An aprotic Tamao oxidation/syn-selective tautomerization reaction for the efficient synthesis of the C(1)-C(9) fragment of fludelone

Article abstract of DOI:10.1021/ol302221s

An efficient synthesis of the C(1)-C(9) fragment of fludelone has been developed. The key step is a tandem silylformylation-crotylsilylation/ Tamao oxidation sequence that establishes the C(5) ketone, the C(6), C(7), and C(8) stereocenters, and the C(9) alkene in a single operation from a readily accessed starting material. The stereochemical outcome at C(6) depends critically on the development of an aprotic Tamao oxidation, which leads to a reversal in the intrinsic diastereoselectivity observed using standard Tamao oxidation conditions.

Full text of DOI:10.1021/ol302221s

ORGANIC
LETTERS
2012
Vol. 14, No. 18
4890–4893
An “Aprotic” Tamao Oxidation/
Syn-Selective Tautomerization Reaction
for the Efficient Synthesis of the
C(1)ÀC(9) Fragment of Fludelone
Tyler J. Harrison, Philippe M. A. Rabbat, and James L. Leighton*
Department of Chemistry, Columbia University, New York, New York 10027,
United States
leighton@chem.columbia.edu
Received August 9, 2012
ABSTRACT
An efficient synthesis of the C(1)ÀC(9) fragment of fludelone has been developed. The key step is a tandem silylformylationÀcrotylsilylation/
Tamao oxidation sequence that establishes the C(5) ketone, the C(6), C(7), and C(8) stereocenters, and the C(9) alkene in a single operation from a
readily accessed starting material. The stereochemical outcome at C(6) depends critically on the development of an “aprotic” Tamao oxidation,
which leads to a reversal in the intrinsic diastereoselectivity observed using “standard” Tamao oxidation conditions.
Polyketide natural products continue to influence small
molecule drug development efforts. A prominent recent
case in point is the epothilone family of natural products
(Figure 1), potent anticancer compounds that operate by a
microtubule-stabilizing mechanism of action.1 Significant
effort has been devoted to the development of analogs2
with improved pharmacokinetic profiles, and one of the
more promising compounds to emerge from these efforts
is fludelone (C(12)-trifluoromethyl-(E)-9,10-dehydrodes-
oxyepothilone B). Developed by Danishefsky and co-
workers,3 this compound is both more potent than epothi-
lone B and more effective in vivo, resulting in the shrinking
of tumors in mice to the point of nondetectability. Given
the extraordinary activity and the structural and stereo-
chemical complexity of fludelone, and given that it is
increasingly likely that other polyketide natural product
analogs will be developed as drug candidates, it is impera-
tive that ever more efficient synthetic chemistry is advanced
to keep pace. The Danishefsky team assembled fludelone in
three operations from fragments 1, 2, and 3 (Figure 1).3a
From a stereochemical standpoint, the C(1)ÀC(9) frag-
ment 1 is the most complex, and its synthesis is the subject
of this report.
We have been engaged for some time in the develop-
ment of tandem reaction strategies for the synthesis of
polyketide fragments4 and recently described the one-step
(1) (a) Altmann, K.-H.; Pfeiffer, B.; Arseniyadis, S.; Pratt, B. A.;
Nicolaou, K. C. ChemMedChem 2007, 2, 396. (b) Altmann, K.-H.;
(4) (a) Zacuto, M. J.; Leighton, J. L. J. Am. Chem. Soc. 2000, 122,
8587. (b) O’Malley, S. J.; Leighton, J. L. Angew. Chem., Int. Ed. 2001, 40,
2915. (c) Zacuto, M. J.; O’Malley, S. J.; Leighton, J. L. J. Am. Chem. Soc.
2002, 124, 7890. (d) Wang, X.; Meng, Q.; Nation, A. J.; Leighton, J. L.
J. Am. Chem. Soc. 2002, 124, 10672. (e) Schmidt, D. R.; O’Malley, S. J.;
Leighton, J. L. J. Am. Chem. Soc. 2003, 125, 1190. (f) Zacuto, M. J.;
O’Malley, S. J.; Leighton, J. L. Tetrahedron 2003, 59, 8889. (g) Schmidt,
D. R.; Park, P. K.; Leighton, J. L. Org. Lett. 2003, 5, 3535. (h) Wang, X.;
Meng, Q.; Perl, N. R.; Xu, Y.; Leighton, J. L. J. Am. Chem. Soc. 2005,
127, 12806. (i) Spletstoser, J. T.; Zacuto, M. J.; Leighton, J. L. Org. Lett.
2008, 10, 5593.
€
€
Hofle, G.; Muller, R.; Mulzer, J.; Prantz, K. In Progress in the Chemistry
of Organic Natural Products, Vol. 90: The Epothilones: An Outstanding
Family of Anti-Tumor Agents. From Soil to the Clinic; Kinghorn, A. D.,
Falk, H., Kobayashi, J., Eds.; Springer: Vienna, 2009; p 260.
(2) For a recent review and lead references, see: Watkins, E. B.;
Chittiboyina, A. G.; Avery, M. A. Eur. J. Org. Chem. 2006, 4071.
(3) (a) Rivkin, A.; Yoshimura, F.; Gabarda, A. E.; Cho, Y. S.; Chou,
T.-C.; Dong, H.; Danishefsky, S. J. J. Am. Chem. Soc. 2004, 126, 10913.
(b) Rivkin, A.; Chou, T.-C.; Danishefsky, S. J. Angew. Chem., Int. Ed.
2005, 44, 2838.
r
10.1021/ol302221s
Published on Web 09/05/2012
2012 American Chemical Society

