Total synthesis of the immunosuppressant FR901483 via an amidoacrolein cycloaddition

Article abstract of DOI:10.1021/ol015506g

Matrix presented The total synthesis of the potent immunosuppressant FR901483 is described. In a key step, the intermolecular Diels-Alder cycloaddition of an amidoacrolein with 2-(triisopropylsilyloxy)-1,3-butadiene produced the desired 3-cyclohexene-1-ca

Full text of DOI:10.1021/ol015506g

ORGANIC
LETTERS
2001
Vol. 3, No. 8
1125-1128
Total Synthesis of the
Immunosuppressant FR901483 via an
Amidoacrolein Cycloaddition
Jun-Ho Maeng and Raymond L. Funk*
Department of Chemistry, PennsylVania State UniVersity,
UniVersity Park, PennsylVania 16801
rlf@chem.psu.edu
Received January 2, 2001 (Revised Manuscript Received February 23, 2001)
ABSTRACT
The total synthesis of the potent immunosuppressant FR901483 is described. In a key step, the intermolecular Diels−Alder cycloaddition of
an amidoacrolein with 2-(triisopropylsilyloxy)-1,3-butadiene produced the desired 3-cyclohexene-1-carboxaldehyde. This compound was subjected
to basic followed by acidic conditions which effected two sequential aldol cyclizations to deliver the tricyclic ring system of the natural
product, suitably functionalized for completion of the total synthesis.
Organ transplant recipients can attribute their survival, in
part, to the discovery of the immunosuppressive agents
cyclosporin A and FK-506 (tacrolimus). However, these
compounds are toxic at high doses1 and, consequently, the
identification of additional immunosuppressants which func-
tion by mechanisms of actions different from these two
drugs2 remains an ongoing concern. To that end, the Fujisawa
company screened a number of microbial culture broths for
the inhibition of 12-O-tetradecanoylphorbol 13-acetate stimu-
lated T-cell proliferation in the presence of exogenous IL-2,
conditions which suppress the antiproliferative activity of
tacrolimus.3 An active compound, FR901483 (1), was
isolated from the fermentation broth of a fungal strain,
Cladobotryum sp. No. 11231, that was retrieved from litter
collected at Iwaki, Japan.3 It was further demonstrated that
FR901483 significantly prolongs graft survival time in the
rat skin allograft model, and moreover, evidence was
obtained which is suggestive of a different mechanism of
action, namely, inhibition of purine nucleotide biosyn-
thesis.4,5
(4) Addition of adenosine or deoxyadenosine (but not deoxyguanosine,
deoxycytidine, uridine or thymidine) resulted in elimination of the immu-
nosuppressive activity by FR901483.3 Moreover, the desphosphoryl com-
pound is not biologically active.3 Thus, FR901483 may inhibit one of the
key steps (shown below) for adenosine biosynthesis. Indeed, certain
structural similarities between these metabolites and FR901483 are intrigu-
ing.
(1) Fujita, T.; Hirose, R.; Yoneta, M.; Sasaki, S.; Inoue, K.; Kiuchi, M.;
Hirase, S.; Chiba, K.; Sakamoto, H.; Arita, M. J. Med. Chem. 1996, 39,
4451 and references therein.
(2) These drugs function by inhibiting the phosphatase activity of
calcineurin which, in turn, prevents the activation of T-cell-specific
transcription factors that are involved in the expression of the lymphokine
interleukin-2 (IL-2), see: Schreiber, S. L. Cell 1992, 70, 365 and references
therein.
(3) Sakamoto, K.; Tsujii, E.; Abe, F.; Nakanishi, T.; Yamashita, M.;
Shigematsu, N.; Izumi, S.; Okuhara, M. J. Antibiot. 1996, 49, 37.
(5) Inhibitors of purine biosynthesis are selectively cytotoxic toward
lymphocytes vis-a`-vis other cells since the former lack the purine salvage
pathway and must rely on de novo purine biosynthesis, see: Robinson, C.
J. Drugs Future 1995, 20, 356 and references therein.
10.1021/ol015506g CCC: $20.00 © 2001 American Chemical Society
Published on Web 03/23/2001

