Stereoselective synthesis of 12,13-cyclopropyl-epothilone B and side-chain-modified variants

Article abstract of DOI:10.1021/ol200114k

A general strategy has been devised for the stereoselective synthesis of 12,13-cyclopropyl-epothilone B and side-chain-modified variants thereof, which relies on late stage introduction of the heterocycle through Wittig olefination of ketone 14. Formation of the macrocycle was achieved through RCM-based ring closure and introduction of the cyclopropane moiety involved a highly selective Charette cyclopropanation of allylic alcohol 7.

Full text of DOI:10.1021/ol200114k

ORGANIC
LETTERS
2011
Vol. 13, No. 6
1436–1439
Stereoselective Synthesis of 12,13-
Cyclopropyl-Epothilone B and Side-Chain-
Modified Variants
†
Raphael Schiess, Jurg Gertsch, W. Bernd Schweizer, and Karl-Heinz Altmann*
‡
†
,†
€
Swiss Federal Institute of Technology (ETH) Zu€rich, HCI H405, Wolfgang-Pauli-Str. 10,
CH-8093 Zu€rich, Switzerland, and University of Bern, Bu€hlstrasse 28,
CH-3012 Bern, Switzerland
karl-heinz.altmann@pharma.ethz.ch
Received January 14, 2011
ABSTRACT
A general strategy has been devised for the stereoselective synthesis of 12,13-cyclopropyl-epothilone B and side-chain-modified variants thereof,
which relies on late stage introduction of the heterocycle through Wittig olefination of ketone 14. Formation of the macrocycle was achieved
through RCM-based ring closure and introduction of the cyclopropane moiety involved a highly selective Charette cyclopropanation of allylic
alcohol 7.
Epothilones are 16-membered bacterial macrolides with
potent antitumor activity1 that have served as lead struc-
tures for a series of clinical candidates for cancer treatment
(Figure 1).2 Among these, the lactam analog of epothilone
B (Epo B) recently has been approved by the FDA for
clinical use in breast cancer patients (ixabepilone, Ixempra).3
The chemistry and biology of epothilones have been
extensively studied, and numerous analogs or derivatives
havebeendiscoveredthatretain potentinvitro activity.1 Of
particular interest is the observation that the 12,13-cyclo-
propyl (CP) analogs of Epo A and B have been found to be
equally potent, or even somewhat more potent than the
respective epoxide-based parent compounds.4 At the same
time the replacement of the natural 12,13-epoxide moiety
by a cyclopropane ring should lead to enhanced chemical
and, in particular, metabolic stability, given the suscepti-
bility of the oxirane ring to undesired chemical trans-
formations5 and metabolic attack.6 Based on these con-
siderations, we have become interested in CP-Epo B
analogs as active drug moieties for antibody-drug conju-
gates (ADC)7 and, thus, have sought efficient synthetic
access to such structures. The initial results of this work are
described in this communication.8
† Swiss Federal Institute of Technology (ETH) Z€urich.
‡ University of Bern.
(1) (a) Altmann, K.-H.; Pfeiffer, B.; Arseniyadis, S.; Pratt, B. A.;
€
Nicolaou, K. C. ChemMedChem 2007, 2, 396. (b) Hofle, G.; Reich-
(5) Xu, X.; Zeng, J.; Mylott, W.; Arnold, M.; Waltrip, J.; Iacono, L.;
Mariannino, T.; Stouffer, B. J. Chromatogr. B 2010, 878, 525.
enbach, H. Epothilone, a myxobacterial metabolite with promising
antitumor activity. In Anticancer Agents from Natural Products; Cragg,
G. M., Kingston, D. G. I., Newman, D. G., Ed.; Taylor & Francis: Boca
Raton, FL, 2005; pp 413.
€
ꢀ
(6) (a) Comezoglu, S. N.; Ly, Van T.; Zhang, D.; Humphreys, W. G.;
Bonacorsi, S. J.; Everett, D. W.; Cohen, M. B.; Gan, J.; Beumer, J. H.;
Beunen, J. H.; Schellens, J. H. M.; Lappin, G. Drug Metab. Pharmaco-
(2) For a review, see: Coley, H. M. Cancer Treat. Rev. 2008, 34, 378.
(3) http://www.cancer.gov/cancertopics/druginfo/fda-ixabepilone
(4) Johnson, J.; Kim, S. H.; Bifano, M.; DiMarco, J.; Fairchild, C.;
Gougoutas, J.; Lee, F.; Long, B.; Tokarski, J.; Vite, G. Org. Lett. 2000, 2,
1537.
