Polyhydroxylated bicyclic isoureas and guanidines are potent glucocerebrosidase inhibitors and nanomolar enzyme activity enhancers in gaucher cells

Article abstract of DOI:10.1021/ja111480z

Four diastereomeric series of N-alkylated [6+5] bicyclic isoureas having hydroxyl substituents mimicking glucose hydroxyl groups have been synthesized as potential β-glucocerebrosidase (GCase) inhibitors with the aim of developing pharmacological chaperon

Full text of DOI:10.1021/ja111480z

ARTICLE
pubs.acs.org/JACS
Polyhydroxylated Bicyclic Isoureas and Guanidines Are Potent
Glucocerebrosidase Inhibitors and Nanomolar Enzyme Activity
Enhancers in Gaucher Cells
Ana Trapero,† Ignacio Alfonso,‡ Terry D. Butters,§ and Amadeu Llebaria*,†
†Research Unit on Bioactive Molecules (RUBAM), Departament de Química Biomꢀedica, Institut de Química Avanc-ada de Catalunya
(IQACꢀCSIC), Jordi Girona 18-26, 08034 Barcelona, Spain
‡Departament de Química Biolꢀogica y Modelitzaciꢁo Molecular, IQACꢀCSIC, Jordi Girona 18-26, 08034, Barcelona, Spain
§Oxford Glycobiology Institute, Department of Biochemistry, University of Oxford, South Parks Road, Oxford OX1 3QU, United
Kingdom
S Supporting Information
b
ABSTRACT: Four diastereomeric series of N-alkylated [6þ5]
bicyclic isoureas having hydroxyl substituents mimicking glucose
hydroxyl groups have been synthesized as potential β-glucocere-
brosidase (GCase) inhibitors with the aim of developing pharma-
cological chaperones for enzyme deficiency in Gaucher disease
(GD). The bicyclic compounds differ either by the configuration
of the ring fusion carbon atoms or by the nature of the N-alkyl
substituents. When assayed for effects on GCase activity, the
isoureas displayed selective inhibition of GCase with low micro-
molar to nanomolar IC50’s in isolated enzyme experiments. One of the series of isoureas, a family having a specific cis ring fusion,
exhibited strong inhibition of recombinant GCase activity with Ki values in the 2ꢀ42 nM range. In addition, the [6þ5] bicyclic
guanidine derivatives with a substitution pattern analogous to the most active isoureas were also found to be potent inhibitors of
GCase with Ki values between 3 and 10 nM. Interestingly, the active bicyclic isoureas and guanidines also behaved as GCase
inhibitors in wild-type human fibroblasts at nanomolar concentrations. The potential of these compounds as pharmaceutical
chaperones was determined by analyzing their capacity for increasing GCase activity in GD lymphoblasts derived from N370S and
L444P variants, two of the most prevalent Gaucher mutations. Six compounds were selected from the different bicyclic isoureas and
guanidines obtained that increased GCase activity by 40ꢀ110% in N370S and 10ꢀ50% in L444P cells at low micromolar to
nanomolar concentrations following a 3 day incubation. These results describe a promising series of potent GCase ligands having the
cellular properties required for pharmacological chaperones.
’ INTRODUCTION
alleviate the CNS symptoms of severe disease.8 A drug that
inhibits the production of GlcCer (N-butyl-deoxynojirimycin),
thus decreasing the substrate load on GCase, has recently
become available as a therapy for type 1 patients.9 However, its
efficacy against CNS Gaucher variants is unknown.10 Since this
therapy reduces GlcCer biosynthesis nonselectively, cellular
pathways utilizing GlcCer and other metabolites are likely to
be adversely affected by the drug, which may explain its side
effects.
An approach that has recently attracted much interest for the
treatment of GD and other lysosomal storage disorders is enzyme
enhancement therapy with pharmacological chaperones.11,12 These
are small molecules that specifically bind to and stabilize the
functional form or three-dimensional shape of a misfolded protein in
the endoplasmic reticulum (ER) of a cell. When misfolded because
Lysosomal storage diseases (LSDs) are a group of genetic
disorders due to defects in some aspect of lysosomal biology.1
Among them, Gaucher disease (GD) is the most prevalent, and it
is caused by the reduced activity of the lysosomal enzyme acid
β-glucosidase, also known as β-glucocerebrosidase (GCase).2
This leads to the accumulation of the substrate glucosylceramide
(GlcCer) in the lysosomes of macrophages. GlcCer storage results
in hepatomegaly, splenomegaly, anemia, bone lesions, and central
nervous system (CNS) involvement.3 Patients without CNS symp-
toms are classified as type 1, whereas those with CNS symptoms
are classified as either type 2 (acute infantile) or type 3 (juvenile
or early adult onset). Current treatment approaches4 for GD include
replacement of the defective enzyme with recombinant enzyme5
(imiglucerase or velaglucerase alfa) or inhibition of GlcCer
production.6 Type 1 disease can be managed with enzyme
replacement therapy.7 However, the treatment is expensive,
and the enzyme is unable to cross the bloodꢀbrain barrier to
Received: December 21, 2010
Published: March 17, 2011
r
2011 American Chemical Society
5474
dx.doi.org/10.1021/ja111480z J. Am. Chem. Soc. 2011, 133, 5474–5484
|

Journal of the American Chemical Society
ARTICLE
Figure 1. Chemical structures of pharmacological chaperones
for GCase.
of a genetic mutation, the enzyme is unable to adopt the correct
functional shape. This misfolded protein is recognized by the
quality control system in the cell and destroyed, leading to decreased
amounts of the enzyme transported from the ER to the lysosome
and, as a consequence, a reduction of the effective enzyme activity
results. The binding of the chaperone molecule is proposed to
help during the protein folding and trafficking from the ER and is
distributed to the lysosome in the cell, thereby increasing enzyme
activity and cellular function, reducing substrate and stress on
cells. Thus, iminosugar-type glycomimetics with GCase inhibi-
tory activity, such as N-nonyl-deoxynojirimycin (NN-DNJ),13
R-1-C-octyl-1-deoxynojirimycin (CO-DNJ),14 R-1-C-nonyl-1-deoxy-
nojirimycin (R-1-C-nonyl-DIX),15 valienamine derivatives (NOV),16
isofagomine derivatives,17ꢀ19 N-substituted δ-lactams,20 and bicy-
clic nojirimycin analogues with the structure of sp2 iminosugars
(NOI-NJ)21 have all shown high promise as pharmacological
chaperones for the treatment of GD by increasing cellular GCase
catalytic activity (Figure 1).
The discovery of selective and potent glycosidase inhibitors
has been intensely stimulated by this novel perspective for
therapeutic intervention in the LSDs. Important members of
this group originate from the structural modification of natural
sugars to increase their metabolic stability and recognition by
enzymes or other carbohydrate binding proteins.22 These com-
pounds mimic either enzyme substrates or products or even the
enzymatic hydrolysis reaction intermediates.23
Figure 2. Chemical structures of bicyclic isoureas and guanidines
(1ꢀ5) evaluated in this work. The structures of the cyclopentitols
trehazolin and allosamidin are also depicted.
series 1, 3, and 4 in racemic form and the enantiomerically pure 2,3-
disubstituted isoureas (1dꢀ4d) were also prepared and evaluated.
It is interesting to note the structural relationship of these
bicyclic compounds with the R-glucosidase inhibitor trehazolin
and the chitinase inhibitor allosamidin (Figure 2),33 which contain
cyclic isourea moieties although in a different structural arrange-
ment to that present in our molecules.
The biological activity of these new bicyclic compounds was
determined using recombinant GCase, and the selectivity in-
hibitory profile of compounds examined for other glycosidases
and glucosylceramide synthase (GCS) from cell homogenates.
