Structure-based design of selective inhibitors of dihydrofolate reductase: Synthesis and antiparasitic activity of 2,4-diaminopteridine analogues with a bridged diarylamine side chain

Article abstract of DOI:10.1021/jm990331q

As part of a larger search for potent as well as selective inhibitors of dihydrofolate reductase (DHFR) enzymes from opportunistic pathogens found in patients with AIDS and other immune disorders, N-[(2,4-diaminopteridin-6- yl)methyl]dibenz[b,f]azepine (4a) and the corresponding dihydrodibenz[b,f]azepine, dihydroacridine, phenoxazine, phenothiazine, carbazole, and diphenylamine analogues were synthesized from 2,4-diamino-6- (bromomethyl)pteridine in 50-75% yield by reaction with the sodium salts of the amines in dry tetrahydrofuran at room temperature. The products were tested for the ability to inhibit DHFR from Pneumocystis carinii (pcDHFR), Toxoplasma gondii (tgDHFR), Mycobacterium avium (maDHFR), and rat liver (rlDHFR). The member of the series with the best combination of potency and species selectivity was 4a, with IC₅₀ values against the four enzymes of 0.21, 0.043, 0.012, and 4.4μM, respectively. The dihydroacridine, phenothiazine, and carbazole analogues were also potent, but nonselective. Of the compounds tested, 4a was the only one to successfully combine the potency of trimetrexate with the selectivity of trimethoprim. Molecular docking simulations using published 3D structural coordinates for the crystalline ternary complexes of pcDHFR and hDHFR suggested a possible structural interpretation for the binding selectivity of 4a and the lack of selectivity of the other compounds. According to this model, 4a is selective because of a unique propensity of the seven-membered ring in the dibenz[b,f]azepine moiety to adopt a puckered orientation that allows it to fit more comfortably into the active site of the P. carinii enzyme than into the active site of the human enzyme. Compound 4a was also evaluated for the ability to be taken up into, and retard the growth of, P. carinii and T. gondii in culture. The IC₅₀ of 4a against P. carinii trophozoites after 7 days of continuous drug treatment was 1.9μM as compared with previously observed IC₅₀ values of >340μM for trimethoprim and 0.27 μM for trimetrexate. In an assay involving [³H]uracil incorporation into the nuclear DNA of T. gondii tachyzoites as the surrogate endpoint for growth, the IC₅₀ of 4a after 5 h of drug exposure was 0.077 μM. The favorable combination of potency and enzyme selectivity shown by 4a suggests that this novel structure may be an interesting lead for structure-activity optimization.

Full text of DOI:10.1021/jm990331q

J . Med. Chem. 1999, 42, 4853-4860
4853
Str u ctu r e-Ba sed Design of Selective In h ibitor s of Dih yd r ofola te Red u cta se:
Syn th esis a n d An tip a r a sitic Activity of 2,4-Dia m in op ter id in e An a logu es w ith a
Br id ged Dia r yla m in e Sid e Ch a in
Andre Rosowsky,*,† Vivian Cody,‡ Nikolai Galitsky,‡ Hongning Fu,† Andrew T. Papoulis,† and
Sherry F. Queener§
Dana-Farber Cancer Institute and Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School,
Boston, Massachusetts 02115, Hauptman-Woodward Medical Research Institute, Buffalo, New York 14203, and Department of
Pharmacology and Toxicology, Indiana University School of Medicine, Indianapolis, Indiana 46202
Received J une 25, 1999
As part of a larger search for potent as well as selective inhibitors of dihydrofolate reductase
(DHFR) enzymes from opportunistic pathogens found in patients with AIDS and other immune
disorders, N-[(2,4-diaminopteridin-6-yl)methyl]dibenz[b,f]azepine (4a ) and the corresponding
dihydrodibenz[b,f]azepine, dihydroacridine, phenoxazine, phenothiazine, carbazole, and diphe-
nylamine analogues were synthesized from 2,4-diamino-6-(bromomethyl)pteridine in 50-75%
yield by reaction with the sodium salts of the amines in dry tetrahydrofuran at room
temperature. The products were tested for the ability to inhibit DHFR from Pneumocystis carinii
(pcDHFR), Toxoplasma gondii (tgDHFR), Mycobacterium avium (maDHFR), and rat liver
(rlDHFR). The member of the series with the best combination of potency and species selectivity
was 4a , with IC50 values against the four enzymes of 0.21, 0.043, 0.012, and 4.4 µM, respectively.