€
€
kinet. 2009, 24, 511. (b) Blum, W.; Aichholz, R.; Ramstein, P.; Kuhnol,
€
€
J.; Bruggen, J.; O’Reilly, T.; Florsheimer, A. Rapid Commun. Mass
Spectrom. 2001, 15, 41.
(7) For a recent review on ADCs, see: Senter, P. Curr. Opin. Chem.
Biol. 2009, 13, 235.
r
10.1021/ol200114k
Published on Web 02/21/2011
2011 American Chemical Society

Scheme 1. Retrosynthesis of CP-Epo B 1a
Figure 1. Structures of natural epothilones A and B and of target
structures 1a-e.
CP analogs of Epo A and B have been previously
prepared by semisynthesis from Epo C and D, respectively,
albeit in low yield;4 CP-Epo A has also been obtained by
total chemical synthesis9 and so have some side-chain-
modified variants thereof.9 In addition, the Nicolaou
group has reported a series of side-chain-modified analogs
of 12,13-trans CP-Epo A9 and CP-Epo B;10 in contrast,
only two examples of side-chain-modified variants of (cis)
CP-Epo B are known11 and no chemical synthesis of CP-
Epo B itself has in fact been documented in the literature.12
Nicolaou’s synthesis of CP-epothilones is based on a
convergent strategy that relies on the late stage addition of
a side chain vinyl iodide to a C15 aldehyde (epothilone
numbering) by means of a Nozaki-Hiyama-Kishi cou-
pling as a key step.9 This strategy has been instrumental in
obtaining analogs for biological testing, but, unfortu-
nately, the Nozaki-Hiyama-Kishi coupling is essentially
nonselective, thus leading to a substantial loss of material
at a late stage of the synthesis. In an attempt to overcome
this limitation we have evaluated an alternative strategy to
CP-Epo B analogs, where the heteroaryl-vinyl side chain
was to be established in the last step of the synthesis (safe
for the final deprotection) through Wittig-type chemistry
with a ketone I-1 (Scheme 1). An analogous approach has
been employed by Danishefsky and co-workers in the
synthesis of 9,10-dehydro-Epo D13 and its 26-trifluoro-
methyl variant (fludelone)13,14 as well as by Avery and
co-workers in their synthesis of Epo A (to generate the
epoxidation precursor Epo C);15 at the same time, Hofle
has been unable to re-establish Epo A from the corre-
sponding (epoxide-containing) side chain ketone, in spite
of significant optimization attempts.16,17
€
As illustrated in Scheme 1, ketone I-1 was to be obtained
through ring-closing metathesis (RCM) between C9 and
C10 of the desired macrocycle followed by selective double
bond reduction. The requisite diene precursor for the
macrocyclization reaction would in turn be assembled
from alcohol I-2 and acid I-3; the bis-TBS protected
version of the latter (Scheme 1, PG = TBS) had been
previously synthesized in our laboratory.18 The stereo-
selective establishment of the cyclopropane moiety in I-2
was to be achieved through Charette cyclopropanation of
allylic alcohol I-4, which would be derived from keto
aldehyde I-5 by means of Still-Gennari olefination and
subsequent two-carbon extension. Based on literature
precedence it was felt that Still-Gennari olefination of
the aldehyde group would be feasible selectively in the
presence of the methyl ketone,19 while the keto group
would have to be protected in subsequent steps. Lastly,
aldehyde I-5 was planned to be prepared from S-malic acid
as a defined source of chirality at C15.
The synthesis of intermediate 11 (i.e., synthon I-2) is
summarized in Scheme 2. Starting from S-malic acid,
hydroxy lactone 2 was prepared in a 3-step literature
sequence.20 Treatment of 2 with MeLi gave a mixture of
cyclic hemiacetal 3 and the corresponding open chain
hydroxy ketone. Synthetically useful conversion of this
mixture into aldehyde 4 could only be achieved by
(8) Heterocycles were selected based on previous SAR data on other
types of potent side-chain-modified epothilones1 and the availability of
the corresponding epoxides for biological comparison (for 16e and d).
ꢁ
(9) Nicolaou, K. C.; Namoto, K.; Ritzen, A.; Ulven, T.; Shoji, M.; Li,
J.; D’Amico, G.; Liotta, D.; French, C. T.; Wartmann, M.; Altmann
K.-H.; Giannakakou, P. J. Am. Chem. Soc. 2001, 123, 9313.
