Selective Pneumocystis carinii dihydrofolate reductase inhibitors: Design, synthesis, and biological evaluation of new 2,4-diamino-5- substituted-furo[2,3-d]pyrimidines

Article abstract of DOI:10.1021/jm970537w

Nonclassical antifolates, 2,4-diamino-5-substituted-furo[2,3- d]pyrimidines 3-12 with bridge region variations of C8-S9, C8-N9, and C8-O9 and 1-naphthyl, 2-naphthyl, 2-phenoxyphenyl, 4-phenoxyphenyl, and 2-biphenyl side chains were synthesized as phenyl r

Full text of DOI:10.1021/jm970537w

J . Med. Chem. 1998, 41, 1263-1271
1263
Selective P n eu m ocystis ca r in ii Dih yd r ofola te Red u cta se In h ibitor s: Design ,
Syn th esis, a n d Biologica l Eva lu a tion of New
2,4-Dia m in o-5-su bstitu ted -fu r o[2,3-d ]p yr im id in es1
Aleem Gangjee,*,† Xin Guo,† Sherry F. Queener,‡ Vivian Cody,§ Nikolai Galitsky,§ J oe R. Luft,§ and
Walter Pangborn§
Division of Medicinal Chemistry, Graduate School of Pharmaceutical Sciences, Duquesne University,
Pittsburgh, Pennsylvania 15282, Department of Pharmacology and Toxicology, Indiana University School of Medicine,
Indianapolis, Indiana 46202, and Hauptman-Woodward Medical Research Institute, Inc., 73 High Street,
Buffalo, New York 14203
Received August 13, 1997
Nonclassical antifolates, 2,4-diamino-5-substituted-furo[2,3-d]pyrimidines 3-12 with bridge
region variations of C8-S9, C8-N9, and C8-O9 and 1-naphthyl, 2-naphthyl, 2-phenoxyphenyl,
4-phenoxyphenyl, and 2-biphenyl side chains were synthesized as phenyl ring appended
analogues of previously reported 2,4-diamino-5-(anilinomethyl)furo[2,3-d]pyrimidines. The
phenyl ring appended analogues were designed to specifically interact with Phe69 of
dihydrofolate reductase (DHFR) from Pneumocystis carinii (pc) to afford selective inhibitors of
pcDHFR. Additional substituted phenyl side chains which include 2,5-dichloro, 3,4-dichloro,
3,4,5-trichloro, 3-methoxy, and 2,5-dimethoxy analogues 13-17 were also synthesized. The
compounds were prepared by nucleophilic displacement of 2,4-diamino-5-(chloromethyl)furo-
[2,3-d]pyrimidine(2) with the appropriate thiol, amine, or naphthol. Compound 2 was obtained
from 2,4-diamino-6-hydroxypyrimidine and 1,3-dichloroacetone. The compounds were evaluated
as inhibitors against DHFR from P. carinii, Toxoplasma gondii, and rat liver. Two analogues,
2,4-diamino-5-[(2′-naphthylthio)methyl]furo[2,3-d]pyrimidine (5) and 2,4-diamino-5-[(2′-phen-
ylanilino)methyl]furo[2,3-d]pyrimidine (11) showed significant selectivity and potency for
pcDHFR compared to trimethoprim. The X-ray crystal structure of 5 with pcDHFR was also
carried out, which corroborated the design rationale and indicated a hydrophobic interaction
of the naphthalene ring of 5 and Phe69 of pcDHFR which is responsible, in part, for the more
than 18-fold selectivity of 5 for pcDHFR as compared with rat liver DHFR.
In tr od u ction
Pneumocystis carinii (pc) and Toxoplasma gondii (tg)
are opportunistic organisms which are responsible for
significant morbidity and mortality in immunocompro-
mised patients such as those with AIDS.2 Selectivity
of inhibitors for dihydrofolate reductase (DHFR) from
P. carinii and/or T. gondii over mammalian DHFR, such
as rat liver (rl) DHFR along with high potency against
these pathogenic DHFR, is a desirable goal. Currently
used DHFR inhibitors against P. carinii and T. gondii
include the weakly inhibitory monocyclic agents tri-
methoprim (TMP) and pyrimethamine, both of which
need augmentation with sulfonamides for clinical util-
ity, and the highly potent but nonselective bicyclic
trimetrexate (TMQ), which is significantly toxic and
must be used with leucovorin, a rescue agent for host
cells.3 The toxicities of these DHFR inhibitors or their
combination with sulfonamides are severe enough in
many cases to force discontinuation of therapy.4
and T. gondii. It was found that the furo[2,3-d]pyrim-
idines were in general more selective for pcDHFR while
pyrrolo[2,3-d]pyrimidines were more selective for tgDH-
FR. Though agents with selectivity for tgDHFR along
with potency have been reported in the 6-6 and 6-5
fused bicyclic antifolates, selectivity for pcDHFR has
been elusive in these series of antifolates. Several 6-6
fused analogues which include pteridines, quinazolines,
and pyridopyrimidines and 6-5 fused systems which
include thienopyrimidines and pyrrolopyrimidines show
reasonable selectivity for tgDHFR vs rlDHFR but
significantly lower or no selectivity for pcDHFR.7 Thus
the selectivity for pcDHFR vs rlDHFR of the series of
2,4-diamino-5-substituted-furo[2,3-d]pyrimidines re-
ported by Gangjee et al.5 along with a complete lack of
selectivity of these analogues for tgDHFR indicated a
Gangjee et al.5,6 recently reported the inhibitory
effects of a series of novel nonclassical and classical 2,4-
diamino-5-substituted-furo[2,3-d]pyrimidine5 antifolates
and nonclassical 2,4-diamino-5-substituted-pyrrolo[2,3-
d]pyrimidine6 antifolates against DHFR from P. carinii
† Duquesne University.
‡ Indiana University.
§ Hauptman-Woodward Medical Research Institute, Inc.
S0022-2623(97)00537-2 CCC: $15.00 © 1998 American Chemical Society
Published on Web 03/18/1998

1264 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 8
Gangjee et al.
Sch em e 1
reversal in the general trend observed in the previously
reported bicyclic 6-6 and 6-5 ring fused antifolates. It
was possible that the furo[2,3-d]pyrimidine ring of the
analogues reported was responsible, in part, for the
pcDHFR selectivity. In the reported series of furo[2,3-
d]pyrimidines, the most selective pcDHFR analogue was
the classical N9-methyl analogue, 1, with a selectivity
ratio of 12.3 for pcDHFR. The most selective bicyclic
pcDHFR inhibitor reported to date is 2,4-diamino-6-
[(thiophenyl)methyl]pteridine which has a selectivity
ratio of 25.9 (vs rlDHFR).8 Since P. carinii lacks the
active transport systems required for folates and re-
duced folate uptake, compound 1 would not be taken
up by P. carinii cells, and hence was not a viable anti
P. carinii agent. The lipophilic, nonclassical 2,4-diami-
nofuro[2,3-d]pyrimidine antifolates reported in the same
study had selectivity ratios of 1.4 to 3.1. It was there-
fore of interest to explore further the 2,4-diaminofuro-
[2,3-d]pyrimidine ring system with different side chain
substituents to improve the selectivity for pcDHFR.
