Synthesis of the polycyclopropane antibiotic FR-900848 via the Horeau gambit

Full text of DOI:10.1021/ja961093g

6096
J. Am. Chem. Soc. 1996, 118, 6096-6097
Scheme 1
Synthesis of the Polycyclopropane Antibiotic
FR-900848 via the Horeau Gambit
J. R. Falck,* Belew Mekonnen, Jurong Yu, and Jing-Yu Lai
Departments of Molecular Genetics and Pharmacology
UniVersity of Texas Southwestern Medical Center
Dallas, Texas 75235
ReceiVed April 3, 1996
StreptoVerticillium ferVens elaborates a unique nucleoside
antibiotic, FR-900848 (1), whose structure was published
without assignment of relative or absolute configurations.1 It
displays potent and highly specific inhibitory activity against
filamentous fungi including several pathogens responsible for
significant human morbidity/mortality but is almost inactive
against Gram-positive and Gram-negative bacteria. In light of
its low toxicity in mammals (murine LD50 > 1g/kg), 1 represents
a promising new lead to counter the alarming increase in the
incidence of systemic fungal infections as well as the concomi-
tant appearance of drug resistant strains.2 The most distinctive
structural feature of 1 is its lipidic proboscis endowed with five
cyclopropane rings. Such unprecedented functionality poses a
daunting synthetic challenge and accordingly has engendered
considerable attention3 that has culminated in a recent total
synthesis.4 Herein, we describe a conceptually distinct approach
to 1 that (a) independently confirms the complete architecture
of FR-900848, (b) validates methodology for the stereocon-
trolled assembly of polycyclopropanes,5 and (c) illustrates a
variant of the Horeau principle6 leading to material of high
enantiomeric enrichment.
A retrosynthetic analysis, outlined in Scheme 1, bisected 1
into fatty acid 2 and dihydrouridine 3. The former was
provisionally assigned an all-trans stereochemistry based on
biogenetic considerations7 and the latter was presumed to have
the configuration typical of nucleosides. Moiety 2 was simpli-
fied further by dismantling into monocyclopropane 4, tetracy-
clopropane 5, and the Horner-Emmons reagent 6.8 Additional
insight into the configuration of 1 came from its ozonolytic
degradation by Fujisawa scientists.9 The 13C NMR spectrum
of the serial tetracyclopropane fragment revealed seven reso-
nances only and was most consistent with a meso or C2-
symmetric product. Combined with extensive NMR compari-
sons with the syn/anti-bicyclopropanes generated from 2,4-
hexadiene-1,6-diol (mucondiol) via nonselective cyclopropanation,
an all-trans, all-syn geometry for 2 was targeted. The absolute
configuration was selected arbitrarily and would be confirmed
en route by comparison with a suitable degradation fragment
from natural material.
A reiterative dimerization strategy10,11 (Scheme 2) was
embraced for the preparation of the tetracyclopropane core of
2 and commenced with a moderately stereospecific (88-90%
ee) Charette-Juteau12 asymmetric cyclopropanation of trans-
allylic alcohol 7.13 Silylation of the derived cyclopropylmetha-
nol14 under standard conditions furnished stannane 8 which was
transmetalated with sec-BuLi. The newly generated lithium
15
anion was added to [ICuPBu3]4 and then subjected to an O2-
induced16 dimerization17 at low temperature to give syn-
trans,trans-bicyclopropane 9 (98% ee18), [R]D23 -41.5° (c 0.23,
absolute EtOH). The enrichment in enantiomeric composition
is a manifestation of the statistical distribution of products and
represents a variant of the Horeau amplification principle.6
Classical resolution techniques are obviated and greater latitude
regarding optical purity of precursors is possible.
(1) Yoshida, M.; Ezaki, M.; Hashimoto, M.; Yamashita, M.; Shigematsu,
N.; Okuhara, M.; Kohsaka, M.; Horikoshi, K. J. Antibiotics 1990, 43, 748-
754.
(2) Sternberg, S. Science 1994, 266, 1632-1634. Georgopapadakou, N.
H.; Walsh, T. J. Ibid. 1994, 264, 371-373.