The dihydroacridine, phenothiazine, and carbazole analogues were also potent, but nonselective.
Of the compounds tested, 4a was the only one to successfully combine the potency of
trimetrexate with the selectivity of trimethoprim. Molecular docking simulations using
published 3D structural coordinates for the crystalline ternary complexes of pcDHFR and
hDHFR suggested a possible structural interpretation for the binding selectivity of 4a and the
lack of selectivity of the other compounds. According to this model, 4a is selective because of
a unique propensity of the seven-membered ring in the dibenz[b,f]azepine moiety to adopt a
puckered orientation that allows it to fit more comfortably into the active site of the P. carinii
enzyme than into the active site of the human enzyme. Compound 4a was also evaluated for
the ability to be taken up into, and retard the growth of, P. carinii and T. gondii in culture.
The IC50 of 4a against P. carinii trophozoites after 7 days of continuous drug treatment was
1.9 µM as compared with previously observed IC50 values of >340 µM for trimethoprim and
0.27 µM for trimetrexate. In an assay involving [3H]uracil incorporation into the nuclear DNA
of T. gondii tachyzoites as the surrogate endpoint for growth, the IC50 of 4a after 5 h of drug
exposure was 0.077 µM. The favorable combination of potency and enzyme selectivity shown
by 4a suggests that this novel structure may be an interesting lead for structure-activity
optimization.
In tr od u ction
phylaxis, is recognized to have a key role in extending
and improving the quality of life of AIDS patients.
A high incidence of potentially life-threatening op-
portunistic infections in patients with AIDS has been a
marker of this disease ever since it was first defined as
a clinical syndrome in the early 1980s. Indeed, clinical
diagnosis of acute AIDS is typically made when an
HIV-1 positive patient presents with a Pneumocystis
carinii or Toxoplasma gondii infection.1 A number of
other opportunistic pathogens which are not ordinarily
a problem in healthy individuals can wreak havoc when
normal immune function has been damaged via sys-
tematic eradication of T-cells by the AIDS virus. Thus,
in addition to the use of reverse transcriptase and
protease inhibitors to keep viral replication in check,2
adjunctive antimicrobial therapy, including chemopro-
Drugs currently prescribed for the therapy and pro-
phylaxis of P. carinii pneumonia fall into several classes,
of which a prominent group consists of lipophilic anti-
folates (reviewed in refs 3a,b). Although other lipophilic
dihydrofolate reductase (DHFR) inhibitors such as
trimetrexate (1)4 and piritrexim (2)5 (Chart 1) have been
tested in controlled clinical settings to treat or prevent
pneumocystis pneumonia in AIDS patients, the one used
most routinely for the treatment of this disease is the
antibacterial agent trimethoprim (TMP, 3).6
Because TMP is not very effective against P. carinii
and other parasitic organisms, a sulfonamide inhibitor
of de novo folate synthesis is generally also included.6
A problem with the use of sulfa drugs in this context,
however, is the occurrence in some patients of cutaneous
side effects which may not respond to corticosteroids and
can be severe enough to require interruption of therapy.7
† Dana-Farber Cancer Institute.
‡ Hauptman-Woodward Medical Research Institute.
§ Indiana University.
10.1021/jm990331q CCC: $18.00 © 1999 American Chemical Society
Published on Web 10/22/1999

4854 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 23
Rosowsky et al.
Ch a r t 1
Sch em e 1
Thus there have been active efforts to develop alterna-
tive lipophilic DHFR inhibitors which are potent enough
to be used without a sulfa drug and yet are selective
enough in their affinity for P. carinii versus mammalian
DHFR to not require leucovorin co-administration to
protect the hematopoietic system of the host.4,5 Although
a large number of lipophilic DHFR inhibitors featuring
the monocyclic or condensed diaminopyrimidine motif
have been synthesized and a number of them have been
quite potent and in some cases fairly selective, an
antifolate to challenge the clinical supremacy of the
trimethoprim/sulfonamide combination remains to be
found.
Synthetic programs directed toward the design of new
lipophilic DHFR inhibitors8-10 have most often followed
an empirical medicinal chemistry approach involving,
for example, replacement of ring nitrogens by carbon
or replacement of a six-membered ring by a five-
membered ring. However recent successes in the clon-
ing, expression, and characterization of P. carinii DHFR
(pcDHFR),11 and ultimately in the elucidation of the 3D
structure of enzyme-inhibitor complexes by X-ray
crystallography,12 have paved the way for more rational
structure-based approaches to the design of inhibi-
tors.13,14 In the present paper we offer an example of
the use of structure-based design to generate a novel
series of DHFR inhibitors with two rotationally re-
stricted phenyl groups on the bridge nitrogen.
