A modular synthesis of annonaceous acetogenins

Article abstract of DOI:10.1021/jo026433x

A synthesis of four Annonaceous acetogenins, asiminocin, asimicin, asimin, and bullanin, by a modular approach from seven fundamental subunits, A-G, is described. The approach employs a central core aldehyde segment, C, to which are appended an aliphatic terminus, A or B, a spacer subunit, D or E, and a butenolide terminus, F or G. Coupling of the A, B, D, and E segments to the core aldehyde unit is effected by highly diastereoselective additions of enantiopure allylic indium or tin reagents. The butenolide termini are attached to the ACD, BCE, or BCD intermediates by means of a Sonogashira coupling. The design of the core, spacer, and termini subunits is such that any of the C30, C10, or C4 natural acetogenins or stereoisomers thereof could be prepared. IC₅₀ values for the four aforementioned acetogenins against H-116 human colon cancer cells were found to be in the 10^-3 to 10^-4 μM range. The IC₉₀ activities were ca. 10^-3 μM for asimicin and asimin but only 0.1^-1 μM for bullanin and asiminocin.

Full text of DOI:10.1021/jo026433x

A Mod u la r Syn th esis of An n on a ceou s Acetogen in s
J ames A. Marshall,*,† Arnaud Piettre,† Mikell A. Paige,† and Frederick Valeriote‡
Department of Chemistry, University of Virginia, Charlottesville, Virginia 22904,
and The J osephine Ford Cancer Center, Detroit, Michigan 48202
jam5x@virginia.edu
Received September 10, 2002
A synthesis of four Annonaceous acetogenins, asiminocin, asimicin, asimin, and bullanin, by a
modular approach from seven fundamental subunits, A-G, is described. The approach employs a
central core aldehyde segment, C, to which are appended an aliphatic terminus, A or B, a spacer
subunit, D or E, and a butenolide terminus, F or G. Coupling of the A, B, D, and E segments to
the core aldehyde unit is effected by highly diastereoselective additions of enantiopure allylic indium
or tin reagents. The butenolide termini are attached to the ACD, BCE, or BCD intermediates by
means of a Sonogashira coupling. The design of the core, spacer, and termini subunits is such that
any of the C30, C10, or C4 natural acetogenins or stereoisomers thereof could be prepared. IC50
values for the four aforementioned acetogenins against H-116 human colon cancer cells were found
to be in the 10-3 to 10-4 µM range. The IC90 activities were ca. 10-3 µM for asimicin and asimin but
only 0.1-1 µM for bullanin and asiminocin.
In tr od u ction
Genera of the plant family Annonaceae have proven
to be rich sources of a unique family of acetogenins with
a wide range of bioactivities.1 The leaves, bark, and seeds
of the edible fruits of these plants have been used in folk
medicine as insecticides, fungicides, antiparasitics, and
a remedy for snake bite. More recently a significant
number of the nearly 400 members of this family have
been shown to exhibit high levels of cytotoxicity against
human tumor cell lines, including those that exhibit
multiple drug resistance (MDR).2 Interestingly, normal
epithelial and bone marrow cells are significantly less
affected than tumor cells. Although the acetogenins are
found in relatively abundant plants, they are present in
only minute amounts as complex mixtures of structurally
similar isomers. Thus isolation of a pure component
presents a major challenge. A typical protocol yields only
10-20 mg of material from 15 kg of stem bark and then
only after multiple separation steps involving partition
extraction and chromatography on several different
columns.3 Final purification often requires preparative
HPLC or radial chromatography. Furthermore, the prod-
F IGURE 1. Structural features of adjacent bistetrahydrofu-
ran Annonaceous acetogenins with high antitumor activity.
ucts isolated are waxes or gums, thus precluding struc-
ture determination through X-ray diffraction.
Str u ctu r es of An n on a ceou s Acetogen in s
The Annonaceous acetogenins contain an unbranched
chain of 30-32 carbon atoms attached to the 3-position
of (S)-5-methyl-(5H)-2-furanone. This chain typically
contains 3-5 or more oxygen substituents with at least
one of these located near the center embedded in a
tetrahydrofuran moiety. Some acetogenins possess ke-
tonic groups or (Z) double bonds in the chain, reflecting
their presumed fatty acid origins. Those with the highest
antitumor activity feature a core structure of two tet-
rahydrofuran rings connected at the R-position with
flanking CHOH substituents at the R′-positions and a
third OH substituent, often located at C30, C29, C28,
C10, or C4 (Figure 1).1 Structures with one, two, or three
tetrahydrofuran rings are commonly found. These rings
need not be adjacent. Core structures containing a
tetrahydropyran ring are also known.1 To date, the great
majority of acetogenins with an adjacent bistetrahydro-
furan core unit have been shown to have the (R) config-
uration at C23.
