Structure-based design of lipophilic quinazoline inhibitors of thymidylate synthase

Article abstract of DOI:10.1021/jm9502652

To develop novel lipophilic thymidylate synthase (TS) inhibitors, the X- ray structure of Escherichia coli TS in ternary complex with FdUMP and the inhibitor 10-propargyl-5,8-dideazafolic acid (CB3717) was used as a basis for structure-based design. A total of 31 novel lipophilic TS inhibitors, lacking a glutamate residue, were synthesized; 26 of them had in common a N-((3,4- dihydro-2-methyl-6-quinazolinyl)methyl)-N-prop-2-ynylaniline structure in which the aniline was appropriately substituted with simple lipophilic substituents either in position 3 or 4, or in both. Compounds were tested for their inhibition of E. coli TS and human TS and also for their inhibition of the growth in tissue culture of a murine leukemia, a human leukemia, and a thymidine kinase-deficient human adenocarcinoma. The crystal structures of five inhibitors complexed with E. coli TS were determined. Five main conclusions are drawn from this study. (i) A 3-substituent such as CF₃, iodo, or ethynyl enhances binding by up to 1 order of magnitude and in the case of CF₃ was proven to fill a nearby pocket in the enzyme. (ii) A simple strongly electron-withdrawing substituent such as NO₂ or CF₃SO₂ in the 4- position enhances binding by 2 orders of magnitude; it is hypothesized that the transannular dipole so induced interacts favorably with the protein. (iii) Attempts to combine the enhancements of i and ii in the same molecule were generally unsuccessful. (iv) A 4-C₆H₅SO₂ substituent provided both electron withdrawal and a van der Waal's interaction of the phenyl group with a hydrophobic surface at the mouth of the active site. The inhibition (K(is) = 12 nM) of human TS by this compound, 7n, showed that C₆H₅SO₂ provided virtually as much binding affinity as the CO-glutamate which it had replaced. (v) The series of compounds were poorly water soluble, and also the potent TS inhibition shown by several of them did not translate into good cytotoxicity. Compounds with large cyclic groups linked to position 4 by an SO or SO; group did, however, have IC₅₀'s in the range 1-5 μM. Of these, 4-(N-((3,4- dihydro-2-methyl-6-quinazolinyl)methyl)-N-prop-2-ynylamino)phenyl phenyl sulfone, 7n, had IC₅₀'s of about 1 μM and was chosen for further elaboration.

Full text of DOI:10.1021/jm9502652

904
J . Med. Chem. 1996, 39, 904-917
Str u ctu r e-Ba sed Design of Lip op h ilic Qu in a zolin e In h ibitor s of Th ym id yla te
Syn th a se
Terence R. J ones,* Michael D. Varney, Stephen E. Webber, Kathleen K. Lewis, Gifford P. Marzoni,
Cindy L. Palmer, Vinit Kathardekar, Katharine M. Welsh, Stephanie Webber, David A. Matthews,
Krzysztof Appelt, Ward W. Smith, Cheryl A. J anson, J . E. Villafranca, Russell J . Bacquet,† Eleanor F. Howland,
Carol L. J . Booth, Steven M. Herrmann, Robert W. Ward, J ennifer White, Ellen W. Moomaw,
Charlotte A. Bartlett, and Cathy A. Morse
Agouron Pharmaceuticals, Inc., 3565 General Atomics Court, San Diego, California 92121
Received April 10, 1995X
To develop novel lipophilic thymidylate synthase (TS) inhibitors, the X-ray structure of
Escherichia coli TS in ternary complex with FdUMP and the inhibitor 10-propargyl-5,8-
dideazafolic acid (CB3717) was used as a basis for structure-based design. A total of 31 novel
lipophilic TS inhibitors, lacking a glutamate residue, were synthesized; 26 of them had in
common a N-((3,4-dihydro-2-methyl-6-quinazolinyl)methyl)-N-prop-2-ynylaniline structure in
which the aniline was appropriately substituted with simple lipophilic substituents either in
position 3 or 4, or in both. Compounds were tested for their inhibition of E. coli TS and human
TS and also for their inhibition of the growth in tissue culture of a murine leukemia, a human
leukemia, and a thymidine kinase-deficient human adenocarcinoma. The crystal structures
of five inhibitors complexed with E. coli TS were determined. Five main conclusions are drawn
from this study. (i) A 3-substituent such as CF3, iodo, or ethynyl enhances binding by up to 1
order of magnitude and in the case of CF3 was proven to fill a nearby pocket in the enzyme.
(ii) A simple strongly electron-withdrawing substituent such as NO2 or CF3SO2 in the 4-position
enhances binding by 2 orders of magnitude; it is hypothesized that the transannular dipole so
induced interacts favorably with the protein. (iii) Attempts to combine the enhancements of
i and ii in the same molecule were generally unsuccessful. (iv) A 4-C6H5SO2 substituent
provided both electron withdrawal and a van der Waal’s interaction of the phenyl group with
a hydrophobic surface at the mouth of the active site. The inhibition (Kis ) 12 nM) of human
TS by this compound, 7n , showed that C6H5SO2 provided virtually as much binding affinity as
the CO-glutamate which it had replaced. (v) The series of compounds were poorly water soluble,
and also the potent TS inhibition shown by several of them did not translate into good
cytotoxicity. Compounds with large cyclic groups linked to position 4 by an SO or SO2 group
did, however, have IC50’s in the range 1-5 µM. Of these, 4-(N-((3,4-dihydro-2-methyl-6-
quinazolinyl)methyl)-N-prop-2-ynylamino)phenyl phenyl sulfone, 7n , had IC50’s of about 1 µM
and was chosen for further elaboration.
In tr od u ction
In the past decade there have been several develop-
ments which have stimulated and refocused antifolate
cancer chemotherapy.1 One has been the discovery,2
development,3 and clinical trial4 of 10-propargyl-5,8-
dideazafolic acid (1), a specific inhibitor of thymidylate
synthase (TS). Clinically active but renally toxic,5 the
drug was withdrawn. Development of the more soluble6
to antifolates owing to defective transport is well
and nontoxic7 desamino analog 2 led to the 2-methyl-
2-desamino compound 3 with improved pharmacological
properties.8,9 Active cellular uptake of classical TS
inhibitors followed by enzymic polyglutamation10 to
tighter binding, noneffluxable forms9,11,12 is an impor-
tant aspect of their pharmacology, full recognition of
which was given in the further development of 3 to the
second-generation thiophene-containing D169413,14 and
also in the development of 5,8-dideazaisofolate15 and
benzoquinazoline compounds.16
described,17 and reduced capacity for polyglutamation
has been seen with methotrexate,18,19 edatrexate,20 and
compound 3 itself.21 It is important therefore to develop
a potent lipophilic TS-inhibiting drug. This type of
agent would pass into cells by diffusion and efflux
similarly. The duration of enzyme blockade would thus
be dependent upon plasma levels which in turn, pro-
vided the drug had a reasonably long half-life, would
be under clinical control. The full rationale for this work
is stated in a preliminary communication.22 Briefly, the
task was to enhance the potency of the CO-glutamate-
lacking lipophilic analog 7a (Scheme 1) of 3 by substi-
tuting lipophilic groups into the phenyl ring. Three
series of compounds were synthesized, the first of which
was 4-substituted. Since in the crystal structure of 1
with TS there was apparent a small unfilled pocket of
the enzyme in the position meta to the N10 nitrogen, a
However, a drug requiring both active transport and
polyglutamation begins to look less favorable when
thought is given to drug resistance. Acquired resistance
* Present address: Atelier Pharmaceuticals, Inc., c/o 4909 Murphy
Canyon Rd., Suite 440, San Diego, CA 92123.
† Deceased, and to whose memory this paper is dedicated.
X Abstract published in Advance ACS Abstracts, J anuary 1, 1996.
0022-2623/96/1839-0904$12.00/0 © 1996 American Chemical Society

Lipophilic Quinazoline Inhibitors of TS
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4 905
F igu r e 1. Stereoview showing opening of the E. coli active site. The active site is outlined with a Connolly surface in yellow, and
the protein is shown in purple. The modeled ligand 7d is shown in green for carbon, red for oxygen, and blue for nitrogen with
the hydroxy OH hydrogen bonded to the Ile79 carbonyl oxygen.
F igu r e 2. Stereoview showing the experimentally determined structure of compound 7d complexed to E. coli TS. The active site
is outlined with a Connolly surface in yellow, and both the protein and the ligand are shown in green for carbon, red for oxygen,
and blue for nitrogen. The ligand hydroxy group points toward bulk solvent and makes a hydrogen bond to a bound water molecule.
second series of compounds with 3-substituents was
designed to exploit it; the third series comprised 3,4-
disubstituted compounds. During the course of these
studies, and indeed central to our approach, we exam-
ined five X-ray structures of the Escherichia coli TS
complexed with FdUMP and compounds newly synthe-
sized. This ability to “see” the newly synthesized drug
candidate bound to the targeted receptor, either to
confirm the mode of binding hypothesized for it or to
show otherwise, is a powerful alternative to more
traditional methods of inhibitor design. In this early,
relatively conservative work, three out of four hypoth-
eses were confirmed.
hydrogen bonds could be made. The model for com-
pound 7d is shown in Figure 1 with 2.82 Å separating
the oxygens. But again when the compounds were
made and tested, there was no improvement in inhibi-
tion for either compound against either enzyme. When
the crystal structure of the complex of 7d in E. coli TS
was solved, it became apparent that the hydroxyl group
preferred to point out of the cavity and make a hydrogen
bond to a water molecule that also interacted with
Glu82 seen beneath (Figure 2). In the human enzyme
there is Ala in place of this Glu82. Thus the Ki’s
measured for compound 7d against E. coli and human
TS, while similar, could reflect different modes of
binding, and it is possible that the model of Figure 1 is
correct for the human enzyme. The trifluoromethyl
compound 7f, the first significantly better TS inhibitor,
was next developed and confirmed to bind to the enzyme
as predicted.22 Three further substituents were used
to probe the meta pocketsbromine, iodine, and ethynyl,
as found in compounds 7g-i. For these it was found
that bromine and ethynyl functioned similarly to CF3,
while, relative to hydrogen, iodine clearly conferred
tighter binding by 1 order of magnitude for both
enzymes.
In h ibitor Design , Resu lts, a n d Discu ssion
The structure of the ternary complex of E. coli TS,
FdUMP, and 123 contains meta to the nitrogen in the
p-aminobenzoyl (pAB) ring a predominantly hydropho-
bic pocket formed by residues Leu172 and Val262 (found
as Met in human TS), the edge of Trp80, and the
backbone carbonyl of Ile79 (all residue numbers refer
to the E. coli enzyme). Our first attempts to fill this
were with methyl and ethyl groups as in compounds24
7b,c. However, these when made and tested offered no
improvement of inhibition for either enzyme. Next it
was noticed that the carbonyl oxygen of Ile79 was
appropriately positioned to accept a hydrogen bond from
a hydroxyl group, and when the hydroxymethyl and
hydroxyethyl compounds 7d ,e were modeled to place the
hydroxyl group facing into the pAB-binding cavity,
At the para position, chlorine (compound 7j), tried
first, gave a modest enhancement of inhibition. Next,
on the hypothesis25 that a transannular dipole in the
pAB ring would increase binding to TS, compounds 7k ,l
containing the strongly electron-withdrawing cyano and
nitro groups were synthesized. The cyano group in 7k

906 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4
J ones et al.
Sch em e 1a
F igu r e 3. Correlation of experimental dipoles of para-
substituted anilines and TS inhibition. Compounds with
increased dipoles have increased TS inhibition. E. coli TS
inhibition is represented by a square, and the human TS
inhibition is represented by a diamond.
for the human enzyme relative to 7a . The crystal
structure for this compound was eventually solved, but
the density for the distal phenyl group was disordered,
and the group could not be positioned. It was doubtful
that the desired interaction of this group with Phe176
was taking place, and consideration was given to a
different linking groupssulfonyl. The sulfone 7n was
thus modeled, synthesized, and found to potently inhibit
TS, to which it bound as predicted.22 In that stereo-
graphic depiction, it is apparent that the uppermost of
the sulfone oxygens in 7n is facing a hydrophobic wall
constituted by the side chain of Leu172 and has likely
in the process of binding undergone desolvation at cost
of binding energy. The corresponding sulfoxide would
not have to pay this penalty, but there would be an
opposing cost of less electron withdrawal. When the
sulfoxide 21 was prepared and tested, it was found to
be 5-fold less inhibitory to either enzyme. The com-
pound22 18 (Scheme 2) combining the 3-CF3 and
4-SO2C6H5 substituents was found no better an inhibitor
than 7n . Four other compounds (7q-s,u combining the
m-CF3 substituent respectively with cyano, nitro, sul-
famoyl, and SO2-glycine substituents at the para posi-
tion were synthesized. Compound 7r was modeled as
the 5-amino-2-nitrobenzotrifluoride fragment to assess
conformational effects. AM1 calculations showed that
the nitro group needed to turn 13.5° out of the ring plane
in relief of steric interaction with the trifluoromethyl
group; this conformation was accommodated by the
enzyme. When the enzyme inhibition results for three
of these disubstituted compounds are analyzed to assess
the additional effect of the CF3 group on the inhibition
of the corresponding 4-monosubstituted compound,
there is seen for the nitrile 7q a 10-fold enhancement
for E. coli and no effect on human; for the nitro
compound 7r , a 2-fold enhancement for E. coli and no
effect on human; and for the sulfanilamide 7s, a 4-fold
enhancement for E. coli and 2-fold enhancement for
human. It is possible that the attenuation of the
inhibitory effect of the nitro group relative to that of
cyano is due to a lessening of through resonance by the
rotation of the nitro group since the modern view that
an aromatic nitro group withdraws inductively may not
apply for polar solvents.29 It is interesting that the
a
(a) Propargyl bromide/K2CO3/∆H (propargyl tosylate for 5d ,e)
(method not used for 5j,l); (b) 6/CaCO3/DMF or DMA/∆H; (c) LiOH
to convert 7t into 7u .
enhanced binding for both enzymes by 30-fold, and the
nitro group in 7l caused 60- and 40-fold enhancements
for bacterial and human TS, respectively. Since on
another front sulfones had become of interest (see
below), two further compounds with sulfonyl-containing
substituents were synthesized, the sulfanilamide 7o and
the trifluoromethyl sulfone 17 (Scheme 2). The former
was about 40-fold more active than 7a and the latter
100-fold more active, having a Ki of 21 nM for the
human enzyme. Electron withdrawal by the 4-substitu-
ent clearly conferred inhibitory potency. Since the CF3-
SO2 substituent is the most powerful substituent in this
category commonly found in drugs, further exploration
on this front alone was considered unnecessary. Figure
3 shows the relationship between the experimental
dipole26 of the corresponding substituted aniline and the
TS inhibition for each of the enzymes.
A stabilizing edge to face interaction27,28 between the
pAB ring of 1 and Phe176 was a noticeable feature in
the working structure;23 this gave the idea of setting
up an additional aromatic residue, with its face to the
outer edge of Phe176, by appending a phenyl group at
position 4 through a strongly electron-withdrawing
linker. A carbonyl group was first considered as the
linker. AM1 calculations on a 4-aminobenzophenone
model fragment showed a stable conformation with 24°
and 38° angles, respectively, between the substituted
and unsubstituted rings and the carbonyl plane; the
angle between the ring planes was 56°. This was not
90° as required, but nonetheless compound 7m was
synthesized, and it showed a 20-fold stronger affinity
for the bacterial enzyme and a 40-fold stronger affinity

