Synthesis and biological evaluation of a focused mixture library of analogues of the antimitotic marine natural product curacin A

Article abstract of DOI:10.1021/ja002213u

The marine natural product curacin A served as the lead compound for the combinatorial synthesis of 6-compound mixture libraries. Fluorous trapping with a vinyl ether tag was used to streamline purification of the heterogeneous multicomponent reaction pro

Full text of DOI:10.1021/ja002213u

J. Am. Chem. Soc. 2000, 122, 9391-9395
9391
Synthesis and Biological Evaluation of a Focused Mixture Library of
Analogues of the Antimitotic Marine Natural Product Curacin A
Peter Wipf,*,§ Jonathan T. Reeves,§ Raghavan Balachandran,† Kenneth A. Giuliano,‡
Ernest Hamel, and Billy W. Day*,†
Contribution from the Department of Chemistry, UniVersity of Pittsburgh, Pittsburgh, PennsylVania 15260,
Department of EnVironmental and Occupational Health, and Department of Pharmaceutical Sciences,
UniVersity of Pittsburgh, Pittsburgh, PennsylVania 15238, Cellomics, Inc., Pittsburgh, PennsylVania 15238,
and Screening Technologies Branch, DeVelopmental Therapeutics Program, DiVision of Cancer Treatment
and Diagnosis, National Cancer Institute, Frederick Cancer Research and DeVelopment Center,
Frederick, Maryland 21702
ReceiVed June 21, 2000
Abstract: The marine natural product curacin A served as the lead compound for the combinatorial synthesis
of 6-compound mixture libraries. Fluorous trapping with a vinyl ether tag was used to streamline purification
of the heterogeneous multicomponent reaction products and provide chemically clearly defined mixtures. The
screening profile of one mixture library, 17mix, was attractive enough to warrant the re-synthesis of the individual
compounds, and an evaluation of their biological effects validated the composite data previously obtained on
the product mixture. The most active of these compounds inhibited tubulin polymerization with an IC50 of ca.
1 µM, showed an average growth inhibition activity GI50 of ca. 250 nM, inhibited [3H]colchicine binding to
tubulin, and blocked mitotic progression at nanomolar concentrations. These compounds represent some of
the most potent synthetic curacin A analogues identified to date but have simplified structures, greater water
solubility, and increased chemical stability.
Introduction
Natural products remain a significant source of promising
lead structures for drug development.1 In recent years, an
expansion of the structural diversity pool by preparation of
libraries of natural products or natural product-like molecules
has become a major focus of combinatorial chemistry.2 After
completion of the total synthesis of the strongly antimitotic
Lyngbya majuscula metabolite curacin A in 1996,3 we remained
intrigued by the impressive antiproliferative profile4 of this
marine natural product and the potential to use it as a scaffold
for the development of novel tubulin polymerization inhibitors.5,6
Curacin A promotes arrest of the cell cycle at the G2/M
checkpoint and competitively inhibits the binding of [3H]-
colchicine to tubulin, and it can therefore be considered a
colchicine site agent.5,7 In addition to a large number of total
syntheses of curacin A,8 the attractive biological properties of
this compound have led to numerous analogue studies. However,
even minor changes in the structure of curacin A can lead to
essentially inactive derivatives (Figure 1).7,9 Critical issues for
further pharmaceutical development of any sufficiently active
analogue are increases in chemical stability and hydrophilicity
and improved availability versus the natural product. Curacin
* To whom correspondence should be addressed.
§ Department of Chemistry, University of Pittsburgh.
† Department of Environmental and Occupational Health, and Department
of Pharmaceutical Sciences, University of Pittsburgh.
‡ Cellomics, Inc.
National Cancer Institute.
(6) For recent evaluations of small-molecule antimitotic agents, see: (a)
Haggarty, S. J.; Mayer, T. U.; Miyamoto, D. T.; Fathi, R.; King, R. W.;
Mitchison, T. J.; Schreiber, S. L. Chem. Biol. 2000, 7, 275. (b) Owa, T.;
Okauchi, T.; Yoshimasa, K.; Sugi, N. H.; Ozawa, Y.; Nagasu, T.; Koyanagi,
N.; Okabe, T.; Kitoh, K.; Yoshino, H. Bioorg. Med. Chem. Lett. 2000, 10,
1223. (c) Uckun, F. M.; Mao, C.; Vassilev, A. O.; Navara, C. S.; Narla, K.