Consideration of the known coordinates for the 3D
structures of crystalline pcDHFR and human DHFR
(hDHFR)12 suggests that the hydrophobic cavity defined
by the van der Waals surfaces of the active site is
somewhat larger in the P. carinii enzyme and can
therefore accommodate a ligand with a bulkier side
chain. Molecular docking simulations with ligands
containing a variety of bulky hydrophobic side chains
were carried out using the known crystallographic
coordinates for the ternary complex of pcDHFR with
NADPH and a tight-binding classical inhibitor. An
intriguing hypothesis which sprang from these simula-
tions was that compounds with two phenyl groups on
N10, instead of one as is more typically found in small-
molecule antifolates, might fit better into the active site
of pcDHFR than into that of hDHFR. Moreover, as
discussed in greater detail below, these simulations
suggested that locking the two phenyl rings into a
rotationally restricted configuration by a bridge might
favor selectivity. To test this hypothesis we synthesized
a pilot series of 2,4-diaminopteridines of general struc-
ture 4 (Scheme 1), in which the X bridge would define
the dihedral angle between the phenyl rings. With the
phenyl rings joined directly (X ) direct link) the tricyclic
moiety would be completely planar, whereas a one-atom
(X ) CH2, O, S) or two-atom (X ) CH2CH2, CHdCH)
bridge would lead to a chairlike conformation as dictated
by the least-strain requirement for six- and seven-
membered rings.
Ch em ica l Syn th esis
Facile access to compounds of general structure 4 was
made possibly by a straightforward modification of the
method of Piper and Montgomery.15 As shown in
Scheme 1, alkylation of the nucleophilic nitrogen anion
of dibenz[b,f]azepine (iminostilbene), 9,10-dihydrodiben-
z[b,f]azepine (iminodibenzyl), 9,10-dihydroacridine (ac-
ridan), phenothiazine, phenoxazine, or carbazole with
2,4-diamino-6-(bromomethyl)pteridine hydrobromide (5‚
HBr) in dry THF at room temperature afforded the
bridged N,N-diarylamine derivatives 4a -f, respectively.
The same reaction using N,N-diphenylamine led to the
nonbridged analogue 4g. Because of the low nucleophi-
licity of the diarylamines, it was necessary to first
convert them to the corresponding anions with NaH. A
good solvent for the reaction was found to be THF,
though care had to be taken to dry the solvent and keep
the reaction mixture under a blanket of dry N2 in order
to exclude moisture. The molar ratio of the amine to
the bromide salt 5‚HBr was 2:1, and the NaH was
likewise used in excess (5.4 mol/mol of 5‚HBr). Purifica-
tion was satisfactorily achieved by flash chromatogra-
phy and/or preparative TLC on silica gel. The products
were yellow powders except in the case of the phenox-
azine 4d and the phenothiazine 4e, whose colors had
an orange-yellow and greenish-yellow tinge, respec-
tively. Purified yields ranged from 39% (4b) to 78% (4a )
1
and were most often around 50%. H NMR spectra of
the products, measured in DMSO-d6 solution, displayed
the expected C9 bridge protons as a singlet at ca. 5.3
ppm and the C7 pteridine ring proton as a singlet
between 8.5 and 8.9 ppm. In the case of 4a -c, 4f, and
4g, structural assignments were also corroborated by
mass spectrometric analysis, which revealed the ex-
pected M + 1 peak as the major species.
