Further studies on 2,4-diamino-5-(2′,5′-disubstituted benzyl)pyrimidines as potent and selective inhibitors of dihydrofolate reductases from three major opportunistic pathogens of AIDS

Article abstract of DOI:10.1021/jm020466n

As part of an ongoing effort to discover novel small-molecule antifolates combining the enzyme-binding species selectivity of trimethoprim (TMP) with the potency of piritrexim (PTX), 10 previously unreported 2,4-diamino-5-(2′-methoxy-5′-substituted)benzylpyrimidines (2-11) containing a carboxyl group at the distal end of the 5′-substituent were synthesized and tested as inhibitors of dihydrofolate reductase (DHFR) from Pneumocystis carinii (Pc), Toxoplasma gondii (Tg), and Mycobacterium avium (Ma), three of the opportunistic pathogens frequently responsible for life-threatening illness in people with impaired immune systems as a result of HIV infection or immunosuppressive chemotherapy. The selectivity index of DHFR inhibition was evaluated by comparing the potency of each compound against the parasite enzymes with its potency against rat liver DHFR. 2,4-Diamino-5-[5′-(5-carboxy-1-pentynyl)-2′- methoxybenzyl]pyrimidine (3) inhibited Pc DHFR with a selectivity index of 79 and was 430 times more potent than TMP. 2,4-Diamino-5-[5′-(4-carboxy-1-butynyl)-2′-methoxybenzyl] pyrimidine (2), with one less carbon than 3 in the side chain, had a selectivity index of 910 against Ma DHFR and was 43 times more potent than TMP. 2,4-Diamino-5-[5′-(5-carboxypentyl)-2′-methoxybenzyl]pyri midine (6) had a selectivity index of 490 against Tg DHFR and was 320 times more potent than TMP. 2,4-Diamino-5-[5′-(6-carboxy-1-hexynyl)-2′-methoxybenzyl] pyrimidine (4), with one more carbon than 3, was less potent against all three of the parasite enzymes than either 3 or 6 and also had a lower selectivity index than 3 against the Pc enzyme. However, 4 was the only member of the series with a selectivity index of >300 against both Tg and Ma DHFR. Given that PTX is at least 10 times more potent against rat DHFR than against P. carinii or T. gondii DHFR and that the selectivity index of several of the compounds matches or exceeds that of TMP as well as PTX, our results suggest that it may be possible to develop clinically useful nonclassical antifolates that are both potent and selective against the major opportunistic pathogens of AIDS.

Full text of DOI:10.1021/jm020466n

1726
J . Med. Chem. 2003, 46, 1726-1736
F u r th er Stu d ies on 2,4-Dia m in o-5-(2′,5′-d isu bstitu ted ben zyl)p yr im id in es a s
P oten t a n d Selective In h ibitor s of Dih yd r ofola te Red u cta ses fr om Th r ee Ma jor
Op p or tu n istic P a th ogen s of AIDS
Andre Rosowsky,*,† Ronald A. Forsch,† and Sherry F. Queener‡
Dana-Farber Cancer Institute and Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School,
Boston, Massachusetts 02115, and Department of Pharmacology and Toxicology, Indiana University School of Medicine,
Indianapolis, Indiana 46202
Received October 18, 2002
As part of an ongoing effort to discover novel small-molecule antifolates combining the enzyme-
binding species selectivity of trimethoprim (TMP) with the potency of piritrexim (PTX), 10
previously unreported 2,4-diamino-5-(2′-methoxy-5′-substituted)benzylpyrimidines (2-11) con-
taining a carboxyl group at the distal end of the 5′-substituent were synthesized and tested as
inhibitors of dihydrofolate reductase (DHFR) from Pneumocystis carinii (Pc), Toxoplasma gondii
(Tg), and Mycobacterium avium (Ma), three of the opportunistic pathogens frequently
responsible for life-threatening illness in people with impaired immune systems as a result of
HIV infection or immunosuppressive chemotherapy. The selectivity index of DHFR inhibition
was evaluated by comparing the potency of each compound against the parasite enzymes with
its potency against rat liver DHFR. 2,4-Diamino-5-[5′-(5-carboxy-1-pentynyl)-2′-methoxybenzyl]-
pyrimidine (3) inhibited Pc DHFR with a selectivity index of 79 and was 430 times more potent
than TMP. 2,4-Diamino-5-[5′-(4-carboxy-1-butynyl)-2′-methoxybenzyl]pyrimidine (2), with one
less carbon than 3 in the side chain, had a selectivity index of 910 against Ma DHFR and was
43 times more potent than TMP. 2,4-Diamino-5-[5′-(5-carboxypentyl)-2′-methoxybenzyl]pyri-
midine (6) had a selectivity index of 490 against Tg DHFR and was 320 times more potent
than TMP. 2,4-Diamino-5-[5′-(6-carboxy-1-hexynyl)-2′-methoxybenzyl]pyrimidine (4), with one
more carbon than 3, was less potent against all three of the parasite enzymes than either 3 or
6 and also had a lower selectivity index than 3 against the Pc enzyme. However, 4 was the
only member of the series with a selectivity index of >300 against both Tg and Ma DHFR.
Given that PTX is at least 10 times more potent against rat DHFR than against P. carinii or
T. gondii DHFR and that the selectivity index of several of the compounds matches or exceeds
that of TMP as well as PTX, our results suggest that it may be possible to develop clinically
useful nonclassical antifolates that are both potent and selective against the major opportunistic
pathogens of AIDS.
In tr od u ction
As part of a longstanding interest in the design and
synthesis of small-molecule inhibitors of the dihydro-
folate reductase (DHFR) enzymes from Pneumocystis
carinii (Pc), Toxoplasma gondii (Tg), and Mycobacterium
avium (Ma), three of the pathogenic organisms most
often responsible for opportunistic morbidity and mor-
tality in people with AIDS and other immune disorders,1
we recently reported2,3 that the 2,4-diamino-5-[2′-meth-
oxy-5′-(4-carboxyalkyloxy)benzyl]pyrimidines 1a and 1b
selectively inhibit DHFR from these three organisms
compared with the enzyme from rat liver and that their
potency is much higher than that of trimethoprim
(TMP), the 2,4-diamino-5-(substituted benzyl)pyrimi-
rare among the hundreds of diaminopyrimidine deriva-
dine most widely used for the treatment and prophylaxis
tives studied to date (reviewed in ref 2). Interestingly,
of pneumocystis pneumonia and toxoplasmosis. The
the 2′,5′-substitution pattern of 1a and 1b was remi-
most prominent feature of 1b as a Pc DHFR inhibitor
was that it surpassed TMP in terms of both potency and
selectivity, a combination that has proved to be quite
niscent of piritrexim (PTX),4 which is more potent than
TMP but is strongly selective for mammalian versus Pc
DHFR. Thus, we had discovered in these compounds
two rare examples in which the potency of PTX is
combined with the desirable species selectivity of tri-
methoprim, a feature that was of considerable practical
interest because PTX can only be given to patients in
* To whom correspondence should be addressed. Phone: 617-632-
3117. Fax: 617-632-2410. E-mail: andre_rosowsky@dfci.harvard.edu.
† Harvard Medical School.
‡ Indiana University School of Medicine.
10.1021/jm020466n CCC: $25.00 © 2003 American Chemical Society
Published on Web 03/26/2003

Pyrimidines as Inhibitors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 9 1727
Ch em istr y
As shown in Scheme 1, the well-known Pd-catalyzed
Sonogashira reaction between an aryl halide and a
terminal acetylene10 provided convenient access to the
analogues of 1a and 1b with a 5′-(ω-carboxy-1-alkynyl),
and ultimately a 5′-(ω-carboxyalkyl), side chain. The key
intermediates needed for the coupling reaction were the
terminal acetylenic esters 15-17 and the iodide 19, the
latter of which was obtained from 2,4-diamino-5-(2′-
methoxybenzyl)pyrimidine (18) by reaction with iodine
monochloride as described by Calas and co-workers.11
The esters were readily prepared from the Cs salts of
the commercially available acids 12-14 with benzyl
bromide in dry DMF. Without purification, 15-17 were
condensed with 19 in the presence of Et3N and catalytic
amounts of (Ph3P)2PdCl2 and CuI in dry DMF at 60 °C
under N2. Because of their limited solubility in nonpolar
solvents, the resulting acetylenic esters 20-22 did not
lend themselves easily to column chromatography on
silica gel and therefore were converted directly to the
acids 2-4 by brief treatment with NaOH in DMSO at
room temperature. The yield of 2 and 4 after purification
by preparative HPLC on C18 silica gel using an isocratic
mixture of MeCN and 0.1 M NH4OAc, pH 7.4, as the
eluent was 32% and 36%, respectively. In the case of 3,
we did not use HPLC but instead attempted to recrys-
tallize the compound from MeOH. This proved to be a
bad choice, inasmuch as heating the acetylenic acid in
MeOH led to a sharp drop in final purified yield because
of extensive degradation to a resinous, base-insoluble
product of unknown composition (the crude yield had
been comparable to that of 2 and 4).
F igu r e 1. Structures of 2,4-diamino-5-(substituted benzyl)-
pryrimidines 2-11.
combination with a second drug, leucovorin (5-formyl
5,6,7,8-tetrahydrofolate), in order to prevent unaccept-
able side effects.5 Computer modeling of the predicted
binding of 1b to Pc DHFR,6 together with the ideas
elegantly put forward by Kuyper and co-workers7 as a
rationale for the synthesis of 3′- and 4′-O-(ω-carboxy-
alkyl) analogues of TMP as selective inhibitors of E. coli
versus rat DHFR, led us to hypothesize that selective
binding of 1b to Pc versus rat or human DHFR may be
due to the ability of the terminal COOH group to
interact with strongly basic residues, Arg 75 and Lys
37, in the active site of the Pc enzyme.8,9 In the case of
human DHFR, this interaction would not be possible;
indeed, when 1b was docked into the active site of
human DHFR,6 only a weak interaction with a nonbasic
Gln35 residue seemed possible.
