Studies directed towards the refinement of the pancratistatin cytotoxic pharmacophore

Article abstract of DOI:10.1016/S0960-894X(00)00614-4

Two deoxy-analogues of the anticancer/antiviral agent pancratistatin containing functionality complementary to the minimum structural pharmacophore were synthesized and subjected to anticancer screening. One of the analogues exhibited selective inhibition of certain tumor cell lines but was significantly less potent than the natural products. The minimum structural pharmacophore has now been refined from eight to three possible structures.

Full text of DOI:10.1016/S0960-894X(00)00614-4

Bioorganic & Medicinal Chemistry Letters 11 (2001) 169±172
Studies Directed Towards the Re®nement of the Pancratistatin
Cytotoxic Pharmacophore
James McNulty,a,Ã Justin Mao,a Romelo Gibe,a Ruowei Mo,a Sonja Wolf,a
George R. Pettit,b Delbert L. Heraldb and Michael R. Boydc
aInstitute of Molecular Catalysis, Department of Chemistry, Brock University, St. Catharines, Ontario, Canada, L2S 3A1
bCancer Research Institute and Department of Chemistry, Arizona State University, Tempe, AZ 85287-2404, USA
cLaboratory of Drug Discovery Research and Development, Division of Basic Sciences, National Cancer Institute,
Frederick Cancer Research and Development Center, Building 1052, Room 12, Frederick, MD 21702-1201, USA
Received 8 September 2000; accepted 30 October 2000
AbstractÐTwo deoxy-analogues of the anticancer/antiviral agent pancratistatin containing functionality complementary to the
minimum structural pharmacophore were synthesized and subjected to anticancer screening. One of the analogues exhibited selec-
tive inhibition of certain tumor cell lines but was signi®cantly less potent than the natural products. The minimum structural
pharmacophore has now been re®ned from eight to three possible structures. # 2001 Elsevier Science Ltd. All rights reserved.
The Amaryllidacae1 alkaloid pancratistatin 1, isolated in
small quantities from the Hawaiian Hymenocallis littor-
alis2 along with some of its closely related natural con-
geners including notably trans-dihydrolycoricidine 2
(Scheme 1), has demonstrated antiviral activity as well
as potent cytotoxicity against certain tumor cell lines.3,4
The promising pharmacological pro®le of the com-
pounds, in addition to the complex polyoxygenated
phenanthridone skeleton, has attracted considerable
synthetic interest, and pancratistatin has been the sub-
ject of several elegant total syntheses.5 Despite the
potent activity of 1 and 2, no systematic study has been
undertaken to de®ne minimum structural requirements
of the pharmacophore or to identify the biological tar-
get of these compounds. Structural similarity to the
related alkaloid narciclasine has led to the suggestion
that the compounds may be inhibitors of protein bio-
synthesis,3,4 although more de®nitive studies have been
hampered in part by the limited supply of these alkaloids.
cytotoxicity activity against the NCI panel of 60 human
tumor cell lines, indicating that the C-1 and C-7
hydroxyl substituents are not requirements of the phar-
macophore.6 In addition, the C-7 hydroxyl was not
required for antiviral3 activity. On the other hand, the
trans-B/C ring junction stereochemistry was found to be
Limited structure±activity correlations of the natural
compounds 1 and 2 and a few minor natural and semi-
synthetic analogues have de®ned some of the structural
requirements of the cytotoxic pharmacophore. Com-
pounds 1 and 2 shared a similar potency and pro®le of
ÃCorresponding author. Tel.: +1-905-688-5550; fax: +1-905-682-
9020; e-mail: jmcnulty@chemiris.labs.brocku.ca
Scheme 1.
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00614-4