† University of Virginia.
‡ The J osephine Ford Cancer Center.
(1) (a) Gu, Z.-M.; Zhao, G.-X.; Oberlies, N. H.; Zeng, L.; McLaughlin,
J . L. In Recent Advances in Phytochemistry; Arnason, J . T., Mata, R.,
Romeo, J . T., Eds.; Plenum Press: New York, 1995; Vol. 29, pp 249-
310. (b) Zeng, L.; Ye, Q.; Oberlies, N. H.; Shi, G.; Gu, Z.-M.; He, K.;
McLaughlin, J . L. Nat. Prod. Rep. 1996, 275. (c) Cave´, A.; Figade´re,
B.; Laurens, A.; Cortes, D. In Progress in the Chemistry of Natural
Products; Herz, W., Kirby, G. W., Moore, R. E., Stegilch, W., Tamm,
C., Eds.; Springer-Verelag Chemistry: New York, 1997; Vol. 70, pp
81-288.
(2) (a) Oberlies, N. H.; Croy, V. L.; Harrison, M. L.; McLaughlin, J .
L. Cancer Lett. 1997, 115, 73. (b) Oberlies, N. H.; Chang, C.-J .;
McLaughlin, J . L. J . Med. Chem. 1997, 40, 2102.
(3) Zhao, G.-X.; Miesbauer, L. R.; Smith, D. L.; McLaughlin, J . L.
J . Med. Chem. 1994, 37, 1971.
It should be noted that despite the extreme cytotoxicity
reported for the aforementioned acetogenins (IC50’s of
10-12 µg/mL in cell culture assays against human tumor
10.1021/jo026433x CCC: $25.00 © 2003 American Chemical Society
Published on Web 02/08/2003
J . Org. Chem. 2003, 68, 1771-1779
1771

Marshall et al.
cell lines), studies on bullatacin2, asimin, and several
analogues4 indicate that they exhibit only modest growth
inhibition of normal cells. Hence the differential cyto-
toxicity is quite high. Preliminary investigations have
shown that acetogenins selectively inhibit ATP synthesis
in tumor cells but not in noncancerous cells. This mode
of action is unique and offers a previously unexplored
mechanism for cancer chemotherapy.
Ba ck gr ou n d
Numerous publications on the synthesis of Annona-
ceous acetogenins and their possible precursors have
appeared in recent years.5-8 The sequences developed in
those investigations provided adequate amounts of the
final products for structure confirmation, but few, if any,
could be judged both efficient and versatile. Motivated
by the reported high levels of activity against human
tumor cell lines and low toxicity toward normal cells, we
initiated a program to develop a modular synthetic
approach that could be used to prepare the most active
naturally occurring structural types and analogues thereof
for structure-activity evaluation.
F IGURE 2. Core tetrahydrofuran assembly.
hydroxyl substituents at C30, C10, and C4 as a test of
concept. Natural acetogenins with these features have
been found to strongly inhibit the growth of human colon,
breast, and lung cancer cells.3 The high convergency and
modular nature of the synthetic plan would provide more
efficient access to these bioactive compounds, and non-
natural analogues thereof, than previous more linear
routes. Following the attainment of these initial synthetic
goals we would be well-positioned to interchange the
various modules to produce unnatural acetogenins with
the potential for improved drug properties.
The central element of our plan entails the synthesis
of four major segments consisting of two pairs of allylic
stannanes, equivalent to the aliphatic termini A and B,
and the spacer units D and E (Figure 3). These would
be joined sequentially to the dialdehyde precursor C by
the aforementioned allylation methodology to assemble
stereochemically homogeneous core precursors, which
would then be cyclized to the bistetrahydrofuran core
units. Sonogashira coupling9 of the butenolide termini F
or G would provide the fully elaborated acetogenin
structures. The syntheses would be completed by hydro-
genation of the side chain multiple bonds and deprotec-
tion of the OH functions.
Syn th etic P la n
Our plan was based on previous success in constructing
the bistetrahydrofuran core structure through stereose-
lective additions of enantiomerically enriched γ-alkoxy
allylic stannanes, or the related allylic indium halide
analogues, to enantioenriched γ-silyloxy aldehydes, fol-
lowed by tetrahydrofuran ring closure of sulfonic ester
derivatives of the adducts (Figure 2).5,6 The allylation
methodology provides a versatile route to either enanti-
omer of the syn or anti diastereomeric adduct. Addition-
ally it is possible to prepare enantiomeric and diastere-
omeric γ-silyloxy aldehyde substrates, thereby allowing
reasonably direct access to any of the 64 possible bistet-
rahydrofuran core isomers. Of course, the efficiency of
this approach would expectedly show some dependence
on stereochemistry such that all isomers might not be
equally accessible.
