in the iso-
fagomine (IFG)-GCase complex structure.14,15 Interestingly,
these two valleys between L3 and L4 and between L1 and L3
have been modeled as possible binding locations for the two
alkyl chains of the ceramide moiety in the natural substrate.15
Over the past years, we have been working actively on the
development of new aminocyclitols with potential applicability
have been reported. All these structures show a very similar
3D-arrangement for the protein with four loops (L1-L4)
located at the entrance to the catalytic site that control open
†Dedicated to Prof. Pelayo Camps on the occasion of his 65th
anniversary.
*To whom correspondence should be addressed. Phone: þ34-
934006108. Fax: þ34-932045904. E-mail: adelgado@cid.csic.es.
a Abbreviations: A-C10, N-decylaminocyclitol; CC50, cytotoxic con-
centration for 50% of the cell population under study; clogP, calculated
logarithm of the partition coeficient; GCase, β-glucocerebrosidase; IFG,
isofagomine; NBDNJ, N-butyldeoxynojirimycin; NNDNJ, N-nonyl-
deoxynojirimycin ; NOV, N-octylvalienamine; SAR, structure-activity
relationship.
r
pubs.acs.org/jmc
Published on Web 06/17/2010
2010 American Chemical Society

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 14 5249
Scheme 1a
Table 1. Aminocyclitols Arising from Click Chemistry Reaction
between ω-Alkynyl Alkylaminocyclitols and Azidesa
alkyne azide (RN3)
R (in RN3)
n-butyl
n-octyl
click adduct aminocyclitol
5
5
5
5
5
5
5
5
7
7
7
9
9
9
23
25
26
27
29
30
31
32
25
26
28
24
25
27
10a
10b
10c
10d
10e
10f
1a
1b
1c
1d
1e
1f
n-nonyl
n-decyl
n-dodecyl
n-tetradecyl
6-(propoxy)hexyl
2-phenylethyl
n-octyl
10g
10h
11a
11b
11c
12a
12b
12c
1g
1h
2a
2b
2c
3a
3b
3c
a Reagents and conditions: (a) amine (propargylamine, 20, or 22),
LiClO4, CH3CN; (b) KF, DMSO.
n-nonyl
n-undecyl
n-hexyl
n-octyl
afforded the expected fully benzylated triazolylalkylamino-
cyclitols 10-12 (Scheme 3). Massive O-benzyl deprotection
was carried out either by hydrogenolysis or by reaction with
BCl3, following reported protocols.17,27
n-decyl
a See Scheme 3.
as chemical chaperones of the enzyme GCase.17,18 In our
previous work, wefound that aminocyclitol A-C10 (Figure 1),
which stabilized GCase under thermal denaturation condi-
tions19 and inhibited GCase in wild type fibroblasts, behaved
as a pharmacological chaperone in Gaucher disease patient
fibroblasts bearing the L444P/G202R and L444P;E326K/
G202R genotypes.20 However, exploration of the chemical
diversity around the amino group has met with limited success
so far.21 Thus, the most promising results arose from the
regio- and stereoselective opening22,23 of conduritol B epoxide
with a series of aliphatic amines of different chain lengths.19 In
order to explore the role of the aminoalkyl side chain on the
enzyme stabilization under thermal denaturation conditions,24
a fast and convenient methodology aimed at the parallel
synthesis of small to medium sized libraries of N-substituted
aminocyclitol derivatives would be desirable. In this context,
click-chemistry approaches based on the Cu-catalyzed 1,3-
dipolar Huisgen cycloaddition of alkynes and azides are being
widely used in drug design.25,26 However, the electronic pro-
perties of the resulting 1,2,3-triazole system make possible the
operation of additional interactions of uncertain effects at the
active site level. In order to unravel such effects, we have
synthesized and tested a small collection of N-alkylamino-
cyclitols arising from the formal placement of the triazole ring
along the N-alkyl side chain. From a structural standpoint, the
aminocyclitols presented in this study can be grouped into
three different subsets (1-3) as a function of the spacer length
(n = 1-3) between the aminocyclitol core and the triazole
system (Figure 1).
Results and Discussion
A pharmacological chaperone binds to the active site of the
target enzyme at the neutral pH of the endoplasmic reticulum
(ER) to assist in protein folding and also to enhance the enzyme
transport to lysosomes. However, it partly dissociates at the low
pH of the lysosomal environment while still stabilizing the
enzyme at that acidic pH. Consequently, compounds were
evaluated as inhibitors of recombinant GCase (imiglucerase
(Cerezyme)) at both neutral and acidic pH. Compounds 1,
bearing the shorter spacer (n= 1) (see Tables 1 and 2) gave IC50
values lower than those of compounds 2 and 3, with a longer
spacer (n = 2 and 3, respectively), resulting from the formal
placement of the triazole ring along the aliphatic chain
(compare the couples 1b/3a and 1c/2a and the triplets 1d/2b/
3b and 1e/2c/3c). Inhibition constants (Ki) for the most active
compounds in each series showed, in all cases, a competitive
inhibition pattern, as illustrated in Figure 2 for 1d (Table 2).28
It is worth mentioning that the low Ki values found for
compounds 1d-1f were in the 60-90 nM range. Interestingly,
none of the above aminocyclitols showed significant activity
on glucosyl ceramide synthase (GCS) from cell homogenates
at 250 μM, thus indicating a selectivity toward the hydrolytic
enzyme.29 Moreover, the above aminocyclitols were inactive
at 0.1 mM against a panel of commercially available R- and
β-glycosidases,30 reinforcing the selectivity against GCase.