170
J. McNulty et al. / Bioorg. Med. Chem. Lett. 11 (2001) 169±172
crucial for potent cytotoxicity in both the natural series6
and in a fully oxygenated synthetic derivative.7a Lastly,
oxygenated seco-analogues (B/C ring opened) of these
compounds were devoid of biological activities.7b These
studies indicate that the conformation of the C-ring, or
more speci®cally, the correct spatial orientation of one
or more of the hydroxyl substituents at positions C-2,
C-3, and C-4 are essential for recognition. At present,
compound 2 is the structurally simplest analogue
known that exhibits potent cytotoxicity. Based on this
information, eight possible structures could embody the
overall minimum pharmacophore; one triol (i.e., com-
pound 2), three diols (2a,3b; 2a,4b; or 3b,4b), three
monoalcohols (2a; 3b; or 4b) or the fully deoxygenated
cyclohexane derivative. As a means of systematically
evaluating the role of the three hydroxyl groups at C-2,
C-3, and C-4 in 2, we divided the functionality into pairs
of complementary analogues. We de®ne these as pairs of
analogues of the active compound having the same
structural skeleton with di€ering functionality but
between them covering all of the functional groups on
the minimum known pharmacophore. Compounds 3
and 4 may therefore be considered complementary ana-
logues to the presently known minimum pharmaco-
phore in 2. We previously reported the synthesis of
compound 48 using
a
stereoselective nitroaldol
approach. We now report the preparation of the com-
plementary analogue 3, utilizing a Diels±Alder approach
to the core structure, as well as the cytotoxic screening
of both of these analogues.
The most direct synthetic route to diol 3 is outlined in
Scheme 2. Diels±Alder reaction of nitroalkene 59 with
butadiene using a modi®cation of the method reported
by Bryce10 provided the cyclohexene derivative 6 and
allowed us to secure the necessary future trans-B/C ring
junction early in the synthesis. In our hands, the reac-
tion proceeded faster and gave cleaner product under
Lewis acid catalysis (ZnCl2) in a sealed bomb. Chemo-
selective reduction of the nitro group was best achieved
Scheme 2. Reagents and conditions: (i) Butadiene sulfone, PhMe, 127 ꢀC, ZnCl2, 12 h, 85%; (ii) Al±Hg, THF/H2O, 22 ꢀC, 12 h, then (iii) ClCO2Me
(1.5 equiv), CH2Cl2, Et3N, 0 ꢀC, 5 h, 96% from 6; (iv) MCPBA (3.0 equiv), CH2Cl2 0 ꢀC, 5 h, 93%; (v) PhCO2Na, H2O, 100 ꢀC, 12 h, then (vi) Ac2O/
Py, 0 ꢀC, 12 h, 55±62% from 9; (vii) Tf2O (5.0 equiv), DMAP (3.0 equiv), CH2Cl2, 0±15 ꢀC, 15 min; then 2 M HCl (aq), dioxane, 20 ꢀC, 18 h; then
Ac2O/Py, 0 ꢀC, 6 h, 85%; (viii) NaOMe (2.1 equiv), THF:MeOH (1:3), 0±18 ꢀC, 12 h, 96%.
Scheme 3. X-ray structure of 10 (R1=0.0855).