S.; Jan, S.-T. Bioorg. Med. Chem. Lett. 2000, 10, 1015.
(7) Verdier-Pinard, P.; Lai, J.-Y.; Yoo, H.-D.; Yu, J.; Marquez, B.; Nagle,
D. G.; Nambu, M.; White, J. D.; Falck, J. R.; Gerwick, W. H.; Day, B. W.;
Hamel, E. Mol. Pharmacol. 1998, 53, 62.
(8) (a) White, J. D.; Kim, T.-S.; Nambu, M. J. Am. Chem. Soc. 1995,
117, 5612. (b) Hoemann, M. Z.; Agrios, K. A.; Aube´, J. Tetrahedron Lett.
1996, 37, 953. (c) Ito, H.; Imai, N.; Takao, K.; Kobayashi, S. Tetrahedron
Lett. 1996, 37, 1799. (d) Onoda, T.; Shirai, R.; Koiso, Y.; Iwasaki, S.
Tetrahedron Lett. 1996, 37, 4397. (e) Lai, J.-Y.; Yu, J.; Mekonnen, B.;
Falck, J. R. Tetrahedron Lett. 1996, 37, 7167. (f) White, J. D.; Kim, T.-S.;
Nambu, M. J. Am. Chem. Soc. 1997, 119, 103. (g) Hoemann, M. Z.; Agrios,
K. A.; Aube´, J. Tetrahedron 1997, 53, 11087. (h) Muir, J. C.; Pattenden,
G.; Ye, T. Tetrahedron Lett. 1998, 39, 2861.
(1) (a) Cragg, G. M.; Newman, D. J.; Snader, K. M. J. Nat. Prod. 1997,
60, 52. (b) Shu, Y.-Z. J. Nat. Prod. 1998, 61, 1053. (c) Nicolaou, K. C.;
Vourloumis, D.; Winssinger, N.; Baran, P. S. Angew. Chem., Int. Ed. 2000,
39, 44.
(2) (a) Tan, D. S.; Foley, M. A.; Shair, M. D.; Schreiber, S. L. J. Am.
Chem. Soc. 1998, 120, 8565. (b) Nicolaou, K. C.; Winssinger, N.;
Vourloumis, D.; Oshima, T.; Kim, S.; Pfefferkorn, J.; Xu, J.-Y.; Li, T. J.
Am. Chem. Soc. 1998, 120, 10814. (c) Meseguer, B.; Alonso-Diaz, D.;
Griebenow, N.; Herget, T.; Waldmann, H. Angew. Chem., Int. Engl. 1999,
19, 2902.
(3) Wipf, P.; Xu, W. J. Org. Chem. 1996, 61, 6556.
(4) (a) Verdier-Pinard, P.; Sitachitta, M.; Rossi, J. V.; Sackett, D. L.;
Gerwick, W. H.; Hamel, E. Arch. Biochem. Biophys. 1999, 370, 51. (b)
Gerwick, W. H.; Proteau, P. J.; Nagle, D. G.; Hamel, E.; Blokhin, A.; Slate,
D. J. Org. Chem. 1994, 59, 1243.
(5) For a review of agents that interact with the mitotic spindle, see:
Jordan, A.; Hadfield, J. A.; Lawrence, N. J.; McGown, A. T. Med. Res.
ReV. 1998, 18, 259-296.
10.1021/ja002213u CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/19/2000

9392 J. Am. Chem. Soc., Vol. 122, No. 39, 2000
Wipf et al.
Scheme 1
exposure of 5-7 as well as derivatives with other aryl sub-
stituents to a cocktail of 3-6 aryllithium reagents. The most
biologically active mixtures were obtained from reaction of 5-7
with 2-lithiated furan, thiophene, benzofuran, benzothiophene,
anisole, and 1,4-dimethoxybenzene (Scheme 2). The lithiation
was performed directly on the mixture and was heteroatom-
directed. Since a large excess (3 equiv each) of organolithium
reagents in a 1:1:1:1:1:1 ratio was used to ensure fast and
unselective aldehyde addition, the resulting alcohols 8mix-
10mix were heavily contaminated with excess Nu-H after
aqueous workup. Fluorous quenching of the crude product with
an excess of the recently developed vinyl ether 11 provided a
convenient means to ensure rapid and thorough purification.12,13
After simple liquid/liquid extraction with a mixture of the
perfluorinated solvent FC-72 and MeCN/H2O, the partially
fluorinated acetals 12mix-14mix were collected in the FC-72
layer; organic impurities remained in the organic phase and
inorganic salts preferred the aqueous environment. Methanolysis
of the fluorous extract and renewed postreaction fluorous/
organic/aqueous liquid-liquid extraction finally provided pure
6-compound mixtures 15mix-17mix in the organic phase,14
while the fluorous phase contained the solvolysis products of
protective group 11.
