Convenient synthesis and evaluation of enzyme inhibitory activity of several N-alkyl-, N-phenylalkyl, and cyclic isourea derivatives of 5a-Carba-α-DL-fucopyranosylamine

Article abstract of DOI:10.1016/S0960-894X(02)00627-3

Convenient synthesis and chemical modification of the potent α-L-fucosidase inhibitor, 5a-carba-α-DL-fucopyranosylamine (1), are described. Among seven N-substituted and three cyclic isourea derivatives newly prepared, the N-octyl derivative was found to be the strongest inhibitor of α-L-fucosidase (bovine kidney) more potent (K_i=0.016 μM) than deoxyfuconojirimycin (K_i=0.031 μM) with p-nitrophenyl-α-L-fucopyranoside as the substrate.

Full text of DOI:10.1016/S0960-894X(02)00627-3

Bioorganic & Medicinal ChemistryLetters 12 (2002) 2811–2814
Convenient Synthesis and Evaluation of Enzyme Inhibitory
Activity of Several N-Alkyl-, N-Phenylalkyl, and Cyclic Isourea
Derivatives of 5a-Carba-ꢀ-DL-fucopyranosylamine
Seiichiro Ogawa,* Makiko Mori, Goichi Takeuchi, Fuminao Doi,
Maiko Watanabe and Yuko Sakata
Department of Life Sciences and Informatics, Faculty of Science and Technology, Keio University, Hiyoshi,
Kohoku-ku, Yokohama 223-8522, Japan
Received 27 May2002; accepted 24 July2002
Abstract—Convenient synthesis and chemical modification of the potent a-l-fucosidase inhibitor, 5a-carba-a-dl-fucopyrano-
sylamine (1), are described. Among seven N-substituted and three cyclic isourea derivatives newly prepared, the N-octyl derivative
was found to be the strongest inhibitor of a-l-fucosidase (bovine kidney) more potent (Ki=0.016 mM) than deoxyfuconojirimycin
(Ki=0.031 mM) with p-nitrophenyl-a-l-fucopyranoside as the substrate.
# 2002 Elsevier Science Ltd. All rights reserved.
Mutant enzyme proteins have been found to be labile
and rapidlydegraded in somatic cells from patients with
lysosomal storage diseases. They are stabilized and
transported to lysosomes by competitive inhibitors of
low molecular weight (chemical chaperones), expressing
catalytic activity.1 This phenomenon was confirmed for
the mutant enzyme causing Fabry disease (a-galacto-
sidase deficiency).2 Veryrecently, N-octyl derivatives3 of
b-valienamine, an unsaturated 5a-carba-glucopyrano-
sylamine, were shown4 to elevate some mutant enzyme
activities in human patients with b-galactosidase or
b-glucosidase deficiency(Fig. 1).
compounds of this group mayfind application in the
near future in new therapeutic trial against genetic defi-
ciencydisorders.
Treatment of 2,3,4-tri-O-acetyl-6-bromo-6-deoxy-5a-
carba-b-dl-glucopyranosyl bromide9 (3) with metha-
nolic sodium methoxide at room temperature gave the
1,2:3,6-dianhydride,10 which was then treated without
isolation with NaH–benzyl bromide in DMF to give the
benzyl ether 4 (65%). Reaction of 4 with sodium azide,
followed byconventional tosylation, led to the azido
tosylate 5 (75%), hydrogenation of which with Raney
nickel in ethanol containing acetic anhydride gave the
N-acetyl derivative 6 (80%). Nucleophilic cleavage of
the anhydro ring followed by simultaneous O-debenzyl-
ation was effected bytreatment with hydrobromic acid
in acetic acid to give N-acetyl-3,4-di-O-acetyl-6-bromo-
6-deoxy-2-O-tosyl-5a-carba-b-dl-glucopyranosylamine 7
(80%), dehydrobromination of which with silver fluo-
ride in pyridine afforded the alkene 8 in 48% yield,
together with 7 (46%) unchanged. Hydrogenation of 8
with 5% Pd/C proceeded selectivelyto give the 6-deoxy
derivative 9 (90%), having an a-ido configuration.
