A concise synthesis of silanediol-based transition-state isostere inhibitors of proteases.

Article abstract of DOI:10.1021/ol026195s

[reaction: see text] An efficient synthesis of silanediol-based transition-state inhibitors of proteases is described. A new convergent synthesis has been optimized by using a two-step sequence of hydrosilylation followed by the addition of a silyllithio

Full text of DOI:10.1021/ol026195s

ORGANIC
LETTERS
2002
Vol. 4, No. 16
2683-2685
A Concise Synthesis of
Silanediol-Based Transition-State
Isostere Inhibitors of Proteases
,†
Michael G. Organ,* Christophe Buon,† Carl P. Decicco,‡ and
,‡
Andrew P. Combs*
The Department of Chemistry, York UniVersity, 4700 Keele Street, Toronto, Ontario
M3J 1P3, Canada, and DiscoVery Chemistry, Bristol-Myers Squibb Company,
Experimental Station, P.O. Box 80500, Wilmington, Delaware 19880-0500
organ@yorku.ca
Received May 16, 2002
ABSTRACT
An efficient synthesis of silanediol-based transition-state inhibitors of proteases is described. A new convergent synthesis has been optimized
by using a two-step sequence of hydrosilylation followed by the addition of a silyllithio species to an imine. The method should be applicable
to the synthesis of a wide variety of silanediol isosteres to probe the utility of this unique transition-state isostere.
Introduction. Medicinal chemists design and synthesize
peptidiomimetic inhibitors to mimic the topography of the
enzyme-bound peptide conformations in efforts to identify
novel molecules with pharmacokinetic properties suitable for
once-a-day oral dosing. The most common uses of peptido-
mimetics are as selective inhibitors of proteolytic enzymes
(proteases).1 There are four major classes of proteases:
Figure 1. Representative peptide bond hydrolysis isosteres.
aspartic, serine, cysteine, and metallo. One approach to
protease inhibition that has proven very successful is the
incorporation of a non-hydrolyzable isostere of the tetrahedral
intermediate of amide hydrolysis (Figure 1, structure 1).
Ketone or aldehyde hydrates (2) are examples which ef-
fectively mimic the tetrahedral intermediate, but such
structures are often reactive and form undesirable covalent
bonds with other nucleophilic species in ViVo.2 The hydroxyl
group (3) has been incorporated successfully into many
potent peptidomimetic inhibitors (norstatines, statines, cy-
clicureas, etc.) to alleviate this untoward reactivity.3
Other molecular structures that mimic the “geminal diol
motif” (e.g., 4, 5, or 6) have also been incorporated into
peptide-derived peptidomimetics and also have been shown
to serve as effective isosteres. Silicon in the form of a
dialkylsilanediol (e.g., 6) is a stable isostere providing the
diol component is hindered enough to prevent siloxane
polymer formation. In fact, it has been demonstrated recently
that silanediol-based dipeptide analogues are potent inhibitors
of metallo and aspartic proteases.4 Compound 7 (Figure 2)
† York University.
‡ Bristol-Myers Squibb Company.
(3) Bursavich, M. G.; Rich, D. H. J. Med. Chem. 2002, 45 (3), 541-
558.
(4) McN. Sieburth, S.; Nittoli, T.; Mutahi, A.; Guo, L. Angew. Chem.,
Int. Ed. Engl. 1998, 37, 811-814.
(1) Leung, D.; Abbenante, G.; Fairlie, D. P. J. Med. Chem. 2000, 43
(3), 305-341.
(2) Sanderson, P. E. Med. Res. ReV. 1999, 19, 179-97.
10.1021/ol026195s CCC: $22.00 © 2002 American Chemical Society
Published on Web 07/13/2002

