Synthetic studies on taxol: Highly stereoselective construction of the taxol C-ring via SN2′ reduction of an allylic phosphonium salt

Article abstract of DOI:10.1021/ol0608606

The highly stereoselective construction of the C3 stereogenic center of the taxol C-ring is described. The trans isomer at the C3-C8 position of the taxol C-ring, which is required for the total synthesis, as well as its diastereomeric cis isomer were suc

Full text of DOI:10.1021/ol0608606

ORGANIC
LETTERS
2006
Vol. 8, No. 14
2973-2976
Synthetic Studies on Taxol: Highly
Stereoselective Construction of the
Taxol C-Ring via SN2′ Reduction of an
Allylic Phosphonium Salt
Masayuki Utsugi, Masayuki Miyano, and Masahisa Nakada*
Department of Chemistry, Faculty of Science and Engineering,
Waseda UniVersity 3-4-1 Ohkubo, Shinjuku-ku, Tokyo 169-8555, Japan
mnakada@waseda.jp
Received April 10, 2006
ABSTRACT
The highly stereoselective construction of the C3 stereogenic center of the taxol C-ring is described. The trans isomer at the C3
of the taxol C-ring, which is required for the total synthesis, as well as its diastereomeric cis isomer were successfully synthesized with highly
diastereoselective SN2 reduction of the allylic phosphonium salts.
−C8 position
′
Recently, we reported asymmetric synthesis of a taxol model
4 via the B-alkyl Suzuki-Miyaura intramolecular coupling
of 3,1a which was prepared by coupling of the enantiopure
fragments, A-ring 1, and C-ring 2 (Scheme 1).1b,c Although
involved at its C3-C4 position, requiring stereoselective
introduction of a hydrogen to the C3 position from its R-face
to complete the total synthesis of taxol. However, construc-
tion of the C3 tertiary stereogenic center with C3R-H was
surmised to be difficult because 4 has a cagelike structure,
thereby impeding stereoselective introduction of a hydrogen
from its R-face, which is the concave side of this molecule.2
Furthermore, even at the stage before an attempted formation
of the B-ring, the stereoselective construction of the C3
stereogenic center starting from the compound with the C3-
C4 alkene is limited.3
Scheme 1
We studied stereoselective construction of the C3 stereo-
genic center4 of the taxol C-ring and report herein highly
(1) (a) Kawada, H.; Iwamoto, M.; Utsugi, M.; Miyano, M.; Nakada, M.
Org. Lett. 2004, 6, 4491-4494. (b) Iwamoto, M.; Kawada, H.; Tanaka, T.;
Nakada, M. Tetrahedron Lett. 2003, 44, 7239-7243. (c) Watanabe, H.;
Iwamoto, M.; Nakada, M. J. Org. Chem. 2005, 70, 4652-4658.
(2) Stereoselective introduction of C3R-H via the elegant intramolecular
protonation. See: Hara, R.; Furukawa, T.; Horiguchi, Y.; Kuwajima, I. J.
Am. Chem. Soc. 1996, 118, 9186-9187.
(3) (a) Nakai, K.; Miyamoto, S.; Sasuga, D.; Doi, T.; Takahashi, T.
Tetrahedron Lett. 2001, 42, 7859-7862. (b) Magnus, P.; Westwood, N.
Tetrahedron Lett. 1999, 40, 4659-4662.
compound 4 possesses a taxane skeleton, an alkene is
10.1021/ol0608606 CCC: $33.50
© 2006 American Chemical Society
Published on Web 06/07/2006