After studying the toxicity of compounds, the inhibitory effect of
bicyclic isoureas and guanidines on GCase was examined in wild-
type human fibroblasts. Next, the effect of some of the active
compounds on the stabilization of recombinant GCase activity
after thermal denaturation was determined as a measure for
potential pharmacological chaperone activity. Finally, N-nonyl
derivatives 1cꢀ5c and guanidine 5e were further investigated as
pharmacological chaperones in human lymphoblasts derived
from Gaucher patients homozygous for N370S (non-neurono-
phatic form) or L444P (neuronophatic form) variants, which are
the two most common mutations associated with GD.34
Several polyhydroxylated bicyclic systems are efficient glyco-
side hydrolase inhibitors. These include products of natural
origin, such as castanospermine,24 swainsonine,25 or the calystegins,26
or compounds of synthetic origin, such as the glycoside-aromatic
heterocycle hybrids developed by Vasella27 or the cyclic amidine
sugars described by Mioskowski.28 To the best of our knowledge,
the above sp2-iminosugar analogues of the nojirimycin, such as
the compound NOI-NJ, are the only bicyclic compounds which
have shown promise as pharmacological chaperones for the
treatment of GD.21
’ RESULTS
Synthesis. Bicyclic isoureas and guanidines 1ꢀ5 were ob-
tained by coupling the amino alcohols 6ꢀ13 or diamine 14 with
an isothiocyanate and subsequent cyclization of the resulting
thioureas using an excess of yellow mercury oxide (II) provided
the bicyclic isourea and guanidine analogues. The final products
1ꢀ5 were obtained after the O-benzyl deprotection by hydro-
genolysis or by reaction with BCl3, following reported protocols.35,36
The N-nonyl protected aminocyclitols 7,29 9, 11, and 13 were
obtained by reductive amination of nonanal with the correspond-
ing protected aminocyclitol 6,37 8,37,38 10,39 and 12,35,37 which
were obtained from conduritol B epoxide according to literature
procedures. The above protected aminocyclitols (6ꢀ13) were
Based on bibliographic data,29ꢀ31 improvement of potency
and selectivity against GCase was observed when substituents
close to the nitrogen atom were preferentially hydrophobic,
aliphatic, or aromatic. Accordingly, we now report the synthesis
of four families of N-nonyl bicyclic isourea derivatives32 having a
relative trans (isoureas 1c and 2c) or cis (isoureas 3c and 4c)
disposition between the C3a and C7a substituents to explore
their effects on GCase activity (Figure 2). Since the N-nonyl
derivative 4c was found to be a strong inhibitor of GCase,
guanidines 5c and 5e in enantiomerically pure forms were also
synthesized. Furthermore, N-hexyl and N-octyl derivatives of
5475
dx.doi.org/10.1021/ja111480z |J. Am. Chem. Soc. 2011, 133, 5474–5484

Journal of the American Chemical Society
ARTICLE
Scheme 1. Synthesis of Bicyclic Isoureas 1ꢀ4 from Amino Alcohols 6, 8, 10, and 12a
a Reagents and conditions: (a) C8H17CHO, NaBH3CN, AcOH, MeOH; (b) RNCS, THF, 45 °C; (c) HgO, toluene, 115 °C; (d) C8H17NCO, CH2Cl2;
(e) Tf2O, pyridine, CH2Cl2, ꢀ40 °C; (f) Pd(OH)2, EtOH, H2 (2 atm), 60 °C; (g) BCl3, CH2Cl2, ꢀ78 °C; and (h) Pd/C, MeOH, H2 (2 atm).
coupled with commercial N-alkyl isothiocyanates, which afforded
the thiourea adducts 15ꢀ18 (Scheme 1) in a reaction that proceeds
with total chemoselectivity in the presence of unprotected hydroxyl
group.40
Scheme 2. Synthesis of Bicyclic Guanidines 5c, e from Dia-
mine 14a
Treatment of the resulting thioureas with an excess of yellow
mercury oxide (II) afforded the corresponding bicyclic isoureas
19ꢀ22 (Scheme 1).
Similarly, reaction of diamine 1441 with 1 equiv of nonyl or
3-phenylpropyl isothiocyanate gave a diastereoisomeric mixture
of the thioureas 23 and 24, which were easily cyclized with yellow
mercury oxide (II) to give bicyclic guanidine analogues 25c, e
(Scheme 2).
It has been reported42 the reaction of N-alkyl urea compounds
with a vicinal hydroxyl group with triflic anhydride in the
presence of pyridine afforded the corresponding 2-aminooxazo-
line with concomitant inversion of stereochemistry at the center
bearing the secondary hydroxyl group through the intramolecu-
lar SN2 displacement of a transient triflate by vicinal carbonyl
group. We initially sought to apply this procedure for the
synthesis of bicyclic isoureas 22aꢀc from amino alcohol 6. Thus
urea 26 was obtained by reaction of amino alcohol 6 with octyl
isocyanate. However, the treatment of the above urea under
similar reaction conditions as previously reported42 led to the
bicyclic isourea 19b with retention of stereochemistry at the
carbon bearing the secondary hydroxyl group (Scheme 1).43
a Reagents and conditions: (a) C9H19NCS or Ph(CH2)3NCS, THF,
45 °C; (b) HgO, toluene, 115 °C; and (c) BCl3, CH2Cl2, ꢀ78 °C.
The final step of the synthesis would require the removal of all
O-benzyl protective groups in isoureas 19aꢀc. This was accom-
plished by hydrogenolysis to give the desired bicyclic compounds
1aꢀc in good yields (Scheme 1). However, hydrogenolysis of 21
and 22 afforded complex mixtures. Gratifyingly, alternative benzyl
removal could be efficiently achieved with BCl3 at low tempera-
ture (Schemes 1 and 2).37 It is noteworthy that monobenzylated
compounds 27aꢀd and 28aꢀc were obtained as a byproduct,
5476
dx.doi.org/10.1021/ja111480z |J. Am. Chem. Soc. 2011, 133, 5474–5484

Journal of the American Chemical Society
ARTICLE
all final compounds were first evaluated as inhibitors of recom-
binant GCase (imiglucerase, Cerezyme) at lysosomal pH (5.2).
Moreover, bicyclic compounds 1ꢀ5 were also tested at lumenal
ER pH (7.0) to determine a possible change in the affinity for
the enzyme with pH changes and to obtain information on
the possible affinity for the protein at the different cellular
pH values relevant for chaperone effects. The inhibitory acti-
vities of all compounds against imiglucerase are summarized
in Table 1.
Compounds 1ꢀ5 inhibited the enzyme activity at both neutral
and acidic pH, as shown in Table 1. Remarkably, the potency for
inhibiting imiglucerase activity was increased by more than three-
fold when measured at pH 7.0 compared to pH 5.2 for several
compounds.46 A similar dependence of the inhibitory activity
from the pH has been previously reported for other compounds,
including ambroxol, which showed enzyme enhancement on
human fibroblasts with various GCase mutations.47,48 As ex-
pected, compounds with one of the hydroxyl group protected as a
benzyl ether (27 and 28) were less potent inhibitors of imiglu-
cerase than the corresponding bicyclic compounds with all free
hydroxyl groups (3 and 4), but they still showed inhibition with
Ki’s in the nanomolar range for (()-27c and 28bꢀc.
Figure 3. Observed (NMR) chair conformations of the studied com-
pounds and inhibition constants (Ki) of the N-nonyl derivatives (R1dH
and RdC9H19) for the inhibition of imiglucerase.
which could be easily separated by column chromatography.44
Surprisingly, the BCl3 O-benzyl deprotection in isourea 21d led
exclusively to monobenzylated compound 27d, which was
transformed into the final compound 3d after removing the
remaining O-benzyl protecting group by hydrogenolysis (see
Scheme 1).
The N-nonyl derivatives 1c, 3c, and 4c were found to be better
inhibitors for this enzyme than the corresponding N-hexyl and
N-octyl derivatives in each series, a fact that is in agreement with
the correlation between lipophilicity and the inhibitory activity
that was observed in other glycomimetic families in this enzyme.29,30
Remarkably, the simple extension of the N-alkyl chain length,
from octyl to nonyl, resulted in more than a three-fold increase in
enzyme activity. In contrast, the introduction of second alkyl
chain at the endocyclic nitrogen greatly decreased the inhibitory
activity (IC50 values around 10 μM at pH 5.2 for compounds 1d,
3d, and 4d and 160 μM for 2d).