Figure 2. Silanediol-based inhibitors of proteases.
Scheme 1a
a Reagents and conditions: (a) Ph2(Cl)SiH, Karstedt’s catalyst, 65 °C (98% yield). (b) Li, THF, rt. (c) LHMDS, THF, -30 °C (used
directly in d). (d) 11, -78 °C (47% yield for steps b, c, d).
inhibits HIV protease with a Ki of 2.4 nM and isomeric
mixtures of 8a and 8b were tested as ACE inhibitors (IC50
values, 57 and 14 nmol, respectively). On the basis of these
results, we were encouraged to investigate the preparation
of a library of such inhibitors to probe the binding sites of
several proteases currently being investigated in our labo-
ratories.
A review of the literature revealed that the existing
synthetic route(s) to structures resembling 7 or 8 were long
(11 or more steps) and contained difficult chemistry, too
much so to consider adapting such routes to high-throughput
synthesis.4,5 The most common strategy for introducing
silicon into an organic structure, that is to form a C-Si bond,
is via an electrophilic silane species such as a chlorosilane.
While simple allylsilanes can be prepared effectively by using
allyl anion chemistry, more elaborate structures cannot as
remote alkyl (sp3) anions are not generated readily. Further,
enolates O-silylate and dithiane-anion chemistry, while
effective for the preparation of acylsilanes, is quite unpleasant
to carry out, especially on a large scale. The route developed
by Sieburth’s group in fact employs both allylmagnesium
bromide and a dithiane to desymmeterize difluorodiphenyl-
silane.4
silicon atom, rather than adding simple units that have to be
subsequently elaborated with a linear strategy. We have
optimized a new convergent synthesis using a two-step
sequence of hydrosilylation followed by the generation of a
silyllithio species, i.e., nucleophilic silicon, and adding it to
an imine.6 The approach is quite general and thus should
allow for the construction of a library of the desired silanediol
peptide-derived peptidomimetics discussed above.
Results and Discussion. Hydrosilylation of allyl ether
9 with chlorodiphenylsilane and Karstedt’s catalyst (Pt2-
[(H2CdCHSiMe2)O]3)7,8 proceeded very cleanly to provide
chlorosilane 10 (Scheme 1). The reaction also proceeded well
with methyl acrylate and with methyl 2-methylacrylate. The
existing method used to generate 2-substituted amides (e.g.,
R ) alkyl on structure 8), which are required for protein
recognition and tight binding, is to alkylate the prerequisite
carboxylic acid derivative. This is an additional step that also
runs the risk of over-alkylating and/or alkylating the amide
on the opposite side of the molecule as well.
The addition of organolithium, organocopper, and Grignard
reagents to the carbon of an imine is a useful method for
the preparation of amines.6 However, steric hindrance and
the reduced electrophilicity of the imine carbon often
We felt that a much more practical route could be realized
if more complex structure could be added directly to the
(6) Campbell, K. N.; Sommers, A. H.; Campbell, B. K. J. Chem. Soc.
1944, 66, 82-84.
(7) Hitchcock, P. B.; Lappert, M. F.; Warhurst, N. J. W. Angew. Chem.,
Int. Ed. Engl. 1991, 30, 438-440.
(5) Chen, C.; McN. Sieburth, S.; Glekas; A.; Hewitt, G.; Trainor, G.;
Erickson-Viitanen, S.; Garber, S.; Cordova, B.; Jeffry, S.; Klabe, R. Chem.
Biol. 2001, 8, 1161-1166.
(8) Kuhnen, T.; Ruffolo, R.; Stradiotto, M.; Ulbrich, D.; McGlinchey,
M. J.; Brook, M. A. Organometallics 1997, 16, 5042-5047.
2684
Org. Lett., Vol. 4, No. 16, 2002

