7896
T.-P. Loh et al. / Tetrahedron Letters 42 (2001) 7893–7897
Ph
LiHMDS / THF
–78 °C, 15
Ph
Ph
CO Me
H
2
N
O
N
COOMe
N
S
R
R
3%
O
OH
O
11 and 11a
16
Scheme 4. Cross aldol reaction.
a-carbon is opposite to the natural product. In addi-
tion, our initial derivatization of aldol 14 which entails
functionalization of the ester at C-8 of 14 would invert
the stereochemistry about C-4 of 1, rendering the
chirality incorrect. In order to circumvent this problem,
with no signs of the bis-oxazolidine skeleton of 14.
Apparently, a retro-aldol cleavage had occurred to
regenerate 11. Subsequent trials with silyl based TIPS
(TIPSOTf, 2,6-lutidine) and TMS (TMSCN; TMSCl,
Mg) failed with recovery of starting material, pre-
sumably due to steric encumbrance about the C-3
hydroxyl. Finally, the MTM (methylthionylmethyl)
group proved successful. However, due to the similarity
of the conditions used (DMSO/AcOH/Ac2O) with that
of Swern oxidation, an approximately 50:50 mixture of
MTM protected 17 and the ketone side-product 18 was
obtained (Scheme 5). The mixture was then subjected to
LiBH4 reduction in MeOH/THF, to give the primary
alcohol 19 and the diol 20 quantitatively, which were
then separated by silica gel column chromatography.
Unfortunately, ensuing efforts to transform the primary
hydroxyl into a good leaving group by both sulfonyla-
tion (MsCl, pyr; TsCl, pyr) and bromination (NBS,
CBr4) proved futile.
we would have to employ the enantiomeric
D-serine
instead, so as to get the enantiomer of serine aldol 14,
although the stereochemistry about C-3 would be incor-
rect in this case. Nevertheless, post-aldol oxidation and
stereoselective reduction should afford the desired stereo-
chemically ‘correct’ isomer.
In our initial effort to establish the versatility of this
route for the synthesis of all the possible stereoisomers
of kaitocephalin, the enantiomeric aldehyde 15 was
synthesized in the same way as for 13 by commencing
from
D-serine methyl ester hydrochloride (Scheme 4).
Surprisingly, the cross aldol reaction between 15 and
ester 11 was found to give a major product with relative
stereochemistry identical to aldol 14, as revealed by a
single-crystal X-ray diffraction analysis of the cross
aldol product 16 (Scheme 4). In an attempt to rational-
ize this unexpected observation, it was noticed that the
yield of 16 (3%) is much lower than that of 14. Also
bearing in mind that the 1,3-oxazolidine ester used is a
9:1 mixture of diastereomeric cis:trans isomers, the
isolated product 16 may have been formed from the
reaction of 15 with the (Z)-enolate derived from the
trans-substituted 1,3-oxazolidine 11a, that is 16 is the
desired enantiomer of 14. This is also supported by the
fact that the selectivity of serine aldol 16 toward all the
other diastereomers decreased drastically from 92:8 in
the case of 14, to 67:33.14 The aldol reaction for reasons
not very clear at this moment, plausibly the stereochem-
ical-controlled conformational match of the two react-
ing molecules exhibits a preference for 14 or 16.
In conclusion, we succeeded in the key aldol reaction of
oxazolidine based ester 11 and aldehyde 13, both
derived from serine methyl ester hydrochloride, for the
construction of the C-1–C-5 fragment of kaitocephalin
1. Unfortunately, further derivatization for coupling
with the known C-6–C-9 fragment 6 via an anticipated
enolate addition met with some difficulties and further
work toward the second key step is currently in pro-
gress in our laboratory. In addition, the absolute
configuration of kaitocephalin 1 was unveiled very
recently, which dictates a post-aldol oxidation–reduc-
tion sequence in order to generate the correct stereo-
chemistries in the C-1–C-4 fragment, while commencing
from
D-serine instead.
As we were initially unaware of the absolute configura-
tion of natural kaitocephalin 1, subsequent synthetic
work was carried out on 14. Continuing with the
synthetic studies, our next task is to protect the newly
formed C-3 secondary hydroxyl of 14, prior to reduc-
tion of the ester function. The benzyl group was chosen
initially, but again to our surprise, upon treating 14
Acknowledgements
We gratefully acknowledge the National University of
Singapore for the generous financial support. We would
also like to thank A/P Jagadese J. Vittal and Ms. Tan
Geok Kheng of the X-ray Diffraction Laboratory,
National University of Singapore, for determining the
X-ray structure for serine aldols 14 and 16.
1
with NaH followed by BnBr the crude H NMR indi-
cated clearly the presence of signals from the ester 11,