An efficient, stereocontrolled synthesis of a potent omuralide-salinosporin hybrid for selective proteasome inhibition

Article abstract of DOI:10.1021/ja052376o

A short and highly stereocontrolled synthesis of the potent proteasome inhibitor 3 from the (S)-threonine-derived oxazoline 4 has been developed. The synthetic sequence is summarized in Scheme 1. Aldol coupling of the zinc enolate of 4 with isobutyraldehyde and subsequent silylation provided the TBS ether 5 diastereoselectively (10:1). Reductive cleavage of the oxazoline ring of 5 followed by Swern oxidation of the resulting amino alcohol afforded amino ketone 6, converted further by N-acylation to the acrylamide 7, whose structure was confirmed by X-ray crystallographic analysis. Acrylamide 7 was cyclized to 8 by a novel application of the Kulinkovich Ti(II)-cyclopentene complex. Silylation of 8 to 9 and radical cyclization at low temperature produced the bicyclic lactam 10 with complete control of all stereocenters. Hydroxy desilylation and N-deprotection of 10 gave the dihydroxy ester 11, which was converted to 3 by a novel three-step sequence: (1) demethylation with [Me₂AlTeMe]₂, (2) combined β-lactonization and chlorination, and (3) desilylation to effect cleavage of the TBS ether. Copyright

Full text of DOI:10.1021/ja052376o

Published on Web 06/02/2005
An Efficient, Stereocontrolled Synthesis of a Potent Omuralide-Salinosporin
Hybrid for Selective Proteasome Inhibition
Leleti Rajender Reddy, Jean-Franc¸ois Fournier, B. V. Subba Reddy, and E. J. Corey*
Department of Chemistry and Chemical Biology, HarVard UniVersity, Cambridge, Massachusetts 02138
Received April 12, 2005; E-mail: corey@chemistry.harvard.edu
Omuralide (1)1 and the corresponding thiolester with N-acetyl-
(S)-cysteine (lactacystin, a terrestrial microbiol product2) have been
shown to be remarkably effective and selective inhibitors of
proteasomes, the biomolecular machines responsible for the break-
down of polyubiquitin-tagged proteins. For this reason and also
because proteasome inhibition is a promising approach to anticancer
therapy,3 there have been numerous studies on the synthesis of
omuralide, lactacystin, and analogues leading to a useful structure-
activity correlation.1a This field has been further advanced by the
recent discovery of the marine microbiol product salinosporamide
A (2), which displays even greater potency than omuralide in vitro
in whole cell assays of antiproteasome and antitumor activity.4 We
recently reported the first total synthesis of 2 by a stereocontrolled
process that started from (S)-threonine and employed an internal
Baylis-Hillman reaction to establish the γ-lactam core.5 We report
in this paper a new approach to the synthesis of omuralide-
salinosporamide-type proteasome inhibitors as applied specifically
to the construction of new hybrid molecules based on the structures
1 and 2, specifically, in this instance, the â-lactone 3 (Scheme 1).
The synthesis of 3 is unusually concise and includes novel
methodology for appendage introduction, diastereoselective γ-lac-
tam ring formation, methyl ester cleavage, and â-lactone formation.
a colorless solid, mp 118 °C. The structure of 7 was confirmed by
X-ray crystallographic analysis (see Supporting Information).