All the newly synthesized compounds were fully characterized,
1
being their corresponding H and 13C NMR signals assigned
with the help of 2D NMR experiments (see Supporting In-
formation). The knowledge of the solution structure of the final
compounds (1ꢀ5) could be important for the rationalization of
the biological results (see below). From the NMR analysis (see
Figure S1, Supporting Information), we concluded that all the
derivatives are protonated at neutral or slightly acidic pH, which
are the conditions for the biological tests. Besides, we unambigu-
ously determined that compounds 1aꢀd, 3aꢀd, 4aꢀd, and
5c,e have their cyclohexane ring in a chairlike conformation with
the hydroxyl groups at positions 4ꢀ7 in equatorial (see Figures 3
and S2, Supporting Information). Thus, the fused cyclic isourea/
guanidine must adopt a trans diequatorial disposition for 1aꢀd,
and a cis disposition for 3aꢀd, 4aꢀd, and 5c,e (see Figures 3 and
S2, Supporting Information). The cis arrangement of the fused
heterocycle locates one of the heteroatoms in axial (N for 3aꢀd,
O for 4aꢀd, and N for 5c,e) and the other one in equatorial (O
for 3aꢀd, N for 4aꢀd, and N for 5c,e). In the special case of the
2aꢀd family, the configuration of the isourea cycle forced a
chair-like geometry with all the OH groups in axial, although an
all-equatorial twisted-boat conformer should be energetically
accessible (for further NMR and modeling details, see Support-
ing Information).
All tested compounds were found competitive inhibitors of
imiglucerase, as illustrated in Figure 4 for (þ)-4c (see also
Figures S4ꢀS8, Supporting Information). It is worth mentioning
the low Ki values found for isoureas 4aꢀc and guanidines 5c,e
that were in the 2ꢀ42 nM range.
In spite of the good correlation between lipophilicity and Ki
values, the configuration of the ring junction carbon atoms
bearing the oxygen and nitrogen atoms of the fused heterocycle
had a decisive effect on the inhibition of imiglucerase (see
Figure 3). Compounds 2c and 3aꢀc exhibited Ki values between
0.4 and 11 μM, whereas 1bꢀc, 4aꢀc, and 5c,e were effective
inhibitors of recombinant GCase with Ki values in the mid- to
low-nanomolar range.
An interesting point is the hydroxyl matching between the
series obtained and that present in the β-gluco configuration of
GlcCer substrate. Except for compounds 2cꢀd, all the other
series have equatorial hydroxyl substituents in the cyclohexane
ring that can match exactly the 2, 3, and 4-OH substituents
present in glucose. As pointed out above, the series 2 showed a
chair conformation with all axial hydroxyl arrangement according
to NMR experiments performed with these compounds and
therefore a very poor match with hydroxyl groups in glucose
results (see Figures 3 and S2, Supporting Information). How-
ever, when the compounds of this series were tested, compounds
(()-2c or (þ)-2c have been found to be low micromolar
inhibitors of the enzyme, and although are less potent than 1,
4, and 5 series, a similar activity to 3c resulted, in spite of the fact
of the very different orientation of the hydroxyl groups. It is
known that GCase has a high selectivity for lipidic β-gluco
’ BIOLOGICAL RESULTS
Inhibition of Imiglucerase and GCase in Wild-Type
Human Fibroblasts by Bicyclic Isoureas and Guanidines.
As pointed out above, we seek to develop enzymatic inhibitors
with potential as pharmacological chaperones for GCase. One
of the first conditions is that compound binds the protein
with high affinity. Most of pharmacological chaperones
are competitive inhibitors of their target enzyme.45 Therefore
5477
dx.doi.org/10.1021/ja111480z |J. Am. Chem. Soc. 2011, 133, 5474–5484

Journal of the American Chemical Society
ARTICLE
Table 1. Inhibitory Activity of Compounds against Imiglucerase and GCase Inhibition in Wild-Type (wt) Human Fibroblastsa
imiglucerase
IC50 (μM)
compound
pH 7.0
pH 5.2
Ki (μM)b
% GCase inhibition (in wt human fibroblasts)c
(()-1a
(()-1b
(()-1c
(ꢀ)-1c
(ꢀ)-1d
(()-2c
(þ)-2c
(þ)-2d
(()-3a
(()-3b
(()-3c
(ꢀ)-3c
(ꢀ)-3d
(()-4a
(()-4b
(()-4c
(þ)-4c
(þ)-4d
(ꢀ)-5c
(ꢀ)-5e
(()-27a
(()-27b
(()-27c
(ꢀ)-27c
(ꢀ)-27d
(()-28a
(()-28b
(()-28c
(þ)-28c
NN-DNJ
1.6
0.08
0.04
0.02
5.6
6.4
0.60
0.22
0.12
11.1
2.9
3.4
0.23
0.07
0.04
12.5
2.4
4
67
72
87
24d
0.60
0.32
56.7
4.4
34
1.9
1.5
33
40d
160
12.9
2.5
NDe
11.1
1.7
0
0.54
0.19
1.7
2
0.71
6.0
0.44
1.8
53
46
63d
1.2
10.4
0.10
0.02
0.009
0.008
6.8
4.8
0.04
0.01
0.008
0.005
2.4
0.04
0.007
0.002
0.002
2.8
77
97
99
100
17d
99
0.004
0.01
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.30
0.01
0.04
64.4
6.4
0.003
0.01
25.9
3.6
100
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
1.3
0.53
8.8
17.3
157
3.9
ND
1.3
0.74
1.7
0.13
0.47
0.18
0.30f
0.35
0.66f
a Typical results differed by less than 6%. See Supporting Information, Tables S2 and S5. b The inhibition was competitive in all cases (determined at pH
5.2). c Incubation for 24 h at 50 μM inhibitor and 5 mM substrate in wt human fibroblast. d Toxic at 50 μM. Incubation for 24 h at 5 μM inhibitor and
5 mM substrate in wt human fibroblast. e ND: not determined. f See ref 15.
substrates, so it is not easy to understand why these compounds
are competitive inhibitors of the enzyme. If the all axial OH chair
would be the binding conformation, then this would probably
preclude the recognition of the inhibitor by the enzyme, and no
activity would be expected. We suspected that probably a conforma-
tion other than the chair was the binding conformation of (þ)-2c
compound, which should match the gluco OH tridimensional
arrangement and also must be energetically accessible from the
chair conformation. A calculation performed with a model showed
that a twist conformation is found approximately 1.5 kcal molꢀ1
above the chair conformation and when compared to a β-glucose
chair showed a reasonable match (see Figure S3, Supporting
Information). Moreover, the barrier between these conforma-
tions is about 7 kcal molꢀ1 and is the same order (or even lower)
to that found in cyclohexanes for chair to chair conformational
transitions through a twist intermediate. Although there is no
evidence of this conformation as actually being the binding
conformation for (þ)-2c, this would explain why this molecule,
having a structure apparently different from a β-glucoside, is an
enzyme inhibitor for GCase. The higher value of Ki for this compound
would be in agreement with a high-energy conformer binding.
Apart from the hydroxyl substituents, it is also interesting to
note that the more active series (compounds 1, 4, and 5) have the
endocyclic nitrogen atom of the fused heterocycle in an equato-
rial arrangement matching the β-glycoside anomeric oxygen present
in the substrate GlcCer. The oxygen by nitrogen replacement is
commonly found in other aminocyclitol inhibitors.29,36,49 The
reasons for the increase in potency in isoureas 4aꢀc, which have
a cis ring fusion due to the axial configuration of the isourea
oxygen atom, over diastereomeric 1aꢀc having a trans ring
fusion and an equatorial oxygen are presently unclear and will
be addressed in future structural studies. However, a better
recognition by the enzyme for the cis compounds due to selective
interactions with active site amino acids is hypothesized.
The data also indicate that isourea 4c and guanidine 5c
displayed similar GCase inhibitory potencies, suggesting that
5478
dx.doi.org/10.1021/ja111480z |J. Am. Chem. Soc. 2011, 133, 5474–5484

Journal of the American Chemical Society
ARTICLE
Figure 5. GCase inhibition of 4aꢀc and 5c,e in wild-type human
fibroblasts after 24 h incubation time at the indicated inhibitor
concentrations.
Figure 4. Type of inhibition of imiglucerase by (þ)-4c. Double
reciprocal plot of imiglucerase incubated at different concentrations of
substrate and compound (þ)-4c [empty square (0): 0 nM, circle (b):
5 nM, solid square (9): 9 nM, triangle (2): 12 nM). Regression lines
arise from data obtained in two different experiments with duplicates.
compounds are powerful GCase inhibitors in live cells, reflecting
good membrane permeability and cellular stability properties to
inhibit the cellular enzyme.
