Cyclotrimerization strategy toward analogues of amaryllidaceae constituents. Synthesis of deoxygenated pancratistatin core

Article abstract of DOI:10.1021/ol052372o

(Chemical Equation Presented) A derivative of pancratistatin having no oxygenation in the aromatic ring was synthesized by a new strategy based on the cobalt-catalyzed cyclotrimerization of acetylenes as a prelude to diversity-oriented synthesis of furthe

Full text of DOI:10.1021/ol052372o

ORGANIC
LETTERS
2005
Vol. 7, No. 25
5669-5672
Cyclotrimerization Strategy toward
Analogues of Amaryllidaceae
Constituents. Synthesis of
Deoxygenated Pancratistatin Core
Michael Moser, Xuetong Sun, and Tomas Hudlicky*
Department of Chemistry and Centre for Biotechnology, Brock UniVersity,
500 Glenridge AVenue, St. Catharines, Ontario L2S 3A1, Canada
thudlicky@brocku.ca
Received September 30, 2005
ABSTRACT
A derivative of pancratistatin having no oxygenation in the aromatic ring was synthesized by a new strategy based on the cobalt-catalyzed
cyclotrimerization of acetylenes as a prelude to diversity-oriented synthesis of further analogues.
Activity in the total synthesis of Amaryllidaceae constituents1
has been intense for over 20 years. The major targets in this
group of natural products, pancratistatin (1), 7-deoxypan-
cratistatin (2), narciclasine (3), and lycoricidine (4), have all
been synthesized several times by diverse and creative
strategies.2 Driven by promising antitumor activities espe-
cially of the phenol-containing compounds 1 and 3, major
effort has been devoted to the preparation of many unnatural
analogues in order to ameliorate the bioavailability of these
poorly soluble agents.3 The precise mode of action of the
most potent constituent, pancratistatin, is unknown, and only
limited information is available concerning the interactions
of narcilasine with RNA.4 The search for the minimum
pharmacophore and the mode of action resulted in the
syntheses and biological screening of many unnatural deriva-
tives, yielding essential information about some of the
functional groups required for activity. Those regions that
are crucial to maintaining the activity of pancratistatin are
indicated in Figure 1.
(1) At the suggestion of Professor R. Pettit (Arizona State University)
we no longer use the descriptor “alkaloids” when referring to the members
of the amaryllidaceae family. With the exception of lycorine-type alkaloids
none contain a basic nitrogen.
(2) For reviews of synthetic strategies, see: (a) Prabhakar, S.; Tavares,
M. R. Alkaloids Chem. Biol. 2001, 15, 433. (b) Rinner, U.; Hudlicky, T.
Synlett 2005, 365. (c) Polt, R. Organic Synthesis: Theory and Applications;
Hudlicky, T., Ed.; JAI: Greenwich, CT, 1997; Vol. 3, p 109. (d) Martin,
S. F. In The Alkaloids; Bross, A. R., Ed.; Academic: New York, 1987;
Vol. 40, p 251. (e) Jin, Z. Nat. Prod. Rep. 2003 20, 606. (f) Jin, Z. Nat.
Prod. Rep. 2005, 22, 111.
Figure 1. Amaryllidaceae constituents.
First, the presence of phenol and phenanthridone func-
tionalities as a potential donor-acceptor pair is essential;
the absence of phenolic hydroxyl, as in 7-deoxypancratistatin,
10.1021/ol052372o CCC: $30.25
© 2005 American Chemical Society
Published on Web 11/09/2005

