Total Synthesis of ClaVepictines A and B
J. Am. Chem. Soc., Vol. 121, No. 43, 1999 10017
Scheme 13
To probe the origin of the observed diastereocontrol in the
key cyclization, the 1:1 diastereomeric allenes were also
prepared starting from the racemic alcohol 47. The otherwise
identical cyclization of these diastereomeric allenes in place of
55 produced a ca. 1:2 mixture of 56 and 57 in 48% yield. In
additional studies, poor selectivity was also observed in the
cyclization of several piperidines bearing structurally related,
racemic (i.e., 1:1 diastereomeric) allenes. It is, therefore, clear
that diastereofacial control generated by the allene functionality
is exceptionally high and overrides the inherent conformational
bias of the substituted piperidine ring.
Having demonstrated the utility of electrophilic cyclizations
of ω-amino allenes in the synthesis of multi-substituted quino-
lizidines, our next goal was to optimize the yield and diaste-
reoselectivity of the cyclization reaction. At the same time, we
sought to reduce the overall number of steps used in our first-
generation synthesis of clavepictines, most of which involved
diastereocontrolled installation of the allene function after the
cross-coupling reaction. Utilization of a prefabricated allene
would lend itself to a convergent, efficient synthesis, but is not
compatible (i.e., see 49 f 50 in Scheme 12) with the alkyne
group employed for Sonogashira coupling. In lieu of the
operationally simple Sonogashira reaction, cross-coupling of the
vinyl triflate 12 and a primary alkyl group was required: the
more challenging coupling reaction of this type has been
achieved by the Suzuki procedure involving use of alkylboranes
derived from hydroboration with 9-BBN or by action of
organozinc reagents.26 As summarized in Scheme 14,27 â-hy-
dride elimination proved to be the major pathway under several
reaction conditions. This unsatisfactory result prompted us to
explore Beak’s deprotonation-substitution chemistry for the
second-generation synthesis of 1A and 1B.
Second-Generation Synthesis. Beak and co-workers have
developed an efficient diastereoselective functionalization of
BOC-protected pyrrolidines and piperidines by R′-lithiation and
subsequent electrophilic substitution.7 Of particular relevance
is the highly diastereoselective preparation of trans-2,6-BOC
piperidines from N-BOC-2-methylpiperidine.28 However, elec-
trophilic substitution of the dipole stabilized carbanion inter-
mediates is limited to acylation, condensation with aldehydes,
or methylation. Indeed, lithiation (sec-BuLi/TMEDA) of the
N-BOC-piperidine 62 and subsequent alkylation with a simple
alkyl iodide [e.g., (E)-6-iodo-2-hexene] proved to be diastereo-
selective (∼5:1 ds), but sluggish, affording up to 45% of the
correct alkylation product, along with recovered starting material
(45%) (Scheme 15). In contrast, use of alkyl iodides bearing
an allene gave none of the alkylation product, presumably
because of the presence of relatively acidic protons associated
with the allene functionality. We subsequently employed an
aldehyde bearing an enantiopure allene unit as the electrophile,
since the alcohol function of the resulting adduct should be
readily reduced at an appropriate stage.
then introduced (86%) in a one-pot operation by sequential
treatment of 53 with DIBAL-H and n-hexylmagnesium bromide
to afford alcohol 54 (Scheme 13). Not surprisingly, 54 was
obtained as an inseparable 1:1 diastereomeric mixture of the
C-14 alcohols, but since the alcohol was destined to undergo
dehydration, lack of stereocontrol here was inconsequential.
Subsequent alcohol protection and removal of the allyl carbam-
ate group by standard methods provided the δ-allenic amine
55 (70% overall). Silver nitrate-mediated cyclization of 55
produced a 7:1 mixture of the desired cis-quinolizidine 56 and
its C-10 epimer 57 in 54% yield. Selective removal of the
triethylsilyl group, followed by treatment with the Martin
sulfurane,25 gave a 10:1 mixture of TIPS-protected clavepictine
and the corresponding E,Z-isomer. Finally, separation and
treatment with n-Bu4NF afforded (+)-clavepictine B (1B). (-)-
Clavepictine A (1A) was then prepared by acetylation of 1B.
The spectroscopic and physical properties of 1A and 1B were
in excellent agreement with literature data;1 the 1H and 13C NMR
spectra of 1A and 1B were found to be identical with those of
authentic samples which were kindly provided by Professor
Momose.4 For additional characterization, the trans-quinolizidine
57 was converted to epi-clavepictine B, which exhibits drasti-
cally different spectra from 1B.
(26) (a) Miyaura, N.; Ishikawa, M.; Suzuki, A. Tetrahedron Lett. 1992,
33, 2571. (b) Oh-e, T.; Miyaura, N.; Suzuki, A. J. Org. Chem. 1993, 58,
2201. (c) Miyaura, N.; Ishiyama, T.; Sasaki, H.; Ishikawa, M.; Satoh, M.;
Suzuki, A. J. Am. Chem. Soc. 1989, 111, 314. (d) Oh-e, T.; Miyaura, N.;
Suzuki, A. Synlett 1990, 221. (e) Kobayashi, M.; Negishi, E. J. Org. Chem.
1980, 45, 5223. (f) Negishi, E.; Valente, L. F.; Kobayashi, M. J. Am. Chem.
Soc. 1980, 102, 3298. (g) Tamaru, Y.; Ochiai, H.; Nakamura, T.; Yoshida,
Y. Angew. Chem., Int. Ed. Engl. 1987, 26, 1157. (h) Stadtmu¨ller, H.; Lentz,
R.; Tucker, C. E.; Stu¨demann, T.; Do¨rner, W.; Knochel, P. J. Am. Chem.
Soc. 1993, 115, 7027. (i) Organocopper Reagents; Taylor, R. J. K., Ed.;
Oxford University Press: Oxford, UK, 1994.
(24) (a) Seyferth, D.; Marmor, R. S.; Hilbert, P. J. Org. Chem. 1971,
36, 1379. (b) Gilbert, J. C.; Weerasooriya, U. J. Org. Chem. 1982, 47, 1837.
(c) Ohira, S. Synth. Commun. 1989, 19, 561. (d) Brown, D. G.; Velthuisen,
E. J.; Commerford, J. R.; Brisbois, R. G.; Hoye, T. R. J. Org. Chem. 1996,
61, 2540.
(25) Martin, J. C.; Franz, J. A.; Arhart, R. J. J. Am. Chem. Soc. 1974,
96, 4604.
(27) Yields given in Scheme 14 are based on recovered starting triflate
(13-18%).
(28) (a) Reference 7b. (b) Chackalamannil, S.; Davies, R. J.; Asberom,
T.; Doller, D.; Leone, D. J. Am. Chem. Soc. 1996, 118, 9812.