Synthesis of the sialidase inhibitor siastatin B.

Article abstract of DOI:10.1021/ol0066680

[structure] The resolved piperidinecarboxylate (R)-7 was converted to siastatin B (1) by an efficient and stereoselective sequence that includes a bromo-beta-lactonization and an N-acyliminium azidation. Two analogues (3 and 4) of siastatin were also prepared.

Full text of DOI:10.1021/ol0066680

ORGANIC
LETTERS
2000
Vol. 2, No. 25
4037-4040
Synthesis of the Sialidase Inhibitor
Siastatin B
Spencer Knapp* and Dalian Zhao
Department of Chemistry; Rutgers - The State UniVersity of New Jersey,
610 Taylor Road, Piscataway, New Jersey 08854-8087
knapp@rutchem.rutgers.edu
Received September 28, 2000
ABSTRACT
The resolved piperidinecarboxylate (R)-7 was converted to siastatin B (1) by an efficient and stereoselective sequence that includes a bromo-
â-lactonization and an N-acyliminium azidation. Two analogues (3 and 4) of siastatin were also prepared.
Sialidases are enzymes that cleave N-acetylneuraminic acid
from the nonreducing ends of glycoconjugates and thereby
mediate a variety of cell surface recognition events.1 Sialidase
inhibitors have proven to be important tools for understand-
ing and controlling sialidases:2 the most prominent health-
related application is the development of the various sialidase
inhibitors (Zanamivir, Oseltamivir, RWJ-270201) that are
currently in use and in clinical trials for the treatment of
influenza virus infection.3 As with any infection, however,
the emergence of drug-resistant strains remains a concern.4
Siastatin B (1) is a broad spectrum sialidase inhibitor
isolated from a Streptomyces culture and characterized as
an unusual 6-acetamido-3-piperidinecarboxylate by Umeza-
wa and co-workers in 1974.5 The charge distribution in the
zwitterion resembles that in the N-acetylneuraminate oxo-
carbenium ion 2, the putative intermediate in the enzyme-
catalyzed reaction,6 and this may account for its effectiveness
in binding sialidases. A synthesis of 1 from L-ribose was
(1) Colman, P. M. Protein Sci. 1994, 3, 1687-1696. Taylor, G. Curr.
Opin. Struct. Biol. 1996, 6, 830-837, and references therein.
(2) El Ashry, E. S. H.; Rashed, N.; Shobier, A. H. S. Pharmazie 2000,
reported by Nishimura and co-workers in 1988,7 and in the
55, 251-262. von Itzstein, M.; Thomson, R. T. Top. Curr. Chem. 1997,
years since they have prepared a number of synthetic
analogues and modified siastatin derivatives with an aston-
186, 119-170. Luo, Y.; Li, S. C.; Li, Y. T.; Luo, M. J. Mol. Biol. 1999,
285, 323-332.
(3) Gubareva, L. V.; Kaiser, L.; Hayden, F. G. Lancet 2000, 355, 827-
835, and references therein.
(4) McKimm-Breschkin, J. L. AntiViral Res. 2000, 47, 1-17, and
(6) Yang; J. S.; Schenkman, S.; Horenstein, B. A. Biochemistry 2000,
39, 5902-5910. Horenstein, B. A.; Bruner, M. J. Am. Chem. Soc. 1998,
120, 1357-1362. Thomas, A.; Jourand, D.; Bret, C.; Amara, P.; Field, M.
J. J. Am. Chem. Soc. 1999, 121, 9693-9702.
references therein.
(5) Umezawa, H.; Aoyagi, T.; Komiyama, H.; Morishima, H.; Hamada,
M.; Takeuchi, T. J. Antibiot. 1974, 27, 963-969.
10.1021/ol0066680 CCC: $19.00 © 2000 American Chemical Society
Published on Web 11/17/2000

