Synthesis and properties of some novel anti-calmodulin drugs

Article abstract of DOI:10.1016/S0968-0896(99)00092-9

The preparation and properties of some novel inhibitors of calmodulin function are described. The compounds are cationic derivatives of phenyl-substituted thiazoles which inhibit the calmodulin stimulation of cyclic-AMP phosphodiesterase and are active against animal tumor cells in culture. These derivatives form the basis for the preparation of new, more potent inhibitors of calmodulin function which could take advantage of the reported elevated levels of calcium-bound calmodulin in tumor cells and show preferential anti-tumor activity. Copyright (C) 1999 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(99)00092-9

Bioorganic & Medicinal Chemistry 7 (1999) 1559±1565
Synthesis and Properties of Some Novel Anti-calmodulin Drugs
Ted T. Sakai a,* and N. Rama Krishna a, b
aComprehensive Cancer Center, NMR Facility, CH19-B31, University of Alabama at Birmingham, UAB Station, Birmingham,
AL 35294-2041, USA
bDepartment of Biochemistry and Molecular Genetics, University of Alabama at Birmingham, UAB Station, Birmingham,
AL 35294-2041, USA
Received 27 August 1998; accepted 29 January 1999
AbstractÐThe preparation and properties of some novel inhibitors of calmodulin function are described. The compounds are
cationic derivatives of phenyl-substituted thiazoles which inhibit the calmodulin stimulation of cyclic-AMP phosphodiesterase and
are active against animal tumor cells in culture. These derivatives form the basis for the preparation of new, more potent inhibitors
of calmodulin function which could take advantage of the reported elevated levels of calcium-bound calmodulin in tumor cells and
show preferential anti-tumor activity. # 1999 Elsevier Science Ltd. All rights reserved.
Introduction
which can bind di€erent aromatic molecules, including
the anti-psychotic drug tri¯uoperazine (TFP) and 8-
anilino-1-naphthalenesulfonate (ANS), a ¯uorescence
probe normally used to study protein structure. They
found that cationic derivatives such as TFP, when
bound to Ca2+±CaM, inhibited the calmodulin activa-
tion of cyclic-AMP phosphodiesterase, whereas the
anionic ANS could bind to Ca2+±CaM but did not
a€ect phosphodiesterase activity. Subsequently, various
aromatic molecules have been found to act as anti-
calmodulin compounds, including most of the other
phenothiazine drugs,14 the anti-estrogen agent tamox-
ifen,15,16 and some synthetic naphthalenesulfonamides
of a,o-diaminoalkanes.17
Calmodulin (CaM) is the main protein involved in cal-
cium bu€ering and in the Ca2+-dependent regulation of
speci®c cell functions. It uses changes in intracellular
calcium levels to modulate a variety of biological and
biochemical processes, a€ecting no fewer than two
dozen enzymes.1±3 Higher eukaryotes have several CaM
genes which are di€erentially regulated while encoding
identical proteins. Because of the importance of Ca2+ in
progression through the cell cycle, CaM plays a critical
role in the regulation of cell proliferation.4±6 It has been
shown that disease states characterized by unregulated
growth, such as cancer, are correlated with elevated
levels of Ca2+-bound CaM.8±11 Thus, CaM is a pro-
mising target for the development of drugs that show
preferential activity in tumors and, as a result, exhibit
fewer toxic side e€ects in normal cells.
Whereas these compounds have been shown to inhibit
calmodulin function, their anti-calmodulin e€ects may
be secondary to other properties, for example, as anti-
psychotic drugs (in the case of the phenothiazines) or
anti-estrogens (in the case of tamoxifen and related
compounds). Little has been done to assess detailed
structure±activity relationships in these anti-calmodulin
drugs, with the phenothiazine drugs14 and the relatively
simple naphthalenesulfonamides17 probably the best-
characterized. Because the structures of compounds in
each of these groups are similar, comparisons have been
limited primarily to the role of substituent e€ects rather
than of structural or conformational variations per se.
The minimum requirements for an anti-CaM compound
appear to be an aromatic group and a cationic group.
