Enantiocontrolled synthesis of (1S,2S)-6-desmethyl- (methylaziridino)mitosene [8]

Full text of DOI:10.1021/ja994504c

J. Am. Chem. Soc. 2000, 122, 5401-5402
5401
the aziridine and ensures the absolute configuration shown in the
key oxazole intermediate 3 if subsequent events occur without
disruption of the aziridine C-N bonds. According to precedents
for generation of azomethine ylides from oxazoles,5 3 would be
converted into an oxazolium salt 4 by internal alkylation, followed
by nucleophilic addition of cyanide to afford the 4-oxazoline 5.
Electrocyclic ring opening to the azomethine ylide 6 and internal
2+3 trapping would then produce the tetracyclic product 8 via
the 3-pyrroline 7.6 None of these steps had been demonstrated in
the presence of an aziridine, so it was difficult to anticipate
whether intermediates such as 4 and 6 would be viable. Our prior
work had also not explored the issue of dipolarophile reactivity
with the eventual C10 carbon at the correct alcohol oxidation state.
Ester-activated acetylenic dipolarophiles had been shown to
intercept simpler azomethine ylides,6 but reduction of a C10 ester
in the aziridinomitosene setting would pose a difficult challenge
because of the presence of vinylogous amide and quinone
functionality. It was therefore necessary to test an unactivated
alkyne as the dipolarophile for ylide trapping from 6 to 8.
The monosubstituted oxazole 97 was converted into the borane
complex 10 to allow lithiation at C2 without complications due
to ring opening.8 Deprotonation produced 11 and trapping with
the protected serinal 129 gave the adduct 13 as a 6:1 diastereomer
mixture, 93% from 9. The isomers could not be separated, but
Mitsunobu cyclization afforded a major aziridine 14 that was
relatively easy to purify and was obtained in 71% yield based on
the mixture of precursors 13. The cis stereochemistry of 14
follows from the 6.1 Hz coupling between aziridine protons, and
the absolute configuration is assured if the serinal precursor 12
couples with 11 without racemization. For confirmation, racemic
14 was prepared, and HPLC comparisons on a chiral stationary
phase established 98.1% ee for 14 prepared from methyl (S)-N-
tritylserinate (Scheme 1).
Enantiocontrolled Synthesis of
(1S,2S)-6-Desmethyl-(methylaziridino)mitosene
E. Vedejs,* A. Klapars, B. N. Naidu, D. W. Piotrowski, and
F. C. Tucci
Department of Chemistry
UniVersity of Michigan
Ann Arbor Michigan 48109
ReceiVed December 30, 1999
Aromatization of the pyrroline subunit of mitomycin B (1)
results in the formation of the corresponding pyrrole, the
aziridinomitosene 2a.1 A similar elimination occurs in the
sequence of events that causes DNA cross-linking and is
associated with the antitumor effects of the mitomycin antiobi-
otics. One of the alkylation steps involved in cross-linking occurs
at the allylic aziridine C-N bond, and the ease of this process is
reflected in the high solvolytic reactivity of 2a or 2b.2 As a
consequence, it has been difficult to access the aziridinomitosene
skeleton including the correct C1,C10 functionality. Several
syntheses of mitomycins have been completed,3 and there are a
number of examples where the tetracyclic core of the aziridi-
nomitosenes has been prepared.4 Only one of the prior studies
was able to incorporate the sensitive natural substitution pattern
in rings B, C, D in a structure that also contains the fully
elaborated ring A quinone,4b resulting in the total synthesis of
racemic aziridinomitosene 2d.
Exploratory work had shown that the bulky N-trityl group is
important for diastereoselectivity in the addition of 11 to 12, and
that removal of the trityl group late in the synthesis is problematic.