Lipophilic Quinazoline Inhibitors of TS
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4 907
F igu r e 4. Stereoview showing the experimentally determined structure of compound 7u complexed to E. coli TS. The active site
is outlined with a Connolly surface in white, and the protein is in blue. The ligand is shown in pale green for carbon, red for
oxygen, blue for nitrogen, green for fluorine, and yellow for sulfur. The side chain of His51 is shown to have moved from its
original position in magenta to make a hydrogen bond with one of the carboxylate oxygens of the inhibitor.
compound 7q selects for inhibition of the bacterial
enzyme, presumably due to its CF3 group. In explora-
tion of this, the 3-chloro analog 7p was synthesized, but
the test results for it showed an ablation of the selective
effect. Considered in the round, in disubstituted com-
pounds the 3-CF3 group moderately favors binding to
the bacterial enzyme.
The glycine 7u was a potent inhibitor of enzyme from
both species. The crystal structure of this compound
was solved, and the binding of the glycine unit is shown
in Figure 4. A notable feature is that the side chain of
His51 has shifted markedly from its normal position,
by rotation on ø1 from -174° to -84°, so as to donate a
hydrogen bond (3.02 Å) to a carboxylate oxygen of 7u ;
there is probably a charge interaction as well. In the
human enzyme there is a phenylalanine in place of this
histidine so the interaction observed in Figure 4 cannot
take place. Other conformations of 7u are possible in
the human enzyme, and without structural illumina-
tion, no explanation of the species similarity of the Ki’s
is possible.
compounds 25, 30, and 31, wherein the distal phenyl
ring was replaced by saturated cyclic amines attached
via nitrogen in a sulfonamide linkage. These amine
residues were expected to bind hydrophobically to the
Ile79 side chain and also increase aqueous solubility.
The trio 25, 30, and 31 showed slightly less inhibition
than the sulfone 7n with no discrimination between the
enzymes. The N10-unsubstituted compound 32 was
required as a synthetic intermediate; a comparison of
its TS inhibitions with those of compound 7n shows that
the propargyl group tightens binding by 80-fold in this
series. The analog of compound 1 containing the N10
ethyl group in place of propargyl was known to be only
3-fold less potent as an inhibitor of murine TS.3 The
ethyl group was thus exemplified in two compounds, 35
and 36, the latter containing the hitherto untried
electron-withdrawing group trifluoroacetyl. Compari-
son of the inhibitions of 35 with those of 7n show also
a 3-fold difference between the N10 substituents in
conferring inhibitory potency against human TS. AM1
calculations on diphenylmethane revealed many avail-
able conformations, but that mimicking the conforma-
tion of compound 7n seen in ref 22 was of the lowest
energy. Compound 38, however, with a methyl group
at N10 was thus synthesized. The methyl group is
known6 to give about 10-fold less inhibition of TS than
the propargyl; nonetheless, compound 38 failed to give
any inhibition when measured at 1 µM, the limit of its
solubility. This loss of activity was attributed to the
absence of electron withdrawal by the 4-benzyl sub-
stituent. Lastly, since sulfur was well accommodated
when modeled in place of oxygen in the heterocyclic ring,
the 4-thioquinazoline derivative 39 was synthesized.
This compound had slight changes in TS inhibitions
compared to compound 7f that together gave a small
selective inhibition of E. coli TS.
The activity conferred by the 4-phenylsulfonyl sub-
stituent prompted the synthesis of the three related
Finally, attention is drawn to position 2 of the
quinazoline ring where for all compounds tested a

908 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4
J ones et al.
Sch em e 2a
methyl substituent was present in place of the amino
substituent found in compound 1; this change of sub-
stituent causes only a slight change (1 vs 3, Table 2) in
inhibition. It is thus interesting to observe that in five
crystal structures the arrangement of the protein at this
locus was the same for all the 2-methylquinazoline
ligands seen accommodated and was identical with that
seen for 1.
Ch em istr y. Twenty lipophilic antifolates, 7a -s,u ,
were prepared from the corresponding propargylanilines
5 (5t for 7u ) by N-alkylation (Scheme 1). Eighteen of
the propargylanilines 5 were prepared from the primary
anilines 4, all of which were commercially available with
four exceptions. Thus, 4n was prepared as described,30
4q was prepared via a von Braun reaction,31 and 4s,t,
containing sulfonamide moieties, were prepared from
2-amino-5-nitrobenzotrifluoride by diazotization, sulfo-
nyl chloride formation, amination, and reduction. Pro-
pargylation of the anilines 4a -c,f-i,k ,m -t utilized
propargyl bromide and a metal carbonate as auxiliary
base (Scheme 1), and details of these reactions are
collected in the Supporting Information. In general, and
as expected, the conditions and yields (ranging from 73%
for 5a to 16% for 5r ) varied with the electron withdrawal
capability of the substituent(s). In the purification of
5r , a strange byproduct eluted last that was proven to
be the amidine 10. A repeat of the experiment but
omitting propargyl bromide left the amine 4r un-
changed, even after 16 h. The propargylamine 5r when
heated in DMF/K2CO3 gave no 10. These two control
experiments implicated the involvement of propargyl
bromide. A repeat of the original experiment with
4-nitroaniline gave no amidine formation. We hypoth-
esize that the sufficiently acidic aniline 4r forms an
anion with K2CO3 which attacks the complex formed
from propargyl bromide and DMF to give the amidine
10 with expulsion of propargyl alcohol. Two propargy-
lanilines with hydroxyl-containing substituents (5d ,e)
were prepared using propargyl tosylate.32 Use was
made of the trifluoroacetyl activating/protecting group33
to give, in one pot, the chloro derivative 5j. The
propargylaniline 5l was prepared from 4-fluoronitroben-
zene,34 and this method was used for a convenient
alternative preparation of 5n .
a
Conditions: (a) C6H5SO2Na/135 °C; (b) SnCl2/HCl; (c) prop-
argyl bromide/DMF/K2CO3/110 °C; (d) (CF3SO2)2O/AlCl3/25 °C; (e)
propargylamine/DMSO/K2CO3/25 °C; (f) 1. NaH/DMF, 2. 16, 3.
LiOH.
In Scheme 2, the synthetic routes to the 4-sulfonyl-
derivatized propargylanilines 14 and 15 incorporated
increasingly useful aromatic nucleophilic reactions.
Thus, the diphenyl sulfone 11 was prepared by displace-
ment of fluoride ion using benzenesulfinate, and this
product was taken on by reduction to give the aniline
12, propargylation of which gave 15. Nucleophilic
displacement of fluorine from 4-fluorophenyl trifluo-
romethyl sulfone (13) is known,35 but rather than
embarking on the difficult chemistry involved in the first
preparation of this compound,36 a newer method was
found.37 Thus 13 was prepared in low yield, although
conveniently, from fluorobenzene in a Friedel-Crafts
sulfonation reaction. The product was proven by NMR
spectroscopy to be a mixture of para and meta isomers
in 88/12 ratio, and they proved in our hands impossible
to separate. The mixture was therefore committted to
a nucleophilic displacement reaction with propargy-
lamine, in which only the activated para isomer 13
participated and from which the desired product 14 was
easily and cleanly obtained. Alkylation of the anions
of 14 and 15, generated with sodium hydride in DMF,
was conducted using the quinazoline 16,38 protected at
its lactam nitrogen with a (pivaloyloxy)methyl group,
The quinazoline derivatives 7a -t were prepared by
N-alkylation of the corresponding propargylanilines 5
with the (bromomethyl)quinazoline 6, prepared as pub-
lished,8 and details of these reactions are collected in
Table 1. As in the first set of alkylations, yields upon
introducing the quinazoline moiety reflected electron
withdrawal by the aniline ring substituents but were
now, in certain cases, 7q,r , very low. The realization
that the reduced nucleophilicity of the nitrogen in 5r ,
occasioned by three electron-withdrawing groups allow-
ing a yield of only 2.3% of 7r , together with the likely
genesis of the amidine 10 suggested an alternative
approach wherein the hydrogen of the aniline could be
ionized off to give a more reactive anionic nitrogen. This
approach is evident in Scheme 2.