The aromatic portion interacts with the hydrophobic
surfaces of CaM that are exposed upon the binding of
CaM, like other Ca2+-binding proteins, contains the
helix-coil-helix Ca2+-binding motif, the so-called EF
hand. The structure of Ca2+-CaM is an elongated
dumbbell in which two structurally similar domains are
connected by a helical linker region,12 with each domain
containing two EF hands (i.e. 4 Ca2+ sites per protein
molecule). LaPorte et al.13 found that the binding of
calcium to calmodulin exposes a hydrophobic surface
Key words: Antagonists; antineoplastics; antiproliferative agents;
proteins.
* Corresponding author. Tel.: +1-205-934-5696; fax: +1-205-934-
6475; e-mail: ted@mozart.nmrcore.uab.edu
0968-0896/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(99)00092-9

1560
T. T. Sakai, N. R. Krishna / Bioorg. Med. Chem. 7 (1999) 1559±1565
Ca2+ to the protein, and the cationic group presumably
interacts with anionic side chains of CaM residues. It
is possible that the charged group merely aids in
solubilizing the compound; however, the observation
that ANS, an anionic aromatic compound, binds to CaM
without inhibiting the CaM stimulation of cyclic-AMP
phosphodiesterase activity13 indicates that association
with CaM alone is not enough to cause an anti-CaM e€ect
and that the nature of the ionic group is important.
The syntheses of the candidate inhibitors is fairly
straightforward and uses literature techniques18 as well
as methods developed in this laboratory.19,20 These
compounds, in fact, constitute an extension and expan-
sion into protein inhibitors using techniques initially
used in the preparation of nucleic acid-binding
ligands.19,20 The substituted thiazoles are prepared by
the Hantzsch method in which the appropriate thioa-
mide and a-haloketone are condensed18 (Scheme 1(a)).
The cationic side chain is introduced by aminolysis of
the desired ester with 3-(methylthio)propylamine
(Scheme 1(b), followed by methylation with iodomethane
(Scheme 1(c)). For ease of identi®cation, compound 11 is
called pDPT (para-diphenylthiazole), compound 12 is
called mDPT (meta-diphenylthiazole) and compound 13
is called YDPT (`Y-shaped' diphenylthiazole).
As part of an on-going NMR-based drug design study,
we have prepared several lead compounds as possible
inhibitors of calmodulin activity and are studying the
three-dimensional structures of their complexes with
calmodulin. We have monitored the e€ects of these
compounds on the stimulation by CaM of cyclic-AMP
phosphodiesterase activity to measure their anti-CaM
e€ects, as well as their e€ects on tumor cells in culture.
This report presents an evaluation of some of the pro-
perties of these inhibitors. The detailed information
obtained from these studies will aid in the design and
synthesis of modi®ed derivatives with enhanced activity
against calmodulin function and tumor cell viability.
Results and Discussion
We have initially used substituted thiazoles as the basis
of our inhibitors for several reasons: (1) The `¯exible'
aromatic system provided by the ability of pendant
Scheme 1. (a) Reactions to form thiazole ring systems, (b) aminolysis reactions forming methylthio ethers, and (c) methylation reactions forming
sulfonium derivatives.

T. T. Sakai, N. R. Krishna / Bioorg. Med. Chem. 7 (1999) 1559±1565
1561
phenyl rings to rotate relative to the thiazole ring allows
the inhibitor to more readily accommodate variations in
the hydrophobic surface of CaM. This should allow
greater speci®city and anity in derivatives which take
advantage of this ¯exibility, particularly when com-
pared to derivatives which have rigid, fused ring systems
(e.g. phenothiazine, naphthalene). (2) The unambiguous
positioning of substituents about the thiazole ring
inherent in the chemistry of thiazole ring formation
makes design of derivatives easier. (3) The synthesis of
such ring systems is usually straightforward and a con-
siderable number of possible aromatic substituents are
possible. (4) In addition, we are hopeful that the natural
occurrence of thiazoles in biomolecules may allow for
more ready disposition of metabolic by-products of the
inhibitors, hopefully minimizing toxicities which might
arise from such metabolites. Similarly, the dimethylsul-
fonium side chain should be metabolized into non-toxic
products.
noticeably less e€ective than pDPT and mDPT in inhi-
biting the phosphodiesterase activity. Under the assay
conditions no conversion of cyclic-AMP to AMP was
observed for the ligands alone (no PDE or CaM), for
ligand plus and CaM (no PDE), or for ligand plus PDE
(no CaM), showing the inhibition is caused by the
ligand±CaM combination.