ishing array of biological and enzyme inhibitory activities.8
One modification, the trifluoroacetamido analogue 3, exhibits
inhibition of bovine liver â-glucuronidase (IC50 6.5 × 10-8
M)9 and effectively suppresses tumor cell metastasis in
mice.10
Our interest in the siastatin family of glycosidase inhibitors
stems from their unusual mixed aminal structures, their
diverse activities, the possibility of using them to improve
the understanding of glycosidase binding and mechanism,
and the challenge of devising a flexible synthesis from non-
carbohydrate starting material. We report efficient and
stereoselective syntheses of siastatin B (1, 13 steps) and its
trifluoroacetamido analogue (3, 14 steps) from the resolved
piperidine carboxylate 7 (Scheme 1). A racemic 3-C-benzyl
analogue 4 was also prepared.
already been described,11 the synthesis of siastatin reduces
to three parts: (1) an effective deconjugation/resolution
procedure, (2) the stereoselective functionalization of the
piperidine ring, including introduction of the unusual mixed
aminal, and then (3) careful deprotection. The deconjugation/
resolution procedure is shown in Scheme 1.
Hydrolysis of ethyl ester 6 led to a crystalline unsaturated
carboxylic acid, which was converted to its γ-extended
enolate by treatment with LDA. Quenching with methanolic
acetic acid gave the crystalline racemic â,γ-unsaturated
carboxylic acid (R,S)-7. Several attempts to quench this
enolate enantioselectively12 were unsuccessful, but this
remains a worthwhile objective. The deconjugated acid could
be conveniently resolved by combining it with (R)-(+)-R-
methylbenzylamine and then crystallizing the resulting salt
from ethyl acetate. A second crystallization followed by
breaking the salt with HCl afforded in 43% overall yield
the unsaturated acid (R)-7 with ee ) 87%. The absolute
configuration and minimum optical purity of (R)-7 were
established by converting it to an amide (8) of (S)-(-)-R-
Scheme 1
1
methylbenzylamine. The H NMR spectrum of 8 exhibits
signals for two diastereomers in the approximate ratio 12:1
(the dd signals of H-2 at 3.3-3.5 ppm are most easily
integrated). Hydrogenation of 8 gave the amide 9 as a 12:1
mixture of diastereomers, and the major diastereomer of 9
was independently prepared from ethyl (R)-nipecotate, which
is commercially available as its L-tartrate salt 10 (Scheme
1). Examination of the 1H NMR spectrum of 9 prepared from
10 revealed no trace of the (S)-diastereomer, suggesting that
no racemization takes place during the formation of the
amide. Furthermore, amide formation from the racemic acid
(R,S)-7 gave a 1:1 mixture of 8 and its diastereomer,
suggesting that preferential kinetic formation of the (R)-
diastereomer is not occurring.
Bromolactonization of unsaturated acid (R)-7 was envi-
sioned as an effective tool for initial functionalization at
C-4,5. Because literature methods for halocyclization to
â-lactones13 gave poor results, considerable effort was spent
in trying to find an effective halonium source and solvent
for this reaction. Prior conversion of (R)-7 to its tetra-n-
butylammonium carboxylate salt 11 allowed bromination
under homogeneous conditions and at low temperature
(Scheme 2). Thus, 11 was carefully dried and then treated
(9) Nishimura, Y.; Shitara, E.; Adachi, H.; Toyoshima, M.; Nakajima,
M.; Okami, Y.; Takeuchi, T. J. Org. Chem. 2000, 65, 2-11, and references
therein.
(10) Nishimura, Y.; Kudo, T.; Kondo, S.; Takeuchi, T. J. Antibiot. 1994,
47, 101-107.
(11) Krogsgaard-Larsen, P.; Jacobsen, P.; Brehm, L.; Larsen, J. Eur. J.
Med. Chem. 1980, 15, 529-535. Krogsgaard-Larsen, P.; Thyssen, K.;
Schumberg, K. Acta Chem. Scand. 1978, B32, 327-334. Krogsgaard-Larsen,
P.; Hjeds, H. Acta Chem. Scand. 1974, B28, 533-538. Allan, R. D.;
Johnston, A. R. Med. Res. ReV. 1983, 3, 91-118.
Given that an easy route to unsaturated piperidine ester 6
(Scheme 1) from the commercially available keto-ester 5 has
(12) For a recent leading reference, see: Vedejs, E.; Kruger, A. W.; Lee,
N.; Sakata, S. T.; Stec, M.; Suna, E. J. Am. Chem. Soc. 2000, 122, 4602-
4607.
(7) Nishimura, Y.; Wang, W.; Kondo, S.; Aoyagi, T.; Umezawa, H. J.
Am. Chem. Soc. 1988, 110, 7249-7250. Nishimura, Y.; Wang, W.; Kudo,
T.; Kondo, S. Bull. Chem. Soc. Jpn. 1992, 65, 978-986. This route proceeds
in about 24 steps and 11% overall yield from L-ribose.
(8) Shitara, E.; Nishimura, Y.; Kojima, F.; Takeuchi, T. Bioorg. Med.
Chem. 2000, 8, 343-352. Nishimura, Y. In Studies in Natural Products
Chemistry; Atta-ur-Rahman, Ed.; Elsevier: Amsterdam, 1995; Vol 16, pp
75-121. Nishimura, Y. Yuki Gosei Kagaku Kyokaishi 1997, 55, 142-151.
(13) Frederickson, M.; Grigg, R. Org. Prep. Proc. Int. 1997, 29, 35-
62. Barnett, W. E.; Needham, L. L. J. Org. Chem. 1975, 40, 2843-2844.
Ganem, B.; Holbert, G. W.; Weiss, L. B.; Ishizumi, K. J. Am. Chem. Soc.
1978, 100, 6483-6491. Barnett, W. E.; Sohn, W. H. J. Chem. Soc., Chem.
Commun. 1972, 472. Mead, K. T.; Park, M. Tetrahedron Lett. 1995, 36,
1205-1208. Cook, C.; Chung, Y. Arch. Pharm. Res. 1981, 4, 133-135.
Homsi, F.; Rousseau, G. J. Org. Chem. 1999, 64, 81-85.
4038
Org. Lett., Vol. 2, No. 25, 2000