Lipophilic Quinazoline Inhibitors of TS
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4 909
Ta ble 1. Preparation of Quinazoline Antifolates 7
quantity of reactant (mmol)
reactn
reactn
yield
compd aniline quinazoline CaCO3 solvent mL temp (°C) time (h) purif (%)
mp (°C)
formula
anal.
C,H,N
C,H,N
C,H,N
C,H,N
C,H,N
C,H,N,F
C,H,N,Br
C,H,N,I
C,H,N
C,H,N,Cl
C,H,N
7a
7b
7c
7d
7e
7f
7.9
10
10
4.65
5.71
10
7.9
10
10
4.65
5.71
10
7.9
20
20
4.65d
5.71d
20
DMF
DMF
DMF
DMF
DMF
DMF
DMF
DMF
DMA
DMA
DMF
DMF
DMA
DMA
DMA
DMF
DMA
DMF
DMA
DMA
30
30
20
25
25
30
30
30
15
40
30
30
35
30
30
30
30
50
30
25
95
95
105
90
105
100
110
100
80
2
2
1.8
2
2
1.8
1.8
2
a
b
c
e
f
g
h
i
33
44
33
58
46
32
32
34
29
65
242.5-243.5
229-230.5 dec
195.5-197
215-217
C19H17N3O
20H19N3O
21H21N3O
C20H19N3O2
C21H21N3O2
20H16F3N3O
19H16BrN3O
19H16IN3O
21H17N3O
19H16ClN3O
C20H16N4O
19H16N4O3
26H21N3O2
C
C
171-180
228-229.5
227.5-229 dec
249.5-250.5
234-236 dec
240-240.5
C
C
C
C
C
7g
7h
7i
9.5
5
3.22
9.5
5
3.22
19
10
6.44
3
j
7j
7k
7l
7m
7n
7o
7p
7q
7r
20
20
40
20
20
20
20
20
20
20
95
2
3
3.3
3.5
4
2
3.6
6
24
3.5
9
k
l
10
10
10
10
10
10
10
3.85
10
16
10
10
10
10
10
10
10
3.85
10
32
110
110
115
115
105
110
115
115
90
13m 266-267.5 dec
n
o
p
q
r
s
t
v
x
20
23
13
35
18
5
278-279 dec
213-214.5 dec
191-193
C
C
C
C
C
C
C
C
C,H,N
C,H,N
25H21N3O3S
19H18N4O3S
20H15ClN4O
21H15F3N4O
20H15F3N4O3
C,H,N,S
C,H,N,S
C,H,N,Cl
C,H,N,F
u
>278 dec
267-268
270.5-271.5
7.7
20
32
2.3 261-262 dec
21
19
7s
>286 dec
20H17F3N4O3S C,H,N,F,S
7tw,y
110
247.5-248.5 dec C22H19F3N4O5S C,H,N,S
a
1.87 g recrystallized from acetone to give off-white microneedles. b 2.62 g twice recrystallized from EtOAc to give off-white microneedles.
c 2.63 g twice recrystallized from EtOAc to give white microneedles. K2CO3 was used in place of CaCO3 and the Celite filtration omitted.
d
e 1.09 g taken up in hot MeOH (150 mL); the filtered, cooled solution was diluted with H2O (250 mL) to precipitate the pure product as
a beige crystalline solid. f 1.37 g taken up in hot MeOH (100 mL); the filtered, cooled solution was diluted with H2O (250 mL) to precipitate
g
h
the pure product as a beige crystalline solid. 2.82 g twice recrystallized from EtOAc to give off-white microneedles. 3.00 g twice
recrystallized from EtOAc to give off-white microneedles. 2.07 g twice recrystallized from EtOAc to give white crystals. 0.42 g flash
chromatographed eluting sequentially with 90% CH2Cl2 in hexane, CH2Cl2, and 1.5% CH3CN in CH2Cl2 to provide white flakes. 6.01 g
twice recrystallized from 1,2-dichloroethane to give white microneedles. 2.03 g flash chromatographed with 40% CH3CN in CH2Cl2 to
i
j
k
l
m
n
give a white solid; the sample was applied in the eluant with the aid of a little DMF. Not optimized. 2.78 g coated onto SiO2 from
CHCl3 and flash chromatographed eluting sequentially with CH3CN, and 15% MeOH in CH3CN to give a yellow solid. o 3.81 g twice
p
recrystallized from 1,2-dichloroethane to give off-white microneedles. 3.83 g coated onto SiO2 from 1,2-dichloroethane and flash
chromatographed using 10% MeOH in benzene to give 2.19 g of impure material which was rechromatographed using 5% MeOH in
q
CH2Cl2 to give 0.56 g which was recrystallized from benzene/hexane to give off-white crystals. 2.94 g coated onto SiO2 from DMF and
r
flash chromatographed using 5% MeOH/EtOAc to give a white solid. Extracted with CHCl3 to give 3.03 g which was coated onto SiO2
s
from CHCl3 and flash chromatographed using 40% CH3CN in CH2Cl2 as eluant to give a white solid. Extracted with EtOAc to give 1.96
g which was coated onto SiO2 from CH3CN and flash chromatographed using CH3CN as eluant to give a white solid. The analytical
t
sample was recrystallized from CH3CN. Extracted with EtOAc to give 0.65 g which was flash chromatographed using 40% CH3CN in
CH2Cl2 to give 0.1 g of material 84% pure by HPLC. This was coated onto SiO2 from CH3CN and rechromatographed using 30% ether in
u
EtOAc to give a yellow solid 98% pure by HPLC. This compound failed to microanalyze well. C: calcd, 57.70; found, 58.83, 58.62. H:
calcd, 3.63; found, 4.17, 4.18. N: calcd, 13.46; found, 12.46, 12.54. High-resolution mass spectrometry proved the structure: calcd for
C
20H15F3N4O3, 416.1096; found, 416.1100. v 3.05 g coated onto SiO2 from DMF and flash chromatographed using 80% CH3CN in CH2Cl2
to give a white solid. The saponification of this methyl glycinate ester to give the carboxylic acid 7u is given in footnote y. x 0.40 g
coated onto SiO2 from DMF and flash chromatographed using 80% CH3CN in CCl4 to give a white solid. y N-(N4-((3,4-Dihydro-2-methyl-
4-oxo-6-quinazolinyl)methyl)-N4-prop-2-ynyl-2-(trifluoromethyl)sulfanilyl)glycine (7u ). The methyl glycinate 7t (0.138 g, 0.264 mmol)
was suspended in a 1:1 mixture of THF:H2O (8 mL) and treated with 1 N LiOH (1.32 mL, 1.32 mmol) to give a solution which was kept
at 20 °C for 3 h. The solution was filtered and brought to pH 2.0 with 1 N HCl to produce a gelatinous white precipitate. This was
centrifuged and washed by three cycles of resuspension (H2O, 15 mL)-centrifugation-decantation. The product was dried over P2O5 in
vacuo at 20 °C and obtained as an amorphous white solid (0.087 g, 65%): mp 237 °C dec. NMR (Me2SO-d6) satisfactory. Anal.
(C22H19F3N4O5S) C,H,N.
w
to prevent anion exchange. Removal of the protecting
group with LiOH during workup gave the desired
antifolates 17 and 18.
34. Coupling of these amines with 6 gave the targeted
N10-ethylated compounds 35 and 36. Coupling of 4-(phe-
nylmethyl)aniline with 6 gave the amine 37 which on
reductive methylation gave the target compound 38
containing the 4-benzyl substituent. Finally, the 4-thio-
quinazoline derivative 39 was prepared by P2S5 treat-
ment of the antifolate 7f.
Displacement of fluoride from 4-fluorophenyl phenyl
sulfoxide 19 gave 4-(propargylamino)phenyl phenyl
sulfoxide 20 that was coupled with the quinazoline 6 to
provide the target sulfoxide 21. Morpholine and pip-
erazine (the latter protected by tert-butyloxycarbonyl)
were treated with 4-nitrobenzenesulfonyl chloride, and
the resulting sulfonamides were taken on in familiar
reduction, propargylation, and coupling steps to give the
quinazolines 25 and 30, the latter following removal of
the t-BOC with trifluoroacetic acid. The piperazine 30
was N-methylated to give the additional target com-
pound 31. The N10-unsubstituted diphenyl sulfone 32
was prepared by N-alkylation. Alkylation of the aniline
4n gave 4-(ethylamino)phenyl phenyl sulfone 33, a
compound that was alternatively prepared from 4-fluo-
rophenyl phenyl sulfone. A similar fluoride displace-
ment reaction gave 4-(ethylamino)trifluoroacetophenone
In h ibition of Cellu la r Gr ow th . The target com-
pounds were assessed for cytotoxicity by measuring
their effect on the growth in tissue culture of three cell
lines: murine (L1210) and human (CCRF-CEM) leuke-
mias and a thymidine kinase-deficient human adeno-
carcinoma cell line (GC3/M TK-). The results are shown
in Table 2. The insolubility of certain compounds
limited or prevented their proper assay, which was an
unwelcome first sign. Potent growth inhibition was not
expected of the weaker TS inhibitors, nor was it found;
but it was disappointing that it was also not found for
most of the stronger TS inhibitors, for example, 7k ,l,o-s
and 17. However, a group of five inhibitors (7n , 18, 21,

910 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4
J ones et al.
Ta ble 2. Inhibition of Thymidylate Synthase and of Growth of Cells in Tissue Culture
Kia (µM)
b
IC50 (µM)
compd
E. coli TS
human TS
L1210
CCRF-CEM
GC3/M TK-
1c
0.0030 ( 0.0003
0.0051 ( 0.0017
0.0048 ( 0.0011
0.013 ( 0.003
4.0 ( 1.2
0.012 ( 0.003
0.011 ( 0.003
0.0085 ( 0.0044
0.0071 ( 0.0035
2.2 ( 1.6
1.57
0.75
2.13
3d
7a
7b
7c
7d
7e
7f
0.045
0.029
0.048
15 ( 7
5.7 ( 1.9
>6.5 (35%)
>5 (30%)
>5 (25%)
48
>10.7 (29%)
>5 (20%)
>5 (26%)
>54 (30%)
>49 (33%)
>4.3 (41%)
>5.7 (20%)
>5.3 (10%)
>2.4 (15%)
>4 (20%)
>5 (10%)
>4.8 (8%)
>7.7 (40%)
1.44
>10.7 (24%)
>5 (34%)
>5 (20%)
>54 (22%)
49
3.4 ( 1.1
1.4 ( 0.4
16 ( 7
3.4 ( 0.3
5.0 ( 2.4
2.7 ( 0.7
24 ( 8
3.8 ( 0.2
1.7 ( 0.7
0.62 ( 0.25
1.2 ( 0.6
5.1 ( 1.3
2.4 ( 1.1
1.7 ( 0.3
8.3 ( 4.3
3.6 ( 0.7
>49 (35%)
>2.2 (35%)
1.8
0.48 ( 0.22
2.2 ( 0.8
0.39 ( 0.4
1.1 ( 0.4
>4.3 (25%)
5.3
7g
7h
7i
0.92 ( 0.35
3.3 ( 1.9
0.65 ( 0.16
0.84 ( 0.07
0.16 ( 0.02
0.44 ( 0.03
0.41 ( 0.08
0.68 ( 0.19
0.36 ( 0.02
0.77 ( 0.19
0.076 ( 0.063
0.17 ( 0.08
0.056 ( 0.013
0.098 ( 0.018
0.11 ( 0.04
0.16 ( 0.07
0.012 ( 0.005
0.021 ( 0.009
0.064 ( 0.015
0.084 ( 0.027
0.037 ( 0.016
0.079 ( 0.028
0.081 ( 0.023
0.14 ( 0.06
0.074 ( 0.035
0.15 ( 0.02
0.036 ( 0.011
0.070 ( 0.021
0.0096 ( 0.0045
0.019 ( 0.007
0.021 ( 0.003
0.028 ( 0.006
0.050 ( 0.031
0.084 ( 0.001
0.061 ( 0.023
0.11 ( 0.02
0.045 ( 0.02
0.091 ( 0.033
0.063 ( 0.013
0.071 ( 0.02
0.083 ( 0.012
0.19 ( 0.03
1.0 ( 0.2
0.33 ( 0.08
1.4 ( 0.5
>5.3 (none)
>2.4 (5%)
>4 (25%)
>5 (35%)
4.7
>5.3 (35%)
NT
0.86 ( 0.46
2.7 ( 1
7j
0.85 ( 0.12
3.7 ( 0.8
NT
7k
7l
0.15 ( 0.05
0.33 ( 0.1
>5 (none)
>4.8 (none)
6.2
0.068 ( 0.02
0.17 ( 0.04
0.094 ( 0.049
0.47 ( 0.22
0.025 ( 0.006
0.068 ( 0.05
0.10 ( 0.04
0.41 ( 0.17
0.035 ( 0.01
0.10 ( 0.03
0.015 ( 0.007
0.047 ( 0.006
0.038 ( 0.019
0.073 ( 0.057
0.024 ( 0.007
0.061 ( 0.024
0.0059 ( 0.0021
0.015 ( 0.002
0.042 ( 0.015
0.064 ( 0.008
0.037 ( 0.019
0.040 ( 0.008
0.16 ( 0.08
0.55 ( 0.04
0.039 ( 0.001
0.15 ( 0.05
0.10 ( 0.02
0.28 ( 0.11
0.13 ( 0.08
0.28 ( 0.08
2.0 ( 0.6
7m
7n
7o
7p
7q
7r
7.7
0.86
1.76
20
>26 (none)
>6.7 (10%)
>2.8 (none)
10.5
>26 (none)
>6.7 (25%)
NT
4.5
>2.8 (15%)
>5 (none)
insoluble
>75 (none)
>2.75 (45%)
2.0
NT
7s
7u
17
18
21
25
30
31
32
35
36
38
39
insoluble
>77 (none)
>2.75 (none)
3.5
insoluble
>77 (28%)
>2.75 (none)
7.5
1.7
4.5
3.9
3.1
4.5
5.2
>50 (47%)
>10 (27%)
>14.7 (40%)
0.66
48
>50 (47%)
>10 (17%)
>14.7 (10%)
2.6
>10 (24%)
>14.7 (33%)
2.7
17 ( 4
2.2 ( 0.4
0.077( 0.044
0.24 ( 0.09
0.18 ( 0.03
1.0 ( 0.1
0.034 ( 0.018
0.13 ( 0.08
0.13 ( 0.01
0.49 ( 0.08
>1
3.4
>4 (10%)
3.4
>4 (10%)
>5 (25%)
>5.6 (none)
>1
>1
>1
2.8
0.18 ( 0.07
0.36 ( 0.2
1.3 ( 0.6
2.3 ( 1.4
>5.6 (none)
>5.6 (30%)
a
For each compound the uppermost value is Kis and the lower Kii. The insolubility of compound 38 prevented any value from being
b
determined. For entries of the form >x (y%), x is the limiting solubility of the compound and y the percent inhibition at that concentration.
NT means not tested. c Reference 2. Reference 8.
d
25, and 35) containing large cyclic groups linked to
position 4 by SO or SO2 groups did show reasonable
cytotoxicity with most IC50 values falling in the range
1-5 µM. The cytoxicity of all of these compounds was
reversible by thymidine. The IC50 ratios (see the
Experimental Section) were >5.8, 2.25, >5.3, 10, and
10, respectively, showing that their locus of action was
indeed TS. The most potent of these compounds was
the diphenyl sulfone 7n , which had IC50 values of about
1 µM.