Cyclization of the keto acrylamide 7 was next effected stereo-
selectively and, in the required sense, by an unprecedented
application of the Kulinkovich reagent6 (3.5 equiv) from the reaction
of Ti(i-PrO)4 (4 equiv) and cyclopentylmagnesium chloride (7
equiv) in t-BuOMe at -40 °C for 30 min. The resulting black-
brown Ti(II)-cyclopentene complex at -40 °C was treated with
the keto amide 7, and the mixture was maintained at -40 °C for
an additional 30 min after which time 5 equiv of I2 was added to
effect iodination of the intermediate R-titanamethyl-γ-lactam (at
-40 °C for 2 h and 23 °C for 2 h). Acidification of the reaction
mixture with 1 N hydrochloric acid and extraction with methylene
chloride afforded an iodomethyl-γ-lactam, which after treatment
with Et3N (30 min at 23 °C) underwent elimination of the iodine
substituent â to the γ-lactam carbonyl to give lactam 8 as the only
detectable stereoisomer.7,8 Pure 8 was obtained as a colorless solid,
mp 59-63 °C, after flash column chromatography in 85% overall
yield from 7. Silylation of the tertiary hydroxyl group in 8 by
BrCH2Si(CH3)2Cl and replacement of bromine by iodine provided
the iodo olefin 9.
The cyclization of the unsaturated iodide 9 (or the corresponding
bromide) turned out to be unexpectedly challenging. Under the
radical cyclization conditions that had been employed in our original
synthesis of salinosporamide A (Bu3SnH, AIBN, C6H6, 80 °C),5
the required product 10 was formed in only minor amount and the
principal products were the two diastereomeric esters (A) that
correspond to cleavage of the carbon-carbon bond connecting the
R-silyloxyisobutyl appendage to the lactam ring. This highly unusual
cleavage, which was not observed with the cyclization product 10
under the reaction conditions, appears to be due to a thermal
fragmentation of the same R-carbonyl stabilized radical cyclization
intermediate that leads to 10 (perhaps driven by steric repulsion
between the substituents on the â and γ quaternary carbons).
Support for this idea came from cyclization experiments at lower
temperatures. Thus, when the cyclization was conducted at -78
°C, after addition of Et3B to a mixture of 9, Bu3SnH, and Et2AlCl
in air-containing toluene, the required product 10 was obtained in
93% yield with essentially complete stereoselectivity.9
The synthesis of the omuralide-salinosporamide hybrid 3, as
outlined in Scheme 1, started with oxazoline 4, which is readily
available from (S)-threonine as indicated previously.5 Reaction of
4 in THF at -78 °C with 1.2 equiv of potassium hexamethyl-
disilazide in toluene at -78 °C for 1 h provided an enolate which
was treated successively with ZnCl2 (1.3 equiv, 10 min, -78 °C)
and isobutyraldehyde (1.5 equiv, 8 h, -78 °C) to produce the
required aldol product in 95% yield with 10:1 diastereoselectivity.
Recrystallization from hot ethyl acetate afforded a major product,
mp 118-120 °C (81%), which was converted to the corresponding
tert-butyldimethylsilyl (TBS) ether 5 in 95% yield. Reductive
cleavage of the oxazoline 55 followed by Swern oxidation (at -78
°C with the reagent from Me2SO and ClCOCOCl in CH2Cl2
followed by Et3N) generated the R-amino ketone 6 in 89% yield.