as the ketal 15 by treating the crude oxidation mixture with
2,2-dimethoxypropane. The behavior of 14 under these
conditions is perhaps analogous to pyranose glycal reactiv-
ity,15 with the N-acylpiperidine nitrogen standing in for the
pyranose ring oxygen.16 Thus, 15 is analogous to a methyl
glycoside. On the basis of its 1H NMR spectrum (two BOC
rotamers, JH5-H6 ∼ 0 Hz, no H-2/H-3 trans-diaxial coupling),
we assign a chairlike conformation to 15 in which the C-6
methoxy prefers an axial position to avoid allylic 1,3-strain
with the N-BOC oxygens.
Scheme 2
Exchange of methoxy for azido was accomplished by
treating 15 with azidotrimethylsilane and boron trifluoride,
with azido entering from the less hindered R-face. Hydro-
genolytic reduction of the azide in the presence of acetic
anhydride led directly to the acetamido product 16. The
pivaloate was cleaved with tetra-n-butylammonium hydrox-
ide to expose the primary hydroxyl, which then was oxidized
to the carboxylic acid (17) with pyridium dichromate.17
Nearly quantitative deprotection of 17 was carried out in a
one-pot procedure by sequentially adding trifluoroacetic acid
(to remove the N-tert-butoxycarbonyl), adding water (to
remove the acetonide), concentrating, adding ethanolic
hydrochloric acid (to exchange the counterion), concentrating
again, and finally adding hexane to solidify the hydrochloride
1
of siastatin B (1‚HCl). The H and 13C NMR spectra, and
the optical rotation, [R]25 +52° (c ) 0.25, H2O), closely
D
match the values reported for the natural product.18 Although
the optical purity of late synthetic intermediates was not
checked, the small amount of unnatural enantiomer present
in the resolved acid (R)-7 had apparently been removed by
crystallization at two intermediate stages along the synthetic
route: the diol derived from 12 and the acetate derived from
13.
The trifluoroacetamido analogue 3 was synthesized by
modifying the siastatin route at the stage of the N-acetylation
(Scheme 3). Thus, hydrogenation of azide 18 gave a primary
Scheme 3
with bromine in dichloromethane solution at -78 °C,
resulting in a 71% isolated yield of the bromo-â-lactone 12.
IR analysis of the crude reaction mixture indicated exclusive
formation of the â-lactone (1840 cm-1) without detectable
γ-lactone. The lactone ring was reduced to the diol with
lithium borohydride (no ring closure to the epoxide occurred
under these conditions), and then the primary hydroxyl was
selectively protected as its pivaloate (13).
For dehydrobromination of 13, the derived acetate was
treated with 1,8-diazabicyclo[5.4.0]undec-7-ene, and then the
acetate was cleaved with ethanolic guanidine.14 The resulting
allylic alcohol 14, which is very acid-sensitive, is the
appropriate substrate for a hydroxyl-directed syn-epoxidation
at C-4,5. Several attempts to isolate a C-4,5 epoxide were
unsuccessful, as were attempts to trap the epoxide with
various azide nucleophiles. However, the product of metha-
nolic m-chloroperoxybenzoic acid oxidation could be trapped
(14) Kunesch, N.; Miet, C.; Poisson, J. Tetrahedron Lett. 1987, 28, 3569-
3572.
Org. Lett., Vol. 2, No. 25, 2000
4039

