Application of ring-closing metathesis to the formal total synthesis of (+)-FR900482

Article abstract of DOI:10.1021/ja0013879

A formal, enantioselective synthesis of the antitumor antibiotic (+)-FR900482 (1) has been completed using an approach that featured the ring-closing metathesis of the diene 37 to give the key intermediate benzazocine 38. Although several initial protecting-group strategies unexpectedly failed at various stages of the endeavor, the successful approach to 1 involved the conversion of commercially available 5-nitrovanillin (10) into the prochiral diol 24. The manipulations of the residues on the aromatic ring of 10 were straightforward, and the diol array in 24 was introduced by the hydride reduction of the malonate 23, which was in turn prepared by a nucleophilic substitution of the triflate 12. Adjustment of alcohol-protecting groups to give 27 and refunctionalization of the aromatic nitro group led to the protected N-allylamine 36. Elaboration of the diol array via a highly stereoselective Grignard addition furnished the diene 37. Ring-closing metathesis of 37 using the Grubbs catalyst 34 cleanly afforded the benzazocine 38. A tactic originally conceived for preparing 42 by introduction of the aziridine ring onto 38 was impractical because the iodo cyclization of the allylic tosylcarbamate 39 was neither efficient nor selective to give 40. Hence, 38 was transformed into 49, which was a key intermediate in Fukuyama's elegant synthesis of racemic FR900482, thereby completing a formal synthesis of the alkaloid. The prochiral diol 24 was enzymatically desymmetrized using Pseudomonas species lipase to give 25 in 94% enantiomeric excess. Inasmuch as subsequent adjustment of the alcohol-protecting groups gave the intermediate 26 (cf the racemic analogue 27) in enantiomerically pure form, an enantioselective synthesis of (+)-FR900482 has also been completed in a formal sense.

Full text of DOI:10.1021/ja0013879

J. Am. Chem. Soc. 2000, 122, 10781-10787
10781
Application of Ring-Closing Metathesis to the Formal Total Synthesis
of (+)-FR900482
Ingrid M. Fellows, David E. Kaelin, Jr., and Stephen F. Martin*
Contribution from the Department of Chemistry and Biochemistry, The UniVersity of Texas at Austin,
Austin, Texas 78712
ReceiVed April 20, 2000
Abstract: A formal, enantioselective synthesis of the antitumor antibiotic (+)-FR900482 (1) has been completed
using an approach that featured the ring-closing metathesis of the diene 37 to give the key intermediate
benzazocine 38. Although several initial protecting-group strategies unexpectedly failed at various stages of
the endeavor, the successful approach to 1 involved the conversion of commercially available 5-nitrovanillin
(10) into the prochiral diol 24. The manipulations of the residues on the aromatic ring of 10 were straightforward,
and the diol array in 24 was introduced by the hydride reduction of the malonate 23, which was in turn prepared
by a nucleophilic substitution of the triflate 12. Adjustment of alcohol-protecting groups to give 27 and
refunctionalization of the aromatic nitro group led to the protected N-allylamine 36. Elaboration of the diol
array via a highly stereoselective Grignard addition furnished the diene 37. Ring-closing metathesis of 37
using the Grubbs catalyst 34 cleanly afforded the benzazocine 38. A tactic originally conceived for preparing
42 by introduction of the aziridine ring onto 38 was impractical because the iodo cyclization of the allylic
tosylcarbamate 39 was neither efficient nor selective to give 40. Hence, 38 was transformed into 49, which
was a key intermediate in Fukuyama’s elegant synthesis of racemic FR900482, thereby completing a formal
synthesis of the alkaloid. The prochiral diol 24 was enzymatically desymmetrized using Pseudomonas species
lipase to give 25 in 94% enantiomeric excess. Inasmuch as subsequent adjustment of the alcohol-protecting
groups gave the intermediate 26 (cf the racemic analogue 27) in enantiomerically pure form, an enantioselective
synthesis of (+)-FR900482 has also been completed in a formal sense.
Introduction
bond followed by sequential cyclization, dehydration, and
tautomerization to give a mitosene species of the general type
5. Such mitosenes are known to exhibit cytotoxic activity
because of their ability to cross-link duplex DNA in the minor
groove both in vitro and in vivo.5
FR900482 (1) is a novel antitumor antibiotic that was isolated
from the fermentation broth of Streptomyces sandaensis No.