The initial goals of the present investigation were to
develop a modular route to representative adjacent
bistetrahydrofuran Annonaceous acetogenins with the
threo, trans, threo, trans, threo core stereochemistry and
We selected asiminocin, asimin, and asimicin as the
initial synthetic targets. These acetogenins have been
reported to exhibit inhibitory activities of ∼10-12 µg/mL
against HT-29 human colon cancer cell lines.3
(4) Valeriote, F. Unpublished results.
(5) (a) Marshall, J . A.; Hinkle, K. W. J . Org. Chem. 1996, 61, 4247.
(b) Marshall, J . A.; Hinkle, K. W. J . Org. Chem. 1997, 62, 5989. (c)
Marshall, J . A.; Chen, M. J . Org. Chem. 1997, 62, 5996.
(6) (a) Marshall, J . A.; Hinkle, K. W. Tetrahedron Lett. 1998, 39,
1303. (b) Marshall, J . A.; J iang, H. Tetrahedron Lett. 1998, 39, 1493.
(c) Marshall, J . A.; J iang, H. J . Org. Chem. 1998, 63, 7066. (d) Marshall,
J . A.; J iang, H. J . Org. Chem. 1999, 64, 971. (e) Marshall, J . A.; J iang,
H. J . Nat. Prod. 1999, 62, 1123.
Segm en t Syn th esis
Our synthesis of the aliphatic terminus A of asiminocin
is depicted in Scheme 1. The known ester aldehyde 210
was treated with dibutylzinc in the presence of the
bissulfonamide catalyst 3 by the Knochel protocol11 to
afford the alcohol 4 as a 95:5 mixture (Mosher ester
analysis)12 of enantiomers in high yield. Sequential
reduction and oxidation of the DPS protected derivative
5 yielded the enal 7, which was subjected to a stannation-
(7) Reviews: (a) Figade´re, B. Acc. Chem. Res. 1995, 28, 359. (b)
Hoppe R.; Scharf, H.-D. Synthesis 1995, 1447. (c) Marshall, J . A.;
Hinkle, K. W.; Hagedorn, C. E. Isr. J . Chem. 1997, 37, 97. (d) Casiraghi,
G.; Zanardi, F.; Battistina, L.; Rassu, G.; Appendino, G. Chemtracts:
Org. Chem. 1998, 11, 803.
(8) (a) Hoye, T. R.; Hanson, P. R.; Kovelesky, A. C.; Ocain, T. D.;
Zhuang, Z. J . Am. Chem. Soc. 1991, 113, 9369. (b) Naito, H.; Kawahara,
E.; Maruta, K.; Maeda, M.; Sasaki, S. J . Org. Chem. 1995, 60, 4419.
(c). Sinha, S.; Sinha, A.; Yazbak, A.; Keinan, E. J . Org. Chem. 1996,
61, 7640. (d) Hoye, T. R.; Ye, Z. J . Am. Chem. Soc. 1996, 118, 1801. (e)
Sinha, S. C.; Sinha, A.; Keinan, E. J . Am. Chem. Soc. 1997, 119, 12014
and references therein. (f) Emde, U.; Koert, U. Tetrahedron Lett. 1999,
40, 5979. (g) Avedissian, H.; Sinha, S. C.; Yazbak, A.; Sinha, A.; Neogi,
P.; Sinha, S. C.; Keinan, E. J . Org. Chem. 2000, 65, 6035.
(9) Sonogashira, K.; Tohda, Y.; Hagihara, N. Tetrahedron Lett. 1975,
4467.
1772 J . Org. Chem., Vol. 68, No. 5, 2003

Modular Synthesis of Annonaceous Acetogenins
F IGURE 3. A modular synthesis of bistetrahydrofuran Annonaceous acetogenins.
SCHEME 1 a
SCHEME 3 a
a
(a) Swern; (b) Ph3PdCHCO2Et (67%); (c) DIBAL-H (83%); (d)
MnO2; (e) (1) Bu3SnLi, (2) ADD, (3) (S)-BINOL, EtOH, LAH, (4)
MOMCI, i-Pr2NEt.