Treatment of 9 with excess methanolic sodium methoxide
gave the 2,3-epoxide 10, which was hydrolyzed with
5a-Carba-a-l-fucopyranosylamine5 (1) and its b-anomer6
2 have been demonstrated to be verypotent and specific
inhibitors of a-l-fucosidase (bovine kidney), the effect
of the former being essentiallycomparable to those of
mammalian
a-l-fucosidase,
deoxyfuconojirimycin
(DFJ),7,8 the most powerful inhibitor identified. The
present paper describes alternative synthesis of 1 and its
chemical modification to prepare several N-alkyl,
phenylalkyl and cyclic isourea derivatives (20a–g and
22a–c), aimed at improving both the inhibitorypoten-
tial and biochemical features in vivo. We expect that
acid, followed byacetlyation, to afford the
N-acetyl-
2,3,4-tri-O-acetyl-5a-carba-a-dl-fucopyranosylamine11 11
(80%). Removal of the protecting groups with 4 M
*Corresponding author. Tel.: +81-45-566-1559; fax: +81-45-566-
1559; e-mail: ogawa@bio.keio.ac.jp
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00627-3

2812
S. Ogawa et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2811–2814
Figure 1.
Scheme 1. For convenience, the formulas depict onlyone enantiomer
of the respective racemates. Reagents and conditions: (a) MeONa (4.5
molar equiv), MeOH, rt; NaH, BnBr, DMF; (b) NaN3, (2 molar
equiv), 90% aq 2-methoxyethanol, 100 ꢁC; TsCl (4 molar equiv), pyr-
idine; (c) H2, RaneyNi, EtOH, Ac 2O; (d) 15% HBr–AcOH, 35 h,
85 ꢁC; (e) AgF (2 molar equiv), pyridine; (f) H2, PtO2, EtOH;
(g) MeONa, MeOH; (h) acetone, 10% H2SO4, 60 ꢁC; Ac2O, pyridine;
(i) 4 M HCl, reflux, Dowex 50 Wꢁ2 (H+) resin.
hydrochloric acid gave, after purification over a column
of Dowex 50 Wꢀ2 (H+) resin with aqueous 5%
ammonia, the free base11 1 quantitatively. Some 3-sub-
stituted derivatives of 1 could be obtained using inter-
mediates 8 or 9 (Scheme 1).
On the other hand, acetolysis of 5 with acetic acid–ace-
tic anhydride–concd sulfuric acid (40:20:1, v/v) at 85 ꢁC
overnight gave the triacetate 12 (48%). Treatment of 12
with excess sodium acetate resulted in preferential
inversion at C-2 and C-3 to produce, after acetylation,
the 5a-carba-b-altropyranosyl azide derivative 13
(70%). O-Deacetylation of 13 under Zemplen condi-
tions, followed by selective tosylation in pyridine, gave
the 6-tosylate 14 (75%). An elimination reaction of 14
with DBU in toluene proceeded readilyto give the
alkene 15 (65%), which was hydrogenated with 5% Pd/C
to give the undesired 2,3,4-tri-O-acetyl-6-deoxy-5a-
carba-b-dl-altropyranosyl azide12 16 (80%) as the sole
product. This sequence could, however, provide the
5-epimer of 1 (Scheme 2).
The N-alkyl derivatives13 20a, b, d of 1 were initially
prepared bylithium aluminum hydride reduction of the
corresponding amides 19a–c prepared from the pro-
tected amine 18 derived from the azide11 17. Alter-
natively, direct reductive alkylation of 1 with the
corresponding aldehydes proceeded smoothly on treat-
ment with sodium cyanoborohydride in THF under
slightlyacidic conditions, leading to the N-alkyl and
Scheme 2. For convenience, the formulas depict onlyone enantiomer
of the respective racemates. Reagents and conditions: (a) AcOH/
Ac2O/H2SO4 (40:20:1, v/v), 30 h, 85 ꢁC; (b) MeONa, MeOH; Ac2O,
pyridine; 10% H2SO4, acetone, 60 ꢁC; Ac2O, pyridine; (c) MeONa,
MeOH; TsCl (2 molar equiv), pyridine; Ac2O, pyridine; (d) DBU
(4 molar equiv), toluene, rt; (e) H2, PtO2, EtOH.