6897 and characterized by workers at Fujisawa Pharmaceutical
Co. in 1987.1 Extensive studies have shown that FR900482 and
related compounds such as FK973 (2) and FK317 (3) display
highly promising potency against a number of tumor cell lines.2,3
Indeed, 3 is presently undergoing advanced clinical trials in
Japan and may replace the widely prescribed anticancer drug
mitomycin C (4). The basis for the biological activity of
compounds 1-3 has been the subject of considerable interest,4
and it is now generally accepted that these compounds are
activated by two-electron reduction of the nitrogen-oxygen
(1) (a) Imanaka, H.; Aoki, H.; Kohsaka, M.; Terano, H.; Kiyoto, S.;
Iwami, M J. Antibiot. 1987, 40, 589-593. (b) Terano, H.; Takase, S.;
Hosoda, J.; Kohsaka, M J. Antibiot. 1989, 42, 145-148. (c) Imanaka, H.;
Aoki, H.; Kohsaka, M.; Terano, H.; Okuhara, M.; Komori, T.; Yamashita,
M.; Shibata, T.; Kiyoto, S. J. Antibiot. 1987, 40, 594-599. (d) Hashimoto,
M.; Kiyoto, S.; Kayakiri, H.; Takase, S.; Uchida, I. J. Am. Chem. Soc. 1987,
109, 4108-4109. (e) Kohsaka, M.; Fujita, T.; Takase, S.; Otsuka, T.; Terano,
H. J. Antibiot. 1988, 41, 392-394.
(2) (a) Kikuchi, H.; Shimomura, K.; Hirai, O.; Mizota, T.; Matsumoto,
S.; Mori, J.; Shibayama, F. J. Antibiot. 1987, 40, 600-606. (b) Kikuchi,
H.; Shimomura, K.; Hirai, O.; Mizota, T.; Matsumoto, S.; Mori, J.;
Shibayama, F. J. Antibiot. 1987, 40, 607-611. (c) Shibayama, F.;
Shimomura, K.; Manda, T.; Mukumoto, S.; Masuda, K.; Nakamur, T.;
Mizota, T.; Matsumoto, S.; Nishigaki, F.; Oku, T.; Mori, J. Cancer Res.
1988, 48, 1166-1172.
(3) (a) Naoe, Y.; Inami, M.; Matsumoto, S.; Takagaki, S.; Fujiwara, T.;
Yamazaki, S.; Kawamura, I.; Nishigaki, F.; Tsujimoto, S.; Manda, T.;
Shimomura, K. Jpn. J. Cancer Res. 1998, 89, 1306-1317. (b) Naoe, Y.;
Kawamura, I.; Inami, M.; Matsumoto, S.; Nishigaki, F.; Tsujimoto, S.;
Manda, T.; Shimomura, K. Jpn. J. Cancer Res. 1998, 89, 1318-1325.
Like the mitomycins, FR900482 (1) contains an aziridine and
a carbamoyloxymethyl group, but it also possesses a rare
hydroxylamine hemiketal, which is the seat of the biological
triggering mechanism. The unusual and compact structure of 1
and its high degree of functionality, coupled with the stereo-
chemical relationship between the center at C(7) and the
aziridine moiety at C(9) and C(10), render 1 an intriguing and
challenging target for synthesis. Indeed, a number of approaches
10.1021/ja0013879 CCC: $19.00 © 2000 American Chemical Society
Published on Web 10/24/2000

10782 J. Am. Chem. Soc., Vol. 122, No. 44, 2000
Fellows et al.
Scheme 1
rings could be formed by RCM reactions.11 We envisioned that
the diene 8 would be derived from the diol 9, which in turn
would be prepared from commercially available 5-nitrovanillin
(10). Because it has been shown that numerous prochiral diols
may be efficiently enzymatically desymmetrized,12 the inter-
mediacy of 9 afforded the significant opportunity of preparing
1 in optically pure form.
Results and Discussion
First Generation Approach. Having set forth the basic route
to FR900482 in Scheme 1, we still needed to identify the precise
tactics that would eventuate in the successful completion of the
synthesis. In addition to exploring new chemistry, one of our
goals was to develop an efficient approach to 1 that could be
applied to the practical synthesis of analogues for biological
testing. Toward this end, it would be essential to minimize
unproductive refunctionalization and protecting maneuvers, and
after analyzing various possibilities, we concluded that the
prochiral diol 15 could be most quickly transformed into the
natural product according to our general plan.
Hence, commercially available 5-nitrovanillin (10) was de-
methylated using HBr/HOAc to provide an intermediate diphe-
nol, the dianion of which was selectively monobenzylated on
the more basic oxygen to give 11. The phenol 11 was converted
to the corresponding triflate 12, which underwent facile nu-
cleophilic aromatic substitution with NaCH(CO2Me)2 in DMF
to provide 13, together with small quantities (<5%) of aldehyde
addition product.13 Interestingly, we discovered that the choice
of solvent for this reaction proved crucial, for when THF was
used, nucleophilic addition to the aldehyde was competitive with
substitution of the triflate. The aldehyde moiety of 13 was then
protected as its dimethyl acetal to furnish 14. The ordering of
steps in converting 12 to 14 was critical as displacement of the
triflate by malonate anion after protection of the aldehyde as
the dimethyl acetal proceeded at a much slower rate and afforded
the substituted product in only about 60% yield.
toward 1 have been described,6 and although three total
syntheses of 1 have been reported,7,8 only one of these produced
enantiomerically pure material.9 Herein we summarize our own
efforts in the area that have culminated in a formal total synthesis
of racemic FR900482, and we also detail studies directed toward
a simple modification of the approach that leads to an enanti-
oselective synthesis of (+)-FR900482.