Regarding neutral pH, IC50 values for compounds 1 were
similar to or slightly lower than those at acidic pH, with the
exception of 1e and 1h (Table 2). Interestingly, compounds 2
and 3 showed lower IC50 values atneutral pH, an indication of
the ability of the inhibitors to interact with the enzyme at the
cellular pH of both ER and lysosomes.
Enzyme stabilization under thermal denaturation con-
ditions, expressed as the stabilization ratio (see Table 2),
was used as an indication of the potential of the target
compounds to behave as pharmacological chaperones.24
An IC50 threshold of 15 μM at pH 7.4 was set as selection
criteria for the thermal stabilization assay. Recovery of
GCase activity was measured at 48 °C in the presence and
in the absence of increasing concentrations (from 0.50 to
150 μM or from 50 to 150 μM, as appropriate) of selected
aminocyclitols at different incubation times (Figure 3).
For comparative purposes, NNDNJ and our previously
reported aminocyclitol A-C1019 (Figure 1) were also as-
sayed under the same experimental conditions.
Synthesis
Aminocyclitols 1-3 were obtained by the Cu-catalyzed 1,3-
dipolar cycloaddition (Huisgen reation) of ω-alkynylamino-
cyclitols 5 (n = 1), 7 (n = 2), and 9 (n = 3) with a set of azides
23-32, followed by massive benzylremoval (see Scheme3 and
Table 1).
Aminocyclitols 5, 7, and 9 were obtained, in turn, by the
regio- and diastereoselective opening22 of epoxide 4 with the
commercially available propargylamine or amines 20 and 22,
followed by TMS removal (Schemes 1 and 2). The use of the
TMS protecting group in amines 20 and 22 increased their
boiling point, thus favoring their isolation, purification, and
subsequent manipulation.
Copper-catalyzed Huisgen cycloaddition reaction of scaf-
folds 5, 7, and 9 with the aliphatic azides shown in Table 1

´
Dıaz et al.
5250 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 14
Scheme 2. Synthesis of Aminesa
a Reagents and conditions: (a) TMSCl, BuLi, THF, -78 °C; (b) MsCl, Et3N, THF, 0 °C; (c) NaN3, DMF; (d) LiAlH4, THF.
Scheme 3a
a Reagents and conditions: (a) R-N3 (23-32; see Table 1), Cu2SO4, sodium ascorbate, H2O/THF (1:1); (b) BCl3, CH2Cl2 (-78 °C); (c) H2, Pd/C.
Table 2. Inhibitory Activity of Aminocyclitols 1a-h, 2a-c, and 3a-c
lization ratio and the concentration used in the experiment
(Figure 3B and Supporting Information). Interestingly, en-
zyme thermal stabilization correlates well with Ki values in
compounds 1a-h, while most of the aminocyclitols tested in
this study were not cytotoxic in wild type human fibroblasts at
concentrations up to 300 μM.31
Finally, some of the aminocyclitols also behaved as GCase
inhibitors in wild type human fibroblasts after 24 h of
incubation at 50 μM (Table 2). A good correlation between
Ki and GCase inhibition was observed for compounds 1b-d
in this assay. Since compounds 1e and 1f could not be tested
at the above concentration because of their toxicity,31 they
were tested at 20 and 10 μM, together with 1d for the sake of
comparison (Table 2). Interestingly, compound 1e was an
even better inhibitor than 1d at 20 μM and almost equipotent
to 1f at 10 μM. However, the high efficacy of 1d,31 together
with its acceptable biological profile, low toxicity, and high
enzyme affinity and enzyme thermal stabilization properties,
makes this compound a promising candidate for further
studies.
Activity data for the compounds described in this work
reflect a clear dependence between GCase inhibition and
clogP (see Supporting Information). Thus, in each series,
activity increases with hydrophobicity, as shown in Figure 4.
An exception to this trend is found for compound 1f, with the
longest N-triazolylalkyl chain (C14) and the highest clogP
value, which shows the same or lower activity than its less
hydrophobic and shorter homologues 1d (C10) and 1e (C12). It
is noteworthy that compounds 1e and 1f are toxic at high
concentrations (>50 μM) and that this has also been observed
for other analogues containing highly hydrophobic moieties,19
which suggests a lipophilicity-related toxic effect. Further-
more, since the activity of aminocyclitols with N-alkyltriazol-
ylalkyl chains of identical combined length is higher for
compounds with the shortest spacer (n = 1, Figure 1) between
the aminocyclitol and the triazole moieties (see above), a
restriction to the length of the N-triazolylalkyl chain R and
to the position of the triazole ring for optimum binding to the
enzyme can be envisaged. Therefore, docking studies were
against Imiglucerasea
IC50 (μM)
stabilization % inhibition GCase
ratioc,d
compd pH 7.4 pH 5.2 Ki (μM)b
in human wt fibroblastse
1a
1b
1c
1d
1e
1f
nd
178
nd
1.8
nd
<5
20
1.1 1.0
0.20 0.20
0.09 0.10
0.06 0.03
0.10 0.09
6.2 7.4
25.0 10.8
55.9 167
11.0 44.7
4.9 25.4
44.6 257
18.2 35.4
1.2 3.5
1.0
13.0
21.0
28.7
13.2
1.1
nd
0.33
0.09
0.06
0.08
2.4
50
90/71f/58g
h/90 f /87g
h, i/93g
<5
1g
1h
2a
2b
2c
3a
3b
3c
19.1
nd
<5
<5
nd
nd
7.6
11.6
nd
<5
50
20.2
nd
<5
40
19.9
7.6
nd
1.4
9.2
4.9
20
nd
NNDNJ 0.30 1.3, 0.66 j 0.30 j
A-C10
1.8
0.30
nd
a See also Table 1 and Scheme 3 for structures). nd: not determined.