Bioch em ica l a n d Biologica l Assa ys
The ability of 4a -g to inhibit partly purified pcDHFR
and T. gondii DHFR (tgDHFR), as well as their selec-
tivity for these enzymes versus rat liver DHFR (rlDH-
FR), was determined as previously described.16 As with
many lipophilic DHFR inhibitors, the low aqueous
solubility of these compounds necessitated the use of
DMSO to prepare concentrated stock solutions. Rat liver
enzyme was used because (i) immunosuppressed rats
are typically the primary in vivo model for drugs against
P. carinii;17 (ii) a full sequence for the rat enzyme has
not been reported, but comparative N-terminal analysis
of the purified rat18a and human18b proteins indicates
substantial sequence homology (including especially the

Synthesis, Antiparasitic Activity of Diaminopteridines
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 23 4855
Ta ble 1. Inhibition of DHFR by N-[(2,4-Diaminopteridin-6-yl)methyl]dibenz[b,f]azepine (4a ) and Related Compounds with a Bulky
Bridged N,N-Diarylamine Side Chain
DHFR inhibition (IC50, µM)b
selectivity ratioc
compd
X
potencya
low
selectivitya
high
pc
12
tg
ma
rl
rl/pc
rl/tg
rl/ma
type 1
trimethoprim (3)f
2.7
0.19e
130
11
48
680
type 2
low
low
4b
4d
4g
CH2CH2
O
none
1.4
3.4
4.9
0.91
2.2
1.3
5.1
13
4.0
3.6
3.8
0.82
5.6
5.9
3.1
2.5
3.7
5.2
1.1
type 3
high
low
trimetrexate (1)d
0.042
0.042
0.12
0.01
0.029
0.11
0.003
0.027
0.20
0.07
0.64
1.7
0.30
0.93
1.8
4c
4e
CH2
S
0.017
0.029
1.6
6.9
4f
direct
0.10
0.012
0.055
0.55
4.6
type 4
4a
high
high
CHdCH
0.21e
0.043e
0.012
4.4f
21
102
367
a
Low potency is arbitrarily defined here as an IC50 of >1.0 µM, high potency as an IC50 of <0.1 µM; the difference between low and
high selectivity is arbitrarily set at a selectivity index of 10. Replicate assays performed as previously described.16 c Defined as the
b
ratio IC50(rat)/IC50(P. carinii or T. gondii). Data cited, for example, in ref 10b. e Mean of two sets of titrations done on different days.
d
f Mean of four sets of titrations on different days.
Asp21 residue which distinguishes pcDHFR from mam-
malian DHFRs); and (iii) inhibitors generally bind
similarly to both enzymes.19 Thus, even though a crystal
structure for the rat enzyme is not yet available and
our simulations were based on published coordinates
for ternary complexes of the human protein,12 we
assume that binding affinities of 4a -g for rlDHFR are
a reasonable portrayal of their interaction with hDHFR.
As shown in Table 1, a rather wide range of potency
was observed among compounds 4a -g against both
pcDHFR and tgDHFR. The most potent members of the
series were the dihydroacridine 4c, the phenothiazine
4e, and the carbazole 4f, with IC50 values of 0.01-0.1
µM; the least potent were the dihydrodibenz[b,f]azepine
4b (CH2CH2 bridge), the phenothiazine 4d (O bridge),
and the diphenylamine 4g (no bridge), with IC50 values
of >1.0 µM. The IC50 of 4e (S bridge) was lower than
that of either 4d (O bridge) or 4b (CH2CH2 bridge) but
higher than that of 4c (CH2 bridge). When the effect of
a bridge versus no bridge was considered, potency was
increased by direct covalent linkage of the two phenyl
rings or by insertion of a CH2, S, or CHdCH bridge,
but not by insertion of a CH2CH2 or O bridge.
The pattern of activity of 4a -g against rlDHFR
paralleled their pattern of activity against pcDHFR, but
4a showed a 20-fold preference for the latter enzyme
even though it was less potent than several other
analogues (Table 1). To the extent that it tended to
support the hypothesis that differences in bulk tolerance
within the active site of DHFR from different species
can be exploited in the structure-based design of new
antifolates selective for P. carinii, this finding was very
encouraging. However an even more gratifying outcome
of the studies was that 4a , apart from being moderately
selective for pcDHFR, displayed 100-fold selectivity
against tgDHFR. Moreover, when 4a was tested against
Mycobacterium avium DHFR (maDHFR), the enzyme
of another opportunistic pathogen often found in AIDS
patients and other immunosuppressed individuals, its
IC50 was found to be 0.012 µM. Thus the enzyme
selectivities of 4a for maDHFR and tgDHFR appear to
be of the same order. Two other members of the series,
4c and 4e, likewise showed excellent activity against
maDHFR (Table 1), with IC50 values of 0.017 and 0.029
µM. However because these compounds were also potent
against rlDHFR, they lacked the selectivity of 4a . It is
of interest that the possibility that small-molecule
antifolates which are active against P. carinii and T.