In Table 1, the biochemical and cell growth activity of
15mix-17mix is summarized. Each of the mixtures inhibited
by >50% the 30 °C GTP-induced polymerization of glutamate-
containing isolated bovine brain tubulin15a at concentrations <5
µM (summed concentration based on average molecular weight
of the components), with both 16mix and 17mix showing high
potency. Assays to determine if these compounds inhibited
binding of [3H]colchicine to tubulin at both 30 and 37 °C15b
showed both 15mix and 16mix were significantly weaker at
displacement than curacin A, whereas one or more of the
Figure 1. Examples of inactive curacin A analogues.
A is sensitive to oxidation, acids, and bases, and is readily
irreversibly absorbed into plastic containers such as polyethylene
96-well plates.
Results and Discussion
Since curacin A binds to the colchicine site on tubulin, and
the alkenyl thiazoline moiety in 1 is largely responsible for the
chemical instability of the natural product, we decided to
substitute the heterocycle with electron-rich aromatics reminis-
cent of the A-ring in 2. Furthermore, we planned to substitute
the homoallylic methyl ether terminus of curacin A with a broad
range of more hydrophilic benzylic alcohols but maintain the
diene spacer unit, which has proven to be essential in the limited
SAR studies of curacin A published to date.4,7,9 Accordingly,
three key building blocks 5-7 were prepared by standard
solution synthesis from the readily available3 aldehyde 3
(Scheme 1). After protection with ethylene glycol, desilylation,
and mesylation, Finkelstein reaction with sodium iodide pro-
vided dioxolane 4 in 57% yield. The corresponding Wittig
reagent was condensed with 2,4-dimethoxy-, 2,5-dimethoxy-,
and 3,4,5-trimethoxybenzaldehyde, respectively, to give alde-
hydes 5-7 in 50-59% overall yields from 4 after acetal
cleavage.10
The solution-phase combinatorial synthesis of mixtures for
biological screening is not commonly used, but represents an
efficient approach for rapid lead evaluation if the composition
of the mixture is clearly defined and the throughput of the assay
is limited.11 In an effort to establish the validity of our analogue
design, we prepared 7 mixtures of 6-9 compounds each by
(11) (a) Boger, D. L.; Lee, J. K.; Goldberg, J.; Jin, Q. J. Org. Chem.
2000, 65, 1467. (b) Houghten, R. A.; Pinilla, C.; Appel, J. R.; Blondelle,
S. E.; Dooley, C. T.; Eichler, J.; Nefzi, A.; Ostresh, J. M. J. Med. Chem.
1999, 42, 3743. (c) Nazarpack-Kandlousy, N.; Zweigenbaum, J.; Henion,
J.; Eliseev, A. V. J. Comb. Chem. 1999, 1, 199.
(12) Wipf, P.; Reeves, J. T. Tetrahedron Lett. 1999, 40, 5139.
(13) (a) Studer, A.; Hadida, S.; Ferritto, R.; Kim, S.-Y.; Jeger, P.; Wipf,
P.; Curran, D. P. Science 1997, 275, 823. (b) Curran, D. P. Angew. Chem.,
Int. Ed. 1998, 37, 1174.
(14) The three mixtures were characterized by LC-MS using a reverse-
phase (C18) column and positive ionization electrospray mass spectral
detection as well as by LC NMR. Each mixture contained the six expected
products in close-to-equimolar ratio, with minor amounts (ca. 15-20%) of
the (E)-alkene derivatives from the Wittig reaction. No organic impurities
>5% were detected. Details of the LC NMR analysis will be reported
elsewhere: Wipf, P.; Reeves, J. T.; Wilkinson, P. S. Bruker Report, in
press.
(9) (a) Marquez, B.; Verdier-Pinard, P.; Hamel, E.; Gerwick, W. H.
