Enantioselective synthesis of the ent-lomaiviticin A bicyclic core

Article abstract of DOI:10.1021/ol302567f

The bicyclic core of ent-lomaiviticin A was prepared in 11 operations from (S)-1-phenyl-2-propyn-1-ol in a two-directional route that features (1) a double Ireland Claisen rearrangement and (2) a double olefin metathesis reaction to form the key C-C bonds of the target.

Full text of DOI:10.1021/ol302567f

ORGANIC
LETTERS
2012
Vol. 14, No. 21
5484–5487
Enantioselective Synthesis of the
ent-Lomaiviticin A Bicyclic Core
Ken S. Feldman* and Brandon R. Selfridge
Chemistry Department, The Pennsylvania State University, University Park,
Pennsylvania 16802, United States
ksf@chem.psu.edu
Received September 18, 2012
ABSTRACT
The bicyclic core of ent-lomaiviticin A was prepared in 11 operations from (S)-1-phenyl-2-propyn-1-ol in a two-directional route that features (1) a
double Ireland Claisen rearrangement and (2) a double olefin metathesis reaction to form the key CꢀC bonds of the target.
Lomaiviticins A and B are glycosylated dimeric marine
actinomycetes isolates that extend the growing family
of diazoparaquinones originally formulated around the
structurally related but monomeric kinamycins (Figure 1).1,2
The lomaiviticins exhibit remarkably potent cytotoxicity
against several cancer cell lines, and the observation that
they induce dsDNA cleavage under reducing conditions
may underscore their biological mechanism-of-action, and
perhaps that of the other structurally related diazopara-
quinones as well.2,3 The challenging structural intricacies
of the lomaiviticin core and their intriguing biological
activity has fueled several synthesis projects,4 eventually
culminating in a landmark 11-step total synthesis of the
lomaiviticin aglycone along with its C(2)ꢀC(20) diastereo-
mer (∼2:1 mixture, lomaiviticin numbering) via a late-
stage dimerization reaction.5 Several of the other lomaivi-
ticin synthesis approaches also recognized the expedience,
and likewise the risk, of pursuing a formal dimerization
strategy to this bipartite target,4b,d,e whereas other ap-
proaches that build outward from a central bicyclic core
have been explored as well.4a,c
(1) (a) He, H.; Ding, W.-D.; Bernan, V. S.; Richardson, A. D.;
Ireland, C. M.; Greenstein, M.; Ellstad, G. A.; Carter, G. T. J. Am.
Chem. Soc. 2001, 123, 5362–5363. (b) Woo, C. M.; Beizer, N. E.; Janso,
J. E.; Herzon, S. B. J. Am. Chem. Soc. 2012, 134, 15285–15288.
(2) Herzon, S. B.; Woo, C. M. Nat. Prod. Rep. 2012, 29, 87–118.
(3) (a) Moore, H. W. Science 1977, 197, 527–532. (b) Arya, D. P.;
Jebaratnam, D. J. J. Org. Chem. 1995, 60, 3268–3269. (c) Laufer, R. S.;
Dmitrienko, G. I. J. Am. Chem. Soc. 2002, 124, 1854–1855. (d) Feldman,
K. S.; Eastman, K. J. J. Am. Chem. Soc. 2005, 127, 15344–15345. (e)
Feldman, K. S.; Eastman, K. J. J. Am. Chem. Soc. 2006, 128, 12562–
12573. (f) Hasinoff, B. B.; Wu, X.; Yalowich, J. C.; Goodfellow, V.;
Laufer, R. S.; Adedayo, O.; Dmitrienko, G. I. Anti-Cancer Drugs 2006,
17, 825–837. (g) Zeng, W.; Ballard, T. E.; Tkachenko, A. G.; Burns,
V. A.; Feldheim, D. L.; Melander, C. Biorg. Med. Chem. Lett. 2006, 16,
5148–5151. (h) O’Hara, K. A.; Wu, X.; Patel, D.; Liang, H.; Yalowich,
J. C.; Chen, N.; Goodfellow, V.; Adedayo, O.; Dmitrienko, G. I.;
Hasinoff, B. B. Free Radical Biol. Med. 2007, 43, 1132–1144. (i) Ballard,
T. E.; Melander, C. Tetrahedron Lett. 2008, 49, 3157–3161. (j) Khdour,
O.; Skibo, E. B. Org. Biomol. Chem. 2009, 7, 2140–2154. (k) Heinecke,
C. L.; Melander, C. Tetrahedron Lett. 2010, 51, 1455–1458. (l) O’Hara,
K. A.; Dmitrienko, G. I.; Hasinoff, B. B. Chem.-Biol. Interact. 2010, 184,
396–402. (m) Mulcahy, S. P.; Woo, C. M.; Ding, W.; Ellestad, G. A.;
Herzon, S. B. Chem. Sci. 2012, 3, 1070–1074.