We had originally chosen benzyl esters in Scheme 1
with the intent of using catalytic hydrogenation (H2/
Pd-C) to reductively cleave the benzyl group and
simultaneously reduce the triple bond to form the
saturated acids 5-7. However, when hydrogenation was
attempted with 21 in DMF solution, the triple bond was
reduced but the ester group survived. When reduction
was performed in a 1:1 mixture of DMF and glacial
AcOH with a view to protonating the diaminopyrimidine
moiety, which we suspected might be deactivating the
Pd-C catalyst enough to prevent debenzylation, the
product proved to be totally insoluble in base and thus
could not be a carboxylic acid. Fortunately, conversion
of the acetylenic esters to the desired saturated acids
was achieved satisfactorily by first cleaving the ester
group with NaOH and hydrogenating the Na salt in situ
in aqueous solution or, in the case of 7, by isolating the
free acid and carrying out the reduction in DMF
solution. In contrast to 3 (see above), purification of 5
was achieved satisfactorily by recrystallization from
MeOH. Preparative HPLC (C18 silica gel, 20% MeCN
in 0.1 M NH4OAc, pH 7.4) was used in the case of 6;
however, analytically pure 7 could be obtained by simply
dissolving the product of the hydrogenation reaction in
dilute NaOH and acidifying the solution with AcOH.
The overall two-step yields of the purified acids 5, 6,
and 7 was 35%, 46%, and 77%, respectively. The 1H
NMR spectra of all three compounds were consistent
with reduction of the CtC bond in that there were four
additional CH2 hydrogens and a slight upfield shift for
the aromatic protons ortho to the ω-carboxyalkyl side
chain.
To further explore the scope of this finding, we have
now synthesized a series of 10 analogues of 1a and 1b
in which the nature of the linker between the COOH
group and the 5′-position of the benzyl ring was altered
as depicted in Figure 1. Three of the new analogues
were of particular interest in terms of structure-activity
correlation. One compound (3), in which the oxygen
atom at the 5′-position was replaced by a carbon-carbon
triple bond, was found to be twice as potent as 1b
against both Pc and rat DHFR, demonstrating that (i)
a polar atom at the 5′-position is not necessary and (ii)
a rigid linear bond between the first two atoms of the
side chain is well tolerated. A second analogue contain-
ing this feature, compound 2, was weaker and less
selective than 3 against the Pc enzyme but was none-
theless a very potent and selective inhibitor of Ma
DHFR despite its shorter side chain. The third ana-
logue, compound 6, in which the triple bond of 3 was
reduced to a saturated carbon-carbon bond, was a
weaker inhibitor of Pc and Ma DHFR than either 1b or
3 but proved to be among the most selective inhibitors
of the Tg enzyme reported to date.

1728 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 9
Rosowsky et al.
Sch em e 1a
a
Reactants: (a) PhCH2Br/Cs2CO3/DMF; (b) ICl/AcOH; (c) (Ph3P)2PdCl2, CuI, Et3N, DMF; (d) NaOH, DMSO; (e) H2/10% Pd-C, DMF.
Sch em e 2a
a
Reactants: (a) K2CO3, 18-crown-6; (b) HCl; (c) 3-morpholinopropionitrile, NaOEt; (d) PhNH2OHCl; (e) H2NC(dNH)NH2; (f) NaOH.
Although we had originally planned to prepare com-
pounds 8 and 9 from the Na salt of 2,4-diamino-5-(2′-
methoxy-4′-hydroxybenzyl)pyrimidine, as we had done
earlier to obtain 1a and other 5′-(ω-carboxyalkoxy)
analogues from ω-bromoalkanoic acids,2 efforts to apply
this method using methyl 4-(bromomethyl)benzoate
were thwarted by several side reactions. To circumvent
this problem, we chose to reverse the synthetic sequence
so that formation of the diaminopyrimidine ring would
be the last step. Thus, we carried out the synthesis of
analogues 8 and 9 according to Scheme 2. Condensation
of 3-hydroxy-6-methoxybenzaldehyde (23)12 with methyl
3-(bromomethyl)benzoate (24) in the presence of K2CO3
and a catalytic amount of a crown ether as described in
our earlier paper2 unexpectedly gave a 79% yield of the
previously unknown acetal 25 after recrystallization
from MeOH; however, the latter was readily converted
to the desired aldehyde 26 by treatment with dilute HCl.
The structure of 25 was deduced from its elemental
analysis, the absence of absorbance at 1685 cm-1 in the
infrared spectrum, and the presence, in the 1H NMR
spectrum, of characteristic signals at δ 3.35 and δ 5.64
for the CH(OMe)2 group. That O-alkylation had occurred
in the expected manner was also evident from the
presence of a peak at δ 5.07 corresponding to the OCH2
protons. The overall two-step yield of 26 from 24 was
69%. Interestingly, when 23 was condensed with tert-
butyl 4-(bromomethyl)benzoate (27), which we had used
earlier in another synthesis,13 only the expected alde-
hyde 28 was obtained even though the product was
recrystallized from MeOH. The probable reason for this
difference in behavior between the meta and para esters
is that dilute citric acid was used during the workup of
the reaction of 24 but not 27. Thus, it is possible that
some residual citric acid was present during the recrys-
tallization from MeOH. The remaining steps were
performed according to one of the standard routes for
TMP analogues, in which the aldehyde is treated
sequentially with 3-morpholinopropionitrile, aniline
hydrochloride, guanidine,14 and in this case NaOH to
cleave the ester groups to acids. Although none of the
intermediates in the four-step sequence from 26 and 28
to the final products 8 and 9, respectively, were isolated
in the pure state before proceeding to the next step, the
final acids were purified by preparative HPLC on C18
silica gel using 18% MeCN in 0.l M NH4OAc, pH 7.4,
as the eluent. Large amounts of nonpolar impurities
were easily removed in this manner, and the identity
and purity of the products were confirmed by micro-
1
chemical and H NMR analyses. As expected, 8 and 9
were soluble in dilute aqueous base and precipitated at
a weakly acidic pH.
The final two compounds in this series of analogues,
10 and 11, were synthesized according to Scheme 3,
starting from ethyl salicylate (29) and ethyl 3-hydroxy-
benzoate (30), respectively. Condensation of the phenols
with propargyl bromide in the presence of K2CO3 and a
crown ether catalyst in DMF afforded the O-propargyl
ethers 31 and 32. Further reaction of the ethers with
iodide 19 in the presence of (Ph3P)2PdCl2, CuI, and
Et3N, followed by saponification, afforded the desired
acids in overall two-step yields of 21% and 24%. Neither
31 and 32 nor the ester intermediates 33 and 34 from
the Sonogashira reaction were chromatographed, but
the final products were carefully purified by preparative
HPLC (C18 silica gel, 20% MeCN in 0.1 M NH4OAc, pH
7.4) and characterized by their 1H NMR spectra in
DMSO-d6 solution, which displayed the expected reso-
nance signals at δ 3.50 (benzylic CH2), 3.82 (OMe), 5.04
(CH2O), and 7.40 (pyrimidine 6-H) in addition to appro-
priate peaks for the various protons of the phenyl rings.

Pyrimidines as Inhibitors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 9 1729
Sch em e 3a
a
Reactants: (a) HCtCCH2Br, K2CO3, 18-crown-6; (b) (Ph3P)2PdCl2, Cul, Et3N; (c) NaOH.
Ta ble 1. Inhibition of P. carinii, T. gondii, M. avium, and Rat Dihydrofolate Reductase by 2,4-Diamino-5-(2′-methoxy-5′-substituted
benzyl)pyrimidines
a
IC50 (µM)
selectivity indexb
T. gondii
compd
P. carinii
0.25
T. gondii
0.18
(0.15-0.20)
0.11
M. avium
rat
P. carinii
M. avium
1a
0.0048
2.7
11 (8.5-13)
16 (12-19)
560 (430-730)
(0.21-0.30)
0.0040-0.0059) (2.5-2.9)
1b
0.054c
0.058c,d
4.6e
(3.2-6.0)
4.0
85 (79-97)
31 (24-39)
79 (63-100)
28 (19-42)
21 (14-31)
27 (18-41)
8.6 (6.4-11)
21 (13-33)
17 (9.6-30)
2.7 (2.1-3.5)
2.4 (1.6-3.7)
14 (10-20)
42 (19-72)
79 (19-230)
(0.045-0.068) (0.066-0.17)
0.13
(0.12-0.15)
0.028
(0.025-0.03)
0.87
(0.77-0.98)
0.15
(0.13-0.18)
0.15
(0.13-0.17)
0.53
(0.43-0.65)
0.89
(0.73-1.1)
1.2
(0.02-0.16)
0.0044
(0.0040-0.0049) (3.5-4.6)
0.0078 2.2
(0.0064-0.0095) (1.9-2.5)
0.041
(0.31-0.56)
0.035
(0.025-0.048)
0.016
(0.011-0.022)
0.089
(0.071-0.11)
0.12
(0.1-0.14)
0.060
(0.053-0.069)
0.036
(0.032-0.041)
0.057
0.050-0.065)
0.30
(0.26-0.35)
0.00061
2
0.097
41 (34-39)
910 (700-1100)
280 (2000-380)
590 (330-1100)
91 (51-160)
(0.093-0.10)
3
0.032
69 (52-91)
(0.027-0.037)
0.072
4
25
340 (230-510)
54 (39-77)
(0.064-0.081)
0.058
(0.052-0.065)
0.0084
0.075-0.009)
0.030
(0.026-0.034)
0.60
(0.53-0.69)
2.0
(1.7-2.2)
0.50
(0.44-0.56)
1.6
(1.1-2.4)
2.8
(19-32)
5
3.2
(2.5-4.0)
4.1
6
490 (340-700)
150 (120-190)
31 (21-46)
260 (140-460)
52 (38-70)
(3.1-5.3)
4.6
(4.2-5.0)
19
(14-24)
21
(13-29)
6.2
(5.2-7.4)
14
(11-16)
180
(160-210)
0.0033
7
8
150 (100-230)
340 (210-560)
170 (130-230)
240 (170-250)
610 (460-810)
9
11 (6.4-18)
13 (9.3-17)
8.3 (4.6-15)
65 (48-87)
(0.098-1.5)
2.3
10
(2.1-2.5)
11
5.6e
(4.4-7.2)
13
(10-16)
0.013
TMPf
PTXg
(2.4-3.3)
0.0043
0.26 (0.17-0.42) 0.76 (0.63-0.97) 5.4 (4.1-7.2)
(0.009-0.017) (0.0040-0.0046) (0.0053-0.0007) (0.0029-0.0039)
a
Numbers in parentheses are 95% confidence limits. The difference in IC50 between rat liver DHFR and each of the parasite enzymes
b
was determined to be statistically significant at P < 0.01. SI ) IC50(rat liver)/IC50(Pc, Tg, or Ma). Numbers in parentheses are 95%
confidence intervals, and represent a range calculated by dividing the lower end of the 95% confidence interval for rat liver DHFR by the
high end of the 95% confidence interval for Pc, Tg, or Ma DHFR. c Mean of three experiments on different days. The IC50 of 1b against
d
Ma DHFR was inadvertently recorded as 0.0058 µM in ref 2. As a result, the SI calculated therein for this compound was off by a factor
of 10. e Mean of two experiments on different days. f TMP ) trimethoprim, 2,4-diamino-5-(3′,4′,5′-trimethoxybenzyl)pyrimidine. Data are
g
taken from ref 2. PTX ) piritrexim, 2,4-diamino-5-methyl-6-(2′,5′-dimethoxybenzyl)pyrido[2,3-d]pyrimidine. Historical IC50 values for
PTX, obtained with an older sample and cited previously for comparison purposes (e.g., see ref 2), were the following: Pc DHFR, 0.031
µM; Tg DHFR, 0.017 µM; rat liver, 0.0015 µM. The data in Table 1 were obtained with a newer sample of PTX kindly provided by Dr.
