Structure-based design of substituted diphenyl sulfones and sulfoxides as lipophilic inhibitors of thymidylate synthase

Article abstract of DOI:10.1021/jm960613f

Six new diphenyl sulfoxide and five new diphenyl sulfones were designed, synthesized, and tested for their inhibition of human and Escherichia coli thymidylate synthase (TS) and of the growth of cells in tissue culture. The best sulfoxide inhibitor of hum

Full text of DOI:10.1021/jm960613f

J . Med. Chem. 1997, 40, 677-683
677
Str u ctu r e-Ba sed Design of Su bstitu ted Dip h en yl Su lfon es a n d Su lfoxid es a s
Lip op h ilic In h ibitor s of Th ym id yla te Syn th a se
Terence R. J ones,*,† Stephen E. Webber, Michael D. Varney, M. Rami Reddy,‡ Kathleen K. Lewis,
Vinit Kathardekar, Hormoz Mazdiyasni, J udith Deal, Dzuy Nguyen, Katharine M. Welsh, Stephanie Webber,
Amanda J ohnston, David A. Matthews, Ward W. Smith, Cheryl A. J anson, Russell J . Bacquet,§
Eleanor F. Howland, Carol L. J . Booth, Steven M. Herrmann, Robert W. Ward, J ennifer White,
Charlotte A. Bartlett, and Cathy A. Morse
Agouron Pharmaceuticals, Inc., 3565 General Atomics Court, San Diego, California 92121
Received August 26, 1996X
Six new diphenyl sulfoxide and five new diphenyl sulfones were designed, synthesized, and
tested for their inhibition of human and Escherichia coli thymidylate synthase (TS) and of the
growth of cells in tissue culture. The best sulfoxide inhibitor of human TS was 3-chloro-N-
((3,4-dihydro-2-methyl-4-oxo-6-quinazolinyl)methyl)-4-(phenylsulfinyl)-N-(prop-2-ynyl)-
aniline (7c) that had a Ki of 27 nM. No sulfone improved on TS inhibition by the previously
reported 4-(N-((3,4-dihydro-2-methyl-6-quinazolinyl)methyl)-N-prop-2-ynylamino)phenyl phenyl
sulfone (Ki ) 12 nM). Nevertheless, one sulfone, 4-((2-chlorophenyl)sulfonyl)-N-((3,4-dihydro-
2-methyl-4-oxo-6-quinazolinyl)methyl)-N-(prop-2-ynyl)aniline, was selected, on the basis of its
inhibition of both TS and cell growth, for antitumor testing; it gave a 61% increase in life span
to mice bearing the thymidine kinase-deficient L5178Y (TK-) lymphoma. A crystal structure
of N-((3,4-dihydro-2-methyl-4-oxo-6-quinazolinyl)methyl)-4-((2-methylphenyl)sulfinyl)-N-(prop-
2-ynyl)aniline complexed with E. coli TS was solved and revealed selective binding of one
sulfoxide enantiomer. AMBER calculations showed that the enantioselection was due to
asymmetric electrostatic effects at the mouth of the active site. In contrast, a similar crystal
structure of the sulfoxide 7c, along with AMBER calculations, indicated that both enantiomers
bound, but with different affinities. The side chain of Phe176 shifted in order to structurally
accommodate the chlorine of the more weakly bound enantiomer.
The first papers in this series1,2 described the devel-
opment of the quinazoline-based diphenyl sulfone 8a
(Scheme 1) as a 12 nM lipophilic inhibitor of the enzyme
thymidylate synthase. This enzyme is crucially involved
in DNA synthesis, and inhibitors of it have potential
as cytotoxic therapeutics. Because some simple meta-
substituted propargylanilines had enhanced the inhibi-
tion of thymidylate synthase (TS), we synthesized and
tested a meta-substituted analogue of 8a (8d , Table 1),
but it inhibited TS no better than 8a .1,2 In this paper
we further explore structure-activity relationships in
the diphenyl sulfone series and describe new substituted
sulfones and also, given the flexibility of the synthetic
route used, the corresponding series of diphenyl sulfox-
ides.
Therefore the fluoro and chloro compounds 8b and 8c,
respectively, were modeled and minimized by AM1, and
their low-energy structures were docked and compared
with the X-ray-derived structure of 8a in the mouth of
the enzyme active site. Both compounds fitted well with
the substituents, neatly filling the meta pocket. Earlier
consideration had been given to substituents rooting
from the distal phenyl ring. From a close examination
of the three-dimensional structure of E. coli TS3 (with
10-propargyl-5,8-dideazafolic acid removed and with 8a
modeled in its place since this was before the ternary
complex of 8a had actually been solved) specifically in
the region where the SO2 group linked the distal phenyl
ring, there was seen an unoccupied pocket created by
the ring of Phe176, the side chain of Ile79, and the
methylene of His51. This small region could be accessed
from the ortho position of the distal phenyl ring of 8a ,
and it was found that either a methyl group or a chlorine
had the optimum steric and hydrophobic requirements
for van der Waals interactions. AM1 minimization
calculations were performed on compounds 8e and 8f,
and their low-energy structures were docked into the
enzyme active site and compared with the model of the
parent diphenyl sulfone 8a . Only slight differences
were observed between the two sets of C-SO2-C torsion
angles. It should be noted that in each case these ortho
substituents could adopt an alternate configuration with
respect to the enzyme. A rotation of 180° about the
SO2-distal phenyl bond would place the methyl or
chlorine toward the side chain of Ala262. The disub-
stituted compound 8g with inner CF3 and outer CH3
substituents had been successfully modeled by AM1 as
Design of In h ibitor s
The report1,2 of the X-ray crystal structure of the
trifluoromethyl-substituted compound 8d bound to Es-
cherichia coli TS revealed that, relative to the unsub-
stituted parent compound, the entire diphenyl sulfone
moiety had moved upward toward residues Phe176 and
His51, a movement that might possibly explain the
unimproved inhibition observed for 8d . It seemed
possible that the CF3 substituent might not be fitting
properly into the so-called meta pocket1,2 and that
smaller substituents X might show some advantage.
* To whom correspondence should be addressed.
† Present address: A° ngstrom Pharmaceuticals, Inc., 11772 Sorrento
Valley Rd., #101, San Diego, CA 92121.
