Total synthesis of tricyclic azaspirane derivatives of tyrosine: FR901483 and TAN1251C

Article abstract of DOI:10.1021/ja016030z

A solution to the long-standing problem presented by the oxidative cyclization of a phenolic 3-arylpropionamide to a spirolactam has been developed in this laboratory via oxazoline chemistry. This research was motivated by our interest in some novel tricyclic azaspirane natural products formally derived from tyrosine, such as FR901483 and TAN1251C. In this paper, we disclose full details of the total synthesis of these substances.

Full text of DOI:10.1021/ja016030z

7534
J. Am. Chem. Soc. 2001, 123, 7534-7538
Total Synthesis of Tricyclic Azaspirane Derivatives of Tyrosine:
FR901483 and TAN1251C
Malika Ousmer,† Norbert A. Braun,‡ Claude Bavoux,§, Monique Perrin,§, and
Marco A. Ciufolini*,†,‡
Contribution from the Laboratoire de Synthe`se et Me´thodologie Organiques (LSMO), UMR CNRS 5078,
UniVersite´ Claude Bernard Lyon 1 and Ecole Supe´rieure de Chimie, Physique, Electronique de Lyon, 43,
BouleVard du 11 NoVembre 1918, 69622 Villeurbanne Cedex, France, Department of Chemistry, MS60,
Rice UniVersity, 6100 Main Street, Houston, Texas 77005-1892, and Laboratoire de Cristallographie,
UMR CNRS 5078, UniVersite´ Claude Bernard Lyon 1, 43, BouleVard du 11 NoVembre 1918,
69622 Villeurbanne Cedex, France
ReceiVed April 17, 2001. ReVised Manuscript ReceiVed June 7, 2001
Abstract: A solution to the long-standing problem presented by the oxidative cyclization of a phenolic
3-arylpropionamide to a spirolactam has been developed in this laboratory via oxazoline chemistry. This research
was motivated by our interest in some novel tricyclic azaspirane natural products formally derived from tyrosine,
such as FR901483 and TAN1251C. In this paper, we disclose full details of the total synthesis of these
substances.
Introduction
Scheme 1
FR901483, 1, and the TAN1251 family, exemplified by
TAN1251C, 2, are architecturally interesting natural products
that display a novel tricyclic azaspirane core (Scheme 1). The
TAN1251 compounds are muscarinic antagonists of potential
interest as antispasmodic or antiulcer agents, and were described
in 1991 by researchers at Takeda Industries.1 FR901483 is a
powerful immunosuppressant that appears to act by inhibiting
the biosynthesis of purines, especially adenine. The substance
was reported in 1996 by a team of Fujisawa scientists.2
Natural products 1 and 2 appear to share a common
biosynthetic antecedent in the form of a tyrosinyl tyrosine
dipeptide, 3 (Scheme 2). A hypothetical oxidative cyclization
of 3 may lead to spirolactam 4, further biosynthetic elaboration
of which could produce keto aldehyde 5. This intermediate
would advance to 1 or 2 via formation of a third ring
incorporating the erstwhile carboxy carbon, a (Scheme 2). Thus,
establishment of a C-C bond between atoms a and b leads to
1, whereas formation of a C-N bond between atoms a and c
would produce 2.
Noteworthy bioactivity and structural novelty have elicited
substantial interest in 1-2 at a synthetic chemical level;3 to wit,
four total syntheses of FR901483 and two of TAN1251 have
been recorded to date. The first route to 1 was disclosed by
Snider,4 who also determined the absolute configuration of the
molecule. Indeed, this important structural aspect had been left
unresolved in the original Fujisawa communication. In rapid
succession, Funk,5 Sorensen,6 and ourselves7 announced alterna-
tive synthetic avenues to the target molecule. The synthesis of
TAN1251 substances has been achieved by the groups of
Kawahara (racemate)8 and of Snider (enantiocontrolled).9 Once
again, the latter workers also determined the absolute config-
uration of these interesting molecules. A formal synthesis of
TAN1251A has been described by Wardrop.10
The biosynthetic hypothesis of Scheme 2 is central to the
synthetic plan that we formulated for 1-2 at the beginning of
the present effort. We felt that the synthesis of the target
(3) FR901483: (a) Yamazaki, N.; Suzuki, H.; Kibayashi, C. J. Org.