leads to a 8- to 32-fold decrease in activity in those cell lines
against which both pancratistatin and 7-deoxypancratistatin
were tested.4f,5 Up to now, it has been assumed that
compounds lacking the amide carbonyl are inactive, as has
recently been demonstrated with lactone analogues of ly-
coricidine and pancratistatin.6
Second, changes in the substitution patterns or function-
alities elsewhere in the aromatic core lead to compounds with
decreased activity, as has recently been shown with several
analogues,4f including an indole-containing mimic of 1.4e,7
Third, the aminoinositol moiety must remain essentially
intact, except for variations of substituents, functionalities,
and configuration at C-1.4a,4b Finally the trans-ring junction
is essential to activity; the cis-fused 7-deoxypancratistatin
is inactive.8 Thus the potential for variation in structure
appears to be the greatest in the region between C-1 and
C-8/C-9/C-10 portion of the aromatic core. Examination of
all recent strategies reveals that no general method has yet
been implemented that would permit access to a large variety
of structures with systematic variations in the aromatic
nucleus only. In this manuscript we report the successful
synthesis of the deoxygenated aromatic core of pancratistatin
by cyclotrimerization of acetylene derivatives with scaffolds
that will allow also for the synthesis of heteroaromatic
variants by incorporation of nitriles into the coupling
sequence. Our strategy to access analogues of the title
compounds such as 5 from fully oxygenated scaffold 7 is
shown in Figure 2.
Figure 2. Retrosynthetic view of analogue synthesis by cyclotri-
merization.
equipped with substituents capable of yielding oxygen or
other functionalities through further manipulations. A review
of literature suggests that cyclotrimerizations are, in general,
not high yielding (an exception being the above-mentioned
estrone synthesis), and the choice of catalysts and precise
reaction engineering require substantial research effort.
Despite such limitations we viewed the cyclotrimerization
approach as a viable route to a large number of compounds
with extensive variations in the aromatic core, attainable from
a simple scaffold such as 7. Because functional and
configurational changes at C-1 of the natural products do
not greatly diminish biological activity, we chose to pursue
the synthesis of both the natural series as well as the series
epimeric at C-1 in order to provide possibilities of further
functionalization in that region of the targets. The synthesis
required a medium-to-large scale optimization for the
preparation of tosylaziridine 8,12 which is available by a
three-step protocol (49% over three steps) from the optically
pure diol 9,13 Scheme 1.
Cyclotrimerization of acetylenes, discovered by Berthelot
in 1866,9 has been used extensively in the syntheses of carbo-
and heterocyclic compounds10 and occasionally applied to
natural product synthesis. Vollhardt’s estrone preparation is
perhaps the best known application.11 The proposed strategy
for the synthesis of Amaryllidaceae analogues depends on
the cyclotrimerization of scaffold 7 with acetylene derivatives
Exposure of 8 to the aluminum complex derived from
lithium trimethylsilylacetylene provided tosylamide 10 in
69% yield after reprotection of the cis-diol. The cyclic sulfate
12 was prepared from diol 11 (OsO4/NMO; 44%, 76% by
conversion) by treatment with SO2Cl2 and NEt3 in 82%
yield.14 Reaction of 12 with ammonium benzoate provided
a mixture of 13 and enyne 14 (1:2), the latter compound
possessing the features of narciclasine or lycoricidine in the
aminoconduritol ring. A study of this reaction revealed that
14 is not derived from a syn elimination of the cyclic sulfate
from C-1 (pancratistatin numbering) but rather by trans
elimination of the intermediate sulfate anion derived from
13. The sulfate anion is situated syn to the proton at C-10b
and facilites the elimination of the benzoate group intramo-
lecularly. The cyclic sulfate 12 and the alcohol 13 are stable
to reaction conditions that exclude ammonium benzoate. It
will eventually be desirable to optimize this reaction in order
(3) (a) Pettit, G. R.; Freeman, S.; Simpson, M. J.; Thompson, M. A.;
Boyd, M. R.; Williams, M. D.; Pettit, G. R., III; Doubek, D. L. Anti-Cancer
Drug Des. 1995, 10, 243. (b) McNulty, J.; Mao, R.; Mo, R.; Wolf, S.;
Pettit, G. R.; Herald, D. L.; Boyd, M. R. Bioorg. Med. Chem. Lett. 2001,
11, 169. (c) Hudlicky, T.; Rinner, U.; Gonzales, D.; Akgun, H.; Schilling,
S.; Siengalewicz, P.; Martinot, T. A.; Pettit, G. R. J. Org. Chem. 2002, 67,
8726. (d) Phung, A. N.; Zannetti, M. T.; Whited, G.; Fessner, W. D. Angew.
Chem., Int. Ed. 2003, 42, 4821. (e) Rinner, U.; Hudlicky, T.; Gordon, H.;
Pettit, G. R. Angew. Chem., Int. Ed. 2004, 43, 5342. (f) Rinner, U.;
Hillebrenner, H. L.; Adams, D. R.; Hudlicky, T.; Pettit, G. R. Bioorg. Med.
Chem. Lett. 2004, 14, 2911.
(4) (a) Carrasco, L.; Fresno, M.; Vazquez, D. FEBS Lett. 1975, 52, 236.
(b) Jimenez, A.; Sanchez, L.; Vasquez, D. FEBS Lett. 1975, 55, 53.
(5) Gabrielsen, B.; Monath, T. P.; Huggins, J. W.; Kefauver, D. F.; Pettit,
G. R.; Groszek, G.; Hollingshead, M.; Kirsi, J. J.; Shannon, W. M.; Schubert,
E. M.; Dare, J.; Ugarkar, B.; Ussery, M. A.; Phelan, M. J. J. Nat. Prod.
1992, 55, 1569.
(6) Ibn-Ahmed, S.; Khaldi, M.; Chretien, F.; Chapleur, Y. J. Org. Chem.
2004, 69, 6722.
(7) Hudlicky, T.; Rinner, U.; Finn, K. J.; Ghiviriga, I. J. Org. Chem.
2005, 70, 3490.
(8) Rinner, U.; Siengalewicz, P.; Hudlicky, T. Org. Lett. 2002, 4, 115.
(9) Berthelot, M. C. R. Acad. Sci. 1866, 62, 905.
(10) For reviews, see: (a) Vollhardt, K. P. C. Angew. Chem., Int. Ed.
Engl. 1984, 23, 539. (b) Bo¨nnemann, H. Angew. Chem., Int. Ed. Engl. 1985,
24, 248. (c) Lautens, M.; Klute, W.; Tam, W. Chem. ReV. 1996, 96, 46. (d)
Schore, N. E. Chem. ReV. 1988, 88, 1081. (e) Saito, S.; Yamamoto, Y.
Chem. ReV. 2000, 100, 2901. (f) Varela, J. A.; Saa´, C. Chem. ReV. 2003,
103, 3787.
(11) (a) Funk, R. L.; Vollhardt, K. P. C. J. Am. Chem. Soc. 1977, 99,
5483. (b) Funk, R. L.; Vollhardt, K. P. C. J. Am. Chem. Soc. 1979, 101,
215.
(12) Hudlicky, T.; Tian, X.; Ko¨nigsberger, K.; Rouden, J. J. Org. Chem.
1994, 59, 4037.
(13) (a) Hudlicky, T.; Gonzales, D.; Gibson, D. T. Aldrichimica Acta
1999, 32, 35. (b) Zylstra, G. J.; Gibson, D. T. J. Bio. Chem. 1989, 264,
14940; Available from Aldrich: catalog# 489492-5G.; 5 g 262.50$Cdn.
(14) (a) Bragg, P. D.; Jones, J. K. N.; Turner, J. C. Can. J. Chem. 1959,
37, 1412. (b) Rinner, U. Dissertation, University of Florida, 2003. (c)
Lohray, B. B. Synthesis 1992, 1035.
5670
Org. Lett., Vol. 7, No. 25, 2005