In developing our own approach to the synthesis of FR900482
(1), we viewed the keto aziridine 6 as a key intermediate that
would be accessible from the benzazocine 7 (Scheme 1). We
were intrigued by the possibility of exploiting a ring-closing
metathesis (RCM) reaction of the highly functionalized diene
8 for constructing the unsaturated eight-membered ring in 7.10
Indeed, we were the first to demonstrate that eight-membered
The reduction of the dimethyl malonate group to the desired
1,3-diol in 15 proved problematic, most likely because of the
acidity of the benzylic proton situated alpha to the two ester
functions. Indeed, the difficulty associated with reducing aryl
malonates is well-documented.14 For example, when 14 was
(4) For some examples, see: (a) Shimomura, K.; Masuda, K.; Nakamura,
T.; Mizota, T.; Mori, J. Cancer Res. 1988, 48, 5172-5177. (b) Fukuyama,
T.; Goto, S. Tetrahedron Lett. 1989, 30, 6491-6494. (c) Williams, R. M.;
Rajski, S. R. Tetrahedron Lett. 1992, 33, 2929-2932. (d) Hopkins, P. B.;
Woo, J.; Sigurdsson, S. T. J. Am. Chem. Soc. 1993, 115, 1199-1200. (e)
Hopkins, R. B.; Williams, R. M.; Rajski, S. R.; Huang, H. Tetrahedron
Lett. 1994, 35, 9669-9672. (f) Hopkins, P. B.; Pratum, T. K.; Huang, H.
J. Am. Chem. Soc. 1994, 116, 2703-2709. (g) Hopkins, P. B.; Paz, M. M.
J. Am. Chem. Soc. 1997, 119, 5999-6005. (h) Williams, R. M.; Rollins, S.
B.; Rajski, S. R. Chem. Biol. 1997, 4, 127-137. (i) Williams, R. M.; Rajski,
S. R. J. Am. Chem. Soc. 1998, 120, 2192-2193. (j) Paz, M. M.; Sigurdsson,
S. T.; Hopkins, P. B. Bioorg. Med. Chem. 2000, 8, 173-179.
(5) For an excellent review see: Williams, R. M.; Rajski, S. R. Chem.
ReV. 1998, 98, 2723-2795.
(6) (a)Williams, R. M.; Yasuda, N. Tetrahedron Lett. 1989, 30, 3397-
3400. (b) Rapoport, H.; Jones, R. J. J. Org. Chem. 1990, 55, 1144-1146.
(c) Danishefsky, S. J.; McClure, K. F. J. Org. Chem. 1991, 56, 850-853.
(d) Danishefsky, S. J.; McClure, K. F. J. Am. Chem. Soc. 1993, 115, 6094-
6100. (e) Martin, S. F.; Wagman, A. S. Tetrahedron Lett. 1995, 36, 1169-
1170. (f) Grubbs, R. H.; Miller, S. J.; Kim, S. H.; Chen, Z. R. J. Am. Chem.
Soc. 1995, 117, 2108-2109. (g) Lim, H. J.; Sulikowski, G. A. Tetrahedron
Lett. 1996, 30, 5243-5246. (h) Ziegler, F. E.; Belema, M. J. Am. Chem.
Soc. 1997, 622, 1083-1094. (i) Mithani, S.; Drew, D. M.; Rydberg, E. H.;
Taylor, N. J.; Mooibroek, S.; Dmitrienko, G. I. J. Am. Chem. Soc. 1997,
119, 1159-1160. (j) Williams, R. M.; Rollins, S. B.; Judd, T. C. Tetrahedron
2000, 56, 521-532. (k) Zhang, W.; Wang, C.; Jimenez, L. S. Synth.
Commun. 2000, 30, 351-366.
(9) (a) Terashima, S.; Katoh, T.; Itoh, E.; Yoshino, T.; Hashimoto, M.;
Nagata, Y. Tetrahedron 1997, 53, 10229-10238. (b) Terashima, S.; Katoh,
T.; Itoh, E.; Yoshino, T.; Hashimoto, M.; Nagata, Y. Tetrahedron 1997,
53, 10239-10252. (c) Terashima, S.; Katoh, T.; Itoh, E.; Yoshino, T.;
Nagata, Y.; Nakatani, S. Tetrahedron 1997, 53, 10253-10270. (d) Katoh,
T.; Terashima, S. J. Synth. Org. Chem., Jpn. 1997, 55, 946-957.
(10) For reviews of RCM, see: (a) Grubbs, R. H.; Miller, S. J.; Fu, G.
C. Acc. Chem. Res. 1995, 28, 446-452. (b) Blechert, S.; Schuster, M.
Angew. Chem. Int. Ed. Engl. 1997, 36, 2036-2056. (c) Armstrong, S. K.
J. Chem. Soc., Perkin Trans. 1 1998, 371-388. (d) Grubbs, R. H.; Chang,
S. Tetrahedron 1998, 54, 4413-4450. (e) Schrock, R. R. Tetrahedron 1999,
55, 8141-8153. (f) Phillips, A. J.; Abell, A. D. Aldrichimica Acta 1999,
32, 75-89.
(11) (a) Martin, S. F.; Liao, Y.; Wong, Y.; Rein, T. Tetrahedron Lett.