Scheme 1. Proposal for a Method to Reverse the Intrinsic
Selectivity in the Tandem Tamao Oxidation/Diastereoselective
Tautomerization Reaction
Figure 1. Epothilones A and B, fludelone, and the three principal
fragments from which fludelone is assembled.
conversion of 4 to 5 as the centerpiece of an efficient
synthesis of zincophorin methyl ester (Scheme 1).5 This
complex series of chemical events transforms a propynyl
fragment, a butenyl fragment, a silyl hydride, CO, and
H2O2 into a six-carbon chain comprised of a ketone, three
stereocenters, and an alkene, and it accomplishes this
with no external stereochemical control, no protecting
group manipulations, and no nonstrategic redox steps.6
While the possibility that we might apply such a reaction
to a synthesis of 1;wherein the entire C(5)ÀC(9) segment
would be directly installed with everything in the right
oxidation state and orientation and without recourse to
any protecting groups (cf. 6 to 7);was exciting, it was also
clear that the reaction would provide the wrong diaste-
reomer at C(6). That stereocenter is established upon
tautomerization of the immediate product of the Tamao
oxidation,7 enol 8, and the observed anti selectivity
(relative to the vicinal hydroxyl group) is proposed to
arise from the illustrated enol conformation 8A in which
allylic 1,3 strain and related considerations dictate that
the terminal crotyl unit effectively blocks the back face
of the enol and that the highlighted hydroxyl group is
well positioned to deliver the proton to the front face of
the enol.4i Importantly, this model assumes that in the
polar, protic environment of the Tamao oxidation condi-
tions (aqueous H2O2 and MeOH as solvent), internal
hydrogen bonding within 8 does not play a signifi-
cant role. We hypothesized that if we could develop an
“aprotic” Tamao oxidation procedure, the conformation
of enol 8 would be dictated by internal hydrogen bonding
as in 8B. If so, tautomerization would then be expected
to produce the desired stereochemistry at C(6), by virtue
of the back face of the enol being significantly more
exposed in this conformation. We describe here the
development of an “aprotic” Tamao oxidation and its
application to an efficient synthesis of the fludelone
C(1)ÀC(9) fragment 1.
Our synthesis of silane 6 commenced from 9,8 which
was epoxidized with m-CPBA to give 109 in 97% yield
(Scheme 2). Regioselective opening of epoxide 10 followed
the procedure of Pagenkopf10 to deliver 11. Removal of the
TBS group was followed by oxidative cleavage of the diol
to give aldehyde 12, which was subjected to the enantiose-
lective ketene cycloaddition procedure of Nelson11 using
catalyst 13 to give β-lactone 14 in 56% overall yield (from 10)
and 87% ee. Alcoholysis with tert-butanol catalyzed
by KCN provided ester 15 in 77% yield, and finally, silane
alcoholysis with di-cis-crotylsilane4c provided 6 in 88%
yield. While more step-economical routes to 6 may be
imagined, this route served our purposes well in that it
provided for reliable access to 6 in 39% overall yield
from 9.
With access to 6 secured we were poised to develop a
syn-selective Tamao oxidation/tautomerization reaction.
Because unraveling the complex diastereoselectivity of
these tandem reactions is not trivial, however, it was
important first to examine in detail the performance of 6
in the tandem silylformylation/crotylsilylation reaction.
Thus, when 6 was subjected to the silylformylation reac-
tion conditions and the resulting mixture was treated with
(8) Volkert, M.; Uwai, K.; Tebbe, A.; Popkirova, B.; Wagner, M.;
Kuhlmann, J.; Waldmann, H. J. Am. Chem. Soc. 2003, 125, 12749.
(9) Wan, S.; Gunaydin, H.; Houk, K. N.; Floreancig, P. E. J. Am.
Chem. Soc. 2007, 129, 7915.
(5) Harrison, T. J.; Ho, S.; Leighton, J. L. J. Am. Chem. Soc. 2011,
133, 7308.
(6) Gaich, T.; Baran, P. S. J. Org. Chem. 2010, 75, 4657.
(7) (a) Tamao, K.; Kakui, T.; Akita, M.; Iwahara, T.; Kanatani, R.;
Yoshida, J.; Kumada, M. Tetrahedron 1983, 39, 983. (b) Jones, G. R.;
Landais, Y. Tetrahedron 1996, 52, 7599.
(10) Zhao, H.; Engers, D. W.; Morales, C. L.; Pagenkopf, B. L.
Tetrahedron 2007, 63, 8774.
(11) Zhu, C.; Shen, X.; Nelson, S. G. J. Am. Chem. Soc. 2004, 126,
5352.
Org. Lett., Vol. 14, No. 18, 2012
4891