‡ Present address: Gensia, Inc., 9360 Towne Center Drive, San
Diego, CA 92121.
§ Deceased.
X Abstract published in Advance ACS Abstracts, February 1, 1997.
S0022-2623(96)00613-9 CCC: $14.00 © 1997 American Chemical Society

678 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 5
J ones et al.
Sch em e 1
above and selected for synthesis but before the detri-
mental effect of the CF3 group in the monosubstituted
compound 8d became fully apparent from the X-ray
crystal structure.1
Biologica l Resu lts
The results for the inhibition of human and bacterial
TS, listed in Table 1, show that the substituents X or Y
generally induced minor changes in the Ki values, some
of which were improvements. However, the two disub-
stituted compounds 7g and 8g were clearly less potent
than their respective parent compounds. In the case of
8g against the human enzyme, this loss was by 1 order
of magnitude. Taking the sulfoxide series first, there
is seen a slight selectivity for inhibition of the human
enzyme across the whole series, parent compound 7a
included, and most observable for the methyl compound
7e. The most potent sulfoxide inhibitor for both en-
zymes was the chloro compound 7c, which, for the
human enzyme, had a Ki of 27 nM. In the sulfone
series, no compound improved upon the parent 8a in
inhibiting either enzyme, which was a disappointment.
Compound 8f with a distal chlorine substituent was the
second best inhibitor of the human enzyme, having a
Ki of 21 nM.
Ch em istr y
Six diphenyl sulfoxides, 7b-g, and five diphenyl
sulfones 8b,c,e-g, were synthesised as shown in Scheme
1. The sulfide 3, prepared in high yield from com-
mercially available starting materials, was reduced to
the corresponding substituted aniline 4. Succesive
alkylations, in good yields, with propargyl bromide and
6-(bromomethyl)-3,4-dihydro-2-methyl-4-oxoquinazo-
line (“QCH2Br”)4 rendered the key intermediate 6. By
careful selection of conditions, the sulfide 6 was oxidized
to the sulfoxide 7 and separately to the sulfone 8, each
isolated by flash chromatography in a high state of
purity. These new compounds were tested for their
inhibition of human and E. coli TS and for their
inhibition of the growth of three cell lines in tissue
culture using methods previously described.2 The re-
sults are shown in Table 1, which also contains our
previous results for the sulfone 8d and, as controls, the
parent compounds 7a and 8a .2
Turning to the tissue culture results, again structural
variation induced little variation in the IC50 values.
However, there were two exceptions to this generaliza-
tion. First, disubstituted compounds 7g and 8g were

Lipophilic Inhibitors of Thymidylate Synthase
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 5 679
Ta ble 1. Inhibition of Thymidylate Synthase and of Growth of Cells in Tissue Culture
Y
O
N
S[O]n
CH2
HN
CH3
X
N
Kia (µM)
human TSd
IC50b (µM)
thymidinec
shift
compd
X
Y
n
E. coli TSd
L1210
CCRF-CEM
GC3/M TK-
7a e
H
H
1
0.16 ( 0.08
0.55 ( 0.04
0.088 ( 0.042
0.29 ( 0.16
0.054 ( 0.018
0.13 ( 0.03
0.11 ( 0.06
0.16 ( 0.09
0.25 ( 0.11
0.78 ( 0.28
0.19 ( 0.09
0.76 ( 0.27
0.29 ( 0.1
0.92 ( 0.41
0.025 ( 0.006
0.068 ( 0.05
0.043 ( 0.019
0.074 ( 0.044
0.033 ( 0.021
0.043 ( 0.01
0.037 ( 0.019
0.040 ( 0.008
0.056 ( 0.025
0.16 ( 0.06
0.033 ( 0.009
0.069 ( 0.022
0.16 ( 0.08
0.39 ( 0.21
0.061 ( 0.023
0.11 ( 0.02
1.7
2.1
1.0
1.1
1.7
2.1
1.0
1.0
1.2
1.6
2.0
0.6
0.3
1.0
4.5
1.4
2.3
3.0
3.5
5.8
3.0
1.6
2.0
1.5
3.5
2.7
0.6
2.0
3.9
4.2
>5.3
>2.3
5.0
7b
7c
7d
7e
7f
F
H
1
1
1
1
1
1
2
2
2
2
2
2
2
0.050 ( 0.018
0.12 ( 0.06
Cl
CF3
H
H
0.027 ( 0.008
0.037 ( 0.016
0.070 ( 0.039
0.15 ( 0.06
3.8
H
3.3
2.9
CH3
Cl
0.045 ( 0.021
0.077 ( 0.016
0.046 ( 0.019
0.073 ( 0.025
0.26 ( 0.12
4.0
8.7
H
4.3
3.8
7g
8a e
8b
8c
8d e
8e
8f
CF3
H
CH3
H
0.39 ( 0.11
9.1
2.0
0.012 ( 0.005
0.021 ( 0.009
0.020 ( 0
1.03
2.8
18.9
>4.6
4.6
F
H
0.046 ( 0.013
0.024 ( 0.012
0.055 ( 0.018
0.050 ( 0.031
0.084 ( 0.001
0.026 ( 0.01
0.058 ( 0.013
0.021 ( 0.007
0.025 ( 0.012
0.15 ( 0.04
Cl
CF3
H
H
3.0
H
7.5
2.3
CH3
Cl
3.4
10.0
6.1
H
1.2
8g
CF3
CH3
0.38 ( 0.23
>5.0 (10%)
1.3
a
b
For each compound the upper value is Kis and the lower Kii. For entries of the form >x (y%), x is the limiting solubility of the
compound and y the percent inhibition at that concentration. c Expressed as the ratio of the IC50 for L1210 cells in the presence of 10 µM
thymidine divided by the IC50 for the same cell line with no thymidine added. Recombinant enzyme produced in E. coli. e Reference 2.
d
obviously less effective against the thymidine kinase-
deficient human adenocarcinoma cell line GC3/M TK-,
and second, the chloro sulfone 8f was 3-fold more
effective against each of the leukemic cell linessmurine
L1210 and human CCRF-CEM. In each case these
changes correlate qualitatively with changes in the
inhibition of human TS as noted above. Considering
both its inhibition of human TS and of the growth of
cells, the chloro sulfone 8f was clearly the best com-
pound in the series. Thymidine reversal studies on the
L1210 cell line were performed to investigate whether
the compounds were acting upon TS when inhibiting
growth. All compounds showed some degree of reversal.