Scheme 1
to produce 13 and 14 selectively, as the latter compounds
for a benzoate group led to compound 26, whose oxidation
under the same conditions gives the imide 27 in 9-10%
yield. Buffering this acidic reaction with Na2CO3 to a neutral
suspension increases the yield to 33%. Oxidation of the
aromatic product obtained from cyclotrimerization of scaffold
19 (“cis series”) with benzyltriethylammonium permanganate
yielded the corresponding imide in 53% yield, whereas the
oxidation of the C-1 epimeric scaffold 26 (“trans series”)
with this oxidizing agent produced the imide 27 in only 12-
14% yield.
will be used in intramolecular variants of the cyclotrimer-
ization to access directly the narciclasine and lycoricidine
analogues. The synthesis of scaffold 17 was completed by
desilylation of the acetylene moiety, protection of C-2 alcohol
(pancratistatin numbering), and alkylation of the tosylamide
with propargyl bromide providing the key intermediate for
the “natural” or “trans series” in seven operations (eight
steps) from aziridine 8. The cis-diol 11 was protected as a
dibenzoate and converted to 19 in two steps to serve as a
scaffold for the unnatural, or “cis series” of analogues
epimeric at C-1.
The literature reports on cyclotrimerization involving bis-
trimethylsilylacetylene (BTMSA) mention moderate yields
as well as desilylation problems and incorporation of CO
ligands into reaction products.15 The optimum conditions for
the formation of tetracycle 20 (83% yield) were found by
slow addition (syringe pump, 36 h) of a mixture of 17, CpCo-
(CO)2, and BTMSA to a heated solution of BTMSA in
xylene, Scheme 2. After distillation and recovery of BTMSA,
the product was isolated by chromatography on silica.
Similary, the “cis” scaffold 19 was converted to the core of
pancratistatin in 87% yield.16
To examine the most effective approach to either 24 or
27 we tested the cyclotrimerization of a propargylimide,
prepared in two steps from the “cis” scaffold 18 by
desilylation and acylation with propargylic anhydride. Under
identical conditions used in cyclotrimerizations leading to
20, the imide provided the aromatic product in only 5% yield,
perhaps because of unfavorable rotamer population derived
from the propargyl imide or possible imide resonance forms
(E/Z forms). For these reasons the sequence involving the
cyclotrimerization of propargylamines followed by benzylic
oxidation proved vastly superior. Following the reductive
removal of the tosyl group, methanolysis, and deprotection,
the core of pancratistatin containing the trimethylsilyl groups
was attained.
Oxidation of20 to imide 24 was accomplished in 10-15%
yield with NaIO4/RuCl3.17 An exchange of the TBS group
Some of the intermediates and the two fully hydroxylated
analogues, 23 and 28, were subjected to biological evaluation
in order to determine what effect the lack of aromatic
oxygenation and the lipophilic silyl groups have on the
potency of these analogues. The biological evaluation of this
series proved interesting. Of the compounds tested against
the cancer cell lines the only inactive one was the fully
(15) (a) Rainier, J. D.; Imbriglio, J. E. Org. Lett. 1999, 1, 2037. (b)
Rinner, U.; Hudlicky, T. Unpublished results.
(16) The details of the reaction sequence in the “cis series” (i.e., 19 f
28) will be reported in an upcoming full paper in due course. The overall
yield of fully hydroxylated derivative 23 via the “cis series” was 2.5-fold
more efficient because the problems of elimination to 14 were not
encountered in this series.
(17) Elango, S.; Yan, T. H. Tetrahedron 2002, 58, 7335.
Org. Lett., Vol. 7, No. 25, 2005
5671

