Rational design and synthesis of highly potent β-glucocerebrosidase inhibitors

Article abstract of DOI:10.1002/anie.200502662

(Chemical Equation Presented) Highly potent human β-glucocerebrosidase inhibitors have been synthesized that appear to recognize both carbohydrate and hydrophobic binding sites. The most potent inhibitor, 6-nonyl isofagomine (see formula), shows an IC

Full text of DOI:10.1002/anie.200502662

Communications
Drug Design
DOI: 10.1002/anie.200502662
Rational Design and Synthesis of Highly Potent
b-Glucocerebrosidase Inhibitors
Xiaoxiang Zhu, Kamlesh A. Sheth, Shihong Li, Hui-
Hwa Chang, and Jian-Qiang Fan*
b-Glucocerebrosidase (GCase or acid b-glucosidase) is a
lysosomal hydrolase that catalyzes the hydrolytic cleavage of
glucose from glucosylceramide.[1] A deficiency in GCase
activity results in the progressive accumulation of glucosyl-
ceramide,a normal intermediate in the catabolism of globo-
sides and gangliosides,in the lysosomes of macrophages. This
deficiency leads to Gaucher disease,the most common
lysosomal storage disorder.[2] The disease severity is based
mainly on the level of the residual enzyme activity in the
tissues of the affected patients.[3] Enzyme-replacement ther-
apy is effective for type 1 Gaucher disease,the non-neuro-
nopathic form; however,this therapy is not available for
types 2 and 3 of Gaucher disease,the acute/subacute neuro-
nopathic forms,because of the difficulty in delivering the
replacement enzyme to the central nervous system.
We have reported that the residual enzyme activity can be
increased in lymphoblasts from patients with Fabry disease,
another lysosomal storage disease,when the cells were
incubated with potent inhibitors of the mutant enzyme at
subinhibitory concentrations.[4] The potent inhibitors serve as
active-site specific chaperones (ASSCs) to assist the efficient
folding process in the endoplasmic reticulum,thus accelerat-
ing the biosynthesis,processing,and maturation of the
mutated protein.[5] A correlation demonstrated that the
more potent inhibitors are the more effective ASSCs,because
they retain high affinity with the catalytic domain of the
[*] Dr. J.-Q. Fan
Department of Human Genetics
Mount Sinai School of Medicine
5th Avenue at 100th street, New York, NY 10029 (USA)
Fax : (+1)212-849-2508
E-mail: jian-qiang.fan@mssm.edu
Dr. X. Zhu, Dr. K. A. Sheth, Dr. S. Li, H.-H. Chang, Dr. J.-Q. Fan
Amicus Therapeutics, Inc.
Cranbury, NJ 08512 (USA)
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
7450
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7450 –7453