Gary Smith, Glaxo-Wellcome Co., Research Triangle Park, NC. The IC50 of PTX against Ma DHFR had not been determined prior to this
work.
En zym e In h ibition Assa ys
organism, 3 was slightly more potent than 1b. However,
because potency increased by about the same extent
against rat DHFR, the selectivity index (SI) as defined
in Table 1, footnote b, was not changed significantly.
Although there was no substantial improvement in
potency or selectivity, this finding was nonetheless of
interest because it demonstrated that the oxygen atom
at the 5′-position of 1b can be safely replaced by a
carbon-carbon triple bond where Pc DHFR binding is
concerned. It may be noted that the CtC(CH2)3CO2H
side chain in 3 contains the same number of atoms as
the O(CH2)4CO2H side chain in 1b but is less flexible
Compounds 2-11 were tested for the ability to inhibit
Pc, Tg, and Ma DHFR according to the standardized
spectrophotometric assays used in the Queener labora-
tory,15,16 and their potencies and selectivities were
compared with updated values for 1a , 1b, TMP, and
PTX. The results are presented in Table 1 and discussed
for each of enzyme in the following sections.
P n eu m ocystis ca r in ii DHF R. As shown in Table 1,
the most potent inhibitor of Pc DHFR in this group was
the 5′-(5-carboxy-1-pentynyl) analogue 3, with an IC50
of 0.028 µM. Thus, against the enzyme from this

1730 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 9
Rosowsky et al.
F igu r e 2. Structures of T. gondii DHFR inhibitors previously reported to be more potent as well as more selective than
trimethoprim.2 Data for compound 5 are taken from Table 1. The calculated SI value given for 40 is based on titrations performed
with human rather than rat DHFR and on assay conditions different from those used for the other analogues. When 40 was
tested against Tg and rat liver DHFR according to the standardized method used by the Queener laboratory,15 its selectivity was
much lower.
because of the geometric constraints imposed by the
triple bond. From the enzyme binding data, it appears
that the terminal carboxyl group may nonetheless be
adequately positioned in the active site to interact with
Arg75. Of interest in this regard is that the shorter
analogue 2 was a slightly weaker inhibitor of both Pc
and rat DHFR than 3, a pattern also observed in the
previous series of 5′-(ω-carboxyalkoxy) analogues.2 How-
ever, compound 4 displayed a more striking decrease
in potency than either 2 or 3 against both the Pc and
rat enzymes. Finally, with regard to the potency and
selectivity of the alkynes 5-7 versus the alkanes 2-4,
a dramatic pattern of structure-activity relationships
did not emerge, and we were not able to determine
whether a carboxyalkyne or carboxyalkane side chain
was superior.With respect to compounds 8-11 as in-
hibitors of Pc versus rat DHFR, it can be seen from
Table 1 that they are all considerably weaker than 1b
or 3 against the Pc enzyme and that the potency
difference is greater against this enzyme than against
the rat enzyme, resulting in an overall decrease in
selectivity. At least with these examples, therefore, it
appears that the introduction of a bulky second phenyl
ring in the side chain, together with replacement of the
aliphatic COOH group by an aromatic one, was unfa-
vorable. It may be noted that when the total number of
side chain atoms is added, the number of atoms sepa-
rating the carboxyl group from the 2′-methoxyphenyl
ring is five in 8, six in 9 and 10, and seven in 11. The
corresponding number of atoms is four in 2 and 5, five
in 3 and 6, and six in 4 and 7. Thus, among the
analogues with an extra phenyl ring, 8-10 most closely
approximate 1b and 2-7 in terms of the total number
of atoms (aromatic and aliphatic) separating the car-
boxyl group from the rest of the molecule, whereas in
11 there is one additional atom.
Toxopla sm a gon d ii DHF R. In contrast to Pc DHFR,
the most potent inhibitor of the Tg enzyme was 6, with
an IC50 of 0.0084 µM (Table 1). Whereas this cor-
responded to a roughly 1-log increase in binding relative
to 1b, a comparable effect was not observed with rat
DHFR. As a result, the selectivity index of 6 was 490
compared with only 42 in the case of 1b. Thus, replace-
ment of the oxygen atom at the 5′-position of 1b by a
saturated carbon resulted in a considerable improve-
ment in both potency and selectivity. Interestingly, this
was in contrast to the decreased potency and selectivity
of 6 relative to 1b against Pc DHFR and of 6 relative to
1a against Ma DHFR.
The published literature on Tg DHFR inhibitors
contains a large number of mono-, di-, and tricyclic 2,4-
diaminopyrimidines that are much more potent than
TMP but only a handful that are both more potent and
more selective against Tg DHFR versus rat DHFR under
our standardized assay conditions. As summarized in
Figure 2, these compounds include 35,2 a two-carbon
homologue of 1a , the TMP analogue 36 (epiroprim),17
the 2,4-diamino-5-aryl-6-ethylpyrimidines 37 and 38,18
the 2,4-diamino-6-substituted pteridines 3919 and 40,3g
the 2,4-diamino-6-benzylaminopyrido[2,3-d]pyrimidine
41 (albeit only when tested against Tg versus human
instead of rat DHF and under assay conditions different
from those typically used with the other compounds
reported),17 and the 2,4-diamino-5-substituted pyrrolo-
[2,3-d]pyrimidines 42 and 43.20 Very surprisingly, three
2,4-diamino-6,7-disubstituted pteridines (44-46) have
also been reported to be among this select group.16
Bulky substitution at the 7-position of 2,4-diamino-

Pyrimidines as Inhibitors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 9 1731
pteridines is generally considered very unfavorable for
DHFR binding. However, it is possible that in the case
of the Tg enzyme the selectivity of these compounds
relative to rat DHFR reflects the fact that the Tg
enzyme is a larger difunctional protein containing both
a thymidylate synthase (TS) and DHFR domain.21 This
question will presumably remain unanswered until the
3D structure of the Tg enzyme is fully solved. On the
basis of its 330-fold increase in potency and 7.5-fold
increase in selectivity relative to TMP against Tg DHFR
and in light of its very good profile when compared to
the inhibitors shown in Figure 2, compound 6 may be
viewed as a most interesting lead for further structure-
activity optimization.
Mycoba cter iu m a viu m DHF R. As indicated in
Table 1, compound 2 was a much better inhibitor of Ma
DHFR than TMP and was similar to the previously
reported 5′-O-(5-carboxybutyloxy) analogue 1a . Indeed,
with its nearly 3-log selectivity for Ma DHFR relative
to the rat enzyme and its approximately 70-fold supe-
riority over TMP in terms of potency, 2 ranks among
the best inhibitors of Ma DHFR described to date.
Although it stands out among the other compounds
tested, there were also several analogues with respect-
able SI values in the 200-600 range. However, it should
be noted that high SI values were not always ac-
companied by high potency. Thus, while the calculated
SI of 590 listed for compound 4 had a range (330-1100)
that overlapped that of 2 (700-1100), its potency
against both Ma and rat DHFR was considerably lower.
Suling and co-workers22 recently published data on a
library of almost 80 2,4-diaminopyrido[2,3-d]pyrim-
idines with either arylamino or arylthiomethyl groups
at the 6-position as inhibitors of Ma versus human
DHFR. The most potent member of their series against
Ma DHFR was 2,4-diamino-6-(2′-methyl-4′-chloro-
anilino)methyl-5-methylpyrido[2,3-d]pyrimidine (47),
with an IC50 of 0.000 19 µM. However, because 47 was
also a potent inhibitor of a mammalian DHFR (human
in this case), its SI value was only 15. A second
compound, 2,4-diamino-6-(2′,5′-diethoxyanilino)methyl-
5-methylpyrido[2,3-d]pyrimidine (48), on the other hand,
had IC50 values of 0.000 84 and 2.3 µM against Ma and
human DHFR, respectively. Unfortunately an IC50 value
Ma DHFR do not require the inhibitor to be a fused-
ring 2,4-diaminopyrimidine.
The 3D structure of the ternary complex of Ma DHFR
with NADPH and TMP has not been reported. However,
it is noteworthy that in a recent study of the closely
related enzyme from M. tuberculosis by X-ray crystal-
lography Li and co-workers24 suggested that TMP
analogues with extended hydrophobic substitution on
the benzyl ring might be designed to selectively fit into
a binding pocket that does not appear to be available
in the active site of mammalian DHFR enzymes.
Although they are 2′,5- rather than 3′,4′,5-trisubstituted
benzyl derivatives, it would be of interest to determine
whether 2-4 can bind to the Ma enzyme in such a way
as to allow the hydrocarbon portion of the 5′-(ω-carboxy-
1-alkynyl) side chain to make van der Waals contact
with this hydrophobic domain.
Su m m a r y
Four new examples of small-molecule antifolates of
the 2,4-diamino-5-(substituted benzyl)pyrimidine fam-
ily, each more potent and/or more selective than TMP
against at least one of three different opportunistic
pathogens of AIDS, were discovered in this work. One
of them, 2,4-diamino-5-[5′-(4-carboxy-1-pentynyl)-2′-
methoxybenzyl]pyrimidine (3) inhibited Pc DHFR with
a selectivity index of 79 relative to the rat liver enzyme
and was 430 times more potent than TMP. A second
compound, 5-[5′-(5-carboxy-1-butynyl)-2′-methoxyben-
zyl]pyrimidine (2), with one less carbon than 3 in the
side chain, had a selectivity index of 910 against Ma
DHFR and was 43 times more potent than TMP. The
third compound, 2,4-diamino-5-[5′-(5-carboxypentyl)-2′-
methoxybenzyl]pyrimidine (6) had a selectivity index of
490 against Tg DHFR and was 320 times more potent
than TMP. The fourth compound, 2,4-diamino-5-[5′-(6-
carboxy-1-hexynyl)-2′-methoxybenzyl]pyrimidine (4), was
less potent against all three of the parasite enzymes
than either 3 or 6 and also had a lower selectivity index
than 3 against the Pc enzyme. However, it was the only
member of the series with a selectivity index of >300
against both Tg and Ma DHFR. Attesting to the likeli-
hood that subtle differences exist in the 3D structure
of the active site of these three enzymes, each inhibitor
was optimally active against only one of the enzymes
despite the fact that there were only minor differences
in molecular structure among the three compounds.