Lipophilic Quinazoline Inhibitors of TS
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4 911
broad singlet; br d, broad doublet; br, broad signal; m,
multiplet. Mass spectra were determined at the University
of California, Riverside, using a VG 7070E-HF high-resolution
mass spectrometer using the direct insertion method, an
ionizing voltage of 70 eV, and an ion source temperature of
200 °C. For HRMS the measured mass was within (13
millimass units of the theoretical value. Infrared absorption
spectra were taken on either a Perkin-Elmer 457 spectrometer
or a MIDAC M2000 FTIR instrument. Elemental microanaly-
ses, performed by MHW Laboratories (Phoenix, AZ) or Atlantic
Microlab Inc. (Norcross, GA), gave results for the elements
stated within (0.4% of the theoretical values. Propargyl
bromide was used as an 80% (w/w) solution in toluene. N,N-
Dimethylformamide (DMF) was dried over activated (250 °C)
3 Å molecular sieves; N,N-dimethylacetamide (DMA) (Aldrich
Gold Label grade) was similarly dried. 6-(Bromomethyl)-3,4-
dihydro-2-methyl-4-oxoquinazoline, 6, was prepared as pub-
lished.8 Hu¨nig’s base refers to N,N-diisopropylethylamine.
Tetrahydrofuran (THF) was distilled from sodium benzophe-
none ketyl under nitrogen. Ether refers to diethyl ether.
Petrol refers to petroleum ether of bp 36-53 °C. Flash
chromatography was performed using silica gel 60 (Merck Art
9385). Where the crude solid (x g) was insoluble in the chosen
eluant, it was dissolved in a more polar solvent and Merck
Art 7734 silica gel (4x g) added. The slurry was evaporated
to dryness on a rotary evaporator fitted with a coarse glass
frit to prevent spraying of the silica gel. The coated silica gel
was then applied to the column. Thin layer chromatographs
(TLC) were performed on precoated sheets of silica gel 60F254
(Merck Art 5719). Extracts were dried over Na2SO4 or MgSO4.
Melting points were determined on a Mel-Temp apparatus and
are corrected.
4-Cya n o-3-(t r iflu or om et h yl)a n ilin e (4q ). This com-
pound was prepared by a published31 method: A mixture of
5-amino-2-bromobenzotrifluoride (12.00 g, 50 mmol), CuCN
(5.37 g, 60 mmol), and DMF (7.5 mL) was heated under reflux
for 2 h. It was poured into 30% (w/v) NaCNaq (150 mL) and
extracted with ether (75 mL). The ether layer was washed
sequentially with 10% NaCNaq (100 mL), H2O (100 mL), and
brine (100 mL), dried (Na2SO4), and evaporated to dryness to
give the crude product (8.00 g). This was flash chromato-
graphed on silica gel (700 g) using 20% ether in CH2Cl2 as the
eluant to give the pure product as a light tan solid (7.03 g,
76%) suitable for further use. An analytical sample recrystal-
lized from EtOH/H2O as off-white needles: mp 142-143 °C;
NMR satisfactory. Anal. (C8H5F3N2) C,H,N,F.
4-Nitr o-2-(tr iflu or om eth yl)ben zen esu lfon yl Ch lor id e
(8). 2-Amino-5-nitrobenzotrifluoride (82.45 g, 0.4 mol) was
added to a mechanically stirred, almost boiling mixture of
concentrated HCl (120 mL) and H2O (120 mL) in a 2 L beaker.
Glacial HOAc (200 mL) was added to give a clear yellow
solution which was then cooled to -10 °C; the yellow hydro-
chloride salt precipitated. A solution of NaNO2 (29.67 g, 0.43
mol) in H2O (200 mL) was added during 15 min keeping the
temperature below -5 °C, and the resulting mixture was
stirred at -5 to -10 °C for 15 min. Meanwhile, glacial HOAc
(400 mL) contained in a 4 L beaker was saturated (45 min)
with SO2 gas and CuCl (9.90 g, 0.1 mol) then added. Passage
of SO2 was continued for 30 min until most of the solid had
dissolved to give a thin blue-green suspension which was then
cooled to 5 °C. The diazonium preparation was added to the
cuprous preparation during 11 min, causing vigorous evolution
of N2; the resulting green mixture was stirred for 1.5 h while
it attained ambient temperature. Crushed ice (1 kg) was
added to give the product as a flocculent yellow solid which
was filtered off, washed well with water, and dried in vacuo
(51.36 g). Careful recrystallization from cyclohexane (400 mL)
with seeding gave the pure product as cubic crystals (38.26 g,
33%): mp 81-83 °C; NMR (CDCl3) satisfactory. Anal. (C7H3-
ClF3NO4S) C,H,N,Cl.
Con clu sion s
This manuscript describes our first study in which
we have exploited the three-dimensional architecture
of thymidylate synthase in the design of novel inhibitors.
There are five main findings to report. (1) A substituent
put meta to the N10 nitrogen in the glutamate-lacking
antifolate structure 7a exploits a pocket in TS and
enhances binding to by up to 1 order of magnitude. The
best substituents are iodo, trifluoromethyl, and ethynyl.
(2) A simple, strongly electron-withdrawing substituent,
such as nitro or (trifluoromethyl)sulfonyl, put para to
N10 in 7a enhances binding by 2 orders of magnitude.
This effect was independently discovered by another
research group39 while this work was in progress. The
enhanced binding is attributed to a favorable interaction
between the protein and the transannular dipole that
the substituent generates. (3) An attempt at additively
combining the results of 1 and 2 by incorporating two
substituents in one molecule gave mixed results for E.
coli TSsunsuccessful in the case of the (trifluorometh-
yl)phenyl sulfone 18 but successful for the trifluorom-
ethyl nitrile 7q. Disubstituted compounds all failed to
show the expected binding for the human enzyme,
perhaps revealing a failing of using the bacterial
structure as a surrogate. (4) Strong electron withdrawal
allied with a well-positioned bulky hydrophobic group
as embodied in a 4-phenylsulfonyl substituent brought
enhancement of binding by over 2 orders of magnitude.
Thus a comparison of the human TS Kis values for
compounds 3 and 7n (0.0085 and 0.012, respectively)
shows that the doubly ionic glutamate residue can be
replaced essentially by a hydrophobic phenyl ring with
no change in inhibition. Unsuccessful variants of the
glutamate residue had been previously reported40 before
the TS structure became known. Once the ligand-bound
enzyme structure was on the graphics, it became pos-
sible to devise quickly a totally different inhibitor
structuresan illustration of the remarkable power of
structure-based design. This result achieved the stated
goal of winning back the inhibition lost by removal of
the glutamate. (5) The inhibition of the growth of cells
in tissue culture by these quinazoline compounds was
not impressive. As a class, they had poor water solubil-
ity, and moreover, the reasonably potent enzyme inhibi-
tion shown by several of them did not translate into good
cytotoxicity. Possible reasons for this outcome are that
the compounds did not penetrate cells sufficiently well
and that if they did, did not remain within. The former
explanation is the more probable. Many additional
properties are required of an enzyme inhibitor before it
can begin to qualify as a drug candidate, and unfortu-
nately this series of compounds seemed generally to lack
them. However, the results for the diphenyl sulfone 7n
showed that it was an exception since it was both the
most potent enzyme inhibitor and also the most cyto-
toxic to cells. Attempts to develop it into an even more
potent TS inhibitor having better cytotoxicity will be
described in forthcoming manuscripts.
Exp er im en ta l Section
Proton magnetic resonance spectra were determined using
a General Electric QE-300 spectrometer operating at a field
strength of 300 MHz. Chemical shifts are reported in parts
per million (δ) downfield from tetramethysilane as an internal
standard, and peak multiplicities are designated as follows:
s, singlet; d, doublet; dd, doublet of doublets; t, triplet; br s,
4-Nitr o-2-(tr iflu or om eth yl)ben zen esu lfon a m id e (9a ).
4-Nitro-2-(trifluoromethyl)benzenesulfonyl chloride (11.58 g,
40 mmol) dissolved in EtOH (150 mL) was treated with
concentrated NH4OHaq (13.5 mL, 200 mmol). The mixture was
kept for 1 h at ambient temperature, the reaction quenched
with H2O (600 mL), and the mixture acidified to pH 4 with 1

912 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4
J ones et al.
N HCl to give a precipitate which was kept thus overnight at
4 °C. The product was filtered off, washed, and dried to give
pale yellow needles (7.54 g, 69.8%): mp 163-164 °C; NMR
(Me2SO-d6) satisfactory. Anal. (C7H5N2F3O4S) C,H,N,F,S.
Meth yl N-((4-Nitr o-2-(tr iflu or om eth yl)ph en yl)su lfon yl)-
glycin a te (9b). 4-Nitro-2-(trifluoromethyl)benzenesulfonyl
chloride (13.03 g, 45 mmol) dissolved in EtOH (175 mL) was
added during 3 min to a slurry formed from methyl glycinate
hydrochloride (11.30 g, 90 mmol) in a mixture of EtOH (30
mL) and Et3N (25.1 mL, 180 mmol) keeping the temperature
between 0 and 5 °C. The mixture was stirred for 50 min at
20 °C and then poured into H2O (1600 mL) to precipitate the
product. The pH was adjusted to 1.5 with 1 N HCl, and the
resulting white needles were filtered off, washed with water,
and dried (9.28 g, 60%): mp 139.5-140 °C; NMR (Me2SO-d6)
satisfactory. Anal. (C10H9F3N2O6S) C,H,N,F,S.
15 °C. The mixture was stirred at 10 °C for 45 min. Propargyl
bromide (74.36 g, 0.50 mol) was then added during 10 min at
8 °C. The mixture was stirred for 2 h as it attained ambient
temperature and then left to stand thus overnight for 19.5 h.
The mixture was partitioned between ether (2 L) and H2O (2
L). The ether layer was washed with H2O (3 × 2 L) and
evaporated to give a brown oil (139.7 g). This oil was treated
with a solution of KOH (84.16 g, 1.50 mol) in CH3OH (750 mL)
and the whole then heated under reflux for 3 h. Most (500
mL) of the CH3OH was removed in vacuo leaving a residue
which was partitioned between ether (1.5 L) and H2O (1.5 L).
The ether extract was washed with saturated brine (2 × 1 L),
dried (MgSO4), and evaporated to dryness in vacuo. The
resulting oil (79.3 g) was vacuum distilled to give the product
(bp 126 °C/5 mmHg) as a colorless oil (46.9 g, 56.6%). Anal.
(C9H8ClN) C,H,N,Cl.
2-(Tr iflu or om eth yl)su lfa n ilim id e (4s). 4-Nitro-2-(trif-
luoromethyl)benzenesulfonamide (9a ; 6.89 g, 25.5 mmol) was
added during 4 min to a warm solution of SnCl2‚2H2O (28.77
g, 127.5 mmol) in EtOH (35 mL). The mixture was heated to
55 °C for 10 min to give a solution to which concentrated HCl
(35 mL) was added. The EtOH was boiled away during 20
min, and the mixture was cooled and treated with 2.5 N NaOH
to pH 6.5 and then extracted with EtOAc (3 × 250 mL). The
combined extracts were washed with brine, dried, and evapo-
rated to dryness to give the product (5.44 g, 89%). An
analytical sample was recrystallized from H2O to give beige
plates: mp 186-187 °C; NMR (Me2SO-d6) satisfactory. Anal.
(C7H7F3N2O2S) C,H,N,S.
Meth yl N-((4-Am in o-2-(tr iflu or om eth yl)p h en yl)su lfo-
n yl)glycin a te (4t). The above procedure was repeated start-
ing from methyl N-((4-nitro-2-(trifluoromethyl)phenyl)sulfonyl)-
glycinate (9b; 7.87 g, 23 mmol), SnCl2‚2H2O (25.95 g, 115
mmol), MeOH (35 mL), and concentrated HCl (35 mL).
Basification to pH 6.0 was carefully done. The product was
obtained as a white solid (3.46 g, 48%). An analytical sample
was recrystallized from H2O to give beige crystals: mp 133-
134 °C; NMR (Me2SO-d6) satisfactory. Anal. (C10H11F3N2O4S)
C,H,N,F,S.
4-(P h en ylsu lfon yl)-N-p r op -2-yn yla n ilin e (5n ). A solu-
tion of 4-(phenylsulfonyl)fluorobenzene (40.00 g, 0.169 mol)
and propargylamine (79.6 g, 1.445 mol, 8.55 equiv) in Me2SO
(900 mL) containing CaCO3 (18.6 g, 0.186 mol) in suspension
was stirred at 125 °C for 32 h under argon. The mixture was
cooled and filtered through Celite. The filtrate was evaporated
to dryness in vacuo to leave a residue which was partitioned
between H2O (2 L) and CH2Cl2 (2 L). The organic layer was
dried and evaporated to dryness to give the crude product (89.6
g). This was thrice flash chromatographed with CH3CN on
silica gel (3 × 1 kg) taking pure product out in the first two
stages and recycling fractions containing less pure material.
The total yield was 34.53 g (75.3%) of a pale yellow solid
identical by NMR with the product (Supporting Information)
prepared as in Scheme 1.
N 1 ,N 1 -D i m e t h y l-N 3 -(4-n i t r o -3-(t r i flu o r o m e t h y l)-
p h en yl)for m a m id in e (10). During the purification of the
amine 5r (Supporting Information) this byproduct eluted last
and was characterized as a yellow oil (2.56 g, 33%): NMR
(CDCl3) δ 3.08 (s, 3H, CH3), 3.11 (s, 3H, CH3), 7.12 (dd, 1H, J
) 8.7, 2.4 Hz, H6), 7.32 (d, 1H, J ) 2.4 Hz, H2), 7.63 (s, 1H,
CHdN), 7.93 (d, 1H, J ) 8.7 Hz, H5). Anal. (C10H10F3N3O2)
C,H,N.
Gen er a l Meth od for P r ep a r in g th e P r op a r gyla n ilin es.
The starting aniline and propargyl bromide were allowed to
react in the solvent stated in the presence of K2CO3. Condi-
tions re given in the Supporting Information. The correspond-
ing dipropargylaniline was present as a byproduct along with
some unreacted starting aniline. 4-(Dimethylamino)cinnama-
ldehyde spray reagent was useful in differentiating these
amines to allow an optimal yield. The reaction mixture was
poured into water and the organic material extracted with
ether. The washed and dried extract was evaporated in vacuo
to provide the crude product that was flash chromatographed.
In each case, the pure propargylamine thus obtained had a
satisfactory NMR spectrum and a satisfactory elemental
microanalysis for the elements stated.
N-P r op -2-yn yl-3-(h yd r oxym eth yl)a n ilin e (5d ). A mix-
ture of 3-aminobenzyl alcohol (3.42 g, 24.76 mmol), propargyl
tosylate32 (5.21 g, 24.76 mmol), K2CO3 (3.42 g, 24.76 mmol),
and DMF (75 mL) was stirred at ambient temperature for 24
h. The reaction mixture was partitioned between ether (250
mL) and H2O (250 mL). The organic layer was washed with
H2O (2 × 250 mL) and then extracted with 0.5 N HCl (200
mL). The acidic extract was basified with 5 N NaOH and
extracted with ether (2 × 200 mL). The combined ether layers
were dried (Na2SO4) and evaporated to yield the crude solid
product (1.11 g, 28%). Flash chromatography (5% MeOH in
CHCl3) gave the pure product as a solid (0.83 g): mp 77-79
°C; NMR satisfactory. Anal. (C10H11NO) C,H,N.
Gen er a l Meth od for P r ep a r in g th e An tifola tes 7.
A
solution of the propargylaniline 5 and the (bromomethyl)-
quinazoline 68 in the solvent stated was stirred over CaCO3.
Conditions are collected in Table 1. SiO2GF/EtOAc was in
most cases a useful TLC system, and the Epstein spray
reagent41 was used to monitor consumption of the (bromom-
ethyl)quinazoline. The mixture was filtered through Celite,
and the filtrate was poured into iced water (1 L). The product
precipitated and, in most cases, was filtered off on a Whatmans
no. 5 filter, washed with H2O, and then dried over P2O5 in
vacuo. In three cases filtration was unsuccessful, and the
product was therefore obtained by extractive workup. Eight
of the compounds were purified by recrystallization, 10 by flash
chromatography, and two by reprecipitation. In some of the
recrystallizations, MgSO4 in equal weight to the crude product
was used to dry the hot solution (and removed at the filtration
stage) to improve crystal formation. All of the compounds in
the table gave satisfactory 1H NMR spectra, and these are not
recorded, with the exception of the spectrum of the 4-chloro
compound 7j to serve as an example: (Me2SO-d6) δ 2.33 (s,
3H, CH3), 3.22 (t, 1H, J ) 2.0 Hz, tCH), 4.24 (d, 2H, J ) 2.0
Hz, CH2-Ct), 4.66 (s, 2H, CH2), 6.81 (d, 2H, J ) 9.1 Hz,
aromatic), 7.21 (d, 2H, J ) 9.1 Hz, aromatic), 7.54 (d, 1H, J )
8.3 Hz, H8), 7.69 (dd, 1H, J ) 8.3, 1.7 Hz, H7), 7.95 (d, 1H, J
) 1.7 Hz, H5), 12.20 (br s, 1H, NH). In like vein, the mass
spectrum of this compound typifies the series: m/z 339 and
337 (M+), 173 (M - C9H7ClN), 132 (173 - CH3CN), 104 (132
- CO).
N-P r op-2-yn yl-3-(1-h ydr oxyeth yl)an ilin e (5e). The above
reaction was repeated with 3-(1-hydroxyethyl)aniline (4.00 g,
29.15 mmol), propargyl tosylate32 (6.13 g, 29.15 mmol), K2CO3
(4.03 g, 29.15 mmol), and DMF (75 mL): thick oil which
crystallized when stored at 0 °C (1.83 g, 36%); NMR satisfac-
tory. Anal. (C11H13NO) C,H,N.
4-Ch lor o-N-p r op -2-yn yla n ilin e (5j). 4-Chlorotrifluoro-
acetanilide33 (111.79 g, 0.50 mol) dissolved in DMF (100 mL)
was added to sodium hydride (15.00 g, 0.625 mol) slurried in
DMF (200 mL) during 30 min keeping the temperature at 10-
4-Nit r o-2-(t r iflu or om et h yl)d ip h en yl P h en yl Su lfon e
(11). A mixture of sodium benzenesulfinate (18.06 g, 0.11
mol), ethylene glycol (75 mL), and 2-(2-ethoxyethoxy)ethanol
(120 mL) was heated to 125 °C in a flask equipped with an
overhead stirrer. 2-Fluoro-5-nitrobenzotrifluoride (20.91 g, 0.1
mol) was added, and the mixture was heated at 135 °C for 3.5
h. The resulting solution was cooled, and H2O (20 mL) was
added to throw down an oil which eventually solidified. The
pasty solid was filtered off, washed with hot H2O (75 mL),

