The synthesis and evaluation of a solution-phase indexed combinatorial library of non-natural polyenes for multidrug resistance reversal

Full text of DOI:10.1021/jo982423d

2978
J . Org. Chem. 1999, 64, 2978-2979
Th e Syn th esis a n d Eva lu a tion of a
Solu tion -P h a se In d exed Com bin a tor ia l
Libr a r y of Non -n a tu r a l P olyen es for
Mu ltid r u g Resista n ce Rever sa l
Merritt B. Andrus,*,1 Timothy M. Turner,
Davoud Asgari, and Wenke Li
Department of Chemistry and Biochemistry, Brigham Young
University, C100 Benson Building, Provo, Utah 84602-5700
Received December 11, 1998
While small molecule libraries have gained great promi-
nence as a means of identifying new inhibitors for biological
targets,2 unified approaches that allow for efficient synthe-
sis, direct screening, and facile individual member identi-
fication are still needed. Solid-phase libraries that provide
for synthetic efficiency are limited to screening with isolable
receptors and are not applicable to cell assays.3 In contrast
to the split-pool peptide approach, solid-phase small mol-
ecule libraries have been made almost exclusively using
spatially localized, serial methods. Solution-phase libraries
offer the advantages of using standard reagents, no linker
or support, and the ability to directly screen with isolated
receptors or cellular assays.4 Natural product libraries, while
more focused with less total numbers compared to peptides
and nucleotides, are inherently more diverse in that a great
variety of functionality can be explored within a core
template.5 We now report a solution-phase library,6 based
on the multidrug resistance (MDR) reversing polyene (-)-
stipiamide, that consists of mixtures indexed in two dimen-
sions that provides for efficient combinatorial synthesis,
direct screening with a cellular assay, and the isolation and
testing of individual compounds.
able in that primary and secondary amines would give
widely varying rates of reaction. Thus, the convergent
Sonogashira coupling route to 3, involving formation of a
central carbon-carbon bond remote from the two ends, was
selected as the key step.7b This provides the indexed pools
with the compounds present in equal amounts and allows
for direct analysis in the cellular assay. False positive and
negative results, synergistic effects of weak compounds or
a potent compound being masked by weak members within
a pool, were a concern. It remained to be seen if the inherent
redundancy of the library, in which each individual member
is present in two distinct pools, would minimize these effects.
The identification and isolation of individual compounds
addresses this concern verifying the results from the pools.
End groups were selected to provide a range of steric, as
with the adamantyl groups, polar, as with the dimethox-
yphenyl and bis-hydroxyethyl amides, and nonpolar func-
tionality with some compounds, as in 3 (10b-l), expected
to be potent while others would be expected to be ineffective.
The library was restricted to 6 × 7 with 42 compounds to
allow for rapid characterization of the mixtures and easy
separation of individual members.
Vinyl iodides 7a -f were prepared individually beginning
with the enals 4a -f, obtained from the corresponding
propanals, following the previous route (Scheme 1).7b Alde-
hydes 4a -f were reacted with crotylborane11 derived from
(-)-(R)-pinene to give the antihomoallylic alcohols (34-76%
yield) with selectivities ranging from 4:1 to 8.5:1.7b After
protection, the dienes were subjected to the AD-mix-â
reagent, according to the procedure of Sharpless,12 to
selectively dihydroxylate the terminal alkene.13 The diols
were reacted with sodium periodate to afford aldehydes 6a -
f, which were transformed to the (E)-vinyl iodides (E:Z >
20:1) using the conditions of Takai.14 Removal of the TBS
group gave the desired vinyl iodides 7a -f.
Recently, we showed that synthetic (-)-stipiamide (1)7 had
only moderate reversal activity with colchicine resistant
cells,8 while 2 (ED50 6.5 µM) and 3 (4 µM) were potent with
a variety of drugs with resistant MCF7-adrR cells that
express Pgp (P-glycoprotein), the membrane-bound small-
molecule MDR pump.9 Importantly, 2 and 3 were far less
toxic (4 and 14 µM) compared to stipiamide 1 (0.01 nM). Also,
3 was shown to bind Pgp by monitoring ATPase activity (10
µm) and through displacement of a known label arylazi-
doprazosin (12 µm).10
Polyene 3 was selected as the template due to its activity
and convenience for varying the end groups (Ph and amide).