In the sulfoxide series the thymidine shift approxi-
mately tracked the human Ki potency (and less the IC50),
although the best compound was 7e not 7c. In the
sulfone series, again the thymidine shift correlated more
closely with Ki rather than IC50; compound 8a was
clearly the most TS-directed. It is possible that the
compounds act elsewhere in the folate enzyme network
or even outside it; however, further mechanistic studies
were not pursued.
intraperitoneal injection for 10 days. The group treated
with compound 8f displayed a mean increase in life span
of more than 61% compared to the control-treated group.
Unfortunately, in vivo testing of this compound was
complicated by its poor aqueous solubility and its short
half-life of approximately 30 min in mice.
Cr ysta llogr a p h y a n d Com p u ta tion a l Ch em istr y
Two compounds only, 7e and 7c from the diphenyl
sulfoxide series, cocrystallized with E. coli TS suf-
ficiently well to be used for X-ray structural analysis.
Figure 1 shows the 2.5 Å crystal structure of the TS
complex with the methyl sulfoxide 7e. Electron density
for the sulfoxide oxygen is seen at a single location,
indicating that E. coli TS binds one of the two enanti-
omers preferentially. A second feature of the binding
of 7e is that the distal methyl substituent has taken
up the position pointing downward, not quite toward
Ala262, but in the recognized alternative direction
envisaged when it was modeled in the corresponding
sulfone 8e.
A clue to an explanation for this enantioselective
binding of 7e was obtained by considering the residues
in the active site. These are shown in Figure 1 with a
Connolly surface color coded for electrostatic potential
calculated using the DELPHI program. The negative
potential arising from Glu82, at the center-bottom of
the view, extends upward and to the left, offering a
possible destabilizing interaction to a sulfoxide oxygen
Compound 8f, noted above, was tested for in vivo
antitumor activity. The compound was active in an
antitumor model using mice inoculated with the thy-
midine kinase-deficient L5178Y (TK-) murine lym-
phoma. A variety of doses and schedules of adminis-
tration were tested. Optimal antitumor activity was
found with a dose of 50 mg/kg given twice daily by

680 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 5
J ones et al.
F igu r e 1. Stereopicture showing the experimentally determined structure of compound 7e complexed to E. coli TS. The active
site is outlined with a Connolly surface color coded to show electrostatic potential derived from DELPHI calculations. Surface of
positive potential is colored red; neutral, green; and negative, blue. Both ligand and protein are shown in pale green for carbon,
red for oxygen, blue for nitrogen, and yellow for sulfur. The oxygen of the sulfoxide is seen pointing toward the positive side chain
nitrogen of Lys48 at top left. The negative side chain of Glu82 is at bottom right.
Ta ble 2. Calculated Interaction Energies of Four Conformers
of Compound 7e with E. coli TS
stage, only hydrogens in the system were allowed to
relax by 100 steps of steepest descent followed by 1000
interaction energy of ligand
with protein (kcal/mol)
steps of conjugate gradient. This step relaxes the
hydrogen atoms prior to relaxing heavy atoms. It was
performed because the hydrogen locations are not
specified in the X-ray structure, and their adjustment
is necessary to improve hydrogen bond geometries. In
the third stage, only ligand and solvent water are moved
(100 steps steepest descent plus 1000 steps conjugate
gradient) by keeping the protein fixed. Finally, all the
atoms of the residues within 25 Å from the center of
the ligand are moved. This stage was performed using
100 steps of steepest descent and 5000 steps of conjugate
gradient methods.
In the minimized structure for each complex, the
interaction energy of the ligand with the protein atoms
was calculated. This energy was decomposed into
electrostatic and van der Waals contributions. These
results are given in Table 2. The calculations show that
the interaction energy of system I (-109.1 kcal/mol) is
lower than that of its cis conformer, system II (-104.5
kcal/mol), but both are much lower than the energy of
system III (-96.0 kcal/mol) or system IV (-98.2 kcal/
mol), together representing the other enantiomer. The
van der Waals contribution to the interaction energy
varies by only 3.3 kcal/mol across the four systems; the
systems are mostly differentiated by electrostatic en-
ergy. The 4.6 kcal/mol difference in interaction energy
between conformers I and II of the bound enantiomer
is not the only factor that favors binding of trans
conformer I since cis isomer II has a higher internal
energy (2.9 kcal by AMBER minimization).
conformation
of moleculeb
van der
Waals
systema
I
electrostatic
total
-64.8
-44.3
-109.1
O
R
S
CH3
II
-66.1
-62.8
-65.4
-38.4
-33.2
-32.8
-104.5
-96.0
-98.2
CH3
O
R
R
S
S
III
IV
O
CH3
CH3
R
S
O
a
b
System I is shown in Figure 1. Structure shows only the
diphenyl sulfoxide moiety.
with a partial negative charge at that position. Con-
versely, Lys48, and the surrounding ordered solvent
molecules, diametrically opposite in the site, offers a
possible stabilizing interaction for the oxygen of the
enantiomer that bound. This hypothesis was further
explored by conformational analysis of 7e in the active
site of the enzyme. Molecular mechanics calculations
(energy minimizations) on four different orientations of
the ligand (Table 2) were performed using the AMBER
force field5,6 to determine the relative conformational
preferences of the sulfoxide oxygen in the bound com-
plex. The details of the systems setup, the calculation
of force field parameters, and the partial atomic charges
used are similar to those published in our earlier paper.7
For the energy minimization of the protein inhibitor
complex, a four-stage protocol was used. In the first
stage, only the waters were minimized using 100 steps
of steepest descent followed by 1000 steps of conjugate
gradient methods, keeping the inhibitor and the protein
fixed. The purpose of this step was to relieve any bad
contacts in the initially solvated system. In the second
The most reasonable interpretation of these results
is that the most stable bound conformation (system I)
is characterized by the polar sulfoxide oxygen engaging
in a favorable electrostatic interaction with the posi-
tively charged residue Lys48. In contrast, oriented as
in systems II and III, the sulfoxide oxygen loses this
interaction and instead engages in an unfavorable
electrostatic interaction with the negatively charged
residue Glu82. This calculated lowest energy conforma-
tion (system I) for the ligand 7e is in agreement with
the observed experimental crystallographic result.