b Competitive inhibitors (determined at pH 5.2). c Defined as the ratio of
relative enzymatic activities (inhibitor vs control) at a given inhibitor
concentration and incubation time. Tabulated values for 100 μM
inhibitor and 60 min of incubation time. d See also Supporting Informa-
tion. e Incubation for 24 h at 50 μM inhibitor and 5 mM substrate.
f Incubation for 24 h at 20 μM inhibitor and 5 mM substrate. g Incuba-
tion for 24 h at 10 μM inhibitor and 5 mM substrate. h Toxic at 50 μM.
i Toxic at 20 μM. j According to Compain et al. ChemBioChem 2006, 7,
1356-1359.
Among the compounds tested as GCase stabilizers, com-
pounds 1b, 1g, and 3c exhibited a similar, albeit almost
negligible, activity pattern at the different times and concen-
trations tested. On the other hand, compounds 1c, 1f, 2b, and
2c showed an enzyme stabilization profile comparable to or
even superior to that of NNDNJ, with stabilization ratios
between 7.6 and 13.2 at 100 μM after 1 h of incubation time at
48 °C(Table2andFigure 3A). Thisparameterwasevenhigher
for aminocyclitols 1d and 1e, with values ranging from 20 to
30 at the above concentrations and incubation times. Com-
pounds 1d and 1e were also assayed at lower concentrations
(0.5-25 μM), showing a good correlation between the stabi-

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 14 5251
Figure 2. Type of inhibition of glucocerebrosidase by 1d. Double
reciprocal plot of imiglucerase incubated at different concentrations
of substrate and compound 1d (square, 0 μM; circle, 0.1 μM;
triangle, 0.25 μM). Regression lines arise from data obtained in
two different experiments with triplicates.
Figure 4. GCase inhibitory activity, expressed as -log(IC50), vs clogP
for compounds of series 1-3.
placed inside the GCase active site and the triazole and alkyl
chain moieties adopt an “extended” conformation on the
exposed surface of the enzyme between loops 3 and 4. The
results obtained using the IFG-GCase structure (Figure 5D),
however, showed an alternative binding mode for most of the
compounds in which the cyclitol moiety still occupies the
active site of the enzyme, but the triazole and the alkyl chain
adopt a “bent” conformation to reach the narrow groove
between loops 1 and 3. Parts E and F of Figure 5 show the best
poses obtained for compound 1d docked against 2V3E and
2NSX, respectively, and illustrates the main interactions
between this compound and the protein in each particular
conformation (corresponding interaction diagrams in Sup-
porting Information). Similar interactions were observed for
the analogue compounds 1e and 1f.
Hence, the best pose obtained for 1d docked against struc-
ture 2V3E (Figure 5E) shows the hydroxyl groups of the
cyclitol moiety establishing hydrogen bonds with residues
D127, W179, W381, and N396. An additional strong interac-
tion between the charged nitrogen atom of 1d and the carboxy-
late group of E235 is observed. The charge on this nitrogen is
also stabilized by a π-cation interaction with Y313. Finally,
the “extended” alkyl chain is stabilized by less specific hydro-
phobic contacts with residues L240, L241, L286, L287, and
L314. On the other hand, the best pose obtained for the same
compound docked against structure 2NSX (Figure 5F) shows
similar hydrogen bonding and polar interaction patterns, but
the “bent” conformation of the alkyl moiety is now stabilized
by hydrophobic interactions with residues W312, Y313, L314,
and F316. Different poses obtained for compound 1d, and also
for the rest of the compounds studied, show slight variations in
their binding mode to the different structures of GCase
considered, which are most probably derived from the large
flexibility of the alkyl substituentsand from the high number of
polar residues in the active site of GCase, which allows for
alternative hydrogen bonding patterns.
Figure 3. (A) and (B) Relative enzymatic activity after thermal
denaturation (48 °C) for 20 min (white bar), 40 min (gray bar), and
60 (black bar) at the indicated inhibitor concentrations (μM)
compared to the corresponding assay at 37 °C. Data for control
(C) are obtained as above except that no inhibitor is present.
conducted to shed light on the way of binding of the amino-
cyclitols described in this work to the active center of GCase.
In order to explore all the potential binding modes of
compounds 1-3, the structures of the GCase-inhibitor com-
plexeswithNNDNJ (PDB code2V3E, Figure5A), IFG(PDB
code 2NSX, Figure 5B), and NBDNJ (PDB code 2V3D, Sup-
porting Information) were chosen as target structures for
docking. Parts C and D of Figure 5 show the best docked
conformations obtainedusingthe 2V3E and2NSXstructures,
respectively, as targets. The results obtained with the 2V3D
structure were similar to those shown in Figure 5C (see Sup-
porting Information). In the docked poses obtained using the
NNDNJ-GCase complex (Figure 5C), the cyclitol moiety is
Although obtaining a good correlation between docking
scores and experimentally determined binding affinities is
frequently not possible because of, inter alia, the limitations
of scoring functions and sampling algorithms,32,33 a reason-
able correlation was obtained for the compounds in series 1
(n = 1) when the scores from docking to the structure of the
“bent” IFG-GCase complex were considered (Figure 6).