Chem. 1997, 62, 8280 (synthetic studies). (b) Snider, B. B.; Lin, H.; Foxman,
B. M. J. Org. Chem. 1998, 63, 6442 (synthesis of desmethylamino
FR901483). (c) Bonjoch, J.; Diaba, F.; Puigbo´, G.; Sole´, D.; Segarra, V.;
Santamar´ıa, L.; Beleta, J.; Ryder, H.; Palacios, J.-M. Bioorg. Med. Chem.
1999, 7, 2891 (analogues). (d) Brummond, K. M.; Lu, J. Org. Lett. 2001,
3, 1347 (synthetic studies). (e) Suzuki, H.; Yamazaki, N.; Kibayashi, C.
Tetrahedron Lett. 2001, 42, 3013 (synthetic studies).
(4) Snider, B. B.; Lin, H. J. Am. Chem. Soc. 1999, 121, 7778.
(5) (a) Funk, R. L.; Maeng, J.-H. Abstracts of the 220th ACS National
Meeting of the American Chemical Society, Washington, DC, August 20-
24, 2000; American Chemical Society: Washington, DC, 2000; Abs. ORGN
264. (b) Maeng, J.-H.; Funk, R. L. Org. Lett. 2001, 3, 1125.
(6) Scheffler, G.; Seike, H.; Sorensen, E. J. Angew. Chem., Int. Ed. Engl.
2000, 39, 4593.
(7) Ousmer, M.; Braun, N. A.; Ciufolini, M. A. Org. Lett. 2001, 3, 765.
(8) Nagumo, S.; Nishida, A.; Yamazaki, C.; Murashige, K.; Kawahara,
N. Tetrahedron Lett. 1998, 39, 4493.
* Address correspondence to this author: (e-mail) ciufi@cpe.fr.
† Universite´ Claude Bernard Lyon 1 and Ecole Supe´rieure de Chimie,
Physique, Electronique de Lyon.
‡ Rice University.
§ Universite´ Claude Bernard Lyon 1.
To whom inquiries regarding X-ray crystallographic studies should be
addressed.
(1) Shirafuji, H.; Tsubotani, S.; Ishimaru, T.; Harada, S. International
Patent PCT WO91/13887, 1999.
(2) Sakamoto, K.; Tsujii, E.; Abe, F.; Nakanishi, T.; Yamashita, M.;
Shigematsu, N.; Izumi, S.; Okuhara, M. J. Antibiot. 1996, 49, 37.
(9) Snider, B. B.; Lin, H. Org. Lett. 2000, 2, 643.
(10) Wardrop, D. J.; Basak, A. Org. Lett. 2001, 3, 1053.
10.1021/ja016030z CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/13/2001

Total Synthesis of Tricyclic Azaspirane DeriVatiVes of Tyrosine
J. Am. Chem. Soc., Vol. 123, No. 31, 2001 7535
Scheme 2
Scheme 4
Scheme 3
is subsequently acetylated to furnish 10, so as to suppress facile
Michael cyclization of 9.14
The ideas contained in Scheme 2 were clearly recognized
also by Snider and by Sorensen. Thus, Snider formulated an
analogous biosynthetic surmise in a landmark 1998 publication3b
on the synthesis of desmethylamino FR901483, even though at
a synthetic level he relied on chemical technology that does
not involve the key C-N bond forming process of Scheme 3.
On the other hand, Sorensen devised a particularly interesting
variant of our chemistry, in which a free secondary amine,
instead of an oxazoline, functions as the nucleophilic component
of the reaction, and went on to incorporate this noteworthy
transformation in his synthesis of 1. The Sorensen contribution
is significant because in our own work we were unable to devise
conditions suitable for the oxidative cyclization of secondary
amine substrates. Attempts in this sense resulted only in
formation of octahydroquinoline-7-ones,13,15 through Michael
cyclization of an 4-(ω-aminoalkyl)dienone intermediate, in a
fashion reminiscent of results obtained by Kita16 and Wipf.17
A strategic plan for the synthesis of 1-2 was formulated on
the basis of the foregoing chemical developments as shown in
Scheme 4. Spirolactam 13 (PAN ) p-anisyl) would be advanced
to intermediate 12, which represent the point of divergence of
the syntheses of FR901483 and TAN 1251C. These would arise
through intramolecular aldol cyclization of 11, or intramolecular
enamine formation from 14, respectively. Details of how these
plans were translated into practice are provided below.
alkaloids could be substantially simplified if it were possible
to duplicate the oxidative conversion of their presumed forerun-
ner 3 to spirolactam 4 by chemical means. A number of
interesting applications of this unusual transformation could be
envisioned well beyond the FR901483/TAN1251 problem.