Figure 1. Lomaiviticin A aglycone and the bicyclic core synthe-
sis target.
We speculated that relative stereochemical control in the
sterically crowded C(3)ꢀC(2)ꢀC(20)ꢀC(30) core region of
r
10.1021/ol302567f
Published on Web 10/17/2012
2012 American Chemical Society

the lomaiviticins might be easier to achieve via the “inside-
out” approach than through the “monomer dimerization”
approach since the former chemistry focuses on establish-
ing the C(2)ꢀC(20) bond early in the route. Toward this
end, a synthesis plan for the lomaiviticin core, distinct from
earlier approaches, can be developed (Scheme 1). In this
plan, the bicyclic core 2 can be prepared by oxidation from
the bis cyclohexenone 3, which in turn should be available
from a double ring-closing olefin metathesis reaction
(RCM) on tetraene 4.6 Tetraene 4 should be available
from two-directional chain extension of the bis ester 5, the
double Ireland Claisen rearrangement product of a bis silyl
ketene acetal. Applying the standard chairlike transition
state model7 for this rearrangement with an equatorial
phenyl ring anchor to these (sequential) Claisen rearrange-
mentsleads to the conclusionthata Z,Z diene 6 isrequired.
In this plan, the Claisen rearrangements are responsible for
setting the central C(3)ꢀC(2)ꢀC(20)ꢀC(30) relative and
absolute stereochemistry. This divergent synthesis plan,
like any two-directional approach, has the advantage of
halving the steps of the route while at the same time
fighting the unavoidable disadvantage of the arithmetic
demon, squared.
(Scheme 2).8 The cheaper of the two enantiomers of 7
was employed for convenience, even though the enantio-
mer of the natural core would result. Glaser coupling of
this propargyl alcohol unites the two “halves” of the target
by forging the C(2)/C(20) bond within the product diyne
8.9 Reduction of the bis diyne 8 to a Z,Z-diene failed
with Lindlar catalyst/H2 under a variety of conditions/
additives, as typically only an enyne product was isolated.
The failure of the second alkyne reduction under Lindlar
hydrogenation conditions, whereas disappointing, is not
without precedent.10 Hence, recourse was made to the
more exotic Zn/Cu/Ag-mediated alkyne reduction proto-
col of Boland,10 which in this instance worked splendidly
to deliver only the Z,Z-diene containing product 9 in good
yield. Double acylation of the crystalline diene diol 9 with
the PMB ether of 2-hydroxybutanoic acid11 proceeded
uneventfully to deliver the Ireland Claisen precursor bis
ester 10.