Stoichiometry and binding constants
The association of mDPT with CaM was monitored by
the changes in the absorption spectrum of the ligand as
the CaM-to-ligand ratio was changed gradually at a
®xed ligand concentration (0.05 mM). The spectral
intensity changes at 330 nm showed a simple satura-
tion curve (Fig. 2), and Scatchard analysis of the data
indicated a 1:1 stoichiometry with an association con-
stant of 1.15Â105 M (Fig. 3). The spectrophotometric
behavior of pDPT is more complex: there is an initial
decrease in ligand absorbance (monitored at 336 nm)
(for CaM-to-ligand ratios less than about 1); subse-
quently, there is a gradual increase which nears satura-
tion but has not saturated at ratios of about 4 or 5. This
is suggestive of a degree of non-speci®c binding with
CaM. Interestingly, this di€erence appears to be the
result of a simple change in the location of the cationic
side chain relative to the initial phenyl ring. The spectral
changes observed upon binding of the YDPT to CaM
are similar to those observed with pDPT in that there is
an initial decrease in ligand absorbance (for ligand-to-
CaM ratios <1), however, this is followed by a slow
linear increase in absorbance, with no indication of
saturation at a CaM-to-ligand ratio of 5 or 6. Again,
this di€erent behavior of YDPT compared to mDPT
arises from a relocation of the cationic side chain rela-
tive to the orientation of the aromatic system. We have
not analyzed the binding of either pDPT or YDPT any
further at this time.
The sulfonium side chain was used initially in previous
work done on DNA binding ligands, and the derivatives
described herein continue its use, in part for the reason
indicated above. Moreover, our feeling at this point is
that while a cationic group is necessary for the anti-
CaM properties of a ligand, the nature of the cationic
group may not be as critical as how it is disposed in the
complex. However, re®nements in ligand structure may
lead to situations in which speci®c features of the
charged group might be altered minutely to provide
enhanced binding of the ligand to CaM.
Anti-CaM e€ects
The three diphenylthiazole (DPT) derivatives inhibit the
calmodulin activation of cyclic-AMP phosphodiesterase
(PDE) to varying degrees (Fig. 1): at 0.1 mM, pDPT
and mDPT inhibit the conversion of cyclic-AMP to
AMP by about 70±75%; under the same conditions, the
known CaM antagonist tri¯uoperazine (TFP) shows
about 95% inhibition. The `Y-shaped' DPT (YDPT) is
The relative ecacies of our derivatives on CaM stimu-
lation of phosphodiesterase activity compared to TFP
presumably re¯ect the relative anities for CaM
1
1
(Kaꢀ1.15Â105 M
for mDPT versus ꢀ106 M
for
TFP21). These data indicate that relatively small chan-
ges in the structure of the ligand molecule can have
relatively large e€ects on its properties, changing the
binding properties from fairly speci®c (in the case of
mDPT) to less speci®c (pDPT) to relatively nonspeci®c
(YDPT).
Cytotoxicity of inhibitors
Preliminary measurements of the e€ects of these com-
pounds on cell viability were performed on L1210 mur-
ine leukemia cells growing in spinner culture. All
compounds showed modest cytotoxicity as measured
by dye exclusion, with about 25% inhibition of viable
cell number occurring at about 5 mM for mDPT and
5±10 mM for pDPT and YDPT. More detailed evalua-
tions of cytotoxic e€ects are being conducted.
Figure 1. E€ects of pDPT, mDPT and YDPT on the calmodulin-sti-
mulation of cyclic-AMP phosphodiesterase. Control (no additions)
is taken as 100%. The e€ect of tri¯uoperazine (TFP) is shown for
comparison.
While most of the calmodulin antagonists (particu-
larly the phenothiazines) exhibit cytotoxicity against a

1562
T. T. Sakai, N. R. Krishna / Bioorg. Med. Chem. 7 (1999) 1559±1565
Figure 2. Changes in absorbance at 335 nm of mDPT (0.05 mM) upon addition of aliquots of a solution of calmodulin (5 mM) containing 0.05 mM
mDPT.