The selectivity of the bicyclic inhibitors for GCase was evaluated
using a panel of commercial glucosidases and galactosidases (see
Table S3, Supporting Information). Some compounds showed
evidence of inhibitory activity on R-glucosidase from baker’s
yeast and on β-galactosidase from bovine liver at concentrations
similar to those required for GCase inhibition. However, with regard
to other possible glycosidase cellular targets of these compounds,
no essential inhibitory activity was found for lysosomal β-galacto-
sidase or R-glucosidase enzymes at concentrations of active
compound where GCase was almost totally inhibited (see Table
S6, Supporting Information), indicative of a high selectivity for
the GCase enzyme.
Stabilization of Recombinant GCase (imiglucerase) under
Thermal Denaturation Conditions. One of the main features of
pharmacological chaperones is the thermodynamic stabilization
they provide when bound to the protein, preventing it from aggregat-
ing and keeping it in a foldable state until it is properly folded and
exported from the ER. Resistance to thermal denaturation in the
presence of a potential pharmacological chaperone is a method
used to assess the stabilization effect.13,29,47,49 In this context, the
effect of N-nonyl derivatives 1cꢀ5c, N-3-phenylpropyl guanidine
5e, and 2,3-disubstituted isoureas 1d and 4d on the stabilization
of recombinant GCase activity after thermal denaturation at
48 °C was determined in the presence and in the absence of
increasing concentrations of compounds at different incubation
times (see Figures S10 and S11, Supporting Information) as a
measure for potential pharmacological chaperone activity. For
comparative purposes, NN-DNJ was also assayed under the same
experimental conditions.
The stabilization ratio values found for each compound at
5 μM (for 1c, 4c, and 5c,e), 10 μM (for 1d, 2c, 3c, and 4d) or
50 μM for NN-DNJ and 60 min incubation time are summarized
in Table 2.
Interestingly, bicyclic compounds 1c, 4c, 5c, and 5e, which
were effective inhibitors of recombinant GCase with Ki values in
the low-nanomolar range, showed stabilization ratios between 9
and 27 at 5 μM and 60 min incubation time. Compounds 1d and
4d, which displayed relatively weak imiglucerase inhibition,
showed an enzyme stabilization profile comparable to or even
superior to that of NN-DNJ, with stabilization ratios around 10 at
50 μM after 1 h of incubation at 48 °C (see Figure S10,
Supporting Information). Collectively, these results reflect a
positive effect of the compounds on enzyme stability and provide
the replacement of isourea oxygen by an NH group is not essentially
affecting the binding of these compounds to the enzyme. Interest-
ingly, guanidine 5e having an N-3-phenylpropyl substituent was a
potent competitive inhibitor of GCase with a Ki value of 10 nM.
This result would indicate that nitrogen substituents other than
linear alkyl can be tolerated without affecting the inhibition of the
enzyme to a large extent. It is noteworthy that the racemic isourea
(()-3c was found to be more active than the enantiopure
(ꢀ)-3c, this would point toward a better activity for the
enantiomer of (ꢀ)-3c, especially at neutral pH. However, the
activity in cellular assays was similar for (()-3c and (ꢀ)-3c.
These results show that these compounds are potent GCase
inhibitors in isolated enzyme assays. To further progress in their
activity profiling, the cellular GCase inhibition of bicyclic iso-
ureas and guanidines 1ꢀ5 was also studied in wild-type human
fibroblasts after 24 h of incubation at 50 μM. Cytotoxicity assays
were done previously to the enzyme cellular assay in wild-type
human fibroblasts to ensure that the measured activity was not
affected by the possible toxicity of the compounds. Since fibroblasts
incubated with compounds 1dꢀ4d at 50 μM displayed some
signs of cytotoxicity, these isoureas were tested at 5 μM, well
below the CC50 determined for these compounds (see Table S4,
Supporting Information). However, the rest of compounds were
not cytotoxic at concentrations up to 300 μM. The GCase
inhibition in wild-type human fibroblasts at above concentrations
is shown in Table 1. In general, a good correlation was observed
between the chain length in each series and the inhibition in cell
culture as well as between Ki values against imiglucerase and
lysosomal β-glucosidase inhibition. The N-nonyl derivatives
(þ)-2c and (ꢀ)-3c showed moderate inhibition (33 and 46%,
respectively) at 50 μM, whereas isoureas (ꢀ)-1c, (þ)-4c, and
guanidines (ꢀ)-5c,e showed more than 87% of GCase inhibition
at the same concentration. Surprisingly, isoureas 1d, 2d, and 4d,
which exhibited moderate inhibitory activities against imiglucer-
ase, showed between 17 and 40% of inhibition at 5 μM, whereas
3d displayed an inhibition of 63% at the same concentration.
The high potency of isoureas 1bꢀc, 3c, 4aꢀc, and guanidines
5c,e required analysis at lower concentrations (see Figure S9,
Supporting Information). Interestingly, some of them also
behaved as GCase inhibitors at nanomolar to low-micromolar
concentrations, as shown in Figure 5. These results show that the
5479
dx.doi.org/10.1021/ja111480z |J. Am. Chem. Soc. 2011, 133, 5474–5484

Journal of the American Chemical Society
ARTICLE
Table 2. Enzyme Stabilization Ratio of Imiglucerase after Thermal Denaturation at 48 °C and Maximum Observed Increase in
GCase Activity Using Pharmacological Chaperones and NN-DNJ
compound
stabilization ratioa (imiglucerase)
N370S GCase activity increaseb
L444P GCase activity increaseb
(ꢀ)-1c
(ꢀ)-1d
(þ)-2c
(()-3c
(ꢀ)-3c
(()-4c
(þ)-4c
(þ)-4d
(ꢀ)-5c
(ꢀ)-5e
NN-DNJ
20c
1.7
13
1.4 ( 0.1 (1 μM)
NDd
1.1 ( 0.1 (1 μM)
ND
1.7 ( 0.2 (10 μM)
1.7 ( 0.1 (5 μM)
1.8 ( 0.1 (10 μM)
1.7 ( 0.1 (0.2 μM)
2.1 ( 0.1 (0.2 μM)
ND
1.3 ( 0.1 (1 μM)
ND
11
5
1.4 ( 0.1 (0.1 μM)
ND
ND
23c
1.5
27c
9c
1.5 ( 0.1 (0.01 μM)
ND
1.8 ( 0.1 (0.1 μM)
1.9 ( 0.1 (0.05 μM)
1.4 ( 0.1 (5 μM)
1.4 ( 0.1 (0.01 μM)
1.4 ( 0.1 (0.01 μM)
no activity
7.6e
a Defined as the ratio of relative enzymatic activities (inhibitor vs control) at a given inhibitor concentration and incubation time. Tabulated values for
10 μM inhibitor and 60 min incubation time. b N370S and L444P lymphoblasts from Gaucher patients were incubated with test compounds for 3 days
before being used for enzyme assay. Data in parentheses correspond to the concentration of the tested compound. Experiments were performed in
triplicate, and the mean ( SD is shown. The relative activity was obtained by normalizing the activity corresponding to each compound concentration
tested to the activity of untreated cells. c Tabulated values for 5 μM inhibitor and 60 min incubation time. d ND: not determined. e Tabulated values for
50 μM inhibitor and 60 min incubation time.
support for the evaluation of these bicyclic derivatives as phar-
macological chaperones for GD.
activity increased in a time-dependent manner in the presence of
compounds, reached a peak on days 3ꢀ4, then decreased slightly.
The effects of enantiomerically pure compounds 1cꢀ5c and
5e on L444P GCase activity were measured using GD type 2
patient-derived lymphoblasts cell line after incubation with above
compounds for 3 days. Compound 1c only slightly increased
(10% at 1 μM) GCase activity, which probably is not significant,
and therefore, it can be concluded that this compound is not
promising as a pharmacological chaperone for this mutation.
Under similar conditions, NN-DNJ was found to be completely
inactive as reported previously in the L444P enzyme.13 However,
2cꢀ5c and 5e showed a moderate increase in GCase activity
(1.3- to 1.5-fold) at low concentrations, as shown in Table 2.
