A concise, stereocontrolled synthesis of (-)-saframycin A by the directed condensation of α-amino aldehyde precursors [1]

Full text of DOI:10.1021/ja993079k

10828
J. Am. Chem. Soc. 1999, 121, 10828-10829
Communications to the Editor
Scheme 1
A Concise, Stereocontrolled Synthesis of
(-)-Saframycin A by the Directed Condensation of
r-Amino Aldehyde Precursors
Andrew G. Myers* and Daniel W. Kung
Department of Chemistry and Chemical Biology
HarVard UniVersity, Cambridge, Massachusetts 02138
ReceiVed August 24, 1999
In this work we describe a short and enantioselective synthetic
route to the potent antitumor agent (-)-saframycin A (1), a
bisquinone alkaloid of microbial origin.1,2 The route employs a
new and powerful synthetic strategy involving the directed
condensation of optically active R-amino aldehydes. This strategy
evolved from retrosynthetic analysis of 1, as shown, where a series
of transformations initiated by the condensation of an aldehyde
with an amine (e.g., reductive amination, Pictet-Spengler, and
Strecker reactions) was envisioned to assemble the target 1 from
five simple components: hydrogen cyanide, formaldehyde, and
three R-amino aldehydes [two of which (structure 3) are the same,
hence the latent symmetry of 1]. The complexity of the analysis
arises in the determination of the precise order and stereochemistry
of bonding events that will link the precursors (seven bonds must
be formed), and upon consideration of the fundamental issues of
stability, reactivity, and protection strategies surrounding the
proposed use of optically active R-amino aldehydes as synthetic
intermediates. Recently, we reported the development of a series
of “C-protected” optically active R-amino aldehydes that incor-
porates an amino nitrile group as a masked aldehyde.3 Morpholino
nitrile derivatives, exemplified by structure 5, were found to be
particularly useful synthetic intermediates, undergoing condensa-
tion reactions with optically active N-protected R-amino aldehydes
with little to no epimerization of either component, thus establish-
ing the basis for the directed assembly of (-)-saframycin A
detailed herein.
Compounds 4 and 5, N- and C-protected versions of the same
chiral R-amino aldehyde (3), were prepared in high enantiomeric
excess from the same product of asymmetric alkylation of (-)-
pseudoephedrine glycinamide, as previously described.3,4 Addition
of N-protected R-amino aldehyde 4 (96% ee, 1.05 equiv) to
C-protected R-amino aldehyde 5 (92% ee, 1 equiv)5 in dichlo-
romethane at 23 °C in the presence of sodium sulfate cleanly
provided the imine 6 (presumed trans) without detectable epimer-
ization of either R-stereocenter (1H NMR analysis, >90% yield.
Addition of a saturated solution of anhydrous lithium bromide in
dimethoxyethane to the imine intermediate and warming to 35
°C brought about Pictet-Spengler cyclization to provide a ∼5:1
mixture of cis and trans tetrahydroisoquinolines, respectively.
Flash column chromatography afforded the desired cis product
(7) in 65-72% yield and 99% ee. The optical purity of 7 was
assayed by HPLC analysis (Chiralcel OD) of the corresponding
bis(benzoyl) derivative against an authentic sample of its enan-
tiomer, derived from (+)-pseudoephedrine via ent-4 and ent-5.
Lithium ion proved to be optimal for mild and selective Lewis
acid activation of the imine function without reaction of the
morpholino nitrile. The cis-trans selectivity of the cyclization
reaction varied markedly as a function of solvent and activating
agent; for example, use of lithium perchlorate in diethyl ether
provided the trans product exclusively. It is also noteworthy that
the transformation of 6 to 7 is the only step in the synthetic route
that was conducted above ambient temperature.
Introduction of the N-methyl group at this stage of the synthesis
was found to be optimal. Stirring 7 at 23 °C in the presence of
formalin (2.0 equiv) and sodium triacetoxyborohydride (1.5 equiv)
in acetonitrile provided the corresponding N-methylated com-
pound in 94% yield; the morpholino nitrile function was unaf-
fected by the reductive conditions. The N-Fmoc and O-TBS
protective groups were then cleaved. While these deprotections
could be performed simultaneously by the action of fluoride or
hydroxide, sequential removal of the silyl ether with acetic acid-
buffered tetrabutylammonium fluoride (2.4 and 1.1 equiv, re-
spectively) followed by cleavage of the carbamate with DBU (1.3
equiv) provided 8 with greater efficiency (92%). Notably,
compound 8 showed no propensity for the primary amine to add
to the masked aldehyde under such conditions as exposure to silica
gel or upon standing in the protic medium 2,2,2-trifluoroethanol,
further highlighting the stability of the morpholino nitrile protec-
tive group.
Addition of the third and final R-amino aldehyde component,
N-Fmoc glycinal (1.5 equiv), to amine 8 (1 equiv) in the presence
of sodium sulfate in deoxygenated dichloromethane at 23 °C
(2) Previous syntheses of the saframycins: (()-Saframycin B: (a) Fuku-
yama, T.; Sachleben, R. A. J. Am. Chem. Soc. 1982, 104, 4957-4958. (b)
Kubo, A.; Saito, N.; Yamato, H.; Masubuchi, K.; Nakamura, M. J. Org. Chem.
1988, 53, 4295-4310. (()-Saframycin A: (c) Fukuyama, T.; Yang, L.; Ajeck,
K. L.; Sachleben, R. A. J. Am. Chem. Soc. 1990, 112, 3712-3713.
(-)-Saframycin A: (d) Martinez, E. J.; Corey, E. J. Org. Lett. 1999, 1, 75-
77. See also: (e) Zhou, B.; Edmondson, S.; Danishefsky, S. J. Tetrahedron
Lett. Submitted for publication. (f) Zhou, B.; Danishefsky, S. J. Tetrahedron
Lett. Submitted for publication.
(3) Myers, A. G.; Kung, D. W.; Zhong, B.; Movassaghi, M.; Kwon, S. J.
Am. Chem. Soc. 1999, 121, 8401-8402.
(4) Myers, A. G.; Schnider, P.; Kwon, S.; Kung, D. W. J. Org. Chem.
1999, 64, 3322-3327.
(5) The sequence of reactions shown in Scheme 2 has been conducted
independently with both the (R)- [depicted] and (S)-morpholino nitrile
diastereomers. The yields and selectivities in both sequences were nearly
identical and in no event was scrambling of the morpholino nitrile stereo-
chemistry observed. We have also successfully executed the sequence using
a ∼1:1 mixture of the (R)- and (S)-diastereomers, converging on the
intermediate 2, this being the preferred route for large-scale synthesis of 1
and 2.
(1) Reviews: (a) Arai, T.; Kubo, A. In The Alkaloids; Brossi, A., Ed.;
Academic Press: New York, 1983; Vol. 21, Chapter 3. (b) Remers, W. A. In
The Chemistry of Antitumor Antibiotics; Wiley-Interscience: New York, 1988;
Vol. 2, Chapter 3.
10.1021/ja993079k CCC: $18.00 © 1999 American Chemical Society
Published on Web 11/05/1999