While the compounds described in this and our
preceding paper2 on 2,4-diamino-5-[2′-methoxy-5-(ω-
carboxyalkoxy)benzyl]pyrimidines (e.g., 1a and 1b )
show promise in terms of in vitro potency and selectiv-
ity, it has not been shown that they are efficiently taken
up by P. carinii, T. gondii, or M. avium. Thus, before
any practical clinical utility for such compounds can be
established, it must be shown that they are able to
penetrate these organisms in at least one stage of the
life cycle or, if they cannot, that a suitable prodrug or
other mode of cellular drug delivery can be developed.
Although they were outside the scope of the work
reported here, future studies of this nature would be of
potential therapeutic interest.
for 48 against rat DHFR was not reported, and the
assay conditions differed in some respects from those
used routinely by the Queener laboratory.16 The binding
of inhibitors to DHFR can vary substantially depending
on the pH, the temperature, and other variables.23 Thus,
it remains to be determined how 2 and 48 would
compare if they were tested under identical assay
conditions in the same laboratory. Irrespective of the
outcome of such a comparison, the excellent combination
of potency reported here over a range of 2,4-diamino-
5-(substituted benzyl)pyrimidines, including some that
are not particularly potent or selective against Pc or Tg
DHFR, demonstrates that potency and selectivity against
Exp er im en ta l Section
IR spectra were obtained on a Perkin-Elmer model 781
double-beam recording spectrophotometer. Only peaks with
wavenumbers above 1200 cm-1 are reported. 1H NMR spectra

1732 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 9
Rosowsky et al.
were recorded in DMSO-d6 solution at 200 MHz on a Varian
VX200 instrument. Each peak is denoted as a singlet (s), broad
singlet (br s), doublet (d), doublet of doublets (dd), triplet (t),
doublet of triplets (dt), or pentet (p). Integrated peak areas
are not listed when the resonance signal was partly obscured
by water or DMSO or in the case of NH2 groups on the
pyrimidine ring. Signals for the aromatic protons in compounds
with two phenyl rings are identified according to the number-
ing in Schemes 2 and 3. TLC analyses were on Whatman
MK6F silica gel plates with UV illumination at 254 nm.
Column chromatography was on Baker 7024 flash silica gel
(40 µm particle size). HPLC separations were performed on
C18 radial compression cartridges (Millipore, Milford, MA;
analytical, 5 µm particle size, 5 mm × 100 mm; preparative,
15 µm particle size, 25 mm × 100 mm). Melting points were
measured in Pyrex capillary tubes in a Mel-Temp apparatus
(Fisher, Pittsburgh, PA) and are not corrected. 3-Morpholino-
propionitrile was prepared by adding acrylonitrile dropwise
to an equimolar amount of morpholine in an ice bath and
stirring the mixture at room temperature for 1 h. The resulting
light-yellow oil did not require purification and was used
directly. 2,4-Diamino-5-(2′-methoxybenzyl)pyrimidine (18) and
2,4-diamino-5-(5′-iodo-2′-methoxybenzyl)pyrimidine (19) were
prepared according to the literature: 18, 42% yield, mp 160-
161 °C (lit.11 160 °C); 19, 66% yield, mp 207-208 °C (lit.11 205
°C). Other chemicals were purchased from Sigma-Aldrich (St.
Louis, MO), Acros Organics (Pittsburgh, PA), and Lancaster
Synthesis (Windham, NH). Elemental analyses were per-
formed by Quantitative Technologies, Inc. (Whitehouse, NJ )
and were within (0.4% of theoretical values. Where microana-
lytical results were consistent with residual acetic acid, its
presence in the sample was confirmed by the finding of a
methyl signal at δ 1.9 in the 1H NMR spectrum.
2,4-Dia m in o-5-[5′-(4-ca r b oxy-1-b u t yn yl)-2′-m et h oxy)-
ben zyl]p yr im id in e (2). Step 1. Benzyl 4-pentynoate (15) was
prepared by stirring a solution of 4-pentynoic acid (12) (1.96
g, 0.02 mol) in dry DMF (25 mL) with Cs2CO3 (3.25 g, 0.01
mol) for 10 min, followed by addition of benzyl bromide (2.38
mL, 3.42 g, 0.02 mol). After 18 h of stirring at room temper-
ature, the solvent was removed by rotary evaporation (vacuum
pump) and the residue was partitioned between EtOAc and
H2O. Evaporation of the organic layer gave 15 as an oil whose
NMR spectrum showed that it still contained some DMF and
a trace of benzyl bromide but was suitable for use in the next
step.
Step 2. A stirred mixture of 15 (1.3 g, estimated by NMR
to contain ca. 6.0 mmol), iodide 19 (1.42 g, 4.0 mmol), (Ph3P)2-
PdCl2 (15 mg), CuI (1 mg), and Et3N (5 mL) in dry DMF (5
mL) was heated at 60 °C for 20 h. A clear solution formed after
ca. 1 h. The solvent was removed by rotary evaporation, and
the residue was triturated successively with isooctane and
H2O. Recrystallization from EtOH afforded 20 as a light-yellow
powder (1.59 g, mp 208 °C), which was used in the next step
without additional purification.
0.01 mol) in dry DMF (25 mL) with Cs2CO3 (1.63 g, 0.005 mol)
for 10 min, followed by addition of benzyl bromide (1.19 mL,
1.71 g, 0.01 mol). After 18 h of stirring at room temperature,
the solvent was removed by rotary evaporation (vacuum pump)
and the residue was partitioned between EtOAc and H2O.
Evaporation of the organic layer gave 16 as an oil suitable for
use directly in the next step.
Step 2. A stirred mixture of 16 (0.65 g, estimated by NMR
to contain ca. 36.0 mmol), iodide 19 (1.42 g, 4.0 mmol), (Ph3P)2-
PdCl2 (10 mg), CuI (1 mg), and Et3N (3 mL) in dry DMF (3
mL) was heated at 60 °C for 72 h. The solvent was removed
by rotary evaporation, and the residue was triturated succes-
sively with isooctane and H2O. Attempted recrystallization
from EtOH gave only a gum. Therefore, the solvent was
evaporated, and the residue was redissolved in hot THF (40
mL). A small amount of insoluble material was filtered off,
the filtrate was concentrated to dryness, and the residue,
containing ester 21, was taken up directly in dry DMSO (6
mL). The solution was treated dropwise with 1 N NaOH (6
mL) and then diluted with H2O (100 mL) and chilled. A small
amount of residual nonsaponifiable material was removed by
filtration, and the filtrate was acidified with 10% AcOH. The
precipitate was collected and dried in a lyophilization ap-
paratus to obtain a beige solid. Attempted recrystallization
from hot MeOH led to separation of a brown insoluble gum,
suggesting that decomposition was occurring. The methanolic
solution was immediately decanted and left to cool passively
in the hood at room temperature, producing the first crop of
analytically pure 3 as white crystals weighing 32 mg (5%): mp
121-124 °C (softening without giving a true melt); IR (KBr) ν
3500-2800 (broad), 3350, 3210, 2930, 1665, 1560-1530, 1505,
1460, 1405, 1295, 1115 cm-1; 1H NMR (DMSO-d6) δ 1.72 (p, J
) 7 Hz, 2H, CtCCH2CH2CH2), 2.37 (two overlapping t, 4H,
CtCCH2CH2CH2), 3.50 (s, 2H, bridge CH2), 3.80 (s, 3H, OMe
partly overlapped by H2O), 5.86 (br s, NH2), 6.20 (br s, NH2),
6.95 (m, 2H, aryl 3′- and 6′-H), 7.24 (dd, J ) 8 Hz, J ) 2 Hz,
1H, aryl 4′-H), 7.39 (s, 1H, pyrimidine 6-H). Another 153 mg
(24%) of less pure 3 was obtained from the mother liquor and
used for hydrogenation as described below. Anal. (C18H20N4O3‚
0.5AcOH‚0.2 H2O) C, H, N.
2,4-Dia m in o-5-[5′-(6-ca r b oxy-1-h exyn yl)-2′-m et h oxy)-
ben zyl]p yr im id in e (4). Step 1. Benzyl 6-heptynoate (17) was
prepared by stirring a solution of 6-heptynoic acid (14) (1.26
g, 0.01 mol) in dry DMF (15 mL) with Cs2CO3 (1.63 g, 0.005
mol) for 10 min, followed by addition of benzyl bromide (1.19
mL, 1.71 g, 0.01 mol). After 18 h of stirring at room temper-
ature, the solvent was removed by rotary evaporation (vacuum
pump) and the residue was partitioned between EtOAc and
H2O. Evaporation of the organic layer gave 17 as an oil suitable
for use directly in the next step.
Step 2. A stirred mixture of 17 (0.65 g, estimated by NMR
to contain ca. 3.0 mmol), iodide 19 (0.71 g, 2.0 mmol), (Ph3P)2-
PdCl2 (20 mg), CuI (2 mg), and Et3N (5 mL) in dry DMF (5
mL) was heated at 65 °C for 2.5 h. Additional amounts of
(Ph3P)2PdCl2 (10 mg) and CuI (1 mg) were added, and heating
was continued for 18 h. The solvent was evaporated under
reduced pressure, and the residue was triturated successively
with isooctane and H2O. The solid was taken up in DMSO (5
mL) and treated with a solution of NaOH (220 mg, 4.4 mmol)
in H2O (1 mL). The solution was diluted to 80 mL with H2O,
then chilled, adjusted to ca. pH 8 with 10% AcOH, and filtered
to remove a trace of insoluble material. The filtrate was
subjected to preparative HPLC (C18 silica gel, 20% MeCN in
0.1 M NH4OAc, pH 7.4, 10 mL/min) and fractions eluting in
the major peak (ca. 12 min on the analytical column at 1.0
mL/min) were pooled and freeze-dried to obtain 4 as a white
solid: yield 292 mg (36%); mp 187-189 °C (softening without
giving a true melt); IR (KBr) ν 3370, 3220, 2950, 2880, 2850br,
1690sh, 1670, 1565, 1540, 1505, 1465, 1450sh, 1430, 1405,
Step 3. A solution of 20 (416 mg, 1.0 mmol) in dry DMSO
(4 mL) was treated dropwise with 1 N NaOH with swirling.