As a prelude to the next level of dimerization (and its
attendant Horeau amplification), 9 was converted to carboxylic
acid 10 via selective fluoride cleavage of one silyl ether and
RuCl3-catalyzed oxidation of the liberated alcohol. The one-
(3) Theberge, C. R.; Zercher, C. K. Tetrahedron Lett. 1994, 35, 9181-
9184. Armstrong, R. W.; Maurer, K. W. Ibid. 1995, 36, 357-360. Barrett,
A. G. M.; Kasdorf, K.; Williams, D. J. J. Chem. Soc., Chem. Commun.
1994, 1781-1782. Barrett, A. G. M.; Doubleday, W. W.; Tustin, G. J.;
White, A. J. P.; Williams, D. J. Ibid. 1994, 1783-1784. Barrett, A. G. M.;
Tustin, G. J. Ibid. 1995, 355-356. Barrett, A. G. M.; Doubleday, W. W.;
Kasdorf, K.; Tustin, G. J.; White, A.J. P.; Williams, D. J. Ibid. 1995, 407-
408. Barrett, A. G. M.; Kasdorf, K.; White, A. J. P.; Williams, D. J. Ibid.
1995, 649-650. Barrett, A. G. M.; Kasdorf, K.; Tustin, G. J.; Williams, D.
J. Ibid. 1995, 1143-1144.
(4) Barrett, A. G. M.; Kasdorf, K. Chem. Commun. 1996, 325-326.
(5) Examples of polycyclopropane syntheses: Conia, J. M.; Denis, J.
M. Tetrahedron Lett. 1969, 3545-3546. Ripoll, J. L.; Limasset, J. C.; Conia,
J. M. Tetrahedron 1971, 27, 2431-2452. Kostikov, R. R.; Molchanov, A.
P. Zh. Org. Khim. 1978, 14, 1108-1109. de Meijere, A.; Jaekel, F.; Simon,
A.; Borrmann, H.; Kohler, J.; Johnels, D.; Scott, L. T. J. Am. Chem. Soc.
1991, 113, 3935-3941.
(6) For a thorough discussion of the Horeau amplification principle,
see: Rautenstrauch, V. Bull. Soc. Chim. Fr. 1994, 131, 515-524.
(7) We speculate that the odd-numbered C23-acid 2 is derived biogene-
ticially from an even-numbered C18-polyunsaturated fatty acid and that the
cyclopropanes arise from the addition of one-carbon units donated by
S-adenosylmethionine or its equivalent. Since the thermodynamically most
stable form of this hypothetical precursor is all-trans, then the cyclopropanes
and olefins in 2 would also be trans.
(9) Dr. Hirokazu Tanaka (Fujisawa Pharmaceutical Co., Ltd.), personal
communication.
(10) Dimerizations of cyclopropane anions have precedence: Slabey, V.
A. J. Am. Chem. Soc. 1952, 74, 4928-4930. Kai, Y.; Knochel, P.;
Kwiatkowski, S.; Dunitz, J. D.; Oth, J. F. M.; Seebach, D. HelV. Chim.
Acta 1982, 65, 137-161. O'Bannon, P. E.; Dailey, W. P. J. Am. Chem.
Soc. 1989, 111, 9244-9245.
(11) The geometric progression (2n units/dimerization) inherent in this
strategy allows one to rapidly accrue repeating functionality. This would
be especially important for the preparation of higher homologs of 1
containing, for example, eight serial cyclopropanes.
(12) Review: Charette, A. B.; Marcoux, J.-F. Synlett 1995, 1197-1207.
(13) Imai, N.; Sakamoto, K.; Takahashi, H.; Kobayashi, S. Tetrahedron
Lett. 1994, 35, 7045-7048. Jung, M. E.; Light, L. A. Ibid. 1982, 23, 3851-
3854.
(14) All isolated intermediates were fully characterized by 1H/13C NMR
and MS analysis. The elemental composition of an analytical sample was
confirmed by combustion analysis or high-resolution mass spectroscopy.
(15) Whitesides, G. M.; Casey, C. P.; Krieger, J. K. J. Am. Chem. Soc.
1971, 93, 1379-1389.
(16) Lipshutz, B. H.; Kayser, F.; Maullin, N. Tetrahedron Lett. 1994,
35, 815-818.
(17) cf.: Walborsky, H. M.; Banks, R. B.; Banks, M. L. A.; Duraisamy,
M. Organometallics 1982, 1, 667-674.
(18) Chiral phase HPLC analysis was performed on a Chiralcel OD
column (Daicel, 4.6 × 250 mm) using 0.8% i-PrOH/hexane at a flow rate
of 1 mL/min. The bis-(S)-Mosher ester of 9 had a Rt ≈ 16.7 min and the
bis-(S)-Mosher ester of its enantiomer had a Rt ≈ 29 min.
(8) Available from Aldrich Chem. Co.