This promising biological activity coupled with an intrigu-
ing structure, secured by X-ray crystallographic analysis,3
elicited our interest in FR901483 as a target for total
synthesis.6 In particular, the relatively uncommon 1-alkyl-
1-aminocyclohexane substructure embodied in 1 might be
assembled by application of cycloaddition reactions of
clizations. We report herein the realization of this plan in a
total synthesis of racemic FR901483.
The precursor to amidoacrolein 6, 1,3-dioxin 9, was
prepared using our standard procedure.7 Thus, the imine
derived from the condensation of 2,2-dimethyl-1,3-dioxan-
5-one (8)9 with aminoacetaldehyde dimethyl acetal was
acetylated with acetic anhydride/triethylamine to afford
dioxin 9 in 83% yield (Scheme 2). Retro Diels-Alder
2-amidoacroleins, recently developed in our group.7
A
retrosynthetic analysis which outlines the implementation of
this synthetic protocol is shown in Scheme 1. Thus, it was
Scheme 2
Scheme 1
envisaged that lactam 2 possessed the necessary functionality
for the introduction of the C(1) p-methoxybenzyl and C(10)
methylamino substituents via enolate alkylation and amina-
tion reactions, respectively. A number of scenarios, albeit
multistep, are conceivable for the transposition of the
â-hydroxy ketone functionality present in lactam 3 to the
isomeric version in lactam 2. Lactam 3 could be constructed
by a “biomimetic” aldol ring closure of 4,8 which in turn
could be fashioned by another aldol-type cyclization within
aldehyde 5. Finally, a Diels-Alder cycloaddition of silyl-
oxydiene 7 with the 2-amidoacrolein 6 was expected to
furnish 5. It should be noted that this intermolecular
cycloaddition was expected to install not only the central
1-alkyl-1-aminocyclohexane substructure of FR901483 but
also both of the electrophilic and nucleophilic components,
appropriately tuned, for the pending sequential aldol cy-
reaction of dioxin 910 in warm benzonitrile (120 °C, 16 h)
generated the amidoacrolein 6 which was trapped in situ with
the silyloxydiene 711 to afford the desired cycloadduct 5
(64%).12 An aldol cyclization between the acetamide and
proximate aldehyde functionalities within 5 proceeded
smoothly (2 equiv of KOt-Bu, 10 equiv of EtOAc, THF, 0
(9) Commercially available from Acros Organics or prepared in two steps
from tris(hydroxymethyl)aminomethane hydrochloride. Hoppe, D.; Schmincke,
H.; Kleemann, H.-W. Tetrahedron 1989, 45, 687.
(10) (a) Funk, R. L.; Bolton, G. L. J. Am. Chem. Soc. 1988, 110, 1290.
(b) Funk, R. L.; Bolton, G. L. Tetrahedron Lett. 1988, 29, 1111. (c) Funk,
R. L.; Yost, K. J., III. J. Org. Chem. 1996, 61, 2598.
(11) Prepared following the procedure used for the preparation of 2-(tert-
butyldimethylsilyloxy)butadiene. Vedejs, E.; Eberlein, T. H.; Mazur, D. J.;
McClure, D. A.; Perry, D. A.; Ruggeri, R.; Schwartz, E.; Stults, J. S.; Varie,
D. L.; Wilde, R. G.; Wittenberger, S. J. Org. Chem. 1986, 51, 1556.
(12) The heterocycloadduct i was also isolated in 12% yield and was
the major product when 2-(trimethylsilyloxy)-1,3-butadiene was employed,
most likely a reflection of the relative steric bulk of the two silyl substituents.
For other examples of this steric effect, see: Rucker, C. Chem. ReV. 1995,
95, 1009.
(6) For previous synthetic work, see: (a) Quirante, J.; Escolano, C.;
Massot, M.; Bonjoch, J. Tetrahedron 1997, 53, 1391. (b) Yamazaki, N.;
Suzuki, H.; Kibayashi, C. J. Org. Chem. 1997, 62, 8280. (c) Braun, N. A.;
Ciufolini, M. A.; Peters, K.; Peters, E.-M. Tetrahedron Lett. 1998, 39, 4667.
(d) Braun, N. A.; Ousmer, M.; Bray, J. D.; Bouchu, D.; Peters, K.; Peters,
E.-M.; Ciufolini, M. A. J. Org. Chem. 2000, 65, 4397. (e) Snider, B. B.;
Lin, H.; Foxman, B. M. J. Org. Chem. 1998, 63, 6442. (f) Snider, B. B.;
Lin, H. J. Am. Chem. Soc. 1999, 121, 7778. (g) Bonjoch, J.; Diaba, F.;
Puigbo, G.; Sole, D.; Segarra, V.; Santamaria, L.; Beleta, J.; Ryder, H.;
Palacios, J.-M. Bioorg. Med. Chem. 1999, 7, 2891. (h) Sole, D.; Peidro,
E.; Bonjoch, J. Org. Lett. 2000, 2, 2225.
(7) The preliminary account of this work will be reported separately.
(8) The biosynthetic pathway to FR901483 most likely involves oxidative
cyclization of a tyrosyltyrosine derivative to a spirocyclic lactam and
subsequent aldol or Dieckman cylization to afford the tricyclic ring system.
For the preparation of spirolactams from tyrosine amides, see refs 6c and
6d. For the preparation of desmethylamino FR901483 and FR901483 via a
related internal aldol reaction, see refs 6e and 6f, respectively.
1126
Org. Lett., Vol. 3, No. 8, 2001

