A biosynthetic proposal for ring formation in the antitumor agent halichomycin. Asymmetric synthesis of the AB-carbon backbone of halichomycin.

Article abstract of DOI:10.1021/ol017266a

[reaction: see text] A biosynthetic proposal for ring formation in the antitumor agent halichomycin is presented in which macrocyclization of the putative prehalichomycin intermediate 1 is the first step. Compound 2 then undergoes dehydration to the alpha

Full text of DOI:10.1021/ol017266a

ORGANIC
LETTERS
2002
Vol. 4, No. 6
897-900
A Biosynthetic Proposal for Ring
Formation in the Antitumor Agent
Halichomycin. Asymmetric Synthesis of
the AB-Carbon Backbone of
Halichomycin
,†
Karl J. Hale,* Paschalis Dimopoulos,† Maxine L. F. Cheung,† Mark Frigerio,†
Jonathan W. Steed,‡ and Philip C. Levett§
The Christopher Ingold Laboratories, UniVersity College London, 20 Gordon Street,
London WC1H 0AJ, England, and The Department of Chemistry, King’s College
London, Strand, London WC2R 2LS, England, and Chemical R & D,
Pfizer Global R & D, Sandwich, Kent CT13 9NJ, England
k.j.hale@ucl.ac.uk
Received December 19, 2001
ABSTRACT
A biosynthetic proposal for ring formation in the antitumor agent halichomycin is presented in which macrocyclization of the putative
prehalichomycin intermediate 1 is the first step. Compound 2 then undergoes dehydration to the r-keto N-acylimine 3 followed by tandem
nucleophilic addition of the C(16)-hydroxyl to form the hemimacrolactam. A stereospecific Michael ring closure and enol protonation complete
C-ring assembly. So far, synthetic efforts toward 1 have resulted in 8.
Halichomycin is a structurally unprecedented tricyclic hemi-
macrolactam produced by a strain of Streptomyces hygro-
scopicus, obtained from the gastrointestinal tract of Hali-
choeres bleekeri, a well-known marine fish.1 Halichomycin
displays powerful antitumor effects in vitro, exhibiting an
ED50 of 0.13 µg/mL against a murine P388 lymphocytic
leukemia cell line. As such, it is of potential interest for the
future treatment of human cancer.
of its skeleton looks propionate- and acetate-derived, it is
not at all clear how the bonds adjacent to the C(8)-C(9)-
bond are formed by such a pathway. It is also not readily
apparent how nature closes the three intersecting ring systems
that are present, which include a fully functionalized 11-
membered ether ring and a structurally unique 13/11-
membered bicyclic hemimacrolactam. These combined con-
ceptual difficulties recently led Kobayashi and Ishibashi to
comment that the biosynthetic provenance of halichomycin
“appeared to be strange”.2 Our biosynthetic proposal for ring
assembly in halichomycin invokes the branched precursor 1
The extraordinary molecular complexity of halichomycin
naturally raises questions about its biosynthesis. While much
† The Christopher Ingold Laboratories.
‡ King’s College London.
§ Pfizer Global R & D.
(2) Kobayashi, J.; Ishibashi, M. ComprehensiVe Natural Product Chem-
istry; Barton D., Nakanishi, K., Meth-Cohn, O., Eds.; Pergamon: Oxford,
1999; Vol. 8, Chapter 8.07, p 416.
(1) Takahashi, C.; Takada, T.; Yamada, T. Minoura, K.; Uchida, K.;
Matsumura, E.; Numata, A. Tetrahedron Lett. 1994, 35, 5013.
10.1021/ol017266a CCC: $22.00 © 2002 American Chemical Society
Published on Web 02/22/2002