Angewandte
Chemie
enzyme.[6] Sawkar et al. reported that the addition of an
inhibitor of GCase to the culture medium of fibroblasts from
patients with Gaucher disease led to a twofold increase in the
activity of N370S GCase,which suggests that a potent
inhibitor of GCase could be of therapeutic interest.[7] There-
fore,an approach to the design and synthesis of potent
inhibitors of GCase was undertaken.
The natural substrates for GCase are N-acylsphingosyl-1-
O-b-d-glucosides,which have different fatty acid acyl and
sphingosyl moieties depending upon the tissue source.[2] It is,
therefore,reasonable to postulate that GCase may contain
two substrate-binding sites in the catalytic domain: one that
recognizes the glucosyl residue and the other that recognizes
the hydrophobic ceramide moiety. Transition-state mimics are
frequently potent inhibitors of the enzyme. Thus,our strategy
to design potent GCase inhibitors focused on molecules that
not only closely resemble both glucose and ceramide,but also
mimic the transition state of enzymatic glycosidic cleavage. It
is established that GCase cleaves the b-glycosidic bond to
release glucose with the retention of the anomeric config-
uration via a covalent glucosyl enzyme intermediate with
Glu340 acting as the nucleophile and Glu235 as the acidic/
basic species.[8] Known as a potent inhibitor of sweet-almond
b-glucosidase (Ki = 110 nm),isofagomine (IFG, 1) closely
resembles glucose with a nitrogen atom in the pseudo-
anomeric position,thus presumably acting as a mimic of the
glycosyl-enzyme intermediate.[9] The X-ray crystal structure
of GCase indicated the existence of an annulus of hydro-
phobic residues around the entrance to the glucose binding
site,[10] which could serve as a hydrophobic bind site. On the
basis of these findings,derivatization of IFG with a hydro-
phobic group may lead to a highly potent inhibitor. We
decided to synthesize a series of novel IFG analogues 2–6 with
a hydrophobic alkyl group at C6 and IFG analogues 7–8 with
an alkyl group at N1 in an attempt to demonstrate this
concept and discover novel GCase inhibitors.
Numerous syntheses of IFG and its derivatives have been
reported,[11] but the synthesis of 6-alkyl IFGs remains unex-
plored. As previously described,[11c] the synthesis of 1 was
difficult because of the lack of a suitably branched carbohy-
drate precursor. In this regard,6-alkyl IFGs with an addi-
tional chiral center might yet be more difficult to synthesize in
a straightforward fashion. We envisaged a concise synthetic
route for the preparation of 2–6 and disclose herein the
success of our strategy with the stereocontrolled introduction
of the 6-alkyl group by the addition of a Grignard reagent to
the nitrile group.
Compounds 2–6 were synthesized from benzyl a-l-
xylopyranoside (Scheme 1). Benzyl a-l-xylopyranoside was
treated with 2-methoxypropene and para-toluenesulfonic
acid (TsOH) in THF to afford 10 in 53% yield. Triflation of
the free 4-hydroxy group of 10 and subsequent treatment with
KCN in dimethylformamide (DMF) in the presence of
[18]crown-6 led to nitrile 11 in 80% yield (two steps from
10). The addition of n-C4H9MgCl to 11,followed by reduction
with NaBH4,stereoselectively afforded a single stereoisomer
12 in 74% yield.[12] The same procedure was used for the
synthesis of 13–16,and the yield ranged from 65 to 74%.
Catalytic hydrogenation of amine 12 in the presence of HCl
=
Scheme 1. Synthesis of IFG (1) and IFG derivatives 2–8. a) CH2
C(OMe)Me, TsOH·H2O, THF, 1.5 h, 08C (53%); b) Tf2O, pyridine,
CH2Cl2, 2 h, ꢀ78!08C; then KCN, 18[crown]-6, DMF, 16 h, RT (80%);
c) RMgX, Et2O, 2 h, RT; then NaBH4, overnight, RT (65–74%); d) H2,
20% Pd(OH)2/C, AcOH, MeOH, 50 psi, overnight, RT; then 1n HCl
(52–81%); e) H2, 20% Pd(OH)2/C, HCl (conc.), MeOH, RT (81%);
f) aldehyde, NaBH3CN, MeOH (79–83%). Tf=trifluoromethanesul-
fonyl.
over 20% Pd(OH)2 on charcoal at atmospheric pressure gave
2 in 52% yield. Alternatively,amine 16 was hydrogenated in
the presence of AcOH over 20% Pd(OH)2 on charcoal under
50 psi,followed by acid hydrolysis of the protective group to
afford 6 in 81% yield. The same procedure was successfully
applied to 13–15 to give 3–5. The hydrogenation of 11 over
20% Pd(OH)2 on charcoal led to debenzylation,intramolec-
ular cyclization,and concurrent deacetonation in one step to
give 1 in 81% yield. In synopsis,although a number of
methods have been reported for the synthesis of 1,[11] the
present method provides an efficient and short synthetic route
for the synthesis of 1 in a total of four steps and an overall
yield of 34%. In addition, N-alkyl IFGs 7 and 8 were prepared
according to the previously reported procedure.[13]
The configuration of the newly formed stereogenic center
in 12 was established based on 6-butyl IFG (2),whose
stereochemistry was determined by 2D NMR spectroscopic
analyses. The NOESY spectrum of 2 showed significant
correlations of H6 with both H4 and H2a,and the coupling
constant J(5,6) was observed to be 10.5 Hz, thus confirming
that H6 is situated at an axial orientation. Thus,the
configuration of the new asymmetric center in 12 can be
assigned as S (see the Supporting Information for the COSY,
TOCSY,and NOESY spectra of 2). Analysis of the 1H NMR
spectra of 2–6 indicates that they all possess the same
configuration at the asymmetric C6 position.
The high stereoselectivity observed above can be
explained by a chelation mechanism (Scheme 2). After
addition of the Grignard reagent to the nitrile moiety,the
magnesium atom of the magnesioimine could be chelated to
the oxygen atom on the pyranose ring,thus resulting in the
formation of a six-membered cyclic intermediate 17. Further
Angew. Chem. Int. Ed. 2005, 44, 7450 –7453
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7451