°C, 40 min) and directly afforded the corresponding conju-
gated lactam.13 This product was of sufficient purity for the
second aldol reaction, which was best accomplished under
acidic conditions (1:1 TFA, H2O, 0 °C f rt, 4 h), presumably
proceeding through the keto aldehyde intermediate 4 en route
to the desired â-hydroxy ketone in 79% yield for the two
steps. Thus, the potential of amidoacrolein cycloaddition
reactions in polycycle synthesis is validated by the assembly
of tricycle 3 in just four steps from dioxanone 8.
Scheme 3
The next phase of the synthesis involved the transposition
of aldol adduct 3 to the protected “aldol” adduct 2.
Accordingly, â-hydroxyketone 3 was subjected to conditions
(2.5 equiv of NaBH4, AcOH, rt, 30 min) which effected a
directed reduction14 of the carbonyl moiety of 3 and thereby
introduced the axial C(4) hydroxyl functionality of 10 (92%)
with complete stereocontrol. All attempts to either selectively
oxidize or protect the C(2) hydroxyl of 10 were unsuccessful.
However, the C(2) equatorial silyl ether of the bis-tert-
butyldimethylsilyl ether derivative of 10 could be selectively
cleaved (1 equiv of TBAF, rt, 1 h, 93% for the two steps)
and the resulting hydroxyl oxidized to afford ketone 2 and
thereby complete the aldol switch transformation. The enolate
derivative of 10 (1 equiv of KOt-Bu, THF, 0 °C, 20 min)
could be stereoselectively p-methoxybenzylated, provided
that an inverse addition protocol was employed (addition of
the enolate to 3 equiv of p-MeOBnBr, DMF, 0 °C, 20 min,
97%) in order to suppress dialkylation. A number of reducing
agents were examined (e.g., L-Selectride, DIBAL-H, AlH3,
NaBH4) in an attempt to stereoselectively reduce the resulting
ketone to the desired C(2) axial alcohol 14 (Scheme 3).
However, only the equatorial alcohol 11 was observed and,
consequently, the C(4) stereogenic center was inverted by
treatment of the nosylate derivative of 11 with rubidium
acetate (5 equiv) in the presence of 18-crown-6 (5 equiv) to
afford the desired acetate (64%) accompanied by, surpris-
ingly, the syn elimination product, alkene 13 (15%), and
recovered nosylate (16%).
(3 equiv of LDA, 3 equiv of trisyl azide, -78 °C, 5 min; 12
equiv of AcOH, -78 °C to rt, 1 h) to afford a diastereomeric
mixture of azides. Although the diastereomers were easily
separated by column chromatography (∆Rf ) 0.3), stereo-
chemical assignments derived from 1H NMR spectroscopic
analysis were not straightforward. Fortunately, X-ray crystal-
lographic analysis of the major isomer showed that it
possessed the desired stereochemistry at C(10) and, in
addition, that the other stereogenic centers had been correctly
assigned on the basis of the anticipated equatorial/axial
cyclohexane proton coupling patterns for the diagnostic
resonances.
Simultaneous reduction of the lactam and azide functional
groups16 with concomitant removal of the triethylsilyl
protecting group was accomplished by subjecting lactam 15
to LiAlH4 (5 equiv) to afford a diamino alcohol whose
primary amine functionality was selectively acylated with
benzyl chloroformate to deliver carbamate 16 (71% yield
for two steps). The remaining carbon atom of the natural
product was introduced by methylation of carbamate 16, and
deprotection of the TBS ether gave diol 17. The selective
phosphorylation of the less encumbered hydroxyl was
capricious when tetrabenzyl pyrophosphate was employed
but was reproducible if diol 17 was first converted to the
The final stage of the synthesis (Scheme 3) required three
major operations: introduction of the C(10) nitrogen atom,
reduction of the C(11) carbonyl, and substitution of the C(4)
phosphate moiety for the silyl ether functionality of alcohol
14, in turn prepared by straightforward saponification of
acetate 12 (7 equiv of KOH, MeOH, rt, 12 h, 74%). The
first of these objectives was accomplished by protection of
the hydroxyl group of 14 as the corresponding triethylsilyl
ether, reduction of the unsaturated lactam (PtO2, H2, EtOH,
12 h, 95% for two steps), and azidification of the enolate
derivative of the saturated lactam following Evans’ protocol15
(13) The â-hydroxylactam was obtained when the cyclization was
performed in the absence of ethyl acetate and could be dehydrated by
standard procedures (MesCl, NEt3). However, we serendipitously discovered
in one experiment that ethyl acetate (a contaminant from the column
chromatography of amide 5) remarkably facilitated the in situ dehydration,
presumably via transesterification to a â-acetoxylactam. The generality of
this procedure is under investigation.
(14) (a) Saksena, A. K.; Mangiaracina, P. Tetrahedron Lett. 1983, 24,
273. (b) Nutaitis, C. F.; Gribble, G. W. Tetrahedron Lett. 1983, 24, 4287.
(c) Turnbull, M. D.; Hatter, G.; Ledgerwood, D. E. Tetrahedron Lett. 1984,
25, 5449. (d) Gribble, G. W.; Nutaitis, C. F. Org. Prep. Proc. Int. 1985,
17, 317.
(16) For additional examples of the preparation of vicinal diamines by
reduction of R-azido lactams/amides, see: (a) Hu, J.; Jagdmann, G. E.
Tetrahedron Lett. 1995, 36, 3659. (b) Lopez-Herrera, F. J.; Sarabia-Garcia,
F.; Heras-Lopez, A.; Pino-Gonzalez, M. S. J. Org. Chem. 1997, 62, 6056.
(15) Evans, D. A.; Britton, T. C.; Ellman, J. A.; Dorow, R. L. J. Am.
Chem. Soc. 1990, 112, 4011.
Org. Lett., Vol. 3, No. 8, 2001
1127