At the time the previous study had been completed,
the crystal structure of pcDHFR was not available. The
binding of the 2,4-diamino-5-substituted-furo[2,3-d]py-
rimidine was expected to parallel that of other 6-6
bicyclic compounds such as TMQ and the 2,4-diamino-
6-substituted-pyrido[2,3-d]pyrimidine analogue piri-
trexim (PTX). The principal active site residues of
pcDHFR were known to be the same as in human (h)
DHFR or only conservatively different.9 For example,
the Glu30 of hDHFR, which forms the salt bridge with
the N1 and the 2-amino group of the 2,4-diaminopyri-
midine moiety of the antifolate, is replaced by Glu32 in
pcDHFR. The Phe31 is replaced by Ile33 (pcDHFR) and
the Phe34 is maintained as Phe36 in pcDHFR. Other
residues which interact with the bicyclic system are all
maintained or conserved in pcDHFR as evident from
the crystal structures of TMP and PTX.10 Replacement
of Asn64 in hDHFR by Phe69 in pcDHFR is of particular
interest because this residue lies just outside the
diaminopyrimidine binding site.9 Appropriate muta-
tions of hDHFR to provide a pseudo pcDHFR were
carried out by using SYBYL 5.511 and its BIOPOLY-
MER option. On the basis of this pseudo pcDHFR and
molecular modeling with SYBYL, it was reasoned that
by designing an additional phenyl ring into the side
chain phenyl of the lipophilic nonclassical 2,4-diamino-
5-[(substituted anilino)methyl]furo[2,3-d]pyrimidine an-
tifolates reported previously5 it may be possible to reach
the Phe69 in pcDHFR in a hydrophobic “edge on”
interaction of the added phenyl ring with Phe69. Such
an interaction could provide selectivity for pcDHFR vs
mammalian DHFR where an Asn649 resides to which
the added phenyl ring was not expected to provide
conducive binding.
the crystal structure of compound 5 in a ternary complex
with pcDHFR and NADPH.
Ch em istr y
To develop a structure-activity/selectivity relation-
ship for the 2,4-diamino-5-substituted-furo[2,3-d]pyri-
midines, it was necessary to use a versatile intermediate
from which a series of nonclassical analogues 3-17 with
variations in both the bridge and the 5-substituted side
chain aromatic moiety could be obtained in a minimum
of steps. The key intermediate chosen for this purpose
was 2,4-diamino-5-(chloromethyl)furo[2,3-d]pyrimidine
(2), reported by Secrist and Liu in 1978.12 The chlo-
romethyl moiety of 2 could be suitably transformed by
nucleophilic displacements to afford the target mol-
ecules in a convergent manner similar to the previous
report by Gangjee et al.5
Compound 2 was prepared from 2,6-diamino-4-hy-
droxypyrimidine and 1,3-dichloroacetone in anhydrous
DMF at room temperature using an improved modifica-
tion of the published method.12 Analysis of the mixture
by TLC after 24 h showed complete disappearance of
the pyrimidine and the formation of a mobile product
spot along with byproducts which remained at the
origin. However, workup of the reaction proved difficult
because 2 was unstable at room temperature and de-
composed on prolonged exposure to air. It was re-
ported12 that filtration of the reaction mixture afforded
the product as a precipitate and that chromatography
of the filtrate afford some additional product (total yield
78%). In our hands, improved results were obtained by
rapidly purifying compound 2 before condensing it with
the appropriate nucleophiles. This was achieved by the
addition of silica gel directly to the reaction mixture
along with additional DMF once the reaction was over.
The mixture was then sonicated and rapidly evaporated
in vacuo at 50-55 °C. The resultant plug was kept in
the dark and dried over P2O5 in vacuo overnight at room
temperature. The plug was then ground into a fine
powder and poured on top of a short, broad column of
silica gel (15 w/w) which was eluted with methanol/
chloroform, 1:7. The pure, dried compound 2 was imme-
diately sealed and stored at low temperature (<-20 °C).
Thus we designed compounds 4-12 (Scheme 1) which
contain an additional phenyl ring either fused to or
attached to the normal phenyl side chain. This took the
form of a 1- or 2-naphthyl, biphenyl, or diphenyl ether
substituent. Compounds 3 and 13-17 were also in-
cluded as new examples of compounds with a single
phenyl ring side chain with N9-CH3 and ring substit-
uents which had not been reported previously.5
To validate our hypothesis concerning the role of
Phe69 and pcDHFR selectivity, we have also determined

P. carinii Dihydrofolate Reductase Inhibitors
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 8 1265
F igu r e 1. Stereoview of active site of pcDHFR complex with NADPH and compound 5 highlighting the intermolecular contacts
of the 2-naphthyl group (yellow) with Phe-69 (green). The van der Waals surface is shown for NADPH, 5, and Phe-69.
Nucleophilic displacement reactions of the chloride 2
with appropriate anilines, phenol, or thiophenols were
accomplished in DMSO in the presence of potassium
carbonate. The product was precipitated by quenching
the reaction mixture with water and was purified by
column chromatography. Analogues 3, 5, and 13-16
were synthesized from crude 2, and the yields were very
low (3-16% for two steps). However, when purified 2
was used, the reactions were much cleaner and the
yields improved significantly (16-48%).
The N9-methyl analogues of this series of furo[2,3-d]-
pyrimidines were prepared by the following two meth-
ods. For compounds 15 and 17, the appropriate com-
mercially available primary anilines were transformed
into their N-methylated analogues, by alkylation with
methyl iodide at 40 °C in CH3CN in the presence of
potassium carbonate. Reaction of the N-methylated
anilines with 2 afforded 15 and 17. In the second
approach, reductive methylation using formaldehyde
and sodium borohydride in an acidic medium was used
to synthesize 12 directly from 7. This method is well
documented for the synthesis of N10-methyl bicyclic
antifolates.13 Since the furo[2,3-d]pyrimidines are not
very stable under acidic conditions, previous attempts
using acetic acid as both the acidic catalyst and the
solvent resulted in slow reaction and yielded multiple
products. However, when 2,4-diamino-5-(2-naphthyl-
aminomethyl)furo[2,3-d]pyrimidine 7, formaldehyde,
and sodium cyanoborohydride were suspended in ac-
etonitrile and concentrated HCl was added dropwise
until a clear solution formed, the reaction was fast and
completed within 1 h, which precluded significant
degradation or dimethylation. The reaction afforded the
desired monomethylated analogue 12 in 64% yield.
region is homologous to those of the mammalian and
bacterial enzymes, reemphasizing the conserved nature
of the active site despite changes in enzyme size.
The binding orientation of 5 is similar to that ob-
served for the furo[2,3-d]pyrimidine classical analogue,
1.14 The carboxylate oxygens of Glu32 protonate the
furopyrimidine ring forming hydrogen bonds to N1 by
OE2 and to the N2-amino nitrogen by OE1. In addition,
OE1 and OE2 form hydrogen bonds to a conserved
threonine and a structural water. The 2-amino group
also hydrogen bonds to a structurally conserved water.
In contrast, no structural water was observed near O8,
as is seen near N8 of the pteridine antifolates.15,16
Because the geometry of the furo[2,3-d]pyrimidine
differs significantly from that of pteridine, the 5-sub-
stituent of the furo[2,3-d]pyrimidine is oriented differ-
ently in the active site.15,16 This change places the C8-
N9 bridge of 5 in a different orientation from that
observed for the C9-N10 bridge of MTX.
Analysis of the intermolecular interactions involving
the 2-naphthyl group shows that it makes hydrophobic
contacts (<4.0 Å) to Phe69 (Figure 1) and Pro66.
Similar hydrophobic interactions with Phe69 and the
glutamate methylene carbons of the classical furopyri-
midine 1 have been observed in its ternary complex with
pcDHFR and NADPH.14
Biologica l Resu lts a n d Discu ssion
Compounds 3-17 were evaluated as inhibitors of
DHFRs from P. carinii, T. gondii, and rat liver which
served as the mammalian reference.8,17 The inhibitory
concentration (IC50) values and the selectivity ratios
(IC50 rlDHFR/IC50 pcDHFR or tgDHFR) for compounds
3-17 are listed in Table 1 along with epiroprim, TMP,
and TMQ as reference standards. The low inhibitory
activity with marginal pcDHFR selectivity previously
observed5 with the di- and tri-OCH3Ph and di- and tri-
ClPh nonclassical furo[2,3-d]pyrimidine analogues was
again evident in most instances in the present study.
Significantly, however, the 2-naphthylthio analogue 5
X-r a y Cr ysta l Str u ctu r e
The overall characteristics of the pcDHFR ternary
complex with NADPH and 5 are similar to those
reported for the TMP complex10 and for the classical
furo[2,3-d]pyrimidine analogue 1.14 The active site

1266 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 8
Gangjee et al.
Ta ble 1. Inhibitory Concentrations (IC50, µM) and Selectivity Ratios of 5-Substituted Furo[2,3-d]pyrimidines against pcDHFR,
tgDHFR, and rlDHFR
DHFR
no.