However, the oxidative conversion of a phenolic amide to a
spirolactam was generally thought to belong to the realm of
“impossible” reactions. As early as 1987, Kita and collaborators
disclosed that oxidative treatment of generic substrates 5 does
not yield lactams 6, but rather lactones 7.11 These probably arise
through capture of an electrophilic intermediate produced by
activation of the phenol by the oxygen atom of the nucleophili-
cally ambident carboxamide group, followed by hydrolysis of
the emerging iminolactone upon workup (Scheme 3). Upon
completion of extensive methodology studies, we disclosed in
1998 that the desired transformation may be achieved by
oxidation of phenolic oxazolines 8 with iodobenzene diacetate
(“DIB”).12,13 The free OH group in the resulting spirolactams 9
Discussion
The construction of an oxazoline substrate suitable for the
conduct of the synthesis required fragments of 16 and 17, both
of which may be made from L-tyrosine, 15. A large quantity of
17 was secured in excellent chemical yield and stereochemical
purity by a 1915 procedure by Fischer.18 Conversely, the
(14) The structure of a related Michael cyclization product was ascer-
tained by X-ray diffractometry: Peters K.; Peters, E.-M.; Braun, N. A.;
Ciufolini, M. A. Z. Kristallogr. NCS 1999, 214, 555.
(15) Structures ascertained by X-ray crystallography: Peters, K.; Peters,
E.-M.; Braun, N. A.; Ciufolini, M. A. Z. Kristallogr. NCS 1999, 214, 273.
(16) Kita, Y.; Tohma, H.; Kikuchi, K.; Inagaki, M.; Yakura, T. J. Org.
Chem. 1991, 56, 435.
(11) Tamura, Y.; Yakura, T.; Haruta, J.-I.; Kita, Y. J. Org. Chem. 1987,
52, 3927.
(12) (a) Braun, N. A.; Ciufolini, M. A.; Peters, K.; Peters, E.-M.
Tetrahedron Lett. 1998, 39, 4667. See also: (b) Braun, N. A.; Bray, J. D.;
Ciufolini, M. A. Tetrahedron Lett. 1999, 40, 4985.
(13) Braun, N. A.; Ousmer, M.; Bray, J. D.; Bouchu, D.; Peters, K.;
Peters, E.-M.; Ciufolini M. A. J. Org. Chem. 2000, 65, 4397.
(17) Wipf, P.; Kim, Y. Tetrahedron Lett. 1992, 33, 5477.

7536 J. Am. Chem. Soc., Vol. 123, No. 31, 2001
Ousmer et al.
Scheme 5a
Scheme 6a
a Reagents: (a) MeOOCCl, NaHCO3, H2O/THF (1:1), 99%; (b)
Me2SO4, K2CO3, acetone, 99%; (c) LAH, THF, 85%; (d) aqueous KOH,
86%; (e) EtOH, HCl(g), 100%; (f) TsCl, aqueous Na2CO3, CHCl3, 98%;
(g) aqueous NaOH, 98%.
a Reagents: (a) CCl4, PPh3, NEt3, MeCN/pyridine (1:1), 73%; (b)
PhI(OAc)2, CF3CH2OH, then solid NaHCO3; (c) Ac2O, pyridine, 41%
(for b + c); (d) H2, PtO2, EtOAc, 96%. PAN ) p-anisyl
preparation of 16 by literature methods19 afforded material of
mediocre optical purity. We thus devised an alternative route
(Scheme 5), which proceeded efficiently and caused no erosion
of stereochemical integrity. The optical quality of 16 was readily
readily cleavable Fukuyama-type nitrosulfonamide,24 was mo-
tivated by our desire to retain the original N-protecting group
during all transformation requiring reductants or nucleophiles.