stereoselective construction of the trans isomer at the C3-
C8 position of taxol, which is required for the total synthesis,
as well as its cis isomer by the diastereoselective SN2′
reduction of the allylic phosphonium salts.
Scheme 3
Reduction of the compounds possessing a leaving group
at their allylic positions usually affords a SN2 product and/
or a SN2′ product.5 SN2′ reduction of 5 would produce 6
(Scheme 2), and the trans isomer could be the major
Scheme 2
diastereomer by use of a certain substrate and/or under the
suitable reaction conditions. Nevertheless, the diastereotropic
prochiral C3 position was surmised to be difficult to
discriminate if only the steric factors which arose from the
substituents at the adjacent C8 quaternary stereogenic center
were used to control the stereoselectivity.
Consequently, first, we studied the SN2′ reduction of 5
(R1 ) H) because it would react with the reducing reagent
to afford the corresponding alkoxide, which could work as
a tethered reducing reagent, delivering a hydride to the
desired diastereofacial side at the C3 position to produce 6
(R1 ) H) stereoselectively via the SN2′ pathway. To the best
of our knowledge, no precedents exist for this hydroxyl group
directed SN2′ reduction.6
Preparation of 14, which corresponds to 5 (R1 ) H, L )
Br), commenced with the previously reported enantiopure
ketol 71a (Scheme 3); Vilsmeier reaction7 of 7 was expected
to afford the bromide 8 possessing a one-carbon unit at its
C4 position.4 Preliminary studies on the Vilsmeier reaction
of 7 with POBr3/DMF7c and PBr3/DMF7d afforded 8 in 57%
(at 52% conversion) and 50% yields, respectively. Because
benzylidene acetal 7 was sensitive to the acidic conditions,
optimization of the reaction conditions was focused on the
additive which improved the yield. To this end, we found
that 4 Å molecular sieves improved the yield of 8 up to 70%.
Aldehyde 8 was then converted to the bromide 14 (Scheme
3). Reduction of 8 with NaBH4 produced the corresponding
alcohol, which was converted to DMPM (3,4-dimethoxyphe-
nylmethyl)8 ether 9 in 90% yield (two steps). Reaction of 9
with t-BuLi and then with DMF afforded the aldehyde, which
was reduced with NaBH4 to afford alcohol 10 in 95% yield
(two steps). Benzylation of 10 followed by DIBAL-H
reduction produced 11 (64%, two steps). Alcohol 11 was
protected as the TBS ether, which was then treated with DDQ
to give alcohol 12 in 66% yield (two steps). Direct conversion
of 12 to the bromide 13 was fruitless,9 but this conversion
was successfully achieved via the mesylate to produce 13
in 99% yield. The bromide 13 was exposed to acidic
conditions to remove the TBS group affording 14 in 90%
yield.
(4) Numbering for taxol was applied to the C-ring fragments.
(5) Entwistle, I. D.; Wood, W. W. In ComprehensiVe Organic Synthesis;
Trost, B. M., Fleming, I., Eds.; Pergamon: Oxford, 1991; Vol. 8, pp 955-
981 and references therein.
(6) For the hydroxyl group directed stereoselective 1,4-reduction of R,â-
unsaturated carbonyl compounds, see: (a) Kuethe, J. T.; Wong, A.; Wu,
J.; Davies, I. W.; Dormer, P. G.; Welch, C. J.; Hillier, M. C.; Hughes, D.
L.; Reider, P. J. J. Org. Chem. 2002, 67, 5993-6000. (b) Solomon, M.;
Jamison, W. C.; Cherry, D. A.; Mills, J. E.; Shah, R. D.; Rodgers, J. D.;
Maryanoff, C. A. J. Am. Chem. Soc. 1988, 110, 3702-3704. (c) Salomon,
R. G.; Sachinvala, N. D.; Raychaudhuri, S. R.; Miller, D. B. J. Am. Chem.
Soc. 1984, 106, 2211-2213.
(7) For reviews, see: (a) Jones, G.; Stanforth, S. P. Org. React. 2000,
56, 355-659. (b) Marson, C. M. Tetrahedron 1992, 48, 3659-3726. For
use of POBr3, see: (c) Paquette, L. A.; Johnson, B. A.; Hinga, F. M. Org.
Synth. Collect. Vol. 1973, 5, 215-217. For use of PBr3, see: (d) Rajamannar,
T.; Balasubramanian, K. K. Tetrahedron Lett. 1988, 29, 5789-5792. (e)
Huang, A. X.; Xiong, Z.; Corey, E. J. J. Am. Chem. Soc. 1999, 121, 9999-
10003.
Initially, various reagents were surveyed to reduce 14, but
only SN2 reduction product 16 was obtained. These results
suggested that the leaving group should be examined.
(8) (a) Howell, S. J.; Spencer, N.; Philp, D. Tetrahedron 2001, 57, 4945-
4954. (b) Oikawa, Y.; Horita, K.; Yonemitsu, O. Tetrahedron Lett. 1985,
26, 1541-1544.
(9) Use of CBr4/PPh3 gave unidentified products.
2974
Org. Lett., Vol. 8, No. 14, 2006