ꢁ
(10) Nicolaou, K. C.; Ritzen, A.; Namoto, K.; Buey, R. M.; Dıaz,
´
J. F.; Andreu, J. M.; Wartmann, M.; Altmann, K.-H.; O’Brate, A.;
Giannakakou, P. Tetrahedron 2002, 58, 6413.
(11) Nicolaou, K. C.; Sasmal, P. K.; Rassias, G.; Reddy, M. V.;
Altmann, K.-H.; Wartmann, M.; O’Brate, A.; Giannakakou, P. Angew.
Chem., Int. Ed. 2003, 42, 3515.
(15) Hindupur, R. M.; Panicker, B.; Valluri, M.; Avery, M. A.
Tetrahedron Lett. 2001, 42, 7341.
(16) Sefkow, M.; Kiffe, M.; Schummer, D.; Hofle, G. Bioorg. Med.
(12) Based on literature data for two examples 12,13-trans configured
CP epothilones of type B (methyl group on C12) appear to be signifi-
canly less active than the corresponding cis derivatives.10,11
(13) (a) Rivkin, A.; Yoshimura, F.; Gabarda, A. E.; Cho, Y. S.;
Chou, T.-C.; Dong, H. J.; Danishefsky, S. J. J. Am. Chem. Soc. 2004,
126, 10913. (b) Rivkin, A.; Yoshimura, F.; Gabarda, A. E.; Chou, T.-C.;
Dong, H. J.; Tong, W. P.; Danishefsky, S. J. J. Am. Chem. Soc. 2003,
125, 2899.
€
Chem. Lett. 1998, 8, 3025.
(17) The preparation of side-chain-modified analogs of 9,10-dehy-
dro-Epo D from the corresponding ketone has been disclosed in a PCT
patent application (WO2004043954 (A2)), without any comments on the
stereochemical outcome of the Wittig reactions.
(18) Feyen, F.; Jantsch, A.; Hauenstein, K.; Pfeiffer, B.; Altmann,
K.-H. Tetrahedron 2008, 64, 7920.
(19) Seifert, A.; Vomund, S.; Grohmann, K.; Kriening, S.; Urlacher,
(14) Chou, T.-C.; Dong, H. J.; Rivkin, A.; Yoshimura, F.; Gabarda,
A. E.; Cho, Y. S.; Tong, W. P.; Danishefsky, S. J. Angew. Chem., Int. Ed.
2003, 42, 4761.
V. K.; Laschat, S.; Pleiss, J. ChemBioChem 2009, 10, 853.
Org. Lett., Vol. 13, No. 6, 2011
1437

Scheme 2. Synthesis of Building Block 11
Scheme 3. Assembly of Macrolactone 14
saturated alcohol 10. An X-ray crystal structure of the
allylic alcohol intermediate obtained from 9 confirmed the
predicted stereochemical outcome of the Charette cyclo-
propanation reaction.25 Finally, seleniumoxide pyrolysis26
and TBS deprotection provided alcohol 11 in a total of 12
steps and excellent overall yield (25%) from lactone 2.
As illustrated in Scheme 3 a high yielding Yamaguchi
esterification27 of alcohol 11 and acid 12 (i.e., I-3 with
PG = TBS, Scheme 1)18 gave the requisite diene for RCM-
based macrocyclization (DCC- or EDCI-based protocols
were less effective); this diene underwent smooth RCM in
the presence of second generation Grubbs catalyst in
refluxing DCM to provide macrolactone 13 as a 12/1
mixture of E/Z isomers. In contrast, the use of toluene as
a solvent only gave traces of product.
Reduction of the C9, C10 double bond in macrolactone
13 proved to be highly challenging, due to simultaneous
cyclopropane ring-opening. After extensive optimization,
hydrogenation over Lindlar catalyst at a hydrogen pres-
sure of 7.5 bar was found to provide the saturated macro-
cycle most efficiently (80% yield). While cyclopropane
ring-opening could not be avoided completely even under
these conditions, it occurred at a much slower rate than
double bond reduction.
Dess-Martinoxidation, while all otheroxidation methods
investigated did not provide any of the aldehyde (PDC,
Swern) or gave only low yields (py SO3). Subsequent
3
Still-Gennari olefination21 with phosphonate 522 furn-
ished the desired Z-isomer 6 exclusively; the geometry of
the double bond was firmly established by means of
NOESY experiments. Acetal protection of the ketone
functionality in 6 by treatment with ethylene glycol and
triethyl orthoformate followed by reduction of the ester
moiety with DIBAL-H led to allylic alcohol 7 in excellent
overall yield (94% for two steps); the latter underwent
highly stereoselective Charette cyclopropanation (dr 18:1)
to afford alcohol 8 in high yield (88% for single isomer).23
It should be noted that violent explosions have been
reported for Charette cyclopropanations carried out on
scales of 8 mmol or higher,23 due to the exothermicity
associated with the formation of Zn(CH2I)2. However,
careful temperature control during the addition of CH2I2
tothe (Et)2Zn solution allowed the cylopropanation of7 to
be carried out safely also on a larger scale.