Since no such correlation was observed from the docking runs

´
Dıaz et al.
5252 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 14
Figure 5. Crystal structures of (A) the NNDNJ-GCase complex (PDB 2V3E) and (B) the IFG-GCase complex (PDB 2NSX). Loops 1, 2, 3,
and 4 (loop numbering based on ref 14) are colored in yellow, light blue, red, and purple, respectively. Best docked poses obtained for
compounds 1-3 using structures 2V3E (C) and 2NSX (D) as targets. The orientation of the protein structures is similar to that shown in panels
A and B, respectively. L1, L2, L3, and L4 denote the position of loops 1-4. In panel D, a black arrow shows the accessible groove between loops
1 and 3. Best poses obtained for compound 1d docked against structures 2V3E (E) and 2NSX (F) as targets.
with the “extended” NNDNJ-GCase complex (see dock-
ing scores in the Supporting Information), this suggests the
“bent” mode of binding as the preferred one for compounds 1.
Compounds of series 2 and 3 also followed an approximate
linear dependence between both calculated and experimental
parameters, similar to that observed when hydrophobicity was
considered (Figure 4). However, these docking results do not
clarify the role of the triazole moiety in the binding process,
since no clear difference in the binding mode among the three
series (spacer n = 1-3, Figure 1) was observed and no clear
interaction between the triazole ring and the receptor was
detected in any case. In this context, the triazole group has
been considered as a good amide bond mimic, being able to act
as a hydrogen bond acceptor.34-36

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 14 5253
Conclusion
New N-alkylaminocyclitols bearing a 1,2,3-triazole system
at different positions of the alkyl chain have been prepared
using click chemistry protocols. The resulting compounds
have been tested as GCase inhibitors and also for their ability
to increase the enzyme stability against thermal denaturation,
as an in vitro assessment of their potential as pharmacological
chaperones. Among them, compounds 1d and 1e have emerged
as promising candidates for further development, in light of the
comparative results with NNDNJ and our previously reported
aminocyclitol A-C10.
In spite of the good correlation between lipophilicity
and IC50 values, the position of the triazole system along the
N-alkyl chain is crucial for enzyme inhibition and thermal
stabilization, since members with the shorter spacer between
the aminocyclitol core and the alkyltriazolyl system exhibited
the maximum potency. Modeling studies using three different
reported GCase-inhibitor complexes (PDB codes 2V3D,
2V3E, and 2NSX) showed two alternative modes of binding
with an apparently preferred “bent” mode (based on docking
scores) for the most active compounds of this study. Although
the docking results do not clarifythe role of the triazolemoiety
in this binding process, the accepted ability of triazole to act as
a hydrogen bond acceptor would allow the operation of a new
hydrogen bondbetween triazole N3 andamino acidQ284by a
slight rearrangement of the conformation of the side chain of
this residue found in the crystal structure. This additional
interaction would likely be precluded in compounds with a
longer spacer (series 2 and 3). Studies addressed at clarifying
the role of triazole in these systems are currently underway in
our group and will be published in due course.
Figure 6. Correlation of docking scores vs experimentally deter-
mined -log(IC50) values for compounds 1-3 docked against
IFG-GCase complex (structure 2NSX).
Experimental Section
Chemistry: General Protocols. Solvents were distilled prior to
use and dried by standard methods. FT-IR spectra are reported
in cm-1. 1H and 13CNMR spectra were obtained in CDCl3 solu-
tions at 300 MHz (for 1H) and 75 MHz (for 13C), respectively,
unless otherwise indicated. Chemical shifts are reported in δ
units, parts per million (ppm) relative to the singlet at 7.24 ppm
of CDCl3 for 1H and in ppm relative to the center line of a triplet
at 77.0 ppm of CDCl3 for 13C. ESI/HRMS spectra were
recorded on a Waters LCT Premier mass spectrometer. 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.
Figure 7. Best pose obtained for compound 1d when docked
into the active site of structure 2NSX (green) and pose obtained
after minimization allowing receptor flexibility (cyan), which
shows movement of the side chain of residue Q284 (yellow) and
of the triazole ring, leading to the formation of a new hydrogen
bond.
Minimization of some of the “bent” dockingposes obtained
for compound 1d, as well as those of other analogs from series
1, was carried out allowing for movement of the closest
residues of the protein. Interestingly, the results showed that
the Q284 side chain can rotate to form a new hydrogen bond
with the triazole N3 nitrogen atom (Figure 7). Compounds of
series 2 and 3, having a longer spacer between the triazole and
the cyclitol moieties, would require a larger movement of this
residue, and perhaps others, to establish such an interaction.
This might be energetically more expensive and would explain
their lower affinity. Clearly, additional modeling studies are
needed to verify such hypotheses and to understand how the
triazole ring can influence the binding strength of these
compounds.