Lipophilic Quinazoline Inhibitors of TS
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4 913
partially dried, and recrystallized from EtOH-H2O with
carbon treatment to provide a tan-colored solid (7.85 g, 23.7%).
Flash chromatography (70% CH2Cl2 in petrol) of a small
portion gave the analytical sample as a cream-colored solid:
mp 167-168 °C; NMR satisfactory. Anal. (C13H8F3NO4S)
C,H,N,F,S.
2.33 (s, 3H, CH3), 3.34 (t, 1H, J ) 2.2 Hz, tCH), 4.50 (d, 2H,
J ) 2.2 Hz, CH2-Ct), 4.93 (s, 2H, CH2), 7.07 (d, 2H, J ) 9.2
Hz, aromatic), 7.56 (d, 1H, J ) 8.4 Hz, H8), 7.68 (dd, 1H, J )
8.4, 2.0 Hz, H7), 7.83 (d, 2H, J ) 9.2 Hz, aromatic), 7.96 (d,
1H,
J
)
2.0 Hz, H5), 12.22 (br s, 1H, NH). Anal.
(C20H16F3N3O3S) C,H,N,F.
4-Am in o-2-(tr iflu or om eth yl)ph en yl P h en yl Su lfon e (12).
The nitro compound 11 (7.00 g, 21 mmol) was added during 4
min to a warm solution of SnCl2‚2H2O (28.43 g, 126 mmol) in
EtOH (33 mL). Heating for 8 min at 54 °C gave a solution to
which concentrated HCl (33 mL) was added. The ethanol was
boiled away during 20 min; the residue was cooled in ice and
treated with 2.5 N NaOH to pH 14 to precipitate the product
which was filtered off, washed with water, and dried (5.62 g,
89%)-suitable for further use. The analytical sample was
obtained by flash chromatography (3% CH3CN in CH2Cl2) as
a white solid: mp 186-188 °C; NMR satisfactory. Anal.
(C13H10F3NO2S) C,H,N,F,S.
N-P r op -2-yn yl-4-(p h en ylsu lfon yl)-3-(tr iflu or om eth yl)-
a n ilin e (15): prepared from 12 (Supporting Information).
4-((Tr iflu or om eth yl)su lfon yl)flu or oben zen e (13). Tri-
fluoromethanesulfonyl anhydride (50 g, 177 mmol) was sy-
ringed during 10 min into a mixture of fluorobenzene (49.84
mL, 531 mmol) and AlCl3 (29.37 g, 220 mmol) stirred at 0 °C
under argon. The resulting orange-colored mixture was stirred
for 10 min at 0 °C and then for 23 h at 25 °C. The mixture
was partitioned between H2O (1000 mL) and ether (2 × 400
mL). The combined ether layers were washed with H2O (2 ×
500 mL) and brine (800 mL) and dried (Na2SO4) to give the
crude product as an orange-colored oil (17.5 g). This was
chromatographed on SiO2 (1 kg) using ether/CCl4 (1.5/98.5)
as the eluant. Fractions containing the second component to
elute were pooled and evaporated to give the technical quality
product (3.69 g, 9.1%). This was 88% pure: NMR (CDCl3) δ
7.37 (dd, 2H, J H ) 8.9 Hz, J F ) 8.1 Hz, H2, H6), 8.09 (dd, 2H,
J H ) 8.9 Hz, J F ) 4.9 Hz, H3, H5). The single impurity was
12% of 3-((trifluoromethyl)sulfonyl)fluorobenzene: NMR δ 7.55
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 lfon yl)-N-p r op -2-yn yl-3-(t r iflu or om et h yl)a -
n ilin e (18). The above procedure was repeated starting from
the aniline 15 (1.27 g, 3.74 mmol) in DMF (20 mL) with
reagents in the following quantities: 60% sodium hydride (0.15
g, 3.70 mmol), 1638 (1.36 g, 3.70 mmol) in DMF (20 mL),
NaHCO3aq (400 mL), and MeOH (90 mL)/1 N LiOH (9 mL, 9
mmol). Exhaustive extraction with CH2Cl2 (11 × 100 mL) was
needed to obtain the crude product as an amber oil (1.37 g).
Flash chromatography (50% CH3CN in CCl4) gave the pure
product as an off-white solid (0.207 g, 10.9%): mp 235.5-237
°C; NMR satisfactory. Anal. (C26H20F3N3O3S) C,H,N,F,S.
4-F lu or op h en yl P h en yl Su lfoxid e (19). To a rapidly
stirred solution of p-fluorodiphenyl sulfide42 (11.0 g, 53.9 mmol)
in HOAc (125 mL) at 25 °C was added 30% H2O2 (6.10 g, 53.9
mmol). The mixture was stirred for 28 h and then poured into
H2O (500 mL); extractive workup (Et2O) followed by flash
chromatography with 50% Et2O in petrol gave the desired
sulfoxide (11.8 g, 100%) identical with that previously de-
scribed.43
4-(N-P r op -2-yn yla m in o)p h en yl P h en yl Su lfoxid e (20).
To a rapidly stirred solution of 19 (2.0 g, 9.1 mmol) in dry
DMSO (15 mL) were added propargylamine (6.2 mL, 91 mmol,
10 equiv) and CaCO3 (1.0 g, 9.99 mmol). The mixture was
heated at 95 °C in a sealed tube for 18 h, at which time further
propargylamine (6.2 mL, 91 mmol) was added. The mixture
was heated at 125 °C for an additional 24 h and then poured
into H2O (500 mL); extractive workup (EtOAc) followed by
flash chromatography with 20% EtOAc in CH2Cl2 gave an off-
white solid (444 mg, 19%) which was then recrystallized from
CH2Cl2: mp 106-107 °C; NMR (CDCl3) δ 2.23 (t, 1H, J ) 2.4
Hz, CH), 3.93 (dd, 2H, J ) 6.0, 2.4 Hz, CH2), 4.29 (br s, 1H,
NH), 6.68 (d, 2H, J ) 8.8 Hz), 7.45 (m, 5H), 7.60 (d, 2H, J )
8.8 Hz); IR (KBr, cm-1) 3395 (s), 3250 (m), 1585 (m), 1320 (m),
1085 (m), 1025 (m), 823 (m). Anal. (C15H13NOS) C,H,N,S.
4-(N-((3,4-Dih yd r o-2-m eth yl-4-oxo-6-qu in a zolin yl)m e-
t h yl)-N-p r op -2-yn yla m in o)p h en yl P h en yl Su lfoxid e
(21). To a rapidly stirred solution of 20 (4.0 g, 15.7 mmol) in
DMF (40 mL) were added 6 (4.76 g, 18.8 mmol) and diisopro-
pylethylamine (3.4 mL, 23.6 mmol). The resulting mixture
was heated at 90 °C for 4 h. It was cooled and poured into
H2O (700 mL), and the resulting solid was filtered off. It was
coated onto SiO2 (20 g), from DMF (20 mL), and chromato-
graphed on SiO2 (500 g) with a gradient of 0-3.5% MeOH in
CH2Cl2 to give the product (4.4 g, 66%) as a white solid.
Recrystallization from 20% CH2Cl2 in Et2O gave an analyti-
cally pure sample: mp 219-221 °C; NMR (Me2SO-d6) δ 2.30
(s, 3H, CH3), 3.21 (br s, 1H, CH), 4.29 (s, 2H, CH2), 4.72 (s,
2H, CH2), 6.87 (d, 2H, J ) 9.0 Hz), 7.40-7.75 (m, 9H), 7.91 (s,
1H), 12.17 (s, 1H, NH); IR (KBr, cm-1) 3160 (m), 1670 (s), 1620
(m), 1590 (m), 1090 (w), 1018 (w). Anal. (C25H21N3O2S)
C,H,N,S.
(dddd, 1H, J 6,F ) 10.7 Hz, J 6,5 ) 7.7 Hz, J 6,4 ) 2.5 Hz, J 6,2
)
1.0 Hz, H6), 7.70 (ddd, 1H, J 5,4 ) J 5,6 ) 7.7 Hz, J 5,F ) 5.0 Hz,
H5), 7.76 (ddd, 1H, J 4,5 ) 7.7 Hz, J 4,6 ) 2.5 Hz, J 4,2 ) 2.2 Hz,
H4), 7.87 (br d, 1H, J 2,F ) 7.8 Hz, H2).
N-P r op-2-yn yl-4-((tr iflu or om eth yl)su lfon yl)an ilin e (14).
A solution of 13 (technical quality, 88%, 1.16 g, 4.47 mmol)
and propargylamine (0.51 g, 9.26 mmol) in DMSO (15 mL) was
stirred over K2CO3 (0.80 g, 5.79 mmol) at 25 °C for 24 h. The
mixture was poured into H2O (400 mL) and basified with 0.1
N NaOH to pH 11; extractive workup (Et2O) gave an es-
sentially pure amber oil which slowly crystallized (0.79 g, 67%).
A small sample was flash chromatographed using CH2Cl2 to
give the analytical sample as a white solid: mp 86-87 °C;
NMR (CDCl3) δ 2.31 (t, 1H, J ) 2.4 Hz, tCH), 4.04 (d, 2H, J
) 2.4 Hz, CH2-Ct), 4.4-5.2 (br, 1Η, ΝΗ), 6.77 (d, 2Η, J ) 8.9
Ηz, Η2, H6), 7.82 (d, 2H, J ) 8.9 Hz, H3, H5). Anal. (C10H8F3-
NO2S) C,H,N,F,S.
N-((3,4-Dih ydr o-2-m eth yl-4-oxo-6-qu in azolin yl)m eth yl)-
N-p r op -2-yn yl-4-((tr iflu or om eth yl)su lfon yl)a n ilin e (17).
Sodium hydride (60% oil dispersion, 0.71 g, 1.77 mmol) was
placed in a flame dried flask and covered with argon. The
propargylaniline 14 (0.453 g, 1.72 mmol) in DMF (9 mL) was
cannulated in during 1 min with stirring. An orange-red
solution formed which was stirred for 30 min. To it was added
6-(bromomethyl)-3, 4-dihydro-2-methyl-4-oxo-3-((pivaloyloxy)-
methyl)quinazoline (16;38 0.316 g, 0.86 mmol) in DMF (10 mL),
and the solution was stirred at 25 °C for 22.5 h. The mixture
was poured into saturated NaHCO3aq (100 mL) and extracted
with CH2Cl2 (4 × 100 mL). The combined organic layers were
dried (Na2SO4) and evaporated to dryness under oil pump
vacuum. The brown residue was treated with MeOH (20 mL)
followed by 1 N LiOH (2 mL, 2 mmol), and the resulting
mixture was stirred for 37 min, acidified to pH 3 with 1 N
HCl, poured into H2O (100 mL), and extracted with CH2Cl2 (4
× 50 mL). The combined, dried extract was taken to dryness
to give the crude product (0.350 g) which was loaded onto a
silica gel column (60 g) in 1,2-dichloroethane and eluted off
with 40% CH3CN in CCl4 to give the pure product as a white
solid (0.127 g, 34%): mp 250-251.5 °C; NMR (Me2SO-d6) δ
N-((4-Nitr op h en yl)su lfon yl)m or p h olin e (22). Morpho-
line (8.6 mL, 98.73 mmol) was added dropwise to a stirred
solution of 4-nitrobenzenesulfonyl chloride (10.4 g, 46.93 mmol)
in CH2Cl2 (150 mL) at 0 °C. After stirring for 10 min, 0.5 N
HCl (150 mL) was added; extractive workup with CH2Cl2 gave
the product as a light brown solid (12.13 g, 95%): mp 172-
174 °C (lit.44 mp 172.5-173 °C); NMR (CDCl3) δ 3.05 (m, 4H),
3.76 (m, 4H), 7.94 (d, 2H, J ) 9.0 Hz), 8.41 (d, 2H, J ) 9.0
Hz). Anal. (C10H12N2O5S) C,H,N,S.
N-((4-Am in op h en yl)su lfon yl)m or p h olin e (23). A solu-
tion of 22 (9.2 g, 33.79 mmol) in CH3OH (250 mL) containing
concentrated HCl (1.4 mL, 16.90 mmol) and with 5% Pd on
activated carbon (1.0 g) in suspension, in a Parr apparatus,
was shaken under 40 psi of hydrogen gas for 18 h. The
reaction mixture was filtered through Celite, the catalyst
washed with CH3OH (500 mL), and the solvent removed under
reduced pressure. Extractive workup (EtOAc) gave the prod-
uct as a pale yellow solid (7.5 g, 92%): mp 210-213 °C (lit.45