Scheme 2. Synthesis of the Z,Z-diene Ireland Claisen
Rearrangement Precursor
Scheme 1. Retrosynthesis of Bicycle 2
The Ireland Claisen rearrangement of 10 into the bis
ester 12 required much optimization (on 10 and on related
model systems11) in order to achieve the high yield shown
(Scheme 3). The silyl source (TMSCl, TBSCl, TIPSCl,
TMSOTF, TBSOTf, TIPSOTf), base (Li, Na, K salts of
N(TMS)2, LDA), Lewis acid additive (none, SnCl2, TiCl4,
ZnCl2), and solvent (THF, CH3CN, Et2O) defined the
parameter space for this optimization. Whereas the for-
mation of Z-silyl ketene acetals from simple 2-unsubsti-
tuted glycolate ethers via chelation-controlled enolization
is well established,12 the same level of predictability does
not necessarily attend 2-substituted (i.e., 2-ethyl) versions
suchas10.11,13 ThedoubleClaisen rearrangement depicted
in transition state model 11 is illustrated as a convenience
only; these rearrangements presumably occur sequentially.
The preparation of the key bis Ireland Claisen precursor
6 commenced with the (commercially available) chiral
secondary alcohol 7, which is prepared inexpensively in
10-g batches through the chiral auxiliary-mediated ad-
dition of zinc trimethylsilylacetylide to benzaldehyde
(4) (a) Nicolaou, K. C.; Denton, R. M.; Lenzen, A.; Edmonds, D. J.;
Li, A.; Milburn, R. R.; Harrison, S. T. Angew. Chem., Int. Ed. 2006, 45,
2076–2081. (b) Krygowski, E. S.; Murphy-Benenato, K.; Shair, M. D.
Angew. Chem., Int. Ed. 2008, 47, 1680–1684. (c) Zhang, W.; Baranczak,
A.; Sulikowski, G. A. Org. Lett. 2008, 10, 1939–1941. (d) Nicolaou,
K. C.; Nold, A. L.; Li, H. Angew. Chem., Int. Ed. 2009, 48, 5860–5863. (e)
Lee, H. G.; Ahn, J. Y.; Lee, A. S.; Shair, M. D. Chem.;Eur. J. 2010, 16,
13058–13062.
(9) Smith, C. D.; Tchabanenko, K.; Adlington, R. M.; Baldwin, J. E.
Tetrahedron Lett. 2006, 47, 2309–3212.
(5) Herzon, S. B.; Lu, L.; Woo, C. M.; Gholap, S. L. J. Am. Chem.
Soc. 2011, 133, 7260–7263.
(10) Boland, W.; Schroer, N.; Sieler, C. Helv. Chim. Acta 1987, 70,
1025–1040.
(6) (a) Bedel, O.; Haudrechy, A.; Langlois, Y. Eur. J. Org. Chem.
2004, 3813–3819. (b) Franc-ais, A.; Bedel, O.; Picoul, W.; Meddour, A.;
Courtieu, J.; Haudrechy, A. Tetrahedron: Asymmetry 2005, 16, 1141–
1155.
(7) Ireland, R. E.; Mueller, R. H.; Willard, A. K. J. Am. Chem. Soc.
1976, 98, 2868–2877.
(8) Qiu, L.; Wang, Q.; Lin, L.; Liu, X.; Jiang, X.; Zhao, Q.; Hu, G.;
(11) Feldman, K. S.; Selfridge, B. R. Tetrahedron Lett. 2012, 53, 825–
828.
(12) (a) Burke, S. D.; Fobare, W. F.; Pacofsky, G. J. J. Org. Chem.
1983, 48, 5221–5228. (b) Gould, T. J.; Balestra, M.; Wittman, M. D.;
Gary, J. A.; Rossano, L. T.; Kallmerten, J. J. Org. Chem. 1987, 52, 3889–
3901.
(13) Picoul, W.; Urchegui, R.; Haudrechy, A.; Langlois, Y. Tetra-
hedron Lett. 1999, 40, 4797–4800.
Wang, R. Chirality 2009, 21, 316–323.
Org. Lett., Vol. 14, No. 21, 2012
5485

The product diester 12 is isolated as a single stereoisomer
without any evidence for a minor diastereomer (1H NMR
detection limit <5%).