N-Acylation of 6 with acrylyl chloride (3 equiv) and triisobutyl-
amine in CH3CN solution at 23 °C for 24 h resulted in incomplete
reaction, and so additional acrylyl chloride (3 equiv) and tertiary
amine (3 equiv) were added on day 2 and day 3. Isolation of the
product after day 4 afforded the keto acrylamide 7 (79% yield) as
Hydroxy desilylation and N-deprotection of 10 followed using
the methodology previously described for the synthesis of salino-
sporamide A5 to afford the dihydroxy lactam 11. Attempts to cleave
9
8974
J. AM. CHEM. SOC. 2005, 127, 8974-8976
10.1021/ja052376o CCC: $30.25 © 2005 American Chemical Society

C O M M U N I C A T I O N S
Scheme 1
either the methyl ester or the silyl ether groups of 11 met with
complete failure under a wide variety of conditions, clearly because
of the great ease of retroaldol cleavage of the same carbon-carbon
bond that was involved in the conversion of 9 to the undesired
product A under radical cyclization conditions. Because every one
of available general reagents10 for the conversion of methyl esters
to the corresponding carboxylic acids failed with substrate 11, a
totally new method was sought that could deal with the issues of
(a) strong steric screening of the methoxy carbonyl group in 11
and (b) strong driving force for retroaldol cleavage from the
tetrahedral alkoxide intermediate that results from nucleophilic
addition to the methoxy carbonyl group of 11. Our attention was
directed to a reagent that could effect COOCH3 demethylation by
selective attack on the methyl group rather than the ester carbonyl
and that could operate under mild conditions in terms of temperature
and basicity. Success was realized with a new reagent, [Me2-
AlTeMe]2, that is readily generated by heating tellurium powder
(1.2 equiv) and trimethylaluminum (1 equiv) in toluene at reflux
for 6 h and cooling the product to ambient temperature to give a
0.8 M solution of [Me2AlTeMe]2 in toluene. Treatment of the
dihydroxy ester 11 with a freshly prepared solution of [Me2-
AlTeMe]2 (0.8 M in toluene) at 23 °C for 12 h under nitrogen
followed by quenching of the reaction mixture with 1 N hydro-
chloric acid and extractive isolation with ethyl acetate afforded the
dihydroxy acid corresponding to 11 cleanly. CAUTION: Organo-
tellurium reagents should be used only in a well-ventilated hood;
treatment with 1 N hydrochloric acid (or bleach) effects their
destruction and deodorization. When the crude acid was subjected
to reaction with 4 equiv of Ph3PCl2 in dry 1:1 CH3CN/pyridine at
23 °C for 12 h, it was transformed directly into the TBS ether of
3, which was obtained as a colorless oil in 81% yield after extractive
isolation with ethyl acetate and flash chromatography on silica gel.
This very efficient operation combines side chain chlorination with
a novel method of â-lactone formation in a single step. Finally,
desilylation of the TBS ether of 3 using 3:1 CH3CN and 48%
hydrofluoric acid at 23 °C for 3 days afforded the target omuralide-
salinosporamide hybrid 3 as a colorless solid in 82% yield after
extractive isolation (EtOAc) and flash chromatography on silica
gel. Found for pure 3: Rf ) 0.45 (silica gel plate, EtOAc-hexane
1:1); mp 144-145 °C [R]23 -22.5 (c 0.5, CHCl3). FTIR (film)
D
ν
max: 3222, 2960, 2944, 2867, 1833, 1710, 1254, 1090, 1059, 852,
825, 777 cm-1. H NMR (CDCl3, 500 MHz): δ 6.43 (1H, (brs)),
3.97 (1H, m), 3. 84 (1H, t, J ) 6.5 Hz), 3.77 (1H, m), 2.82 (1H,
t, J ) 7.5 Hz), 2.27 (1H, m), 2.12 (1H, m), 1.93 (1H, m), 1.82
(3H, s), 1.12 (3H, d, J ) 7.0 Hz), 1.08 (3H, d, J ) 7.0 Hz). 13C
NMR (CDCl3, 125 MHz): δ 177.72, 167.47, 85.89, 79.04, 71.94,
44.92, 42.49, 31.53 28.24, 19.92, 19.74, 18.76. HRMS (ESI) calcd
for C12H19ClNO4 [M + H]+ 276.1002, found 276.1006.
