Synthesis of a novel C-10 spiro-epoxide of paclitaxel

Article abstract of DOI:10.1016/S0960-894X(01)00280-3

New analogues of paclitaxel (1a, active constituent of Taxol) were synthesized containing an epoxide at the C-10 position. The introduction of the epoxide was carried out by selective removal of the C10-acetate followed by protection of the C2′- and C7-hydroxyl groups. After oxidation to yield a ketone at the C10-position, this intermediate was reacted with dimethylsulfonium ylide. Deprotection and further manipulations provide the C10-spiro epoxide of paclitaxel (1b) and the corresponding C7-MOM ether (1c).

Full text of DOI:10.1016/S0960-894X(01)00280-3

Bioorganic & Medicinal Chemistry Letters 11 (2001) 1683–1685
Synthesis of a Novel C-10 Spiro-Epoxide of Paclitaxel
Michael A. Walker,* Timothy D. Johnson, Stella Huang, Dolatorai M. Vyas
and John F. Kadow
Bristol-Myers Squibb, Pharmaceutical Research Institute, Richard L. Gelb Center for Pharmaceutical Research and Development,
5 Research Parkway, Wallingford, CT 06492, USA
Received 22 February 2001; accepted 13 April 2001
Abstract—Newanalogues of paclitaxel ( 1a, active constituent of Taxol1) were synthesized containing an epoxide at the C-10
position. The introduction of the epoxide was carried out by selective removal of the C10-acetate followed by protection of the C20-
and C7-hydroxyl groups. After oxidation to yield a ketone at the C10-position, this intermediate was reacted with dimethylsulfo-
nium ylide. Deprotection and further manipulations provide the C10-spiro epoxide of paclitaxel (1b) and the corresponding C7-
MOM ether (1c). # 2001 Elsevier Science Ltd. All rights reserved.
Taxol1 (paclitaxel, 1a as an active ingredient) has been
found to be very beneficial in the treatment of ovarian
and breast cancer as well as AIDS-related Karposi’s
sarcoma. Due to its unique mechanism of action and the
impact it has had on these forms of cancer it is currently
being studied in nonsmall cell lung cancer (NSCLC),
small cell lung cancer (SCLC), and head and neck can-
cer.1 It has also been the subject of numerous medicinal
chemistry studies aimed at optimizing its activity and
pharmacological profile through synthetic modification
of the structure. Indeed a number of newpaclitaxel
analogues have been synthesized showing improvements
in both parameters.2
1b. This analogue is interesting in that not only does it
possess an alkyl group at C10, it also has a spiro-fused
ring system, which might in theory affect the conforma-
tion of the cyclooctane ring. This modification is also
interesting from another perspective in that it provides a
handle for the introduction of additional functionality.
The route we envisioned for the synthesis of the C-10
epoxide of paclitaxel proceeds through the C-10 ketone
2. Methods for the synthesis of this intermediate have
previously been described.3 Our synthesis of this inter-
mediate, shown in Scheme 1, is similar to the literature
method with a few minor variations. Thus, the C-10
acetate group was selectively removed by treatment with
ꢀ
hydrazine at lowtemperature (0 C). Following this, the
C-20 hydroxyl group was protected with the t-butyldi-
methylsilyl (TBDMS) group, followed by masking of
the C-7 hydroxy group as the corresponding triethylsilyl
(TES) ether. Once these positions were protected, oxi-
dation of the C-10 position was achieved using TPAP.4
The resulting ketone group serves as a convenient
synthetic handle on which to build the C-10 epoxide.
As part of our efforts in this area, we were interested in
expanding the SAR-investigation of the C10-position.
Previous studies at this position have been primarily
focused on modifications of the ester group and manip-
ulating the oxidation state of C10. Despite the amount
of work carried out at this position, little attention has
been paid to introducing alkyl substituents at C10. As a
preliminary entry into this area we decided to synthesize
Of the available procedures for converting a carbonyl
group to the corresponding epoxide we viewed the
dimethylsulfonium methylide approach to be the most
readily accessible. As shown in Scheme 1, complete
consumption of the starting material was achieved when
5 equiv of this reagent were used. Reaction with the
ylide yielded two products having equal molecular
weight. We initially assumed these to be the two possi-
ble isomeric-epoxides resulting from attack of the
*Corresponding author. Fax: +1-203-677-7702; e-mail: michael.a.
walker@bms.com
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00280-3