914 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4
J ones et al.
mp 217 °C); NMR (Me2SO-d6) δ 2.77 (t, 4H, J ) 4.6 Hz), 3.62
(t, 4H, J ) 4.6 Hz), 6.15 (s, 2H), 6.67 (d, 2H, J ) 8.7 Hz), 7.35
(d, 2H, J ) 8.7 Hz). Anal. (C10H14N2O3S) C,H,N,S.
4-((4-(N-((3,4-Dih yd r o-2-m eth yl-4-oxo-6-qu in a zolin yl)-
m eth yl)-N-p r op -2-yn yla m in o)p h en yl)su lfon yl)-1-(ter t-bu -
tyloxyca r bon yl)p ip er a zin e (29). A solution of 28 (0.678 g,
1.79 mmol), 6 (0.452 g, 1.79 mmol), and 2,6-lutidine (0.31 mL,
2.66 mmol) in DMF (10 mL) was heated at 100 °C. At 1 h,
further 6 (0.452 g, 1.79 mmol) and lutidine (0.21 mL, 1.80
mmol) were added. At 2 h, further 6 (0.226 g, 0.89 mmol) and
lutidine (0.1 mL, 0.90 mmol) were added. At 4 h, the reaction
mixture was poured into 0.5 N HCl (75 mL); extractive workup
(EtOAc) followed by flash chromatography with benzene-
EtOAc (1:2) gave the desired product as an off-white solid (330
mg, 33%): mp 193-195 °C; IR (KBr, cm-1) 1680, 1590, 1160;
NMR (CDCl3) δ 1.4 (s, 9H), 2.31 (t, 1H, J ) 2.3 Hz), 2.55 (s,
3H), 2.95 (m, 4H), 3.49 (t, 4H, J ) 4.8 Hz), 4.18 (d, 2H, J )
2.3 Hz), 4.77 (s, 2H), 6.88 (d, 2H, J ) 9.0 Hz), 7.58 (d, 2H, J
) 9.0 Hz), 7.68 (s, 2H), 8.15 (s, 1H), 10.72 (s, 1H). Anal.
(C28H33N5O5S) C,H,N,S.
N-((4-(N-((3,4-Dih yd r o-2-m eth yl-4-oxo-6-qu in a zolin yl)-
m et h yl)-N-p r op -2-yn yla m in o)p h en yl)su lfon yl)p ip er a -
zin e (30). CF3COOH (0.50 mL, 6.49 mmol) was added to a
stirred solution of 29 (0.200 g, 0.363 mmol) in CH2Cl2 (4 mL).
After 1 h, volatile components were removed under reduced
pressure and the resulting residue was dissolved in EtOAc (25
mL) and washed with 2 N NaOH (10 mL). Exhaustive
extractive workup (EtOAc) followed by flash chromatography
with 10% CH3OH in CH2Cl2 gave the product as a white,
hygroscopic solid (150 mg, 92%): mp indeterminate; NMR
(CDCl3) δ 2.30 (t, 1H, J ) 2.3 Hz), 2.52 (s, 3H), 3.02 (br s, 8H),
4.13 (d, 2H, J ) 2.3 Hz), 4.82 (s, 2H), 6.93 (d, 2H, J ) 9.0 Hz),
7.62 (d, 2H, J ) 9.0 Hz), 7.66 (m, 2H), 7.88 (m, 1H), NH’s not
discernable; HRMS (C23H25N5O3S)+ calcd 451.1688, found
451.1678. HCl salt: mp 225-227 °C.
4-((4-(N-((3,4-Dih yd r o-2-m eth yl-4-oxo-6-qu in a zolin yl)-
m et h yl)-N-p r op -2-yn yla m in o)p h en yl)su lfon yl)-1-m et h -
ylp ip er a zin e (31). Iodomethane (0.017 mL, 0.273 mmol) was
added to a stirred solution of 30 (123 mg, 0.272 mmol) and
N,N-diisopropylethylamine (0.048 mL, 0.276 mmol) in DMF
(5 mL). After 18 h at 25 °C, the mixture was poured into H2O
(20 mL); extractive workup (CH2Cl2) followed by chromatog-
raphy with CH2Cl2-CH3OH (10:1) gave the product as a white
solid (75 mg, 59%): mp 139-141 °C; NMR (CDCl3) δ 2.26 (s,
3H), 2.31 (t, 1H, J ) 2.3 Hz), 2.47 (br t, 4H), 2.99 (br t, 4H),
4.18 (d, 2H, J ) 2.3 Hz), 4.77 (s, 2H), 6.88 (d, 2H, J ) 7.2 Hz),
7.61 (d, 2H, J ) 7.2 Hz), 7.68 (d, 2H, J ) 1.2 Hz), 8.19 (s, 1H),
11.09 (br s, 1H); HRMS (C24H27N5O3S)+ calcd 465.1835, found
465.1830. Anal. (C24H27N5O3S) C,H,N,S.
N-((4-(N-P r op -2-yn yla m in o)p h en yl)su lfon yl)m or p h o-
lin e (24). A solution of 23 (1.001 g, 4.14 mmol), N,N-
diisopropylethylamine (0.86 mL, 4.93 mmol), and propargyl
bromide (80% (w/w) in toluene, 0.46 mL, 4.14 mmol) in DMF
(10 mL) was heated at 75 °C. At 5 h, additional diisopropy-
lethylamine (0.36 mL, 2.5 mmol) and propargyl bromide (0.23
mL, 2.07 mmol) were added, and the mixture was heated for
3 h more. It was poured into 0.5 N HCl (200 mL), and
extractive workup (EtOAc) followed by flash chromatography
with CH2Cl2-EtOAc (20:1) gave the desired product as a pale
yellow solid (0.47 g, 41%): mp 117-120 °C; NMR (CDCl3) δ
2.27 (t, 1H, J ) 2.5 Hz), 2.97 (m, 4H), 3.73 (m, 4H), 4.00 (dd,
2H, J ) 2.5, 6.0 Hz), 4.47 (m, 1H), 6.71 (d, 2H, J ) 8.8 Hz),
7.58 (d, 2H, J ) 8.8 Hz). Anal. (C13H16N2O3S) C,H,N,S.
N-((4-(N-((3,4-Dih yd r o-2-m eth yl-4-oxo-6-qu in a zolin yl)-
m et h yl)-N-p r op -2-yn yla m in o)p h en yl)su lfon yl)m or p h o-
lin e (25). A solution of 24 (1.314 g, 4.69 mmol), 6 (1.185 g,
4.69 mmol), and 2,6-lutidine (0.85 mL, 7.3 mmol) in DMF (15
mL) was heated at 100 °C. At 1.5 h, further 6 (0.885 g, 3.52
mmol) and 2,6-lutidine (0.41 mL, 3.52 mmol) were added. At
4.5 h, the reaction mixture was poured into 0.5 N HCl (150
mL), and extractive workup (EtOAc) followed by flash chro-
matography with 50% EtOAc in CH2Cl2 gave the desired
product as an off-white solid (0.630 g, 30%): mp 222-225 °C
slight dec; NMR (Me2SO-d6) δ 2.31 (s, 3H), 2.77 (m, 4H), 3.26
(t, 1H, J ) 1.9 Hz), 3.59 (m, 4H), 4.37 (d, 2H, J ) 1.7 Hz),
4.81 (s, 2H), 6.95 (d, 2H, J ) 9.0 Hz), 7.52 (m, 3H), 7.68 (dd,
1H, J ) 1.9, 8.3 Hz), 7.96 (d, 1H, J ) 1.5 Hz), 12.18 (s, 1H);
HRMS (C23H24N4O4S)+ calcd 452.1518, found 452.1497. Anal.
(C23H24N4O4S).
4-((4-Nitr op h en yl)su lfon yl)-1-(ter t-bu tyloxyca r bon yl)-
p ip er a zin e (26). To a stirred solution of tert-butyl 1-pipera-
zinecarboxylate (3.069 g, 16.48 mmol) and diisopropylethy-
lamine (3.16 mL, 18.13 mmol) in CH2Cl2 (80 mL) at 0 °C was
slowly added a solution of 4-nitrobenzenesulfonyl chloride
(3.652 g, 16.48 mmol) in CH2Cl2 (50 mL). When the addition
was complete, the reaction mixture was allowed to warm to
room temperature, and then it was poured into 0.5 N HCl (75
mL); extractive workup (CH2Cl2) gave the desired sulfonamide
as a tan solid (5.78 g, 94%): mp 169-171 °C; NMR (CDCl3) δ
1.40 (s, 9H), 3.03 (t, 4H, J ) 5.0 Hz), 3.53 (t, 4H, J ) 5.1 Hz),
7.94 (d, 2H, J ) 8.8 Hz), 8.39 (d, 2H, J ) 8.8 Hz). Anal.
(C15H21N3O6S) 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-(p h en ylsu lfon yl)a n ilin e (32). A mixture of 4-aminophenyl
phenyl sulfone (4n ; 7.0 g, 30 mmol), 6 (technical grade, 7.91
g, 25 mmol), 2,6-lutidine (2.95 g, 28 mmol), and DMA (50 mL)
was stirred under argon at 60 °C for 17 h. The solvent was
evaporated (90 °C/2 mmHg) to leave a yellow solid which was
flash chromatographed on SiO2 (1 kg) with 6% MeOH in EtOAc
to give the pure product as a tan powder (4.33 g, 43%): mp
4-((4-Am in oph en yl)su lfon yl)-1-(ter t-bu tyloxycar bon yl)-
p ip er a zin e (27). A solution of 26 (5.00 g, 13.46 mmol) in
EtOAc (50 mL) with 5% Pd on activated carbon (1.0 g) in
suspension, in a Parr apparatus, was shaken under 44 psi of
hydrogen gas for 18 h. At this time further catalyst (0.60 g)
was added, and the mixture was hydrogenated at 44 psi for a
further 4 h. The reaction mixture was filtered through Celite,
the catalyst washed with EtOAc (1 L), and the solvent removed
under reduced pressure to give the desired aniline as a white
solid (4.58 g, ∼100%): mp 194-197 °C; NMR (CDCl3) δ 1.40
(s, 9H), 2.92 (t, 4H, J ) 5.0 Hz), 3.49 (t, 4H, J ) 5.1 Hz), 4.15
(s, 2H), 6.69 (d, 2H, J ) 8.7 Hz), 7.50 (d, 2H, J ) 8.7 Hz);
HRMS (C15H23N3O4S)+ calcd 341.1409, found 341.1419. This
material was used without further purification.
4-((4-(P r op -2-yn yla m in o)p h en yl)su lfon yl)-1-(ter t-b u -
tyloxyca r bon yl)p ip er a zin e (28). A solution of the aniline
27 (1.50 g, 4.39 mmol), propargyl bromide (80% (w/w) in
toluene, 0.49 mL, 4.39 mmol), and N,N-diisopropylethylamine
(0.80 mL, 4.61 mmol) in DMF (20 mL) was heated at 75 °C.
At 4 h, additional propargyl bromide (0.49 mL, 4.39 mmol)
and diisopropylethylamine (0.80 mL, 4.61 mmol) were added,
and the mixture was heated for 4 h more. It was poured into
0.5 N HCl (100 mL); extractive workup (EtOAc) followed by
flash chromatography with CH2Cl2-EtOAc (20:1) gave the
desired propargylamine as a pale yellow solid (0.923 g, 55%):
mp 142-143 °C; IR (KBr, cm-1) 3390, 1680, 1595, 1325, 1160;
NMR (CDCl3) δ 1.41 (s, 9H), 2.27 (t, 1H, J ) 2.5 Hz), 2.94 (t,
4H, J ) 5.0 Hz), 3.49 (t, 4H, J ) 5.0 Hz), 4.00 (dd, 2H, J )
2.3, 5.6 Hz), 4.45 (t, 1H, J ) 5.7 Hz), 6.71 (d, 2H, J ) 8.8 Hz),
7.57 (d, 2H, J ) 8.8 Hz). Anal. (C18H25N3O4S) C,H,N,S.
300-301 °C; NMR (Me2SO-d6) satisfactory.
(C22H19N3O3S) C,H,N,S.
Anal.
4-(P h en ylsu lfon yl)-N-eth yla n ilin e (33). Meth od A. A
mixture of 4-aminophenyl phenyl sulfone (4n ; 5.83 g, 25 mmol),
iodoethane (7.80 g, 50 mmol), K2CO3 (3.45 g, 25 mmol), and
DMA (125 mL) was stirred at 110 °C for 3.5 h. The DMA was
removed in vacuo at 80 °C leaving a white residue; extractive
workup (EtOAc/H2O) (500 mL) gave an oil that was flash
chromatographed using 10% MeOH in toluene to give the pure
product as a white solid (3.17 g, 48.5%): mp 134.5-136 °C;
NMR satisfactory. Anal. (C14H15NO2S) C,H,N,S.
Meth od B. A stirred solution of 4-(phenylsulfonyl)fluo-
robenzene (25.00 g, 0.106 mol) and C2H5NH2 (20.8 mL, 0.318
mol, 3 equiv) in DMSO (125 mL) in a flask fitted with a dry
ice condenser was heated at 130 °C for 3 h under argon. The
resulting amber solution was allowed to cool, poured into H2O
(1.3 L), and extracted with CH2Cl2 (2 × 200 mL) and finally
with EtOAc (2 × 200 mL). The organic layers were combined,
dried, and evaporated in vacuo to give the crude beige-colored
product. Recrystallization from ethanol gave off-white crystals
(23.03 g, 83.1%) identical by TLC with the product prepared
as above.