combine to make this a difficult reaction. Despite these
limitations, the one-pot synthesis of a primary amine resulting
from the reaction between an organometallic compound and
an N-silylimine9 would eliminate several functional group
manipulations and protection/deprotection steps while pro-
viding the requisite aminomethylsilane (12) directly. To this
end, the silyllithium derivative was formed from chlo-
rodiphenylsilane 10 that was quenched with silylimine 11,
which was itself prepared in situ by treatment of isobutyral-
dehyde with lithium hexamethyldisilylamide (LHMDS).10
The preparation of any silanol typically requires that the
hydroxyl groups be masked or protected along the synthetic
route. Phenyl groups were selected for this purpose because
of their ability to assist in stabilizing the intermediate silyl
anion and the possibility of removing these groups at the
end with a strong acid.4
Benzoylation of 12 provides a representative amide moiety
on the left-hand side of the structure by substitution chemistry
(12a). This same transformation can be effected by using
condensation chemistry as well, thus broadening the synthetic
flexibility. Amide formation was followed by deprotection
of the methoxymethyl ether to provide 12b and PDC
oxidation of the resulting alcohol provided carboxylic acid
13 (Scheme 2). Condensation of 13 with diethylamine gave
the silanediol precursor 14.
lent yield (Scheme 3). Hydrolysis with potassium hydroxide
in iPrOH14 afforded the silanediol 16 in moderate yield
Scheme 3a
a Reagents and conditions: (a) BF3‚2AcOH, CH2Cl2, reflux (96%
yield). (b) KOH, iPrOH, rt (35% yield).
(35%), but provided the product cleanly. The balance of the
material in this transformation is simply returned difluoride.
We are continuing our efforts to further optimize this
deprotection.
In conclusion, a general and efficient method for the
synthesis of silanediol-based transition-state isosteres has
been developed. Noteworthy is the effective use of hydro-
silylation to install the carboxyl-containing portion of the
structure essentially intact and ready to diversify with a
variety of amines. The choice to utilize a nucleophilic silicon
species to effect Si-C bond formation on the amino side
accomplished a number of goals. It eliminates the need to
generate nonstabilized alkyl anions that are not only difficult
to create, but to control the reactivity due to their strong
basicity. Further, this approach allows for the one-pot
incorporation of a variety of amines via the corresponding
silylimine species. This strategy is currently being adapted
to a parallel synthesis format for the preparation of a variety
of silanediol peptidomimetics.
Scheme 2a
a Reagents and conditions: (a) PhCOCl, Et3N, CH2Cl2, rt (68%
yield). (b) HCl (cat.), MeOH, reflux (94% yield). (c) PDC, DMF,
rt (63% yield). (d) Et2NH, EDCI, CH2Cl2, 0 °C to rt (67% yield).
Acknowledgment. This work was funded by a research
grant from the Ontario Research and Development Challenge
Fund (ORDCF).
Supporting Information Available: Experimental pro-
cedures and spectral data for compounds 10, 12, 12a, 12b,
13, 14, 15, and 16. This material is available free of charge
via the Internet at http://pubs.acs.org.
The phenyl protecting groups can be removed with triflic
acid.4,11 However, low, irreproducible yields for this reaction
in our own hands encouraged us to seek an alternative
method to liberate the silanediol.12 Treatment of 14 with 50
equiv of BF3‚2AcOH13 provided difluorosilane 15 in excel-
OL026195S
(12) For Tamao oxidation, see: (a) Tamao, K.; Kakui, K.; Akita, M.;
Iwahara, T.; Kanatani, R.; Yoshida, J.; Kumada, M. Tetrahedron 1983, 39,
983-990. (b) Fleming, I.; Henning, R. R.; Parker, D. C.; Plaut, H. E.;
Sanderson, P. E. J. J. Chem. Soc., Perkin Trans. 1 1995, 317-337.
(13) Kno¨lker, H.-J.; Jones, P. G.; Wanzl, G. Synlett 1998, 613-616.
(14) Nakadaira, Y.; Oharu, K.; Sakurai, H. J. Organomet. Chem. 1986,
309, 247-256.
(9) Leboutet, L.; Courtois, G.; Miginiac, L. J. Organomet. Chem. 1991,
420, 155-161.
(10) Cainelli, G.; Giacomini, D.; Panunzio, M.; Martelli, G.; Spunta, G.
Tetrahedron Lett. 1987, 28, 5369-5372.
(11) (a) Uhlig, W. Chem. Ber. 1996, 129, 733-739. (b) Bassindale, A.
R.; Stout, T. J. Organomet. Chem. 1984, 271, C1-C3.
Org. Lett., Vol. 4, No. 16, 2002
2685