Scheme 2. Synthesis of Di-cis-crotylsilyl Ether 6
Scheme 3. Performance of 6 in the Tandem Silylformylation/
Crotylsilylation Reaction and in the Tandem Tamao Oxidation/
Diastereoselective Tautomerization Reaction
n-Bu4NF, a 5:1 mixture of protodesilylated products 16
and 17 was produced, by way of a 5:1 mixture of initial
products 18 and 19 (Scheme 3). The 5:1 mixture of 18 and
19 was therefore to be the starting point for our survey of
Tamao oxidation conditions. In principle, Tamao oxida-
tion of this mixture can produce four diastereomers,
complicating our efforts to directly and clearly measure
the selectivity of the Tamao oxidation/tautomerization
of 18. When our “standard” oxidation conditions were
employed and resulted in the production of 20 as the major
product of a 5:1 mixture of diastereomers, it was clear that,
as expected, 18 had been transformed into 20 with high
levels of anti-diastereoselectivity, at C(6). Once we ob-
tained a sample of 7 (see below), we were then able to
screen, but we did find that nonpolar solvents generally led
to significantly more efficient reactions and eventually
settled on cyclohexane/carbonitrile13 (CyCN) as our stan-
dard solvent (entry 2). That left the fluoride source as
the remaining variable, and it was found that the use
of tetrabutylphosphonium fluoride (formed in situ from
tetrabutylphosphonium chloride and silver(I) fluoride14)
reproducibly led to slightly improved selectivity (entry 3).
We decided to prepare a series of phosphonium salts with
bulkier alkyl groups, and although this line of inquiry did
result in improved selectivity (entries 4 and 5), we could
achieve no better than ∼5:1 dr. The breakthrough came
when a repeat of entry 4 gave a 6:1 dr, and we determined
that the phosphonium salt used in that experiment was
contaminated with PCy3•HCl. Among other things, this
led us to examine amine•HF salts,14 and we were delighted
to find that Et3N•HF resulted in a 6.5:1 dr (entry 6). It
quickly became apparent that the structure of the amine
had a direct and significant impact on the diastereoselec-
tivity, with the HF salts of i-Pr2NEt and other bulkier
amines giving reduced selectivities. Me3N•HF was there-
fore examined, and our delight at the improved (10:1)
selectivity was tempered only by a dramatic reduction
in reaction efficiency (entry 7). Reasoning that the poor
oxidation efficiency in this case was likely due to poor
1
develop a simple H NMR assay for the reaction of 18
without interference from the presence of the oxidation
product(s) of 19.
As discussed in detail above, our central hypothesis was
that if the oxidation was carried out in an aprotic and
nonpolar environment, the enol resulting from oxidation
of 18 would undergo tautomerization from a hydrogen-
bonded conformation such as 8B, leading to desired
product 7. The first task was to identify an oxidant that
could be generated safely in a nonaqueous solution and
that would be effective in a relatively nonpolar solvent,
and we were intrigued by a report from Tamao that small
amounts of H2O2 could be generated in a controlled
fashion by the use of methylhydroquinone (MeHQ) in
concert with 1 atm of O2.12 When this procedure was
employed under otherwise “normal” Tamao conditions
(with n-Bu4NF in THF), we were delighted to find that the
near-total selectivity for 20had indeed been reversed with 7
produced as the major product with 2:1 selectivity, albeit
with poor efficiency (Table 1, entry 1). We were unable
to perturb this ratio significantly in an extensive solvent
(13) It is interesting to note that nitriles and H2O2 can combine to
form peroxyimidic acids (with PhCN, the product is Payne’s reagent;
see: Payne, G. B. Tetrahedron 1962, 18, 763). Because the formation of
the peroxyimidic acid seems to require mildly basic conditions (pH 8),
and because we have never observed any epoxidation of the terminal
alkene in these reactions, we do not believe this is relevant to the present
oxidation reaction, but we cannot definitively rule it out.
(14) For convenience, the phosphonium and ammonium fluoride
salts were in every case formed in situ by using the corresponding
chloride salts in combination with AgF. Control experimentsestablished
that the oxidation reactions do not work at all in the absence of either
component.
(12) Tamao, K.; Hayashi, T.; Ito, Y. Tetrahedron Lett. 1989, 30,
6533.
4892
Org. Lett., Vol. 14, No. 18, 2012