Lipophilic Quinazoline Inhibitors of TS
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4 915
4-(E t h yla m in o)t r iflu or oa cet op h en on e (34). A clear
solution of 4-fluorotrifluoroacetophenone46 (10.0 g, 52 mmol)
and C2H5NH2 (7.04 g, 156 mmol, 3 equiv), in DMSO (50 mL)
in a flask fitted with a dry ice condenser was stirred at 140
°C for 1 h 40 min under an inert atmosphere. The mixture
was poured into H2O (500 mL); extractive workup (CH2Cl2)
gave a yellow oil which was flash chromatographed with CH2-
Cl2 as eluant to give the pure product as a yellow solid (6.87
g, 61%): mp 80-81 °C; NMR satisfactory. Anal. (C10H10F3-
NO) C,H,N,F.
4-(N-((3,4-Dih yd r o-2-m eth yl-4-oxo-6-qu in a zolin yl)m e-
th yl)-N-eth yla m in o)p h en yl P h en yl Su lfon e (35). A mix-
ture of 33 (1.41 g, 5.4 mmol), 6 (1.51 g, 6 mmol), CaCO3 (0.060
g, 6 mmol), and DMA (25 mL) was stirred at 100 °C. Further
6 was added at 2 h (0.1 g) and at 3 h (0.15 g). At 6 h the
mixture was cooled and then filtered through Celite. The
solution was slowly pipetted into H2O (700 mL) to produce a
cream-colored flocculent solid. This was filtered off, washed
carefully with water, dried over P2O5 in vacuo, and purified
by flash column chromatography twice (300 g of SiO2, 15%
MeOH in benzene; 200 g of SiO2, 6% MeOH in CH2Cl2) to
give the product as a cream-colored solid (0.87 g, 37%): mp
CH3CN (4 mL), Et3N (0.56 mL, 4.0 mmol), and P2S5 (0.445 g,
2.0 mmol) while under an argon atmosphere. The stirred
mixture was heated to 50 °C under reflux whereupon most of
the P2S5 dissolved. Heating was continued overnight. The
mixture was cooled and poured into H2O (100 mL). A solid
precipitated which was filtered off, washed with cold H2O,
dried over P2O5 in vacuo, and flash chromatographed (2% CH3-
CN in CH2Cl2) to give the thiolactam as a pale yellow solid
(0.368 g, 95%): mp 232-234 °C; Rf ) 0.34 (silica gel, 5% CH3-
CN in CH2Cl2); NMR (Me2SO-d6) δ 2.44 (s, 3H, CH3), 3.24 (t,
1H, J ) 2.1 Hz, tCH), 4.33 (d, 2H, J ) 2.1 Hz, CH2-Ct), 4.79
(s, 2H, N-CH2-Ph), 7.00-7.1 (m, 3H, aromatic), 7.39 (t, 1H, J
) 8.25 Hz, aromatic), 7.61 (d, 1H, J ) 8.4 Hz, H8), 7.76 (dd,
1H, J ) 8.4, 2.0 Hz, H7), 8.48 (d, 1H, J ) 2.0 Hz, H5), 13.68
(br s, 1H, NH). Anal. (C20H16F3N3S) C,H,N,S.
P r otein Cr ysta lliza tion . Small crystals (∼0.1 mm) were
grown by vapor diffusion using 1.5 M sodium/potassium
phosphate as the precipitant according to the protocol de-
scribed.24 Fresh protein was then concentrated to 27 mg/mL
in 0.002 M Tris buffer, 0.66 M phosphate, and 1% poly-
(ethylene glycol) (PEG400) at pH 8.0; 10 µL drops were placed
on plastic cover slips and inverted over wells containing 0.005
M Tris, 2.0 M phosphate, and 1% PEG400 at pH 8.0. After
24 h of equilibration at 4 °C, a single 0.1 mm seed crystal was
introduced into each drop. Over the course of 1-2 weeks the
seed crystals grew to final dimensions of around 0.8 mm on
an edge.
Bioch em ica l Assa ys. TS activity was assayed by a modi-
fied procedure of the tritium release method of Lomax and
Greenberg.47 Inhibition constants Kis and Kii were as described
by Cleland48 and determined by steady state analysis against
5,10-methylenetetrahydrofolate as the variable cosubstrate
under conditions of saturating dUMP. Values listed are
averaged over multiple data sets. Reaction conditions in 0.1
mL were 50 mM Tris (pH 7.6), 10 mM dithiothreitol, 1 mM
ethylenediaminetetraacetic acid, 25 mM MgCl2, 15 mM form-
aldehyde, 25 µM dUMP ([5-3H]; specific activity = 2 × 108 cpm/
µmol), and tetrahydrofolate (eight concentrations ranging from
5 to 150 µM). Bovine serum albumin at up to 100 µg/mL was
present when human TS was assayed. These reactions were
conducted either in the absence of inhibitor or in the presence
of inhibitor at concentrations ranging, at a minimum, between
0.5 Ki and 2.0 Ki except when the solubility of the inhibitor
was limiting. Reactions were run at room temperature by
initiating with the addition of enzyme. After 5 min, the
reactions were quenched by the addition of charcoal, the
mixtures were centrifuged to remove unreacted dUMP, and
the supernatant was counted to determine the release of
tritium from the 5-position of dUMP. Experimental results
were analyzed by a nonlinear regression analysis program49
which fit the data to a mixed noncompetitive inhibition
scheme.
Mea su r em en t of Tissu e Cu ltu r e IC50’s. IC50 values for
the inhibition of cellular growth were measured using a
modification50 of the MTT51 colorimetric assay of Mosmann52
using mouse (L1210) and human (CCRF-CEM) leukemia lines
(ATCC) and a human adenocarcinoma (GC3/M TK-) deficient
in thymidine kinase.53 Cells were seeded at 1000 (L1210) or
10 000 (CCRF-CEM, GC3/M TK-) cells/well in 96-well plates,
and growth was measured over a range of nine 2-fold serial
dilutions of each compound. Culture medium (RPMI-1640)
contained 5% (L1210, CCRF-CEM) or 10% (GC3/M TK-) fetal
calf serum and 0.5% DMSO. Following a 3 day (L1210) or 5
day (CCRF-CEM, GC3/M TK-) incubation and a 4 h treatment
with MTT, cells were harvested and growth was measured
spectrophotometrically after dissolution of the deposited for-
mazan in DMSO. IC50 values were determined from semi-
logarithmic plots of compound concentration vs the mean of
the four growth assessments made at each serial dilution of
the agent relative to the growth of control cultures.
193-194 °C; NMR (Me2SO-d6) satisfactory.
(C24H23N3O3S‚0.5H2O) C,H,N,S.
Anal.
4-(N-((3,4-Dih yd r o-2-m eth yl-4-oxo-6-qu in a zolin yl)m e-
th yl)-N-eth yla m in o)tr iflu or oa cetop h en on e (36). A mix-
ture of 34 (2.17 g, 10 mmol), 6 (3.23 g, 12.8 mmol), CaCO3
(1.28 g, 12.8 mmol), and DMA (30 mL) was stirred under argon
at 100-110 °C for 3.5 h. The orange-colored mixture was
filtered through Celite and the solvent then evaporated at 90
°C (2 mmHg), to leave an amber resin which was flash
chromatographed with 40% CH3CN in CCl4 to give the pure
product as an off-white solid (0.870 g, 22%): mp 273.5-275.5
°C; NMR (Me2SO-d6) satisfactory. Anal. (C20H18F3N3O2)
C,H,N,F.
(4-(N-((3,4-Dih yd r o-2-m eth yl-4-oxo-6-qu in a zolin yl)m -
eth yl)a m in o)p h en yl)p h en ylm eth a n e (37). A solution of
4-(phenylmethyl)aniline (0.46 g, 2.53 mmol), 6 (0.64 g, 2.53
mmol), and N,N-diisopropylethylamine (0.36 g, 2.78 mmol, 1.1
equiv) in DMA (15 mL) was stirred at 85-90 °C for 1.5 h and
then pipetted into rapidly stirring H2O (150 mL). The result-
ing beige-colored solid was filtered off, washed with H2O (25
mL), and then dissolved in boiling EtOH (50 mL). This
solution was charcoal treated, filtered, brought back to the boil,
diluted with H2O (25 mL), and cooled in ice for 15 min. The
resulting pale yellow, flocculent solid was filtered off and
washed with EtOH (25 mL) followed by Et2O (25 mL). This
slightly impure product (0.30 g) was coated onto SiO2 (1.50 g)
from 50% MeOH in CH2Cl2 (25 mL) and flash chromato-
graphed with 5% MeOH in EtOAc to give the pure product as
a cream-colored solid (0.118 g, 13%): mp 223-226 °C dec;
NMR (Me2SO-d6) δ 2.32 (s, 3H, CH3), 3.72 (s, 2H, CH2), 4.34
(d, 2H, J ) 5.2 Hz, CH2), 6.26 (t, 1H, J ) 5.2 Hz, NH), 6.48 (d,
2H, J ) 8.4 Hz), 6.88 (d, 2H, J ) 8.4 Hz), 7.10-7.27 (m, 5H,
phenyl), 7.51 (d, 1H, J ) 8.4 Hz, H8), 7.72 (dd, 1H, J ) 8.4,
1.9 Hz, H7), 8.02 (d, 1H, J ) 1.9 Hz, H5), 12.14 (s, 1H, lactam
NH). Anal. (C23H21N3O‚0.3H2O) C,H,N.
(4-(N-((3,4-Dih yd r o-2-m eth yl-4-oxo-6-qu in a zolin yl)m -
eth yl)m eth yla m in o)p h en yl)p h en ylm eth a n e (38). A sus-
pension of 37 (108 mg, 0.3 mmol) in EtOH (5 mL) was treated
with formaldehyde (37%, 142 mg, 1.7 mmol) and then NaC-
NBH3 (27 mg, 0.4 mmol). HC1aq (0.1 N) was added dropwise
during 45 min to maintain the pH at just below 7. The
suspension was stirred at 25 °C for 15 h; 1 N HC1 (2 mL) was
added to quench excessive hydride. The resulting precipitate
was filtered off, washed, and dried to give the product as an
off-white solid (87 mg, 79%): mp 225 °C dec; NMR (Me2SO-
d6) δ 2.27 (s, 3H, CH3), 2.94 (s, 3H, N-CH3), 3.74 (s, 2H, CH2),
4.57 (s, 2H, N-CH2), 6.62 (d, 2H, J ) 8.5 Hz), 6.96 (d, 2H, J )
8.5 Hz), 7.09-7.23 (m, 5H, phenyl), 7.48 (d, 1 H, J ) 8.3 Hz,
H8), 7.55 (dd, 1H, J ) 8.3, 1.5 Hz, H7), 7.82 (d, 1H, J ) 1.5 Hz,
H5), 12.17 (br s, 1H, NH); HRMS (C24H23N3O)+ calcd 369.1841,
found 369.1843. Anal. (C24H23N3O‚0.6H2O) C,H,N.
Mea su r em en t of IC50 Sh ift Du e to Th ym id in e. The
ability of thymidine to reverse cell growth inhibition induced
by inhibitors was assessed in L1210 cells by comparing the
IC50 measured under standard conditions (RPMI-1640 medium
containing 5% fetal calf serum) with that obtained in the
presence of 10 µM thymidine which was replenished daily
N-((3,4-Dih ydr o-2-m eth yl-4-th io-6-qu in azolin yl)m eth yl)-
N-p r op -2-yn yl-3-(tr iflu or om eth yl)a n ilin e (39). In a flame-
dried two-necked flask were placed 7f (0.371 g, 1.0 mmol),