X
Ar
phenyl
pc
>26
19
0.65
13.5
41
14
>12
8.1
7.7
14.8
50.9
44.8
284
>31.3
>27
0.042
12
2.6
rl
252
23
12.3
12
36.5
60.3
>12
16.2
137
14.6
71.9
>27
34.3
>31.3
>27
0.003
133
33.2
rl/pc
tg
rl/tg
3
4
5
6
7
S
S
S
NH
NH
O
NH
NH
NH
ND
1.2
>26
19
11.6
37
38
>42
>12
32.4
45.4
23.6
>47
>27
21.5
ND
1.2
1.1
1-naphthyl
2-naphthyl
1-naphthyl
2-naphthyl
2-naphthyl
2-phenoxyPh
4-phenoxyPh
2-phenylPh
2-naphthyl
2,5-diClPh
3,4-diClPh
3,4,5-triClPh
3-OMePh
18.9
0.89
0.89
4.31
ND
2.00
17.79
0.99
1.4
ND
0.1
ND
ND
0.32
0.96
ND
ND
0.50
3.02
0.62
ND
ND
1.6
ND
ND
0.3
49
70.6
8
9
10
11
12
13
14
15
16
17
N(CH3)
NH
N(CH3)
N(CH3)
NH
N(CH3)
TMQ
TMP
>31.3
>27
2,5-diOMePh
0.07
11.1
12.8
0.010
2.7
0.48
Epiroprim
and the biphenyl-2-ylamino analogue 11 gave selectivity
ratios of 18.9 and 17.8, respectively, against pcDHFR.
These selectivity ratios rank compounds 5 and 11 as
the second and third most selective bicyclic pcDHFR
inhibitors (vs rlDHFR) reported to date.8,18 Compound
5, in addition to its considerable selectivity, was also
reasonably potent against pcDHFR with an IC50 of 0.65
µM.
again underscore the previous conclusion that all three
features, i.e., the heterocyclic ring, the bridge, and the
side chain aryl moiety, combine to provide pcDHFR
selectivity in this series.
Analogues 3 and 13-17, with a single phenyl ring in
the side chain, generally showed low potency and poor
selectivity for pcDHFR compared to the most potent and
selective analogue 5. Methylation of the bridge nitrogen
in bicyclic 6-6 and 6-5 ring systems usually leads to
an increase in DHFR potency.7,11 In the furo[2,3-d]-
pyrimidine series, the N9-CH3 analogues 12, 14, and
17 had IC50 values similar to those of the corresponding
N9-H analogues, and in the case of compound 15, N9-
methylation actually produced a 34-fold decrease in
potency against pcDHFR.5
In contrast to the selectivity obtained for pcDHFR,
these analogues were devoid of any selectivity for
tgDHFR with the exception of 11 and 15, which were
marginally selective. The potency of these analogues
against tgDHFR was also low, with IC50 values greater
than 11 µM.
The idea that an added phenyl ring to the side chain
phenyl of the previously reported 2,4-diaminofuro[2,3-
d]pyrimidines would result in increased selectivity for
pcDHFR was indeed realized in compounds 5 and 11.
It therefore became of interest to determine if this
increased selectivity could be attributed to a favorable
“edge on” interaction of the appended phenyl ring with
the phenyl ring of Phe69 of pcDHFR as was originally
proposed in our design of compounds 5 and 11.
As illustrated in Figure 1, the X-ray crystal structure
of 5 as a ternary complex with pcDHFR and NADPH
verifies the close hydrophobic interactions of the second
fused phenyl ring of the 2-naphthyl substituent of 5 with
the Phe69 of pcDHFR. This interaction may be respon-
sible, in part, for the selectivity of compound 5 for
pcDHFR since such an interaction is not possible for
the human enzyme as there is a sequence change to
Asn64 at this position in human DHFR.
The most selective bicyclic pcDHFR inhibitor (vs
rlDHFR) 2,4-diamino-6-[(thiophenyl)methyl]pteridine
has a pcDHFR selectivity ratio of 25.9 with an IC50 of
9.5 µM.8 Compound 5 is 15-fold more potent and is only
slightly less selective with a selectivity ratio of 18.9.
Compared to epiroprim and TMP, compound 5 is
significantly more potent and more selective against
pcDHFR in our assay, which is performed in the
presence of saturating concentrations of the substrate
and cofactor with a 37 °C incubating temperature.8
Comparison of the thio bridged analogues with the
corresponding NH-bridged analogues (i.e., 4 and 5 with
6 and 7) in the 1- and 2-naphthyl series indicates that
for selectivity and potency against pcDHFR the S-2-
naphthyl moiety was essential. All three of the other
analogues, 4, 6, and 7, not only lacked appreciable
selectivity but also were poorly inhibitory. Compound
8 in which the heteroatom in the bridge was isosterically
substituted by an oxygen has similar potency as the NH
analogue with slightly improved selectivity for pcDHFR.
These results indicate that both potency and selectivity
are determined by the heterocycle as well as the side
chain substituent including the bridge. On the basis
of the significant potency and selectivity of 5, three
analogues, 9-11, with an appended phenyl ring in the
side chain were synthesized in the NH-bridged series.
Compound 11, with 17.8-fold selectivity and an IC50 of
7.7 µM, was comparable to epiroprim and TMP in its
potency but had better selectivity against pcDHFR. The
other analogues were neither potent nor selective for
pcDHFR. The results in the CH2NH-bridged series

P. carinii Dihydrofolate Reductase Inhibitors
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 8 1267
solid: mp 227 °C (dec); TLC Rf 0.48 (CHCl3/MeOH, 7:1, with
one drop of NH4OH, silica gel); 1H NMR (DMSO-d6) δ 4.27 (s,
2 H, 5-CH2), 6.03 (s, 2 H, 4-NH2), 6.56 (s, 2 H, 2-NH2), 7.14 (s,
1 H, 6-CH), 7.18 (m, 1 H, 4′-CH), 7.23-7.40 (m, 4 H, 2′,3′,5′,6′-
CH). Anal. (C13H12N4OS‚0.10CH3OH‚0.22H2O) C, H, N, S.
Compound 5, on the basis of its potency and selectiv-
ity for pcDHFR, was evaluated in inhibitory studies of
the growth of P. carinii cells in culture. Initial studies
indicate that the IC50 against the growth of P. carinii
cells in culture is between 1 and 10 µM and the
compound is not cytotoxic at 10 µM. Compound 5 is
currently undergoing animal studies in the P. carinii
pneumonia mouse model.
In summary, compounds 4-11 were designed as “ex-
tended” phenyl ring analogues of previously reported
pcDHFR selective 2,4-diaminofuro[2,3-d]pyrimidines,
with the hypothesis that this extended phenyl ring
would favorably interact with Phe69 on pcDHFR which
is different from Asn64 in mammalian DHFR and
perhaps provide selectivity for pcDHFR. Of the syn-
thesized analogues, compounds 5 and 11 showed sig-
nificant selectivity for pcDHFR, and the X-ray crystal
structure of compound 5 with pcDHFR indicates that
the B ring of the naphthyl moiety of 5 does indeed
interact in an “edge on” hydrophobic binding with Phe69
of pcDHFR which translates into selectivity. Compound
5 is the second most selective analogue in the bicyclic
series against pcDHFR with an inhibitory potency and
selectivity better than both epiroprim and TMP for
pcDHFR. Compound 5 is currently undergoing studies
in animal models of P. carinii pneumonia, the results
of which will be reported in a future communication.