The second spacer unit E (Scheme 4) was assembled
from the known diepoxy diol 22,15 prepared by a modified
procedure in four steps from 1,5-cyclooctadiene. Treat-
ment of the derived chloride 23 with LDA effected
elimination to the dialkyne diol 24.16 Dihydroxylation of
the bis-DPS ether enediyne 25 followed by diol oxidative
cleavage with lead tetraacetate and subsequent Wittig
homologation yielded the conjugated ester 28, which was
subjected to successive reduction with DIBAL-H and
oxidation with MnO2 to afford enal 30. The synthesis of
the allylic stannane spacer segment 31 (E) was completed
by application of the previously employed four-step
stannylation-oxidation-reduction-protection sequence.
a
(a) Ph3PdCHCO2Et, THF, reflux (88%); (b) Swern (97%); (c)
DIBAL-H, hexanes (97%); (d) (1) Bu3SnLi, (2) ADD, (3) (S)-BINOL,
EtOH, LAH, (4) MOMCl, i-Pr2NEt.
SCHEME 2 a
Our initial route to the OTBS analogue 39 of ester 28
(Scheme 5) commenced with 4-pentenol (32) and pro-
ceeded by a route analogous to that employed in Scheme
4. However, intermediates 33-37 proved somewhat
volatile, resulting in diminished yields. Ultimately, this
sequence failed because the ozonolysis of enyne 38 could
not be achieved satisfactorily. The successful route
outlined in Scheme 4 circumvented these two problems.
Initially, the TBS ether was employed in this sequence,
but conversion of 30 (TBS in place of DPS) to stannane
31 (TBS in place of DPS) proceeded in only 13% yield,
possibly the result of TBS cleavage in the reduction step.
a
(a) (MeO)3CMe, EtCO2H (96%); (b) AD-mix â (99%); (c)
MeONHMe‚HCl, AlMe3, (99%); (d) TBSCl, Im (99%); (e) DIBAL-H
(99%).
oxidation-reduction-protection sequence13 leading to
the (R)-allylic R-OMOM stannane 10 (A) of high enan-
tiomeric purity. The commercially available enal 814 was
carried through the same stannane sequence to give
the related (R)-allylic R-OMOM stannane 12 (B) in four
steps.
Our synthesis of the known core precursor C (16) was
effected as previously described6e from allylic alcohol 13
in the five steps outlined in Scheme 2. The spacer unit
D (19) was synthesized along the same lines as the
stannane termini A and B starting from 5-hexyn-1-ol (17)
as in Scheme 3.
The two butenolide termini, segments F and G, were
prepared according to Schemes 6 and 7. Both utilized the
known R-SPh lactone 41.17 For the former, alkylation of
41 with iodide 42 and subsequent sulfoxide elimination
afforded butenolide 44. The alcohol derivative 45 was
oxidized to the unstable aldehyde 46, which was con-
verted without purification to vinyl iodide 47 (F ) upon
Takai homologation.18
(10) Pandey, G.; Hajra, S.; Ghorai, M. K.; Kamur, K. R. J . Org.
Chem. 1997, 62, 5966.
(11) Lutz, C.; Knochel, P. J . Org. Chem. 1997, 62, 7895.
(12) Dale, J . A.; Dull, D. L.; Mosher, H. S. J . Org. Chem. 1969, 34,
2543.
(15) Hoye, T. R.; Suhadolnik, J . C. Tetrahedron 1986, 42, 2855.
(16) Yadav, J . S.; Deshpande, P. K.; Sharma, G. V. M. Tetrahedron
1990, 46, 7033.
(17) White, J . D.; Somers, T. C.; Reddy, G. N. J . Org. Chem. 1992,
(13) Marshall, J . A.; Garofalo, A. W.; Hinkle, K. W. Org. Synth. 1999,
77, 98.
(14) Aldrich Flavors and Fragrances, Milwaukee, WI.
57, 4991.
(18) Takai, K.; Nitta, K.; Utimoto, K. J . Am. Chem. Soc. 1986, 108,
7408.
J . Org. Chem, Vol. 68, No. 5, 2003 1773

Marshall et al.
SCHEME 4
SCHEME 5
SCHEME 7 a
a
(a) CHI3, CrCl2 (50%); (b) Dowex 50W, MeOH (93%); (c) Tf2O,
2,6-lut; (d) TBSOTf (85%); (e) m-CPBA; (f) PhCH3 (g) HF‚py, THF.
SCHEME 6
of the secondary alcohol as the TBS ether completed the
synthesis of the alkylating agent 52.21 The alkylation of
lactone 41 was effected in 54% yield. The sulfoxide
derivative underwent pyrolysis in refluxing toluene to
yield the butenolide 54, which was desilylated with HF-
pyridine in THF.
Segm en t Cou p lin g
The next phase of our studies was directed at the
coupling of appropriate A-G intermediates and hydro-
genation-deprotection of the coupling products to the
selected threo, trans, threo, trans, threo core compounds,
asiminocin, asimin, and asimicin, with (R)-C4, (R)-C10,
and (S)-C30 OH substituents, as depicted in Figure 1.