The mixture was diluted immediately with H2O (40 mL),
acidified with 10% AcOH, and chilled in ice. The precipitate
was collected and dried to obtain a beige solid: crude yield
239 mg. Analytically pure 2 for bioassay was obtained by
preparative HPLC (C18 silica gel, 13% MeCN in 0.1 M NH4-
OAc, pH 7.4). Appropriately pooled fractions were concentrated
and freeze-dried: yield 123 mg (32% based on the amount of
19 used); mp 162-163 °C (softening without giving a true
melt); IR (KBr) ν 3500-2800 (broad), 3350, 3210, 2930, 1665,
1560-1530, 1505, 1460, 1405, 1295, 1115 cm-1 1H NMR
;
(DMSO-d6) δ 2.50 (m, 4H, CH2CH2, partly overlapped by
DMSO peak), 3.49 (s, 2H, bridge CH2), 3.87 (s, 3H, OMe), 5.84
(br s, NH2), 6.16 (br s, NH2), 7.00 (m, 2H, aryl 3′- and 6′-H),
7.22 (d, J ) 9 Hz, 1H, aryl 4′-H), 7.39 (s, 1H, pyrimidine 6-H).
Anal. (C19H22N4O5‚0.2AcOH‚2.5H2O) C, H, N.
1
1375s, 1340, 1320, 1295, 1350, 1300 cm-1; H NMR (DMSO-
d6) δ 1.63 (m, 4H, CH2CH2CH2CO2H), 2.18 (t, J ) 7 Hz, 2H,
CH2CO2H), 2.39 (t, J ) 7 Hz, CtCCH2), 3.80 (s, OMe, partly
overlapped by broad H2O peak), 5.74 (br s, NH2), 6.09 (br s,
NH2), 6.93 (d, J ) 8 Hz, 1H, 3′-H), 7.02 (d, J ) 2 Hz, 1H, 6′-
2,4-Dia m in o-5-[5′-(4-ca r boxy-1-p en tyn yl)-2′-m eth oxy)-
ben zyl]p yr im id in e (3). Step 1. Benzyl 5-hexynoate (16) was
prepared by stirring a solution of 5-hexynoic acid (13) (1.12 g,

Pyrimidines as Inhibitors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 9 1733
H), 7.22 (dd, J ) 8 Hz, J ) 2 Hz, 1H, 4′-H), 7.41 (s, 1H,
pyrimidine 6-H). The signal for the CH2 bridge was completely
obscured by a strong H2O peak. Anal. (C19H22N4O3‚2.75H2O)
C, H, N.
3.74 (s, OMe, partly overlapped by broad H2O peak), 5.65 (br
s, NH2), 6.03 (br s, NH2), 6.88 (m, 2H, 3′- and 6′-H), 6.98 (d, J
) 8 Hz, 1H, 4′-H), 7.32 (s, 1H, pyrimidine 6-H). The signal for
the CH2 bridge was obscured by a strong H2O peak. Anal.
(C19H26N4O3‚AcOH‚1.3H2O) C, H, N.
2,4-Dia m in o-5-[5′-(4-ca r boxybu tyl)-2′-m eth oxy)ben zyl]-
p yr im id in e (5). A solution of recrystallized 20 (416 mg, 1.0
mmol) in dry DMF (20 mL) containing 10%% Pd-C (100 mg)
was shaken under H2 (50 psi initial pressure) for 18 h. After
filtration to remove the catalyst, the solution was evaporated
to a solid that was not soluble in NaOH, indicating that the
ester group had survived. The solid, assumed to be the benzyl
ester of 5, was dissolved directly in hot EtOH (20 mL) and
treated with 1 N NaOH (2 mL). The solvent was evaporated
under reduced pressure, and the residue was taken up in H2O.
A trace of cloudiness, indicating that a trace of ester was still
present, was discharged by adding another small portion of 1
N NaOH and some EtOH. The volume was reduced by rotary
evaporation, and 10% AcOH was added dropwise until a solid
formed, which was filtered and dried on a lyophilizer to obtain
5 as a white powder (117 mg, 35%): mp 134-137 °C (softening
without giving a true melt); IR (KBr) ν 3330, 3180, 2930, 2850,
Met h yl 3-(3-F or m yl-4-m et h oxyp h en oxym et h yl)b en -
zoa te (26). Step 1. A stirred mixture of 3-hydroxy-6-meth-
oxybenzaldehyde (23) (456 mg, 3.0 mmol),2 methyl 3-bromo-
methylbenzoate (24) (687 mg, 3.0 mmol), K2CO3 (1.04 g, 7.5
mmol), and 18-crown-6 (79 mg, 0.3 mmol) in dry DMF (10 mL)
was heated at 70 °C for 20 h. The mixture was cooled to room
temperature, the salts were filtered off, and the solvent was
removed by rotary evaporation using a vacuum pump. The
residue was partitioned between EtOAc and dilute aqueous
citric acid. Evaporation of the organic layer gave a tan solid
(1.03 g), which on recrystallization from MeOH proved unex-
pectedly to be the dimethyl acetal 25: yield 816 mg (79%);
mp 65-66 °C; IR (KBr) ν 3010, 2970, 2920, 2850, 1730, 1690w,
1615, 1595, 1505, 1470, 1455, 1440, 1405, 1380, 1370, 1290,
1235, 1210 cm-1 1H NMR (CDCl3) δ 3.35 (s, 6H, two OMe
;
groups of acetal), 3.81 (s, 3H, ether OMe), 3.93 (s, 3H, ester
OMe), 5.07 (s, 2H, benzylic CH2), 5.64 (s, 1H, acetal CH), 6.82
(d, J ) 9H, 1H, aryl 5-H), 6.91 (dd, J ) 9 Hz, J ) 3 Hz, 1H,
aryl 6-H), 7.21 (d, J ) 3 Hz, 1H, aryl 4-H), 7.45 (t, J ) 8 Hz,
1H, aryl 5′-H), 7.64 (d, J ) 7 Hz, 1H, aryl 6′-H), 7.99 (d, J )
7 Hz, 1H, aryl 4′-H), 8.11 (s, 1H, aryl 2′-H). Anal. (C19H22O6)
C, H.
1660, 1560, 1505, 1455, 1400, 1290, 1250 cm-1 1H NMR
;
(DMSO-d6) δ 1.46 (m, 4H, CH2CH2CH2CO2H), 1.77 (m, 4H,
CH2CH2CH2CH2CO2H), 3.7 (s, 2H, bridge CH2), 3.75 (s, OMe,
partly overlapped by broad H2O peak), 5.66 (br s, NH2), 6.03
(br s, NH2), 6.89 (m, 3H, 3′-, 4′-, and 6′-H), 7.34 (s, 1H,
pyrimidine 6-H). Anal. (C17H22N4O3‚0.5AcOH) C, H, N.
Step 2. The acetal 25 from the preceding step was dissolved
in THF (10 mL), and the solution was cooled in an ice bath
and stirred while cold 1 N HCl (10 mL) was added dropwise,
After 25 min at 0 °C, the mixture was diluted with isooctane.
Partial precipitation occurred, but when TLC showed that both
the solid and the solution contained an identical single spot
(Rf ) 0.5, silica gel, 1:1 EtOAc-isooctane), they were recom-
bined in EtOAc and the solution was washed with 5%
NaHCO3. Evaporation of the organic layer afforded a solid (0.7
g crude yield), which on recrystallization from MeOH with
three drops of added Et3N afforded aldehyde 26 as off-white
flakes (617 mg, 69%): mp 102-103 °C; IR (KBr) ν 2960w,
2880w, 1730, 1685, 1615, 1640w, 1500, 1475, 1455, 1430, 1405,
2,4-Diam in o-5-[5′-(5-car boxypen tyl)-2′-m eth oxy)ben zyl]-
p yr im id in e (6). Benzyl 5-hexynoate (16) (0.65 g of nonpurified
ester, ca. 3.0 mmol, prepared from 11 as described above) was
heated with iodide 19 (0.71 g, 2.0 mmol), (Ph3P)2PdCl2 (10 mg),
CuI (1 mg), and Et3N (3 mL) in dry DMF (3 mL) under N2 at
60 °C for 3.5 h. Solution occurred after ca. 1 h. The solvent
was evaporated under reduced pressure, the residue was taken
up in warm 95% EtOH (60 mL), and the solution, containing
ester 21, was treated with 1 M NaOH (6 mL). The EtOH was
evaporated and replaced with H2O (50 mL). The mixture was
chilled and filtered, and the filtrate was transferred directly
to a Parr apparatus and subjected to catalytic hydrogenation
(42 psi initial pressure) in the presence of 10% Pd-C (85 mg).
The catalyst was filtered off, and the filtrate was acidified with
10% AcOH. The precipitated solid was collected and dried on
a lyophilizer: yield ca. 0.4 g. Analytical HPLC (C18 silica gel,
20% MeCN in 0.1 M NH4OAc, pH 7.4, 1 mL/min) showed a
major peak eluting at 12 min, along with several unidentified
impurities. Preparative HPLC using the same eluent system,
pooling of appropriate fractions, and freeze-drying afforded 6
as a white powder (334 mg, 46%): mp 96-98 °C (softening
without giving a true melt); IR (KBr) ν 3350, 3200, 2930, 2860,
1385, 1320, 1295, 1280, 1265, 1225, 1210 cm-1 1H NMR
;
(CDCl3) δ 3.91 (s, 3H, ether OMe), 3.93 (s, 3H, ester OMe),
5.09 (s, 2H, CH2O), 6.96 (d, J ) 9 Hz, 1H, aryl 5-H), 7.21 (dd,
J ) 9 Hz, J ) 3 Hz, 1H, aryl 6-H), 7.43 (d, J ) 3 Hz, 1H, aryl
4-H), 7.48 (d, J ) 7 Hz, 1H, aryl 6′-H), 7.63 (d, J ) 8 Hz, 1H,
aryl 5′-H), 8.01 (d, J ) 8 Hz, 1H, aryl 4′-H), 8.11 (s, 1H, aryl
2′-H), 10.45 (s, 1H, CHdO). Anal. (C17H16O5‚0.1H2O) C, H.
2,4-Dia m in o-5-[2′-m et h oxy-5′-(3′′-ca r b oxyb en zyloxy)-
ben zylp yr im id in e (8). Step 1. Metallic Na (23 mg, 1.0 mmol)
was dissolved in absolute EtOH (30 mL). The solvent was
removed by rotary evaporation, and the residue was redis-
solved in dry DMSO (2 mL). 3-Morpholinopropionitrile (280
mg, 2.0 mmol) was added, and the reaction mixture was placed
in an oil bath preheated to 100 °C. A solution of 26 (600 mg,
2.0 mmol) in DMSO (3 mL, with slight warming as needed)
was added all at once, and heating was continued for 20 min.