916 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4
J ones et al.
(13) J ackman, A. L.; Taylor, G. A.; Gibson, W.; Kimbell, R.; Brown,
M.; Calvert, A. H.; J udson, I. R.; Hughes, L. R. ICI D1694, a
quinazoline antifolate thymidylate synthase inhibitor that is a
potent inhibitor of L1210 tumor cell growth In vitro and in
vivo: A new agent for clinical study. Cancer Res. 1991, 51, 5579-
5586.
(14) Clarke, S.; Ward, J .; Planting, A.; Spiers, J .; Smith, R.; Verweij,
J .; J udson, I. Phase I trial of ICI D1694: A novel thymidylate
synthase inhibitor. Proc. Am. Assoc. Cancer Res. 1992, 33, 406.
(15) Hagan, R. L.; Duch, D. S.; Smith, G. K.; Hanlon, M. H.; Shane,
B.; Freisheim, J . H.; Hynes, J . B. Studies on the mechanism of
antitumor action of 2-desamino-2-methyl-5,8-dideazaisofolic acid.
Biochem. Pharmacol. 1991, 41, 781-787.
during the 3 days of growth. The magnitude of the ratio of
the IC50 measured in the presence of thymidine to that
measured without added nucleoside was used as a measure
of the extent to which the inhibition of growth could be
attributed to intracellular inhibition of thymidylate synthase.
A value of e1.0 under these conditions would suggest a locus
of action other than TS, while larger values probably indicate
a direct relationship between growth inhibition and TS target-
ing.
(16) Duch, D. S.; Banks, S.; Dev, I. K.; Dickerson, S. H.; Ferone, R.;
Heath, L. S.; Humphreys, J .; Knick, V.; Pendergast, W.; Singer,
S.; Smith, G. K.; Waters, K.; Wilson, H. R. Biochemical and
cellular pharmacology of 1843U89, a novel benzoquinazoline
inhibitor of thymidylate synthase. Cancer Res. 1993, 53, 810-
818.
(17) Albrecht, A. M.; Biedler, J . L. Acquired resistance of tumor cells
to folate antagonists. In Folate Antagonists as Therapeutic
Agents, Vol 1; Sirotnak, F. M., Burchall, J . J ., Ensminger, W.
B., Montgomery, J . A., Eds.; Academic Press: San Diego, 1984;
pp 317-353.
Ack n ow led gm en t. We thank Dr. Bill Scott of
Nachem, Inc., for his generous gift of 4-amino-2-chlo-
robenzonitrile. We are grateful to Peter J ohnson and
Dr. David Henry for their support and encouragement
of this work.
Su p p or tin g In for m a tion Ava ila ble: One table giving the
preparation and physical properties of the propargylanilines
5a -c,f-i,k ,m -t, and 15 (3 pages). Ordering information is
given on any current masthead page.
(18) Curt, G. A.; J olivet, J .; Carney, D. N.; Bailey, B. D.; Drake, J .
C.; Clendennin, N. J .; Chabner, B. A. Determinants of the
sensitivity of human small-cell lung cancer cell lines to meth-
otrexate. J . Clin. Invest. 1985, 76, 1323-1329.
(19) McCloskey, D. E.; McGuire, J . J .; Russell, C. A.; Rowan, B. G.;
Bertino, J . R.; Pizzorno, G.; Mini, E. Decreased folylpoly-
glutamate synthetase activity as a mechanism of methotrexate
resistance in CCRF-CEM human leukemia sublines. J . Biol.
Chem. 1991, 266, 6181-6187.
(20) Rumberger, B. G.; Schmid, F. A.; Otter, G. M.; Sirotnak, F. M.
Preferential selection during therapy in vivo by edatrexate
compared to methotrexate of resistant L1210 cell variants with
decreased folylpolyglutamate synthetase activity. Cancer Com-
mun. 1990, 2, 305-310.
(21) O’Connor, B. M.; J ackman, A. L.; Crossley, P. H.; Freemantle,
S. E.; Lunec, J .; Calvert, A. H. Human lymphoblastoid cells with
acquired resistance to C2-desamino-C2-methyl-N10-propargyl-5,8-
dideazafolic acid: A novel folate-based thymidylate synthase
inhibitor. Cancer Res. 1992, 52, 1137-1143.
(22) 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.
Refer en ces
(1) Berman, E. M.; Werbel, L. M. The renewed potential for folate
antagonists in contemporary cancer chemotherapy. J . Med.
Chem. 1991, 34, 479-485.
(2) J ones, T. R.; Calvert, A. H.; J ackman, A. L.; Brown, S. J .; J ones,
M.; Harrap, K.R. A potent antitumour quinazoline inhibitor of
thymidylate synthetase: Synthesis, biological properties and
therapeutic results in mice. Eur. J . Cancer 1981, 17, 11-19.
(3) J ones, T. R.; Calvert, A. H.; J ackman, A. L.; Eakin, M. A.;
Smithers, M. J .; Betteridge, R. F.; Newell, D. R.; Hayter, A. J .;
Stocker, A.; Harland, S. J .; Davies, L.C.; Harrap, K. R. Quinazo-
line antifolates inhibiting thymidylate synthase: Variation of
the N10 substituent. J . Med. Chem. 1985, 28, 1468-1476.
(4) Calvert, A. H.; Alison, D. L.; Harland, S. J .; Robinson, B. A.;
J ackman, A. L.; J ones, T. R.; Newell, D. R.; Siddik, Z. H.;
Wiltshaw, E.; McElwain, T. J .; Smith, I. E.; Harrap, K. R. A
phase I evaluation of the quinazoline antifolate thymidylate
synthase inhibitor, N10-propargyl-5,8-dideazafolic acid, CB3717.
J . Clin. Oncol. 1986, 4, 1245-1252.
(5) J odrell, D. I.; Newell, D. R.; Morgan, S. E.; Clinton, S.; Bensted,
J . P. M.; Hughes, L. R.; Calvert, A. H. The renal effects of N10
-
propargyl-5,8-dideazafolic acid (CB3717) and a non-nephrotoxic
analogue ICI D1694 in mice. Br. J . Cancer 1991, 64, 833-838.
(6) J ones, T. R.; Thornton, T. J .; Flinn, A.; J ackman, A. L.; Newell,
D. R.; Calvert, A. H. Quinazoline antifolates inhibiting thymidy-
late synthase: 2-Desamino derivatives with enhanced solubility
and potency. J . Med. Chem. 1989, 32, 847-852.
(7) J ackman, A. L.; Newell, D. R.; Taylor, G. A.; O’Connor, B.;
Hughes, L. R.; Calvert, A. H. 2-Desamino-10-propargyl-5,8-
dideazafolic acid (desamino-CB3717), a thymidylate synthase
inhibitor devoid of renal and hepatic toxicities in mice. Proc. Am.
Assoc. Cancer Res. 1987, 28, 271.
(8) 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 thymidy-
late synthase inhibitors: Alkyl, substituted alkyl, and aryl
substituents in the C2 position. J . Med. Chem. 1990, 33, 3060-
3067.
(9) J ackman, A. L.; Newell, D. R.; Gibson, W.; J odrell, D. I.; Taylor,
G. A.; Bishop, J . A.; Hughes, L. R.; Calvert, A. H. The biochemi-
cal pharmacology of the thymidylate synthase inhibitor, 2-de-
samino-2-methyl-N10-propargyl-5,8-dideazafolic acid (ICI 198583).
Biochem. Pharmacol. 1991, 42, 1885-1895.
(10) Moran, R. G.; Colman, P. D.; Rosowsky, A.; Forsch, R. A.; Chan,
K. K. Structural features of 4-amino antifolates required for
substrate activity with mammalian folylpolyglutamate syn-
thetase. Mol. Pharmacol. 1985, 27, 156-166.
(11) Sikora, E.; J ackman, A. L.; Newell, D. R.; Calvert, A. H.
Formation and retention and biological activity of N10-propargyl-
5,8-dideazafolic acid (CB3717) polyglutamates in L1210 cells in
vitro. Biochem. Pharmacol. 1988, 37, 4047-4054.
(12) J ackman, A. L.; Taylor, G. A.; O’Connor, B. M.; Bishop, J . A.;
Moran, R. G.; Calvert, A. H. Activity of the thymidylate synthase
inhibitor 2-desamino-N10-propargyl-5,8-dideazafolic acid and
related compounds in murine (L1210) and Human (W1L2)
systems in vitro and in L1210 in vivo. Cancer Res. 1990, 50,
5212-5218.
(23) 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-didea-
zafolate. J . Mol. Biol. 1990, 214, 923-936.
(24) All small molecules were minimized using the semiempirical
AM1 method: Dewar, M. J . S.; Zoebisch, E. G.; Healy, E. F.;
Stewart, J . J . P. AM1: A new general purpose quantum
mechanical molecular model. J . Am. Chem. Soc. 1985, 107,
3902-3909. Modeling was conducted using Polygen Corp. Quanta
program on a Silicon Graphics 4D70GT workstation.
(25) Fitzhugh, A. L.; Fodor, S.; Kaufman, S.; Spiro, T. G. Resonance
Raman spectroscopic evidence for alternative structures in the
native ternary complex formed with thymidylate synthase. J .
Am. Chem. Soc. 1986, 108, 7422-7424.
(26) McClellan, A. L. Tables of Experimental Dipole Moments, Volume
II; Rahara Enterprises: El Cerrito, CA, 1974.
(27) Burley, S. K.; Petsko, G. A. Aromatic-aromatic interaction: A
mechanism of protein structure stabilization. Science 1985, 229,
23-28.
(28) Singh, J .; Thornton, J . M. The interaction between phenylalanine
rings in proteins. FEBS Lett. 1985, 191, 1-6.
(29) Hiberty, P. C.; Ohanessian, G. Valence-Bond Description of
Conjugated Molecules. 3. The through-resonance concept in
para-substituted nitrobenzenes. J . Am. Chem. Soc. 1984, 106,
6963-6968.
(30) Waldron, W. R.; Reid, E. E. The influence of sulfur on the color
of azo dyes. J . Am. Chem. Soc. 1923, 45, 2399-2417.
(31) Friedman, L.; Shechter, H. Dimethylformamide as a useful
solvent in preparing nitriles from aryl halides and cuprous
cyanide; improved isolation techniques. J . Org. Chem. 1961, 26,
2522-2524.
(32) Brandsma, L.; Verkruijsse, H. D. Synthesis of Acetylenes, Allenes
and Cumulenes; Elsevier: Amsterdam, 1981; pp 223-224.
(33) Bourne, E. J .; Henry, S. H.; Tatlow, C. E. M.; Tatlow, J . C.
Studies of trifluoroacetic atcid. Part VI. Trifluoroacetyl deriva-
tives of amines. J . Chem. Soc. 1952, 4014-4019.

Lipophilic Quinazoline Inhibitors of TS
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 4 917
(34) Garito, A. F.; Horner, C. J .; Kalyanaraman, P. S.; Desai, K. N.
Synthesis of N-(nitrophenyl)amine substituted diacetylene mono-
mers. Makromol. Chem. 1980, 181, 1605-1612.
(35) Shein, S. M.; Solodova, K. V.; Govorchenko, V. I. Nucleophilic
substitution in the aromatic series XXXIII, Effect of the nature
of substituted atoms or groups on reaction rates for para-
substituted trifluoromethyl phenyl sulfones with sodium alco-
holates and piperidines. J . Gen. Chem. USSR 1969, 39, 1603-
1609.
(36) Yagupol’skii, L. M.; Gruz, B. E. Synthesis of some derivatives
of phenyl trifluoromethyl sulfide and phenyl trifluoromethyl
sulfone. J . Gen. Chem. USSR 1961, 31, 1219-1223.
(37) Hendrickson, J . B.; Bair, K. W. New methods for the synthesis
of triflones. J . Org. Chem. 1977, 42, 3875-3878.
(38) Hughes, L. R. Eur. Pat. Appl. 239 362 A2, September 30, 1987.
(39) McNamara, D. J .; Berman, E. M.; Fry, D. W.; Werbel, L. M.
Potent inhibition of thymidylate synthase by two series of
nonclassical quinazolines. J . Med. Chem. 1990, 33, 2045-2051.
(40) J ones, T. R.; Smithers, M. J .; Betteridge, R. F.; Taylor, M. A.;
J ackman, A. L.; Calvert, A. H.; Davies, L. C.; Harrap, K. R.
Quinazoline antifolates inhibiting thymidylate synthase: Varia-
tion of the amino acid. J . Med. Chem. 1986, 29, 1114-11189.
(41) Epstein, J .; Rosenthal, R. W.; Ess, R. J . Use of γ-(4-nitrobenzyl)-
pyridine as analytical reagent for ethylenimines and alkylating
agents. Anal. Chem. 1955, 27, 1435-1439.
(42) Leonard, N. J .; Sutton, L. E. The use of fluoro compounds in
the determination of valency angles by electric dipole moment
measurements. J . Am. Chem. Soc. 1948, 70, 1564-1571.
(43) Kaplan, L. J .; Martin, J . C. Sulfuranes. IX. Sulfuranyl substitu-
ent parameters. substituent effects on the reactivity of dialkoxy-
diarylsulfuranes in the dehydration of alcohols. J . Am. Chem.
Soc. 1973, 95, 793-798.
(44) Yung, D. K.; Forrest, T. P.; Manzer, A. R.; Gilroy, M. L. Reactions
of benzenesulfonohydrazides and benzenesulfonamides with
hydrogen chloride or hydrogen bromide in acetic acid. J . Pharm.
Sci. 1977, 66, 1009-1012.
(45) Adams, R.; Long, P. H.; J ohanson, A. J . Sulfanilamide deriva-
tives. J . Am. Chem. Soc. 1939, 61, 2342-2346.
(46) Herkes, F. E.; Burton, D. J . Fluoro olefins. I. The synthesis of
â-substituted perfluoro olefins. J . Org. Chem. 1967 32, 1311-
1318.
(47) Lomax, M. I. S.; Greenberg, G. R. A new assay of thymidylate
synthase activity based on the release of tritium from deox-
yuridylate-5-3H. J . Biol. Chem. 1967, 242, 109-113.
(48) Cleland, W. W. The kinetics of enzyme-catalysed reactions with
two or more substrates or products. II Inhibition: Nomenclature
and theory. Biochim. Biophys. Acta 1963, 67, 173-187.
(49) Perrella, F. W. EZ-FIT; Perrella Scientific: Springfield, PA, 1989.
(50) Alley, M. C.; Scuderio, D. A.; Monks, A.; Hursey, M. L.;
Czerwinski, M. J .; Fine, D. L.; Abbott, B. J .; Mayo, J . G.;
Shoemaker, R. H.; Boyd, M. R. Feasability of drug screeening
with panels of human tumor cell lines using a microculture
tetrazolium assay. Cancer Res. 1988, 48, 589-601.
(51) MTT is 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bro-
mide.
(52) Mosmann, T. Rapid colorimetric assay for cellular growth and
survival: application to proliferation and cytotoxicity assays. J .
Immunol. Methods 1983, 65, 55-63.
(53) Kindly supplied by Dr. P. J . Houghton and Dr. J . A. Houghton,
St. J ude Children’s Research Hospital, Memphis, TN.
J M9502652