gondii may be useful as drugs against M. avium
complex has recently been suggested in two other
reports.14b,20
Overall, compounds 4a -g can be seen to fall into four
categories which we have arbitrarily defined as type 1
(low potency/high selectivity), type 2 (low potency/low
selectivity), type 3 (high potency/low selectivity), and
type 4 (high potency/high selectivity). Prototypical
examples of the first and third category are trimetho-
prim (3) and trimetrexate (1), respectively. The relative
potencies and selectivities of 3 and 1 against pcDHFR
and rlDHFR were consistent with their previously
reported kinetic inhibition constants (Ki values) against
pcDHFR and hDHFR.11d Interestingly, 4b-g were all
either more potent than 3 or more selective than 1,
whereas only 4a displayed both of these characteristics
and therefore could be said to belong to type 4. Thus an
important goal of future work with these compounds
would be to find more examples of type 4.
Because of its promising combination of potency and
selectivity, 4a was also tested as an inhibitor of the
growth of P. carinii organisms cultured on confluent
layers of human embryonic lung (HEL) fibroblasts in
the presence of10 µM folicacid as previouslydescribed.17a,c
In a pilot experiment in which the cells were continu-
ously exposed to 4a for 7 days at a concentration of 1.0

4856 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 23
Rosowsky et al.
F igu r e 1. Computer-generated model of the sterically disfa-
vored ‘a’ orientation of 4a (pink structure) and 4b (red
structure) in the active site of pcDHFR (green) and hDHFR
(white). The illustrations were made with the program SE-
TOR.24 Note that, in this orientation, while pcDHFR has better
contacts, there are many more unfavorable contacts through
this region of the active site.
F igu r e 2. Computer-generated model of the sterically favored
‘b’ orientation of 4a (cyan structure) and 4b (yellow structure)
in the active site of pcDHFR (green) and hDHFR (white). The
illustrations were made with the program SETOR.24 Note that,
in this orientation, 4a has better contacts to the pcDHFR
active site than to the human enzyme and that the contacts
to both enzymes are more favorable than those of the ‘a’
orientation in Figure 1.
or 10 µM, a >75% decrease in growth relative to
untreated controls was observed, whereas at 0.1 µM the
drug was inactive. As expected from the inclusion of a
high concentration of folic acid in the medium, no
evidence of toxicity to the HEL cells was noted. The IC50
was estimated to be 1.9 µM. Trimethoprim (3) alone has
almost no effect on the growth of P. carinii in culture
even at 340 µM.17a In contrast, the reported IC50 of
trimetrexate (1) is 0.27 µM.3a Thus the potency of 4a in
this assay was intermediate between that of 1 and that
of 3, but closer to that of 1. When the effect of the drug
was assessed via metabolic incorporation of [3H]p-
aminobenzoic acid into the total folate pool of P. carinii
organisms as described by Kovacs and co-workers,21
treatment with 4a at a concentration of 18 µM for 5 h
resulted in a 62% decrease in radioactivity relative to
untreated controls. The fact that much more 4a was
required to inhibit the growth of the intact trophozoites
than to inhibit isolated DHFR suggests that uptake into
the parasite is not very efficient, as has been noted for
other lipophilic DHFR inhibitors.17c
Growth inhibition assays were also performed against
T. gondii tachyzoites in culture as previously described,17c
using an assay based on incorporation of [3H]uracil into
the nuclear DNA of the parasite after a drug exposure
time of 24 h. The IC50 value of 4a in this surrogate assay
of growth was 0.077 µM. The IC50 of 4b was 0.69 µM,
in reasonable agreement with the difference in IC50 that
had been observed between 4a and 4b in the tgDHFR
inhibition assay.
binding nor the ability to inhibit parasitemia in culture
can guarantee that a compound will be efficacious in
vivo. Thus, any future assessment of the clinical poten-
tial of 4a or one of its second-generation analogues
would have to address its mode of administration,
pharmacokinetics, tissue distribution, and other phar-
macologic characteristics. While these issues are clearly
important, they were not considered to fall within the
scope of the present work.
