Intramolecular 2 + 2 cycloaddition in a baccatin. From a tetracyclic to a hexacyclic molecule

Full text of DOI:10.1021/jo970645q

4900
J . Org. Chem. 1997, 62, 4900-4901
In tr a m olecu la r 2 + 2 Cycloa d d ition in a
Ba cca tin . F r om a Tetr a cyclic to a
Hexa cyclic Molecu le
Eldon G. Nidy, Nancy A. Wicnienski,
Constance G. Chidester, Paul J . Dobrowolski,
Stephen A. Mizsak, Arthur Toy, Thomas A. Runge,
Samuel J . Qualls, Robert C. Kelly,* and
Roy A. J ohnson*
Research Laboratories, Pharmacia & Upjohn, Inc.,
Kalamazoo, Michigan 49001
Received April 11, 1997
In a previous paper, we described the unexpected
isolation of the baccatin enol 1, a compound to which we
gave the name 12,13-isobaccatin III-7-O-TES. The enol
1 was successfully coupled with the side-chain precursor
2 in the presence of dicyclohexylcarbodiimide to give the
12,13-isotaxol 3.1 Although stable enough to isolate and
F igu r e 1. Stereoview of bisoxetane 6.
manipulate, the enol 1 does ketonize, for example, when
exposed to silica gel, giving the 11,12-dihydrobaccatin III
derivative 4. In the present paper, we describe the
further chemical transformation of ketone 4 and elucida-
tion of the structure of yet another unexpected compound
derived from baccatin.
F igu r e 2. Stereoview of 14(1)-enone 9.
DEPT 13C NMR spectrum of 5 are consistent with a
C-1,C-14 olefin. We speculate that the 14(1)-enone
system of 5 arises as a consequence of ketonization of a
small portion of 1 to 4 followed by dehydration of the C-1
hydroxyl group under the conditions of the coupling
reaction.
When samples of 5 are stored neat or are kept in
solution, a new compound is seen to arise. The new com-
pound is stable under normal conditions, and analytical
data indicated that the compound is isomeric with 5. We
were puzzled in our efforts to derive a plausible structure
from the spectral data for this new compound until we
actually built a Dreiding model of the precursor ketone
5. Whereas a conventional drawing of 5, such as the one
used here, portrays the C-9 ketone as being far removed
from the 14(1)-enone double bond, in the model these two
functional groups are seen to be closely aligned and in
an almost parallel orientation.2 The drawing 5a is an
accurate representation of the conformation of the mol-
In the coupling reaction of 1 with 2, a byproduct of the
reaction is consistently isolated in 2-15% yields. On the
basis of the analytical and spectral properties of this
byproduct, the structure of the compound is formulated
(1) Kelly, R. C.; Wicnienski, N. A.; Gebhard, I.; Qualls, S. J .; Han,
F.; Dobrowolski, P, J .; Nidy, E. G.; J ohnson, R. A. J . Am. Chem. Soc.
1996, 118, 919.
(2) There is a second conformation of the molecule that can be built
with the Dreiding model in which the orientation of the C-9 ketone is
in the opposite direction from that portrayed in the drawing of 4a ;
interconversion between the two conformations appears unlikely from
the Dreiding model.
1
to be the 14(1)-enone 5. Distinctive among the H NMR
signals for 5 are a doublet for the C-18 methyl group, a
five-line signal for H-12, and a doublet (J ) 2.8 Hz) at δ
5.94 for H-14. Signals at δ 163.9 (q) and 133.7 (t) in the
S0022-3263(97)00645-2 CCC: $14.00 © 1997 American Chemical Society