Scheme 2
protected tosylamide 20. As the free alcohol content in-
matic ring. The oxidative transformation of the silyl groups
will be compared with Baeyer-Villiger oxidations of acetyl
groups in similar tetracycles derived by using 2,5-dihydroxy-
hex-3-yne, 1,4-dihydroxybut-2-yne, or hex-3-yne-2,5-dione.
Once the best combination of cyclotrimerization partners is
established, the diversity-oriented synthesis of carbo- and
heterocyclic analogues will be initiated. We will report the
results of these endeavors in due course.
creased (i.e., 21 and 22) so did the activity against cell lines,
with 22 and 23 having similar activity and being only 10-
fold less active than 7-deoxypancratistatin.18 This finding is
both surprising and exciting since none of the compounds
have the amide carbonyl thought to be essential to activity.
Tosylamide 23 is the first active compound that lacks all
oxygenation in the phenanthridone moiety, and this result,
if supported by further testing of similar compounds, may
lead to revision of the proposed mode of action.
The pancratistatin core, lacking aromatic oxygenation, has
been synthesized by a high-yielding cyclotrimerization
strategy in 11 steps from tosylaziridine 8 in both the natural
and the C-1 epimeric series in an overall yield of 5% and
12%, respectively. The sequences proceeded in reasonable
yields and the preparation of the key scaffolds, 17 and 19,
was successfully carried out on medium scale (2.5 g). The
“cis” isomer was also converted to the natural stereoisomer
25 by late-stage opening of the cyclic sulfate after cyclo-
trimerization of scaffold 19.16 What remains to be established
is the identification of the best precursors for the eventual
installation of catechol/phenol functionalities onto the aro-
Acknowledgment. The authors thank Professor John
Rainier (University of Utah) and Professor Peter Vollhardt
(University of California, Berkeley) for helpful discussions
and advice related to cyclotrimerization. The following
agencies are gratefully acknowledged for financial support
of this work: Natural Sciences and Engineering Research
Council of Canada (NSERC), Canada Foundation for In-
novation (CFI), Ontario Innovation Trust (OIT), Brock
University, TDC Research Foundation, Inc. (fellowship to
M.M.) and TDC Research, Inc.
Supporting Information Available: Experimental pro-
cedures and spectral data for key compounds 10-23. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(18) Details of the testing results will be reported in a full paper for all
compounds. We thank Professor George R. Pettit (Arizona State University)
for the biological evaluation of our natural product derivatives.
OL052372O
5672
Org. Lett., Vol. 7, No. 25, 2005