Core assembly for all three would employ allylindium
additions to aldehyde C along the lines of Figure 2.
An analogous sequence19 was employed for the syn-
thesis of butenolide G (Scheme 7). However, incorpora-
tion of the eventual C4 hydroxy substituent required a
somewhat lengthier sequence to prepare the alkylating
agent, triflate 52. To that end, aldehyde 48, available in
three steps from (R)-dimethyl malate,20 was subjected to
Takai homologation18 affording the vinyl iodide 49.
Cleavage of the acetonide and ensuing selective conver-
sion of the primary alcohol to the triflate and protection
Asim in ocin
The synthesis of asiminocin (64) proceeded according
to Scheme 8. Thus, in situ transmetalation of allylic
stannane 10 (A) with InBr3 in the presence of aldehyde
16 (C) afforded the anti adduct 56 in 80% yield virtually
free of isomeric byproducts (1H NMR analysis). The
double bond of the tosylate derivative 57 was reduced
over Pd-C, and the intermediate alcohol was subjected
(19) This sequence closely parallels that of Mosin, B.: Maezaki, N.;
Kojima, N.; Sakamoto, A.; Iwata, C.; Tanaka, T. Org. Lett. 2001, 3,
429.
(20) Saito, S.; Hasegawa, T.; Inaba, M.; Nishida, R.; Fujii, T.;
Nomizu, S.; Moriwake, T. Chem. Lett. 1984, 1389.
(21) Mori, Y.; Yaegashi, K.; Furukawa, H. J . Am. Chem. Soc. 1996,
118, 8158.
1774 J . Org. Chem., Vol. 68, No. 5, 2003

Modular Synthesis of Annonaceous Acetogenins
SCHEME 8
to Swern oxidation22 affording aldehyde 58 in 76% yield
for the two steps. A second in situ transmetalation of
allyllic stannane 19 (D) in the presence of aldehyde 58
proceeded in 92% yield. The resulting anti adduct 59 was
the sole detectable product. Tosylation led to the bisto-
sylate, which was converted to the bistetrahydrofuran
61 upon exposure to TBAF in THF for 23 h. These
conditions also caused partial cleavage of the DPS ether.
This cleavage was quite slow and prolonged exposure to
TBAF was required, resulting in some loss of material.
Conceivably, the mixture of C30 alcohol and DPS ether
could be carried through the subsequent steps with final
cleavage being effected in the last deprotection step.
However, this option was not explored. Sonogashira
coupling9 of the alkynyl THF 61 with vinyl iodide 47 (F )
produced the fully elaborated acetogenin framework 62.
The use of Pd(PPh3)2Cl2 as a precatalyst proved some-
what superior to Pd(PPh3)4 (68% vs 56% yield) for this
coupling reaction. Selective hydrogenation of the dienyne
moieties with diimide generated from tosylhydrazide23
afforded the octahydro product contaminated with tosyl-
hydrazide byproducts. Hydrolysis of the MOM ethers was
effected with aqueous HCl in methanolic THF to afford
asiminocin (64), which was easily separated from the
aforementioned byproducts. The rotation and spectra of
the synthetic material were in close agreement with the
reported values.24
aliphatic terminus was appended to the core precursor
16 (C) through use of the allylic stannane 12 (B) and
1
InBr3. The adduct 65 (dr > 95:5 according to H NMR
analysis) was tosylated, debenzylated, and oxidized to
aldehyde 67. The spacer unit 19 (D) was transmetalated
to the allylic indium bromide with InBr3, and this
intermediate was added to aldehyde 67 in situ. Once
again the addition proceeded with high diastereoselec-
tivity affording the anti adduct 68 (dr > 95:5). The
bistosylate 69 cyclized to the elaborated bistetrahydro-
furan core segment 70 upon exposure to TBAF in THF.
Sonogashira coupling of the terminal alkyne of 70 with
the vinylic iodide 55 (G) led to the fully elaborated
acetogenin structure 71 which was reduced, as before,
with diimide and subjected to aqueous HCl to remove the
MOM ether protecting groups. The product 73 thus
obtained was identified as asimicin through comparison
of the NMR spectra and optical rotation.25
Asim in
The third target of these studies, asimin (79), was
prepared by the sequence outlined in Scheme 10. Alde-
hyde 67, previously employed in the synthesis of asimicin
(Scheme 9), was treated with the allylic indium reagent
derived from spacer stannane 31 (E) to afford the adduct
74 with high diastereoselectivity. Cyclization of the
bistosylate 75 and attendant (slower) cleavage of the DPS
ether with TBAF proceeded in modest yield. Extended
reaction times offered no improvement and appeared to
diminish the yield and purity of the isolated product.