Treatment with 1 μM of isourea 2c or 0.1 μM of 3c caused 1.3-
and 1.4-fold increase in the GCase activity of L444P cell line,
respectively. Furthermore, isoureas 4cꢀ5c and guanidine 5e
maximally increased the activity between 40 and 50% at 10 nM,
in agreement with the high affinity for GCase enzyme in both
isolated protein and cellular assays.
Activity of Bicyclic Isoureas and Guanidines on L444P and
N370S GCase Gaucher Cells. The effects of bicyclic compounds
1cꢀ5c and 5e in enantiomerically pure form on human GCase
activity were further evaluated in human lymphoblasts derived
from Gaucher patients homozygous for N370S50 or L444P muta-
tions at different concentrations.51 The iminosugar NN-DNJ,
which is a known chaperone for N370S GCase, but not L444P
GCase, was used as a control.13 The fold increase in enzyme
activity and the pharmacological chaperone concentration which
maximizes GCase activity after 3 days of incubation are summar-
ized in Table 2. The results obtained with the compounds at
other concentrations are given in the Supporting Information
(Figures S12 and S13).
A good correlation was found between GCase inhibition in
wild-type human fibroblast or recombinant GCase and concen-
tration of maximal N370S or L444P GCase activity, with the
exception of isourea (ꢀ)-1c. In the analysis of N370S GCase acti-
vation, the above isourea, which had a high affinity and enzyme
thermal stabilization properties, only showed a 40% increase in
enzyme activity at 1 μM. In contrast, the rest of compounds
showed more than a 1.7-fold increase over a range of concentra-
tions. Treatment with 10 μM of isoureas 2c and 3c caused 1.7-
and 1.8-fold increase in the GCase activity of N370S cell line,
respectively, whereas bicyclic compounds 4c, 5c, and 5e maxi-
mally increased the activity between 1.7- and 2.1-fold at nano-
molar concentrations (between 50 nM and 200 nM). Using
similar conditions, NN-DNJ gave a maximal enhancement of 1.4-
fold at 5 μM. It is worth mentioning that the enantiomerically
pure compounds 4c, 5c, and 5e led to an approximate 40% increase
in N370S GCase activity at a very low concentration of 10 nM
(see Figure 6A).
’ DISCUSSION
Pharmacological chaperones for the treatment of lysosomal
diseases and singularly for GD have been an object of intense
interest12 as a new therapy for these disorders.52 The bicyclic
isoureas and guanidines obtained have found to be potent inhibitors
of GCase in vitro. Moreover, these compounds maintain a high
cellular inhibitory activity at concentrations similar to those
found effective in isolated enzyme. This would indicate a good
cellular penetration and stability of the compounds. One of the
possible problems associated to the chaperone therapy is how to
manage the opposing role of a molecule acting both as inhibitor
and activator for an enzyme. This is usually faced by administra-
tion of the compound at subinhibitory concentrations, a question
that could be relatively simple in cultured cells, but it is not
evident in vivo. In any case, the use of low concentrations of
chaperones is highly desirable due to the increase in its selectivity
for GCase and expected lower side effects. Therefore the
compounds here described, with activity on the nanomolar range
An additional 5 day time-course analysis of chaperone activities
was carried out at 0.1 μM for enantiomerically pure compounds 4c,
5c, and 5e and the iminosugar NN-DNJ. N370S lymphoblasts
were cultured in the presence of compounds for 5 days, and GCase
activity was measured every day in lysed cells, as described in the
Supporting Information. As shown in Figure 6B, the GCase
5480
dx.doi.org/10.1021/ja111480z |J. Am. Chem. Soc. 2011, 133, 5474–5484

Journal of the American Chemical Society
ARTICLE
efficacy of isofagomine in adult patients with type 1 GD. The lack
of efficacy could be related to the low activity of isofagomine in
intact cell GCase, which requires 3 orders of magnitude higher
concentrations of compound to increase the GCase activity in
cells compared with the concentration needed to inhibit GCase
in vitro.18 This would reflect a low membrane permeability, a
limited ER access, or a poor cellular stability of isofagomine. In
the case of the active bicyclic isoureas and guanidines, we have
found that nanomolar concentrations of compound produce
enzyme activity enhancements on L444P and N370S mutations.
It is clear that, in addition to their high enzyme affinities, these
compounds reflect an advantageous cell permeability, distribu-
tion, and metabolism as required for pharmacological chaperones.
Finally, a comment is warranted on the exact mechanism of
these compounds and other compounds designed as pharmaco-
logical chaperones. With the data currently available for our
compounds and most of the literature reports, which are based
on the enzyme activity enhancement measurements, the use of
the chaperone term is probably inexact, and enzyme activator or
enhancer would be more accurate. Rigorously, a pharmacological
chaperone would bind to the mutant protein assisting to its
folding process and/or stabilize it against denaturation and
degradation in the cellular environment. This can promote the
correct subcellular localization of the mutant enzyme increasing
its cellular activity and preventing the accumulation of substrate.
These subjects have not been addressed in this work. Another
point of interest questions whether the activity enhancement due
to compound treatment is high enough for clinical improvement
or is too low to prevent disease progression. Although this cannot
be fully answered now, there is an interesting study in the
literature that demonstrates in a model of GD that there is a
relatively low threshold activity—about 11ꢀ15% of the wild-
type enzyme—that leads to substrate accumulation and hence to
the conditions for disease development.54 Therefore, it seems
that even relatively small relative increases in GCase activity can
be important in Gaucher treatment specially for those variants
insensitive to current therapies, such as L444P. Research ad-
dressed to answer these questions is beyond this study and will be
subject of future work.
Figure 6. The effects of compounds (þ)-4c, (ꢀ)-5c, and (ꢀ)-5e on
GCase activity in mutant N370S lymphoblasts. Experiments were
performed in triplicate, and each bar represents the mean ( SD. Enzyme
activity is normalized to untreated cells, assigned a relative activity of 1.
(A) Cells were cultured for 3 days in the absence or presence of
increasing concentrations (μM) of compounds before GCase activity
was measured. (B) Lymphoblasts were cultured for 5 days with NN-
DNJ, (þ)-4c, (ꢀ)-5c, and (ꢀ)-5e at 0.1 μM and cells were analyzed
each day for enzyme activity.
in cultured cells are among the best candidates for pharmacolo-
gical chaperones for GD.
’ CONCLUSIONS
The assay of the bicyclic isoureas and guanidines on L444P
GCase, which characterizes a more severe disease phenotype
than N370S, results in a moderate increase of enzymatic activity
(1.3- to 1.5-fold) at nanomolar to micromolar concentrations.
The effect of the compounds on L444P cell assays is a second
remarkable feature, since this mutation is highly reluctant to
chaperone enhancement. Several reports describe the use of
pharmacological chaperones for different GCase mutations that,
however, failed in L444P cell assays.13,16,17,21,47,53 This lack of
activity has been attributed to the distant location of the mutation
site from the enzyme catalytic center. To our knowledge, only
R-1-C-octyl-1-deoxynojirimycin,14 isofagomine,19 and ambroxol48
have been reported to increase activity in L444P Gaucher cells.
However, these compounds must be used at relatively high
concentrations to significantly increase GCase activity. In a
recent study on the use of the isofagomine as a pharmacological
chaperone for Gaucher L444P mutation,19 it has been reported
the failure of obtaining clinically meaningful improvements. The
assay included tests on animal models and a phase 2 randomized,
open-label clinical study to assess the safety, tolerability, and
Several diastereomeric bicyclic isoureas (1ꢀ4) and guanidine
analogues 5 have been synthesized as potential GCase pharma-
cological chaperones. The final compounds have been tested as
GCase inhibitors using isolated enzyme (imiglucerase) and wild-
type human fibroblasts. Among them, compounds 4b,c, 5c, and
5e were found to be potent inhibitors of recombinant GCase
with Ki values between 2 and 10 nM. In spite of the good
correlation between lipophilicity of the N-alkyl chain derivatives
in each series and GCase inhibition, the configurations of the
oxygen and nitrogen atoms of the fused heterocycle are of utmost
importance for bicyclic inhibitors. In contrast, isoureas bearing
an extra alkyl chain at endocyclic nitrogen (1dꢀ4d) were
markedly less potent inhibitors. The most active compounds
were found to stabilize recombinant GCase under denaturing
conditions and inhibit GCase enzyme in human fibroblasts at
nanomolar concentrations. Furthermore, the effects of N-nonyl
derivatives 1cꢀ5c and guanidine 5e on human GCase activities
were evaluated in lymphoblasts derived from Gaucher patients
for N370S or L444P mutations. Among them, compounds 4c,
5c, and 5e are able to increase the activity of N370S GCase
5481
dx.doi.org/10.1021/ja111480z |J. Am. Chem. Soc. 2011, 133, 5474–5484

Journal of the American Chemical Society
ARTICLE
between 1.8- and 2-fold at 100 nM. Moreover, the above
compounds maximally increase the GCase activity of L444P cell
line by about 40% at 10 nM. Further optimization and study of
these compounds on other mutations and in animal models to
lead to clinical candidates is of relevance. In addition, the
structural details of the binding of these inhibitors are also of
evident interest. Work along these lines is currently in progress.