1684
M. A. Walker et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1683–1685
carbonyl from the a- and the b-face. However, compar-
ison of the NMR of these two compounds was not
consistent with this assumption but fit more with 3 and
4.5 For example, the 13C NMR shift observed for the C9
carbonyl of 4 (150 ppm) is upfield compared to the sig-
nal for the same carbon atom in 3 (202 ppm) consistent
Subsequent manipulations of intermediate 3, en route to
1b, uncovered an interesting protocol for selective
removal of either the C2-0 or the C7-silyl protecting
group. Following a published procedure for complete
desilylation of protected paclitaxel intermediates,10 we
observed incomplete removal of the C20-TBDMS.
HFꢁ pyridine appeared to be selective for removing the
triethylsilyl group since the 20-desilylated product was
not observed. Under slightly milder reaction conditions
the reaction could be controlled to deliver the 7-desilyl-
ated product almost exclusively. This result was not
completely unexpected since there is literature precedent
1
with an ester rather than a ketone.6 In addition, the H
NMR signal for the newCH -group of 3 was found at
2
3.60 ppm, while the analogous signal for 4 was at 5.0
ppm suggesting a terminal olefin.7 This unusual by-
product appears to be newin the arena of sulfur ylide
chemistry, although a similar reaction has been descri-
bed for bismuthonium ylides.8 Assignment of the epox-
ide stereochemistry of 3 was established through the
observance of an NOE signal enhancement between the
bridgehead methyl group (C17) and the epoxide meth-
ylene. The high facial selectivity obtained in the epox-
idation step is not surprising considering that
nucleophile approach from the opposite face is sterically
blocked.9
for the removal of the TES protecting group in the pre-
11
.
sence of TBDMS using HF pyridine. However, we
were delighted to discover that we could effect the
reverse selectivity, that is to remove the TBDMS group
while leaving the TES-group in place, by switching to
TBAF as the desilylating agent. Thus as shown in Scheme
2, stirring 3 with 1.1 equiv of TBAF for 15 min yielded 6
almost exclusively. To the best of our knowledge this
Scheme 1. Reagents and conditions: (a) N2H2ꢁ H2O, EtOH, 0 ꢀC, 2 h; (b) 2.0 equiv TBDMSCl, 2.0 equiv imidazole, DMF, 0 ꢀC, 24 h; (c) 2.2 equiv
TESCl, 2.4 equiv iPr2NEt, 0.4 equiv DMAP, CH2Cl2, ꢂ40 to 0 ꢀC, 24 h; (d) TPAP (cat.), 1.5 equiv NMO, 4 A sieves, CH2Cl2; (e) 5.0 equiv
Me3S+Iꢂ, 5.0 equiv KHMDS, THF ꢂ40 ꢀC.
Scheme 2. (a) 10:1, 48%HF/pyridine, CH3CN, 0 ꢀC, 2 h; (b) 2.3:1 48%HF/pyridine, CH3CN, 0 ꢀC, 3 h; (c) 1.1 equiv TBAF, THF, 0 ꢀC, 15 min; (d)
30 equiv Me2S, 7.0 equiv Bz2O2, CH3CN, 0 ꢀC, 0.5 h; (e) 4.0 equiv I2 4 A sieves, MeOH, 7 h; (f) 1.5 equiv TBAF, THF, 0 ꢀC, 0.5 h.