S0002-7863(96)01093-1 CCC: $12.00 © 1996 American Chemical Society

Communications to the Editor
J. Am. Chem. Soc., Vol. 118, No. 25, 1996 6097
Scheme 2
Partial deprotection of 12 evolved 14 which was prepared
for union with the remaining cyclopropyl unit by catalytic
tetrapropylammonium perruthenate (TPAP) oxidation. The
resultant aldehyde smoothly accepted the anion of sulfone 21,
made from alcohol 2012,13 via Mitsunobu condensation22 with
thiophenol and peracid oxidation (eq 1), to yield 19 (R ) H) as
a mixture of diastereomers. However, Julia elimination23 (R
) Ac, Ms) under a variety of conditions resulted in extensive
structural collapse and furnished almost none of the desired
trans-olefin. Alternatively, Peterson-type olefination exploiting
22 secured vinyl sulfone 15 and a variable amount (10-20%)
of the cis-isomer that was removed chromatographically. The
sulfone was stripped away using lithium naphthalenide at -78
°C and aldehyde 16 was isolated after desilylation and oxidation
as described above. Horner-Emmons homologation of 16
utilizing the ylide of 6 furnished the all-trans adduct as the sole
product which was saponified and condensed with 4-nitrophenol
using DCC to afford active ester 17. Acylation of 5′-amino-
5′-deoxy-5,6-dihydrouridine (3)24 with 17 in DMF at room
temperature concluded the synthesis of 1. Synthetic and natural
FR-900848 were identical in all respects (1H/13C NMR, HPLC,
FAB-MS).25
Acknowledgment. This work is dedicated to Prof. Frank R.
Stermitz: an accomplished scientist, a skillful teacher, and a compas-
sionate mentor. The authors express their gratitude to Professors
Marcus Tius (University of Hawaii), Spencer Knapp (Rutgers Univer-
sity), Kurt Mislow (Princeton University), and Andre´ Charette (Uni-
versite´ de Montre´al) for generously sharing their expertise. Dr.
Hirokazu Tanaka (Fujisawa Pharmaceutical Co.) is thanked for a
voucher sample of natural FR-900848 as well as spectra of its
degradation products and Prof. A.G.M. Barrett is thanked for kindly
providing unpublished data. Financial support was provided by the
Robert A. Welch Foundation (I-782) and USPHS NIH (GM31278).
pot preparation and photolytic decarboxylation of the corre-
sponding Barton thiohydroxamic ester19 in BrCCl3 at 0 °C gave
rise to a 14:1 mixture of bromide 11 and its cis-isomer,
respectively, that was readily separated by chromatography.
Repetition of the dimerization sequence, using tert-BuLi for
anion generation, stereospecifically20 transformed 11 into the
isotactic tetracyclopropane 12 (>99.9% ee21 ) in good yield.
bis-Desilylation led to diol 13, [R]D23 -151.8° (c 1.37, absolute
EtOH), whose spectral (1H/13C/CIMS) characteristics were
indistinguishable from those of a sample obtained by degradation
of natural material (O3/NaBH4). Following derivatization to the
bis-(S)-Mosher ester, 13 and the diol from degraded natural
material had identical retention times upon chiral phase HPLC
analysis21 and could be clearly resolved from enantiomer 18
Supporting Information Available: Details of the syntheses and
characterization data for new compounds (9 pages). See any current
masthead page for ordering and Internet access instructions.
JA961093G
(22) Hughes, D. L. Org. React. 1992, 42, 335-656.
(23) Kocienski, P. Phosphorus Sulfur 1985, 24, 97-127.
(24) Prepared from commercial uridine by a conventional sequence as
summarized below.
(prepared as described in Scheme 2 using the R,R-Charette/
Juteau catalyst).
23
(19) Barton, D. H. R.; Crich, D. C.; Motherwell, W. B. Tetrahedron
Lett. 1983, 24, 4979-4982.
(25) The exception was the optical rotation. Synthetic 1 showed [R]D20
-158° (c 0.3, DMSO), whereas the originally reported1 value was [R]D
-36.22° (c 0.5, DMSO). Unfortunately, a sample of natural material suitable
for verification of this measurement could not be obtained. However, the
[R]D -168.1° (c 0.42, DMSO-d6) of synthetic and natural FR-900848
observed by Dr. Krista Kasdorf, in the laboratories of Professor A. G. M.
Barrett, is of a magnitude comparable to ours (Prof. A. G. M. Barrett,
personal communication).
(20) A better indication of the stereospecificity is seen in the dimerization
of the cis-analog of 11. The all-syn-trans,cis,cis,trans-tetracyclopropane
dimer was isolated in 80% yield, and no other stereoisomers were observed.
(21) Chiral phase HPLC analysis was performed as described in ref 18.
The bis-(S)-Mosher ester of 13 and diol from degraded natural material
had Rt ≈ 14.4 min; the bis-(S)-Mosher ester of 18 had a Rt ≈ 17.6 min.