Furthermore, no significant byproducts could be identi-
fied. This situation is to be contrasted with the analogous
higher yield cyclization of the bistosylate 69 in which
there is no protected “remote” OH group.
Asim icin
A parallel sequence of reactions was employed for the
synthesis of asimicin (73, Scheme 9). In that case the
(22) Omura, K.; Swern, D. Tetrahedron 1978, 34, 1651.
(23) Hart, D. J .; Hong, W.-P.; Hsu, L.-Y. J . Org. Chem. 1987, 52,
4665.
(24) Zhao, G.-X.; Chao, J .-F.; Zeng, L.; Rieser, M. J .; McLaughlin,
J . L. Bioorg. Med. Chem. 1996, 4, 25.
(25) Hoye, T. R.; Tan, L. S. Tetrahedron Lett. 1995, 36, 1981.
J . Org. Chem, Vol. 68, No. 5, 2003 1775

Marshall et al.
SCHEME 9
SCHEME 10
The bistetrahydrofuran 76 was coupled to the buteno-
lide terminus 47 (F ) by the Sonogashira methodology to
afford the enyne 77 in satisfactory yield. Selective
hydrogenation of this dienyne with diimide followed
by acidic hydrolysis of the MOM ethers afforded asimin
(79). The NMR spectra and rotation of this material
were in close agreement with those of the natural
product.3
Bu lla n in , a Cor e Ster eoisom er of Asim in ocin
As an initial step toward establishing the applicability
of this modular approach to other core stereoisomers of
bistetrahydrofuran acetogenins, we targeted bullanin
(91), a C24 epimer of asiminocin.26 As noted in Figure 2,
diastereomeric homoallylic alcohol adducts can be pre-
pared from a common R-OMOM allylic stannane through
a change in reaction conditions. Transmetalation with
1776 J . Org. Chem., Vol. 68, No. 5, 2003

Modular Synthesis of Annonaceous Acetogenins
SCHEME 11
present study toward H-116 human colon cancer cells.
These compounds have previously been examined for
antitumor activity against HT-29 colon cancer cells (Table
1).
Remarkably, it was found that the potency of the
synthesized acetogenins against H-116 cells was lower
than that reported for HT-29 cells by a factor of 108 to
109 (!) as measured by the IC50 values. The source of the
apparent discrepancy between the two sets of determina-
tions is unknown. Possibly the two cell lines possess a
significant difference in susceptibility. In any event,
activities in the 10-3 to 10-4 µM range exhibited by our
synthesized acetogenins justify additional exploration of
these compounds for cancer chemotherapy. As a com-
parison, 5-fluorouracil, a current chemotherapeutic agent
for colon cancer,29 shows significantly lower activity than
any of the tested acetogenins toward the H-116 cell line.
Of further interest are the IC90 results for these com-
pounds. Although the IC50 values of the five acetogenins
in Table 1 are comparable, both asimicin and asimin
exhibit significantly lower IC90 levels in comparison to
asiminocin and bullanin. The ratio IC90/IC50 is a measure
of the concentration response. Thus, asiminocin and
bullanin, although quite active in the IC50 assay, show a
rather shallow concentration-response plot (surviving
fraction of tumor cells vs concentration of drug), reflected
in the high IC90/IC50 ratio, and would therefore be
considered less likely candidates for in vivo studies.30
InBr3 in situ affords erythro (anti) adducts, whereas
sequential BF3-promoted isomerization and addition
leads to threo (syn) isomers. Accordingly, we could employ
this latter methodology to add the aliphatic terminus A
to the core precursor aldehyde C (Figure 3). In the course
of our earlier studies we noted that cleavage of the DPS
ether concomitant with tetrahydrofuran formation was
not always complete. Accordingly, we employed the more
labile TBS ether 80, prepared from hydroxy ester 4 along
the lines of Scheme 1, for the present synthesis. 1,3-
Isomerization to 81 could be effected by treatment of 80
with BF3‚OEt2 (Scheme 11).
The addition of stannane 81 to aldehyde 16 (C)
afforded the syn adduct 82 diastereoselectively (dr > 95:5
according to 1H NMR analysis) in 77% yield (Scheme 12).
Isomerization of stannane 80 to 81 could also be carried
out in situ followed by addition of the aldehyde. However,
the isomerization proved to be relatively slow and its
progress was difficult to monitor by TLC, with the result
that the γ adduct 82 was sometimes contaminated with
the isomeric adduct of the R stannane 80. Hence we
elected to isolate the γ stannane before addition of the
aldehyde. The remainder of the synthetic sequence was
identical to that described for asiminocin (Scheme 8).