Nojima et al. reported that SN2′ reduction proceeded
exclusively in the reaction of allylic phosphonium salts.10
Consequently, we expected that the SN2′ reduction of allylic
phosphonium salt 15 could proceed predominantly, so we
prepared 15 and examined its SN2′ reduction. Phosphonium
salt 15 was easily prepared by 14 and PPh3 (Scheme 3).
Reduction of 15 with LiAlH4 (1.5 equiv) in THF at 0 °C
proceeded in 88% yield (entry 1, Table 1), producing only
Figure 1. Proposed models A and B.
Table 1. Diastereoselective Reduction of 15
stereoselective reduction of 15 would produce the major
product 17 (cis) via model A. The benzyloxy group is located
at the pseudoaxial position in model A, but it is located at
the pseudoequatorial position in model B, suggesting that
model B could be energetically more favored. Therefore, it
was interesting that 17 (cis), which would form via the less
favored model A, was formed exclusively. Coordination of
the oxygen atom at the C7 position of 15 to the reducing
reagent might explain this diastereoselectivity.
entry reagent (equiv) T (°C) time (h) yield (%)a cis/transb
1
2
3
LiAlH4 (1.5)
LiAlH4 (1.5)
Red-Al (1.2)
0
-78
-78
0.5
1
1.5
88
87
31
17:1
1:0
1:0
The models in Figure 1 also explain the reason for the
lack of the desired intramolecular reduction.13 The hydroxyl
group held pseudoaxial in model A was unfavorable for the
intramolecular axial attack because of the stereoelectronic
reason, and the intramolecular reduction was also unfavored
in model B because the hydroxyl group was held equatorial.
Models A and B in Figure 1 suggested that protection of
the allylic hydroxyl group as a bulky TBS ether would
increase the ratio of 17 (trans) because the TBS protected
hydroxyl group would take the pseudoequatorial position,
rendering model B more stable.
1
a Isolated yields. b Ratio determined by 400 MHz H NMR.
the SN2′ products with high diastereoselectivity (17:1).
Formation of 17 (trans) was not detected if the reduction
was performed at -78 °C (entry 2), and the reduction by
use of Red-Al at -78 °C also produced only 17 (cis);
however, the yield was lower than that of entry 2 (entry 3).
We determined the stereochemistry of the major product
in the reduction of 15 as shown in Scheme 4 because we
Reduction of phosphonium salt 19, which was easily
prepared from 13 (Scheme 5), produced 20 (dr ) 1.3:1).
Scheme 4
Scheme 5
had previously reported the highly stereoselective synthesis
of 18.11 Thus, 18 was successfully converted to 17 (trans)
via benzylation, followed by removal of the MPM group,
1
and comparison of their H NMR spectra clearly showed
that the major product derived from 15 was 17 (cis). This
result suggested that the hydroxyl group directed SN2′
reduction of 15 did not occur but rather that the reduction
proceeded via the intermolecular process.
The stereoselective reduction of 15 is well explained by
the models depicted in Figure 1. Considering that the axial
attack is favored in the six-membered ring system,12 the
Conversion of 20 to the known 17 unambiguously deter-
mined the stereochemistry of the products from 20. Although
the cis product was still the major product in this reaction,
(12) For a recent example of the axial addition of cyanide to a cyclohexyl
aldehyde derivative, see: Kallan, N. C.; Halcomb, R. L. Org. Lett. 2000,
2, 2687-2690.
(13) The hydroxyl group in 15 reacted with the reducing reagent because
evolution of hydrogen gas was observed.
(10) Hirabe, T.; Nojima, M.; Kusabayashi, S. J. Org. Chem. 1984, 49,
4084-4086.
(11) Nakada, M.; Kojima, E.; Iwata, Y. Tetrahedron Lett. 1998, 39, 313-
316.
Org. Lett., Vol. 8, No. 14, 2006
2975