General Method for Click Reactions: Synthesis of 10e as
Representative Example. A solution of the azide 29 (0.25 mmol)
(see Table 1) in a mixture H2O/THF (1:1) (2 mL) was added
to aminocyclitol 5 (0.20 mmol) in a vial. Then CuSO4 5H2O
3
(5.2 mg, 0.020 mmol) and a catalytic amount of sodium ascor-
bate were added. The mixture was stirred at room temperature
for 2 min, quenched with water (3 mL), and extracted with
EtOAc (3 ꢀ 5 mL). Usual workup afforded a green solid, which
was purified by flash chromatography on elution with ethyl
1
acetate. Compound 10e was obtained in 60% yield. H NMR
(500 MHz, CDCl3): 7.44 (m, 20, 1H), 4.92 (m, 8H), 4.30 (s, 2H),
4.11 (m, 2H), 3.56 (m, 5H), 2.69 (t, 1H), 1.88 (d, 2H, J = 1.2 Hz),
1.32 (m, 18H), 0.92 (t, 3H, J = 6.9 Hz). 13C NMR (400 MHz,
CDCl3): 146.88, 138.72-138.37, 128.59-127.50, 121.21, 84.76,
84.10, 83.078, 81.85, 73.03, 61.54, 50.39, 43.12, 31.97- 29.11,

´
Dıaz et al.
5254 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 14
22.76, 14.22. HRMS calculated for C49H65N4O5 (M þ Hþ):
789.4955. Found: 789.4935.
that location); (iii) up to 50 poses per ligand were kept for the
postdocking minimization. Glide XP scores45 were used to rank
the resulting docked poses. pKa calculations for compounds
1a-h, 2a-c, and 3a-c were performed with the pKa prediction
module46 included in Jaguar47with its default settings. The
coordinates of the complexes of human GCase with different
inhibitors bound in its active site, i.e., isofagomine (PDB
2NSX), NB-DNJ (PDB 2V3D), or NN-DNJ (PDB 2V3E), were
obtained from the Protein Data Bank48 at Brookhaven National
Laboratory. The structures of the proteins were prepared using
the Protein Preparation Wizard included in Maestro to remove
the solvent molecules and ligands, adding hydrogens, setting
protonation states, and minimizing the energy using the OPLS
force field. Compounds 1a-h, 2a-c, and 3a-c were built within
Maestro and preoptimized before docking using the LigPrep49
application included in the software. All ligands were docked
with the secondary amine group in the protonated state based on
the results from pKa calculations (see Supporting Information)
and the fact that preliminary docking runs always yielded poses
where this amino group was located in a highly polar environ-
ment, close to the carboxylate of D235 residue, which would
favor the protonated form of the amine. Interaction diagrams
were built with MOE50 for the best docked poses obtained for
some of the inhibitors.
Debenzylation with Boron Trichloride: Synthesis of 1e as
Representative Example. A solution of 10e (0.05 mmol) in
CH2Cl2 (2 mL) under nitrogen at -78 °C was treated with
1 M BCl3 solution in heptane (0.4 mL). After being stirred for 2 h
at -78 °C, the reaction mixture was allowed to warm to room
temperature and stirred for additional 12 h at room tempera-
ture. The mixture was next cooled to -78 °C, quenched with
methanol (1 mL), and evaporated under reduced pressure. The
resulting residue was taken up in a 1:1 MeOH/H2O mixture,
filtered through a small pad of charcoal, and evaporated to
afford 1e as the corresponding hydrochloride in 87% yield. 1H
NMR (500 MHz, CD3OD): 8.24 (s, 1H), 4.63 (s, 2H), 4.47 (t,
2H, J = 7.1 Hz), 3.64 (m, 2H), 3.30 (m, 3H), 3.13 (t, 1H, J =
10.7 Hz), 1.94 (dd, 2H, J = 6.9 Hz, J = 13.8 Hz), 1.26 (m, 18H),
0.91 (t, 3H, J = 6.9 Hz). 13C NMR (400 MHz, CD3OD): 77.46,
76.06, 71.72, 63.98, 59.18, 42.89, 33.79, 32.02, 31.46, 31.39,
31.29, 31.15, 30.87, 28.29, 24.43, 21.28, 19.17, 15.1. Triazole
13C signals were not observed in any of the click adducts because
of unsuitable relaxation times. However, their presence was
confirmed by a combination of gHSQC and gHMBC experi-
ments from a representative library member (see Supporting
Information for details). HRMS calculated for C21H41N4O5
(M þ Hþ): 429.3077. Found: 429.3014.
GCase Inhibition. In vitro activity was determined with 4 mM
4-methylumbelliferyl-β-D-glucopyranoside in McIlvaine buffer
(pH 5.2 or 7.4). Enzyme solutions (25 μL from a stock solution
containing 0.1 mg of 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 (1 mM) during 30 min, and after addition of
corresponding substrate solution (60 μL), incubations were
maintained at 37 °C for 10 min. Enzymatic reactions were stop-
ped 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 1420 VICTOR2 multilabel counter (Wallac)
fluorometer at 355 nm (excitation) and 460 nm (emission).
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 deter-
mined by Lineweaver-Burk or Dixon plots of assays performed
with different concentrations of inhibitor and substrate.