1994, 35, 691-694. (b) Martin, S. F.; Liao, Y.; Chen, H.-J.; Pa¨tzel, M.;
Ramser, M. Tetrahedron Lett. 1994, 35, 6005-6008. (c) Martin, S. F.; Chen,
H. J.; Courtney, A. K.; Liao, Y.; Pa¨tzel, M.; Ramser, M. N.; Wagman, A.
S. Tetrahedron 1996, 52, 7251-7164. See also: (d) Martin, S. F.;
Humphrey, J. M.; Ali, A.; Hillier, M. C. J. Am. Chem. Soc. 1999, 121,
866-867.
(12) For reviews on enzymatic transformations see: (a) Azerad, R. Bull.
Soc. Chim. Fr. 1995, 132, 17-51. (b) Drauz, D.; Waldmann, H. Enzyme
Catalysis in Organic Synthesis; VCH and Weinheim: New York, 1995.
(c) Johnson, C. R.; Schoffers, E.; Golebiowski, A. Tetrahedron 1996, 52,
3769-3826.
(13) Atkinson, J. G.; Wasson, B. K.; Fuentes, J. J.; Griard, T.; Rooney,
C. S. Tetrahedron Lett. 1979, 31, 2857-2860.
(14) For example, see: Choi, Y. M.; Emblidge, R. W. J. Org. Chem.
1989, 54, 1198-1200.
(7) Fukuyama, T.; Xu, L.; Goto, S. J. Am. Chem. Soc. 1992, 114, 383-
384.
(8) Danishefsky, S. J.; Schkreyantz, J. M. J. Am. Chem. Soc. 1995, 117,
4722-4723.

Formal Total Synthesis of (+)-FR900482
J. Am. Chem. Soc., Vol. 122, No. 44, 2000 10783
treated with hydride reagents such as LiAlH4, NaBH4, and AlH3,
only starting material was recovered. Of all of the reductants
surveyed, DIBAL-H was the only one that provided the requisite
diol 15 and then in modest yield.
Scheme 2
In view of our goal to provide for an asymmetric synthesis
of FR900482, attention was then focused upon the enzymatic
desymmetrization of diol 15.12 After an extensive survey of the
available enzymes, acyl donors, and reaction conditions, we
found that treatment of 15 with vinyl acetate in the presence of
Pseudomonas species lipase (PSL) from Sigma and 4 Å
molecular sieves gave the chiral monoacylated product 16 in
68% yield and >95% ee.15,16 The purity of the enzyme was
important because the PSL enzyme from Amano did not provide
the desired monoacylated product, whereas the PSL enzyme
from Sigma, which was more homogeneous, as evidenced by
SDS page gel electrophoresis, was effective. Temperature was
also a critical factor, and the optimum reaction temperature was
found to be 35-40 °C. At lower temperatures, the reaction was
sluggish, and at higher temperatures, enantioselectivity was
compromised. The ee of 16 was determined by NMR analysis
of its (R)-Mosher ester,17 and the absolute configuration of the
benzylic center in 16 was determined to be S by X-ray analysis
of this Mosher ester.
Preliminary attempts to selectively protect the primary alcohol
function in 16 as its methoxyphenylmethyl (MPM) ether using
methoxyphenylmethyl trichloroacetimidate under acidic condi-
tions led to significant loss of the dimethyl acetal-protecting
group.18 Similarly, basic conditions led to recovery of starting
material (i.e., K2CO3, MPM-Cl) or the loss of the acetyl group
(i.e., NaH, MPM-Cl). Because of a concern regarding possible
racemization of the benzylic position of 16 and derived
compounds,19 we deemed it essential to avoid exposing such
intermediates to more strongly basic conditions. On the other
hand, introduction of the tert-butyldimethylsilyl protecting group
onto 16 proceeded smoothly, and subsequent methanolysis of
the acetate and oxidation of the resulting primary alcohol with
Dess-Martin periodinane20 afforded the aldehyde 17 in 40%
overall yield (unoptimized). Unfortunately, all attempts to add
a vinyl group to the aldehyde function using magnesium,
lithium,21 cerium,22 and chromium23 reagents were unsuccessful,
perhaps because of the acidity of the benzylic hydrogen alpha
to the carbonyl group in 17. It should also be noted that the
configuration at C(7) in 17 is enantiomeric to that found in
FR900482.
Thus, the diol 15 was converted in one step to the methoxy-
phenylmethyl acetal 18 by treatment with p-CH3OC6H4CH2-
OCH3 and DDQ (Scheme 3).24 Selective reductive cleavage of
the cyclic acetal moiety in 18 gave 19.25 Mindful of the problems
previously encountered with the additions of vinyl organome-
tallic reagents to the aldehyde 17, we reasoned that a protected
amino aldehyde, such as 22, might react with vinyl nucleophiles
in a more straightforward fashion. To test this hypothesis, the
primary alcohol group in 19 was protected, and the nitro group
was reduced by catalytic hydrogenation using Raney nickel in
anhydrous THF/MeOH (1:1). Use of Raney nickel under
anhydrous conditions was critical to the success of this reduction
because the derived hydroxylamine is a significant byproduct
when the reaction is conducted in the presence of water.