However, the TBAF-promoted cyclization of the TBS
ether bistosylate 87 proceeded in higher yield than that
of the analogous conversion of the related DPS ether 60
and cleavage of the TBS ether was complete.
It was of interest to compare the activity of asimicin,
one of the most active of the synthesized acetogenins,
with its dihydro derivative 93. This substance was
prepared by hydrogenation of the butenolide 71 over Pd-
on-carbon in ethanol (Scheme 13). The crude hydrogena-
tion product (92) was isolated as a 1:1 mixture of
diastereoisomers,31 but following hydrolysis of the MOM
ethers, it was possible to partially separate the mixture
through column chromatography on silica gel resulting
in a 4:1 mixture. This mixture exhibited an IC50 value
comparable to that of the parent. However, a significantly
increased drug concentration was required to reach an
IC90 of H-116 cell growth. McLaughlin has stated that
bullatacin is “much more active” than its reduction
product dihydrobullatacin at the IC50 level.32 However,
it should be noted that rolliniastatin-1 and its dihydro
derivative show comparable activity.33 IC90 values were
not measured for these compounds. Although it is gener-
ally conceded that the butenolide terminus plays an
important role in the growth-inhibitory activity of the
acetogenins,34 this feature does not seem prerequisite to
cytotoxicity in all acetogenins. Furthermore, the differ-
ences in inhibitory concentrations between the 50% and
90% levels show no obvious correlation with structure.
Biologica l Activity
The Annonaceous acetogenins inhibit intracellular
production of ATP by blocking electron transport in
Complex I, a membrane-bound protein of the mitochon-
drial electron transport system.27 The inhibition is thought
to involve interference with ubiquinone-linked NADH
oxidase in the plasma membrane. It has been proposed
that the hydrophilic core of these compounds binds to the
polar glyceryl end groups of the phospholipid segment
of the membrane, allowing the hydrophobic chains to
penetrate the bilayer and interact through the furanone
terminus with the ubiquinone-dependent site in the
cellular electron transport system.28 The resultant ces-
sation of ATP production in the cellular membrane
inhibits the growth of cancer cells and leads to ultimate
selective cell death.
The modular syntheses described in this report should
permit access to meaningful amounts of natural aceto-
genins and novel analogues for biological evaluation
and for structure-activity studies. As a step in this
direction we were interested in measuring the cytotox-
icity of the four synthetic acetogenins prepared in the
(29) Blumberg, D.; Ramanthan, R. K. J . Clin. Gastroenterol. 2002,
34, 15.
(30) Valeriote, F. Unpublished work.
(31) Cortes, D.; Myint S. H.; Harmange, J . C.; Sahpaz, S.; Figadere,
B. Tetrahedron Lett. 1992, 33, 5225.
(32) Rupprecht, J . K.; Hui, Y.-H.; McLaughlin, J . L. J . Nat. Prod.
1990, 53, 237.
(26) Zhao, G.-X.; Ng, J . H.; Kozlowzki, J . F.; Smith, D. L.; McLaugh-
lin, J . L. Heterocycles 1994, 38, 1897.
(27) Miyoshi, H.; Ohshima, M.; Shimada, H.; Akagi, H.; Iwamura,
H.; McLaughlin, J . L. Biochem. Biophys. Acta 1998, 1365, 443.
(28) Shimada, H.; Grutzner, J . B.; Kozlowski, J . F.; McLaughlin, J .
L. Biochemistry 1998, 37, 854.
(33) Gu, Z.-M.; Zhao, G.-X.; Oberlies, N. H.; Zeng, L.; McLaughlin,
J . L. In Recent Advances in Phytochemistry; Arnason, J . T., Mata, R.,
Romeo, J . T., Eds.; Plenum Press: New York, 1995; Vol. 29, pp 271-
274.
(34) Arndt, S.; Emde, U.; Ba¨urle, S.; Friedrich, T.; Grubert, L.; Koert,
U. Chem. Eur. 2001, 7, 993.
J . Org. Chem, Vol. 68, No. 5, 2003 1777

Marshall et al.