Acknowledgment. Partial financial support from the
ꢁ
“Ministerio de Ciencia e Innovacion”, Spain (Project CTQ-
2008-01426/BQU), and “Generalitat de Catalunya” (Grant
2009SGR-1072) is acknowledged. L.D. is grateful to CSIC for
predoctoral research training support within the JAE-Predoc
program. The authors thank Dr. Meritxell Egido-Gabas for
ꢁ
technical assistance, Nuria Guillem and Laura Planas for
experimental contributions, Eva Dalmau for HRMS analysis,
and Genzyme Corp. for a generous supply of imiglucerase
(Cerezyme). Finally, the authors acknowledge the “Centre de
ꢁ
Supercomputacio de Catalunya” (CESCA) for allowing the
use of its software and hardware resources.
Supporting Information Available: General experimental
protocols and compound characterization, cytotoxicity, inhibi-
tion of GlcCerase (wt fibroblasts), GCS (cell homogenates), and
commercial glycosidases, stabilization ratios for thermal dena-
turation studies, and predicted pKa and best Glide XP docking
scores obtained for compounds 1a-h, 2a-c, and 3a-c docked
against targets 2V3D, 2V3E, and 2NSX. This material is avail-
able free of charge via the Internet at http://pubs.acs.org.
Thermal Stabilization Assay. Following a modification of a
reported method,37 GlcCerase aliquots (48 μL, 2 mg/mL) were
incubated at pH 7.4 with 0 (control), 10, 25, 50, 100, or 150 μM
test compound at 48 °C. 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 and incubated for 10 min at
37 °C in the presence of 0.l % Triton X-100 and 0.2% tauro-
deoxycholic acid. Then an amount of 300 μL of glycine buffer
(200 mM, pH 10.6) was added, and liberated 4-methylumbelli-
ferone was measured. Enzyme activity was reported relative to
that of the enzyme at 37 °C.
References
(1) Brady, R. O. Gaucher’s disease: past, present and future. Baillieres
Clin. Haematol. 1997, 10, 621–634.
(2) Sawkar, A. R.; D’Haeze, W.; Kelly, J. W. Therapeutic strategies to
ameliorate lysosomal storage disorders;a focus on Gaucher dis-
ease. Cell. Mol. Life Sci. 2006, 63, 1179–1192.
(3) 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.
(4) Fan, J.-Q. A contradictory treatment for lysosomal storage dis-
orders: inhibitors enhance mutant enzyme activity. Trends Phar-
macol. Sci. 2003, 24, 355–360.
(5) Butters, T. D. Pharmacotherapeutic strategies using small mole-
cules for the treatment of glycolipid lysosomal storage disorders.
Expert Opin. Pharmacother. 2007, 8, 427–435.
(6) Grabowski, G. A. Treatment perspectives for the lysosomal sto-
rage diseases. Expert Opin. Emerging Drugs 2008, 13, 197–211.
(7) Dvir, H.; Harel, M.; McCarthy, A. A.; Toker, L.; Silman, I.;
Futerman, A. H.; Sussman, J. L. X-ray structure of human acid-
beta-glucosidase, the defective enzyme in Gaucher disease. EMBO
Rep. 2003, 4, 704–709.
Computational Methods. All molecular simulations were con-
€
€
ducted with the package Schrodinger Suite 2008 (Schrodinger
LLC), through its graphical interphase Maestro.38 The program
Macromodel39 with its default force field OPLS 2005, a modi-
fied version of the OPLS-AA force field,40 and GB/SA water
solvation conditions41 were used for all energetic calculations.
The program Glide42-44 was used for the docking calculations
using the default XP precision settings except for the following:
(i) given the high flexibility of most of the compounds, a setting
of 5 000 000 poses per ligand for the initial phase of docking and
a scoring window of 500 for keeping initial poses were chosen;
(ii) constraints were applied to maintain the cyclitol moiety of
the ligands inside the catalytic cavity of the CCase structures
(preliminary docking runs showed, as expected, that the best
scored poses for all compounds always had the cyclitol placed in
(8) Brumshtein, B.; Wormald, M. R.; Silman, I.; Futerman, A. H.;
Sussman, J. L. Structural comparison of differently glycosylated
forms of acid-beta-glucosidase, the defective enzyme in Gaucher

Article
disease. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2006, 62,
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 14 5255
(28) Inhibition constants (Ki) were only determined for compounds
with IC50 values below 40 μM at pH 5.0 (pH for optimal enzyme
activity).
(29) Only compounds 2b, 2c, and 3a showed GCS inhibition (A549
cells), albeit not higher than 30% at the indicated concentration
(250 μM). For a matter of comparison, a 70% inhibition of enzyme
activity was found for NBDNJ at 10 μM. For experimental details,
see the Supporting Information.
1458–1465.
(9) Shaaltiel, Y.; Bartfeld, D.; Hashmueli, S.; Baum, G.; Brill-Almon,
E.; Galili, G.;Dym, O.;Boldin-Adamsky, S. A.;Silman, I.;Sussman,
J. L.; Futerman, A. H.; Aviezer, D. Production of glucocerebro-
sidase with terminal mannose glycans for enzyme replacement
therapy of Gaucher’s disease using a plant cell system. Plant
Biotechnol. J. 2007, 5, 579–590.
(10) Brumshtein, B.; Salinas, P.; Peterson, B.; Chan, V.; Silman, I.;
Sussman, J. L.; Savickas, P. J.; Robinson, G. S.; Futerman, A. H.
Characterization of gene-activated human acid-{beta}-glucosi-
dase: crystal structure, glycan composition and internalization into
macrophages. Glycobiology 2010, 20, 24–32.