Although the nitro group is reduced readily with Raney nickel,
the reaction must be carefully monitored to avoid hydrogenolysis
of the various benzyl groups. Conversion of the amino group
in 20 into its corresponding tert-butyl carbamate using NaHMDS
and Boc2O,26 followed by N-allylation furnished 21. Deprotec-
tion of the silyl ether was accomplished by reaction of 21 with
tetrabutylammonium fluoride (TBAF) that had been pretreated
with activated 4 Å molecular sieves to remove adventitious
water, thereby avoiding hydrolysis of the acetal moiety. Swern
oxidation of the crude alcohol thus obtained provided the
aldehyde 22. Despite our hopes to the contrary, numerous
attempts to add vinylmagnesium, -lithium, -cerium, and -chro-
mium reagents to 22 as described previously were unavailing.
For example, use of excess vinylmagnesium bromide caused
degradation of the dimethyl acetals.27
Because of the cost of the PSL enzyme and the initial
difficulties encountered in advancing 16, we elected to establish
a viable route to FR900482 by first using racemic intermediates.
(15) Wong, C. H.; Wang, Y. F.; Lalonde, J. J.; Momongan, M.;
Bergbreiter, D. E. J. Am. Chem. Soc. 1988, 110, 7200-7205.
(16) (a) Burgess, K.; Jennings, L. J. Am. Chem. Soc. 1991, 113, 6129-
6139. (b) Carrea, G.; Riva, S.; Secundo, F. Tetrahedron: Asymmetry 1992,
3, 267-280. (c) Guanti, G.; Banfi, L.; Riva, R. Tetrahedron: Asymmetry
1994, 5, 9-12.
(17) (a) Mosher, H. S.; Sullivan, G. R.; Dale, J. A. J. Org. Chem. 1969,
34, 2543-2549. (b) Mosher, H. S.; Dale, J. A. J. Am. Chem. Soc. 1973,
95, 512-519.
(18) Yonemitsu, O.; Nakajima, N.; Horita, K.; Abe, R. Tetrahedron Lett.
1988, 29, 4139-4142.
(19) Riva, R.; Gaunti, G.; Banfi, L. Tetrahedron 1996, 42, 13493-13512.
(20) (a) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155-4156.
(b) Meyer, S. D.; Schreiber, S. L. J. Org. Chem. 1994, 59, 7549-7552.
(21) Weiner, M.; Seyferth, D. J. Am. Chem. Soc. 1991, 83, 3583-3586.
(22) Imamato, T.; Takiyama, N.; Nakamura, K.; Hatajima, T.; Kamiya,
T. J. Am. Chem. Soc. 1989, 111, 4392-4398.
(23) (a) Cinta, P. Synthesis 1992, 248-257. (b) Takai, K.; Hiyama, T.;
Nozaki, H.; Kimura, K.; Kuroda, T. Tetrahedron Lett. 1983, 24, 5281-
5284. (c) Kishi, Y.; Haolun, J.; Uenishi, J.; Christ, W. J. J. Am. Chem. Soc.
1986, 108, 5644-5646. (d) Takai, K.; Nozaki, H.; Tagahira, M.; Kuroda,
T.; Oshima, K. J. Am. Chem. Soc. 1986, 108, 6048-6050.
At this stage, it was apparent that the dimethyl acetal
protecting group for the aldehyde at C(3) of FR900482 was
causing problems at various stages because of its acid-sensitivity.
(24) (a) Oikawa, Y.; Yoshioka, T.; Yonemitsu, O. Tetrahedron Lett. 1982,
23, 889-892. (b) Oikawa, Y.; Yoshioka, T.; Yonemitsu, O. Tetrahedron
Lett. 1983, 24, 4037-4040.
(25) (a) Mitsunobu, O.; Asano, H.; Mikami, T. Chem. Lett. 1987, 2033-
2036. (b) Takano, S.; Akiyama, M.; Sato, S.; Ogasawara, K. Chem. Lett.
1983, 1593-1596.
(26) Kelly, T. A.; McNeil, D. W. Tetrahedron Lett. 1994, 35, 9003-
9006.

10784 J. Am. Chem. Soc., Vol. 122, No. 44, 2000
Fellows et al.
Scheme 3
Scheme 4
whereas the (R)-MTPA ester of the racemic acetate 25 exhibited
two well-resolved singlets of equal intensity at δ -71.6 ppm
and -71.8 ppm. The conversion of 25 into 26, which has the
requisite absolute configuration at C(7) for eventual conversion
to FR900482, was achieved by a straightforward series of
hydroxyl protecting group interchanges.
Although it would have been possible to carry enantiomeri-
cally enriched 26 forward, we elected instead to probe the
viability of the key ring-closing metathesis step and develop a
reliable end game for the synthesis of FR900482 in the racemic
series. Thus, the prochiral diol 24 was converted in two steps
into 27 (Scheme 5), which corresponds to the racemic modifica-
Scheme 5
To circumvent this problem, an alternative protecting-group
strategy was developed in devising a new set of tactics for the
synthesis of FR900482.