A second portion of NaOMe (1.0 mmol in DMSO) was added,
and heating was resumed for another 20 min. The reaction
mixture was cooled and partitioned between EtOAc and dilute
aqueous citric acid. The EtOAc layer was evaporated, the
residue was dissolved in absolute EtOH (20 mL), aniline
hydrochloride (389 mg, 3.0 mmol) was added, and the mixture
was heated under reflux for 30 min and set aside until the
next step.
Step 2. Metallic Na (184 mg, 8.0 mmol) was dissolved in
absolute EtOH (25 mL), and guanidine hydrochloride (382 mg,
4.0 mmol) was added. The resulting mixture, containing some
precipitated NaCl, was combined with the ethanolic solution
from the preceding step. The reaction mixture was heated
under reflux for 18 h and then was chilled and filtered. The
filter cake was taken up in H2O, and the pH was neutralized
with 10% AcOH. A trace of solid precipitate was collected and
redissolved in a small volume of 1 N NaOH. This solution was
added back to the filtrate, which was then basified with 1 N
1660, 1560, 1505, 1460, 1405, 1290, 1250 cm-1 1H NMR
;
(DMSO-d6) δ 1.26 (m, 2H, CH2CH2CH2CO2H), 1.45 (p, J ) 7
Hz, 4H, CH2CH2CH2CH2CO2H), 2.13 (t, J ) 7 Hz, 2H, benzylic
CH2), 2.45 (t, CH2CO2H, partly overlapped by DMSO peak),
3.48 (s, bridge CH2, partly overlapped by broad H2O peak),
3.76 (s, OMe, partly overlapped by broad H2O peak), 5.66 (br
s, NH2), 6.04 (br s, NH2), 6.90 (m, 2H, aryl 3′- and 6′-H), 7.00
(dd, J ) 8 Hz, J ) 2 Hz, 1H, aryl 4′-H), 7.35 (s, 1H, pyrimidine
6-H). Anal. (C18H24N4O2‚1.25H2O) C, H, N.
2,4-Dia m in o-5-[5′-(6-ca r boxyh exyl)-2′-m eth oxy)ben zyl]-
p yr im id in e (7). A solution of 4 (120 mg, 0.3 mmol) in DMF
(10 mL) was shaken under H2 (initial pressure 3 atm) in the
presence of 5% Pd-C in a Parr apparatus for 20 h. The catalyst
was removed, and the solvent wasevaporated under reduced
pressure. The residue, consisting of the benzyl ester of 7, was
treated with a small volume of dilute NaOH, and a trace of
insoluble material was filtered off. The filtrate was acidified
with 10% AcOH and chilled, and the precipitate was collected
and dried on a lyophilizer to obtain 7 as a white solid: yield
115 mg (77%); mp 101-105 °C (softening without giving a true
melt); IR (KBr) ν 2860, 3210, 2950, 2870, 1670, 1565, 1510,
1470, 1410, 1295, 1260 cm-1 1H NMR (DMSO-d6) δ 1.23
;
(poorly resolved m, 4H, CH2CH2CH2CH2CO2H), 1.45 (m, 4H,
CH2CH2CH2CH2CH2CO2H), 2.14 (t, J ) 7 Hz, 2H, CH2CO2H),
2.43 (t, J ) 7 Hz, benzylic CH2, partly overlapped by DMSO),

1734 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 9
Rosowsky et al.
NaOH (5 mL) and concentrated to dryness by rotary evapora-
tion. The residue was taken up in H2O (40 mL), the solution
was acidified with 10% AcOH and chilled, and the precipitate
was collected, dried on a lyophilizer, and purified for elemental
analysis and bioassay by preparative HPLC (C18 silica gel, 18%
MeCN in 0.1 M NH4OAc, pH 7.4). Appropriately pooled
fractions were freeze-dried, and the solid was redissolved in 1
N NaOH. A trace of solid was filtered off, the filtrate was
reacidified with 10% AcOH and chilled, and the precipitate
was collected and dried in a lyophilizer to obtain 8 as a white
solid (66 mg, 9% overall yield from 26): mp ca. 140 °C
(softening without giving a true melt); IR (KBr) ν 3325-3380,
2920w, 1655, 1535br, 1495, 1380, 1275 cm-1; 1H NMR (DMSO-
d6) δ 3.50 (s, bridge benzylic CH2, partially overlapped by broad
H2O peak), 3.74 (s, OMe, partially overlapped by broad H2O
peak), 5.06 (s, 2H, CH2O), 6.18 (br s, NH2), 6.37(br s, NH2),
6.83 (m, 3H, aryl 3′-, 4′-, and 6′-H), 7.32 (s, 1H pyrimidine 6-H),
7.47(t, J ) 8 Hz, 1H, aryl 5′′-H), 7.61 (d, J ) 7 Hz, aryl 4′′-H),
7.87 (d, J ) 7 Hz, 1H, aryl 6′′-H), 7.99 (s, 1H, aryl 2′′-H). Anal.
(C20H20N4O4‚0.1AcOH‚2H2O) C, H, N.
pooled fractions were concentrated and freeze-dried, and the
solid was redissolved in dilute NH4OH and filtered. The filtrate
was acidified with 10% AcOH and the precipitate was collected
and freeze-dried to obtain 9 as a white solid (60 mg, 8% overall
yield from 28): mp 250-251 °C; IR (KBr) ν 3330, 3170, 1660,
1615sh, 1595sh, 1535, 1500, 1460, 1380, 1295, 1220 cm-1; 1H
NMR (DMSO)-d6) δ 3.49 (s, bridge CH2 partially overlapped
by broad H2O peak), 3.74 (s, OMe, partially overlapped by
broad H2O peak), 5.08 (s, 2H, benzylic CH2), 5.90 (br s, NH2),
6.20 (br s, NH2), 6.83 (m, 3H, 3′-, 4′-, and 6′-H), 7.37 (s, 1H,
pyrimidine 6-H), 7.50 (d, J ) 8 Hz, 2H, 2′′- and 6′′-H), 7.93 (d,
J ) 8 Hz, 3′′- and 5′′-H). Anal. (C20H20N4O2‚0.7AcOH) C, H,
N.
2,4-Dia m in o-5-[2′-m eth oxy-5′-[3-(2′′-ca r boxyp h en oxy)-
p r op yn -1-yl]ben zyl]p yr im id in e (10). Step 1. A mixture of
ethyl salicylate (29) (1.66 g, 0.01 mol), K2CO3 (2.07 g, 0.015
mol), a catalytic amount of 18-crown-6 (26 mg), and propargyl
bromide (1.12 mL, of 80% solution in toluene, calculated to
contain 1.19 g, 0.01 mol) in dry DMF (10 mL) was stirred at
room temperature for 20 h. The salts were filtered off, and
the solvent was evaporated under reduced pressure in a bath
maintained at 50 °C. The residue was partitioned between
EtOAc and H2O, and the organic layer evaporated to obtain
the propargyl ether 31 (2.1 g) as an oil suitable for use directly
in the next step: 1H NMR (CDCl3) δ 1.38 (t, J ) 7 Hz, 3H,
CH2CH3), 2.53 (t, J ) 2 Hz), 1H, CtCH), 4.36 (q, J ) 7 Hz,
2H, CH2CH3), 4.79 (d, J ) 2 Hz, 2H, tCCH2O), 7.05 (dt, J )
8 Hz, J ) 2 Hz, 1H, aryl 4-H), 7.12 (d, J ) 8 Hz, 1H, aryl
3-H), 7.47 (dt, J ) 8 Hz, J ) 2 Hz, 1H, aryl 5-H), 7.80 (dd, J
) 8 Hz, J ) 2 Hz, 1H, aryl 6-H).
ter t-Bu tyl 4-(3-F or m yl-4-m eth oxyp h en oxym eth yl)ben -
zoa te (28). To a 4:1 THF-DMF mixture (10 mL) were
sequentially added tert-butyl 4-bromomethylbenzoate (27)
(1.08 g, 4.0 mmol),13 3-hydroxy-6-methoxybenzaldehyde (23)
(496 mg, 3.25 mmol),2 K2CO3 (1.38 g, 4.0 mmol), and 18-
crown-6 (106 mg, 0.4 mmol). The mixture was heated at 70
°C for 40 min, after which the THF was blown off with a
stream of air. TLC (silica gel, 1:1 EtOAc-isooctane) showed
some unchanged bromide (Rf ) 0.6) and phenol (Rf ) 0.3),
along with a small new spot at Rf ) 0.5 (blue-fluorescent)
corresponding to the desired product. Replacement of the THF
and heating for another 66 h led to complete disappearance
of the phenol (Rf ) 0.3). The reaction mixture was diluted with
EtOAc, and the combined organic solvents were decanted from
the salts, washed with H2O, and evaporated to dryness under
reduced pressure to obtain a solid (1.3 g). Recrystallization
from MeOH afforded 28 as a light-yellow powder (0.85 g,
76%): mp 91-92 °C; IR (KBr) ν 3000, 2980, 1860, 1700, 1675,
1615, 1585, 1575, 1495, 1470, 1455, 1435, 1425, 1410, 1395,
Step 2. A mixture of propargyl ether 31 (306 mg, 1.5 mmol),
iodide 19 (356 mg, 1.0 mmol), (Ph3P)2PdCl2 (10 mg), CuI (1
mg), and Et3N (3 mL) in dry DMF (3 mL) was heated under
N2 at 65 °C for 3 h. The solvent was evaporated under reduced
pressure, and the residue was triturated with alternating
portions of isooctane and H2O. The residue, consisting of ester
33, was dissolved in 50% EtOH (200 mL) at 50 °C and treated
with Ba(OH)2‚8H2O (947 mg, 3.0 mmol). The mixture was
stirred at room temperature for 20 h, whereupon a solution of
(NH4)2CO3 (500 mg, 5 mmol) in H2O (10 mL) was added. After
10 min of vigorous stirring, the solid was removed by filtration
and the filtrate was concentrated to a small volume under
reduced pressure. Pure product for elemental analysis and
bioassay was isolated by preparative HPLC (C18 silica gel, 20%
MeCN in 0.1 M NH4OAc, pH 7.4, 280 nm). The major peak
(eluting at 13 min on an analytical C18 column using the same
elution system) was concentrated and freeze-dried, and the
residue was dissolved in small volume of dilute NH4OH. A
small amount of insoluble material was filtered off, and the
filtrate was chilled and acidified with 10% AcOH. The pre-
cipitate was collected and freeze-dried to obtain pure 10 as a
white solid (104 mg, 24%): mp 147-153 °C (softening rather
than giving a true melt); IR (KBr) ν 3360, 3210, 2960br, 2230,
1665, 1605, 1595, 1560, 1505, 1355, 1385, 1295, 1275sh, 1250
1365, 1315, 1300, 1280, 1255, 1220, 1200 cm-1 1H NMR
;
(CDCl3) δ 1.60 (s, 9H, t-Bu), 3.90 (s, 3H, OMe), 5.11 (s, 2H
benzylic CH2), 6.94 (d, J ) 9 Hz, 1H, 5-H), 7.21 (dd, J ) 9 Hz,
J ) 3 Hz, 6-H), 7.41 (d, J ) 3 Hz, 1H, 4-H), 7.46 (d, J ) 8 Hz,
2H, 2′- and 6′-H), 8.00 (d, J ) 8 Hz, 2H, 3′- and 5′-H). Anal.