1
The biological activity of 3 was evaluated relative to omuralide
(1) by measurement of the rates of inactivation of the â5-subunit
of the human 20S proteasome. The potency of 3 by this standard
is about 2.5 times that of omuralide (1) but somewhat less than
that of salinosporamide A (2). We anticipate that 3 may have
comparable whole cell potency to salinosporamide, although these
studies are still in progress. One reason for this expectation is the
possibility that the chlorine substituent in 2 and 3 may be critical
to ensuring that the covalent attachment of these agents to the
proteasome is irreversible. This possibility is suggested by chemical
9
J. AM. CHEM. SOC. VOL. 127, NO. 25, 2005 8975

C O M M U N I C A T I O N S
evidence for the facile displacement of chlorine from the R-side
chain of 2 or 3 by a free tertiary hydroxyl group at C-â. Thus,
reaction of 3 with benzylamine in THF at 23 °C leads to rapid
cleavage of the â-lactone ring to form the amide 12 which upon
exposure to K2CO3 in CH3OH affords the bicyclic tetrahydrofuran
derivative 13. In this experiment, benzylamine serves as a chemical
model for the nucleophilic â5-subunit of the proteasome. To the
extent that this model is valid, a simple explanation for the potency
of salinosporamide A in cell-function assays may be found in the
irreversibility of proteasome inactivation associated with tetra-
hydrofuran ring closure, as exemplified by 12 f 13.11
with a wide range of hindered esters. Three further examples of
the new process are given in Table 1.
Acknowledgment. J.-F.F. is grateful to NSERC of Canada for
a postdoctoral fellowship. We thank Millennium Pharmaceuticals
Inc. for a general research grant and for data on proteasome
inhibition.
Supporting Information Available: Experimental procedures for
the steps shown in Scheme 1 are given along with characterization
data for each product and the complete list of authors for ref 3a. The
complete list of authors for ref 3a. This material is available free of
charge via the Internet at http://pubs.acs.org.
References
(1) Reviewed in: (a) Corey, E. J.; Li, W.-D. Z. Chem. Pharm. Bull. 1999,
47, 1-10. (b) Corey, E. J.; Reichard, G. A.; Kania, R. Tetrahedron Lett.
1993, 34, 6977-6980. (c) Corey, E. J.; Reichard, G. A. J. Am. Chem.
Soc. 1992, 114, 10677-10678. (d) Corey, E. J.; Li, W.; Reichard, G. A.
J. Am. Chem. Soc. 1998, 120, 2330-2336. (e) Fenteany, G.; Standaert,
R. F.; Reichard, G. A.; Corey, E. J.; Schreiber, S. L. Proc. Natl. Acad.
Sci. U.S.A. 1994, 91, 3358-3362. (f) Fenteany, G.; Standaert, R. F.; Lane,
W. S.; Choi, S.; Corey, E. J.; Schreiber, S. L. Science 1995, 268, 726-
731.
(2) (a) Omura, S.; Fujimoto, T.; Otoguro, K.; Matsuzaki, K.; Moriguchi, R.;
Tanaka, H.; Sasaki, Y. J. Antibiot. 1991, 44, 113-116. (b) Omura, S.;
Matsuzaki, K.; Fujimoto, T.; Kosuge, K.; Furuya, T.; Fujita, S.; Nakagawa,
A. J. Antibiot. 1991, 44, 117-118.
(3) Bortezomib (Velcade), a peptidyl boronic acid which is a reversible (0.6
nM Ki) proteasome inhibitor, is currently in use and approved for the
treatment of multiple myeloma. In addition, there are numerous ongoing
clinical trials on the use of this agent for treatment of other malignant
diseases. See: (a) Richardson, P. G. et al. N. Engl. J. Med. 2003, 348,
2609-2617. (b) Richardson, P. G.; Hideshima, T.; Anderson, K. C. Cancer
Control 2003, 10, 361-369. (c) Steinberg, D. The Scientist 2003, 17 (S2),
S18-S22. (d) Adams, J. Proteasome Inhibitors in Cancer Therapy;
Humana Press: New York, 2004.
(4) Feling, R. H.; Buchanan, G. O.; Mincer, T. J.; Kauffman, C. A.; Jensen,
P. R.; Fenical, W. Angew. Chem., Int. Ed. 2003, 42, 355-357.
(5) Reddy, L. R.; Saravanan, P.; Corey, E. J. J. Am. Chem. Soc. 2004, 126,
6230-6231.