Figure 3. Scatchard analysis of data in Figure 2.
variety of cells, the overall similarity of structures and
properties makes meaningful evaluation of structure±
activity relationships very dicult. We have begun to
develop some new inhibitors of calmodulin function
which will provide the basis for a more detailed
study of structure±activity relationships in calmodulin
antagonists. The data obtained in this study indicate
that relatively small changes in the placement of sub-
stitutents in this group of compounds have noticeable
e€ects on their binding to CaM as well as on their anti-
CaM e€ects. Lesser perturbations are noticeable on the
toxicities. We are hopeful that these di€erences will be

T. T. Sakai, N. R. Krishna / Bioorg. Med. Chem. 7 (1999) 1559±1565
1563
evident in the three-dimensional structures of com-
plexes with Ca2+±CaM and, thus, provide a starting
point for subsequent alterations in the structure of the
inhibitor.
Methyl 3-thiocarbamoylbenzoate (4). Dimethyl iso-
phthalate (Aldrich) was converted to methyl 3-cyano-
benzoate in 4 steps: (i) hydrolysis to methyl hydrogen
isophthalate with aqueous hydrochloric acid; (ii) con-
version to the acid chloride with thionyl chloride; (iii)
amidation to methyl isophthalamate with concentrated
ammonium hydroxide; and (iv) dehydration of the
amide with phosphorus pentoxide. The nitrile (9.21 g,
57.1 mmol) was dissolved in 35 mL of anhydrous N,N-
dimethylformamide and the solution was heated to 65±
70ꢁC; H2S was slowly bubbled through the solution and
1.0 mL of diethylamine was added. After 1 h, the reac-
tion mixture was poured into 250 mL of water and the
precipitate which formed was collected and dried in
vacuo, yield 9.8 g (88%), mp 133±135ꢁC. A small por-
tion was recrystallized from ethanol to give an analy-
tical sample, mp 137±138.5ꢁC. NMR (CDCl3) d 8.44 (s,
1, phenyl C2H), 8.18 (d, 2, phenyl C4H, C6H), 7.52 (t, 1,
phenyl C5H), 6.46 (s, 2, NH2), 3.95 (s, 3, OCH3) ppm.
Anal. calcd for C9H9NO2S (195.24): C, 55.37; H, 4.65;
N, 7.17. Found: C, 55.44; H, 4.67; N, 7.15.
Experimental
All materials were obtained from Aldrich (Milwaukee,
WI), Sigma (St. Louis, MO) or Fisher Scienti®c (Nor-
cross, GA) unless otherwise noted. Melting points were
determined on a Laboratory Devices Mel-Temp appa-
1
ratus. H NMR measurements were made on a Bruker
WH-400 spectrometer with chemical shifts referenced to
internal tetramethylsilane (organic solvents) or internal
sodium 4, 4-dimethyl-2,2,3,3-tetradeuterio-4-silapent-
anoate (aqueous solvents). Thin-layer chromatography
(TLC) was run on silica gel GF254 plates. All analytical
samples were homogenous upon TLC with ethyl acetate,
ethyl acetate:petroleum ether (bp 35±60ꢁC) (4:1, v/v) or
chloroform:methanol (3:1, v/v) as solvent systems.
Combustion analyses were performed by Galbraith
Laboratories (Knoxville, TN). All compounds gave
combustion analyses and/or NMR spectra consistent
with the indicated structures. Recombinant Drosophila
melanogaster calmodulin was expressed in Escherchia
coli cells received from Dr. Kate Beckingham (Rice
University) and puri®ed using procedures obtained
from her.