The free amine would be released during vigorous hydride
reduction of the spirolactam segment to the corresponding
pyrrolidine. Notice that acetylation of the primary product of
oxidative cyclization results also in N-acylation of the tosyl-
amide. This event is of no import, because both acetyl groups
are removed simultaneously at a later stage of the synthesis.
Conversion of dienone 19 to the corresponding cyclohexanone
was effected by hydrogenation in the presence of PtO2 (Adams
catalyst). Palladium or rhodium catalysts were unsatisfactory
for the present application, due to their tendency to promote
reductive aromatization of the substrate through C-N bond
cleavage. This troublesome side reaction occurs to the extent
of 50-60% when hydrogenation is attempted over supported
Pd, but it constitutes the exclusive outcome when Rh catalysts
are employed. It is also worthy of note that the use of supported
platinum catalysts, e.g., Pt(C), resulted in formation of variable
amounts of the cyclohexanol corresponding to 20. No such
problem was observed with the Adams catalyst. Saturation of
dienones related to 19 may also be effected by hydrogenation
over Raney nickel, but in this case the fully reduced cyclohex-
anol is obtained.6
A straightforward series of reactions advanced intermediate
20 to keto aldehyde 23 (Scheme 7), which constitutes the
substrate for the crucial aldol cyclization leading to the
obligatory intermediate 24. The Snider synthesis of 1 employs
an analogous aldol step, details of which were first disclosed
in a preliminary communication3b that appeared while we were
researching the same transformation. We were thus assisted in
the optimization of our own reaction conditions by the important
observations recorded by these workers, who ultimately chose
tBuOK in tBuOH for the conduct of this step. We favor the
use of sodium methoxide in 90% aqueous methanol to ac-
complish the same transformation.25 Regio- and diastereose-
lectivity seem to be more satisfactory under these conditions.26
Compound 24 was the major, but not the exclusive, product
thus formed, and it was obtained in 44% yield after chromato-
graphic purification. It should be noted that aldehyde 23 is fairly
1
assessed by an H NMR shift study using (+)-Eu(hfc)3 as the
chiral shift reagent.
The union of an amino alcohol and a carboxylic acid to form
an oxazoline20 may be achieved by various methods, among
which the Wipf21 and the Vorbru¨ggen22 techniques are especially
effective. In the present case, the Vorbru¨ggen protocol represents
the method of choice, in that it leads to the desired heterocycle
in one step and it tolerates an unprotected phenolic function in
component 17. Conversely, the Wipf procedure involves cy-
clization of a preformed N-hydroxyethyl amide with the Burgess
reagent,23 and it requires protection of the phenolic OH to
suppress formation of sulfate esters. These are not readily
converted back to the free phenol, to the detriment of overall
yields.
Oxazoline 18 underwent DIB oxidation/acetylation to 19. We
note that the nature of the protecting group applied to the lateral
amino group in 18 is crucial for the success of the cyclization
step. In particular, a carbonyl-type blocking unit, e.g., BOC, is
unsatisfactory because it competes effectively with the oxazoline
for the electrophilic intermediate produced through DIB activa-
tion of the phenol. Products unrelated to the desired spirolactam
are thus obtained.12,13 The protecting group of choice here is a
sulfonamide, as thoroughly detailed in our previous communica-
tions. The sulfonamide does not interfere with the cyclization
step and it facilitates purification of the stereochemically labile
oxazoline by a particularly mild acid-base extraction (cf.
Experimental Section). Significantly, protection of the pendant
amino group as a sulfonamide is also apparent in the Sorensen
synthesis of 1.6 Our choice of a tosylamide, rather than a more
(18) Fischer, E.; Lipschitz, W. Ber. Dtsch. Chem. Ges. 1915, 48, 360. It
is noteworthy that direct tosylation of tyrosine gave only ditosylated tyrosine
(conversion 50%) (McChesney, E. V.; Swann, Wm. K., Jr. J. Am. Chem.
Soc. 1937, 59, 1116), whereas tyrosine ethyl ester gave clean monotosylated
product.
(19) The preparation of 16 has been described byAbarbri et al. and Jung
et al. (Abarbri, M.; Guignard, A.; Lamant, M. HelV. Chim. Acta 1995, 78,-
109. Jung, M. E.; Jachiet, D.; Rohloff, J. R. Tetrahedron Lett. 1989, 30,
4211). Unfortunately, the product thus obtained is essentially racemic (cf.