2,4-Dia m in o-5-[(1-n a p h th ylth io)m eth yl]fu r o[2,3-d ]p y-
r im id in e (4). To a 50 mL three-neck round-bottom flask were
added sequentially anhydrous DMSO (8 mL), 1-naphthale-
nethiol (0.89 g, 5.55 mmol), purified 2 (0.63 g, 4.63 mmol), and
anhydrous potassium carbonate (0.92 g, 6.67 mmol). The
reaction mixture was stirred at room temperature for 2 h and
then quenched with distilled water (16 mL). The resultant
suspension was stirred at room temperature for 4 h, cooled to
5 °C, and filtered. The solid was washed with cold water (2 ×
5 mL) and then stirred in ether (20 mL) for 8 h. The
suspension obtained was filtered and the residue dissolved in
acetic acid (40 mL). The solution was mixed with charcoal
(0.1 g) and stirred at 40 °C for 15 min. The mixture was
filtered and the filtrate evaporated to dryness. The residue
was washed thoroughly with MeOH to yield 0.44 g (29%) of 4
as a light yellow solid which was homogeneous on TLC: mp
215-218 °C (dec); TLC Rf 0.53 (CHCl3/MeOH, 7:1, with one
drop of NH4OH, silica gel); 1H NMR (DMSO-d6) δ 4.36 (s, 2H,
8-CH2), 6.04 (s, 2H, 4-NH2), 6.63 (s, 2H, 2-NH2), 7.02 (s, 1H,
6-CH), 7.50-8.30 (m, 7H, C10H7). Anal. (C17H14N4OS‚CH3-
COOH) C, H, N, S.
2,4-Dia m in o-5-[(2-n a p h th ylth io)m eth yl]fu r o[2,3-d ]p y-
r im id in e (5). To a 50 mL three-neck round-bottom flask were
added sequentially anhydrous DMSO (5 mL), 2-naphthale-
nethiol (1.12 g, 7.0 mmol), crude 2 (1.39 g, 6.98 mmol), and
anhydrous potassium carbonate (0.97 g, 7.0 mmol). The
reaction mixture was stirred at room temperature for 2 h and
then quenched with distilled water (50 mL). The resultant
suspension was stirred at room temperature for 8 h, cooled to
5 °C, and filtered. The solid was washed with cold water (2 ×
5 mL) and then extracted several times with MeOH (100 mL).
Extracts with a major spot of the product on TLC were pooled
and mixed with silica gel (2 g). The mixture was evaporated
under reduced pressure and then dried in vacuo. The residue
was ground into a fine powder and poured on top of a dry
column of silica gel (40 g). The column was washed with
MeOH in CHCl3: 200 mL (0.5%), 100 mL (1%), 100 mL (2%).
The column was then eluted with 3% MeOH in CHCl3.
Fractions corresponding to a single spot of the product (TLC)
were pooled and evaporated to afford 0.11 g of the pure
product. Fractions with a major spot of the product (TLC) and
a faint spot just above the product spot were also pooled and
evaporated. The residue was washed thoroughly with ether
to afford an additional 0.05 g of the pure product. Fractions
with a major spot of the product and a faint spot just below
the product spot were also pooled and evaporated. The residue
was washed thoroughly with MeOH to afford an additional
0.06 g of the pure product (combined yield 10% for two steps)
5 as a light pinkish solid: mp 233-235 °C (dec); TLC, Rf 0.49
(CHCl3/MeOH, 7:1, with one drop of NH4OH, silica gel); 1H
NMR (DMSO-d6) δ 4.41 (s, 2H, 8-CH2), 6.05 (s, 2H, 4-NH2),
6.62 (s, 2H, 2-NH2), 7.21 (s, 1H, 6-CH), 7.40-7.55 (m, 3H,
C10H7), 7.70-7.90 (m, 4H, C10H7). Anal. (C17H14N4OS‚0.25H2O)
C, H, N, S.
2,4-Dia m in o-5-[(1-n a p h th yla m in o)m eth yl]fu r o[2,3-d ]-
p yr im id in e (6). To a 50 mL three-neck round-bottom flask
were added sequentially anhydrous DMSO (5 mL), 1-amino-
naphthalene (0.66 g, 4.64 mmol), purified 2 (0.50 g, 2.51 mmol),
and anhydrous potassium carbonate (0.69 g, 5.00 mmol). The
reaction mixture was stirred at 50 °C for 8 h under nitrogen
in the dark. The reaction was then quenched with distilled
water (15 mL), and the resultant suspension was stirred at
room temperature for 4 h, cooled to 5 °C, and filtered. The
residue was washed with cold water (2 × 5 mL) and then
dissolved in hot MeOH, and the solution was mixed with silica
gel (2 g). The mixture was evaporated under reduced pressure
and then dried in vacuo. The residue was ground into a fine
powder and poured on top of a dry column of silica gel (40 g).
The column was washed with MeOH in CHCl3: 200 mL (0.5%),
Exp er im en ta l Section
Melting points were determined on a Fisher-J ohns melting
point apparatus or
a Mel-Temp II and are uncorrected.
Nuclear magnetic resonance spectra for proton (1H NMR) were
recorded on a Bruker WH-300 (300 MHz). The chemical shift
values are expressed in δ values (parts per million) relative
to tetramethylsilane as an internal standard: s ) singlet, d
) doublet, t ) triplet, m ) multiplet, and dd ) doublet of
doublet. Thin-layer chromatography (TLC) was performed on
silica gel plates with fluorescent indicator that was visualized
with light at 254 or 366 nm. Proportions of solvents used for
TLC and column chromatography are by volume. Column
chromatography was performed on 230-400 mesh silica gel
purchased from Aldrich, Milwaukee, WI. Solvents routinely
used for reactions and purification were purchased from
Aldrich, Milwaukee, WI, or Fisher Scientific, Pittsburgh, PA
and were used as obtained. Samples for microanalysis were
dried in vacuo over phosphorus pentoxide at room tempera-
ture. Elemental analyses were performed by Atlantic Micro-
lab, Inc., Norcross, GA. Element compositions are within
(0.4% of the calculated values. Fractional moles of water or
organic solvents frequently found in some analytical samples
of antifolates were not prevented despite drying in vacuo and
were confirmed by their presence in the 1H NMR spectrum.
2,4-Dia m in o-5-[(p h en ylth io)m eth yl]fu r o[2,3-d ]p yr im i-
d in e (3). To a 100 mL three-neck round-bottom flask were
added sequentially anhydrous DMSO (15 mL), thiophenol (0.85
g, 7.50 mmol), crude 2 (0.75 g, 3.75 mmol), and anhydrous
potassium carbonate (1.03 g, 7.50 mmol). The reaction mix-
ture was stirred at room temperature for 2 h and then
quenched with water (150 mL). The resultant suspension was
stirred at room temperature for 10 h and filtered. The solid
was extracted twice with MeOH (150 mL + 100 mL), and the
extracts were mixed with silica gel (2 g). The mixture was
evaporated under reduced pressure and dried in vacuo. The
residue was powdered and poured on top of a dry column of
silica gel (40 g). The column was washed with MeOH in
CHCl3: 200 mL (0.5%), 100 mL (1%), 100 mL (2%). The
column was then eluted with 3% MeOH in CHCl3. Fractions
corresponding to a single spot (TLC) of the product were pooled
and evaporated. The residue was washed with ether and air-
dried to afford 0.17 g (16% for two steps) of 3 as a light pinkish

1268 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 8
Gangjee et al.
100 mL (1%), 100 mL (2%). The column was then eluted with
3% MeOH in CHCl3. Fractions corresponding to a single spot
of the product (TLC) were pooled and evaporated. The residue
was washed with ether and air-dried to afford 0.25 g (33%) of
6 as a light yellow solid: mp 230 °C (dec); TLC Rf 0.48 (CHCl3/
MeOH, 7:1, with one drop of NH4OH, silica gel); 1H NMR
(DMSO-d6) δ 4.41 (d, 2H, 8-CH2), 6.00 (s, 2H, 4-NH2), 6.60-
6.74 (m, 3H, 2-NH2 and 9-NH), 6.75 (d, 1H, 4′-CH, J ) 7.4
Hz), 7.20 (d, 1H, 2′-CH), 7.28 (m, 1H, 3′-CH), 7.40-7.46 (m,
3H, 6, 6′,7′-CH), 7.78 (d, 1H, 5′-CH) 8.17 (d, 1H, 8′-CH, J )
9.2 Hz). Anal. (C17H15N5O‚0.20H2O) C, H, N.
CH), 7.01 (t, 1H, 4′′-CH, J ) 7.3 Hz), 7.31 (m, 2H, 3′′,5′′-CH),
7.38 (s, 1H, 6-CH). Anal. (C19H17N5O2‚0.20H2O) C, H, N.