Methyl 4-[(4-phenyl)thiazol-2-yl]benzoate (5). A solution
of 3 (0.83 g, 4.2 mmol) and 2-bromoacetophenone
(0.83 g, 4.1 mmol) in 2 mL of anhydrous N,N-dimeth-
ylformamide was heated at 60±65ꢁC for 2 h. The crys-
tals that formed upon cooling the reaction solution
were collected by ®ltration, washed with water and
dried in vacuo; yield 1.04 g, mp 156±158ꢁC. Concentra-
tion of the ®ltrate provided an additional 0.20 g, mp
154±156ꢁC; total yield 97%. A portion was crystallized
from ethanol to give an analytical sample, mp 156.5±
Ethyl thiocarbamoylmethanoate (1). A solution of ethyl
cyanomethanoate (31.4 g, 0.32 mol) (Aldrich) in 50 mL
of benzene was cooled to 10ꢁC in a stainless steel
reaction vessel and saturated with H2S. Diethylamine
(0.5 mL) was added and the addition of H2S continued
at a rate to maintain the temperature of the solution
between 25 and 30ꢁC. After approximately 4 h, the
solution was cooled to 5ꢁC and again saturated with
H2S. The vessel was sealed and allowed to stand for 18 h
at room temperature. The crystals which formed were
collected and washed with carbon tetrachloride; yield
24.5 g (58%), mp 60±63.5ꢁC.
157ꢁC. NMR (CDCl3) d 8.13 (s, 4, phenyl C2H, C3H,
0
thiazole C5H), 7.46 (t, 2, phenyl C3 ,5 H s), 7.37 (d, 1,
0
C5H, C6H), 8.00 (d, 2, phenyl C2 ,6 H 's), 7.56 (s, 1,
0
0
phenyl C4 H) ppm. Anal. calcd for C17H13NO2S (295.36):
0
0
0
C, 69.13; H, 4.44; N, 4.74. Found: C, 69.06; H, 4.46;
N, 4.73.
Methyl 3-[(4-phenyl)thiazol-2-yl]benzoate (6). A solution
of 4 (0.95 g, 4.9 mmol) in 1 mL of anhydrous N,N-
dimethylformamide was added to a solution of 2-bro-
moacetophenone (0.95 g, 4.8 mmol) in 1 mL of the same
solvent. The solution was heated at 60±65ꢁC for 2 h and
then cooled to room temperature. The solution was
poured into 25 mL of water and the resulting solid was
®ltered, washed with water and dried in vacuo, yield
1.17 g (82%), mp 80±81ꢁC. NMR (CDCl3) d 8.66 (s, 1,
phenyl C2H), 8.27 (d, 1, phenyl C06H), 8.11 (d, 1, phenyl
ꢀ-Bromodeoxybenzoin (ꢀ-Bromobenzylphenyl ketone)
(2). Deoxybenzoin (19.6 g, 0.1 mol) (Aldrich) was bro-
minated with bromine in acetic acid as described for 4-
bromoacetophenone22 in 95% yield, mp 55±57ꢁC. This
material was used without further puri®cation.
0 0
C4H), 8.01 (d, 2, phenyl C2 ,6 H s), 7.56 (t, 1, phenyl
Methyl 4-thiocarbamoylbenzoate (3). Methyl 4-cyano-
benzoate (5.0 g, 31.0 mmol) (Aldrich) was dissolved in
25 mL of anhydrous N,N-dimethylformamide and the
resulting solution was saturated with H2S. The solution
was brought to 60±65ꢁC, 0.1 mL of diethylamine was
added and heating was continued for 1 h. The solution
was poured into a mixture of 100 g ice and 100 mL
water with vigorous stirring. The resulting solid was
collected and recrystallized from ethanol, total yield
5.6 g (93%), mp 189±191ꢁC. NMR (CDCl3) d 8.06 (d, 2,
phenyl C2,6 H0s), 7.90 (d, 2, phenyl C3,5H0s), 7.69 (broad
s, 1, NH?), 7.21 (broad s, 1, NH?), 3.95 (s, 3, OCH3).
Anal. calcd for C9H9NO2S (195.24): C, 55.37; H, 4.65;
N, 7.17. Found: C, 55.43; H, 4.69; N, 7.14.
0
C5H), 7.52 (s, 1, thiazole C5H), 7.46 (t, 2, phenyl C3 ,5
0
0
0
H s), 7.37 (d, 2, phenyl C4 H), 3.98 (s, 3, OCH3) ppm.
Anal. calcd for C17H13NO2S (295.36): C, 69.13; H, 4.44,
N, 4.74. Found: C, 69.05; H, 4.45; N, 4.75.