J. McNulty et al. / Bioorg. Med. Chem. Lett. 11 (2001) 169±172
171
using aluminium amalgam in THF, with the resulting
amine being immediately protected as the methoxy-
carbonyl derivative giving cyclohexene 7 in 96% overall
yield from 6.11 Epoxidation with MCPBA gave the a-
and b-epoxides 8 and 9 (ratio 2.7:1) in 93% yield. The
epoxides were readily separable on ¯ash silica and
independently subjected to a surprisingly stereospeci®c
ring opening using the method reported by Magnus and
co-workers.5c Reaction of the a-epoxide 8 with re¯uxing
aqueous sodium benzoate followed by treatment with
acetic anhydride in pyridine provided the diacetoxy deri-
vative 10 in 55% yield. Similar opening of the b-epoxide 9
and likewise protection as the diacetate gave the same
diol derivative 10 in 62% yield. Under these conditions,
the conformationally biased diastereomeric epoxides are
opened exclusively by axial attack. Thus in 8, opening
occurs from the axial direction by attack at C-3 while
the diastereomeric epoxide 9 is opened by axial attack
at C-2 yielding the same 2,3-diaxial diol. The relative
stereochemistry in product 10 was con®rmed by a single
crystal X-ray structural determination (Scheme 3). The
crystal structure clearly shows the chair conformation
of the cyclohexane ring of 10 locked by the equatorial
aryl and methoxycarbonylamino substituents as well as
the diaxial-2,3-diacetoxy substituents.12 All four sub-
stituents on the cyclohexane ring are therefore set up
with the correct relative stereochemistry desired in ana-
logue 3. Cyclization of 10 to give the phenanthridone
skeleton was e€ected in low yield (5±15%) using the
method of Banwell.13 Under these conditions two inter-
mediate products were also isolated in addition to the
desired product 13. These were readily identi®ed as 11
and 12 in comparison with the results of the Magnus
work.5c Further investigation showed that the cycliza-
tion reaction with Tf2O and DMAP was extremely
rapid, complete in 15 min at 0 ꢀC, giving 12 which
slowly converts to 13 on standing. Quenching the reac-
tion after 15 min and acidic hydrolysis of the mixture
led also to partial loss of the acetate groups requiring a
®nal re-acetylation step. The cyclization process is very
sensitive to traces of water, however under the opti-
mized cyclization/hydrolysis/reprotection protocol a
74±90% yield of the cyclized protected diol 13 could be
realized with no detectable trace of the side products.
Removal of the acetate groups completed the synthesis
of the desired diol 3.
mL). Overall, compound 3 was not suciently potent to
generate a useful mean-graph for COMPARE analysis14
(Scheme 3)
The present study helps further de®ne the minimum
cytotoxic pharmacophore shared by pancratistatin 1
and trans-dihydrolycoricidine 2. In addition to the
necessary conformationally locked trans-fused B/C ring
junction, at least two correctly positioned hydroxyl
groups appear to be needed in the C-ring for potent
cytotoxic activity. The results with the complementary
analogues 3 and 4 rule out any monoalcohol analogue
as a minimum pharmacophore and the modest potency
of 3 suggests the importance of either, or both, of the
C-2 and C-3 hydroxyl groups, in addition to the C-4
hydroxyl as a minimum pharmacophore. We tentatively
conclude that the minimum structural pharmacophore
must reside in either of the diols 14 or 15 (Scheme 3) or
may in fact be that depicted in the trihydroxy containing
natural product 2. The synthesis of analogues 2, 14 and
15 is currently under investigation in our laboratories.
Acknowledgements
Financial support from the Natural Sciences and Engi-
neering Research Council of Canada, Research Cor-
poration (Cottrell Scholar Science Award to J.McN)
and the Canada Foundation for Innovation are grate-
fully acknowledged. One of us (G.R.P.) wishes to thank
the Arizona Disease Control Research Commission
and Outstanding Investigator Grant CA-44344-10-12
awarded by the DCTD, National Cancer Institute,
DHHS. We thank Mr. Tim Jones for conducting MS
and 2D-NMR analyses.
References and Notes
1. Lewis, J. R. Nat. Prod. Rep. 2000, 17, 57 and prior reports
in this series.
2. Pettit, G. R.; Gaddamidi, V.; Cragg, G. M.; Herald, D. L.;
Sagawa, Y. Chem. Commun. 1984, 1693.
3. Gabrielsen, B.; Monath, T. P.; Huggins, J. W.; Kefauver,
D. F.; Pettit, G. R.; Groszek, G.; Hollinshead, 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.
4. Mondon, A.; Krohn, K. Chem. Ber. 1975, 108, 445.
5. (a) Rigby, J. H.; Maharoof, U. S. M.; Mateo, M. E. J. Am.
Chem. Soc. 2000, 122, 6624. (b) Keck, G. E.; Wager, T. T.;
Rodriquez, J. F. D. J. Am. Chem. Soc. 1999, 121, 5176.
(c) Magnus, P.; Sebhat, I. K. Tetrahedron 1998, 54, 15509.
(d) Doyle, T. J.; Hendrix, M.; VanDerveer, D.; Javanmard, S.;
Haseltine, J. Tetrahedron 1997, 53, 11153. (e) Schultz, A. G.;
Holoboski, M. A.; Smyth, M. S. J. Am. Chem. Soc. 1996, 118,
6210. (f) Tian, X.; Hudlicky, T.; Konigsberger, K. J. Am.
Chem. Soc. 1995, 117, 3643. (g) Trost, B. M.; Pulley, S. R. J.
Am. Chem. Soc. 1995, 117, 10143. (h) Park, T. K.; Dani-
shefsky, S. J. Tetrahedron Lett. 1995, 36, 195. (i) Banwell,
M. G.; Cowden, C. J.; Gable, R. W. J. Chem. Soc., Perkin
Trans. 1 1994, 3515. (j) Lopes, R. S. C.; Lopes, C. C.; Heath-
cock, C. H. Tetrahedron Lett. 1992, 33, 6775.
Initial screening of complementary analogues 3 and 4
against the P-388 mouse leukemia cell line gave inter-
esting results. The mono alcohol derivative 4 displayed
marginal inhibition (ED50=40.1 mg/mL) while the diol 3
proved to be relatively potent (ED50=0.45 mg/mL),
although less potent than pancratistatin itself. Further
comparative assaying of compound 3 against the NCI
60 panel human tumor cell-line assay was undertaken
simultaneously with authentic pancratistatin. The
results con®rmed that 3 was in fact overall two to three
magnitudes less potent than pancratistatin although
selective inhibition of some of the cell lines was indi-
cated. Compound 3 inhibited proliferation of the non-
small cell lung cancer line NCI-H226 (ED50=0.65 mg/mL),
as well as two of the leukemia cell lines; CCRF-CEM
(ED50=0.55 mg/mL) and HL-60(TB) (ED50=0.89 mg/
6. Pettit, G. R.; Pettit, G. R., III; Backhaus, R. A.; Boyd,
M. R.; Meerow, A. W. J. Nat. Prod. 1993, 56, 1682.
7. (a) Pettit, G. R.; Melody, N.; O'Sullivan, M.; Thompson,
M. A.; Herald, D. L.; Coates, B. Chem. Commun. 1994, 2725.