this result indicated that the trans selectivity was improved
as compared with the results in Table 1.
Table 2. Diastereoselective Reduction of 23
Studies on the reduction of 15 and 19 strongly suggested
that conformation of the substrate was crucial for the
stereoselective SN2′ reduction. Therefore, we next examined
the SN2′ reduction of 23, in which the axial attack of a
hydride from the bottom face was expected to be dominant
because of the rigid conformation of the benzylidene acetal,
resulting in the preferential formation of 24 (trans) (Figure
2).
entry
T (°C)
time
yield (%)a
cis/transb
1
2
3
4
rt
0
-78
-95
20 min
1 h
1 h
quantitative
quantitative
quantitative
84
1:15
1:20
1:27
1:30
1 h
1
a Isolated yields. b Ratio determined by 400 MHz H NMR.
exclusive formation of 24 (trans) in the reduction of the
allylic phosphonium salt 23 occurred as we expected.
In conclusion, the highly stereoselective construction of
the C3 stereogenic center of the taxol C-ring was achieved.
The trans isomer at the C3-C8 position of the taxol C-ring,
which is required for the total synthesis of taxol, as well as
its cis isomer were successfully synthesized by the diaste-
reoselective SN2′ reduction of the allylic phosphonium salts.
To the best of our knowledge, this is the first diastereose-
lective SN2′ reduction of an allylic phosphonium salt, which
constructs a stereogenic tertiary carbon center with high
selectivity. This protocol would be applicable for other cyclic
systems as well as for acyclic systems to generate a new
stereogenic center. Consequently, further studies on this
protocol are now underway in our laboratory.
Figure 2. Model for the diastereoselective reduction of 23.
Alcohol 10 was easily converted to benzyl ether 21, which
was treated with DDQ to afford the corresponding alcohol,
followed by transformation to the bromide 22 via the
mesylate, and finally, treatment of 22 with PPh3 produced
the desired phosphonium salt 23 (Scheme 6).
Scheme 6
Acknowledgment. We thank Dr. Mitsuhiro Iwamoto for
helpful discussions. This work was financially supported in
part by a Waseda University Grant for Special Research
Projects and a Grant-in-Aid for Scientific Research on
Priority Areas (Creation of Biologically Functional Molecules
(No. 17035082)) from The Ministry of Education, Culture,
Sports, Science and Technology (MEXT), Japan. We are also
indebted to 21COE “Practical Nano-Chemistry”.
Reduction of 23 with LiAlH4 proceeded at ambient
temperature to afford products 24 quantitatively with the
diastereomer ratio of 1:15 (entry 1, Table 2). The ratio
changed according to the reaction temperature; thus, the ratio
was 1:20 at 0 °C (entry 2), 1:27 at -78 °C (entry 3), and
1:30 at -95 °C (entry 4). DIBAL-H reduced the major
product of 24 to the known 17 (trans) in 73% yield; hence,
Supporting Information Available: Spectral data for all
new compounds and experimental procedures. This material
is available free of charge via the Internet at http://pubs.acs.org.
OL0608606
2976
Org. Lett., Vol. 8, No. 14, 2006