Ethyl 4,5-diphenylthiazole-2-carboxylate (7) was pre-
pared from 1 and 2 as described for 6 except that the
product was extracted into dichloromethane. Removal
of the solvent provided the product as an oil, yield
40%; recrystallization from ethanol gives a low melting
solid (mp <10ꢁC) which was used without further
puri®cation. NMR (CDCl3) d 7.51±7.53 (m, 4, 4- and 5-
phenyl C2,6H0s), 7.33±7.37 (m, 2, 4- and 5-phenyl

1564
T. T. Sakai, N. R. Krishna / Bioorg. Med. Chem. 7 (1999) 1559±1565
C4H0s), 7.27±7.30 (m, 4, 4- and 5-phenyl C3,5H0s), 4.51
(q, 2, OCH2), 1.45 (t, 3, CH3) ppm.
C4 H), 3.54 (t, 2, NCH2), 3.38 (t, 2, SCH2), 2.94 (s, 6,
S(CH3)2), 2.14 (p, 2, internal CH2) ppm. Anal. calcd for
0
.
C21H23ClN2OS2 0.75 H2O (432.52): C, 58.32; H, 5.71;
N, 6.48. Found: C, 58.28; H, 5.72; N, 6.42.
General procedure for the preparation of 3-(methylthio)-
propylamides. The appropriate ester (usually 2±5 mmol)
is dissolved in 2-5 mL of 3-(methylthio)propylamine
(Eastman) and the solution heated at 70±75ꢁC for 2±3 h.
The solution is allowed to cool to room temperature
and then diluted to 50 mL with dichloromethane. The
organic solution is extracted with N HCl (3Â50 mL) and
water (3Â50 mL) and dried with anhydrous magnesium
sulfate. Evaporation of the solvent in vacuo provides
the crude product.
Dimethyl 3-{3-[(4-phenyl)thiazol-2-yl]benzamido}propyl-
sulfonium chloride (12). Yield 72%, amorphous. NMR
(D2O) 8.29 (s, 1, phenyl C2H), 8.16 (d, 1, phenyl C6H),
0
0
0
7.95 (d, 2, phenyl C2 ,6 H s), 7.88 (d, 1, phenyl C4H),
7.85 (s, 1, thiazol0e C5H), 7.66 (t, 1, phenyl C5H), 7.56 (t,
0
0
0
2, phenyl C3 ,5 H s), 7.49 (d, 1, phenyl C4 H), 3.60 (t, 2,
NCH2), 3.41 (t, 2, SCH2), 2.95 (s, 6, S(CH3)2), 2.19 (p,
2, internal CH2) ppm. Anal. calcd for C21H23ClN2OS2
.
H2O (437.02): C, 57.72; H, 5.77; N, 6.41. Found: C,
57.44, H, 5.85; N, 6.37.
Methyl 3-{[4-(4-phenyl)thiazol-2-yl]benzamido}propyl
sul®de (8). Yield 98%; crystallized from ethanol, mp
143±144ꢁC. NMR (CDC0l3) 8.12 (d, 2, phenyl C2,6H0s),
Dimethyl 3-(4,5-diphenylthiazole-2-carboxamido)propyl-
sulfonium chloride (13). Yield 72%, amorphous
hydrate. NMR (D2O) 7.49±7.52 (m, 4, 4- and 5-phenyl
C2,6H0s), 7.39±7.43 (m, 6, 4- and 5-phenylC3,4,5H0s), 3.63
(t, 2, NCH2), 3.41 (t, 2, SCH2), 2.92 (s, 6, S(CH3)2), 2.19
(p, 2, internal CH2) ppm.
8.00 (d, 2, phenyl C2 ,6 H s), 7.87 (d, 2, phenyl C3,5H00s),
0
0
0
0
7.54 (s, 1, thiazole C5H), 7.46 (t, 2, phenyl C3 ,5 H s),
0
7.37 (d, 1, phenyl C4 H), 6.56 (broad t, 1, NH), 3.62 (q,
2, NCH2), 2.65 (t, 2, SCH2), 2.14 (s, 3, SCH3), 1.99 (p,
2, internal CH2) ppm. Anal. calcd for C20H20N2OS2
(368.52): C, 65.18; H, 5.47; N, 7.60. Found: C, 65.00; H,
5.49; N, 7.52.