Enantiomeric selectivity by TS is apparently less
stringent for the corresponding meta chlorosulfoxide

Lipophilic Inhibitors of Thymidylate Synthase
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 5 681
F igu r e 2. Stereopicture of difference electron density indicating two distinct positions for the chlorine of compound 7c when
complexed with E. coli TS and 5-FdUMP. The map is calculated at 2.5 Å resolution using coefficients [F7c-F8a] where F7c and F8a
are observed structure amplitudes for compounds 7c and 8a , respectively. Model phases were used with no contribution from the
quinazoline inhibitor. The map is contoured at 8σ (red) and 12σ (yellow). The FdUMP molecule is visible below the quinazoline
ring of compound 7c.
analog 7c. Chlorine substitution in 7c is associated
with strong difference density peaks appearing adjacent
to both meta carbon atoms of the inhibitor’s proximal
phenyl ring (Figure 2). This indicates that either (i) a
single enantiomer binds but the proximal phenyl of
bound 7c can flip about its long axis placing Cl cis or
trans to the sulfoxide oxygen or (ii) both enantiomers
bind (probably with their Cl and sulfoxide oxygens
trans) but with different affinities. Since the X-ray
experiment returns an electron density averaged over
all populations in the crystal, we cannot eliminate the
possibility that both enantiomers bind in both confor-
mations. However, as for compound 7e (vide supra),
AMBER calculations indicate that the cis isomer has a
significantly higher internal energy, suggesting that the
two Cl peaks probably arise simply from binding both
trans enantiomers. 2F7c-Fc difference maps (data not
shown) reveal electron density for the sulfoxide oxygen
in two positions consistent with the notion that both
enantiomers of 7c are bound in the cocrystal. J udging
from the relative integrated peaks, the enantiomer with
the sulfoxide oxygen pointed toward Lys48 predomi-
nates roughly 3 to 1. Further inspection of the differ-
ence density indicated a shift in the side chain of Phe176
as a result of 30-50° rotations around ø1 and ø2 in order
to structurally accommodate the Cl of the more weakly
bound enantiomer and such that the side chain had
moved closer to the propargyl group of the inhibitor. No
change in the enzyme structure around Phe176 was
evident.
inhibition measured for these compounds offered no or
minor improvement over their unsubstituted parents.
The probable reason for this outcome is that the pockets
of the enzyme which the substituents were designed to
occupy, particularly those for the distal substituents,
were insufficiently defined and not deep enough to bury
enough hydrophobic surface area to give the improve-
ments in binding that were sought. The best compound,
considering also tissue culture results, was the chloro
sulfone 8f. This compound was advanced to in vivo
testing in lymphoma-bearing mice. It had definite
activity since at a dose of 50 mg/kg it increased the life
span of the mice by more than 61%. However, its poor
aqueous solubility combined with its disadvantageous
short half-life in mice of only 30 min deterred us from
investigating it further.
Two interesting results were obtained from the crys-
tallographic analysis. In the first there was observed
a selective binding to E. coli TS of one enantiomer of
the sulfoxide, 7e. The likely explanation for this
phenomenon is that the positively-charged Lys48 and
the negatively-charged Glu82 at the mouth of the active
site act in concert to favor the binding of one enantiomer
by electrostatic effects. Molecular mechanics calcula-
tions performed with the AMBER force field validated
this explanation. It cannot be extrapolated that such
enantioselective binding occurs with human TS because,
although Lys48 is conserved, the residue corresponding
to Glu82 is Ala in that enzyme. The second crystal
structure revealed that the side chain of Phe 176 had
rotated sideways and back into the enzyme to make
room for the meta chlorine substituent of the sulfoxide
7c to take partial occupancy by hydrophobic interaction.
In three previous crystal structures we presented with
Discu ssion
We have described in this paper a series of new
diphenyl sulfoxide and diphenyl sulfone inhibitors of the
enzyme thymidylate synthase having a substituent in
one or the other ring adjacent to the central sulfur and,
in one case, with substituents in each ring. The enzyme
1
the meta substituents CF3 and CH2OH,2 this phenom-
enon was not observed. Unlike 7e, both enantiomers
of 7c appeared to bind to TS. The clear result was that

682 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 5
J ones et al.
4-(P h en ylt h io)-N-(p r op -2-yn yl)-3-(t r iflu or om et h yl)-
a n ilin e (5d ): yellow oil (53%); NMR (Me2SO-d6). Anal.
(C16H12F3NS) C, H, N, F, S.
4-((2-Meth ylp h en yl)th io)-N-(p r op -2-yn yl)a n ilin e (5e):
oil (68%); NMR (CDCl3). Anal. (C16H15NS) C, H, N, S.
4-((2-Ch lor op h en yl)th io)-N-(p r op -2-yn yl)a n ilin e (5f):
oil (69%); NMR (CDCl3). Anal. (C15H12ClNS) C, H, N, Cl, S.
4-((2-M e t h y lp h e n y l)t h i o )-N -(p r o p -2-y n y l)-3-(t r i -
flu or om eth yl)a n ilin e (5g): oil (51%); NMR (CDCl3). Anal.
(C17H14F3NS‚0.25H2O) C, H, N, S.
Phe176 was a residue with a hitherto unrecognized
flexibility of movement and that the E. coli enzyme, and
by extension the human enzyme, was capable of accept-
ing trisubstituted compounds with substituents at posi-
tions 3, 4, and 5 relative to the N10 aniline nitrogen.
Thus the crystal structure of compound 7c bound to TS
immediately suggested the next iteration of design and
synthesis, and this will be described in a future paper.
Qu in a zolin e Su lfid es 6. Reactions were conducted and
the products purified by methods previously described.2
N-((3,4-Dih ydr o-2-m eth yl-4-oxo-6-qu in azolin yl)m eth yl)-
3-flu or o-4-(p h en ylth io)-N-(p r op -2-yn yl)a n ilin e (6b): white
solid (51%); mp 221-223 °C dec (CH3CN/CCl4); NMR (Me2-
SO-d6). Anal. (C25H20FN3OS) C, H, N, F, S.