Table 1. Optimization of a Syn-Selective Tamao Oxidation/
Diastereoselective Tautomerization Reaction
Scheme 4. Application of Optimized Conditions to Silane 6 for
the Direct and Efficient Synthesis of the C(1)ÀC(9) Fragment of
Fludelone
entry
fluoride source
conversion/efficiencyb dr (7:20)c
1a n-Bu4NF
poor
good
good
good
good
good
poor
good
2:1
2:1
2
3
4
5
6
7
8
n-Bu4NF
n-Bu4PCl/AgF
3:1
Cy3PBuCl/AgF
4.5:1
5:1
Cy3PCH2CH(Et)2Cl/AgF
40% overall yield from 6. Based on experiments in which
7 was purified by more careful chromatography, we can
estimate that the yield for the transformation of 6 to 7 is
∼60%. Given the number of chemical bond making and
breaking events in the former transformation, we view the
60% yield as being a quite reasonable level of efficiency,
and we believe this transformation could serve as the
basis for a step-economical and scalable synthesis of acid
1/1•CyNH2.
The tandem silylformylationÀcrotylsilylation/Tamao
oxidation reaction allows the rapid buildup of molecular
complexity relevant to precious polyketide targets from
very simple building blocks (an alkyne, crotylsilanes, and
carbon monoxide). Prior to the work reported here, how-
ever, only one stereochemical permutation (anti) was ac-
cessible. We have now identified an alternate set of Tamao
oxidation conditions that allow access to the corresponding
syn isomer highly selectively with a substrate that is directly
relevant to the C(1)ÀC(9) fragment of fludelone. Notably,
this constitutes a rare example of the ability to select for
either diastereomer in a substrate-controlled reaction with
no external stereochemical influence simply by varying the
reaction conditions.
Et3N HCl/AgF
6.5:1
10:1
14:1
3
Me3N HCl/AgF
3
quinuclidine HCl/AgF
a This reaction was performed in THF. b Qualitative measure as
judged by 1H NMR analysis of the unpurified product mixture. c De-
termined by 1H NMR analysis of the unpurified product mixture.
solubility of the amine salt and that we therefore needed a
sterically small yet more hydrophobic amine, we turned
finally to the HF salt of quinuclidine and were gratified to
find that this led to a highly diastereoselective (14:1) and
efficient reaction in favor of the desired product 7 (entry 8).
Although these results suggest that the origins of the
diastereoselectivity are more complex than the hypothesis
presented in Scheme 1, it is nevertheless true that that
hypothesis was the origin of these experiments, and we
believe that the idea regarding hydrogen-bonded enol
conformation 8B may well be sound.
We next employed the optimized oxidation conditions in
a preparative reaction (Scheme 4). Thus, 6 was subjected to
the tandem silylformylationÀcrotylsilylation reaction, and
this was followed without purification by the optimized
“aprotic” Tamao oxidation (in preparative scale reactions,
it proved practically advantageous to use PhCN as the
solvent instead of CyCN in the oxidation reaction; no
significant difference in reaction efficiency or selectivity
was noted13). After filtration through a plug of silica gel, 7
was obtained as the major product of a mixture of di-
astereomers (7, 20, and the oxidation product(s) derived
from 19). We were able to confirm at this point that the
selectivity at C(6) for the major diastereomer from the first
part of the reaction (18) was 14:1. Treatment of this
mixture with TBSOTf resulted in TBS protection of
both alcohols and t-Bu ester cleavage to give acid 1. For
convenient purification and isolation, we then prepared
the salt 1•CyNH2 and isolated it by recrystallization in
Acknowledgment. This work was supported by a grant
from the National Institute of General Medical Sciences
(GM58133). T.H. was supported by a National Sciences
and Engineering Research Council of Canada (NSERC)
Postdoctoral Fellowship.
Supporting Information Available. Experimental pro-
cedures, characterization data, and stereochemical proofs
as well as X-ray crystallographic data for compound 20.
This material is available free of charge via the Internet at
http://pubs.acs.org.
The authors declare no competing financial interest.
Org. Lett., Vol. 14, No. 18, 2012
4893