(6) For reviews on applications of titanacyclopropane reagents, see: (a)
Kulinkovich, O. G. Chem. ReV. 2003, 103, 2597-2632. (b) Sato, F.;
Okamoto, S. AdV. Synth. Catal. 2001, 343, 759-784.
(7) For the closest precedent to this cyclization involving δ,ꢀ-enones, see:
(a) Kablaoui, N. M.; Buchwald, S. L. J. Am. Chem. Soc. 1996, 118, 3182-
3191. (b) Mandal, S. K.; Amin, S. R.; Crowe, W. E. J. Am. Chem. Soc.
2001, 123, 6457-6458. (c) Crowe, W. E.; Vu, A. T. J. Am. Chem. Soc.
1996, 118, 1557-1558. (d) Quan, L. G.; Cha, J. K. Tetrahedron Lett.
2001, 42, 8567-8569.
(8) For a detailed study of the pathway for the stereochemistry of the
conversion of 7 to 8, see: Reddy, L. R.; Fournier, J.-F.; Subba Reddy, B.
V.; Corey, E. J. Org. Lett. ASAP.
The efficient synthesis of antiprotealide11 (3) was made possible
by the success of a number of crucial individual steps including:
(1) the doubly diastereoselective aldol coupling of a zinc enolate
via a preferred cyclic chair-formed transition state; (2) the Kulink-
ovich reagent-mediated internal carbo-titanation of the acrylamide
7, which leads to 8 in excellent yield and with very high
diastereoselectivity; (3) the striking effectiveness of low-temperature
conditions for promoting the diastereoselective cyclization 9 f 10;
and, finally (4) the great success of the novel reagent [Me2AlTeMe]2
for the demethylation of ester 11.
Preliminary studies in this laboratory have shown that this mild
method for methyl ester cleavage can be used to good advantage
Table 1. [Me2AlTeMe]2-Promoted Conversion of Esters into
Carboxylic Acids in Toluene at 23 °C
(9) For low-temperature initiation of radical formation with Et3B-O2, see:
(a) Sibi, M. P.; He, L. Org. Lett. 2004, 6, 1749-1752. (b) Kim, K.;
Okamoto, S.; Sato, F. Org. Lett. 2001, 3, 67-69.
(10) For reviews, see: (a) Greene, T. W.; Wuts, P. G. M. ProtectiVe Groups
in Organic Synthesis, 3rd ed.; John Wiley & Sons: New York, 1999. (b)
Kocienski, P. J. Protecting Groups, 3rd ed.; Georg Thieme: Stuttgart,
New York, 2004. (c) Salomon, C. J.; Mata, E. G.; Mascaretti, O. A.
Tetrahedron 1993, 49, 3691-3734. (d) Nicolaou, K. C.; Estrada, A. A.;
Zak, M.; Lee, S. H.; Safina, B. S. Angew. Chem., Int. Ed. 2005, 44, 1378-
1382. (e) Olah, G. A.; Narang, S. C.; Salem, G. F.; Gupta, B. G. B.
Synthesis 1981, 142-143. (f) Marchand, P. S. J. Chem. Soc., Chem.
Commun. 1971, 667-668. (g) Bartlett, P. A.; Johnson, W. S. Tetrahedron
Lett. 1970, 4459-4462.
(11) Because of the availability of the hybrid molecule 3 by the synthetic route
that is summarized in Scheme 1 and also because of its likely irreversible
mode of action in vivo, we assign it the name “antiprotealide” in the
expectation that it represents a convenient and useful tool in experimental
biology and medicine. For a recent review of natural product hybrids,
see: Tietze, L. F.; Bell, H. P.; Chandrasekhar, S. Angew. Chem., Int. Ed.
2003, 42, 3996-4028.
1
a All products were characterized by H NMR, IR, and mass spectros-
copy. b Isolated yield of pure product.
JA052376O
9
8976 J. AM. CHEM. SOC. VOL. 127, NO. 25, 2005