Communications
greatly decreased the inhibitory activity (IC50 ꢁ 44 mm for 7
and 8,respectively). The potent inhibitory activity of IFG is
most likely attributed to the salt bridge between protonated 1
and the catalytic carboxylate of GCase.[9] The alkyl group
attached to N1 might interfere with the formation of the salt
bridge,thus resulting in less potent inhibitors. Incorporation
of the 6-butyl group into 1 generated 6-butyl IFG (2),which
did not exhibit a better inhibitory activity on GCase than 1.
However,6-hexyl IFG ( 3) was apparently more potent than 1
with a 13-fold increase in potency. Further extension of the
length of the alkyl chain improved the potency accordingly
(4–6). The most potent inhibitor 6-nonyl IFG (6) displayed a
remarkable IC50 value of 0.6 nm,which is 93-fold more potent
relative to 1. These data clearly indicate that a long alkyl chain
(greater than four carbon atoms) is required to improve
potency and a longer chain provides a higher potency,thus
suggesting that there is a hydrophobic interaction between
the alkyl group and enzyme and that an appropriate alkyl
chain length is needed to bind the hydrophobic domain. In
recent years,despite many efforts being directed towards the
synthesis and study of IFG analogues,only weaker inhibitors
have resulted from these investigations.[9,11] Based on the
concept we have proposed and the x-ray crystal structure of
GCase,we discovered highly potent GCase inhibitors 3–6,
which are between 13- and 93-fold more potent than the
parent inhibitor 1. The proposed binding of 6 with GCase is
shown in Figure 2. The structural features of the potent
inhibitors as described above can be applied further to find
potent and selective inhibitors of GCase with an improved
biopharmaceutical profile.
Scheme 2. Proposed mechanism for the stereoselective formation of
amine 18.
reduction of the ketimine function with NaBH4 prefers an
attack at the Re face to yield amine 18 with a S configuration.
The inhibitory activities of all compounds against human
GCase was determined with 4-methylumbelliferyl b-glucoside
(4MU-b-Glc) and are summarized in Table 1. To determine
Table 1: Inhibitory activity of 1 and derivatives 2–8 against GCase.
Inhibitors
IC50 [nm][a]
Ki [nm]
IFG (1)
56
160
4.2
1.8
0.8
0.6
44000
>100000
25[b]
6-butyl IFG (2)
6-hexyl IFG (3)
6-heptyl IFG (4)
6-octyl IFG (5)
6-nonyl IFG (6)
N-butyl IFG (7)
N-nonyl IFG (8)
120[b]
[c]
[c]
[c]
[c]
n.d.
n.d.
[a] All inhibitory activities were determined with 4MU-b-Glcat 3 m m.
[b] The kinetic data were fit to a double-reciprocal plot and replotted to
Kmapp versus [I]. [c] Data do not fit to a competitive-inhibition model by
using lineal and nonlineal models (Prism GraphPad). n.d.=not
determined.
the Ki values,a Lineweaver–Burk plot was used to calculate
the Kmapp in the presence of an inhibitor at various concen-
trations and replotted against the concentrations of inhibitor
(Figure 1). IFG (1) was a potent inhibitor against GCase with
IC50 and Ki values of 56 and 25 nm,respectively. In contrast,
modification of the imino group by a hydrophobic group
Figure 2. Proposed binding of 6 with human GCase.
The inhibition mode for 1 and 2 is competitive,with Ki
values of 25 and 120 nm, respectively,using 4MU- b-Glc as the
variable substrate (Table 1). The inhibition mode for the 6-
alkyl IFG derivatives with longer chains was more compli-
cated than being purely competitive but showed character-
istics of a mixed-type inhibition. This difference may be
explained on the basis that these inhibitors bind the enzyme
through two distant binding sites: a carbohydrate binding site
and a hydrophobic binding site (Figure 2). Whereas both
inhibitory binding sites are likely to be competitive,the strong
combination of both acts in a manner to provide a non-
competitive-like inhibitory mechanism in some of the most
potent 6-alkyl IFGs.
Figure 1. Lineweaver–Burk plots of 1 and 6-butyl IFG (2) for the
inhibition of human GCase. The increasing concentrations of substrate
were used to determine the Kmapp and Ki values and the data were
In conclusion,we have demonstrated an efficient and
straightforward methodology for the synthesis of 6-alkyl
IFGs,which represent a new class of highly potent inhibitors
of human GCase. These inhibitors appear to utilize recog-
*
*
plotted as 1/v versus 1/[S]. a) Concentrations of 1 were 0 ( ), 20 ( ),
^
^
*
*
50 ( ), and 150 nm ( ). b) Concentrations of 2 were 0 ( ), 100 ( ),
^
^
200 ( ), and 400 nm ( ).
7452
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7450 –7453