(C20H22O5) C, H.
2,4-Dia m in o-5-(2′-m et h oxy-5′-(4′′-ca r b oxyb en zyloxy)-
ben zylp yr im id in e (9). Metallic Na (12 mg, 0.5 mmol) was
dissolved in absolute EtOH (3 mL), and the solvent was
evaporated under reduced pressure and replaced by dry DMSO
(2 mL). 3-Morpholinopropionitrile (280 mg, 2.0 mmol) was
added, and the mixture was placed in a bath preheated to 100
°C and was slowly treated with a solution of 28 (684 mg, 2.0
mmol) in dry DMSO (3 mL). After another 20 min at 100 °C,
the reaction mixture was cooled and partitioned between
EtOAc and dilute citric acid. The EtOAc layer was concen-
trated to dryness by rotary evaporation, and the residue was
taken up in absolute EtOH (10 mL). Aniline hydrochloride (260
mg, 2.0 mmol) was added, and the reaction mixture was
refluxed for 30 min. Separately, guanidine hydrochloride (478
mg, 5.0 mmol) was added to a solution of metallic Na (161 g,
7.0 mmol) in absolute EtOH (20 mL), and the mixture was
swirled for 5 min and added to the foregoing ethanolic solution
containing the anilino nitrile adduct. The mixture was refluxed
for 18 h and then was chilled and filtered. The filter cake,
which was completely soluble in H2O, was discarded, and the
filtrate was evaporated under reduced pressure. The residue
was taken up in trifluoroacetic acid (5 mL), the solution was
evaporated to dryness, and the residue was suspended in H2O
(30 mL) and dissolved by adding 1 N NaOH and stirring for
several minutes until the pH was 10. Reacidification with 10%
AcOH produced a solid, which was collected, dried in a
lyophilization apparatus (crude yield 720 mg), and purified for
elemental analysis and bioassay by preparative HPLC (C18
silica gel, 18% MeCN in 0.1 M NH4OAc, pH 7.4). Appropriately
cm-1 1H NMR (DMSO-d6) δ 3.51 (s, bridge CH2, partly
;
overlapped by broad H2O peak), 3.82 (s, OMe, partly over-
lapped by broad H2O peak), 5.05 (s, 2H, CH2O), 5.94 (br s,
NH2), 6.23 (br s, NH2), 6.95-7.55 (complex m, 7H, aryl and
pyrimidine protons), 7.62 (d, J ) 7 Hz, 1H, 3′′-H). Anal.
(C22H20N4O4‚1.5H2O) C, H, N.
2,4-Dia m in o-5-[2′-m eth oxy-5′-[3-(3′′-ca r boxyp h en oxy)-
1-p r op yn yl]ben zyl]p yr im id in e (11). Step 1. A mixture of
ethyl 3-hydroxybenzoate (30) (1.66 g, 0.01 mol), K2CO3 (2.07
g, 0.015 mol), a catalytic amount of 18-crown-6 (26 mg), and
propargyl bromide (1.12 mL, 1.49 g of 80% solution in toluene,
calculated to contain 0.01 mol) in dry DMF (10 mL) was stirred
at room temperature for 20 h. The salts were filtered off, and
the solvent was evaporated under reduced pressure in a bath
maintained at 50 °C. The residue was partitioned between
EtOAc and H2O, and the organic layer was evaporated to
obtain the propargyl ether 32 (2.3 g) as an oil suitable for use
directly in the next step: 1H NMR (CDCl3) δ 1.39 (t, J ) 7 Hz,
3H, CH3CH2), 2.53 (d, J ) 1 Hz, 1H, CtCΗ), 4.38 (q, J ) 7
Hz, 2H, CH3CH2), 4.74 (s, 2H, tCCH2O), 7.17 (d, J ) 8 Hz,

Pyrimidines as Inhibitors
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 9 1735
trimetrexate and piritrexim. J . Heterocycl. Chem. 1995, 32, 335-
340. (d) Rosowsky, A.; Forsch, R. A.; Queener, S. F. 2,4-
Diaminopyrido[3,2-d]pyrimidine inhibitors of dihydrofolate re-
ductase from Pneumocystis carinii and Toxoplasma gondii. J .
Med. Chem. 1995, 38, 2615-2620. (e) Rosowsky, A.; Papoulis,
A. T.; Queener, S. F. 2,4-Diaminothieno[2,3-d]pyrimidine lipo-
philic antifolates as inhibitors of Pneumocystis carinii and
Toxoplasma gondii dihydrofolate reductase. J . Med. Chem. 1997,
40, 3694-3699. (f) Rosowsky, A.; Papoulis, A. T.; Queener, S. F.
2,4-Diamino-6,7-dihydro-5H-cyclopenta[d]pyrimidine analogues
of trimethoprim as inhibitors of Pneumocystis carinii and
Toxoplasma gondii dihydrofolate reductase. J . Med. Chem. 1998,
41, 913-918. (g) Rosowsky, A.; Cody, G.; Galitsky, N.; Fu, H.;
Papoulis, A. T.; Queener, S. F. Structure-based design of
selective inhibitors of dihydrofolate reductase: Synthesis and
antiparasitic activity of 2,4-diaminopteridine analogues with a
bridged diarylamine side chain. J . Med. Chem. 1999, 42, 4853-
4860. (h) Rosowsky, A.; Fu, H.; Queener, S. F. Synthesis of 2,4-
diaminopyrido[2,3-d]pyrimidines and 2,4-diaminoquinazolines
with bulky dibenz[b,f]azepine and dibenzo[a,d]cycloheptene sub-
stituents at the 6-position as inhibitors of dihydrofolate reduc-
tases from Pneumocystis carinii, Toxoplasma gondii, and My-
cobacterium avium. J . Heterocycl. Chem. 2000, 37, 921-927. (i)
Rosowsky, A.; Fu, H. Synthesis of new 2,4-diamino-7H-pyrrolo-
[2,3-d]pyrimidines via the Taylor ring transformation/ring an-
nulation strategy. J . Heterocycl. Chem. 2001, 38, 1197-1201.
(j) Rosowsky, A.; Chen, H. A novel method of synthesis of 2,4-
diamino-6-arylmethylquinazolines using palladium(0)-catalyzed
organozinc chemistry. J . Org. Chem. 2001, 66, 7522-7526. (k)
Rosowsky, A.; Chen, H.; Fu, H.; Queener, S. F. Synthesis of new
2,4-diaminopyrido[2,3-d]pyrimidine and 2,4-diaminopyrrolo[2,3-
d]pyrimidine inhibitors of Pneumocystis carinii, Toxoplasma
gondii, and Mycobacterium avium dihydrofolate reductase.
Bioorg. Med. Chem. 2002, 11, 59-67.
1H, aryl 4-H), 7.36 (t, J ) 8 Hz, 1H, aryl 5-H), 7.65 (m, 2H,
aryl 2- and 6-H).
Step 2. A mixture of propargyl ether 32 (306 mg, 1.5 mmol),
iodide 19 (356 mg, 1.0 mmol), (Ph3P)2PdCl2 (10 mg), CuI (1
mg), and Et3N (3 mL) in dry DMF (3 mL) was heated under
N2 at 65 °C for 3.5 h. The solvent was evaporated under
reduced pressure, and the residue was triturated with alter-
nating portions of isooctane and H2O. The solid, consisting of
ester 34, was collected and dissolved in DMSO (5 mL) with
slight warming as needed. This solution was then swirled and
treated dropwise with a solution of NaOH (120 mg, 3.0 mmol)
in H2O (1 mL). The mixture was diluted to 90 mL with H2O
and then brought to approximately pH 8 with 10% AcOH. Pure
9 for microchemical analysis and bioassay was isolated by
preparative HPLC (C18 silica gel, 20% MeCN in 0.1 M NH4-
OAc, pH 7.4, 280 nm). Appropriately pooled eluates were
concentrated and freeze-dried, the residue was redissolved in
0.1 N NaOH, and the solution was filtered. Acidification with
10% AcOH, followed by freeze-drying, yielded pure 9 as a white
powder (100 mg, 21%): mp 149-153 °C (softening without
giving a true melt); IR (KBr) ν 3360, 3200, 2960 (broad), 1670,
1610, 1560, 1505, 1460, 1445, 1390, 1275, 1245 cm-1; 1H NMR
(DMSO-d6) δ 3.51 (s, benzylic CH2, partly overlapped by broad
H2O peak), 3.82 (s, OMe, partly overlapped by broad H2O
peak), 5.04 (s, 2H, CH2O), 5.94 (br s, NH2), 6.21 (br s, NH2),
6.99 (d, J ) 8 Hz, 1H, 3′-H), 7.11 (d, J ) 2 Hz, 1H, 6′-H), 7.21
(dd, J ) 8 Hz, J ) 2 Hz, 1H, 4′-H), 7.28-7.45 (m, 3H, 5′′- and
6′′-H, pyrimidine 6-H), 7.54 (m, 2H, 2′′ and 4′′-H). Anal.
(C22H20N4O4‚AcOH‚1.25H2O) C, H, N.
En zym e Assa ys a n d Da ta An a lysis. Standardized spec-
trophotometric determination of Pc, Tg, Ma, and rat liver
DHFR inhibition was performed as described earlier.15,16 IC50
values were obtained with the aid of the curve-fitting program
Prism 3.0. The data were transformed to percentage of
uninhibited control for enzyme activity (y-axis) and the log of
the molar drug concentration (x-axis). The analyses for indi-
vidual curves had degrees of freedom ranging from 6 (2 curves)
to 42 (1 curve); the mode for degrees of freedom was 12 (14
curves). The r2 value for individual curves, determined using
the program InStat 2.03, ranged from 0.71 (1 curve) to >0.99
(24 curves); all but two curves had r2 values of >0.93. The
range of SI values for each compound was calculated from the
95% confidence intervals for the individual IC50 values for that
compound; all were referenced to rat liver DHFR. As an
example of the reproducibility of these methods, when 1b was
assayed against Pc DHFR on three independent occasions, the
mean ( SEM (standard error of the mean) was found to be
0.054 ( 0.0023 µM (i.e., the SEM was 4.3% of the mean). In
quality control assays using TMP against the same enzyme,
the SEM was 6.3% of the mean.
