Total synthesis of psorospermin

Article abstract of DOI:10.1016/j.tetlet.2004.12.006

The xanthone natural product psorospermin was synthesized in 13 steps with an overall yield of 1.7%. This compound shows potent antineoplastic activity in a variety of cancer cell lines.

Full text of DOI:10.1016/j.tetlet.2004.12.006

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 827–829
Total synthesis of psorospermin
Michael K. Schwaebe,* Terence J. Moran and Jeffrey P. Whitten
Department of Medicinal Chemistry, Cylene Pharmaceuticals Inc., 11045 Roselle St. Suite C, San Diego, CA 92121, USA
Received 14 September 2004; revised 24 November 2004; accepted 1 December 2004
Available online 16 December 2004
Abstract—The xanthone natural product psorospermin was synthesized in 13 steps with an overall yield of 1.7%. This compound
shows potent antineoplastic activity in a variety of cancer cell lines.
ꢀ 2004 Elsevier Ltd. All rights reserved.
Psorospermin 1 is a xanthone natural product first iso-
lated by Kupchan et al. from the bark of the African
plant Psorospermum febrifugum.1 The absolute stereo-
chemistry was later determined by Cassady and co-
workers2. Psorospermin has been of great interest for
the past three decades because of itÕs unique structure
and novel antineoplastic properties.3 Although several
publications report efforts toward the total synthesis of
both chiral and non-chiral analogs of psorospermin, to
date no total synthesis has been achieved.4 We report
the total synthesis of psorospermin, a novel fused tetra-
cyclic xanthone containing two chiral centers and a reac-
tive epoxide.
O
O
O
O
O
O
1
OBn
OBn
OBn
OBn
5
4
3
O
O
OH
The natural product, psorospermin, contains the
(20R,30R)-stereochemsitry and is set by the stereochemis-
try of the olefination and the chirality of the epoxida-
tion. This necessitated the selective formation of the
Z-olefin 3 and its reduction to the allylic alcohol 4 in
the presence of a variety of sensitive groups. Addition-
ally, we needed to differentiate the 1- and 5-hydroxys
(the 5-hydroxy being more reactive and the 1-hydroxy
being more labile) in a manner that enabled selec-
tive methylation, protection and deprotection to form
the final compound. The synthesis is outlined in
Scheme 1.
O
O
O
O
O
O
O
OH
H
H
2'
3'
2
1
O
O
The synthesis begins with the construction of the xan-
thone skeleton by the method of Grover et al.6 Contrary
to much of the xanthone literature we found it unneces-
sary to fuse the zinc chloride prior to the coupling. We
found this extra step actually decreases reaction yield
because of the insolubility of glass-like fused zinc chlo-
ride. To circumvent this step we heated the zinc chloride
in POCl3 to 60 ꢁC for 30 min prior to the addition of
dimethoxy benzoic acid 5. We continued to heat for
an additional 30 min then added phloroglucinol 6.
As the two stereocenters are adjacent to each other and
one center is highly hindered, we examined a similar
route to Cassady and co-workers that utilized a zip-
per-type reaction as in the synthesis of compound 25.
Our main objective was to develop a scaleable route to
provide enough material for preclinical evaluation as a
cancer therapeutic.
Keywords: Psorospermin; Asymmetric synthesis; Anticancer agent;
Xanthone; Natural product.
Selective allylation of the 3-hydroxyl followed by
methylation of the 1-hydoxyl proceeded smoothly to
afford xanthones 8 and 9 in 77% and 95% yields,
*
Corresponding author. Tel.: +1 8588661605; fax: +1 8588661609;
e-mail: mschwaebe@cylenepharma.com
0040-4039/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.12.006