Supporting Information).
(20) Review: Gant, T. G.; Meyers A. I. Tetrahedron 1994, 50, 2297.
(21) (a) Wipf, P.; Miller, C. P. Tetrahedron Lett 1992, 33, 907. (b) Wipf,
P.; Kim, Y.; Goldstein, D. M. J. Am. Chem. Soc. 1995, 117, 11106. (c)
Wipf, P.; Li, W. J. Org. Chem. 1999, 64, 4576.
(24) Fukuyama, T.; Jow, C.-K.; Cheung, M. Tetrahedron Lett. 1995, 36,
6373.
(22) Vorbru¨ggen, H.; Krolikiewicz, K. Tetrahedron 1993, 49, 9353.
(23) (a) Atkins, G. M.; Burgess, E. M. J. Am. Chem. Soc. 1968, 90,
4744. (b) Burgess, E. M.; Penton, H. R., Jr.; Taylor, E. A.; Williams, W.
M. Organic Syntheses; Wiley: New York, 1988; Collect. Vol. VI, p 788.
(25) The terminology “sodium methoxide in aqueous methanol” stands
only to indicate the reagents utilized in this step, and it should not be
interpreted as an inference regarding the precise nature of the basic species
present in solution and/or responsible for the reaction.

Total Synthesis of Tricyclic Azaspirane DeriVatiVes of Tyrosine
J. Am. Chem. Soc., Vol. 123, No. 31, 2001 7537
Scheme 7a
Scheme 8a
a Reagents: (a) K2CO3, MeOH, 79%; (b) MeI, K2CO3, acetone/DMF
97%; (c) catalyst TPAP, NMO, MS 4 Å, CH2Cl2, 77%; (d) NaOMe,
90% aqueous MeOH, 44%. PAN ) p-anisyl
resistant to epimerization at C-6. This property, first recorded
by Snider,4 is consistent with recent observations by Myers27
and earlier ones by Garner28 concerning the configurational
stability of amino acid-derived aldehydes. Stereochemical
stability minimized the probability of formation of aldol isomers
possessing the undesired C-6 (R)-configuration. Aldol products
were best characterized as the corresponding acetates (cf. 25),
obtained in quantitative yield by the standard treatment with
acetic anhydride in pyridine.
a Reagents: (a) Ac2O, pyridine, DMAP, CH2Cl2, 93%; (b) L-
Selectride, THF, 95%; (c) NsCl, NEt3, DMAP, CH2Cl2, 72%; (d)
CsOAc, 18-cr-6, PhH, 73%; (e) LAH, THF; (f) Cbz-Cl, NEt3, DMAP,
CH2Cl2, 70% (for e + f); (g) i-Pr2NP(OBn)2, tetrazole, CH2Cl2, then
t-BuOOH, decane, 29% + 30; (h) H2, Pd(C), aqueous HCl, MeOH,
94%. PAN ) p-anisyl.
byproduct (12%). Global deprotection/reduction of 28 was
achieved in high yield by vigorous LAH treatment, and the
secondary amino group in the emerging 29 was protected as a
benzyl carbamate prior to selective phosphorylation of the C-9
carbinol. Notice that this step required no protection of the C-7
alcohol, which is rather hindered. Final hydrogenolysis of all
benzyl groups in the presence of aqueous HCl provided the bis-
hydrochloride salt of 1, which was identical in all respects to
material prepared from an authentic sample of FR901483, kindly
provided by the Fujisawa Pharmaceutical Company.