3-Ch lor o-N-((3,4-d ih yd r o-2-m et h yl-4-oxo-6-q u in a zo-
lin yl)m eth yl)-4-(p h en ylth io)-N-(p r op -2-yn yl)a n ilin e (6c):
white solid (20%); mp 224 °C dec (CH3CN/CCl4); NMR (Me2-
SO-d6). Anal. (C25H20ClN3OS) C, H, N, Cl, S.
N-((3,4-Dih ydr o-2-m eth yl-4-oxo-6-qu in azolin yl)m eth yl)-
4-(p h e n y lt h io )-N -(p r o p -2-y n y l)-3-(t r iflu o r o m e t h y l)-
a n ilin e (6d ): off-white solid (23%); mp 228-230 °C; NMR
(CDCl3). Anal. (C26H20F3N3OS‚H2O) C, H, N.
N-((3,4-Dih ydr o-2-m eth yl-4-oxo-6-qu in azolin yl)m eth yl)-
4-((2-m eth ylp h en yl)th io)-N-(p r op -2-yn yl)a n ilin e (6e): off-
white solid (90%); mp 192-194 °C (CHCl3/MeOH); NMR
(CDCl3). Anal. (C26H23N3OS‚H2O) C, H, N.
4-((2-Ch lor op h en yl)t h io)-N-((3,4-d ih yd r o-2-m et h yl-4-
oxo-6-qu in a zolin yl)m eth yl)-N-(p r op -2-yn yl)a n ilin e (6f):
pale yellow solid (26%); mp 215-217 °C dec (CHCl3); NMR
(Me2SO-d6). Anal. (C25H20ClN3OS‚0.6H2O) C, H, N, Cl, S.
N-((3,4-Dih ydr o-2-m eth yl-4-oxo-6-qu in azolin yl)m eth yl)-
4-((2-m et h ylp h en yl)t h io)-N-(p r op -2-yn yl)-3-(t r iflu or o-
m eth yl)a n ilin e (6g): off-white solid (45%); mp 219-222 °C
(CHCl3/MeOH); NMR (Me2SO-d6). Anal. (C27H22F3N3OS) C,
H, N, S.
P r ep a r a tion of Su lfoxid es 7: Meth od A. The sulfide 6b
or 6c (y mmol) was dissolved in HOAc (100y mL) and treated
with NaBO3‚4H2O (y mmol). The mixture was stirred (6b/2
h/45 °C; 6c/23 h/60 °C) and then poured into saturated aqueous
NaHCO3. Extractive (CH2Cl2) workup followed by flash chro-
matography on silica using the eluant stated gave the pure
product. Meth od B. The sulfides 6d ,e,f,g (y mmol) were
dissolved in HOAc (20y mL) and treated with H2O2 (30%, 1.5y
mmol, 1.5 equiv). The mixture was stirred 16 h at 25 °C and
then poured into saturated aqueous NaHCO3. Extractive
(CHCl3) workup followed by flash chromatography on silica
using the eluant stated gave the pure product.
N-((3,4-Dih ydr o-2-m eth yl-4-oxo-6-qu in azolin yl)m eth yl)-
3-flu or o-4-(p h en ylsu lfin yl)-N-(p r op -2-yn yl)a n ilin e (7b):
eluant, toluene 50%-CH3CN 48%-MeOH 2%; white solid
(13%); mp 228.5-229.5 °C dec; NMR (Me2SO-d6) δ 2.32 (s, 3H),
3.25 (t, 1H), 4.34 (d, 2H), 4.77 (s, 2H), 6.67 (dd, 1H), 6.76 (dd,
1H), 7.38-7.67 (m, 8H), 7.93 (d, 1H), 12.20 (s, 1H). Anal.
(C25H20FN3O2S) C, H, N, F, S.
3-Ch lor o-N-((3,4-d ih yd r o-2-m et h yl-4-oxo-6-q u in a zo-
lin yl)m eth yl)-4-(p h en ylsu lfin yl)-N-(p r op -2-yn yl)a n ilin e
(7c): eluant, toluene 50%-CH3CN 48%-MeOH 2%; white solid
(33%); mp 233-234 °C dec; NMR (Me2SO-d6) δ 2.32 (s, 3H),
3.27 (t, 1H), 4.35 (d, 2H), 4.77 (s, 2H), 6.87 (d, 1H), 6.95 (dd,
1H), 7.50-7.67 (m, 8H), 7.94 (d, 1H), 12.20 (s, 1H). Anal.
(C25H20ClN3O2S) C, H, N, Cl, S.
Exp er im en ta l Section
Protocols for chemistry, crystallography, and in vitro biology
were previously described.2 Every intermediate (3-6) de-
scribed below had a satisfactory 1H NMR spectrum in the
solvent stated; chemical shifts and couplings typical of each
compound class have been detailed.2
Gen er a l Meth od for P r ep a r a tion of th e Su lfid es 3. A
solution of the thiol 2 (10 mmol) in THF (about 90 mL) in a
flame-dried flask under argon was treated with NaH (10 mmol)
at 0-3 °C with ice cooling. Following the addition, the salt
solution was allowed to warm to room temperature during 30-
45 min and then treated with the fluoride 1 (10 mmol).
Reaction usually took place at room temperature but was
completed at 60 °C. Extractive workup (hexane/H2O) gave the
product 3.
3-Flu or o-4-(ph en ylth io)n itr oben zen e (3b): yellow prisms
(95%); mp 53.5-54.5 °C (EtOH); NMR (Me2SO-d6). Anal.
(C12H8FNO2S) C, H, N, S.
3-Ch lor o-4-(p h en ylth io)n itr oben zen e (3c): large yellow
crystals (73%); mp 117-118.5 °C (EtOH); NMR (Me2SO-d6).
Anal. (C12H8ClNO2S) C, H, N, Cl, S (lit.8).
4-(P h en ylth io)-3-(tr iflu or om eth yl)n itr oben zen e (3d ):
oil (90%); NMR (CDCl3). Anal. (C13H8F3NO2S) C, H, N, F, S
(lit.9).