SCHEME 12
TABLE 1. Cytotoxicity of An n on a ceou s Acetogen in s
tow a r d Hu m a n Colon Ca n cer Cells
TABLE 2. Disk Diffu sion Assa y for An titu m or
Com p ou n d s
H-116a
IC90
HT-29b
IC50
compound
µg/disk
H-116a
CEMb
CFU-GMc
IC50
asiminocin
asimicin
dihydroasimicin
asimin
bullanin
5-fluorouracil
4 × 10-2
1.5 × 10-2
7.5 × 10-2
3 × 10-1
600
450
300
400
350
500
300
350
250
0
200
500
0
300
200
50
150
0
compound
(µg/mL)c (µg/mL)c IC90/IC50 (µg/mL)c ref
-12
asiminocin
asimicin
5 × 10-3
1.5
300
4
<10
22
3
5 × 10-4
2 × 10-3
1 × 10-1
3 × 10-3
1 × 10-1
<10-12
1.5 × 10-2
4 × 10-2
dihydroasimicin 2 × 10-4
500
10
33
10
asimin
3 × 10-4
3 × 10-3
<10-12
3
5 × 10-12 24
a
b
Human colon cancer. Human leukemia. c Human bone mar-
row.
bullanin
5-fluorouracild
1.5 × 10-1 1.5
Present study. See references. c 1 g/mL ≈ 1.5 M. 1 g/mL
a
b
d
inhibition is measured to evaluate the effective cytotox-
icity of the test compound toward a given cell type. A
“zone” is defined as the distance from the edge of the filter
disk to the point of appearance of control-sized colonies.
The diameter of the filter disk is arbitrarily taken as 200
“units” of activity. Each of the selected acetogenins was
assayed against three cell types; H-116 (human colon
cancer), CEM (human leukemia), and CFU-GM (human
bone marrow). On the basis of past experience,35 it has
been determined that compounds exhibiting a differential
activity against solid tumor cancer cells vs either leuke-
mia or normal cells of more than 250 units are good
candidates for in vivo testing and eventual progression
to clinical evaluation. Accordingly, asiminocin and asimin
are the most promising members of this group.
≈ 7 M.
SCHEME 13
Differ en tia l Cytotoxicity-Disk Diffu ssion Assa ys
Su m m a r y
An in vitro cell-based disk diffusion assay was em-
ployed to evaluate differential cytotoxicity of the aceto-
genins synthesized in this study (Table 2).35 In this assay
a 6.5-mm filter disk impregnated with a measured
amount of the test compound is placed on a plate that
has been seeded with either tumor or normal cells. The
plates are incubated for 7 days, and a zone of colony
In summary, the proposed modular synthesis approach
has been reduced to practice for representative members
(35) Valeriote, F.; Grieshaber, C. K.; Media, J .; Pietraszkewicz, H.;
Hoffmann, J .; Pan, M.; McLaughlin, S. J . Exp. Ther. Oncol. 2002, 2,
228.
(36) Valeriote, F.; Corbett, T.; LoRusso, P. Int. J . Pharmacog. 1995,
33, 59.
1778 J . Org. Chem., Vol. 68, No. 5, 2003

Modular Synthesis of Annonaceous Acetogenins
of the bistetrahydrofuran Annonaceous acetogenins. The
stereochemically defining steps proceed with excellent
selectivity. The key bistetrahydrofuran cyclization reac-
tions take place in ca. 50-70% yield, depending on the
nature of the “remote” protected OH substituent. The
relatively slow cleavage of a remote ODPS ether in the
sequences leading to asiminocin (Scheme 8) and asimin
(Scheme 10) required extended reaction times, which
appear to lower the isolated yield of the bistetrahydro-
furan products. The related cyclization reactions leading
to asimicin (Scheme 9) and bullanin (Scheme 12), in
which either no remote silyl ether or a labile OTBS ether
was present, proceed in significantly higher yield.
The synthesized acetogenins were found to possess IC50
values of ∼10-2 to 10-4 µM toward H-116 solid human
colon cancer cells. This value contrasts sharply with
reported concentrations of ∼10-12 µM for the natural
products vs HT-29 colon cancer cells. IC90 values have
not previously been determined for Annonaceous aceto-
genins. These values were significantly higher than the
IC50 values and, depending on the specific acetogenin,
demonstrated either steep or shallow concentration-
response relationships. Differential cytotoxicity, mea-
sured by a disk diffusion assay, showed that the synthe-
sized natural acetogenins are quite selective for H-116
human colon cancer cells vs human bone marrow and
leukemia cells. Future studies will explore these issues
in greater detail, as the synthetic methodology delineated
in this investigation should be broadly applicable to other
bistetrahydrofuran acetogenins and analogues thereof.
Ack n ow led gm en t . Funding was provided from
research grant R01GM CA56769 from the National
Institutes of General Medical Sciences.
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
1
cedures, H NMR spectra of key intermediates, and selected
13C NMR spectra. This material is available free of charge via
the Internet at http://pubs.acs.org.
J O026433X
J . Org. Chem, Vol. 68, No. 5, 2003 1779