Sim u la ted In ter a ction of In h ibitor s w ith DHF R
Models of the binding of 4a -c, 4e, and 4f to the active
site of pcDHFR were generated by using the molecular
graphics program Sybyl22 to ‘replace’ the tight-binding
inhibitor methotrexate (MTX) in the corresponding
enzyme complexes of known structure,12b as was done
recently to simulate the interaction of another series of
small-molecule inhibitors with pcDHFR.23 A significant
feature of compounds of general structure 4 is that
inversion of the pyramidal ring nitrogen can cause the
tricyclic moiety to adopt different spatial orientations
relative to the pteridine ring. Because the topology of
these limiting orientations is especially important in the
dibenzazepine derivatives 4a and 4b, a stereoscopic
image for one of these orientations (called ‘a’) is dis-
played in Figure 1, whereas the other (called ‘b’) is given
in Figure 2. Key amino acids in the active site of the P.
carinii enzyme are marked in green, and their homolo-
gous residues in the human enzyme are marked in
white. As required by the docking strategy, the pteridine
ring in both orientations of 4a and 4b in the ternary
complex with NADPH is essentially superimposable on
On the basis of the foregoing biochemical and biologi-
cal results, it appears that the dibenzazepine derivative
4a may be an interesting lead for further development.
However it is obvious that neither selective DHFR

Synthesis, Antiparasitic Activity of Diaminopteridines
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 23 4857
the diaminopteridine moiety of MTX, allowing the
2-amino group to interact in the usual manner with the
γ-carboxyl group of human Glu30 and P. carinii Glu32.
Prominent differences in the P. carinii enzyme contrib-
uting to its sterically more permissive active site relative
to the human enzyme are the substitution of Ile33 for
Phe31 and of Ser24 for Asp21. Of particular note is that
the Asp21 side chain in hDHFR points toward the
inhibitor, whereas the Ser24 side chain in pcDHFR is
rotated out of the way, creating an empty cavity into
which the inhibitor can comfortably fit. A difference can
also be seen in the permissive arrangement of Ser64
and Ile65 in pcDHFR relative to Ser59 and Ile60 in
hDHFR. Of greatest interest, however, is that in the ‘a’
orientation the tricyclic moiety of 4a (Figure 1, pink
structure) and 4b (Figure 1, red structure) lies prohibi-
tively close to the serine and isoleucine side chains of
each enzyme, whereas in the ‘b’ orientation (Figure 2;
4a , cyan structure; 4b, yellow structure) the edge rather
than the face of the tricyclic moiety is turned toward
the viewer. It was evident from these models that 4a
and 4b should bind better to the enzyme in orientation
‘b’. This distinctive spatial realignment of the tricyclic
moiety via inversion of the pyramidal ring nitrogen,
analogous to the flapping of butterfly wings, is only
possible when the central ring is seven-membered.
Thus, while inversion of this nitrogen does occur in the
other tricyclic systems (4c-f), as well as in the non-
bridged analogue 4g, only in 4a and 4b does this
inversion bring about the same steric relief.
contact distances for the other compounds (4d -g) does
not reval this species difference. Thus we were led to
speculate that 4a and 4b may show greater species
selectivity than the other rotationally restricted com-
pounds in the series, for which the same opportunity to
gain steric relief via inversion of the pyramidal ring
nitrogen does not seem to exist.
Computer simulation was also used to model the
interaction of the Asp21 side chain of hDHFR with 4a
and the other members of the series. As indicated in
Table 2 (Supporting Information), while the aromatic
carbon ortho to the CHdCH bridge of 4a is 3.1 Å from
the â-carbon of Ser59, this aromatic carbon also lies near
the â-COOH of Asp21, with a predicted contact distance
of 2.7 Å in the ‘a’ orientation and 3.1 Å in the ‘b’
orientation. Thus, in the sterically preferred ‘b’ orienta-
tion, this carbon is equidistant from the â-COOH of
Asp21 and the â-carbon of Ser59. However, because
there is no counterpart for the Asp21 interaction in the
P. carinii enzyme, we feel that this interaction may be
less important than the one involving Ser59 in confer-
ring selectivity.
An important qualification has to be made whenever
docking simulations of small molecules to enzymes or
other complex protein structures are performed, namely
that consideration of a protein as a rigid structure may
lead to an oversimplified view of the system. Thus, a
drawback of such a static model is that it does not take
into account the possibility that the DHFR can undergo
a ligand-induced conformational change of its 3D struc-
ture. To rigorously address this issue, a docking experi-
ment needs to be followed up with a real (as opposed to
‘virtual’) structural analysis of the enzyme-inhibitor
complex. To this end, work is currently in progress to
crystallize and solve the structure of the ternary com-
plex of 4a with pcDHFR. Modeling experiments using
second-generation analogues of 4a would likewise be of
potential interest.