2,4-Dia m in o-5-[(2′-p h en yla n ilin o)m et h yl]fu r o[2,3-d ]-
p yr im id in e (11). To a 50 mL three-neck round-bottom flask
containing purified 2 (0.30 g, 1.51 mmol), 2-aminobiphenyl
(0.38 g, 2.26 mmol), and K2CO3 (0.25 g, 1.812 mmol) was added
anhydrous DMSO (3 mL), and the reaction mixture was stirred
for 8 h at 45 °C under nitrogen in the dark. Distilled water
(10 mL) was added to precipitate the product, and the
suspension was stirred for 2 h, cooled to 5 °C, and filtered.
The solid was washed with cold water (2 × 2 mL) and
suspended in MeOH (200 mL). Silica gel (1 g) was added, and
the suspension was sonicated and heated to 50 °C for 15 min,
evaporated under reduced pressure, and dried in vacuo for 10
h. The residue was powdered and poured on top of a wet
column of silica gel (30 g) and eluted with 3% of MeOH in
CHCl3. Fractions corresponding to the product spot (TLC)
were pooled and evaporated. The residue was washed with
ether and air-dried to afford 0.24 g (48%) of 11 as a light yel-
low solid. Recrystallization from MeOH afforded the analyti-
cally pure compound 11: mp 250-260 °C (dec); TLC Rf 0.55
(CHCl3/MeOH, 7:1, with one drop of NH4OH, silica gel); 1H
NMR (DMSO-d6) δ 4.25 (d, 2 H, 8-CH2), 5.10 (t, 1H, 9-NH),
5.59 (s, 2H, 4-NH2), 6.61 (s, 2H, 2-NH2), 6.72 (m, 1H, 4′-CH),
6.87 (d, 1H, 6′-CH), 6.98 (d, 1H, 3′-CH), 7.15 (m, 1H, 5′-CH),
7.30-7.50 (m, 6H, 2′-phenyl and 6-CH). Anal. (C19H17N5O)
C, H, N.
2,4-Dia m in o-5-[(N-m et h yl-2-n a p h t h yla m in o)m et h yl]-
fu r o[2,3-d ]p yr im id in e (12). To a suspension of 7 (0.10 g,
0.32 mmol), 30% (w/w) aqueous solution of formaldehyde (0.1
mL), and NaCNBH3 (0.063 g, 1 mmol) in acetonitrile (10 mL)
was added dropwise concentrated hydrochloric acid until the
suspension dissolved. The solution was stirred at room tem-
perature for 1 h. The solvent was evaporated under reduced
pressure and the residue dissolved in a minimum amount of
distilled water and the pH of the solution adjusted to 7 with
concentrated NH4OH. The resulting suspension was soni-
cated, cooled to 5 °C, and filtered. The residue was washed
with cold water (2 × 3 mL), stirred in MeOH (5 mL) for 12 h,
and filtered. The residue was washed with MeOH (2 × 3
mL) to afford 0.065 g (64%) of 12 as a light yellow solid: mp
234-240 °C (dec); TLC Rf 0.50 (CHCl3/MeOH, 7:1, with one
drop of NH4OH, silica gel); 1H NMR (DMSO-d6) δ 2.95 (s, 3H,
N-CH3), 4.54 (s, 2H, 8-CH2), 6.10 (s, 2H, 2-NH2), 6.60 (s, 2H,
4-NH2), 7.39 (s, 1H, 6-CH), 7.20-7.80 (m, 7H, C10H7). Anal.
(C18H17N5O‚0.20H2O) C, H, N.
2,4-Dia m in o-5-[(2-n a p h th yla m in o)m eth yl]fu r o[2,3-d ]-
p yr im id in e (7). Compound 7 was prepared from anhydrous
DMSO (4 mL), 2-aminonaphthalene (0.54 g, 3.76 mmol),
purified 2 (0.61 g, 3.09 mmol), and anhydrous potassium
carbonate (0.55 g, 4.00 mmol) as described for compound 6 to
afford 0.27 g (29%) of 7 as a white solid: mp 260-265 °C (dec);
TLC Rf 0.52 (CHCl3/MeOH, 7:1, with one drop of NH4OH, silica
1
gel); H NMR (DMSO-d6) δ 4.32 (d, 2H, 8-CH2, J ) 4.9 Hz),
6.04 (s, 2H, 4-NH2), 6.38 (t, 1H, 9-NH), 6.55 (s, 2H, 2-NH2),
6.98 (d, 1H, 1′-CH, J ) 1.5 Hz), 7.08 (dd, 1H, 3′-CH, J ) 8.8,
2.1 Hz), 7.16 (m, 1H, 6′- or 7′-CH), 7.33 (m, 1H, 6′- or 7′-CH),
7.43 (s, 1H, 6-CH), 7.55-7.70 (m, 3H, 4′,5′,8′-CH). Anal.
(C17H15N5O‚0.33H2O) C, H, N.
2,4-Dia m in o-5-[(2-n a p h t h oxy)m et h yl]fu r o[2,3-d ]p yr i-
m id in e (8). To a 50 mL three-neck round-bottom flask were
added sequentially anhydrous DMSO (4 mL), 2-naphthol (0.27
g, 1.89 mmol), purified 2 (0.25 g, 1.26 mmol), and anhydrous
potassium carbonate (0.26 g, 1.89 mmol). The reaction mix-
ture was stirred at 45 °C for 24 h under nitrogen in the dark
and then quenched with distilled water (12 mL). The resultant
suspension was stirred at room temperature for 4 h, cooled to
5 °C, and filtered. The solid was washed with cold water (2 ×
5 mL) and then dissolved in MeOH. The solution was mixed
with silica gel (1 g), and the mixture was evaporated under
reduced pressure and dried in vacuo. The residue was ground
into a fine powder and poured on top of a wet column of silica
gel (30 g) and eluted with 3% MeOH in CHCl3. Fractions
corresponding to a single spot of the product (TLC) were pooled
and evaporated to afford 0.09 g of 8 (23%) as a white solid:
mp 245-247 °C; TLC Rf 0.62 (CHCl3/MeOH, 7:1, with one drop
of NH4OH, silica gel); 1H NMR (DMSO-d6) δ 5.30 (s, 2H,
8-CH2), 6.11 (s, 2H, 4-NH2), 6.47 (s, 2H, 2-NH2), 7.23 (dd, 1H,
3′-CH), 7.36 (m, 1H, 6′- or 7′-CH), 7.48 (t, 1H, 6′- or 7′-CH),
7.51 (s, 1H, 6- or 1′-CH), 7.55 (s, 1H, 6- or 1′-CH), 7.70-7.90
(m, 3H, 4′,5′,8′-CH). Anal. (C17H14N4O2‚0.20H2O) C, H, N.
2,4-Dia m in o-5-[(2′,5′-d ich lor oa n ilin o)m eth yl]fu r o[2,3-
d ]p yr im id in e (13). To a 100 mL three-neck round-bottom
flask was added in sequence anhydrous DMSO (30 mL), 2,5-
dichloroaniline (2.43 g, 15 mmol), crude 2 (1.49 g, 7.5 mmol),
and anhydrous potassium carbonate (2.07 g, 15 mmol). The
reaction mixture was stirred at 45 °C for 8 h and then
quenched with distilled water (150 mL). The resultant
suspension was stirred at room temperature for 8 h and
filtered. The solid was extracted with MeOH (100 mL) several
times, the extracts which gave a major spot of the product on
TLC were pooled and mixed with silica gel (2 g), and the
mixture was evaporated under reduced pressure and then
dried in vacuo. The residue was ground into a fine powder
and poured on top of a dry column of silica gel (40 g). The
column was washed with MeOH in CHCl3: 200 mL (0.5%),
100 mL (1%), 100 mL (2%). The column was then eluted with
3% MeOH in CHCl3. Fractions corresponding to a single spot
of the product (TLC) were pooled and evaporated. The residue
obtained was washed with ether and air-dried to afford 0.067
g (3% for two steps) of 13 as a light yellow solid: mp 258 °C
(dec); TLC Rf 0.49 (CHCl3/MeOH, 7:1, with one drop of NH4-
OH, silica gel); 1H NMR (DMSO-d6) δ 4.39 (d, 2H, 8-CH2, J )
5.2 Hz), 6.02 (s, 2H, 4-NH2), 6.27 (t, 1H, 9-NH), 6.62 (s, 2H,
2-NH2), 6.64 (dd, 1H, 4′-CH, J ) 8.4 Hz, 2.0 Hz), 6.84 (d, 1H,
6′-CH, J ) 2.0 Hz), 7.27 (d, 1H, 3′-CH, J ) 8.4 Hz), 7.39 (s,
1H, 6-CH). Anal. (C13H11N5OCl2‚0.25H2O) C, H, N, Cl.