S. Ogawa et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2811–2814
2813
Table 1. Enzyme inhibitory activity of compounds 1, 2, 20a–g, and
22a–c
Compd
Inhibitoryactivity, IC
(Ki) (mM)
50
a-l-Fucosidase
(bovine kidney)
b-Galactosidase
(bovine liver)
1
2
9.3 (1.0)
25 (2.9)
NI
NI
NI
NI
NT
37
NT
NT
NT
NI
NI
NI
NI
20a
20b
20c
20d
20e
20f
20g
22a
22b
22c
DFJ
21 (0.18)
0.92 (0.074)
0.27 (0.048)
0.11 (0.016)
0.24 (0.11)
1.0 (0.069)
0.24 (0.032)
140
7.6
72
0.41 (0.031)
ꢃ4
NI, no inhibitoryactivitywas observed at 10
M; NT, not tested.
Biological Assay
Results of inhibition assay17 of the newlyprepared
derivatives of 1 toward a-l-fucosidase (bovine kidney)
and four other glycosidases are listed in Table 1. All
compounds, except for 1, 2, and the reference com-
pound DFJ, are racemic, and did not show anysig-
nificant activityagainst either a-glucosidase (baker’s
yeast and rat intestine), a-mannosidase (Jack beans), or
a-galactosidase (green coffee beans and rat liver). As
shown within an onlylimited scope of racemic modifi-
cations18 of the compounds tested, the inhibitoryactiv-
ityagainst a-l-fucosidase was dramaticallyincreased by
incorporation of alkyl and phenylalkyl groups into the
amino function of the parent 1. The changing the
N-ethyl on 1 to a N-nonyl group clearly improved the
inhibitorypotential, reaching a maximum with an octyl
chain: N-octyl-5a-carba-a-dl-fucopyranosylamine (20d)
was shown to possess verystrong and almost specific
19
inhibitoryactivity against a-l-fucosidase (bovine kid-
Scheme 3. For convenience, the formulas depict onlyone enantiomer
of the respective racemates. Reagents and conditions: (a) H2, Raney
Ni, EtOH; (b) RCOCl, pyridine; (c) LAH, THF, reflux; (d) 1 M HCl;
MeOH, MS-4A, RCHO, NaBH3CN, 20 h, rt; Ac2O, pyridine; 4 M
HCl, reflux; Dowex 50 Wꢀ2 (H+) resin, 5% aq NH3; (e) RNCS, 60%
aq EtOH; (f) HgO (yellow), acetone/EtOH (1:1).
ney), with p-nitrophenyl-a-l-fucopyranoside as sub-
strate. The l-enantiomer should surelybe evaluated as 5
times or more potent than DFJ, the strongest fucosidase
inhibitor reported7 so far. The results suggest that the
catalytic site of the enzyme can tolerate addition of
various sizes of aliphatic chain to the basic portion of
the inhibitors involved in binding. Concerning chemical
modification8,20,21 of DFJ, N-alkylation was earlier
found to result in rather a lowering of potential:8
N-methyl DFJ has about one-fifth the potency of the
parent. However, it is noteworthythat, the N-methyl
derivative has anyanti-(human immunodeficiencyvirus)
activity,8 although this is lacking with DFJ.
N-phenylalkyl derivatives 20c, e–g in acceptable 45–
70% yields (Scheme 3).
Furthermore, an attempt was made to improve inhibi-
torypotential byincorporation of a cyclic isourea func-
tion with variable N-substituents,14 therebyinducing
change of the electron distribution and somewhat flat-
tening the chair conformation without affecting the
configurational arrangement of the two hydroxyl and
methyl groups. Thus, the free amine 1 was treated with
the corresponding alkyl and aryl isothiocyanates in
aqueous 60% ethanol to yield the thioureas 21a–c
(ꢂ100%), which were treated with yellow mercuric
oxide15 in a mixture of acetone and ethanol to give the
corresponding cyclic isoureas16 22a–c (ꢂ100%).