Phytochemistry 1998, 49, 2387. (b) Nishikawa, A.; Shirai, R.; Koiso, Y.;
Hashimoto, Y.; Iwasaki, S. Bioorg. Med. Chem. Lett. 1997, 7, 2657. (c)
Martin, B. K. D.; Mann, J.; Sageot, O. A. J. Chem. Soc., Perkin Trans. 1
1999, 2455. (d) Onoda, T.; Shirai, R.; Koiso, Y.; Iwasaki, S. Tetrahedron
1996, 52, 14543. (e) Blokhin, A. V.; Yoo, H. D.; Geralds, R. S.; Nagle, D.
G.; Gerwick, W. H.; Hamel, E. Mol. Pharmacol. 1995, 48, 523.
(10) Aldehydes 5-7 were obtained as all-(E) isomers at the diene moiety
in a ca. 5:1 ratio of (Z):(E) alkene stereoisomers after the Wittig condensation
1
and characterized by H NMR and HRMS.

Antimitotic Marine Natural Product Curacin A
J. Am. Chem. Soc., Vol. 122, No. 39, 2000 9393
Table 1. 50% Colchicine Binding Inhibitory Concentration (CBI), 50% Tubulin Polymerization Inhibition Concentration (TPI), and 50%
Growth Inhibitory Concentrations (GI50) for 15mix-17mix and Discrete Compounds Comprising 17mix17
CBIa (µM)
GI50 (µM)c
TPIb
compd
30 °C
37 °C
IC50 (µM)
MDA-MB231 (48 h)
PC-3 (48 h)
2008 (48 h)
15mix
16mix
17mix
17a
17b
17c
17d
17e
17f
curacin A
>250
>250
4
2
1.4
0.9
1.4
1.2
1.2
1.6
>50
>50
36 ( 5
>50
48 ( 5
26 ( 6
9 ( 6
5.2 ( 2
>50
21 ( 4
>50
0.30 ( 1
0.80 ( 0.5
8.0 ( 4
0.58 ( 0.2
2.7 ( 1
32 ( 4
9 ( 7
39 ( 10
9 ( 5
9 ( 4
>50
0.69 ( 0.1
0.23 ( 0.06
0.28 ( 0.04
0.34 ( 0.03
0.12 ( 0.07
9.7 ( 1
19 ( 2
9 ( 3
0.62 ( 0.05
0.38 ( 0.08
0.62 ( 0.2
0.31 ( 0.2
>50
0.36 ( 0.001
0.28 ( 0.02
0.33 ( 0.02
0.27 ( 0.05
22 ( 5
9 ( 4
28 ( 8
>50
>50
0.78 ( 0.07
0.9 ( 0.3
0.6 ( 0.1
0.096 ( 0.06
0.050 ( 0.009
0.035 ( 0.007
a Values shown are means (N ) 9 over four concentrations) ( standard deviations (SD) for incubation at 30 °C for 15 min and 37 °C for 10 min.
b Average of two determinations except for curacin A, where N ) 7 ((SD). c Means (N ) 4 over 10 concentrations) ( SD.
Scheme 2
compounds comprising 17mix had significant concentration-
dependent inhibitory activity at the colchicine site. Screening
for growth inhibitory activities15c against a short series of human
breast, prostate, and ovarian carcinoma cells confirmed the
trends seen in the biochemical screens and, moreover, showed
that 17mix inhibited cell proliferation at submicromolar con-
centrations suggesting that one or more components might have
potency comparable to that of curacin A.