Therefore, 14 was deprotected to give 15 using the combination
of trifluoroacetic acid and trimethylamine borane, a procedure
where trityl cation is generated and is intercepted reductively to
give triphenylmethane. A related method has been reported for
O-trityl cleavage in nucleosides using silane-reducing agents,10a
but the amine borane variation gave cleaner results.10b The
aziridine 15 (82%) proved easy to N-methylate (BuLi; MeI) and
gave 16 in 91% yield. Subsequent O-allyl cleavage11 to 17 and
conversion to the iodide 18 using modified Mitsunobu conditions12
worked well after considerable optimization,13 and cleared the
way for the incorporation of a tethered alkyne dipolarophile. This
was accomplished by cleavage of the TBS ether to afford the
alcohol 19, Dess-Martin oxidation to the aldehyde, and addition
of a TBS-protected propargyllithium reagent. The resulting
secondary alcohol was protected as the acetate 3 to avoid
complications at the stage of ylide generation.
(5) Vedejs, E.; Grissom, J. W. J. Am. Chem. Soc. 1988, 110, 3238.
(6) (a) Vedejs, E. Dax, S. L. Tetrahedron Lett. 1989, 30, 2627. (b) Vedejs,
E.; Piotrowski, D. W. J. Org. Chem. 1993, 58, 1341. (c) Vedejs, E.; Monahan,
S. D. J. Org. Chem. 1997, 62, 4763.
(7) Vedejs, E.; Luchetta, L. J. Org. Chem. 1999, 64, 1011.
(8) Vedejs, E.; Monahan, S. D. J. Org. Chem. 1996, 61, 5192.
(9) (a) Prepared from methyl N-tritylserinate9b by O-allylation with allyl
bromide/NaH in DMF followed by DIBAL reduction at -78 °C, 94% yield.
(b) Baldwin, J. E.; Spivey, A. C.; Schofield, C. J.; Sweeney, J. B. Tetrahedron
1993, 49, 6309.
We have investigated a route to aziridinomitosenes that
assembles the six-membered carbocycle from nonaromatic precur-
sors. This approach was designed to allow greater potential for
the variation of substituents at C6,C7 than is possible from the
natural products, and specifically to target the 6-desmethyl-
(methylaziridino)mitosene 2c (nomenclature: see ref 1). The latter
has no C6,C7 substituents and therefore has additional electrophilic
sites compared to 2a. The strategy relies on early introduction of
(1) Patrick, J. B.; Williams, R. P.; Meyer, W. E.; Fulmor, W.; Cosulich,
D. B.; Broschard, R. W.; Webb, J. S. J. Am. Chem. Soc. 1964, 86, 1889.
(2) Han, I.; Kohn, H. J. Org. Chem. 1991, 56, 4648.
(3) Review: Danishefsky, S. J.; Schkeryantz, J. M. Synlett 1995, 475.
(4) (a) Lee, S.; Lee, W. M.; Sulikowski, G. A. J. Org. Chem. 1999, 64, 4,
4224. (b) Dong, W.; Jimenez, L. S. J. Org. Chem. 1999, 64, 2520. (c) Edstrom,
E. D.; Yu, T. Tetrahedron 1997, 53, 4549. (d) Wang, Z.; Jimenez, L. S. J.
Org. Chem. 1996, 61, 816. (e) Shaw, K. J.; Luly, J. R.; Rapoport, H. J. Org.
Chem. 1985, 50, 4515.
(10) (a) Ravikumar, V. T.; Krotz, A. H.; Cole, D. L. Tetrahedron Lett.
1995, 36, 6587. Pearson, D. A.; Blanchette, M.; Baker, M. L.; Guindon, C.
A. Tetrahedron Lett. 1989, 30, 2739. (b) The use of the Et3SiH (TES-H)
reagent in place of amine borane risks scrambling of TES with the TBS-
protecting group.
(11) Ito, H.; Taguchi, T.; Hanzawa, Y. J. Org. Chem. 1993, 58, 774.
(12) Kunz, H.; Schmidt, P. Liebigs Ann. Chem. 1982, 1245.
(13) Studies performed by D. Warner.
10.1021/ja994504c CCC: $19.00 © 2000 American Chemical Society
Published on Web 05/17/2000

5402 J. Am. Chem. Soc., Vol. 122, No. 22, 2000
Communications to the Editor
Scheme 1a
seconds, and the desired 8 was obtained in 40% yield after
chromatographic purification. No evidence for the 3-pyrroline
intermediate 7 was found.