828
M. K. Schwaebe et al. / Tetrahedron Letters 46 (2005) 827–829
O
O
OH
OH
OH
b
+
a
O
77%
O
OH
76%
HO
OH
O
O
5
6
7
O
O
O
O
O
O
O
OH
OH
c
e
d
O
O
OH
95%
91%
42-66%
O
O
O
9
O
10
8
O
O
OH
O
O
O
g
f
h
O
OH
O
OBn
OBn
O
OBn
81%
85%
44%
OH
OBn
OBn
13
12
11
O
O
O
O
O
O
O
i
j
k
OBn
O
OBn
H
OBn
76%
78%
42%
OBn
OBn
3
O
4
14
HO
O
O
O
O
O
O
O
O
O
O
O
m
l
OBn
O
OBn
O
O
H
70%
95%
OBn
OBn
OH
1
16
15
HO
MsO
O
Scheme 1. Reagents: (a) ZnCl2, POCl3; (b) allyl bromide, K2CO3; (c) CH3I, K2CO3; (d) mesitylene, 180 ꢁC; (e) BBr3; (f) benzyl bromide, NaH, DMF;
(g) CH3I, DMF; (h) OsO4, NaIO4; (i) (CF3CH2O)2POCHCH3CO2Me, KHMDS, 18-crown-6; (j) DIBALH/CH2Cl2; (k) t-BUOOH, (ꢀ) DIPT,
Ti(i-Opr)4; (l) MsCl, Et3N; (m) raney nickel/K2CO3, ethanol.
respectively. Claisen rearrangement was affected in re-
fluxing mesitylene and the product 10 crystallized out
of the reaction mixture upon cooling to room tempera-
ture. The use of more polar solvents decreased the ratio
of the rearrangement occurring at the 4-position over
the 2-position. The addition of even a catalytic amount
of a weak base (i.e., potassium carbonate or triethyl
amine) results in a significant amount of cyclized mate-
rial that was identified as the dihydrofuran.
with potassium carbonate led to significant amounts of
the cyclized dihydrofuran side-product.
As reported in the literature, the choice of solvent for
the Johnson–Lemieux oxidation of 13 is critical.7 Treat-
ment of 13 with catalytic osmium tetroxide and sodium
periodate in a 3:1 mixture of dioxane/water afforded the
aldehyde 14 in a 44% yield after recrystallization from
THF. At elevated temperatures above 40 ꢁC, over-oxi-
dation to the acid became a significant side product.
Formation of the Z-olefin was efficiently carried out
by the method of Still and Gennari.8 While many other
reagents were examined, this method provided a very
good yield and Z/E ratio of products. We found temper-
ature was critical and, although we could add the alde-
hyde 14 at ꢀ85 ꢁC, its solubility demanded the use of
copious amounts of THF to dissolve the aldehyde. A
better solution was to add the aldehyde 14 as a dry pow-
der all at once at ꢀ78 ꢁC. The reaction proceeded
smoothly as the solubility of 14 slowly increased without
any local exotherm. The result was a consistent, high-
The key to the sequence was whether we could efficiently
perform the di-demethylation and subsequent selective
3,5-bis benzylation. Treatment of 10 with boron tribro-
mide at 0 ꢁC cleanly afforded the trihydroxy xanthone
11 in a 91% yield after treatment with ice water and fil-
tration. The bis-benzylation proceeded efficiently to
afford the xanthone 12 with only 10% of the
tribenzylated material being formed. Methylation of
the 1-hydroxyl with methyl iodide proceeded at room
temperature with the addition of sodium hydride in
DMF in an 85% yield. The use of elevated temperatures