In light of the intriguing activities in the tubulin polymeri-
zation and growth inhibition assays observed for 17mix, we
decided to synthesize the compounds 17a-f individually and
determine their specific potencies.16,17 Analysis of the discrete
(15) (a) Tubulin without microtubule-associated proteins was prepared
from fresh bovine brains (Hamel, E.; Lin, C. M. Biochemistry 1984, 23,
4173). Reactions were carried out as described previously.7 Tubulin (final
concentration 10 µM.; 1 mg/mL) was preincubated with drugs dissolved in
DMSO (4% v/v final concentration) and monosodium glutamate (0.8 M
final concentration) for 15 min at 30 °C. The reaction mixture was cooled
to 0 °C and GTP (0.4 mM final concentration) was added. Reaction mixtures
were transferred to cuvettes at 0 °C in a Beckman-Coulter 7400 spectro-
photometer reading absorbance at 350 nm. Baselines at 0 °C were
established and temperature was quickly raised to 30 °C (in approximately
1 min with an electronically controlled Peltier temperature controller). The
change in absorbance 20 min after samples reached 30 °C was used to
calculate the extent of polymerization. The change in absorbance at this
time point for vehicle plus no GTP was considered 100% assembly
inhibition, while the change in absorbance for GTP plus vehicle was taken
as 0% inhibition. Each series of determinations included positive and
negative control determinations plus one determination made with 5 µM
curacin A. (b) Using methods described previously (Kang, G.-J.; Getahun,
Z.; Muzaffar, A.; Brossi, A.; Hamel, E. J. Biol. Chem. 1990, 265, 10255),7
5 µM [3H]colchicine (2.3 TBq/mmol), drug (1, 5, 10, 50 or 250 µM), or
vehicle (DMSO, 5% v/v) were incubated at 30 °C for 15 min or at 37 °C
for 10 min with 1 µM tubulin in the presence of 1 M monosodium glutamate,
0.1 M glucose-1-phosphate, 1 mM MgCl2, 1 mM GTP, and 0.5 mg/mL
bovine serum albumin. The solutions were filtered through two stacks of
DEAE-cellulose filters and the radioactivity in the filtrate was determined
by scintillation spectrometry. Each series of determinations included positive
controls of 1, 5, and/or 50 µM curacin A. (c) Cells were plated (500-1500
cells/well depending on the cell line) in 96-well plates and allowed to attach
and grow for 48 h. They were then treated with vehicle (DMSO) or drug
[50, 10, 2, 0.4, and 0.08 µM (for the mixtures, these would be the summed,
apparent concentrations, i.e., approximately six times the concentration of
each compound); 10, 2, 0.4, 0.08, and 0.016 µM for curacin A; then 1, 0.2,
0.04, 0.008, and 0.0016 µM for 17a-e and curacin A] for the given times.
One plate consisted entirely of cells used for a time zero cell number
determination. The other plates contained eight wells of control cells, eight
wells of medium, and each drug concentration tested in quadruplicate. Cell
numbers were obtained spectrophotometrically (absorbance at 490 nm minus
that at 630 nm) in a Dynamax plate reader after treatment with 3-(4,5-
dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-
potencies in all of the isolated protein and cell-based growth
tetrazolium (MTS.; Owen’s reagent) using N-methyldibenzopyurazine
compounds revealed that four of them, 17b-e, had considerable
inhibition screens, and the 1-(2,5-dimethoxyphenyl)-containing
17f was essentially inactive (Table 1). The furan-containing 17a,
although potent in the tubulin assays and against the breast
carcinoma cell line, had reduced antiproliferative activity against
the prostate and ovarian carcinoma cells. As comparison of
methyl sulfate (phenazine methosulfonate) as the electron acceptor.
(16) The synthesis of 17a-f proceeded as shown in Schemes 1 and 2,
but the final products were purified by column chromatography, and were
1
characterized by H and 13C NMR, IR, and HRMS.
(17) As a negative control, single compounds of 15mix were also
resynthesized, and biological testing confirmed their lack of activity.

9394 J. Am. Chem. Soc., Vol. 122, No. 39, 2000
Wipf et al.
B
C
Figure 2. The effects of microtubule disrupting drugs on the microtubule cytoskeleton and the cell cycle of human carcinoma cells. (A) multiparameter
fluorescence micrographs of untreated MCF-7 cells (a-c) and cells treated with compound 17b (1 µM) for 16 h (d-f). Each set of three images,
obtained from the same microscopic field, was used to map the size and shape of the cellular nuclei (a, d), a mitosis-specific phosphorylated core
histone protein (b, e), and the microtubule cytoskeleton (c, f). Compound 17b profoundly altered the morphology of the nuclei, the relative number
of cells blocked in mitosis, and the fibrous structure of the microtubule cytoskeleton. (B) Determination of the minimum effective concentration of
a drug to induce a measurable effect on the cell cycle of living tumor cells. MCF-7 carcinoma cells were treated with a range of vinblastine
concentrations for 14 h followed by the measurement of the mitotic index. The minimum effective concentration of vinblastine to alter the cell
cycle was determined graphically as the concentration (x-axis) where the linear portion of the dose response curve equaled the mitotic index of
normally cycling cells (arrow). (C) Minimum effective concentration values of microtubule disrupting drugs.
17mix with 17a-f clearly demonstrates, tubulin polymerization
and cellular growth inhibition reflect the sum of the individual
compounds that make up 17mix. These results validate our
mixture-based screening strategy.17 In contrast, colchicine
binding inhibition for 17mix does not correlate as well with
the activity of 17a-f, possibly due to a combination of high-
affinity (colchicine-site) and other low-affinity binding sites on
tubulin for these compounds. Some less active curacin A
analogues actually enhance the binding of [3H]colchicine to
tubulin, maybe via allosteric interactions. We are currently
investigating the specific binding modes in detail, but these data
confirm that while for some biological assays mixture analysis
is highly relevant for a quick assessment of biological activities
of a collection of compounds, more specific competitive binding
studies can easily fail to adequately reflect the sum of the
individual mixture components.