4-((2-Meth ylp h en yl)th io)n itr oben zen e (3e): solid (96%);
mp 42-46 °C (hexanes); NMR (CDCl3). Anal. (C13H11NO2S)
C, H, N, S (lit.10 64-65 °C).
4-((2-Ch lor op h en yl)th io)n itr oben zen e (3f): solid (92%);
mp 105-107 °C (hexanes); NMR (CDCl3). Anal. (C12H8-
ClNO2S‚0.1C6H14) C, H, N, Cl, S (lit.10 113-114 °C).
4-((2-Meth ylp h en yl)th io)-3-(tr iflu or om eth yl)n itr oben -
zen e (3g): solid (83%); mp 42-45 °C (hexane); NMR (CDCl3).
Anal. (C14H10F3NO2S) C, H, N, S.
Gen er a l Meth od for P r ep a r a tion of th e An ilin es 4.
The nitrobenzene 3 (q mmol) was dissolved in a warm (50-60
°C) solution of SnCl2‚2H2O (5q mmol) in EtOH (1-1.5q mL)
in a beaker. Concentrated HCl (1-1.5q mL) was added and
the mixture heated at 60 °C for 15-30 min to remove most of
the EtOH. The cooled mixture was basified to pH >> 12 with
NaOHaq and the product isolated by filtration and/or extrac-
tion.
3-F lu or o-4-(p h en ylth io)a n ilin e (4b): white needles (92%);
mp 87-88.5 °C (EtOH); NMR (CDCl3). Anal. (C12H10FNS)
C, H, N, F, S.
3-Ch lor o-4-(p h en ylth io)a n ilin e (4c): white solid (96%);
mp 48-49 °C; NMR (Me2SO-d6). Anal. (C12H10ClNS) C, H,
N, Cl, S (lit.11).
4-(P h en ylth io)-3-(tr iflu or om eth yl)a n ilin e (4d ): yellow
oil (100%); NMR (CDCl3). Anal. (C13H10F3NS) C, H, N, F, S
(lit.12).
4-((2-Meth ylp h en yl)th io)a n ilin e (4e): solid (44%); mp
50-51 °C (hexanes); NMR (CDCl3). Anal. (C13H13NS) C, H,
N, S (lit.13 mp 50 °C).
N-((3,4-Dih ydr o-2-m eth yl-4-oxo-6-qu in azolin yl)m eth yl)-
4-(p h en ylsu lfin yl)-N-(p r op -2-yn yl)-3-(t r iflu or om et h yl)-
a n ilin e (7d ): eluant, EtOAc; white solid (90%); mp 226-228
°C dec; NMR (CDCl3) δ 2.30 (t, lH, J ) 2.1 Hz), 2.54 (s, 3H),
4.15 (d, 2H, J ) 2.1 Hz), 4.74 (s, 2H), 7.04-7.07 (m, 2H), 7.40-
7.46 (m, 3H), 7.61-7.65 (m, 4H), 7.88 (d, lH, J ) 8.5 Hz), 8.13
(s, lH), 11.17 (br s, lH). Anal. (C26H20F3N3O2S‚0.5H2O) C, H,
N, S.
N-((3,4-Dih ydr o-2-m eth yl-4-oxo-6-qu in azolin yl)m eth yl)-
4-((2-m eth ylp h en yl)su lfin yl)-N-(p r op -2-yn yl)a n ilin e (7e):
eluant, EtOAc; off-white solid (72%); mp 197-198 °C dec
(EtOAc); NMR (CDCl3) δ 2.23 (s, 3H), 2.26 (t, 1H, J ) 2.3 Hz),
2.55 (s, 3H), 4.11 (d, 2H, J ) 2.3 Hz), 4.70 (s, 2H), 6.85 (d, 2H,
4-((2-Ch lor op h en yl)th io)a n ilin e (4f): solid (71%); mp 72-
74 °C (hexanes); NMR (CDCl3). Anal. (C12H10ClNS‚0.2C6H14
)
C, H, N, Cl, S (lit.10 mp 77-78 °C).
4-((2-Meth ylph en yl)th io)-3-(tr iflu or om eth yl)an ilin e (4g):
oil (96%); NMR (CDCl3). Anal. (C14H12F3NS) C, H, N, S.
P r op a r gyla n ilin es 5. Reactions were conducted and the
products purified by methods previously described.2
3-F lu or o-4-(p h en ylth io)-N-(p r op -2-yn yl)a n ilin e (5b): oil
(42%); NMR (Me2SO-d6). Anal. (C15H12FNS‚0.25H2O) C, H,
N, S.
3-Ch lor o-4-(p h en ylth io)-N-(p r op -2-yn yl)a n ilin e (5c): oil
(36%); NMR (Me2SO-d6). Anal. (C15H12ClNS) C, H, N, Cl, S.

Lipophilic Inhibitors of Thymidylate Synthase
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 5 683
J ) 9.0 Hz), 7.14 (d, 1H, J ) 7.4 Hz), 7.35 (dt, 1H, J ) 7.4, 1.4
Hz), 7.45 (m, 3H), 7.64 (s, 2H), 8.07 (dd, 1H, J ) 7.7, 1.3 Hz),
8.14 (s, 1H), 11.40 (br s, 1H). Anal. (C26H23N3O2S‚0.2H2O)
C, H, N, S.
4-((2-Ch lor op h en yl)su lfin yl)-N-((3,4-d ih yd r o-2-m eth yl-
4-oxo-6-qu in a zolin yl)m eth yl)-N-(p r op -2-yn yl)a n ilin e (7f):
eluant, CH2Cl2 95%-MeOH 5%; pale yellow solid (48%); mp
221-223 °C dec (CH2Cl2/MeOH); NMR (CDCl3) δ 2.27 (t, 1H,
J ) 2.2 Hz), 2.55 (s, 3H), 4.11 (d, 2H, J ) 2.1 Hz), 4.70 (s,
2H), 6.85 (d, 2H, J ) 9.0 Hz), 7.32 (dt, 1H, J ) 8.4, 1.5 Hz),
7.37 (dt, 1H, J ) 7.2, 1.5 Hz), 7.53 (m, 3H), 7.64 (s, 2H), 8.12
(m, 2H), 11.30 (br s, 1H). Anal. (C25H20ClN3O2S) C, H, N, Cl,
S.