The octyl chain seems to act as a structurally efficient
hydrophobic spacer, leading to proper electron-release
to the nitrogen atom for docking at the active site of the
enzyme. Interestingly, the results are in line with those
observed for inhibition of glucocerebrosidase by
N-alkyl-b-valienamines.3 Thus, incorporation of N-alkyl
functions could be clearlydemonstrated to influence the
activityof the inhibitors of this kind, suggesting that

2814
S. Ogawa et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2811–2814
there is much room for development of potent new
forms byfurther modification.
12. 1H NMR (300 MHz, CDCl3) (inter alia): d 4.79 (dd, 1H,
J3,4=3.0, J4,5=10.5 Hz, H-4), 4.38 [ddd, 1H, J1,2 2.7, J1,5a(ax)
12.3, J1,5a(eq)=2.4 Hz, H-1], 0.96 (d, 3H, J 6.4 Hz, Me).
1
13. The structures were confirmed on the basis of H NMR
1
Similar change of the potential has been observed for
three cyclic isourea derivatives 22a–c. The N-phenyl
derivative 22b showed good inhibitoryactivity, due to a
possible stacking effect of the spacer phenyl group,
indicating that a free hydroxyl group at C-2 is not
always a minimum motif for attaining potential.22
spectra: for example H NMR (300 MHz, CD3OD) data for
20b: d 3.64 (dd, 1H, J1,2=4.9, J2,3=10.2 Hz, H-2), 3.63 (br d,
1H, J3,4=3.2, J4,5=ꢂ0 Hz, H-4), 3.49 (dd, 1H, H-3), 2.86 (br
s, 1H, H-1), 2.53 and 2.28 (2 ddd, each 1H, J=7.5, Jgem=11.5
Hz, NHCH2CH2), 1.72 (m, 1H, H-5), 1.48 [ddd, 1H,
J1,5a(eq)=J5,5a(eq)=ꢂ3, J5agem=14.6 Hz, H-5a(eq)], 1.35–1.08
[m, 5H, H-5a(ax), (CH2)2CH3], 0.76 (d, 3H, J 7.0 Hz,
CHCH3), 0.71 (t, 3H, J=7.3 Hz, CH2CH3); for 20d: d 3.83
(dd, 1H, J1,2=4.8, J2,3=9.8 Hz, H-2), 3.71 (br s, 1H,
J3,4=J4,5=ꢂ0 Hz, H-4), 3.48 (br d, 1H, H-3), 3.40 (br s, 1H,
H-1), 2.91 (m, 1H, NCH2CH2), 1.74–1.44 (m, 3H, H-5, H-
5a,5a), 1.25–1.05 [m, 12H, (CH2)6CH3], 0.83 (br d, 3H, J=6.6
Hz, CHCH3), 0.68 (m, 3H, CH2CH3). It is noteworthythat the
chemical shifts of the signals due to H-1 and NCH2CH2 are
shown to change appreciablyaccording to the chain-length of
N-alkyl group: for example in 20a: d 3.10, and d 2.78 and 2.59,
respectively.
The present work documents that an appreciable num-
ber of the derivatives tested have stronger inhibitory
activitythan the parent 1. The N-octyl derivative 20d
has been shown to be a verystrong and specific a-l-
fucosidase inhibitor fullycomparable to DFJ and other
23
structurallyrelated natural and synthetic inhibitors.
The inhibitorypotentials of both enantiomers of 20c–e
clearlywarrant assessment.
14. Uchida, C.; Kimura, H.; Ogawa, S. Bioorg. Med. Chem.
1997, 5, 921, and reference cited therein.
Acknowledgements
15. Uchida, C.; Yamagishi, T.; Ogawa, S. J. Chem. Soc., Per-
kin Trans. 1 1994, 589.
The authors sincerelythank Drs. A. Takahashi and A.
Tomoda (Hokko Chemical Industry, Co. Ltd., Atsugi,
Japan) for the biological assay, and Prof. Yoshiyuki
Suzuki (International Universityof Health and Welfare,
Otawara, Japan) for helpful discussions.