(11) Lieberman, R. L.; D’Aquino, J, A.; Ringe, D.; Petsko, G. A. Effects
of pH and iminosugar pharmacological chaperones on lysosomal
glycosidase structure and stability. Biochemistry 2009, 48, 4816–4827.
(12) Kacher, Y.; Brumshtein, B.; Boldin-Adamsky, S.; Toker, L.;
Shainskaya, A.; Silman, I.; Sussman, J. L.; Futerman, A. H. Acid
beta-glucosidase: insights from structural analysis and relevance to
Gaucher disease therapy. Biol. Chem. 2008, 389, 1361–1369.
(13) Premkumar, L.; Sawkar, A. R.; Boldin-Adamsky, S.; Toker, L.;
Silman, I.; Kelly, J. W.; Futerman, A. H.; Sussman, J. L. X-ray
structure of human acid-beta-glucosidase covalently bound to
conduritol-B-epoxide. Implications for Gaucher disease. J. Biol.
Chem. 2005, 280, 23815–23819.
(30) Compounds were screened against the following glycosidases:
sweet almond β-glucosidase, yeast R-glucosidase, rice R-glucosi-
dase, rice R-galactosidase, and bovine liver β-galactosidase. For
further details, see the Supporting Information.
(31) Among the most active compounds, only 1d was nontoxic at
300 μM in wt human fibroblasts, whereas 1e and 1f showed CC50
values of 55 and 19 μM, respectively. The remaining compounds
were not toxic at 300 μM, with the exceptions of 3b and 3c, with
CC50 values of 97 and 170 μM, respectively. Taking the CC50/Ki
ratio as a measure of compound efficacy, the following values are
obtained: 1c g 900; 1d g 3300; 1e = 916; 1f = 237.
(32) Warren, G. L.; Andrews, C. W.; Capelli, A. M.; Clarke, B.;
LaLonde, J.; Lambert, M. H.; Lindvall, M.; Nevins, N.; Semus,
S. F.; Senger, S.; Tedesco, G.; Wall, I. D.; Woolven, J. M.; Peishoff,
C. E.; Head, M. S. A critical assessment of docking programs and
scoring functions. J. Med. Chem. 2006, 49, 5912–5931.
(33) Guimaraes, C. R.; Cardozo, M. MM-GB/SA rescoring of docking
poses in structure-based lead optimization. J. Chem. Inf. Model.
2008, 48, 958–970.
(34) Brik, A.; Alexandratos, J.; Lin, Y. C.; Elder, J. H.; Olson, A. J.;
Wlodawer, A.; Goodsell, D. S.; Wong, C. H. 1,2,3-Triazole as a
peptide surrogate in the rapid synthesis of HIV-1 protease inhibi-
tors. ChemBioChem 2005, 6, 1167–1169.
(14) Brumshtein, B.; Greenblatt, H. M.; Butters, T. D.; Shaaltiel, Y.;
Aviezer, D.; Silman, I.; Futerman, A. H.; Sussman, J. L. Crystal
structures of complexes of N-butyl- and N-nonyl-deoxynojirimycin
bound to acid beta-glucosidase: insights into the mechanism of
chemical chaperone action in Gaucher disease. J. Biol. Chem. 2007,
282, 29052–29058.
(15) Lieberman, R. L.; Wustman, B. A.; Huertas, P.; Powe, A. C., Jr.;
Pine, C. W.; Khanna, R.; Schlossmacher, M. G.; Ringe, D.; Petsko,
G. A. Structure of acid beta-glucosidase with pharmacological
chaperone provides insight into Gaucher disease. Nat. Chem. Biol.
2007, 3, 101–107.
(16) Brumshtein, B.; Aguilar-Moncayo, M.; Garcia-Moreno, M. I.;
Ortiz Mellet, C.; Garcia Fernandez, J. M.; Silman, I.; Shaaltiel,
Y.; Aviezer, D.; Sussman, J. L.; Futerman, A. H. 6-Amino-6-
deoxy-5,6-di-N-(N0-octyliminomethylidene)nojirimycin: synthesis,
biological evaluation, and crystal structure in complex with acid
beta-glucosidase. ChemBioChem 2009, 10, 1480–1485.
(17) Egido-Gabas, M.; Serrano, P.; Casas, J.; Llebaria, A.; Delgado, A.
New aminocyclitols as modulators of glucosylceramide meta-
bolism. Org. Biomol. Chem. 2005, 3, 1195–1201.
(18) Egido-Gabas, M.; Serrano, P.; Casas, J.; Zucco, M.; Emeric, G.;
Llebaria, A.; Delgado, A. New aminocyclitol libraries as candi-
dates to chemical chaperone therapy for Gaucher disease. Naunyn-
Schmiedeberg’s Arch. Pharmacol. 2007, 374, 340.
(19) Egido-Gabas, M.; Canals, D.; Casas, J.; Llebaria, A.; Delgado, A.
Aminocyclitols as pharmacological chaperones for glucocerebro-
sidase, a defective enzyme in Gaucher disease. ChemMedChem
2007, 2, 992–994.
(20) 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 deriva-
tives on mutant glucocerebrosidases causing Gaucher disease.
Blood Cells, Mol. Dis. 2009, 42, 159–166.