The final sequence that produced our first samples of fully
synthetic FR901483 (Scheme 8) commenced with reduction of
the ketone in 25 to the corresponding alcohol. The shape of the
molecule disfavors the approach of reducing agents from the
top (Re) face of the ketone, so that the desired C-9 axial carbinol
is not directly available from 25. This necessitates an ultimate
inversion of configuration at C-9. In our case, reduction was
effected by the use of L-Selectride, in the hope of obtaining at
least some of the correct carbinol diastereomer,29 but in fact
the reaction occurred with complete diastereocontrol in favor
The Sorensen synthesis of 1 demonstrated the feasibility of
a Mitsunobu-type inversion of configuration of the C-9 carbinol
in substrates similar to 26.6 This observation allowed us to
simplify our own synthesis as shown in Scheme 9. Thus, LAH
reduction of compound 24 produced a 6:1 mixture of equatorial
(compound 33, major component) and axial (compound 29,
minor component) amino alcohols. This somewhat surprising
stereochemical result is attributable to the greater reactivity,
hence the lower selectivity, of LAH relative to other reducing
agents. Separation of the two diastereomeric carbinols was
difficult at this stage, due to their great polarity. Accordingly,
the mixture was treated with excess dibenzyl phosphate, which
presumably converted both secondary and tertiary amino groups
to the corresponding salts, followed by DIAD and tris(4-
chlorophenyl)phosphine. The emerging Mitsunobu products
were again extremely polar and difficult to handle, necessitating
conversion to N-Cbz derivatives prior to chromatography and
characterization. The presumed major product, 34, was thus
retrieved as compound 31. Separation of the minor diastereomer
produced during LAH reduction was readily effected at this
stage, providing us with a small sample of substance 35. Overall
yields were consistent with those described by Sorensen.6 It thus
seems that although the Mitsunobu step may be conducted in
the presence of an unprotected methylamino substituent, techni-
cal difficulties rule in favor of protection of the secondary amine
prior to inversion at C-9, just as described by the Scripps team.
1
of the equatorial alcohol 26 (within the limits of 500 MHz H
NMR spectroscopy). The structure of this intermediate was
verified by X-ray crystallography.30 Inversion of C-9 configu-
ration was achieved by the method of Snider, via p-nitroben-
zenesulfonate ester 27. The resulting 28, the structure of which
was also ascertained by X-ray diffractometry,31 was obtained
in a satisfactory 73% yield, but olefin 32 was a significant
(26) The beneficial effect of protic solvents on the diastereoselectivity
of the aldol step in favor of 23 was first described by Snider (ref 4). We
independently found that solvents of increasingly greater hydrogen bonding
power afforded improved diastereoselectivity in favor of 23. Conduct of
the reaction in aqueous MeOH provided the best selectivity in the desired
sense. A referee suggested the use of trifluoroethanol as the solvent for
this reaction. However, fluorinated alcohols were not explored as possible
aldol solvents.
(27) (a) Myers, A. G.; Kung, D. W.; Zhong B. J. Am. Chem. Soc. 2000,
122, 3236. (b) Myers, A. G.; Zhong, B.; Kung, D. W.; Movassaghi, M.;
Lanman, B. A.; Kwon, S. Org. Lett. 2000, 2, 3337. (c) Myers, A. G.; Kung,
D. W. Org. Lett. 2000, 2, 3019. (d) Myers, A. G.; Kung, D. W. J. Am.
Chem. Soc. 1999, 121, 10828.
(28) (a) Garner, P. Tetrahedron Lett. 1984, 25, 5855. Review: (b)
Jurczak, J.; Golebiowski, A. Chem. ReV. 1989, 89, 149.
(29) The axial carbinol stereoisomer is normally the major product of
reduction of cyclohexanones with Selectride reagents: Brown, H. C.;
Krishnamurthy, S. J. Am. Chem. Soc. 1972, 94, 7159.
(30) Ousmer, M.; Braun, N. A.; Ciufolini, M. A.; Perrin, M. Z.
Kristallogr. NCS 2000, 215, 597.
(31) Ousmer, M.; Braun, N. A.; Ciufolini, M. A.; Perrin, M.; Bavoux,
C. Z. Kristallogr. NCS 2001, in press.

7538 J. Am. Chem. Soc., Vol. 123, No. 31, 2001
Ousmer et al.