Second Generation Approach. After considering several
possibilities, the approach that emerged as most attractive
commenced with the reduction of the aldehyde 13 and the
protection of the primary alcohol thus formed as its benzyl ether
to give 23 (Scheme 4). The benzyl group was selected as the
protecting group of choice because we reasoned that both benzyl
groups could be removed simultaneously in the final stages of
the synthesis. Reduction of the malonate ester of 23 with
DIBAL-H as before yielded the prochiral diol 24, again in
modest yield. Desymmetrization of the prochiral diol using PSL
and vinyl acetate as before yielded the (S)-hydroxy acetate 25
in 94% ee as determined by 19F NMR analysis of the
corresponding (R)-Mosher (MTPA) ester. That the absolute
stereochemistry of 25 was the same as 16 was readily established
by converting the (R)-MTPA ester of 16 into the (R)-MTPA
ester of 25 via sequential acetal hydrolysis, hydride reduction,
and O-benzylation [(i) 3 N aqueous HCl, THF, rt, 1 h; (ii)
NaBH4, THF, rt, 1 h; (iii) Cl3CC(dNH)OBn, TfOH, CH2Cl2,
rt, 15 h]. The 19F NMR spectra of the Mosher ester of 25
prepared by both routes exhibited a singlet at δ -71.6 ppm,
tion of 26. Selective reduction of the nitro group using Raney
nickel in anhydrous methanol proceeded without incident to
deliver the amine 28. Following conversion of the amino group
in 28 to its tert-butyl carbamate derivative using Boc2O in
refluxing THF,28 N-allylation of the derived anion provided 29.
With 29 in hand, the stage was nearly set for the key RCM
reaction using a substrate related to 8. In the event, fluoride-
(27) For examples of Grignard reagent addition to acetals see: (a)
Silverman, G. S.; Rakita, P. E. Handbook of Grignard Reagents; Marcel
Dekker: New York, 1991; p 317. (b) Denmark, S. E.; Willson, T. M.;
Amburgey, J. J. Chem. Soc., Perkin Trans. 1 1991, 2899-2906. (c)
Wakefield, B. J. Best Synthetic Methods Series: Organomagnesium Methods
in Organic Synthesis; Academic Press: New York. 1995; pp 160-172.

Formal Total Synthesis of (+)-FR900482
J. Am. Chem. Soc., Vol. 122, No. 44, 2000 10785
Scheme 6
Scheme 7
31 in the presence of 35 (10 mol %) and found that this
cyclization proceeded quickly to give 33 in 88% yield. Because
the synthetic plan required a free alcohol for future transforma-
tions, the cyclization of the allylic alcohol 30 using the
ruthenium catalyst 34 (10 mol %), which unlike 35 is compatible
with free hydroxyl groups, was examined and found to proceed
to furnish 32 in 78% yield. Starting diene 30 (11%) was also
recovered. As noted previously, an X-ray structure of 32
supported the relative stereochemistry between C(7) and C(8).
The next step in the synthetic plan required replacing the Boc
protecting group on 32 with a hydroxylamine moiety. However,
despite numerous attempts under a wide range of conditions,34
we were unable to selectively deprotect the nitrogen without
concomitant removal of the primary MPM ether. Such facile
loss of the MPM group was not anticipated and may owe its
origin to the proximity of the allylic alcohol at C(8). Indeed,
the X-ray structure of 32 reveals a strong hydrogen bond
between the hydrogen on the allylic alcohol and the oxygen of
the MPM-protected alcohol. Once again, it was necessary to
develop a new protecting-group strategy.
Third-Generation Approach. The trichloroethoxycarbonyl
group was quickly identified as a suitable nitrogen-protecting
group, and its introduction onto the amine 28, followed by
allylation, which required KH rather than NaH as the base, to
give 36 was straightforward (Scheme 7). Deprotection of the
primary TIPS ether was effected by HF‚pyridine in this instance
because the more highly basic reagent TBAF tended to give
significant quantities of byproducts. The subsequent Swern
oxidation and addition of vinylmagnesium bromide to the
intermediate aldehyde was performed in one pot, as described
previously, to provide 37 as the only isolable product. The diene
37 underwent facile RCM with the ruthenium catalyst 34 to
yield the desired benzazocine 38. The relative stereochemistry
at C(7) and C(8) in 38 was initially assigned on the basis of a
comparison of the coupling constants for the protons at C(7)
and C(8) in 38 with the corresponding protons in 32. Namely,
for 32 and 38, the C(8) proton in each was an apparent triplet
with J ) 7.4 and 7.5 Hz, respectively, and the C(7) proton was
induced removal of the silyl protecting group on 29 furnished
an intermediate alcohol that was oxidized to the corresponding
aldehyde under Swern conditions. Rather than isolate the
aldehyde, excess vinylmagnesium bromide was simply added
to the reaction mixture to yield 30 as the only isolated product.29
The stereochemical outcome of this addition, which was
subsequently established by an X-ray analysis of 32 (vide infra),
may be rationalized either on the basis of a chelated transition
state involving the protected â-hydroxyl group or by the Felkin-
Anh model.30 Silylation of alcohol 30 provided diene 31.