Inhibition of CaM-dependent cyclic-AMP phospho-
diesterase activity. The conversion of cyclic-AMP to
AMP by calmodulin-activated bovine brain cyclic-AMP
phosphodiesterase was monitored by TLC of reaction
aliquots on silica gel GF254 and visualizing the chroma-
togram under low wavelength UV light. Typically,
reaction mixtures of 100 mL contained 0.002±0.005 unit
of phosphodiesterase, 0.3 nmol Ca2+±CaM, 5 nmol
CaCl2, 10 nmol of cyclic-AMP, and 100 nmol of ligand
(where used), in a bu€er of 0.01 M HEPES (pH 7.5).
After incubation for various times (0.5 to 1.0 h), the
reaction mixtures were treated with an equal volume of
ice-cold isopropanol and centrifuged. The supernatant is
taken to dryness on a Speed-Vac and the residue was
dissolved in 10 mL of water followed by 10 mL ethanol.
Aliquots of this mixture were applied to silica gel GF254
plates and the applied material completely dried under a
stream of nitrogen gas. Using the solvent system acet-
onitrile:isopropanol:0.1 M ammonium acetate (pH 7.4):
concentrated ammonia: 6:1:2:1, there is a clear separa-
tion of cyclic-AMP (Rf = 0.55) and AMP (Rf = 0.05±
0.10).23 Quantitation is carried out by scraping the
spots, eluting with 50% aqueous ethanol and measuring
the absorbancies using a molar extinction coecient of
12Â103 M 1 cm 1 24 at 262 nm.
Methyl 3-{[3-(4-phenyl)thiazol-2-yl]benzamido}propyl
sul®de (9). Yield 78%; recrystallized from ethanol, mp
125±126ꢁC. NMR (CDCl3) 8.50 (s, 1, phenyl0C2H), 8.16
0
0
(d, 1, phenyl C6H), 8.00 (d, 2, phenyl C2 ,6 H s), 7.87 (d,
1, phenyl C4H), 7.55 (t, 1, phenyl C5H), 7.52 (s, 1, thia-
0
C4 H), 6.67 (broad t, 1, NH), 3.70 (q, 2, NCH2), 2.65 (t,
0
0
zole C5H), 7.46 (t, 2, phenyl C3 ,5 H s), 7.37 (d, 1, phenyl
0
2, SCH2), 2.15 (s, 3, SCH3), 1.99 (p, 2, internal CH2)
ppm. Anal. calcd for C20H20N2OS2 (368.52): C, 65.18;
H, 5.47; N, 7.60. Found: C, 65.23; H, 5.48; N, 7.56.
Methyl 3-(4,5-diphenylthiazole-2-carboxamido)propyl
sul®de (10). Yield 82%, mp 118±119ꢁC after recrystalli-
zation from ethanol. NMR (CDCl3) 7.48±7.51 (m, 4, 4-
and 5-phenyl C2,6H0s), 7.33±7.36 (m, 2, 4-and 5-phenyl
C4H0s), 7.30±7.32 (m, 4, 4- and 5-phenyl C3,5H0s), 3.60
(q, 2, NCH2), 2.61 (t, 2, SCH2), 2.13 (s, 3, SCH3), 1.96
(p, internal CH2) ppm.
General procedure for the preparation of dimethyl-
sulfonium derivatives. The appropriate methyl sul®de
(1±2 mmol) is dissolved in 5 mL of methanol:iodo-
methane (1:1 by volume). The solution is allowed to
stand in the dark for 24 h at room temperature.
Removal of the solvent in a stream of dry nitrogen
a€ords the crude product as the iodide salt. Generally,
the iodide derivatives proved intractable due to dis-
coloration or decomposition and they were converted to
the chloride derivatives by passage of an aqueous solu-
tion through a column (1Â8 cm) of Dowex 1Â8 (chlo-
ride form). Lyophilization of the eluate provides the
chloride derivative.
Cell culture. Murine leukemia cells (L1210) were main-
tained at 37ꢁC in spinner ¯asks in Fischer's medium
containing 10% newborn bovine serum and equilibrated
with a hydrated atmosphere containing 5% CO2. Cul-
tures were exposed to various concentrations of the
compounds being tested and cell viability was mon-
itored by the ability of cells to exclude Trypan Blue.
Cells were counted using a hemacytometer.