as an intermediate and postulates internal macrocyclic
N-acylcarbinolamine formation3 as the first step in ring
formation (Scheme 1). Intermediate 2 is then thought to
metric total synthesis of the putative halichomycin precursor
1, and now report a synthetic strategy for the key AB-carbon
backbone intermediate 8 needed for this venture.
We envisaged constructing the C(18)-C(24) dienone
sector of 1 through a Kishi-Nozaki-Hiyama-Takai coup-
ling5 between 6 and 7, followed by oxidation (Scheme 2).
Scheme 1. A Biosynthetic Proposal for Ring Formation in the
Antitumor Agent Halichomycin
Scheme 2. Retrosynthetic Planning for Halichomycin
undergo dehydration to the R-keto-N-acylimine 3. Molecular
models of 3 show that it can adopt a conformation appropri-
ate for stereospecific internal attack of the C(16)-hydroxyl
upon the N-acylimine carbon,4 which would lead to the
hemimacrolactam 4 after C(14)-ketone reduction and O-
methylation with S-adenosylmethionine. Models further show
that the C(7)-hydroxyl of 4 can readily approach the C(22)-
position of the dienone by a trajectory suitable for acid-
catalyzed Michael ring closure, a process that could later be
followed by stereospecific enol protonation at C(21). To
experimentally test this simple biogenetic hypothesis, by both
chemical and biological means, we embarked on an asym-
(3) It is well-known that primary amides add reversibly to the carbonyl
group of aldehydes and ketones in acidic, neutral, or basic media. In fact,
for many aldehydes, the resulting N-acylcarbinolamines are quite stable
and isolable. According to Weinreb, the equilibria of such additions usually
lies toward the N-acylcarbinolamine to the extent of 5 kcal/mol. See:
Auerbach, J.; Zamore, M.; Weinreb, S. M. J. Org. Chem. 1976, 41, 725.
For reviews on this topic, see: (a) Challis, B. C.; Challis, J. A. Compren-
ensiVe Organic Chemistry; Barton, D., Ollis, W. D., Eds.; Pergamon:
Oxford, 1979; Vol. 2, Chapter 9.9, p 957. (b) Zaugg, H. E.; Martin, W. B.
Org. React. 1965, 14, 52. For other pertinent papers on this topic, see: (a)
Vail, S. L.; Moran, C. M.; Barker R. H. J. Org. Chem. 1965, 30, 1195. (b)
Feuer, H.; Lynch, U. E. J. Am. Chem. Soc. 1953, 75, 5027. (c) Overkleeft,
H. S.; van Wittenberg, J.; Pandit, U. K. Tetrahedron 1994, 50, 4215.
(4) The addition of alcohols to N-acylimines derived from N-acylcarbino-
lamines is well precedented in the chemical literature. See: Zaugg, H. E.;
Martin, W. B. Org. React. 1965, 14, 52.
Aldehyde 6 would be derived from 8 by ammonolysis,
O-debenzylation, and primary alcohol oxidation. A double
(5) Review: Cintas, P. Synthesis 1992, 248.
Org. Lett., Vol. 4, No. 6, 2002
898