Angewandte
Chemie
nition domains for both a carbohydrate and a hydrophobic
binding site. As a result,the most potent GCase inhibitor 6-
nonyl IFG (6) shows an IC50 value at subnanomolar concen-
trations. We have used these compounds as ASSCs for the
rescue of misfolded mutant GCase in cells from Gaucher
patients (results to be published elsewhere). As a conse-
quence,they may be ideal candidates for further development
as small-molecule drugs for the treatment of Gaucher disease,
in particular for the neuronopathic disease form.
Received: July 29,2005
Published online: October 18,2005
Keywords: biological activity · drug design · enzymes ·
.
inhibitors · natural products
[1] R. O. Brady,J. N. Kanfer,D. Shapiro, Biochem. Biophys. Res.
Commun. 1965,18,221 – 225.
[2] E. Beutler,G. A. Grabowski in The Metabolic and Molecular
Bases of Inherited Disease,8th ed. (Eds.: C. R. Scriver,A. L.
Beaudet,W. S. Sly,D. Valle),McGraw-Hill,New York,
p. 3635.
2001,
[3] G. A. Grabowski in Adv. Hum. Genet., Vol. 21 (Eds.: H. Harris,
K. Hirschhorn),Plenum,New York, 1993,p. 377.
[4] J.-Q. Fan,S. Ishii,N. Asano,Y. Suzuki, Nat. Med. 1999, 5,112 –
115.
[5] J.-Q. Fan, Trends Pharmacol. Sci. 2003, 24,355 – 360.
[6] N. Asano,S. Ishii,H. Kizu,K. Ikeda,K. K. Yasuda,A. Kato,
O. R. Martin,J.-Q. Fan, Eur. J. Biochem. 2000, 267,4179 – 4186.
[7] a) A. R. Sawkar,W.-C. Cheng,E. Beutler,C.-H. Wong,W. E.
Balch,J. W. Kelly, Proc. Natl. Acad. Sci. USA 2002, 99,15428 –
15433; b) H. Lin,Y. Sugimoto,Y. Ohsaki,H. Ninomiya,A. Oka,
M. Taniguchi,H. Ida,Y. Eto,S. Ogawa,Y. Matsuzaki,M. Sawa,
T. Inoue,K. Higaki,E. Nanba,K. Ohno,Y. Suzuki,
Biochim.
Biophys. Acta 2004, 1689,219 – 228; c) S. Ogawa,Y. Kobayashi,
K. Kabayama,M. Jimbo,J. Inokuchi, Bioorg. Med. Chem. 1998,
6,1955 – 1962.
[8] a) S. Fabrega,P. Durand,P. Codogno,C. Bauvy,C. Delomenie,B.
Henrissat,B. M. Martin,C. McKinney,E. I. Ginns,J. P. Mornon,
P. Lehn, Glycobiology 2000, 10,1217 – 1224; b) S. Fabrega,P.
Durand,J. P. Mornon,P. Lehn, J. Soc. Biol. 2002, 196,151 – 160;
c) G. Davies,B. Henrissat, Structure 1995, 3,853 – 859.
[9] V. H. Lillelund,H. H. Jensen,X. Liang,M. Bols,
Chem. Rev.
2002, 102,515 – 553,and references therein.
[10] H. Dvir,M. Harel,A. A. McCarthy,L. Toker,I. Silman,A. H.
Futerman, EMBORep. 2003, 4, 1 – 6.
[11] a) T. M. Jespersen,W. Dong,M. R. Sierks,T. Skrydstrup,I.
Jundt,M. Bols, Angew. Chem. 1994, 106,1858 – 1861; Angew.
Chem. Int. Ed. Engl. 1994, 33,1778 – 1779; b) Y. Ichikawa,Y.
Igarashi,M. Ichikawa,Y. Suhara, J. Am. Chem. Soc. 1998, 120,
3007 – 3018; c) J. Andersch,M. Bols, Chem. Eur. J. 2001, 7,3744 –
3747,and references therein; d) P. Jakobsen,J. M. Lundbeck,M.
Kristiansen,J. Breinholt,H. Demuth,J. Pawlas,M. Torres,B.
Anderson,N. Westergaard,K. Lundgren,N. Asano,
Bioorg.
Med. Chem. 2001, 9,733 – 744; e) W. M. Best,J. M. Macdonald,
B. W. Skelton,R. V. Stick,D. M. Tilbrook,A. H. White, Can. J.
Chem. 2002, 80,857 – 865.
[12] HPLC and TLC analysis from the crude mixture and purified
sample detected the formation of 12; the presence of its
diastereoisomer was not observed.
[13] H. R. Mellor,J. Nolan,L. Pickering,M. R. Wormald,F. M. Platt,
R. A. Dwek,G. W. J. Fleet,T. D. Butters, Biochem. J. 2002, 366,
225 – 233.
Angew. Chem. Int. Ed. 2005, 44, 7450 –7453
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7453