Compared to the aziridine-containing precursors 14-18, the
tetracyclic 8 and all subsequent intermediates were sensitive to
nucleophiles and to acidic reagents, presumably due to the
presence of C-O and C-N bonds activated by pyrrole nitrogen.
The cycloaddition product 8 was sufficiently stable for chro-
matographic purification and subsequent saponification and
tetrapropylammonium perruthenate (TPAP) oxidation to give the
diketone 20. Chromatography was more difficult and significant
decomposition occurred on a time scale of ∼1 h on unbuffered
silica gel, and considerably faster in the presence of Et3N buffer.
A possible explanation is that initial enolization would result in
a monoenol that should be highly activated for a second
enolization to afford the aromatized hydroquinone. The latter is
a leucoaziridinomitosene14 and is expected to undergo especially
facile aziridine solvolysis. Fortunately, it was possible to intercept
the hydroquinone oxidatively by treatment with DABCO/oxygen/
Co(II)-salen. The resulting quinone 21 could be deprotected to
22 with the HF-pyridine reagent, provided that triethylamine was
also present to control acidity.
All of the tetracyclic intermediates were sensitive and difficult
to handle, so that it was not a total surprise when conversion of
22 to 2c using the standard Cl3CC(O)NdCdO method failed.15
NMR analysis suggested that the aziridine ring had not survived
exposure to the strongly electrophilic reagent, and therefore the
reaction was repeated using a somewhat deactivated carbamoyl-
ating agent FMOCNdCdO/triethylamine. This gave a protected
carbamate 23 at -40 °C to room temperature, and deprotection
with triethylamine at room temperature afforded 2c. A comparison
with the reported UV and NMR characteristics for 2a leaves no
doubt regarding the structure, and the absolute configuration is
defined by the starting methyl serinate.
These findings confirm that the azomethine ylide strategy is
viable, and establish a route to C6,C7-unsubstituted aziridinomi-
tosenes. The isolation of the diketone 20 also suggests that it may
be possible to access natural leucoaziridinomitosenes from
relatively stable precursors if the corresponding diketone tau-
tomers can be prepared. Studies designed to address these issues
and to probe the in vitro activity of 2c are under way.
a (a) i. n-BuLi, -78 °C. ii. MeI. (b) Cp2ZrCl2, n-BuLi, THF, -78 °C
to rt. (c) PPh3, diisopropyl azodicarboxylate, MeI, PhMe, 70 °C.
Extensive exploratory work had established that there are two
critical stages in the one-pot sequence from 3 to 8. The first is
the conversion to an oxazolium salt 4 without destruction of the
aziridine. This step was carried out from the iodide using AgOTf
activation, 3 h in acetonitrile at 70 °C. Attempts to perform the
internal alkylation thermally, without silver ion assistance, resulted
in aziridine degradation in related structures. The other key was
to use a soluble cyanide source, BnMe3N(+)CN(-) for ylide
generation, as reported in a model study.6c When a solution of 4
in acetonitrile was added to the BnMe3N(+)CN(-) reagent at
room temperature, each drop produced a transient yellow color,
tentatively attributed to the ylide 6. The color faded within
Acknowledgment. This work was supported by the National Institutes
of Health (CA17918), and by a fellowship to F.C.T. from CNPq, Brazil.
Supporting Information Available: Experimental procedures and
characterization data (PDF). This material is available free of charge via
the Internet at http://pubs.acs.org.
JA994504C
(14) Egbertson, M.; Danishefsky, S. J. J. Am. Chem. Soc. 1986, 108, 4648.
Egbertson, M.; Danishefsky, S. J. J. Am. Chem. Soc. 1987, 109, 2204. Cera,
C.; Egbertson, M.; Teng, S. P.; Crothers, D. M.; Danishefsky, S. J.
Biochemistry 1989, 28, 5665.
(15) Minami, N.; Ko, S. S.; Kishi, Y. J. Am. Chem. Soc. 1982, 104, 1109.