Critical to the effort was the selection of a diversity generat-
ing reaction that would produce equal relative amounts of
each compound in the pools. Amide formation was unsuit-
The amides 9g-m were made in a more convergent
manner (Scheme 2). Acid 8, obtained from 4-pentyn-1-ol
using PCC, Wittig coupling, and hydrolysis,7b was reacted
with each of the amines (R′g-m , 1.2 equiv) using PyBrop.15
Widely varying yields (63-100%) and rates confirmed the
decision not to use this reaction as the key step for the
library.
(1) To whom correspondence should be addressed. E-mail: mbandrus@
chemgate.byu.edu.
(2) Thompson, L. A.; Ellman, J . A. Chem. Rev. 1996, 96, 555.
(3) Lam, K. S.; Salmon, S. E.; Hersh, E. M.; Hruby, V. J .; Kazmierski,
W. M.; Knapp, R. J . Nature 1991, 354, 82.
(4) Carell, T.; Wintner, E. A.; Rebek, J ., J r. Angew. Chem., Int. Ed. Engl.
1994, 33, 2061. Boger, D. L.; Chai, W.; Ozer, R. S.; Anderson, C. Bioorg.
Med. Chem. Lett. 1997, 7, 463.
(5) Nicolaou, K. C.; Pastor, J .; Winssinger, N.; Murphy, F. J . Am. Chem.
Soc. 1998, 120, 5132. Tan, D. S.; Foley, M. A.; Shair, M. D.; Schreiber, S. L.
J . Am. Chem. Soc. 1998, 120, 8565.
(6) Pirrrung, M. C.; Chen, J . J . Am. Chem. Soc. 1995, 117, 1240.
(7) (a) Andrus, M. B.; Lepore, S. D. J . Am. Chem. Soc. 1997, 119, 2327.
(b) Andrus, M. B.; Lepore, S. D.; Turner, T. M. J . Am. Chem. Soc. 1997,
119, 12159.
(8) Kim, Y. J .; Furihata, K.; Yamanaka, S.; Fudo, R.; Seto, H. J . Antibiot.
1991, 44, 553.
Iodides 7a -f were coupled with the acetylenes 9g-m to
give enynes 10 in an indexed manner using the optimized
(11) Brown, H. C.; Bhat, K. S. J . Am. Chem. Soc. 1986, 108, 293.
(12) Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, K. B. Chem Rev.
1994, 94, 2483.
(9) Higgins, C. F.; Callaghan, R.; Linton, K. J .; Rosenberg, M. F.; Ford,
R. C. Sem. Cancer Biol. 1997, 8, 135.
(10) Performed by S. V. Ambudkar, NIH, Cell Biology: Dey, S.; Ram-
achandra, M.; Pastan, I.; Gottesman, M. M.; Ambudkar, S. V. Proc. Natl.
Acad. Sci. U.S.A. 1997, 94, 10594.
(13) Andrus, M. B.; Lepore, S. D.; Sclafani, J . A. Tetrahedron Lett. 1997,
38, 4043.
(14) Takai, K.; Nitta, K.; Utimoto, K. J . Am. Chem. Soc. 1986, 108, 7408.
(15) Coste, J .; Ferot, E.; J ouin, P.; Castro, B. Tetrahedron Lett. 1991,
32, 1967.
10.1021/jo982423d CCC: $18.00 © 1999 American Chemical Society
Published on Web 04/15/1999

Communications
J . Org. Chem., Vol. 64, No. 9, 1999 2979
Ta ble 1. 6 × 7 In d exed Libr a r y Syn th esis a n d MDR
Sch em e 1
Activity
b
R
R′
yielda (%) ED50 (µM) STDc
left side
constant
a
b
c
d
e
f
a -f
a -f
a -f
a -f
a -f
a -f
a -f
b
b
e
e
e
e
c
d
d
g-m
90
100
100
100
89
100
99
100
90
71
60
87
42
6.6
3.5
9.7
16
1.3
9.4
23
12
>70
0.9
0.3
1.5
5
0.1
1.2
7
g-m
g-m
g-m
g-m
g-m
right side
constant
g
h
i
j
k
l
m
l
m
j
k
l
m
g
h
i
2
2.3
1.6
0.9
1.8
1.85
8.8
1.69
1.45
1.73
1.48
0.9
0.1
0.1
0.2
0.04
0.1
0.02
0.08
0.05
0.02
2
Sch em e 2
isolated
compounds
12
>70
>70
a
Following silica gel filtration, TLC, NMR, MS, and HPLC
b
identified individual members. MCF7-adrR cells and 37 nM
adriamycin. c Standard deviation.