Unfortunately, we were unable at this stage to accom-
plish the selective hydrolysis/removal of the cyclic acetal in
the presence of the two TBS ethers at C3 and C7, using a
variety of different conditions. As a consequence, the
reduction product was submitted to global deprotection,
With this key reaction successfully implemented, our
efforts were then directed toward installing the terminal
double bond required for RCM-based macrocyclization.
After Swern oxidation of 8, the resulting aldehyde was
subjected to Wittig-olefination with Ph3PCHCO2Et to
furnish R,β-unsaturated ester 9. Subsequent reduction of
the ester moiety using DIBAL-H followed by reduction of
which was best achieved with FeCl3 6H2O.28
3
Not unexpectedly, the resulting (unprotected) ketone
was found to be a poor substrate for olefinations with
heterocycle-containing phosporanes or phosphonate an-
ions, with target structures 1 being obtained in only
low and nonreproducible yields (0-25%). The hydroxyl
groups were thus reprotected as TMS ethers to produce the
the double bond with NaBH4/CoCl2 6H2O24 provided
3
(20) (a) Denmark, S. E.; Yang, S.-M. J. Am. Chem. Soc. 2002, 124,
15196. (b) Mulzer, J.; Mantoulidis, A.; Oehler, E. J. Org. Chem. 2000, 65,
7456. (c) White, J. D.; Hrnciar, P. J. Org. Chem. 2000, 65, 9129. (d)
Green, D. L. C.; Kiddle, J. J.; Thompson, C. M. Tetrahedron 1995, 51,
2865.
(21) Still, W. C.; Gennari, C. Tetrahedron Lett. 1983, 24, 4405.
(22) Phosphonate 5 is accessible from commerically available ethyl-
phosphonic dichloride in a two-step sequence (cf. ref 21).
(23) Charette, A. B.; Juteau, H.; Lebel, H.; Molinaro, C. J. Am.
Chem. Soc. 1998, 120, 11943.
(25) The structure has been deposited in the Cambridge Crystal-
lographic Data Base (deposition number CCDC 808039).
(26) (a) Grieco, P. A.; Gilman, S.; Nishizawa, M. J. Org. Chem. 1976,
41, 1485. (b) Sharpless, K. B.; Young, M. W. J. Org. Chem. 1975, 40, 947.
(27) Inanaga, J.; Hirata, K.; Saeki, H.; Katsuki, T.; Yamaguchi, M.
Bull. Chem. Soc. Jpn. 1979, 52, 1989.
(28) Sen, S. E.; Roach, S. L.; Boggs, J. K.; Ewing, G. J.; Margrath, J.
J. Org. Chem. 1997, 62, 6684.
(29) For experimental details see: Dietrich, S. A.; Lindauer, R.;
Stierlin, C.; Gertsch, J.; Matesanz, R.; Notararigo, S.; Dı
Altmann, K.-H. Chem.;Eur. J. 2009, 15, 10144.
´
az, J. F.;
(24) He, R.; Deng, M. Z. Tetrahedron 2002, 58, 7613.
1438
Org. Lett., Vol. 13, No. 6, 2011

Table 2. Antiproliferative Activity of CP-Epo B Analogs 1a-e
Scheme 4. Synthesis of Target Structures 1a-e (cf. Figure 1)
(IC50 values [nM]a; after 72 h exposure time)29
cell line
1a
1b
1c
1d
1e
A549
0.70 ( 0.2 0.30 ( 0.07 4.7 ( 0.8 0.90 ( 0.13 1.7 ( 0.15
MCF-7 0.80 ( 0.3 0.40 ( 0.06 8.5 ( 1.5 0.80 ( 0.19 1.9 ( 0.23
HCT116 1.30 ( 0.2 0.30 ( 0.03 3.8 ( 0.7 0.40 ( 0.08 0.90 ( 0.07
a IC50 values of 0.33, 0.34, and 0.16 nM have been reported for Epo B
against the A549, MCF-7, and HCT116 cell lines, respectively.29
Deprotection of 16a-e was achieved with citric acid to
provide CP-Epo B analogs 1a-e in excellent yields
(Scheme 4). For 1d, the minor isomer formed in the Wittig
reaction could be removed by preparative HPLC, whereas
1e could only be obtained as a 10:1 mixture of the desired
stucture and its presumed C15 epimer (and was evaluated
as such in cellular experiments).