M. A. Walker et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1683–1685
1685
Table 1. In vitro activity (IC50) of paclitaxel analogues (nM)
found to be 6 times less potent than the parent in a
previous study.15
Compd
Tub. poly. ratioa
HCT 116
HCT VM46
A2780
1a
1b
1c
1
0.5
0.7
2.9
2.0
1.5
>118
>121
14
3.5
2.5
1.8
References and Notes
aThe ratio in the tubulin polymerization assay is the potency of the
analogue/the potency of paclitaxel. Ratios less than 1 reflect analogues
that are more potent than paclitaxel.
1. Goldspeil, B. R. Pharmacotherapy 1997, 17, 125.
2. (a) Lin, S.; Ojima, I. Expert Opin. Ther. Pat. 2000, 10, 869.
(b) Kingston, D. G. I. J. Nat. Prod. 2000, 63, 726.
3. (a) Datta, A.; Vander Velde, D. G.; Georg, G. I.; Himes,
R. H. Tetrahedron Lett. 1995, 36, 1985. (b) Kingston, D. G. I.;
Samaranayake, G.; Ivey, C. A. J. Nat. Prod. 1990, 53, 1.
4. Griffith, W. P.; Ley, S. V. Aldrichim. Acta 1990, 23, 13.
appears to be the first example of TBDMS removal in
the presence of a TES group.12
1
5. (a) H NMR of 3 (500 MHz, CDCl3) d ꢂ0.31 (s, 3), ꢂ0.03
Derivatization of the C20- and C7-positions of paclitaxel
has been exploited in the past to yield prodrugs and
analogues with improved activity or pharmacological
properties.13 Therefore, in order to demonstrate the
utility of our selective deprotection method we decided
to introduce additional modifications to the C10-epox-
ide analogue. Intermediate 5 was readily converted to
the corresponding C7-MOM (1c) analogue. The two-
step procedure depicted in Scheme 2 was found to afford
the MOM-ether in much higher yield than reaction of
the C7-hydroxyl with MOMCl/Et3N.
(s, 3), 0.56 (m, 6), 0.79 (s, 9), 0.91 (t, 9, J=8), 1.11 (s, 3), 1.17
(s, 3), 1.69 (s, 3), 1.77 (m, 2), 2.03 (s, 3), 2.15 (m, 1), 2,46 (m,
1), 2.49 (m, 1), 2.56 (s, 3), 3.60 (d, 1, J=7), 4.15 (d, 1, J=7),
4.20 (dd, 1, J=8, 60), 4.58 (m, 1), 4.66 (m, 1), 4.96 (d, 1, J=9),
1
5.71 (s, 1), 5.72 (s, 1), 6.12 (m, 1), 7.29–8.12 (m, 15). (b) H
NMR spectra for 4 (500 MHz, DMSO) d ꢂ0.01 (s, 3), 0.03 (s,
3), 0.53 (m, 6), 0.75 (s, 9), 0.84 (t, 9, J=8), 0.95 (s, 3), 1.04 (s,
3), 1.45 (s, 3), 1.50 (s, 3), 1.70 (m, 2), 2.17 (m, 1), 2.34 (m, 1),
2.48 (s, 3), 3.17 (d, 1, J=4), 4.14 (dd, 2, J=7, 11), 4.48 (m, 1),
4.74 (d, 1, J=8), 4.88 (m, 1), 4.93 (s, 1), 5.04 (s, 1), 5.28 (d, 1,
J=4), 5.55 (d, 1, J=8), 5.74 (m, 1), 7.24–7.98 (m, 15).
6. The assignment of this shift was established by a 2-D
HMBC spectrum showing correlation between the proton sig-
nal at C19 and the C9 carbon.
7. 2-D HMBC showed correlation between the olefin 1H
NMR signal and the C11-carbon.
8. Rahman, M. M.; Matano, Y.; Suzuki, H. Chem. Commun.
1998, 13, 1359.
Final products 1b and 1c were tested for in vitro activity
against a number of cell lines.14 IC50 values for
HCT116, a colon carcinoma derived cell line and A2780
an ovarian cell line, are comparable to paclitaxel. Inter-
estingly enough, when these compounds were tested
against a paclitaxel-resistant cell line, HCT VM46, 1c
retained activity while 1b was much less potent (Table 1).
9. The steric environment around the C10 is readily apparent
in the X-ray structure of the taxane core. See: Alstadt, T.;
Gao, Q.; Wittman, M. D.; Kadow, J. F.; Vyas, D. M. Tetra-
hedron Lett. 1998, 39, 4965.
10. Chen, S.-H.; Farina, V.; Vyas, D. M.; Doyle, T. W.; Long,
B. H.; Fairchild, C. J. Org. Chem. 1996, 61, 2065.
11. Holton, R. A.; Kim, H.-B.; Somoza, C.; Liang, F.; Bide-
ger, R. J.; Boatman, P. D.; Shindo, M.; Smith, C. C.; Kim, S.;
Nadizadeh, H.; Suzuki, Y.; Tao, C.; Vu, P.; Tang, S.; Zhang,
P.; Murthi, K. K.; Gentile, L. N.; Liu, J. H. J. Am. Chem. Soc.
1994, 116, 1599.
12. Nelson, T. D.; Crouch, R. D. Synthesis 1996, 1031.
13. For an example, see: Golik, J.; Wong, H. S. L.; Chen,
S. H.; Doyle, T. W.; Wright, J. J. K.; Knipe, J.; Rose, W. C.;
Casazza, A. M.; Vyas, D. M. Bioorg. Med. Chem. Lett. 1996,
6, 1837.
14. H NMR of 1c (300 MHz, CDCl3) d 1.14 (s, 3), 1.19 (s, 3),
1.7 (s, 3), 1.94 (s, 3), 2.04 (m, 2), 2.32 (m, 2), 2.37 (s, 3), 2.62 (d,
1, J=7), 2.79 (m, 1), 3.28 (s, 3), 3.68 (d, 1, J=7), 4.17 (m, 3),
4.30 (m, 1), 4.43 (d, 1, J=7), 4.77 (s, 1), 4.79 (s, 1), 4.95 (d, 1,
J=9), 5.71 (d, 1, J=6), 5.81 (m, 1), 6.07 (m, 1), 7.26–8.13 (m,
15).
15. Kant, J.; O’Keeffe, W. S.; Chen, S.-C.; Farina, V.; Fair-
child, C.; Johnston, K.; Kadow, J. F.; Long, B. H.; Vyas, D.
Tetrahedron Lett. 1994, 35, 5543.
A convenient method for converting the C-10 position
of paclitaxel to the corresponding spiro-epoxide has
been developed. The installation of the epoxide func-
tionality using sulfonium ylide chemistry results in ste-
reospecefic ring formation along with a unique by-
product. Further manipulations at C20 or C7 can be
achieved using newly discovered methods for selective
desilylation of either position. The C10-epoxide analo-
gue of paclitaxel and the corresponding C7-MOM deri-
vative, synthesized according to this method, were
found to be equipotent to the parent in in vitro assays
measuring tubulin binding and cytotoxicity tissue cul-
ture. These results suggest that epoxidation of the C10-
position according to the current method is well toler-
ated. This would be in agreement with previous results,
which showed that the C10-position can tolerate a cer-
tain degree of modification. However, it should be
noted that the C10-methyl ether analogue of paclitaxel,
which can be considered an acyclic version of 1b, w as