2,4-Dia m in o-5-[(2′-p h en oxya n ilin o)m eth yl]fu r o[2,3-d ]-
p yr im id in e (9). Compound 9 was prepared from anhydrous
DMSO (4 mL), 2-phenoxyaniline (0.35 g, 1.89 mmol), purified
2 (0.25 g, 1.26 mmol), and anhydrous potassium carbonate
(0.21 g, 1.51 mmol) as described for compound 8, and fractions
corresponding to a single spot (TLC) of the product were pooled
and evaporated. The residue was washed with ether and air-
dried to afford 0.16 g (37%) of 9 as a white solid: mp 245-
250 °C (dec); TLC Rf 0.58 (CHCl3/MeOH, 7:1, with one drop of
NH4OH, silica gel); 1H NMR (DMSO-d6) δ 4.30 (d, 2H, 8-CH2,
J ) 5.3 Hz), 5.79 (t, 1H, 9-NH, J ) 5.3 Hz), 5.98 (s, 2H, 4-NH2),
6.52 (s, 2H, 2-NH2), 6.61 (m, 1H, 4′-CH), 6.75 (dd, 1H, 6′-CH,
J ) 7.9, 1.2 Hz), 6.85-7.00 (m, 4H, 3′,5′,2′′,6′′-CH), 7.08 (t,
1H, 4′′-CH, J ) 7.2 Hz), 7.31 (s, 1H, 6-CH), 7.34 (m, 2H, 3′′,5′′-
CH). Anal. (C19H17N5O2‚0.25H2O) C, H, N.
2,4-Dia m in o-5-[(4′-p h en oxya n ilin o)m eth yl]fu r o[2,3-d ]-
p yr im id in e (10). Using a method similar to that of 9, 0.06 g
of chromatographically pure 10 was obtained as a white solid.
Fractions corresponding to a major spot of the product (TLC)
and a spot of some unknown impurity just below were pooled
and evaporated. The residue was washed with MeOH to afford
an additional 0.12 g of the product which was homogeneous
on TLC (total yield 41%): mp 245-250 °C (dec); TLC Rf 0.50
(CHCl3/MeOH, 7:1, with one drop of NH4OH, silica gel); 1H
NMR (DMSO-d6) δ 4.18 (d, 2H, 8-CH2, J ) 5.5 Hz), 6.04 (s,
2H, 4-NH2), 6.06 (t, 1H, 9-NH, J ) 5.5 Hz), 6.64 (s, 2H, 2-NH2),
6.80 (d, 2H, 2′,6′-CH, J ) 8.9 Hz), 6.85-6.95 (m, 4H, 3′,5′,2′′,6′′-
2,4-Diam in o-5-[(N-m eth yl-3′,4′-dich lor oan ilin o)m eth yl]-
fu r o[2,3-d ]p yr im id in e (14). Compound 14 was prepared

P. carinii Dihydrofolate Reductase Inhibitors
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 8 1269
from anhydrous DMSO (30 mL), N-methyl-3,4-dichloroaniline
(2.64 g, 15.00 mmol), crude 2 (1.49 g, 7.50 mmol), and
anhydrous potassium carbonate (2.07 g,15.00 mmol), as de-
scribed for compound 13 to afford 0.08 g (3% for two steps) of
14 as a white solid: mp 185-190 °C (dec); TLC Rf 0.49 (CHCl3/
MeOH, 7:1, with one drop of NH4OH, silica gel); 1H NMR
(DMSO-d6) δ 2.93 (s, 3H, 9-NCH3), 4.56 (s, 2H, 8-CH2), 6.07
(s, 2H, 4-NH2), 6.42 (s, 2H, 2-NH2), 6.82 (dd, 1H, 6′-CH, J )
9.0 Hz, 2.9 Hz), 7.02 (d, 1H, 2′-CH, J ) 2.9 Hz), 7.06 (s, 1H,
6-CH), 7.38 (d, 1H, 5′-CH, J ) 9.0 Hz). Anal. (C14H13N5OCl2)
C, H, N, Cl.
mL). The organic layer was dried (MgSO4) and evaporated
under reduced pressure. The residue was dissolved in a
minimum amount of ethyl acetate/hexane, 1:8. The solution
was placed on a wet column of silica gel (80 g) and eluted with
ethyl acetate/hexane (300 mL (1:8), 500 mL (1:5)) to afford 0.51
g of N-methyl-2,5-dimethoxyaniline 17a (31%): TLC Rf 0.44
(hexanes/EtOAc, 3:1, silica gel); 1H NMR (CDCl3) δ 2.67 (d,
3H, N-CH3, J ) 5.2 Hz), 3.65 (s, 3H, 2-OCH3), 3.69 (s, 3H,
5-OCH3), 5.04 (q, 1H, NH, J ) 5.2 Hz), 6.02 (s, 1H, 6-CH),
6.03 (m, 1H, 4-CH), 6.65 (d, 1H, 3-CH). To a 50 mL three-
neck round-bottom flask with 2 (0.32 g, 1.61 mmol), N-methyl-
2,5-dimethoxyaniline 17a (0.40 g, 2.42 mmol), and anhydrous
potassium carbonate (0.27 g, 1.93 mmol) was added anhydrous
DMSO (3 mL), and the reaction mixture was stirred at 45 °C
for 8 h under nitrogen in the dark. Distilled water (10 mL)
was added to precipitate the product, and the suspension was
stirred for 2 h, then cooled to 5 °C, and filtered. The solid
was washed with cold water (2 × 2 mL) and suspended in
MeOH (200 mL). Silica gel (1 g) was added, and the suspen-
sion was sonicated and then heated to 50 °C for 15 min. The
suspension was then evaporated under reduced pressure and
dried in vacuo at room temperature for 10 h. The residue was
powdered, poured on top of a wet column of silica gel (40 g),
and eluted with 5% of MeOH in CHCl3. Fractions correspond-
ing to the product spot (TLC) were pooled and evaporated. The
residue was washed with ether and air-dried to afford 0.15 g
(28%) of 17 as a white solid: mp 238-241 °C (dec); TLC Rf
0.54 (CHCl3/MeOH, 7:1, with one drop of NH4OH, silica gel);
1H NMR (DMSO-d6) δ 2.49 (s, 3H, N-CH3), 3.70 (s, 3H, 2- or
5-OCH3), 3.78 (s, 3H, 2- or 5-OCH3), 3.95 (s, 2H, 8-CH2), 6.01
(s, 2H, 4-NH2), 6.61 (dd, 1H, 4′-CH, J ) 2.9, 8.8 Hz), 6.72 (d,
1H, 6′-CH, J ) 2.9 Hz), 6.49 (d, 1H, 3′-CH, J ) 8.8 Hz), 7.20
(bs, 2H, 2-NH2), 7.36 (s, 1H, 6-CH). Anal. (C16H19N5O3) C,
H, N.
Biologica l Eva lu a tion . The P. carinii DHFR used for
these studies was recombinant enzyme produced from a clone
with a sequence identical to that reported by Edman et al.9
The properties of this enzyme were identical to those reported
for the native enzyme.8 Purification of the recombinant
enzyme was with a folte-Sepharose affinity column, which
yielded pure, active enzyme without refolding. DHFR assays
were carried out at 37 °C in the presence of 90 µM dihydrofolic
acid as the substrate and 119 µM NADPH as the cofactor.