Further verification of the biological activity of compounds
17a-e was obtained in living tumor cells by quantifying their
effect on the mitotic index, essentially, counting the number of
cells in the mitotic phase of the cell cycle. Antimitotic
compounds increase the mitotic index, which was quantified
immunohistochemically using an antibody to a mitosis-specific
phosphorylated core histone protein (Mitotic Index HitKit (tm),
Cellomics, Inc. (tm), Pittsburgh, PA).18 MCF-7 human breast
carcinoma cells were treated with increasing concentrations of
compounds 17a-e (7.8-1000 nM), curacin A (3.1-400 nM),
and vinblastine (0.08-10 nM) for 14-16 h. The morphology
(18) Phosphorylation of a core histone occurs during mitosis and is linked
to chromosome condensation. Thus, antimitotic agents cause an accumula-
tion of cells containing the phosphorylated form of this protein. Wei, Y.;
Mizzen, C. A.; Cook, R. G.; Gorovsky, M. A.; Allis, C. D. Proc. Natl.
Acad. Sci. U.S.A. 1998, 95, 7480.

Antimitotic Marine Natural Product Curacin A
J. Am. Chem. Soc., Vol. 122, No. 39, 2000 9395
of the cellular nuclei and the microtubule cytoskeleton, as well
as the phosphorylation of a core histone protein, were visualized
using a multiparameter immunofluorescence approach (Figure
2; A). For example, compound 17b induced a large increase in
the fraction of cells exhibiting condensed chromatin and
increased core histone phosphorylation (Figure 2; A: d and e).
Furthermore, compound 17b induced cell rounding and the
generalized loss of structured microtubules in the same cells
(Figure 2, A: f). The mitotic index in drug-treated cells was
measured automatically using the ArrayScan® II System
(Cellomics, Inc.). To assess the relative activity of compounds
17a-e and two known microtubule disrupting drugs, vinblastine
and curacin A, the minimum effective concentration of each
drug was measured (Figure 2; B). Compounds 17a-d were the
most potent of the new agents with minimum effective
concentrations in the submicromolar range and well within an
order of magnitude of the activity of parent curacin A (Figure
2; C).
data previously obtained on the product mixture. A multipa-
rameter high-content assay showed that compounds 17a-d had
significant activity in altering the cell cycle as well as the
microtubule cytoskeleton in living human carcinoma cells. Five
of the six compounds in 17mix had significant activity in several
other assays. The most active of these compounds, 17c and 17e,
inhibited tubulin polymerization with an IC50 of ca. 1 µM,
showed an average growth inhibition activity GI50 of ca. 250
nM, inhibited [3H]colchicine binding to tubulin, and blocked
mitotic progression at nanomolar concentrations. Compounds
17c and 17e represent therefore some of the most potent
synthetic curacin A analogues identified to date, with an activity
profile rivaling that of the natural product despite their simplified
structure, greater water solubility, and increased chemical
stability. Our analysis demonstrates that, depending on the type
of assay used, the biological evaluation of mixtures can give
highly relevant data that reflect the sum of the individual
components.
Conclusions
Acknowledgment. This work was supported in part by
grants from the National Institutes of Health (CA 78039) and
Merck & Co. We gratefully acknowledge use of the resources
of the Combinatorial Chemistry Center (CCC) at the University
of Pittsburgh.
Despite the relative structural complexity of the lead com-
pound curacin A, we were able to devise an efficient synthetic
strategy for the preparation of 6-compound mixture libraries
that incorporate important structural elements of the marine
natural product. Fluorous trapping with vinyl ether 11 was used
to streamline purification of the heterogeneous multicomponent
reaction products after the diversification step of the library
synthesis and provide structurally defined mixtures. The screen-
ing profile of one mixture library, 17mix, was attractive enough
to warrant the synthesis of the individual components, and an
evaluation of their biological effects validated the composite
Supporting Information Available: Synthetic procedures
and experimental data for 3-7, 15mix, 16mix, 17mix, and
17a-f (PDF). This material is available free of charge via the
Internet at http://pubs.acs.org.
JA002213U