Communications to the Editor
J. Am. Chem. Soc., Vol. 121, No. 46, 1999 10829
Scheme 2a
cyclization, also at 23 °C, to provide a g9:1 mixture of cis and
trans tetrahydroisoquinolines, respectively. The desired cis isomer
(9) was isolated in 66% yield. The reaction solvent was again
critical for selective formation of the desired cis tetrahydroiso-
quinoline; protic solvents afforded predominantly the trans
diastereomer (e.g., in methanol, trans:cis >5:1). With the “trimer”
of R-amino aldehydes assembled (9), the C-terminus morpholino
nitrile blocking group was then cleaved with anhydrous zinc
chloride6 (3.0 equiv) in a mixture of trifluoroethanol and
tetrahydrofuran (2:1) at 23 °C, producing the key pentacyclic
intermediate 2 in 86% yield. This transformation presumably
proceeded by the sequential formation of iminium ion 10,
cyclization (addition of the secondary amine to the iminium ion),
expulsion of morpholine, and trapping of the resultant iminium
ion by cyanide. It was necessary to introduce exogenous cyanide
(trimethylsilyl cyanide, 2.0 equiv) to ensure complete amino nitrile
formation; in the absence of added cyanide, small amounts (5-
10%) of the hemiaminal corresponding to hydrolysis of amino
nitrile 2 were observed, presumably due to adventitious water.
Finally, the N-Fmoc protective group of 2 was cleaved with DBU
(1.3 equiv) at 23 °C in 88% yield, and the resulting primary amine
was acylated with pyruvoyl chloride (3.0 equiv) in the presence
of N,N-diethylaniline (1.1 equiv) at 0 °C to afford 11 (89%).
Oxidative demethylation of the hydroquinones with iodosobenzene
(2.5 equiv) in acetonitrile-water (1:1, 0 °C) furnished synthetic
(-)-saframycin A in 66% yield (127 mg of (-)-1).7 The synthetic
material was found to be identical in all respects (1H NMR, 13C
NMR, IR, HPLC, TLC analysis, and optical rotation) with an
authentic sample of natural saframycin A, kindly provided by
Professor T. Arai.
In summary, we have developed a practical and efficient
synthesis of (-)-saframycin A that proceeds in just 8 steps from
the R-amino aldehyde precursors 4 and 5, in ∼15% overall yield.
The synthesis illustrates a simple strategy for alkaloid assembly
that involves the directed condensation of R-amino aldehyde
precursors in a manner not unlike oligopeptide synthesis (here,
with C f N directionality). The present route is suitable for the
production of 1 in quantity; to date, more than 1 g of 2 and 200
mg of (-)-1 have been prepared.
Acknowledgment. Financial support from the National Institutes of
Health is gratefully acknowledged. D.W.K. acknowledges an NSF
predoctoral fellowship. We thank Professor Tadashi Arai for a sample
of natural saframycin A.
Supporting Information Available: Experimental procedures and
listings of spectral data (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
a Reaction conditions: (a) Na2SO4, CH2Cl2, 23 °C, >90%; LiBr, DME,
35 °C, 65-72%. (b) CH2O-H2O, NaBH(OAc)3, CH3CN, 23 °C, 94%.
(c) HOAc, TBAF, THF, 23 °C; DBU, CH2Cl2, 23 °C, 92%. (d) Na2SO4,
CH2Cl2, 23 °C, 66%. (e) ZnCl2, TMSCN, CF3CH2OH-THF, 23 °C, 86%.
(f) DBU, CH2Cl2, 23 °C, 88%. (g) ClCOCOCH3, PhNEt2, CH2Cl2, 0 °C,
89%. (h) PhIO, CH3CN-H2O, 0 °C, 66%.
JA993079K
(6) Guibe, F.; Grierson, D. S.; Husson, H.-P. Tetrahedron Lett. 1982, 23,
5055-5058.
(7) Two unstable byproducts, tentatively assigned as the regioisomeric o-,p-
and p-,o-quinones, were also isolated in a combined yield of ∼15%.
produced an imine intermediate that underwent Pictet-Spengler