These concentrations saturated the enzyme and gave maximal
activity. Assays for rlDHFR and tgDHFR also contained 150
mM KCl. Testing of DHFR inhibitors against P.carinii in
culture was in a model that has been predictive of the activity
of standard agents such as trimethoprim, pyrimethamine, and
trimetrexate.19
Str u ctu r e Deter m in a tion a n d Refin em en t. Crystals
were grown using the sitting drop vapor diffusion method. The
protein solution of pcDHFR was prepared in 0.1 M imidazole
HCl buffer, pH 7.5, incubated overnight at 4 °C with NADPH,
followed by addition of the inhibitor.5 Protein droplets (5 mL)
containing 3.6 mg/mL pcDHFR in 24% (w/v) PEG 8K, 0.1 M
imidazole HCl buffer, pH 7.5, were utilized for crystal growth.
Crystals grew in 3 weeks time. Crystals of recombinant
pcDHFR ternary complexes are monoclinic, space group P21,
and diffract to 2.5 Å resolution. The lattice constants for the
5-NADPH-pcDHFR ternary complex are a ) 37.552 Å, b )
43.256 Å, c ) 61.389 Å, and â ) 94.97°. The data for the
complex refined to R ) 16.1% for data to 2.5 Å resolution
(Table 2).
The structure was solved by molecular replacement methods
using the coordinates of pcDHFR from the trimethoprim
complex.10 Inspection of the resulting difference electron
density map, using the program CHAIN20 running on a Silicon
Graphics Elan workstation, revealed density for a ternary
complex. However, there was an unusually broad electron
density profile near the diaminopyrimidine ring of compound
5 and Glu32 of pcDHFR which suggested the presence of more
than one mode of binding for the furopyrimidine ring of
compound 5 by rotation of the furopyrimidine ring, or the
presence of a second molecule. This disorder did not extend
2,4-Dia m in o-5-[(N -m e t h yl-3′,4′,5′-t r ich lor oa n ilin o)
m eth yl]fu r o[2,3-d ]-p yr im id in e (15). To a 100 mL three-
neck round-bottom flask was added 3,4,5-trichloroaniline (1.96
g, 10.00 mmol), potassium carbonate (1.66 g, 12.00 mmol),
acetonitrile (25 mL), and iodomethane (1.42 g, 10.00 mmol).
The reaction mixture was stirred at 40 °C for 3 days under
nitrogen in the dark. The reaction mixture was then quenched
with ethyl acetate (50 mL) and filtered. The residue was
washed with ethyl acetate (3 × 20 mL). The filtrate and the
washings were combined and washed with water (3 × 30 mL).
The organic layer was dried (anhydrous MgSO4) and evapo-
rated under reduced pressure. The residue was dissolved in
a minimum amount of ethyl acetate/hexane, 1:12, and the
solution was placed on a wet column of silica gel (80 g) and
eluted with ethyl acetate/hexane (300 mL (1:12), 500 mL (1:
8)) to afford 0.88 g of N-methyl-3,4,5-trichloroaniline 15a (42%)
as a white solid. The product was used directly for the
condensation without characterization. To a 100 mL three-
neck round-bottom flask was added anhydrous DMSO (5 mL),
N-methyl-3,4,5-trichloroaniline (0.88 g, 4.20 mmol), crude 2
(0.20 g, 1.00 mmol), and anhydrous potassium carbonate (0.56
g, 4.09 mmol). The reaction mixture was stirred at 45 °C for
24 h and then quenched with distilled water (15 mL). The
resulting suspension was stirred at room temperature for 4 h
and filtered. The residue was extracted with MeOH (100 mL)
several times, and the extracts which showed a major spot of
the product on TLC were pooled and mixed with silica gel (2
g). The mixture was evaporated under reduced pressure and
dried in vacuo. The residue was ground into a fine powder
and poured on top of a dry column of silica gel (40 g). The
column was eluted with MeOH in CHCl3: 200 mL (0.5%), 100
mL (1%), 100 mL (2%). The column was then eluted with 3%
MeOH in CHCl3. Fractions corresponding to a single spot of
the product were pooled and evaporated. The residue was
washed with ether and air-dried to afford 0.06 g (16% for two
steps) of 15 as a light yellow solid: mp 280 °C (dec); TLC Rf
0.55 (CHCl3/MeOH, 7:1, with one drop of NH4OH, silica gel);
1H NMR (DMSO-d6) δ 2.97 (s, 3H, 9-NCH3), 4.64 (s, 2H,
8-CH2), 6.07 (s, 2H, 4-NH2), 6.39 (s, 2H, 2-NH2), 7.00 (s, 2H,
2′, 6′-CH), 7.04 (s, 1H, 6-CH). Anal. (C14H12N5OCl3‚0.50H2O)
C, H, N, Cl.
2,4-Dia m in o-5-[(3′-m eth oxya n ilin o)m eth yl]fu r o[2,3-d ]-
p yr im id in e (16). Compound 16 was prepared from anhy-
drous DMSO (15 mL), 3-methoxyaniline (0.93 g, 7.50 mmol),
crude 2 (0.75 g, 3.75 mmol), and anhydrous potassium carbon-
ate (1.04 g, 7.50 mmol) as described for compound 13 to afford
0.12 g (11% for the two steps) of 16 as a white solid: mp 225
°C (dec); TLC Rf 0.46 (CHCl3/MeOH, 7:1, with one drop of
NH4OH, silica gel); 1H NMR (DMSO-d6) δ 4.18 (d, 2H, 8-CH2,
J ) 4.8 Hz), 6.04 (s, 2H, 4-NH2), 6.09 (t, 1H, 9-NH, J ) 4.8
Hz), 6.22 (d, 1H, 6′-CH), 6.31 (s, 1H, 2′-CH), 6.34 (d, 1H, 4′-
CH), 6.56 (s, 2H, 2-NH2), 7.00 (m, 1H, 5′-CH), 7.36 (s, 1H,
6-CH). Anal. (C14H15N5O2) C, H, N.
2,4-D ia m in o -5-[(N -m e t h y l-2′,5′-d im e t h o x y a n ilin o )
m eth yl]fu r o[2,3-d ]p yr im id in e (17). To a 100 mL three-
neck round-bottom flask were added 2,5-dimethoxyaniline
(1.53 g, 10.00 mmol), anhydrous potassium carbonate (1.66 g,
12.00 mmol), acetonitrile (25 mL), and iodomethane (1.42 g,
10.00 mmol). The reaction mixture was stirred at 40 °C for 3
days under nitrogen in the dark. The reaction mixture was
quenched with ethyl acetate (50 mL) and filtered. The residue
was washed with ethyl acetate (3 × 20 mL). The filtrate and
the washings were combined and washed with water (3 × 30

1270 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 8
Gangjee et al.
Ta ble 2. Refinement Statistics for MTOS-NADPH-pcDHFR
Medical Sciences [GM 40998 (A.G.) and GM 51670
(V.C.)] and by the National Institute of Allergy and
Infectious Diseases [AI 41743 (A.G.)] and by NIH
contract NO1-AI-35171 (S.F.Q.).
Complex
lattice constants, Å
37.55, 43.26, 61.39,
â ) 94.9
P21
space group
Resolution range, Å
reflections collected
reflections used (8-2.5 Å; 2σ)
R factor, %
8.0-2.5
7117
3220
16.1
1678
Refer en ces
(1) (a) Taken in part from the thesis submitted by X.G. to the
Graduate School of Pharmaceutical Sciences, Duquesne Uni-
versity, in partial fulfilment of the requirements for the degree
of Master of Science, August 1995. (b) Presented in part at the
Eleventh International Symposium Chemistry and Biology of
Pteridines and Folates, Berchtesgaden, Germany, J une 15-20,
1997; Abstr. 17 and 36.
(2) Masur, H. Problems in the Management of Opportunistic
Infections in Patients Infected With Human Immunodeficiency
Virus. J . Infect. Dis. 1990, 161, 858-864.
(3) Masur, H.; Polis, M. A.; Tuazon, C. U.; Ogata-Arakaki, D.;
Kovacs, J . A.; Katz, D.; Hilt, D.; Simmons, T.; Feuerstein, I.;
Lundgren, B.; Lane, H. C.; Chabner, B. A.; Allegra, C. J . Trial
of Trimetrexate-Leucovorin for the Treatment of Cerebral Toxo-
plasmosis. J . Infect. Dis. 1992, 167, 1422-1426.