The antiproliferative activity of CP-Epo B analogs 1a-e
was tested against the human cancer cell lines A549 (lung),
MCF-7 (breast), and HCT116 (colon) (Table 2).
Table 1. Wittig Reactions with Ketone 14
Wittig salt
base
temperature
product E:Z
yield
All compounds are potent inhibitors of cancer cell
proliferation, with IC50 values in the single digit nM or
even sub-nM range. Compared to Epo B, CP-Epo B (1a)
appears tobe2-8-foldlessactive; IC50 valuessimilartothe
corresponding epoxide-based analogs30 were also ob-
served for 1d/1e (<3-fold difference in all cases), although
1d and 1e showed a tendency for slightly enhanced activity.
The most potent compound investigated was isoxazole
derivative 1b, which is in line with previous findings by the
Danishefsky group on the activity of isoxazole-containing
variants of 9,10-dehydro-12,13-deoxy-epothilones.31 In
light of its sub-nM potency 1b is an attractive candidate
for the construction of ADCs.
In conclusion, we have established an efficient synthesis
of CP-based Epo B analogs, which relies on ketone 14 as a
highly advanced precursor for the late stage incorporation
of the side chain heterocycle; as such this strategy enables
convenient access to a wide range of side-chain-modified
derivatives. We are currently pursuing the synthesis of
further CP-Epo B analogs and of conjugates of such
compounds with tumor-targeted antibodies. The results
of these studies will be reported in due course.
15a
15b
15c
15d
15e
KHMDS -78 to -20 °C
KHMDS -78 to -20 °C
n-BuLi
n-BuLi
n-BuLi
16a
16b
16c
16d
16e
6:1 67%a
17:1 74%a
13:1 85%a
10:1 90%b
7:1 54%b,c
-78 to -20 °C
-78 to 25 °C
-78 to 75 °C
a Mixture of E and Z isomers. b Mixture of E isomer and presumed
C15 epimer (see text). c 77% based on recovered starting material.
bis-TMS protected macrolactone 14 (i.e., I-1 with
PG = TMS, Scheme 1) in quantitative yield (Scheme 3).
As illustrated in Scheme 4 and Table 1, methyl ketone 14
proved to be a highly suitable substrate for heterocycle
attachment through Wittig olefination.
In all cases investigated, the reactions proceeded with
good selectivity in favor of the desired E isomers which
could be isolated in acceptable to good yields (Table 1).
However, distinct differences were observed between dif-
ferent phosphoranes with regard to selectivity and also
reactivity (Table 1).
Thus, while 16a-c were already formed upon warming
of the reaction mixture to -20 °C (after deprotonation of
the phosphonium salts at -78 °C and addition of 14 at the
same temperature), the reaction of 14 with the phospho-
ranes derived from 15d and 15e required temperatures of
25 °C and, quite remarkably, 75 °C, respectively. Attempts
to accelerate the reaction of 14 with the phosphorane
derived from 15d by raising the temperature to 75 °C
resulted in decomposition.
Acknowledgment. This work was supported by the
ETH Research Grant CH1-01 08-3. We are indebted to
Dr. Bernhard Pfeiffer (ETHZ) for NMR support, to Kurt
Hauenstein (ETHZ) for advice, and toLouis Bertschi from
the ETHZ-LOC MS-Service for HRMS spectra.
Supporting Information Available. Synthetic proce-
1
Noteworthy, after chromatographic separation of 16d
1
and 16e from the corresponding Z isomers, their H and
dures, complete spectroscopic data, H and 13C NMR
data for all new compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
13C NMR spectra still indicatedthe presence of anisomeric
impurity that could not be separated (ca. 12%). We have
not established the identity of this impurity, but it is well
conceivable that the stereocenter at C15 partially epi-
merizes under the more forcing conditions of the Wittig
reaction required for 16d and 16e.
(30) These compounds were kindly provided by the Novartis Insti-
tute for Biomedical Research in Basel, Switzerland.
(31) Chou, T.-C.; Zhang, X.; Zhong, Z.-Y.; Li, Y.; Feng, L.; Eng, S.;
Myles, D. M.; Johnson, R., Jr.; Wu, N.; Yin, Y. I.; Wilson, R. M.;
Danishefsky, S. J. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 13157.
Org. Lett., Vol. 13, No. 6, 2011
1439