Sonogashira coupling conditions (Table 1).16 This reaction
proved ideal with short reaction times and tolerance of the
wide array of unprotected functionality on the two ends. The
six left-hand iodides 7a -f were reacted individually with a
mixture of the seven acetylenes 9g-m with (Ph3P)2PdCl2
(5 mol %) and CuI (17 mol %) at low temperature (-20 °C)
in ethyl acetate at 1:1.4 stoichiometry with 0.2 equiv of each
acetylene present in the mixture. The scale (∼100 mg)
allowed for multiple MDR assays (1-2 mg per run) and for
subsequent isolation and testing of selected individual
compounds. The yields, following filtration through silica gel
to remove impurities, were quantitative in many cases.
Individual compounds, in both high and lower yielding
reactions, were shown by TLC, NMR, MS, and HPLC to be
present in approximately equal amounts in the pools.
Uniform coupling rates were essential to ensure that each
compound would have an effect in the assay. The reactions
resulted in six pools, constituting the left-hand dimension
of the library, where R was homogeneous within each pool.
Conversely, seven pools homogeneous for the amide end
group were made by reacting each of the individual acety-
lenes 9g-m with a mixture of the six vinyliodides to
generate the right-hand dimension of the library. In total,
13 couplings resulted in the formation of 42 compounds
indexed in two dimensions. The synthesis is thus 3.2 times
more efficient than the serial approach. It is important to
note that each individual compound appears in two distinct
pools and is never found with another compound twice.
MCF-7adrR cell assays with added adriamycin were
performed with the pools in quadruplicate over a range of
concentrations using an average molecular weight (Table
1).17 The R ) Ph(b) pool ED50 was 3.5 µM, while the
cyclohexyl pool at 10 µM was three times less potent. The
dimethoxyphenyl and adamantyl pools were far less active.
At least two times more potent was the naphthyl pool at
1.3 µM. Trends evident in the right-hand amide dimension
include the far less potent adamantyl (g) and the sterate
ester group (i). The hydroxy amide (R′l ) (S)-alaninol) and
morpholine (j) pools were active. Potent individual agents
are readily identified where active pools overlap. Selected
compounds were easily isolated by radial chromatography
from reserved material and additional reversal assays were
performed (Table 1). Isolated compound 3 (10b,l, R ) Ph
and R′ ) alaninol) showed an ED50 of 1.9 µM and 10e,l (2-
naphthyl, alaninol) was 1.7 µM in accord with the pool
assays.18 The most potent new compound was 10e,k (naph-
thyl, phenylaninol) at 1.45 µM. Compound 10e,m (2-naph-
thyl, bishydroxyethylene) was also very active. Less potent
combinations 10c,g (cyclohexyl, adamantyl) at 12 µM and
10d ,h and 10d ,i (adamantyl, dibenzylamine and ethyl
sterate amine) again >70 µM were also in accord with the
pool results. Correlation of these and other individual
activities with the library assays confirms the viability of
the indexed approach using a complex natural product
template. Extended solution-phase polyene libraries and
applications to other natural product templates are now
under investigation.
Ack n ow led gm en t. We wish to thank the National
Institutes of Health (GM57275) and Brigham Young Uni-
versity for funding, Dr. Bruce J ackson for mass spectrom-
etry, Vicki Croy (Purdue University) for the MDR assays,
and Dr. Suresh Ambudkar for Pgp assays.
Su p p or tin g In for m a tion Ava ila ble: Experimental details,
analytical data, and MDR protocols for selected compounds.
(16) See ref 7b and: Sonogashira, K.; Tohda, Y.; Hagihara, N. Tetrahe-
dron Lett. 1975, 4467.
J O982423D
(17) The cells, adriamycin, and reversal agent pools were incubated for
6 d (see Experimental Section). Hall, A. M.; Croy, V., Chan, T.; Ruff, D.;
Kuczek, T.; Chang, C.-j. J . Nat. Prod. 1996, 59, 35.
(18) ED50 for cis-flupenthixol (ref 10) was 5.2 µM performed as a control.