N-((3,4-Dih ydr o-2-m eth yl-4-oxo-6-qu in azolin yl)m eth yl)-
4-((2-m et h ylp h en yl)su lfin yl)-N-(p r op -2-yn yl)-3-(t r iflu o-
r om eth yl)a n ilin e (7g): eluant, CH2Cl2 95%-MeOH 5%; off-
white solid (45%); mp 216-218 °C dec (CHCl3/MeOH); NMR
(Me2SO-d6) δ 2.12 (s, 3H), 2.32 (s, 3H), 3.32 (t, 1H, J ) 2.2
Hz), 4.46 (d, 2H, J ) 1.9 Hz), 4.85 (s, 2H), 7.16 (m, 2H), 7.31
(m, 2H), 7.50 (m, 3H), 7.65 (d, 1H, J ) 7.3 Hz), 7.77 (d, 1H, J
) 7.3 Hz), 7.96 (d, 1H, J ) 1.8 Hz), 12.20 (br, s, 1H). Anal.
(C27H22F3N3O2S) C, H, N, F, S.
Mea su r em en t of in Vivo An titu m or Activity. L5178Y
(TK-), a thymidine kinase-deficient murine lymphoma, was
inoculated by intraperitoneal injection (ip) into B6D2F1 mice
at a concentration of 1 × 106 cells/mouse. Treatment began
on the following day with either compound 8f or vehicle
delivered by ip injection. Compound 8f was dissolved in N,N-
dimethylacetamide (DMA), and this solution was diluted with
a solution of (hydroxypropyl)-â-cyclodextrin (Molecusol, 45%
w/v) to give the injectable solution with a concentration of 5
mg/mL of 8f and 1.7% DMA.
Control-treated animals received vehicle. There were six
animals in each group. The day of death was recorded for each
animal and used to calculate the mean day of death for each
group. The increase in life span (ILS) for the 8f-treated group
was calculated as follows:
ILS (%) )
[8f group - control group]
mean day of death
× 100
[control group]
P r ep a r a tion of Su lfon es 8: Meth od A. The sulfide 6c
or 6e (y mmol) was dissolved in CHCl3 (200y mL) and treated
with MCPBA (50-60%, 2y mmol, 2 equiv). The mixture was
stirred for approximately 17 h at 25 °C and then poured into
saturated aqueous NaHCO3. Extractive (CHCl3) workup fol-
lowed by flash chromatography on silica using the eluant
stated gave the pure product. Meth od B. The sulfides 6b,f,g
(y mmol) were dissolved in HOAc (100y mL) and treated with
H2O2 (30%, 5y mmol, 5 equiv). The mixture was stirred 23 h
at 40-60 °C and then poured into saturated aqueous NaHCO3.
Extractive (CH2Cl2) workup followed by flash chromatography
on silica using the eluant stated gave the pure product.
N-((3,4-Dih ydr o-2-m eth yl-4-oxo-6-qu in azolin yl)m eth yl)-
3-flu or o-4-(p h en ylsu lfon yl)-N-(p r op -2-yn yl)a n ilin e (8b):
eluant, toluene 60%-CH3CN 38%-MeOH 2%; cream-colored
solid (5%); mp 237.5-239 °C dec (CH3CN/CCl4); NMR (Me2-
SO-d6) δ 2.32 (s, 3H), 3.29 (t, 1H), 4.38 (d, 2H), 4.82 (s, 2H),
6.68 (dd, 1H), 6.76 (dd, 1H), 7.52 (m, 9H), 12.21 (s, 1H). Anal.
(C25H20FN3O3S) C, H, N, F, S.
Refer en ces
(1) Appelt, K.; Bacquet, R. J .; Bartlett, C. A.; Booth, C. L. J .; Freer,
S. T.; Fuhry M. A. M.; Gehring, M. R.; Herrmann, S. M.;
Howland, E. F.; J anson, C. A.; J ones, T. R.; Kan, C.-C.;
Kathardekar, V.; Lewis, K. K.; Marzoni, G. P.; Matthews, D. A.;
Mohr, C.; Moomaw, E. W.; Morse, C. A.; Oatley, S. J .; Ogden, R.
C.; Reddy, M. R.; Reich, S. H.; Schoettlin, W. S.; Smith, W. W.;
Varney, M. D.; Villafranca, J . E.; Ward, R. W.; Webber, S.;
Webber, S. E.; Welsh, K. M.; White, J . Design of enzyme
inhibitors using iterative protein crystallographic analysis. J .
Med. Chem. 1991, 34, 1925-1934.
(2) J ones, T. R.; Varney, M. D.; Webber, S. E.; Lewis, K. K.; Marzoni,
G. P.; Palmer, C. L.; Kathardekar, V.; Welsh, K. M.; Webber,
S.; Matthews, D. A.; Appelt, K.; Smith, W. W.; J anson, C. A.;
Villafranca, J . E.; Bacquet, R. J .; Howland, E. F.; Booth, C. L.
J .; Herrmann, S. M.; Ward, R. W.; White, J .; Moomaw, E. W.;
Bartlett, C. A.; Morse, C. A. Structure-based design of lipophilic
quinazoline inhibitors of thymidylate synthase. J . Med. Chem.
1996, 39, 904-917.
3-Ch lor o-N-((3,4-d ih yd r o-2-m et h yl-4-oxo-6-q u in a zo-
li n y l)m e t h y l)-4-(p h e n y ls u lfo n y l)-N -(p r o p -2-y n y l)-
a n ilin e (8c): eluant for first chromatography, CCl4 55%-CH3-
CN 35%-Et2O 10%; eluant for second chromatography, tolu-
ene 50%-CH3CN 48%-MeOH 2%; pale yellow solid (43%); mp
242-243.5 °C dec (CH3CN/CCl4); NMR (Me2SO-d6) δ 2.32 (s,
3H), 3.30 (t, 1H), 4.42 (d, 2H), 4.80 (s, 2H), 6.90 (d, 1H), 6.94
(dd, 1H), 7.53-7.69 (m, 5H), 7.84 (m, 2H), 7.94 (d, 1H), 8.02
(d, 1H), 12.21 (s, 1H). Anal. (C25H20ClN3O3S) C, H, N, Cl, S.