corresponding phosphite ester6d and then oxidized with
m-CPBA in the presence of triethylamine to produce the
dibenzyl phosphate 18. Finally, hydrogenolysis of the diben-
zyl phosphate and benzyl carbamate functionalities provided
(() FR901483 which was characterized as the hydrochloride
salt. The 1H and 13C NMR and IR spectra were identical in
all respects with the previously published spectra as well as
more detailed spectra provided by the Fujisawa company.
In conclusion, we have completed the total synthesis of
the potent immunosuppressant FR901483 in 22 steps and
2.4% overall yield from dioxanone 8. In addition, we have
demonstrated that the easily accessible trifunctional arrays
of 2-amidoacroleins can be exploited in the rapid assembly
of tricyclic ring systems. Additional applications of this
methodology in natural product synthesis are now warranted
and will be reported in due course.
Acknowledgment. We are grateful to the NIH (GM-
28663) for financial support of this research. We thank Dr.
Douglas R. Powell of the University of Wisconsin for the
X-ray crystallographic analysis of compound 15. We thank
Dr. Takase in the Exploratory Laboratory of the Fujisawa
company for providing authentic spectra of FR901483.
Supporting Information Available: Characterization
data, detailed experimental procedures, and 1H and 13C NMR
spectra for all new compounds and X-ray crystallographic
data for 15. This material is available free of charge via the
Internet at http://pubs.acs.org.
OL015506G
1128
Org. Lett., Vol. 3, No. 8, 2001