possessed the (3R,4R) configuration.13 The preferential
formation of (3R,4R)-8a was in good agreement with Lajoie’s
nucleophilic addition to 510a and Ma’s aldol reaction of the
ester enolate derived from proline ester8 with 7. On the basis
of these results, we converted 4b to the TBS ether 4c since
the N-Cbz group in the aldol adduct 8a is incompatible with
the benzyl group for its selective removal. The use of LDA
for the aldol reaction using 4c was found to be superior to
LHMDS to produce the aldol adduct 8b in 69% yield (90%
based on recovery of the starting material (brsm)).14
afford 13 as the sole diastereomer (87%, 3 steps from 10).18
The stereochemistry at C3 was determined to be (S) by
converting it to the bicyclic lactam 9b and its X-ray crystal-
lographic analysis (Scheme 2).13,19 The stereochemical out-
come of the present reduction was in accord with those of
Lajoie’s synthesis of ꢀ-hydroxy-R-amino acids.10a
Installation of the N-acyl-R,ꢀ-dehydroalanyl moiety was
performed by the initial oxidation of 13 and subsequent
tetramethylguanidine (TMG)-assisted HWE olefination of the
resulting aldehyde 3 with 6a to give (Z)-2 as the sole
product.20 The hydrogenation of (Z)-2 using the cationic
achiral rhodium catalyst under 1.0 MPa hydrogen atmosphere
smoothly proceeded to give a mixture of diastereomers at
C9 in which the undesired (9R)-isomer 14a was obtained as
the major product (14a/14b ) 95:5). Contrary to our initial
expectation that the desired (9S)-14b could be prepared using
a chiral catalyst, attempts using the (R,R)- and (S,S)-
QuinoxP*-Rh complex21 resulted in the formation of the
undesired 14a as the major isomer, respectively (14a/14b
) 91:9 and 83:17). These results suggested that the product’s
stereochemistry is attributed to the structure of (Z)-2.
Therefore, the reagent controlled diastereoselective hydro-
Next, we examined the inversion of the configuration at
the C3 hydroxy group by the oxidation-reduction sequence.
Because the C3 hydroxy group of 8b is highly congested
due to the presence of the polar and sterically bulky
functional groups at its proximal position, the oxidation using
conventional reagents15 was not successful at all, resulting
in the decomposition or recovery of 8b. After numerous
attempts, we found that the treatment with TFAA-DMSO
at -15 °C underwent a smooth oxidation to give the desired
ketone 10 (90%).16 The next reduction was again ac-
companied by difficulty, contrary to the previous studies by
Kitahara4 and Chamberlin6 groups in which NaBH4 in
THF-MeOH or DIBAL-THF was the effective reducing
conditions, respectively. The reduction of 10 using these and
other reagents, such as LiBHEt3 and LiBH4, did not react at
all at low temperature or reduced both the ketone and ester
groups at room temperature. During the screening of the
solvent system using NaBH4, we found that a trace amount
of water in DME affected the desired reduction to give 11.
The yield was increased to 48% when 10 equiv of water was
added.17 The byproduct was a hydrolyzed ester and/or lactam
product presumably because the solution was strongly basic,
suggesting that the reduced pH would prevent the side reactions.
Finally, the reaction condition was optimized using DME and
pH 9 phosphate buffer (9:1). The reaction smoothly proceeded
to give the crude 11.17 The product without chromatographic
purification was converted into the methyl ester 13 by the
following sequence of reactions: (1) treatment with aqueous
AcOH, which produced simultaneous partial hydrolysis of the
OBO ester and removal of the TBS group to give 12, and (2)
transesterification of 12 with powdered K2HPO4 in MeOH to
1
genation was not applicable in this case. The H NMR
analysis of (Z)-2 revealed that its vinyl amide proton
appeared at relatively low field (δ 10.7)22 probably as a result
of the hydrogen bonding with the proximal Boc group,
indicating that the re face of (Z)-2 is sterically shielded from
attack by the [Rh-H] complex (Scheme 3).6
Scheme 3. Synthesis of (Z)-2 and Its Catalytic Hydrogenation
(10) (a) Blaskovich, M. A.; Lajoie, G. A. J. Am. Chem. Soc. 1993, 115,
5021–5030. (b) Blaskovich, M. A.; Evindar, G.; Rose, N. G. W.; Wilkinson,
S.; Luo, Y.; Lajoie, G. A. J. Org. Chem. 1998, 63, 3631–3646. (c) Hansen,
D. B.; Wan, X.; Carroll, P. J.; Joullie´, M. M. J. Org. Chem. 2005, 70,
3120–3126.
(11) Garner, P.; Park, J.-M. Org. Synth. 1991, 70, 18–28.
(12) Preparation of the antipode of 4a: Lee, M.; Lee, T.; Kim, E.-Y.;
Ko, H.; Kim, D.; Kim, S. Org. Lett. 2006, 8, 745–748.
(13) For procedures for the conversion of 8a to 9a and 13 to 9b and the
NOE correlation of 9a and X-ray structure of 9b, see Supporting
Information.
(14) A trace amount of the inseparable diastereomers was contaminated.
1
The amount was estimated to be 5-10% from the H NMR spectrum of
the crude 8b, but the exact ratio could not be determined because of their
broard signals. The contaminated diastereomers were chromatographically
removed during its conversion to 11.
To overcome this problem, we considered that the
hydrogenation of the (E)-isomer 2 would give the desired
(15) The oxidation of 8b was attempted using the following reagents:
Dess-Martin periodinane, PDC, (COCl)2-DMSO, or 1-Me-AZADO.
(16) Appell, R. B.; Duguid, R. J. Org. Process Res. DeV. 2000, 4, 172–
174.
(18) Methanolysis of the (3R)-isomer of 12 using K2CO3 (1.5 equiv) in
MeOH gave a sluggish mixture of products containing an N-Cbz group-
removed cyclic carbamate (4.9%), and a small amount of the desired 3-epi-
13 was isolated (1.5%). The extremely mild transesterification using K2HPO4
is under investigation and will be reported in due course.
(17) The yield was calculated after conversion to 13 since the OBO
ester 11 was labile upon chromatographic isolation or storage under ambient
conditions for gradual conversion to the partially hydrolyzed 12. Contrary
to 11, its C3-epimers 8a,b are quite stable under these conditions.
4666
Org. Lett., Vol. 11, No. 20, 2009
Hamada, Makoto