Data for (þ)-28c: white solid; [R]2D5 þ 10.7 (c 0.9, MeOH); 1H NMR
(δ, 500 MHz, CD3OD): 0.90 (t, 3H, J = 7.0 Hz, CH3), 1.20ꢀ1.39 (m,
12H, 6xCH2), 1.54ꢀ1.63 (m, 2H, CH2), 3.22 (t, 2H, J = 7.0 Hz, NCH2),
3.61ꢀ3.67 (m, 3H, H-4, H-5, H-6), 3.98 (dd, 1H, J = 3.6, 7.1 Hz, H-7),
4.14ꢀ4.20 (m, 1H, H-3a), 4.74 (d, 1H, J = 11.4 Hz, PhCH2O), 4.85 (d,
1H, J = 11.4 Hz, PhCH2O), 4.95ꢀ5.05 (m, 1H, H-7a), 7.26ꢀ7.43 (m,
5H, Ph); 13C NMR (δ, 100 MHz, CD3OD): 14.4 (CH3), 23.7 (CH2),
27.7 (CH2), 30.2 (CH2), 30.39 (CH2), 30.42 (CH2), 30.6 (CH2), 33.0
(CH2), 43.8 (NCH2), 62.0 (C3a), 71.0 (C7), 74.3 (C6), 75.4 (C5), 76.5
(PhCH2O), 82.5 (C4), 84.3 (C7a), 128.7 (CHar), 129.1 (2xCHar), 129.3
(2xCHar), 139.7 (Car), 163.4 (C2). HRMS calcd for C23H37N2O5:421.2702
[M þ H]þ. Found: 421.2704.
’ EXPERIMENTAL SECTION
For general methods, see the Supporting Information.
Data for (þ)-4c: white solid; [R]2D5 þ 16.2 (c 0.9, MeOH); 1H NMR
(δ, 500 MHz, CD3OD): 0.90 (t, 3H, J = 6.9 Hz, CH3), 1.25ꢀ1.42 (m,
12H, 6xCH2), 1.57ꢀ1.65 (m, 2H, CH2), 3.25ꢀ3.30 (m, 2H, NCH2),
3.34ꢀ3.38 (m, 1H, H-5), 3.66 (t, 1H, J = 6.0 Hz, H-6), 3.81 (t, 1H, J =
8.3 Hz, H-4), 3.97ꢀ4.04 (m, 1H, H-7), 4.09ꢀ4.14 (m, 1H, H-3a),
5.13ꢀ5.24 (m, 1H, H-7a); 13C NMR (δ, 100 MHz, CD3OD): 14.4
(CH3), 23.7 (CH2), 27.6 (CH2), 30.3 (CH2), 30.4 (CH2), 30.6 (CH2),
33.0 (CH2), 44.0 (NCH2), 61.4 (C3a), 71.4 (C7), 74.5 (C6), 75.8 (C5),
76.5 (C4), 84.5 (C7a), 163.2 (C2). HRMS calcd for C16H31N2O5:
331.2233 [M þ H]þ. Found: 331.2235.
General Procedure for the Preparation of Thioureas:
Synthesis of (þ)-18c as a Representative Example. To a
solution of the amino alcohol (ꢀ)-12 (111 mg, 0.21 mmol) in THF
(6 mL) was added 87 μL of nonyl isothiocyanate (0.42 mmol). The
resulting mixture was stirred at 45 °C for 18 h, and the solvent was
removed under reduced pressure to give a yellow oil, which was purified
by column chromatography using a mixture of hexane/EtOAc (4:1) to
afford 141 mg (0.19 mmol, 95%) of (þ)-18c. [R]2D5 þ 18.6 (c 0.6,
CHCl3); IR (film): υ = 3305, 3088, 3063, 3030, 2926, 2854, 1549, 1097,
1070, 696 cmꢀ1; 1H NMR (δ, 500 MHz, CDCl3): 0.88 (t, 3H, J = 7.2
Hz), 1.10ꢀ1.35 (m, 14H), 2.86ꢀ3.18 (m, 2H), 3.30ꢀ3.45 (m, 1H),
3.47ꢀ3.55 (m, 1H), 3.58 (t, 1H, J = 8.7 Hz), 3.89 (t, 1H, J = 9.8 Hz), 3.94
(t, 1H, J = 9.4 Hz), 4.18ꢀ4.28 (m, 1H), 4.62ꢀ5.00 (m, 8H), 6.40ꢀ6.90
(br s, 1H), 7.26ꢀ7.43 (m, 20H); 13C NMR (δ, 100 MHz, CDCl3): 14.2,
22.8, 27.0, 28.7, 29.3, 29.5, 31.9, 45.6, 56.9, 68.7, 73.0, 75.9, 76.0, 80.4,
81.1, 83.7, 127.7ꢀ128.7, 137.6, 138.4, 138.6, 182.8. HRMS calcd for
C44H57N2O5S: 725.3988 [M þ H]þ. Found: 725.4008.
’ ASSOCIATED CONTENT
S
Supporting Information. Complete refs 6, 16a, and 19
b
general experimental protocols, structural features, and com-
pound characterization data for all new compounds as well as
experimental procedure for glycosidase inhibition assays, GCS
inhibition in cell homogenates, inhibition of GCase (imiglucerase
and wt human fibroblasts), cytotoxicity, stabilization ratios after
thermal denaturation of recombinant GCase (imiglucerase), and
the effects of compounds on L444P or N370S GCase activity in
GD lymphoblasts. This material is available free of charge via the
Internet at http://pubs.acs.org.
General Procedure for the Preparation of Bicyclic Isoureas
and Guanidines: Synthesis of (ꢀ)-22c as a Representative
Example. Yellow mercury oxide (II) (117 mg, 0.54 mmol) was added
in one portion to a solution of the thiourea (þ)-18c (130 mg, 0.18
mmol) in toluene (15 mL). The reaction mixture was stirred at 115 °C
for 24 h under argon atmosphere. After cooling, the mixture was filtered
through Celite, and the plug of Celite was washed several times with
CH2Cl2. The filtrate and washings were combined, evaporated, and
purified by flash chromatography (2:1 to 1:2 hexane/EtOAc gradient) to
afford 112 mg (0.16 mmol, 89%) of (ꢀ)-22c. [R]2D5 ꢀ 2.4 (c 0.8,
CHCl3); IR (film): υ = 3330, 3086, 3063, 3030, 2926, 2854, 1718, 1671,
’ AUTHOR INFORMATION
Corresponding Author
amadeu.llebaria@cid.csic.es
1
1454, 1092, 1069, 1028, 696 cmꢀ1; H NMR (δ, 500 MHz, CDCl3):
0.89 (t, 3H, J = 7.0 Hz, CH3), 1.20ꢀ1.34 (m, 12H, 6xCH2), 1.41ꢀ1.51
(m, 2H, CH2), 3.14 (t, 2H, J = 7.2 Hz, NCH2), 3.73 (t, 1H, J = 5.2 Hz,
H-5), 3.87 (t, 1H, J = 5.0 Hz, H-4), 3.89 (dd, 1H, J = 4.9, 8.2 Hz, H-6),
4.01 (dd, 1H, J = 3.4, 8.2 Hz, H-7), 4.22 (dd, 1H, J = 4.6, 8.7 Hz, H-3a),
’ ACKNOWLEDGMENT
This work was supported by the Spanish MICINN (Project
CTQ2008-01426/BQU), CSIC, and Generalitat de Catalunya
(Grant 2009-SGR-1072). A.T. is grateful to MICINN for a
predoctoral fellowship. The authors thank Dr. J. Casas and
Dr. J.M. García Fernꢁandez for helpful discussions, E. Dalmau
for HRMS analysis, Dr. M. Egido-Gabꢁas, N. Guillem, and
L. Planas for experimental contributions, G. Reinkensmeier for
excellent technical assistance, and Genzyme Corp. for a generous
supply of imiglucerase.