172
J. McNulty et al. / Bioorg. Med. Chem. Lett. 11 (2001) 169±172
(b) Chretien, F.; Ahmed, S. I.; Masion, A.; Chapleur, Y. Tet-
rahedron 1993, 49, 7463.
8. McNulty, J.; Mo, R. Chem. Commun. 1998, 933.
9. McNulty, J.; Steere, J. A.; Wolf, S. Tetrahedron Lett. 1998,
39, 8013.
10. Bryce, M. R.; Gardiner, J. M.; Hursthouse, M. B.; Short,
R. L. Tetrahedron Lett. 1987, 28, 577.
11. The use of Al±Hg was operationally simpler and provided
consistently higher yields than the previously reported method
using zinc in acetic acid, see: Hendrickson, J. B.; Alder, R. W.;
Dalton, D. R.; Hey, D. G. J. Org. Chem. 1969, 34, 2667.
12. Crystallographic data have been deposited with the Cam-
bridge Crystallographic Data Centre, 12 Union Road, Cam-
bridge CB2 1EZ, UK (e-mail: deposit@ccdc.cam.ac.uk).
13. Banwell, M. G.; Bissett, B. D.; Busato, S.; Cowden, C. J.;
Hockless, D. C. R.; Holman, J. W.; Read, R. W.; Wu, A. W.
Chem. Commun. 1995, 2551.
14. Boyd, M. R.; Paull, K. D. Drug Develop. Res. 1995, 34, 91.