N-((3,4-Dih ydr o-2-m eth yl-4-oxo-6-qu in azolin yl)m eth yl)-
4-((2-m eth ylp h en yl)su lfon yl)-N-(p r op -2-yn yl)a n ilin e (8e):
eluant, EtOAc; cream-colored solid (65%); mp 186-188 °C dec
(EtOAc); NMR (CDCl3) δ 2.29 (t, 1H, J ) 2.4 Hz), 2.49 (s, 3H),
2.56 (s, 3H), 4.16 (d, 2H, J ) 2.2 Hz), 4.75 (s, 2H), 6.86 (d, 2H,
J ) 8.8 Hz), 7.20 (d, 1H, J ) 7.4 Hz), 7.33 (t, 1H, J ) 7.5 Hz),
7.47 (dt, 1H, J ) 7.4, 1.4 Hz), 7.65 (d, 2H, J ) 1.5 Hz), 7.72 (d,
2H, J ) 9.2 Hz), 8.14 (m, 2H), 11.35 (br s, 1H). Anal.
(C26H23N3O3S‚0.6H2O) C, H, N, S.
4-((2-Ch lor op h en yl)su lfon yl)-N-((3,4-d ih yd r o-2-m eth -
yl-4-oxo-6-q u in a zolin yl)m et h yl)-N-(p r op -2-yn yl)a n ilin e
(8f): eluant, EtOAc; pale yellow solid (44%); mp 219-220 °C
dec (EtOAc); NMR (CDCl3) δ 2.30 (t, 1H, J ) 2.2 Hz), 2.54 (s,
3H), 4.17 (d, 2H, J ) 2.2 Hz), 4.76 (s, 2H), 6.84 (d, 2H, J ) 9.1
Hz), 7.42 (m, 3H), 7.64 (d, 2H, J ) 1.3 Hz), 7.80 (d, 2H, J )
9.1 Hz), 8.11 (s, 1H), 8.26 (d, 1H, J ) 7.0 Hz), 10.95 (br s, 1H).
Anal. (C25H20ClN3O3S) C, H, N, Cl, S.
(3) Matthews, D. A.; Appelt, K.; Oatley, S. J .; Xuong, Ng. H. Crystal
structure of Escherichia coli thymidylate synthase containing
bound 5-fluoro-2′-deoxyuridylate and 10-propargyl-5,8-dide-
azafolate. J . Mol. Biol. 1990, 214, 923-936.
(4) Hughes L. R.; J ackman, A. L.; Oldfield, J .; Smith, R. C.; Burrows,
K. D.; Marsham, P. R.; Bishop, J . A. M.; J ones, T. R.; O’Connor,
B. M.; Calvert, A. H. Quinazoline antifolate thymidylate syn-
thase inhibitors: Alkyl, substituted alkyl, and aryl substituents
in the C2 position. J . Med. Chem. 1990, 33, 3060-3067.
(5) Weiner, S. J .; Kollman, P. A.; Case, D. A.; Singh, U. C.; Ghio,
C.; Alagona, G.; Profeta, S., J r.; Weiner, P. A new force field for
moecular mechanical simulation of nucleic acids and proteins.
J . Am. Chem. Soc. 1984, 106, 765-784.
(6) Singh, U. C.; Weiner, P. K.; Caldwell, J . K.; Kollman, P. A.
AMBER (Version 3.3); University of California: San Francisco,
1986.
(7) Reddy, M. R.; Bacquet, R. J .; Zichi, D.; Matthews, D. A.; Welsh,
K. M.; J ones, T. R.; Freer, S. Calculation of solvation and binding
free energy differences for folate-based inhibitors of the enzyme
thymidylate synthase. J . Am. Chem. Soc. 1992, 114, 10117-
10122.
(8) Hamada, Y.; Matsuoka, H.; Hayashi, M. Yakugashi Zasshi 1970,
90, 644-646.
(9) Stacy, G. W.; Bresson, C. R.; Harmon, R. E.; Thamm, R. C.
Preparation of some aminotrifluoromethyldiphenyl sulfones as
possible antibacterial agents. J . Org. Chem. 1957, 22, 298-302.
(10) Gilman, H.; Broadbent, H. S. Some basically substituted diaryl
sulfides and sulfones. J . Am. Chem. Soc. 1947, 69, 2053-2057.
(11) Singhal, G. P. U.S. Patent 3,576,872, April 27, 1971.
(12) Loeffler, R. S. T.; Woodcock, D.; Carter, G. A.; J ohnson, D. M.
Structure-activity relationships in fungitoxicants related to
triforine and chloraniformethan. Part IV. Evaluation of the
predictive power of the Hansch analysis by the preparation and
testing of further chloraniformethan analogs. Pestic. Sci. 1981,
12, 627-637.
N-((3,4-Dih ydr o-2-m eth yl-4-oxo-6-qu in azolin yl)m eth yl)-
4-((2-m eth ylp h en yl)su lfon yl)-N-(p r op -2-yn yl)-3-(tr iflu o-
r om eth yl)a n ilin e (8g): eluant, CHCl3 97%-MeOH 3%; off-
white solid (43%); mp 169-171 °C dec (CHCl3/MeOH); NMR
(Me2SO-d6) δ 2.33 (s, 3H), 2.40 (s, 3H), 3.35 (t, 1H, J ) 2.2
Hz), 4.53 (d, 2H, J ) 2.2 Hz), 4.93 (s, 2H), 7.20 (m, 2H), 7.36
(m, 2H), 7.57 (m, 2H), 7.69 (dd, 1H, J ) 8.5, 2.1 Hz), 7.77 (d,
1H, J ) 7.8 Hz), 7.98 (d, 1H, J ) 2.2 Hz), 8.03 (d, 1H, J ) 9.0
Hz), 12.25 (br s, 1H). Anal. (C27H22F3N3O3S) C, H, N, F.
(13) Heiduschka, A.; Langkammerer, H. U¨ ber p- und o-toluolsulfin-
sa¨ure. J . Prakt. Chem., Ser. 2. 1913, 88, 425-442.
J M960613F