Scheme 4. Formation of the Bicyclic Structure by Double Olefin
Metathesis
Scheme 3. Ireland Claisen Rearrangement to Establish the
Pivotal C(2), C(3), C(20), C(30) Stereochemical Tetrad
(ClCH2)2, benzene, toluene). Eventually, reproduc-
ible, moderate yields of the bis cyclohexenone 16 were
obtained by conducting the RCM reaction at 100 °C in
a sealed tube (0.02 M in toluene) after thoroughly
degassing the sample via three freezeꢀthaw cycles.
The conclusion of the ent-lomaiviticin A aglycone core
synthesis involves a complicated solution to a seemingly
simple oxidative transformation (Scheme 5). Conversion
ofthe β,γ-alkenes of 16into the requisite γ-hydroxyenones
of 2 should have been no more challenging than alkene
epoxidation followed by oxirane opening facilitated by
enolization of the ketone.16 Unfortunately, none of that
planned chemistry worked. All efforts at alkene epoxida-
tion within 16 (mCPBA, DMDO, peracetic acid, Mn(ppei)2-
(OAc)6) led to one of two equally unfortunate outcomes:
no chemical reaction or compound destruction. Even a simple
model system (half of 16) was untouched by mCPBA and
could only be epoxidized with DMDO. An initial work-
around, which was designed to exploit hydroxyl-directed
epoxidation methodologies, was set up by liberating the
tertiary hydroxyls with TFA treatment of 16. The caged
compound 18 resulted.
A second approach to γ-hydroxylation focused on
forming a bis dienyl silyl ether 19 from 16 with the hope
that oxidation of this species could be directed to the
γ-position. Attempts to epoxidize a simpler model dienyl silyl
ether (i.e., half of 19) led only to R-hydroxylation, pre-
sumably via an unisolated silyloxyepoxide (i.e., Rubottom
oxidation). Fortunately, the electron-rich dienes of 19 were
competent partners for singlet oxygen-mediated cyclo-
additions,17 and the bis endoperoxide 20 as a stable single
diastereomer was formed in modest overall yield from 16.
The structure and stereochemistry of 20 was determined by
single crystal X-ray analysis (see Supporting Information).14
Attempts to cleave the endoperoxide bond via various
reductants (Me2S, thiourea, tributylphosphine) proved
fruitless, as only compound destruction ensued. However,
The relative stereochemistry of 12 was established via
single crystal X-ray analysis of a downstream intermediate
(vide infra). Thus the relative stereochemical outcome of
this bis Claisen rearrangement is consistent with reaction
through chairlike transition states with Z-silyl ketene
acetal precursors, although boat-like transition states
and E-silyl ketene acetals cannot be excluded. Facile
desilylation of crude diester 12 provided the readily pur-
ified diacid 13.
Continuation of the synthesis plan required double
chain extension of diacid 13 to set up a double RCM
sequence (Scheme 4). The sterically hindered acids of 13
simply cyclized into a seven-membered ring anhydride
upon exposure to HN(OMe)Me and EDC-mediated ami-
dation conditions, presaging what turned out to be a
difficult transformation. Eventually, conversion of the
bis acid 13 into the bis acid chloride and then acylation
with NH(OMe)Me did suffice to form the bis Weinreb
amide 14 in acceptable yield. Exposure of this bis amide to
an allyl Grignard reagent led to monoaddition only. For-
tunately, allyl lithium was servicable for this double chain
extension, leading to a good yield of the bis allyl ketone
product 15. The structure and stereochemistry of 15 was
firmly established by single crystal X-ray analysis (see
Supporting Information).14
All thatremainedfor thisphase ofthesynthesissequence
was a double ring closing metathesis. Once again, extensive
optimization studies were required to overcome problems
with low yields in this transformation: Schrock’s catalyst,
Grubbs I, Grubbs II, and HoyvedaꢀGrubbs catalysts15
were explored at a range of temperatures (rtꢀ110 °C),
concentrations (0.03ꢀ0.002 M), and solvents (CH2Cl2,
(16) Crich, D.; Krishnamurthy, V. Tetrahedron 2006, 62, 6830–6840.