Scheme 9a
Scheme 11
trifluoroacetyl group and cyclization of the intermediate second-
ary amine (not observed) to fully synthetic TAN 1251C, 2,
whose properties were identical to those reported in the literature
for the natural product. Whereas this step achieved the total
synthesis, the overall yield of the final sequence was disap-
pointing. We presume that the basic treatment required to
liberate the secondary amine might have diverted a good portion
of 40 into aldol manifolds. Considerable improvement in overall
efficiency was observed upon protection of the amino group as
a 2,2,2-trichloroethyl carbamate (Troc), followed by TPAP/
NMO oxidation32 of the resulting 39 to 41 and final Troc
deprotection/cyclization by the use of the Cd/Pb couple.33
A summary of our approach to 1 and 2 appears in Scheme
11. The longest linear sequence leading to FR901483 through
a Snider-type inversion encompasses 20 steps from commercial
L-tyrosine and it proceeds with an overall yield of 1%. The
alternative synthesis involving a Sorensen-type Mitsunobu
inversion is shorter (17 steps), and it affords identical overall
yields (1.3%). This compares favorably with the other enan-
tiocontrolled routes of 1, both of which started with N-Boc-
tyrosine and apparently proceeded in a total of 224 and 186 linear
steps from this educt. Likewise, our synthesis of TAN1251C
requires a maximum of 16 linear steps from L-tyrosine (4%
overall yield). The alternative enantiocontrolled route to this
natural product requires 18 steps from N-Boc-tyrosine (7.2%
overall yield).9
a Reagents: (a) LAH, THF, reflux; (b) 6 equiv of (BnO)2P(O)OH,
THF, tris(4-chlorophenyl)phosphine, DIAD, then, NEt3; (c) Cbz-Cl,
NaHCO3/THF, 26% for a-c.
Scheme 10a
In conclusion, our oxazoline-based avenue to spirolactams
appears to be capable of sustaining synthetic efforts toward
molecules of at least moderate complexity, such as 1 and 2,
and to simplify the synthetic problem posed by these structures
to a significant extent. Further applications of oxidative cy-
clizations of oxazolines are currently under study and new
developments in this area will be disclosed in due course.
Acknowledgment. We thank the MENRT (Fellowship to
M. O.), the CNRS, the Re´gion Rhoˆne-Alpes, the NIH (CA-
55268), the NSF (CHE 95-26183), and the R. A. Welch
Foundation (C-1007) for support of our research. M.A.C. is a
Fellow of the A. P. Sloan Foundation (1994-1998) and the
recipient of a Merck & Co. Academic Development Award
(2000, 2001). We are especially grateful to Dr. S. Goto and
Mr. T. Sonoda, both from Fujisawa Pharmaceutical Co., for the
sample of authentic FR901483. The authors also express their
gratitude to Dr. Denis Bouchu and Laurence Rousset, of the
mass spec facility of the LSMO, for measuring the mass spectra.
a Reagents: (a) BBr3, CH2Cl2, -60 °C, 87%; (b) prenyl bromide,
Cs2CO3, acetone, 98%; (c) LAH, THF; (d) Troc-Cl, NaHCO3/THF,
60% (for c + d); (e) catalyst TPAP, NMO, MS 4 Å, CH2Cl2, 63%; (f)
Cd/Pb couple, aqueous NH4OAc/THF, 79%; (g) DMSO, TFAA, Et3N,
CH2Cl2, -78 °C; (h) K2CO3, MeOH/H2O, 10% (for g + h) + 40.
We now turn to the total synthesis of TAN 1251C, 2, which
Snider has shown to represent the precursor to the entire
TAN1251 family of compounds.9 A synthesis of 2 thus amounts
to a formal synthesis of all other TAN1251 substances. Our
avenue to 2 (Scheme 10) evolved from spirolactam 22, which
was subjected to O-demethylation and prenylation of the
intermediate phenol. Substance 37 thus obtained underwent
stereoselective LAH reduction to aminodiol 38. As reported by
Snider for a similar system, this material may be oxidized to
keto aldehyde 40 with DMSO/TFAA without prior protection
of the secondary amine, because the latter apparently forms the
corresponding trifluoroacetamide in situ under these conditions.
Exposure of 40 to methanolic K2CO3 induced release of the
Supporting Information Available: Experimental details
and 1H and 13C NMR spectra of all compounds described herein
(PDF). This material is available free of charge via the Internet
at http://pubs.acs.org. See any current masthead page for
ordering information and Web access instructions.
JA016030Z
(32) Ley, S. V.; Normand, J.; Griffith, W. P.; Marsden, S. P. Synthesis
1994, 639.
(33) Dong, Q.; Anderson, C. E.; Ciufolini, M. A. Tetrahedron Lett. 1995,
36, 5681.