Wittig sequence involving aldehyde 10 and ylide 11 was
envisioned for stereospecific elaboration of the dienoate array
in 8, while a Stille reaction6 with â-stannylenone 9 was
planned for fashioning the dienone perimeter. The syn-
relationship between the C(6) and C(7) stereocenters of 10
could potentially be controlled through an Evans asymmetric
aldol reaction7 between 12 and 13, while a face-selective
alkylation reaction between 15 and 16 could set the C(8)
stereocenter. We imagined deriving lactone 16 from the
known homoallylic alcohol 17, which would be available
from the Roush asymmetric crotylboration of 19 with (R,R)-
18.8
Scheme 3. Synthesis of Vinyl Iodide 26a
The synthesis of vinyl iodide 26 is shown in Scheme 3.
O-Benzylation of the anti-alcohol 17 with O-benzyl trichlo-
roacetimidate and catalytic TfOH9 procured the doubly
protected alkene 20, which was converted to the primary
alcohol by rhodium-catalyzed hydroboration10 and oxidation.
Further oxidation to acid 21 was accomplished with PDC in
DMF. Hydrogenation of acid 21 over a 20% Pd(OH)2 on
carbon catalyst effected a clean, but rather slow, deprotection
of the O-benzyl ether to permit in situ butyrolactonization.
The stereospecific C-alkylation of butyrolactone 16 was
achieved by low-temperature enolization with LDA and
addition of the allylic bromide 15.11 The total stereocontrol
observed in this reaction is attributable to the stereodirecting
influence of the C(25)-Me group (which hinders syn-
approach of the bulky electrophile to the enolate) and
preservation of the reaction temperature at -78 °C through-
out. In this regard, premature warming markedly lowered
the selectivity levels that could be attained. The configuration
of the newly induced stereocenter in 14 was verified by NOE
analysis. An OPMB for OTBDPS protecting group inter-
change was now effected to permit C(19)-C(24) side-chain
elaboration later on in the synthesis; this delivered the PMB-
ether 22.
a Reagents and conditions: (a) BnOC(NH)CCl3 (1 equiv), TfOH
(0.05 equiv), CH2Cl2 (0.4 M), rt, 3.5 h; (b) catecholborane (1.1
equiv), (Ph3P)3RhCl (0.05 equiv), THF (0.2 M), 0 °C for 5 min,
then rt for 14 h; 27.5% H2O2/MeOH/2N NaOH, 0 °C, 2 h; (c) PDC
(7 equiv), DMF (0.3 M), rt, 48 h; (d) H2, 20% Pd(OH)2/C, MeOH
(0.4 M), 3-7 d; (e) 16, LiN(Pr-i)2 (1.3 equiv), THF-HMPA (10:
1) (0.2 M), -78 °C, 1 h, then add 15 (1.2 equiv) in THF at -78
°C dropwise and stir at -78 °C for 2 h; (f) 40% aqueous HF/THF/
MeCN (1:2:1) (concentration of 14 ca. 0.09 M), rt, 24-27 h; (g)
PMBOC(NH)CCl3 (2 equiv), PPTS (0.5 equiv), CH2Cl2 (0.1 M),
rt, 7 h; (h) LiBH4 (10 equiv), THF/MeOH (100:1) (0.2 M), ∆, 3 h;
(i) 23, Imid (2.2 equiv), DMF (concentration of 23, ca. 0.1 M), 0
°C, add Et3SiCl (1.2 equiv) over 5 min, then stir at 0 °C for 1.5 h;
(j) 2,6-lutidine (20 equiv), CH2Cl2, -50 °C, add TBSOTf (3 equiv)
over 5 min, then stir for 0.5 h; (k) 2% aqueous HF, THF/MeCN
(1:1), rt, 1.5 h; (l) TPAP (0.05 equiv), NMO (2 equiv), CH2Cl2
(0.01 M), 4A MS, rt, 40 min; (m) 13 (4 equiv), (n-Bu)2BOTf (4
equiv), Et3N (4.2 equiv), CH2Cl2, 0 °C, 0.5 h, then cool to -78
°C, add 12 (1 equiv) in CH2Cl2, stir for 35 min, then warm to rt
for 1 h; (n) add Et3SiOTf (5 equiv) over 5 min to 25 in CH2Cl2
(0.02M), 2,6-lutidine (20 equiv), at -50 °C, then warm to rt for
45 min.
Having fulfilled its role in stereospecific attachment of
the C(8)-methallyl unit, the butyrolactone ring of 22 was
reductively ring-opened with lithium borohydride and diol
23 differentially O-silylated to obtain 24. Selective cleavage
of the primary OTES group now permitted oxidation to the
aldehyde 12 with TPAP/NMO.12 The Evans aldol addition
between 12 and 13 required the use of a significant excess
of the propionimide enolate (4 equiv) to drive the reaction
to completion, which made the purification of 25 exceedingly
difficult. The subsequent O-triethylsilylation reaction rem-
(6) (a) Stille, J. K.; Groh, B. L. J. Am. Chem. Soc. 1987, 109, 813. (b)
Farina, V.; Krishnamurthy, V.; Scott, W. J. Org. React. 1998, 50, 1.
(7) Evans, D. A.; Bartroli, J.; Shih, T. L. J. Am. Chem. Soc. 1981, 103,
2127.
(8) Roush, W. R.; Palkowitz, A. D.; Ando, K. J. Am. Chem. Soc. 1990,
112, 6355.
(9) Wessel, H.-P.; Iverson, T.; Bundle, D. R. J. Chem. Soc., Perkin. Trans.
1 1985, 2247.
edied this situation, allowing the protected aldol adduct 26
to be isolated pure by simple flash chromatography. Sig-
nificantly, no other aldol adducts were observed in the above
addition. The structure of 26 was verified by X-ray crystal-
lography.
(10) (a) Mannig, D.; Noth, H. Angew. Chem., Int. Ed. Engl. 1985, 24,
878. (b) Evans, D. A.; Fu, G. C.; Hoveyda, A. H. J. Am. Chem. Soc. 1988,
110, 6917. (c) Evans, D. A.; Fu, G. J. Org. Chem. 1990, 55, 2280.
(11) Bromide 15 was synthesized in 72% yield from (Z)-3-iodo-2-
methylpropen-1-ol (prepared according to Rayner, C. M.; Astles, P. C.;
Paquette, L. A. J. Am. Chem. Soc. 1992, 114, 3926) after treatment with
Ph3P (1.5 equiv) and NBS (1.3 equiv) in CH2Cl2 (0.2 M) at 0 °C for 1.5 h.
(12) Ley, S. V.; Norman, J.; Griffith, W. P.; Marsden, S. P. Synthesis
1994, 639.
Attention now shifted toward stereospecific elaboration
of the two diene arrays present within 8 (Scheme 4).
Org. Lett., Vol. 4, No. 6, 2002
899