4.52ꢀ4.82 (m, 9H, H-7a, 4xPhCH2O), 7.25ꢀ7.41 (m, 20H, Ph); 13
C
NMR (δ, 100 MHz, CDCl3): 14.2 (CH3), 22.8 (CH2), 26.9 (CH2),
29.37 (CH2), 29.40 (CH2), 29.6 (CH2), 29.8 (CH2), 32.0 (CH2), 43.1
(NCH2), 67.3 (C3a), 72.7 (PhCH2O), 72.8 (PhCH2O), 73.2
(PhCH2O), 73.7 (PhCH2O), 77.1 (C7), 78.7 (C7a), 79.7 (C4), 81.0
(C6), 82.5 (C5), 127.6ꢀ128.5 (CHar), 138.3 (Car), 138.57 (Car),
138.59 (Car), 160.8 (C2). HRMS calcd for C44H55N2O5: 691.4111
[M þ H]þ. Found: 691.4140.
General Procedure for O-Debenzylation with Boron
Trichloride: Synthesis of (þ)-28c and (þ)-4c as a Represen-
tative Example. A solution of the starting benzylated bicyclic isourea
(ꢀ)-22c (112 mg, 0.16 mmol) in anhydrous CH2Cl2 (10 mL)
at ꢀ78 °C was treated with 1 M BCl3 solution in heptane (1.6 mL,
1.6 mmol, 2.5 equiv for O-benzyl group). The reaction was stirred 2 h
at ꢀ78 °C and 5 h at ꢀ40 °C. The mixture was then cooled to ꢀ78 °C
and quenched with methanol (2 mL), and the solvents were removed
under reduced pressure. Purification by flash chromatography (20:1 to
10:1 CH2Cl2/MeOH gradient) provided (þ)-28c (10 mg, 0.02 mmol,
14%), followed by (þ)-4c (36 mg, 0.11 mmol, 68%).
’ REFERENCES
(1) Futerman, A. H.; van Meer, G. Nat. Rev. Mol. Cell Biol. 2004,
5, 554–565.
(2) Brady, R. O. Baillieres Clin. Haematol. 1997, 10, 621–634.
(3) Zhao, H.; Grabowski, G. A. Cell. Mol. Life Sci. 2002, 59, 694–707.
(4) Sawkar, A. R.; D’Haeze, W.; Kelly, J. W. Cell. Mol. Life Sci. 2006,
63, 1179–1192.
(5) (a) 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. N. Engl. J. Med. 1991, 324, 1464–1470. (b) Brady,
5482
dx.doi.org/10.1021/ja111480z |J. Am. Chem. Soc. 2011, 133, 5474–5484

Journal of the American Chemical Society
ARTICLE
R. O. Annu. Rev. Med. 2006, 57, 283–296. (c) Aerts, J. M.; Yasothan, U.;
Kirkpatrick, P. Nat. Rev. Drug Discovery 2010, 9, 837–838.
(6) Cox, T. M.; et al. J. Inherit. Metab. Dis. 2003, 26, 513–526.
(7) Weinreb, N. J.; Charrow, J.; Andersson, H. C.; Kaplan, P.;
Kolodny, E. H.; Mistry, P.; Pastores, G.; Rosenbloom, B. E.; Scott,
C. R.; Wappner, R. S.; Zimran, A. Am. J. Med. 2002, 113, 112–119.
(8) (a) Aoki, M.; Takahashi, Y.; Miwa, Y.; Iida, S.; Sukegawa, K.;
Horai, T.; Orii, T.; Kondo, N. Eur. J. Pediatr. 2001, 160, 63–64. (b)
Prows, C. A.; Sanchez, N.; Daugherty, C.; Grabowski, G. A. Am. J. Med.
Genet. 1997, 71, 16–21.
(9) (a) Cox, T.; Lachmann, R.; Hollak, C.; Aerts, J.; van Weely, S.;
Hrebicek, M.; Platt, F.; Butters, T. D.; Dwek, R.; Moyses, C.; Gow, I.;
Elstein, D.; Zimran, A. Lancet 2000, 355, 1481–1485. (b) Elstein, D.;
Hollak, C.; Aerts, J. M.; van Weely, S.; Maas, M.; Cox, T. M.; Lachmann,
R. H.; Hrebicek, M.; Platt, F. M.; Butters, T. D.; Dwek, R. A.; Zimran, A.
J. Inherit. Metab. Dis. 2004, 27, 757–766. (c) Zimran, A.; Elstein, D.
Philos. Trans. R. Soc. Lond., B: Biol. Sci. 2003, 358, 961–966.
(10) Schiffmann, R.; Fitzgibbon, E. J.; Harris, C.; DeVile, C.; Davies,
E. H.; Abel, L.; van Schaik, I. N.; Benko, W.; Timmons, M.; Ries, M.;
Vellodi, A. Ann. Neurol. 2008, 64, 514–522.
(28) Heck, M.-P.; Vincent, S. P.; Murray, B. W.; Bellamy, F.; Wong,
C.-H.; Mioskowski, C. J. Am. Chem. Soc. 2004, 126, 1971–1979.
(29) Egido-Gabꢁas, M.; Canals, D.; Casas, J.; Llebaria, A.; Delgado, A.
ChemMedChem 2007, 2, 992–994.
(30) (a) Zhu, X.; Sheth, K. A.; Li, S.; Chang, H. H.; Fan, J. Q. Angew.
Chem., Int. Ed. 2005, 44, 7450–7453. (b) Ogawa, S.; Kobayashi, Y.;
Kabayama, K.; Jimbo, M.; Inokuchi, J. Bioorg. Med. Chem. 1998,
6, 1955–1962. (c) Aguilar, M.; Gloster, T. M.; García-Moreno, M. I.;
Ortiz Mellet, C.; Davies, G. J.; Llebaria, A.; Casas, J.; Egido-Gabꢁas, M.;
García Fernꢁandez, J. M. ChemBioChem 2008, 9, 2612–2618. (d) 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. J. Biol. Chem.
1998, 273, 26522–26527.
(31) Chang, H. H.; Asano, N.; Ishii, S.; Ichikawa, Y.; Fan, J. Q. FEBS
J. 2006, 273, 4082–4092.
(32) The N-nonyl derivatives 1cꢀ4c were synthesized in both
racemic and enantiomerically pure forms.
(33) For a review, see: Berecibar, A.; Grandjean, C.; Siriwardena, A.
Chem. Rev. 1999, 99, 779–844.
(34) Grabowski, G. A. Genet. Test. 1997, 1, 5–12.
(11) (a) Morello, J.-P.; Bouvier, M.; Pet€aj€a-Repo, U. E.; Bichet, D. G.
Trends Pharmacol. Sci. 2000, 21, 466–469. (b) Fan, J. Q. Trends
Pharmacol. Sci. 2003, 24, 355–360. (c) Yu, Z.; Sawkar, A. R.; Kelly,
J. W. FEBS J. 2007, 274, 4944–4950. (d) Butters, T. D. Expert Opin.
Pharmacother. 2007, 8, 427–435. (e) Suzuki, Y.; Ogawa, S.; Sakakibara,
Y. Perspect. Med. Chem. 2009, 3, 7–19.
(35) Gonzꢁalez-Bulnes, P.; Casas, J.; Delgado, A.; Llebaria, A. Carbo-
hydr. Res. 2007, 342, 1947–1952.
(36) Egido-Gabꢁas, M.; Serrano, P.; Casas, J.; Llebaria, A.; Delgado,
A. Org. Biomol. Chem. 2005, 3, 1195–1201.
(37) Serrano, P.; Llebaria, A.; Delgado, A. J. Org. Chem. 2005,
70, 7829–7840.
(12) Parenti, G. EMBO Mol. Med. 2009, 1, 268–279.
(38) Harrak, Y.; Barra, C. M.; Bedia, C.; Delgado, A.; Castano, A. R.;
Llebaria, A. ChemMedChem 2009, 4, 1608–1613.