Both 30 and 31 were viewed as possible substrates for the
RCM reaction, and at the time this work was performed, the
two most popular initiators of RCM were the Grubbs ruthenium
catalyst 34 and the Schrock molybdenum catalyst 35.31,32 Thus,
when the protected alcohol 31 was exposed to the ruthenium
catalyst 34 (10 mol %), the desired benzazocine 33 was obtained
in 22% yield, together with unreacted starting material (48%);
the use of additional quantities of catalyst did not increase the
yield of the product. While this success was certainly gratifying,
a more efficient RCM would be necessary to render this
approach to FR900482 practical. Because the molybdenum
catalyst 35 is more reactive and is frequently superior to 34
with sterically hindered substrates,33 we examined the RCM of
(28) Kondo, Y.; Sakamuto, T.; Kojima, S. J. Org. Chem. 1997, 62, 6507-
6511.
(29) See also: Ireland, R. E.; Norbeck, D. W. J. Org. Chem. 1985, 50,
2198-2200.
(30) (a) Cram, D. J.; Elhafez, F. A. A. J. Am. Chem. Soc. 1952, 74,
5828-5835. (b) Cram, D. J.; Wilson, D. R. J. Am. Chem. Soc. 1963, 85,
1245-1249. (c) Still, W. C.; McDonald, J. H. Tetrahedron Lett. 1980,
1031-1034. (d) Cherest, M.; Felkin, H. Tetrahedron Lett. 1968, 18, 2205-
2208. (e) Cherest, M.; Felkin, H.; Prudent, N. Tetrahedron Lett. 1968, 18,
2199-2204. (f) Anh, N. T.; Eisenstein, O. NouV. J. Chim. 1977, 1, 61-70.
(31) Schwab, P. E.; Grubbs, R. H.; Ziller, J. W. J. Am. Chem. Soc. 1996,
118, 100-110.
(32) Schrock, R. R.; Murdzek, J. S.; Bazan, G. C.; Robbins, J.; DiMare,
M.; O’Regan, M. J. Am. Chem. Soc. 1990, 112, 3875-3886.
(33) Grubbs, R. H.; Kirkland, T. A. J. Org. Chem. 1997, 62, 7310-
7318.
(34) For deprotection methods attempted, see (a) Stirling, I.; Ponsford,
R. J.; Goodacre, J. Tetrahedron Lett. 1975, 3609-3612. (b) Smith, C. W.;
Walter, R.; Stahl, G. L. J. Org. Chem. 1978, 43, 2285-2286. (c) Utz, R.;
Lieberknecht, A.; Griesser, H.; Potzolli, B.; Bahr, J.; Wagner, K.; Fischer,
P.; Schmidt, U. Synthesis 1987, 236-241. (d) Kaiser, E.; Kubiak, T. M.;
Merrifield, R. B.; Tam, J. P. Tetrahedron Lett. 1988, 29, 303-306. (e)
Bradshaw, J. S.; Krakowiak, K. E. Synth. Commun. 1996, 26, 3999-4004.
(f) Wensbo, D.; Apelquist, T. Tetrahedron Lett. 1996, 37, 1471-1472. (g)
Taylor, R. J. K.; Evans, E. F.; Lewis, N. J.; Kapfer, I.; Macdonald, G. Synth.
Commun. 1997, 27, 1819-1825.

10786 J. Am. Chem. Soc., Vol. 122, No. 44, 2000
Fellows et al.
Scheme 8
Scheme 9
a ddd with J ) 10.9, 7.4, 3.0, and 15.0, 7.5, 3.0 Hz, respectively.
This assignment was later verified by the X-ray structures of
40 and 41 (vide infra).
Our intention at this point was to exploit the stereochemistry
of the allylic alcohol to introduce the aziridine via an intramo-
lecular cyclization onto the olefin.35 Indeed, Shibasaki had
applied one such approach to preparing the mitomycin core.36
In accord with this precedent, the allylic alcohol 38 was treated
with tosyl isocyanate to afford the N-tosyl carbamate 39
(Scheme 8). Subsequent exposure of 39 to molecular iodine in
the presence of NaHCO3 gave a mixture (ca. 1:1.4) of the cyclic
carbamates 40 and 41, which were difficult to separate by
conventional chromatography, together with significant quanti-
ties (>30%) of recovered starting material. The structures of
40 and 41 were each established unequivocally by X-ray
crystallography. Because the stereochemistry at C(9) in the
major product 41 was opposite that required for FR900482, the
cyclization of 39 was not optimized.
It was, nevertheless, of interest to ascertain whether the minor
product 40 could be converted into the desired aziridine 42 by
the sequential opening of the carbamate ring, followed by
cyclization. However, when 40 was treated with methanolic
potassium carbonate, a mixture of compounds was produced
from which it was not possible to isolate 42 cleanly. The mass
spectrum of partially separated mixtures suggested 42 was
present, and examination of the 1H NMR spectra of these
mixtures revealed protons corresponding to those expected for
42, as well as signals in the olefinic region. This observation
indicated that base-induced elimination of HI was a significant
competing side reaction. In marked contrast to these results,
the isomeric product 41 underwent facile transformation upon
reaction with methanolic potassium carbonate to give the
aziridine 43 in 80% yield. These experiments, coupled with
those of Shibasaki, suggest that conformational effects arising
from the relative stereochemistry at C(7)-C(9) of benzazocines
related to 39-41 play a significant role in determining chemical
reactivity.