Scheme 4. Synthesis of 8a
Scheme 5. Synthesis of Vinylstannane 9a
a Reagents and conditions: (a) Bu3SnH (1.2 equiv), AIBN (0.05
equiv), PhMe (0.3 M), ∆, 24 h; (b) PPTS (1.5 equiv), MeOH (0.3
M), rt, 24 h; (c) TBSCl (1 equiv) (0.03 M in CH2Cl2) added
dropwise to 34 (1 equiv) and imidazole (2 equiv) in CH2Cl2
(concentration of 34 ca. 0.26 M) at 0 °C; stir 10 min; (d) TPAP
(0.05 equiv), NMO (2.2 equiv), 4A MS, CH2Cl2 (0.26 M), rt, 1 h.
10, which reacted readily with 11 in PhMe at reflux to give
28 with complete stereocontrol.14 DIBAL reduction to 29
and allylic alcohol oxidation with MnO2 generated the
aldehyde 30, which willingly engaged in a second Wittig
reaction with 11. The (E,E)-dienoate 31 was formed as a
single geometrical isomer in 97% yield. The dienone unit
was fashioned by a Stille coupling6 between 31 and 9 (for
the preparation of 9, see Scheme 5). The desired tetraene 8
was isolated as a single geometrical isomer in 33-40% yield,
but was formed alongside a significant quantity of the
stannane homocoupling product. Work is currently underway
to improve the yield of 8 and to reduce the amount of
dimerization that is occurring with stannylenone 9. Future
reports will deal with the synthesis of isotopically labeled 1
from 8, and with our chemical and biological efforts to
convert 1 into halichomycin itself.
Acknowledgment. We thank the EPSRC (Project Grant
GR/M63690), Pfizer (EPSRC INDUSTRIAL CASE QUOTA
studentship to M.L.F.C.), and the BBSRC (M.F.) for financial
support and the ULIRS for HRMS.
a Reagents and conditions: (a) add LiBH4 (10 equiv) in one
portion to 26 in Et2O/H2O (160:1) (0.02 M), at 0 °C, then warm to
rt and stir for 1.5 h; (b) TPAP (0.1 equiv), NMO (2 equiv), CH2Cl2
(0.02 M), 4A MS, rt, 1 h 10 min; (c) 11 (10 equiv), PhMe (0.01
M), ∆, 5 h; (d) i-Bu2AlH (2.2 equiv), PhMe (0.048M), -78 °C,
0.5 h; (e) MnO2 (20 equiv), CHCl3 (0.02 M), ∆, 6 h; (f) 11 (15
equiv), PhMe (0.01 M), ∆, 16 h; (g) 9 (2 equiv), (CH3CN)2PdCl2
(0.5 equiv), i-Pr2NEt (10 equiv), DMF (0.01 M), 5 h.
Supporting Information Available: High-resolution
mass spectra, 500 MHz 1H and 125 MHz 13C NMR spectra
of all new compounds, and X-ray data for 26. This material
is available free of charge via the Internet at http://pubs.acs.org.
Reductive removal13 of the oxazolidinone unit from 26 with
LiBH4 furnished the primary alcohol 27 in excellent yield
(82%). A TPAP oxidation12 converted 27 into the aldehyde
OL017266A
(14) For an alternative synthetic strategy to the C(1)-C(7)-segment of
halichomycin which also utilises ylide 11, see: McGann, E. E.; Janes, G.;
Ortsey, C.; Wood, J. L. Tetrahedron Lett. 1997, 38, 303.
(13) Nicolaou, K. C.; Chakraborty, T. K.; Piscopio, A. D.; Minowa, N.;
Bertinato, P. J. Am. Chem. Soc. 1993, 115, 4419.
(15) Jiang, B.; Ma, P. Synth. Comm. 1995, 25, 3641.
900
Org. Lett., Vol. 4, No. 6, 2002