(4) Sigel, C. W.; Macklin, A. W.; Woolley, J . L, J r.; J ohnson, N. W.;
Collier, M. A.; Blum, R. M.; Clendeninnn, N. J .; Everitt, B. M.
J .; Grebe, G.; Mackars, A.; Foss, R. G.; Duch, D. S.; Bowers, S.
W.; Nichol, C. A. Preclinical biochemical pharmacology and
toxicology of piritrexim, a lipophilic inhibitor of dihydrofolate
reductase. Natl. Cancer Inst. Monogr. 1987, 5, 111-120.
(5) Falloon, J .; Allegra, C. J .; Kovacs, J .; O’Neill, D.; Ogata-Arakaki,
D.; Feuerstein, I.; Polis, M.; Davey, R.; Lane, H. C.; LaFon, S.;
Rogers, M.; Zunich, K.; Turlo, J .; Tuazon, C.; Parenti, D.; Simon,
G.; Masur, H. Piritrexim with leucovorin for the treatment of
pneumocystis pneumonia (PCP) in AIDS patients. Clin. Res.
1990, 38, 361A.
(6) Cody, V. Personal communication. For a preliminary report on
the binary complexes with human DHFR, see the following.
Cody, V.; Rosowsky, A. Crystal structure determination of
human DHFR binary complexes with TMP and modeling studies
of TMP analogues. Am. Assoc. Cancer Res. Proc. 2002, 43, 203.
(7) Kuyper, L. F.; Roth, B.; Baccanari, D. P.; Ferone, R.; Beddell,
C. R.; Champness, J . N.; Stammers, D. K.; Dann, J . G.;
Norrington, F. E.; Baker, D. J .; Goodford, P. J . Receptor-based
design of dihydrofolate reductase inhibitors: comparison of
crystallographically determined enzyme binding with enzyme
affinity in a series of carboxy-substituted trimethoprim ana-
logues. J . Med. Chem. 1985, 28, 303-311.
(8) The highly conserved Arg75 residue of Pc DHFR is similarly
found at position 75 in both rat and mouse DHFR and corre-
sponds to Arg57 of the E. coli enzyme. For a recently completed
sequence analysis of cloned rat DHFR, see the following. Wang,
Y.; Bruenn, J . A.; Queener, S. F.; Cody, V. Isolation of rat
dihydrofolate reductase gene and characterization of recombi-
nant enzyme. Antimicrob. Agents Chemother. 2001, 45, 2517-
2523.
Ack n ow led gm en t. This work was supported by
Research Grant RO1-AI29904 (A.R.) from the National
Institute of Allergy and Infectious Diseases (NIAID),
U.S. Department of Health and Human Services.
Refer en ces
(9) Although the 3D structure for the ternary complex of TMP with
rat DHFR remains unpublished at this time, the hypothesis that
the terminal COOH group in our 5′-modified TMP analogues is
likely to interact with Arg75 is supported by the high degree of
homology between rat and mouse DHFR, and the binding mode
of TMP to mouse and E. coli DHFR is similar. Groom, C. R.;
Thillet, J .; North, A. C. T.; Pictet, R.; Geddes, A. J . Trimethoprim
binds in a bacterial mode to the wild-type and E30D mutant of
mouse dihydrofolate reductase. J . Biol. Chem. 1991, 266, 19890-
19893.
(10) (a) Sonogashira, K.; Tohda, Y.; Hagihara, B. Convenient syn-
thesis of acetylenes. Catalytic substitutions of acetylenic hydro-
gen with bromo alkenes, iodo arenes, and bromopyridines.
Tetrahedron Lett. 1975, 4467-4470. (b) Takahashi, S.; Ku-
royama, Y.; Sonogashira, K.; Hagihara, N. A convenient syn-
thesis of ethynylarenes and diethynylarenes. Synthesis 1980,
627-630.
(1) Klepser, M. W.; Klepser, T. B. Drug treatment of HIV-related
opportunistic infections. Drugs 1997, 3, 40-73.
(2) Rosowsky, A.; Forsch, R. A.; Queener, S. F. Inhibition of
Pneumocystis carinii, Toxoplasma gondii, and Mycobacterium
avium dihydrofolate reductases by 2,4-diamino-5-[2-methoxy-5-
(ω-carboxyalkoxy)benzyl]pyrimidines: marked improvement in
potency relative to trimethoprim and species selectivity relative
to piritrexim. J . Med. Chem. 2002, 45, 233-241.
(3) The following is a selected list of our publications in this area
from 1993 to date. (a) Rosowsky, A.; Mota, C. E.; Wright, J . E.;
Heusner, J . J .; McCormack, J . J .; Queener, S. F. 2,4-Diami-
nothieno[2,3-d]pyrimidine analogues of trimetrexate and piri-
trexim as potential inhibitors of Pneumocystis carinii and
Toxoplasma gondii dihydrofolate reductase. J . Med. Chem. 1993,
36, 3103-3112. (b) Rosowsky, A.; Mota, C. E.; Wright, J . E.;
Queener, S. F. 2,4-Diamino-5-chloroquinazoline analogues of
trimetrexate and piritrexim: synthesis and antifolate activity.
J . Med. Chem. 1994, 37, 4511-4528. (c) Rosowsky, A.; Mota, C.
E.; Queener, S. F. Synthesis and antifolate activity of 2,4-
diamino-5,6,7,8-tetrahydropyrido[4,3-d]pyrimidine analogues of
(11) Calas, M.; Barbier, A.; Giral, L.; Balmayer, B.; Despaux, E.
Synthesis of new trimethoprim analogs. Antibacterial structure-
activity relationship. Eur. J . Med. Chem. 1982, 17, 497-504.

1736 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 9
Rosowsky et al.
(12) Scarpati, M. L.; Bianco, A.; Mascitelli, L.; Passacantilli, P.
Selective formylation of diphenols. Synth. Commun. 1990, 20,
2565-2572.
(13) Rosowsky, A.; Forsch, R. A.; Moran, R. G. (6R,6S)-5,8,10-
Trideaza-5,6,7,8-tetrahydrofolate and (6R,6S)-5,8,10-trideaza-
5,6,7,8-tetrahydropteroyl-L-ornithine as potential antifolates and
antitumor agents. J . Med. Chem. 1989, 32, 709-715.
(14) (a) Kompis, I.; Wick, A. Synthesis of 4-halogen substituted
analogues of trimethoprim. Helv. Chim. Acta 1977, 60, 3025-
3034. (b) Roth, B.; Baccanari, D. P.; Sigel, C. W.; Hubbell, J . P.;
Eaddy, J .; Kao, J . C.; Grace, M. E.; Grace, M. E.; Rauckman, R.
S. 2,4-Diamino-5-benzylpyrimidines and analogues as antibacte-
rial agents. 9. Lipophilic trimethoprim analogues as antigono-
coccal agents. J . Med. Chem. 1988, 31, 122-129.
(15) Broughton, M. C.; Queener, S. F. Pneumocystis carinii dihydro-
folate reductase used to screen potential antipneumocystis drugs.
Antimicrob. Agents Chemother. 1991, 35, 1348-1355.
(16) Chio, L.-C.; Queener, S. F. Identification of highly potent and
selective inhibitors of Toxoplasma gondii dihydrofolate reduc-
tase. Antimicrob. Agents Chemother. 1993, 37, 1914-1923.
(17) Gangjee, A.; Vasudevan, A.; Queener, S. F.; Kisliuk, R. L. 2,4-
Diamino-5-deaza-6-substituted pyrido[2,3-d]pyrimidine anti-
folates as potent and selective nonclassical inhibitors of dihy-
drofolate reductases. J . Med. Chem. 1996, 39, 1438-1446.
(18) Stevens, M. F. G.; Phillip, K. S.; Rathbone, D. L.; O’Shea, D.
M.; Queener, S. F.; Schwalbe, C. H.; Lambert, P. A. Structural
studies on bioactive compounds. 28. Selective activity of triaz-
enyl-substituted pyrimethamine derivatives against Pneumocys-
tis carinii dihydrofolate reductase. J . Med. Chem. 1997, 40,
1886-1893.
(19) Piper, J . R.; J ohnson, C. A.; Krauth, C. A.; Carter, R. L.; Hosmer,
C. A.; Queener, S. F.; Borotz, S. E.; Pfefferkorn, E. R. Lipophilic
antifolates as agents against opportunistic infections. 1. Agents
superior to trimetrexate and piritrexim against Toxoplasma
gondii and Pneumocystis carinii in in vitro evaluations. J . Med.
Chem. 1996, 39, 1271-1280.
(20) Gangjee, A.; Mavandadi, F.; Queener, S. F.; McGuire, J . J . Novel
2,4-Diamino-5-substituted-pyrrolo[2,3-d]pyrimidines as classical
and nonclassical antifolate inhibitors of dihydrofolate reductases.
J . Med. Chem. 1995, 38, 2158-2165.
(21) Roos, D. S. Primary structure of the dihydrofolate reductase-
thymidylate synthase gene from Toxoplasma gondii. J . Biol.
Chem. 1993, 268, 6269-6280.
(22) Suling, W. J .; Seitz, L. E.; Pathak, V.; Westbrook, L.; Barrow,
E. W.; Zywno-Van Ginkel, S.; Reynolds, R. C.; Piper, J . R.;
Barrow, W. W. Antimycobacterial activities of 2,4-diamino-5-
deazapteridine derivatives and effects on mycobacterial dihy-
drofolate reductase. Antimicrob. Agents Chemother. 2000, 44,
2784-2793.
(23) Blakley, R. L. Folates and Pterins: Chemistry and Biochemistry
of Folates; Blakley, R. L., Benkovic, S. J ., Eds.; Wiley-Inter-
science: New York, 1984; Vol. 1, pp 191-253.
(24) Li, R.; Sirawaporn, R.; Chitnumsub, P.; Sirawaporn, W.; Wooden,
J .; Athappilly, F.; Turley, S.; Hol, W. G. Three-dimensional
structure of M. tuberculosis dihydrofolate reductase reveals
opportunities for the design of novel tuberculosis drugs. J . Mol.
Biol. 1999, 295, 307-323.
J M020466N