(21) Serrano, P.; Casas, J.; Zucco, M.; Emeric, G.; Egido-Gabas, M.;
Llebaria, A.; Delgado, A. Combinatorial approach to N-substi-
tuted aminocyclitol libraries by solution-phase parallel synthesis
and preliminary evaluation as glucocerebrosidase inhibitors.
J. Comb. Chem 2007, 9, 43–52.
(22) Serrano, P.; Llebaria, A.; Delgado, A. Regio- and stereoselective
synthesis of aminoinositols and 1,2-diaminoinositols from condur-
itol B epoxide. J. Org. Chem. 2005, 70, 7829–7840.
(35) Appendino, G.; Bacchiega, S.; Minassi, A.; Cascio, M. G.; De
Petrocellis, L.; Di Marzo, V. The 1,2,3-triazole ring as a peptido-
and olefinomimetic element: discovery of click vanilloids and
cannabinoids. Angew. Chem., Int. Ed. 2007, 46, 9312–9315.
(36) Odlo, K.; Hentzen, J.; dit Chabert, J. F.; Ducki, S.; Gani, O. A.;
Sylte, I.; Skrede, M.; Florenes, V. A.; Hansen, T. V. 1,5-Disubsti-
tuted 1,2,3-triazoles as cis-restricted analogues of combretastatin
A-4: synthesis, molecular modeling and evaluation as cytotoxic
agents and inhibitors of tubulin. Bioorg. Med. Chem. 2008, 16,
4829–4838.
(37) 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.
€
(38) Maestro, version 8.5; Schrodinger, LLC: New York, 2008.
€
(39) MacroModel, version 9.6; Schrodinger, LLC: New York, 2005.
(40) Jorgensen, W. L.; Maxwell, D. S.; TiradoRives, J. Development
and testing of the OPLS all-atom force field on conformational
energetics and properties of organic liquids. J. Am. Chem. Soc.
1996, 118, 11225–11236.
(41) Still, W. C.; Tempczyk, A.; Hawley, R. C.; Hendrickson, T.
Semianalytical treatment of solvation for molecular mechanics
and dynamics. J. Am. Chem. Soc. 1990, 112, 6127–6129.
€
(42) Glide, version 5.0; Schrodinger, LLC: New York, 2008.
(43) Friesner, R. A.; Banks, J. L.; Murphy, R. B.; Halgren, T. A.; Klicic,
J. J.; Mainz, D. T.; Repasky, M. P.; Knoll, E. H.; Shelley, M.; Perry,
J. K.; Shaw, D. E.; Francis, P.; Shenkin, P. S. Glide: a new
approach for rapid, accurate docking and scoring. 1. Method
and assessment of docking accuracy. J. Med. Chem. 2004, 47,
1739–1749.
(44) Halgren, T. A.; Murphy, R. B.; Friesner, R. A.; Beard, H. S.; Frye,
L. L.; Pollard, W. T.; Banks, J. L. Glide: a new approach for rapid,
accurate docking and scoring. 2. Enrichment factors in database
screening. J. Med. Chem. 2004, 47, 1750–1759.
(23) Serrano, P.; Llebaria, A.; Vazquez, J.; de Pablo, J.; Anglada, J. M.;
Delgado, A. On the regio- and stereoselective synthesis of amino-
cyclitols from cyclitol epoxides: the effect of Li as a chelating agent.
Chem.;Eur. J. 2005, 11, 4465–4472.
(24) The recovery of recombinant GlcCerase activity after thermal
denaturation is currently used as an in vitro model assay for
potential pharmacological chaperones (see ref 37).
(45) Friesner, R. A.; Murphy, R. B.; Repasky, M. P.; Frye, L. L.;
Greenwood, J. R.; Halgren, T. A.; Sanschagrin, P. C.; Mainz, D. T.
Extra precision glide: docking and scoring incorporating a model
of hydrophobic enclosure for protein-ligand complexes. J. Med.
Chem. 2006, 49, 6177–6196.
(46) Klicic, J. J.; Friesner, R. A.; Liu, S. Y.; Guida, W. C. Accurate
prediction of acidity constants in aqueous solution via density
functional theory and self-consistent reaction field methods.
J. Phys. Chem. A 2002, 106, 1327–1335.
(25) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B. A
stepwise Huisgen cycloaddition process: copper(I)-catalyzed regio-
selective “ligation” of azides and terminal alkynes. Angew. Chem.,
Int. Ed. 2002, 41, 2596–2599.
€
(47) Jaguar, version 7.5; Schrodinger, LLC: New York, 2008.
(48) Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T. N.;
Weissig, H.; Shindyalov, I. N.; Bourne, P. E. The Protein Data
Bank. Nucleic Acids Res. 2000, 28, 235–242.
ꢁ
ꢁ
ꢁ
(26) Gil, M. V.; Arevalo, M. J.; Lopez, O. Click chemistry;What’s in a
name? Triazole synthesis and beyond. Synthesis 2007, 1589–1620.
(27) Gonzalez-Bulnes, P.; Casas, J.; Delgado, A.; Llebaria, A. Practical
synthesis of (-)-1-amino-1-deoxy-myo-inositol from achiral pre-
cursors. Carbohydr. Res. 2007, 342, 1947–1952.
€
(49) LigPrep, version 2.2; Schrodinger, LLC: New York, 2005.
(50) Molecular Operating Environment, version 2008.10; Chemical Com-
puting Group: Montreal, Quebec, Canada, 2008.