(4) Fallon, J .; Masur, H. Infectious Complications of HIV. In AIDS
Etiology, Diagnosis, Treatment and Prevention, 3rd ed.; Devita,
V. T., Hellman, S., Rosenberg, S. A., Eds.; Lippincott Co.:
Philadelphia, 1992; pp 157-167.
(5) Gangjee, A.; Devraj, R.; McGuire, J . J .; Kisliuk, R. L.; Queener,
S. F.; Barrows, L. R. Classical and Nonclassical Furo[2,3-d]-
pyrimidines as Novel Antifolates: Synthesis and Biological
Activities. J . Med. Chem. 1994, 37, 1169-1176.
(6) Gangjee, A.; Mavandadi, F. M.; 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.
(7) Gangjee, A.; Elzein, E.; Kothare, M.; Vasudevan, A. Classical
and Nonclassical Antifolates as Potential Antitumor, Antipneu-
mocystis and Antitoxoplasma Agents. Curr. Pharm. Design 1996,
2, 263-280.
(8) Broughton, M. C.; Queener, S. F. Pneumocystis carinii Dihydro-
folate Reductase Used to Screen Potential Antipneumocystis
Drugs. Antimicrob. Agents Chemother. 1991, 35, 1348-1355.
(9) Edman, J . C.; Edman, U.; Cao, M.; Lundgren, B.; Kovacs, J . A.;
Santi, D. V. Isolation and Expression of the Pneumocystis carinii
Dihydrofolate Reductase Gene. Proc. Natl. Acad. Sci. U.S.A.
1989, 86, 8625-8629.
(10) Champness, J . N.; Achari, A.; Ballantine, S. P.; Bryant, P. K.;
Delves, C. J .; Stammers, D. K. The structure of Pneumocystis
carinii Dihydrofolate Reductase to 1.9 Å Resolution. Structure
1994, 2, 915-924.
(11) Tripos Inc., 1699 South Handley Rd, Suite 303, St. Louis, MO
63144.
(12) Secrist, J . A., III.; Liu, P. S. Studies Directed Toward a Total
Synthesis of Nucleoside Q. The Annulation of 2,4-diaminopyri-
midin-4-one with R-Halo Carbonyls to Form Pyrrolo[2,3-d]-
pyrimidines and Furo[2,3-d]pyrimidines. J . Org. Chem. 1978,
43, 3937-3941.
(13) Rosowsky, A. Chemistry and Biological Activity of Antifolates.
Progr. Med. Chem. 1989, 26, 1-253.
(14) Cody, V.; Galitsky, N.; Luft, J . R.; Pangborn, W.; Gangjee, A.;
Devraj, R.; Queener, S. F.; Blakley, R. L. Comparison of Ternary
Complexes of Pneumocystis carinii and Wild Type Human
Dihydrofolate Reductase With a Novel Classical Antitumor Furo-
[2,3-d]pyrimidine Antifolate. Acta Crystallogr. D 1997, D53,
638-649.
(15) Cody, V.; Luft, J . R.; Ciszak, E.; Kalman, T. I.; Freisheim, J . H.
Crystal Structure Determination at 2.3 Å of Recombinant
Human Dihydrofolate Reductase Ternary Complex With NAD-
PH and Methotrexate-γ-tetrazole, Anticancer Drug Des. 1992,
7, 483-491.
protein atoms
water molecules
66
19.34
B factor (protein average) Å2
root mean
square σ
target σ
distances, Å
bonds
angles
0.020
0.040
0.050
0.020
0.150
0.014
0.051
0.052
0.010
0.173
planar 1-4
planar groups
chiral volume
nonbonded distances
single torsion
multiple torsion
possible hydrogen bonds
torsion angles, deg
planar
0.500
0.500
0.500
0.225
0.325
0.275
3.0
15.0
15.0
1.8
23.9
23.7
staggered
orthonormal
to the naphthyl group of 5 which had a clearly defined electron
density. After several models were tested for the orientation
of the diaminofuropyrimidine ring, which left the position of
the 2-naphthyl group constant, it was concluded that the extra
residual density could be the result of a folate-like molecule
remaining in the original pcDHFR sample. It was noted that
this particular sample had a yellow color which was not
removed on repeated washing. Such contaminants have been
observed in other protocols for pcDHFR purification.21 In our
hands, this is the first crystal structure that has such a
contaminant, as it was not observed in the other crystals grown
samples purified from the same protocols. The final cycles of
refinement were carried out with the inhibitor 5 and folate
present at half occupancy, and cofactor included using the
program PROLSQ.22 The Ramachandran conformational pa-
rameters from the last cycle of refinement, generated by
PROCHECK,23 show that 84.3% of the residues have the most
favored conformation and none in the disallowed regions.
Further restrained refinement was continued for the ternary
complex, including the cofactor and inhibitor which were
modeled using the builder function of InsightII.24 Between
least-squares minimizations, the structures were manually
adjusted to fit its electron density values and verified by a
series of OMIT maps calculated from the current model with
deleted fragments. The highest thermal parameters (30-50
Å2) of the protein residues are near the binding site entrance
and in flexible loop regions. The remaining side chain thermal
parameters range from 10 to 25 Å2. In the final refinement
stages, 66 water molecules were added in accord with the
criteria of good electron density and acceptable hydrogen bonds
to other atoms. The final refinement statistics are sum-
marized in Table 2.
(16) Chunduru, S. K.; Cody, V.; Luft, J . R.; Pangborn, W.; Appleman,
J . R.; Blakley, R. L. Methotrexate-resistant Variants of Human
Dihydrofolate Reductase: Effects of Phe31 Substitutions, J . Biol.
Chem. 1994, 269, 9547-9555.
The broad electron density profile near Glu32 suggested that
the density represented a composite of more than one model
and further analysis of this profile suggested the presence of
folate which filled this density. After several possible orienta-
tions of the naphthalene inhibitor and folate were tested, the
final model was that of folate and the inhibitor, 5, at half
occupancy. However, the resolution and extent of the current
diffraction data preclude further discrimination of these
models.
(17) 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.
(18) While this manuscript was in preparation a report of significant
pcDHFR selectivity in monocyclic pyrimidine antifolates was
reported. Stevens, M. F. G.; Phillip, K. S.; Rathbone, D. L.;
O’Shea, D. M.; Queener, S. F.; Schwalbe, C. H.; Lambert, P. A.
J . Med. Chem. 1997, 40, 1886-1893.
Ack n ow led gm en t. This work was supported, in
part, by grants from The National Institute of General
(19) Bartlett, M. S.; Shaw, M.; Navaran, P.; Smith, J . W.; Queener,
S. F. Evaluation of Potent Inhibitors of Dihydrofolate Reductase

P. carinii Dihydrofolate Reductase Inhibitors
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 8 1271
in a Culture Model for Growth of Pneumocysitis carinii. Anti-
microb. Agents Chemother. 1995, 39, 2436-2441.
Indian Academy of Sciences: Bangalore, India, 1980; p 13.01.
Finzel, B. C. J . Appl. Crystallogr. 1987, 20, 53.
(23) Laksowski, R. A.; MacArthur, M. W.; Moss, D. S.; Thornton, J .
M. PROCHECK: A Program to Check the Stereochemical Quali-
ty of Protein Structures. J . Appl. Crystallogr. 1993, 26, 283-
291.
(20) Sack, J . S. CHAINsA Crystallographic Modeling Program. J .
Mol. Graph. 1988, 6, 224-225.
(21) Margosiak, S. A.; Appleman, J . R.; Santi, D. V.; Blakley, R. L.
Dihydrofolate Reductase from the Pathogenic Fungus Pneu-
mocystis carinii: Catalytic Properties and Interaction With
Antifolates. Arch. Biochem. Biophys. 1993, 305, 499-508.
(22) Hendrickson, W. A.; Konnert J . H. In Computing in Crystal-
lography; Diamond, R., Ramaseshan, S., Venkatesan. K., Eds.;
(24) BIOSYM Technologies, 9685 Scranton Rd, San Diego, CA
92121-2777.
J M970537W