Because our original plan to introduce the aziridine ring
directly and efficiently onto the eight-membered ring of 39 was
unsuccessful, we completed a formal synthesis of FR900482
by intersecting with an advanced intermediate in Fukuyama’s
synthesis.7 In the event, treatment of the carbamate 38 with Zn
in HOAc induced the removal of the Troc protecting group to
afford the aniline 44 (Scheme 9). Oxidation of the amino group
in 44 with MCPBA at 0 °C furnished the sensitive hydroxy-
lamine moiety that was protected immediately by O-acetylation
to give 45. The double bond in 45 was somewhat resistant to
epoxidation and required the use of a saturated solution of
MCPBA at room temperature to afford a mixture of the desired
epoxide 46 (61% yield), together with recovered starting material
(18%). The stereochemistry of the epoxide relative to C(7) and
C(8) in 46 was supported by NOE experiments. Namely, there
was a NOE enhancement of 6.1% (4.3%) between Ha and Hc,
which is suggestive of a cis relationship, whereas the NOE
enhancements between Ha and Hb of 2.1% and between Hc and
Hb of 1.4% (3.2%) suggest trans relationships; these assignments
were confirmed in subsequent work (vide infra).
(35) For example, see: (a) Hirama, M.; Ito, S.; Iwashita, M.; Yamazaki,
Y. Tetrahedron Lett. 1984, 4963-4964. (b) Dreiding, A. S.; Egli, M. HelV.
Chim. Acta 1986, 69, 1442-1460. (c) Bergmeier, S. C.; Stanchina, D. M.
J. Org. Chem. 1997, 62, 4449-4456.
(36) Shibasaki, M.; Ban, Y.; Nakajima, S.; Yoshida, K.; Mori, M.; Date,
T. Heterocycles 1994, 39, 657-667.
Although it would have been possible to convert 46 into
FR900482 by following a sequence of reactions similar to that
(37) We thank Professor Tohru Fukuyama (Tokyo University) for copies
of the NMR spectra of 49.

Formal Total Synthesis of (+)-FR900482
J. Am. Chem. Soc., Vol. 122, No. 44, 2000 10787
employed by Fukuyama,7 little new knowledge would be gained
by such an exercise. We elected, instead, to complete the formal
synthesis of FR900482 by converting 46 into 49 by simply
replacing the protecting groups for the two primary alcohols in
46. Thus, the MPM group in 46 was exchanged for a tert-
butyldimethylsilyl group to yield 47 in a straightforward fashion.
Preferential hydrogenolysis of the benzylic benzyl ether in 47
could be achieved with H2 and Pd/C in ethyl acetate to afford
the alcohol 48 together with variable amounts of 47. Attempts
to force the reaction to completion led to competing hydro-
genolysis of the benzyl ether group on the phenol. Finally,
reaction of 48 with p-methoxyphenol under Mitsunobu condi-
tions delivered 49, the 1H and 13C NMR of which were identical
to that of an authentic sample supplied by Fukuyama.37
alkaloid. The prochiral diol 24 was enzymatically desymme-
trized using Pseudomonas species lipase to give 25 in 94%
enantiomeric excess. Inasmuch as subsequent adjustment of the
alcohol-protecting groups yielded the intermediate 26 in enan-
tiomerically enriched form, a formal enantioselective synthesis
of FR900482 has also been completed. The use of ring-closing
metathesis as a key step further validates the use of such
constructions in the synthesis of complex alkaloids, and other
applications of RCM reactions will be reported in due course.
Acknowledgment. We thank the National Institutes of
Health, The Robert A. Welch Foundation, Merck Research
Laboratories, and Pfizer, Inc. for generous support of this
research. We also thank Dr. Vince Lynch for performing the
X-ray crystallographic analyses and Steve Sorey for performing
2D NMR and NOE experiments.
Conclusion
A formal, enantioselective synthesis of the antitumor antibi-
otic (+)-FR900482 (1) has been completed using an approach
that featured the ring-closing metathesis of a highly function-
alized diene to yield the eight membered ring of the key
intermediate, benzazocine 38. Because a protected aziridine
having the stereochemistry required for 1 could not be cleanly
introduced onto 38, we transformed 38 into 49, which was a
key intermediate in Fukuyama’s elegant synthesis of racemic
FR900482, thereby completing a formal synthesis of the
Supporting Information Available: Experimental proce-
dures and complete characterization (1H and 13C NMR, IR, and
mass spectral data) for all new compounds, copies of 1H NMR
spectra, and X-ray data for the Mosher ester of 16 and for 32,
40, 41, and 43. This material is available free of charge via the
Internet at http://pubs.acs.org.
JA0013879