Communications
J . Org. Chem., Vol. 62, No. 15, 1997 4901
Sch em e 1
ecule as seen in the Dreiding model. The orientation of
the π-orbitals of these two functional groups as displayed
by the model suggests that a 2 + 2 cycloaddition reaction
of the groups is feasible and that the structure of the new
compound may be the hexacyclic bisoxetane 6.
The analogous baccatin enol 7, in which the 7-O-TES
group is replaced by the 7-O-SDMS (7-O-siamyldimeth-
ylsilyl) group, was observed to undergo the same series
of transformations to ketone 8, the 14(1)-enone 9, and
the analogous final product 10.
two symmetry-independent molecules of 9, between car-
bonyl oxygen (O-9) and the R-carbon of the enone (C-14),
are 2.76 and 2.77 Å and between C-9 and the â-carbon
(C-1) are 2.93 and 2.97 Å. In the X-ray structure of 6,
the C14-O9 bond length is 1.49 Å and the C1-C9 bond
length is 1.62 Å. The distances between reacting atoms
in 5 or 9 are only ∼1.3 Å greater than the bond lengths
in the final oxetanes. Clearly, the proximity and strict
orientation of the two key functional groups in 5 and 9
are determining the regiochemical outcome of the cy-
cloaddition reactions observed in these molecules.
The spectral data for both 6 and 10 are consistent with
the bisoxetane structure. Some key elements among
these data are the presence of only one ketone carbonyl
signal in the 13C NMR spectrum; signals at δ 102.8, 86.4,
81.7, and 80.6 are assigned to the carbon atoms adjacent
to the oxetane oxygens with the δ 102.8 signal assigned
to C-14. In the 1H NMR spectrum, distinctive signals
are seen for H-10 and H-14 at δ 5.65 and 4.58, repec-
tively, as well as a five-line signal for H-12 at 2.92 that
is coupled to a doublet for the C-18 methyl group at 1.58.
Irradiation of solutions of 9 with light of only moderate
intensity rapidly accelerated the conversion into 10,
consistent with formulation of the reaction as a 2 + 2 or
Paterno-Buchi cyclization.3 There is brief literature
precedent for photochemical 2 + 2 cycloaddition reactions
between aldehydes or ketones and electron-deficient
olefins such as an enone.4-6
The bisoxetane 6 was crystallized and the structure
determined by X-ray crystallography, confirming the
structure derived from modeling and spectroscopic analy-
sis. A stereo representation of the structure of 6 obtained
by the X-ray study is shown in Figure 1.7 We were
curious about how close to each other are the two reacting
π-orbitals in either 5 or 9 and we obtained crystals of 9
which gave the X-ray structure shown in Figure 2.7 From
the crystallographic data, we find that the distances, in
The unusual structural nature of the modified bacc-
atins 6 and 10 provided us with an opportunity to prepare
new Taxol8 analog molecules (Scheme 1). Reduction9,10
of 10 with NaBH4/CeCl3 was essentially complete within
15 min and produced a major product (11, 73% yield) and
a minor product (12, 15%). Removal of the silyl protect-
ing group from 11 gave 13, shown by X-ray crystal-
lography to have the 13â-alcohol. The major reduction
product 11 therefore is the 13â-alcohol and the minor
product 12 is the 13R-alcohol.
Both 11 and 12 were carried on to fully developed
taxotere side-chain analogs as shown in the chart. The
Taxotere11 side-chain precursor 2 was used for coupling12
with each of the two alcohols and gave the compounds
14 and 15, respectively. The protecting groups were
removed sequentially from each intermediate. The silyl
group protecting the 7-OH was removed with Et3N‚(HF)3,
giving 16 and 17, and the side-chain protecting group
was removed with acidic methanol. Analogs 18 and 19,
generated by these transformations, were devoid of
activity in microtubule binding experiments.
Su p p or tin g In for m a tion Ava ila ble: Experimental data
for compounds 5-7 and 9-19 (6 pages).
J O970645Q
(3) Cf. Porco, J . A., J r.; Schreiber, S. L. In Comprehensive Organic
Synthesis; Trost, B. M., Fleming, I., Paquette, L. A., Eds.; Plenum:
New York, 1991; Vol. 5, p 151.
(8) The name Taxol has been registered as a trademark by Bristol-
Myers Squibb; the generic name for Taxol is paclitaxel.
(9) Se´nilh, V.; Gue´ritte, F.; Gue´nard, D.; Colin, M.; Potier, P. C. R.
Acad. Sci. Paris 1984, 299, Se´rie II, 1039.
(10) Marder, R.; Dubois, J .; Gue´nard, D.; Gue´ritte-Voegelein, F.;
Potier, P. Tetrahedron 1995, 51, 1985.
(11) Taxotere is a registered trademark of Rhoˆne-Poulenc Rorer; the
generic name for Taxotere is docetaxel.
(12) (a) Didier, E.; Fouque, E; Taillepied, I.; Commerc¸on, A. Tetra-
hedron Lett. 1994, 35, 2349. (b) J ohnson, R. A; Nidy, E. G.; Dobrowolski,
P. J .; Gebhart, I.; Qualls, S. J .; Wicnienski, N. A.; Kelly, R. C.
Tetrahedron Lett. 1994, 35, 7893.
(4) (a) Barltrop, J . A.; Carless, H. A. J . Tetrahedron Lett. 1968, 3901.
(b) Barltrop, J . A.; Carless, H. A. J . J . Am. Chem. Soc. 1972, 94, 1951.
(5) Carless, H. A. J .; Halfhide, A. F. E. J . Chem. Soc., Perkin Trans.
1 1992, 1081.
(6) Meier, R.-M.; Ganem, B. J . Org. Chem. 1994, 59, 5436.
(7) Details of the single-crystal X-ray structures obtained for 6, 9,
and 13 will be published as part of the full account of this work. The
author has deposited atomic coordinates for these structures with the
Cambridge Crystallographic Data Centre. The coordinates can be
obtained, on request, from the Director, Cambridge Crystallographic
Data Centre, 12 Union Road, Cambridge, CB2 1EZ, UK.