M. K. Schwaebe et al. / Tetrahedron Letters 46 (2005) 827–829
829
yielding formation of the olefin in a 10:1 ratio of Z/E.
References and notes
Pure Z-olefin 3 could be obtained by crystallization
from ethyl acetate to afford the Z-a,b-unsaturated ester
3 in a 76% yield.
1. Kupchan, S. M.; Streelman, D. R.; Sneden, A. T. J. Nat.
Prod. 1980, 43, 296.
2. Habib, A. M.; Ho, D. K.; Masuda, T. M.; Reddy, M. A.;
McKenzie, A.; Byrn, S. R.; Chang, C. J.; Cassady, J. M. J.
Org. Chem. 1987, 52, 412.
3. (a) Hansen, M.; Lee, S. J.; Cassady, J. M.; Hurley, L. H. J.
Am. Chem. Soc. 1996, 118, 5553; (b) Kwok, Y.; Hurley, L.
H. J. Biol. Chem. 1988, 273, 33020; (c) Kwok, Y.; Zeng,
Q.; Hurley, L. H. Proc. Natl. Acad. Sci. 1998, 95, 13531.
4. (a) Ho, D. K.; McKenzie, A. T.; Byrn, S. R.; Cassady, J.
M. J. Org. Chem. 1987, 52, 342; (b) Ko, O. H. Yakhak
Hoeji 1997, 41, 294.
5. Reddy, K. S.; Ko, O. H.; Ho, D.; Persons, P. E.; Cassady,
J. M. Tetrahedron Lett. 1987, 28, 3075.
6. (a) Grover, P. K.; Shah, G. D.; Shah, R. D. J. Chem. Soc.
1955, 3982; (b) Grover, P. K.; Shah, G. D.; Shah, R. D. J.
Sci. Ind. Res. India 1956, 15B, 629.
7. Papps, R.; Allen, D. S., Jr.; Lemieux, R. U.; Johnson,
W. S. J. Org. Chem. 1956, 21, 478.
8. Still, W. C.; Gennari, C. Tetrahedron Lett. 1983, 24, 4405.
9. Katsuki, T.; Sharpless, K. B. J. Am. Chem. Soc. 1980, 102,
5974.
The 1,2-reduction of ester 3 was extremely troublesome.
After examining a host of reagents and conditions we
could only affect the transformation by careful addi-
tion of 2.0 equiv of DiBALH in methylene chloride at
ꢀ78 ꢁC. Even then, it was necessary to separate the
product 4 from the residual starting ester 3 and the
over-reduced material via silica gel chromatography.
This procedure, although less than optimal, afforded
the Z-allylic alcohol in a 42% yield. Sharpless epoxida-
tion9 proceeded smoothly with stoichiometric addi-
tion of reagents and powdered sieves. The enantio-
selectivity of the reaction was determined by chiral LC
of the 5-methoxy derivative of the final product 1.10
Thus, we had our epoxide 15 in a 78% yield with a
70% ee. Formation of the mesylate proceeded rou-
tinely to afford the psorospermin precursor 16 in a
95% yield.
10. Compound 1 was converted to the 5-methoxy compound
for chiral LCMS by treatment of 1 with methyl iodide in
the presence of potassium carbonate in acetone. The
mixture was evaluated by normal phase chiral chroma-
tography using the Waters 600 pump controller, 2687 dual
wave detector and ZQ mass spectrometer. The mobile
phase was an isocratic mixture of hexane/isopropanol/
1,2-dichloroethane; 30:10:10 at a constant flow of 6.0 mL/
min. The stationary phase was a Chirex (S)-VAL and
DNAn 250 · 10.0 mm column by phenomenex. The tem-
perature was kept constant at 30 ꢁC with a Waters 600
column heater.
The final step involved the selective reduction of the 3-
and 5-benzyl protecting groups and a zipper-type cycli-
zation.11 The key to affecting this transformation was
to find a way to reductively cleave both benzyl groups
without reduction of the resulting epoxide. When the
reduction was carried out with 10% Pd/C in the absence
of base, the reaction stalled with the product being ab-
sorbed onto the catalyst. In the presence of potassium
carbonate, over reduction predominated. An improve-
ment involved the use of 1,4-cyclohexadiene with Pd/
Ba2SO4 in refluxing methanol. Finally, the use of stoi-
chiometric Raney Nickel with potassium carbonate in
a 1:1 mixture of ethyl acetate and ethanol at 60 ꢁC affor-
ded psorospermin in a 70% yield after silica gel chroma-
tography. The 1H NMR spectrum of 1 was in agreement
with an authentic sample of psorospermin and showed
the characteristic doublets of the 40 protons at 2.73
and 2.99 ppm (Dd 0.258).12,13
11. Dolle, R. E.; Nicolaou, K. C. J. Am. Chem. Soc. 1985, 107,
1691.
12. Spectral data for compound 1. 1H NMR (500 MHz,
CDCl3) d 7.82 (1H, dd, J = 7.3, 1.8 Hz), 7.25 (1H,
dd, J = 8.9, 1.8 Hz), 7.20 (1H, t, J = 8.3 Hz), 6.37
(1H, s), 5.61 (1H, s), 4.92 (1H, dd, J = 9.4, 7.1 Hz), 3.97
(3H, s), 3.50 (1H, dd, J = 15, 10 Hz), 3.31 (1H, dd,
J = 15.3, 7.2 Hz), 2.99 (1H, d, J = 4.5 Hz), 2.73 (1H, d,
J = 5 Hz), 1.44 (3H, s); 13C NMR (125 MHz, CDCl3) d
175.8, 165.9, 163.3, 154.1, 144.8, 143.8, 123.9, 119.6, 117.4,
106.8, 104.2, 90.1, 87.0, 58.0, 56.6, 51.2, 49.7, 28.9, 16.8;
LCMS (254 nm): Rt = 3.41 min (100%); EIMS(pos) m/z
341.1 [M+H]+ (100), 363.1 [M+23]+ (17), 271.1
[M ꢀ 70+1]+ (11).
Supplementary data
NMR and LCMS supplementary data associated with
this article can be found, in the online version, at
doi:10.1016/j.tetlet.2004.12.006.
13. An authentic sample of 1 was obtained from the NCI via
Laurence Hurley.