amine that was stable enough to survive until it could be
trifluoroacetylated. The resulting trifluoroaceamide 19 was
selectively deprotected at the pivaloate by treatment with
methanolic tetrabutylammonium hydroxide. The trifluoro-
acetamide, which might have been expected to cleave under
these conditions, likely survives because it is deprotonated
instead. Oxidation and deprotection as before afforded (F3)-
Scheme 4. Synthesis of 3-C-Benzylsiastatina
1
siastatin B as its hydrochloride salt 3, whose H NMR
spectrum and optical rotation, [R]25D +28° (c ) 0.25, H2O),
closely match the literature values.19
The second analogue, 3-C-benzylsiastatin (4), was actually
synthesized before 1 and 3 as a model for improving several
of the transformations needed for the synthesis of the latter
two. Unsaturated ester 6 was converted to its lithium
γ-extended enolate, which reacted with benzyl bromide to
afford the racemic ester 22 (Scheme 4). Treatment of 22 with
tetrabutylammonium hydroxide led to the tetrabutylammo-
nium salt of the corresponding carboxylate, which was
directly bromocyclized at low temperature to give the bromo
â-lactone 23. Exposure of 23 to the lithium salt of benzyl
alcohol caused alcoholysis of the â-lactone and subsequent
ring closure to the epoxide (24). Isomerization of 24 to allylic
alcohol 26 was accomplished by site-selective epoxide
opening with sodium phenylselenylate20 to give 25, followed
by acetylation, oxidation with spontaneous elimination of
PhSeOH, and finally deacetylation. Syn-epoxidation of 26
and subsequent treatment with 2,2-dimethoxypropane pro-
vided the ketal 27, and then azidation, reduction of the
azide,21 acetylation, hydrogenolysis of the benzyl ester, and
finally deprotection gave the 3-C-benzyl analogue 4. The
steps leading to 22, 23, 27, 28, 29, and 4 provided
information valuable for application to the synthesis of 1
and 3. In contrast to these routes, however, the route to 4
did not require reduction of the ester as there was no danger
of elimination at C-3,4. Although it is structurally closely
related to siastatin B, the 3-C-benzyl analogue 4 showed no
inhibitory activity when evaluated against the commercial
sialidase from Salmonella tymphimurium.
a Reagents and conditions: (a) LDA, HMPA, -78 to 0 °C, then
PhCH2Br, -78 to 23 °C; (b) Bu4NOH, MeOH, 23 °C, then Br2,
CH2Cl2, -78 °C; (c) PhCH2OH, n-BuLi, THF, -40 °C; (d)
PhSeSePh, NaBH4, EtOH, 23 °C; (e) Ac2O, pyridine, DMAP, THF,
40 °C; (f) 30% H2O2, i-Pr2NH, THF, 0 to 23 °C; (g) Na2CO3,
MeOH; (h) dimethyldioxirane, aqueous acetone; (i) Me2C(OMe)2,
TsOH; (j) Me3SiN3, BF3‚OEt2, CH2Cl2, -40 °C; (k) H2, Lindlar
catalyst, EtOH; (l) Ac2O, pyridine, DMAP, THF; (m) H2, 10% Pd-
C, EtOH; (n) TFA, H2O, then HCl, EtOH.
Acknowledgment. This work was supported by Merck
& Co. and SynChem Research, Inc. We are grateful to Dr.
Robert Reamer of Merck for assistance with NMR analysis
and to Prof. Benjamin A. Horenstein and Jingsong Yang of
the University of Florida for enzymatic evaluation of 4.
(15) Knapp, S.; Lal, G. S.; Sahai, D. J. Org. Chem. 1986, 51, 380. Sweet,
F.; Brown, R. K. Can. J. Chem. 1966, 44, 1571.
(16) Review of N-acyliminum ion additions: Hiemstra, H.; Speckamp,
W. N. In ComprehensiVe Organic Synthesis: SelectiVity, Strategy, and
Efficiency in Modern Organic Synthesis; Trost, B. M., Ed.; Pergamon
Press: London; 1991; Vol. 2, pp 1047-1082.
(17) Corey, E. J.; Schmidt, G. Tetrahedron Lett. 1979, 20, 399-402.
(18) From ref 7, [R]20 +53° (c ) 0.25, H2O); from ref 5, [R] +57.2°
Supporting Information Available: Experimental pro-
cedures and spectroscopic characterization for all new
compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
D
(H2O).
(19) From ref 10, [R]31 +27° (c ) 0.22, H2O).
D
(20) Sharpless, K. B.; Lauer, R. F. J. Am. Chem. Soc. 1973, 95, 2697-
2699.
(21) Corey, E. J.; Nicolaaou, K. C.; Balanson, R. D.; Machida, Y.
Synthesis 1975, 590-591.
OL0066680
4040
Org. Lett., Vol. 2, No. 25, 2000