Dimethyl 3-{4-[(4-phenyl)thiazol-2-yl]benzamido}propyl-
sulfonium chloride (11). Yield 86%, amorphous. NMR
(D2O) 08.00 (d, 2, phenyl C2,6H0s), 7.90 (d, 2, phenyl
Acknowledgements
0
0
C2 ,6 H s), 7.84 (s, 1, thiazole C5H), 7.80 (d, 2, phenyl
This work is supported by USPHS Grant CA-13148
from the National Cancer Institute. We thank Dr. Kate
C3,5H0s), 7.53 (t, 2, phenyl C3 ,5 H s), 7.46 (d, 1, phenyl
0
0
0

T. T. Sakai, N. R. Krishna / Bioorg. Med. Chem. 7 (1999) 1559±1565
1565
Beckingham for providing the calmodulin-producing
Escherichia coli strain.
13. LaPorte, D. C.; Wierman, B. M.; Storm, D. R. Biochem-
istry 1980, 19, 3814.
14. Motohashi, N. Anticancer Res. 1991, 11, 1125.
15. MacNeil, S.; Wagner, M.; Rennie, I. G. Clin. Exp. Metas-
tasis 1994, 12, 375.
References
16. Renondon, A.; Boucher, J. L.; Sari, M. A.; Delaforge, M.;
Ouazzani, J.; Mansuy, D. Biochem. Pharmacol. 1997, 54, 1109.
17. Craven, C. J.; Whitehead, B.; Jones, S. K. A.; Thulin, E.;
Blackburn, G. M.; Waltho, J. P. Biochemistry 1996, 35, 10287.
18. Metzger, J. V. In Thiazole and its Derivatives, Parts 1±3.
The Chemistry of Heterocyclic Compounds; Weissberger, A.,
Taylor, E. C., Eds.; John Wiley and Sons: New York, 1979,
Vol. 34.
1. Cheung, W. Y. Harvey Lect. 1985, 79, 173.
2. Vogel, H. J. Biochem. Cellul. Biol. 1994, 72, 357.
3. Means, A. R.; VanBerkum, M. F. A.; Bagchi, I.; Lu, K. P.;
Rasmussen, C. D. Pharmacol. Ther. 1991, 50, 255.
4. Rasmussen, C. D.; Lu, K. P.; Means, R. L.; Means, A. R. J.
Physiol. 1992, 86, 83.
5. Lu, K. P.; Means, A. R. Endocrine Rev. 1993, 14, 40.
6. Bachs, O.; Agell, N.; Carafoli, E. Biochim. Biophys. Acta
1992, 1113, 259.
19. Riordan, J. M.; Sakai, T. T. J. Heterocycl. Chem. 1981, 18,
1213.
7. Bachs, O.; Agell, N.; Carafoli, E. Cell Calcium 1994, 16, 289.
8. Wei, J.-W.; Hickie, R. A. Biochem. Biophys. Res. Commun.
1981, 100, 1562.
20. Riordan, J. M.; Sakai, T. T. J. Med. Chem. 1983, 26, 884.
21. Klevit, R. E.; Levine, B. A.; Williams, R. J. P. FEBS Lett.
1981, 123, 25.
9. Wei, J.-W.; Morris, H. P.; Hickie, R. A. Cancer Res. 1982,
42, 2571.
22. Furniss, B. S., Hannaford, A. J., Smith, P. W. G., Tatchell,
A. R. Vogel's Textbook of Practical Organic Chemistry, 5th
ed., Longman: Harlow (UK), 1990, pp 1052±1053.
23. Tomashefski, J. F., Jr; Barrios, R. J.; Sudilovsky, O. Ana-
lyt. Biochem. 1974, 60, 589.
10. Hickie, R. A.; Wei, J.-W.; Blyth, L. M.; Wong, D. Y. W.
Canad. J. Biochem. Cellul. Biol. 1983, 61, 934.
11. Hait, W. N.; Lazo, J. S. J. Clin. Oncol. 1986, 4, 994.
12. Babu, Y.; Bugg, C. E.; Cook, W. J. J. Molec. Biol. 1988,
204, 191.
24. Bergmeyer, H. U., Ed. Methods of Enzymatic Analysis, 2nd
English ed., Verlag±Chemie: Weinheim, 1974, Vol. 1, p 526.