(39) Synthesis of 10 in racemic form: see ref 37. See Supporting
Information for the synthesis of enantiopure 10.
(40) (a) Ogawa, S.; Mori, M.; Takeuchi, G.; Doi, F.; Watanabe, M.;
Sakata, Y. Bioorg. Med. Chem. Lett. 2002, 12, 2811–2814. (b) Maya, I.;
Lꢁopez, O.; Fernꢁandez-Bola~nos, J.; Robina, I.; Fuentes, J. Tetrahedron
Lett. 2001, 42, 5413–5416. (c) Uchida, C.; Yamagishi, T.; Kitahashi, H.;
Iwaisaki, Y.; Ogawa, S. Bioorg. Med. Chem. 1995, 3, 1605–1624.
(41) Synthesis of 14 in racemic form: Azev, V. N.; D’Alarcao, M.
J. Org. Chem. 2004, 69, 4839–4842. See Supporting Information for the
synthesis of enantiopure 14.
(13) Sawkar, A. R.; Cheng, W. C.; Beutler, E.; Wong, C. H.; Balch,
W. E.; Kelly, J. W. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 15428–15433.
(14) Yu, L.; Ikeda, K.; Kato, A.; Adachi, I.; Godin, G.; Compain, P.;
Martin, O.; Asano, N. Bioorg. Med. Chem. 2006, 14, 7736–7744.
(15) Compain, P.; Martin, O. R.; Boucheron, C.; Godin, G.; Yu, L.;
Ikeda, K.; Asano, N. ChemBioChem 2006, 7, 1356–1359.
(16) (a) Lin, H.; et al. Biochim. Biophys. Acta 2004, 1689, 219–228.
(b) Luan, Z.; Li, L.; Ninomiya, H.; Ohno, K.; Ogawa, S.; Kubo, T.; Iida,
M.; Suzuki, Y. Blood Cells Mol. Dis. 2010, 44, 48–54.
(17) Yu, Z.; Sawkar, A. R.; Whalen, L. J.; Wong, C. H.; Kelly, J. W.
J. Med. Chem. 2007, 50, 94–100.
(18) Steet, R. A.; Chung, S.; Wustman, B.; Powe, A.; Do, H.;
Kornfeld, S. A. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 13813–13818.
(19) Khanna, R.; et al. FEBS J. 2010, 277, 1618–1638.
(20) Wang, G.-N.; Reinkensmeier, G.; Zhang, S.-W.; Zhou, J.;
Zhang, L.-R.; Zhang, L.-H.; Butters, T. D.; Ye, X.-S. J. Med. Chem.
2009, 52, 3146–3149.
(21) (a) Luan, Z.; Higaki, K.; Aguilar-Moncayo, M.; Ninomiya, H.;
Ohno, K.; García-Moreno, M. I.; Ortiz Mellet, C.; García Fernꢁandez,
J. M.; Suzuki, Y. ChemBioChem 2009, 10, 2780–2792. (b) Luan, Z.;
Higaki, K.; Aguilar-Moncayo, M.; Li, L.; Ninomiya, H.; Nanba, E.; Ohno,
K.; García-Moreno, M. I.; Ortiz Mellet, C.; García Fernꢁandez, J. M.;
Suzuki, Y. ChemBioChem 2010, 11, 2453–2464.
(42) Storch de Gracia, I.; Bobo, S.; Martín-Ortega, M. D.; Chiara,
J. L. Org. Lett. 1999, 1, 1705–1708.
(43) This assignment was supported by the large coupling constants
(10.3ꢀ12.2 Hz) observed between H3a/H7a, H3a/H4, and H7a/H7 in
compound 19b. This retention of configuration was confirmed by
comparison with the spectral data obtained for 19a,b and the desired
compound 22b, which was later synthesized from amino alcohol 12 via
the thiourea derivatives as depicted in Scheme 1.
(44) The position of the benzyl group in compounds 27aꢀd and
28aꢀc was established on the basis of heteronuclear correlation
measurements (HMBC and HSQC, see Supporting Information) on
selected representative examples.
(45) Desnick, R. J.; Schuchman, E. H. Nat. Rev. Genet. 2002, 3,
954–966.
(46) Ideal chaperone should be less effective as an inhibitor in the
acidic environment of the lysosome, where substrate accumulates but
would bind maximally at the neutral pH of the ER where optimal
pharmacological chaperone-activity is desired.
(47) Maegawa, G. H.; Tropak, M. B.; Buttner, J. D.; Rigat, B. A.;
Fuller, M.; Pandit, D.; Tang, L.; Kornhaber, G. J.; Hamuro, Y.; Clarke,
J. T.; Mahuran, D. J. J. Biol. Chem. 2009, 284, 23502–23516.
(48) Bendikov-Bar, I.; Ron, I.; Filocamo, M.; Horowitz, M. Blood
Cells, Mol., Dis. 2011, 46, 4–10.
(22) Compain, P.; Martin, O. R. From Synthesis Iminosugars to
Therapeutic Applications, Wiley: New York, 2007, p 467. Arjona, O.;
Gꢁomez, A. M.; Lꢁopez, J. C.; Plumet, J. Chem. Rev. 2007, 107, 1919–2036.
(23) Gloster, T. M.; Davies, G. J. Org. Biomol. Chem. 2010, 8,
305–320.
(24) (a) Hohenschutz, L. D.; Bell, E. A.; Jewess, P. J.; Leworthy,
D. P.; Pryce, R. J.; Arnold, E.; Clardy, J. Phytochemistry 1981,
20, 811–814. (b) Reymond, J. L.; Pinkerton, A. A.; Vogel, P. J. Org.
Chem. 1991, 56, 2128–2135.
(25) (a) Colegate, S. M.; Dorling, P. R.; Huxtable, C. R. Aust. J.
Chem. 1979, 32, 2257–2264. (b) Fleet, G. W. J.; Gough, M. J.; Smith,
P. W. Tetrahedron Lett. 1984, 25, 1853–1856.
(49) Díaz, L.; Bujons, J.; Casas, J.; Llebaria, A.; Delgado, A. J. Med.
Chem. 2010, 53, 5248–5255.
(26) Dr€ager, B. Nat. Prod. Rep. 2004, 21, 211–223.
(50) The racemic forms of compounds 3c and 4c were also tested in
N370S GD lymphoblasts.
(51) GCase activities in lymphoblasts were determined as described
in the literature with slight modifications. See ref 20 and Supporting
Information.
(27) (a) Granier, T.; Panday, N.; Vasella, A. Helv. Chim. Acta 1997,
80, 979–987. (b) Heightman, T. D.; Vasella, A. T. Angew. Chem., Int. Ed.
1999, 38, 750–770. (c) Gloster, T. M.; Roberts, S.; Perugino, G.; Rossi,
M.; Moracci, M.; Panday, N.; Terinek, M.; Vasella, A.; Davies, G. J.
Biochemistry 2006, 45, 11879–11884.
5483
dx.doi.org/10.1021/ja111480z |J. Am. Chem. Soc. 2011, 133, 5474–5484

Journal of the American Chemical Society
ARTICLE
(52) Beck, M. Expert Opin. Emerg. Drugs 2010, 15, 495–507.
(53) (a) Lei, K.; Ninomiya, H.; Suzuki, M.; Inoue, T.; Sawa, M.; Iida,
M.; Ida, H.; Eto, Y.; Ogawa, S.; Ohno, K.; Suzuki, Y. Biochim. Biophys.
Acta, Mol. Basis Dis. 2007, 1772, 587–596. (b) Alfonso, P.; Pampín, S.;
Estrada, J.; Rodríguez-Rey, J. C.; Giraldo, P.; Sancho, J.; Pocoví, M. Blood
Cells Mol. Dis. 2005, 35, 268–276. (c) Sawkar, A. R.; Adamski-Werner,
S. L.; Cheng, W. C.; Wong, C. H.; Beutler, E.; Zimmer, K. P.; Kelly, J. W.
Chem. Biol. 2005, 12, 1235–1244.
(54) Schueler, U. H.; Kolter, T.; Kaneski, C. R.; Zirzow, G. C.;
Sandhoff, K.; Brady, R. O. J. Inherit. Metab. Dis. 2004, 27, 649–658.
5484
dx.doi.org/10.1021/ja111480z |J. Am. Chem. Soc. 2011, 133, 5474–5484