16. The structures were confirmed with reference to the 1H
NMR spectra of the corresponding di-O-acetyl derivatives:
e.g., (1RS,2RS,3RS,4RS,6RS)-8-phenylamino-9-oxa-7-azabi-
cyclo[4.3.0]non-7-ene-2,3-diol diacetate derived from 22b: IR
1
(neat): n 1650 (C¼N), 1750 (ester) cmꢃ1: H NMR (300 MHz,
CDCl3): d 7.25–6.96 (m, 5H, Ph), 5.31 (t, 1H, J2,3=J3,4=3.4
Hz, H-3), 5.06 (dd, 1H, J1,2=7.3 Hz, H-2), 4.58 (m, 2H, H-1,
H-6), 2.47 [ddd, 1H, J4,5(eq)=6.1, J5(eq),6=2.9, J5gem=11.0 Hz,
H-5(eq)], 2.17 and 2.04 (2 s, each 3H, 2ꢀAc), 1.85 [ddd, 1H,
J4,5(ax)=9.5, J5(ax),6=5.1 Hz, H-5(ax)], 0.97 (d, 3H, J=6.6 Hz,
CMe).
References and Notes
1. Fan, J.-Q.; Ishii, S.; Asano, N.; Suzuki, Y. Nat. Med. 1999,
5, 112.
2. Fan, J.-Q.; Asano, N.; Suzuki, Y.; Ishii, S. Glycoconjugate
J. 1999, 16, S156.
3. Ogawa, S.; Ashiura, M.; Uchida, C.; Watanabe, S. Bioorg.
Med. Chem. Lett. 1996, 6, 929.
17. Biological assays were carried out in a standard manner
byDr. Akihiro Tomoda (Hokko Chemical Industr,y Co.
Ltd.), to whom our thanks are due.
4. Suzuki, Y.; Nanba, E.; Ohno, K.; Matsuda, J.; Ogawa, S.
Unpublished results.
5. Ogawa, S.; Maruyama, A.; Odagiri, T.; Yuasa, H.; Hashi-
moto, H. Eur. J. Org. Chem. 2001, 967, and references cited
therein.
6. Ogawa, S.; Watanabe, M.; Hisamatsu, S. Bioorg. Med.
Chem. Lett. 2002, 749.
7. Fleet, G. W. J.; Shaw, A. N.; Evans, S. V.; Fellows, L. E.
J. Chem. Soc., Chem. Commun. 1985, 841.
8. Winchester, B.; Barker, C.; Baines, S.; Jacob, G. S.; Nam-
goong, S. K.; Fleet, G. Biochem. J. 1990, 265, 277.
9. Ogawa, S.; Nakamoto, K.; Takahara, M.; Tannno, Y.;
Chida, N.; Suami, T. Bull. Chem. Soc. Jpn. 1979, 52, 1174.
10. Ogawa, S.; Oya, M.; Toyokuni, T.; Chida, N.; Suami, T.
Bull. Chem. Soc. Jpn. 1983, 56, 1441.
18. For convenience, in this work, we used readilyavailable
racemic compounds in order to evaluate preliminaryinhibi-
torypotential.
19. Moderate inhibitoryactivitytoward b-galactosidase seems
to be exerted bythe d-enantiomer, that is N-octyl-6-deoxy-5a-
carba-a-d-galactopyranosylamine.
20. Fleet, G. W.; Karpas, A.; Dwek, R. A.; Fellows, L. E.;
Tyms, A. S.; Petursson, S.; Namgoong, S. K.; Ramsden, N. G.;
Smith, P. W.; Son, J. C.; Wilson, F.; Witty, D. R.; Jacob, G.;
Rademacher, T. W. FEBS Lett. 1988, 237, 128.
21. Paulsen, H.; Matzke, M.; Orthen, B.; Nuck, R.; Reutter,
W. Liebigs Ann. Chem. 1990, 953.
6
22. The 2-deoxyand 2-epimeric derivatives of 1 proved not
to be inhibitors of a-l-fucosidase (bovine kidney): Ogawa, S.;
Watanabe, M. Unpublished results.
11. Ogawa, S.; Sekura, R.; Maruyama, A.; Yuasa, H.; Hashi-
moto, H. Eur. J. Org. Chem. 2000, 2089.
23. Asano, N.; Yasuda, K.; Kizu, H.; Kato, A.; Fan, J.-Q.;
Nash, R. J.; Fleet, G. W. Eur. J. Biochem. 2001, 268, 35.