(17) Saito, I.; Nagata, R.; Kotsuki, H.; Matsuura, T. Tetrahedron
Lett. 1982, 23, 1717–1720.
(18) He, J.; Tchabanenko, K.; Adlington, R. M.; Cowley, A. R.;
Baldwin, J. E. Eur. J. Org. Chem. 2006, 17, 4003–4013.
(14) Cambridge Crystallographic Data Centre deposition number
for 15: CCDC 867255; for 20: CCDC 894557. The data can be obtained
free from Cambridge Crystallographic Data Centre via http://www.
ccdc.cam.ac.uk/data_request/cif.
(15) Prunet, J. Eur. J. Chem. 2011, 3634–3647.
5486
Org. Lett., Vol. 14, No. 21, 2012

Scheme 5. Completion of the Bicyclic Core Synthesis
Scheme 6. Feasibility of Access to the Lomaiviticin B Core from
a C(1)/C(4) Dione Precursor
would be strongly favored. This calculational result, in
conjunction with the 17 f 18 conversion, perhaps points
out a potential weakness of any strategy to the core of
lomaiviticin B that might proceed through a free C(4)
ketone.
18
In summary, the bicyclic core of ent-lomaiviticin A has
been prepared in enantiomerically enriched form over 11
chemical operations from a chiral alkynol starting mate-
rial. Steric hindranceabout the congestedC(2), C(3), C(20),
C(30) sector of various intermediate structures to some
extent governed the success (or not) of several transforma-
tions. Eventually, key CꢀC bond forming steps (double
Ireland Claisen rearrangement, double ring closing metathesis)
were optimized to provide good yields of milestone inter-
mediates along the way. The core bicycle 2 possesses useful
functional handles by way of the γ-hydroxyenone units;
these functionalities may serve as linkage points for attach-
ment of the remaining aromatic portions of the lomaiviti-
cin A aglycone in future synthesis studies.
desilylation under mild conditions with H2SiF6 did
afford the bis peroxide 21, which could be reduced easily
to the diol 2 with PPh3.
The bis peroxide 21 served as the launch point for an
alternative thrust toward the lomaiviticin core (Scheme 6).
The plan was to generate a bis enedione 22 by formal bis
dehydration of the two peroxide moieties within 21 and
then deprotect the hydroxyls (22 ꢀ> 23) in anticipation of
a double cyclization event to form the lomaiviticin B core,
24. This chemistry would pit the desired hemiketal-
forming cyclization (23 f 24) against the undesired but
now precedented (cf. 17 f 18) alternative of C(3)ꢀOH f
C(40) ketonecyclization toform25. With23, there isaC(10)
ketone to offer up a competition with C(3)ꢀOH f C(40)
ketone cyclization; this C(10) ketone was lacking in the 17
f 18 conversion. In the event, the enedione 22 was
prepared from 21 by way of an intermediate bis acetate.
All attempts to remove the PMB protecting groups from
22 met with compound destruction, so the cyclization
selectivity of hypothetical lomaiviticin B core precursor
23 remains unknown. To the extent that such a cyclization
might occur under thermodynamic control, density func-
tional calculations19 suggest that the undesired isomer 25
Acknowledgment. Financial support from the National
Science Foundation (CHE 0956458) is gratefully acknowl-
edged. The X-ray facility is supported by NSF CHE
0131112.
Supporting Information Available. Full experimental
procedures and copies of the 1H and 13C NMR spectra for
2, 8ꢀ10, 13ꢀ16, 18, 20, and 22; X-ray data for 15 and 20.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(19) Spartan 10 was used with an initial MMFF force field confor-
mational search and then density functional calculations (B3LYP/6-
31G**) on the lowest-energy conformer.
The authors declare no competing financial interest.
Org. Lett., Vol. 14, No. 21, 2012
5487