A comparison of the simulated contact distances
between the bound ligand and several key amino acid
residues in the active site of hDHFR (Leu22, Asp21,
Phe31, Ser59, Ile60) and pcDHFR (Leu25, Ile33, Ser64,
Ile65) was carried out for the ‘a’ and ‘b’ orientations of
4a and 4b, as well as for the other analogues in which
nitrogen inversion cannot afford the same degree of
topologic change as it does in the seven-membered ring
structures. A detailed listing of these simulated dis-
tances is given in Table 2 (see Supporting Information).
The Ser24 side chain of pcDHFR was excluded from the
analysis because, in contrast to the Asp21 of the human
enzyme, it faces away from the ligand. Generally
speaking, simulated contact distances of <2.5 Å are
considered to be sterically disfavored, whereas distances
of 2.5-3.0 Å are viewed as borderline, and distances of
3.0-3.5 Å are preferred. Using the ‘a’ and ‘b’ orienta-
tions of 4a as an illustration, a number of the predicted
contact distances turn out to be <2.5 Å in the ‘a’
orientation, whereas in the ‘b’ orientation the number
of such distances is zero. For example the nearest
predicted contact for the aromatic carbons next to the
CHdCH bridge in the ‘a’ orientation of 4a is the
â-carbon of Ser59 in hDHFR (2.1 Å) and Ser64 in
pcDHFR (2.7 Å). The corresponding distances in the ‘b’
orientation are 3.1 and 3.6 Å, respectively. Thus, for the
atoms used in this particular comparison, it can easily
be seen that (i) the ‘a’ orientation is less favorable than
the ‘b’ orientation where both enzymes are concerned,
and (ii) even in the ‘b’ orientation, the P. carinii enzyme
appears to be favored over the human enzyme. Similar
analysis of other predicted distances for 4a and 4b (cf.
Table 2, Supporting Information) leads to the same
general conclusion. In contrast, analysis of predicted
Exp er im en ta l Section
IR spectra were obtained on a Perkin-Elmer model 781
double-beam recording spectrophotometer. Only peaks above
1200 cm-1 are reported; weak peaks and shoulders are omitted.
1H NMR spectra were recorded at 60 MHz on a Varian model
EM360 instrument using Me4Si as the reference or on a 500
MHz Varian VX 500 instrument. When they were seen, NH2
peaks were broad signals at δ 6.6 and 7.6 and could be
eliminated by adding a drop of D2O. Mass spectra in the
electron-impact (EI) or fast-atom bombardment (FAB) mode
were provided by the Molecular Biology Core Facility of the
Dana-Farber Cancer Institute. Reported M + 1 peaks were
major species in each case. Analytical TLC was on Whatman
MK6F silica gel plates, and preparative separations were on
Aldrich TLC plates (fluorescent silica gel, 1000-µm layer, 20
× 20 cm). TLC spots were visualized under a UV lamp at 254
nm. Column chromatography was on Baker 7024 flash silica
gel (40-µm particle size). 2,4-Diamino-6-(bromomethyl)pteri-
dine hydrobromide (5‚HBr) was obtained as the i-PrOH solvate
as reported in the literature15 and for simplicity is referred to
as 5‚HBr. The amount of i-PrOH in each batch was estimated
by microanalysis or examination of the 1H NMR spectrum, and
the amount of material used in each reaction was adjusted to
correct for variations in the estimated amount of i-PrOH in
the sample. 9,10-Dihydroacridine (acridan) was obtained from
acridine in 70% yield by hydrogenation in the presence of
Raney Ni and was recrystallized from MeOH in the form of
colorless needles: mp 169-170 °C (lit.25 mp 170 °C). Other
chemicals were purchased from Aldrich or Fisher; all the
diarylamines were dried in vacuo at 60 °C overnight before

4858 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 23
Rosowsky et al.
use. The THF for the N-alkylation reactions was distilled from
Na and dibenzophenone and was stored over Linde 4A mo-
lecular sieves. Melting points were determined in Pyrex
capillary tubes using a Mel-Temp apparatus (Laboratory
Devices, Inc., Cambridge, MA) and are not corrected. Samples
for microanalysis were dried in a vacuum oven at 70 °C
overnight. Elemental analyses were performed by QTI Labo-
ratories, Whitehouse, NJ , or Robertson Microlit Laboratories,
Madison, NJ , and were within (0.4% of theoretical values
unless otherwise noted.
(s, 2H, bridge CH2), 6.80-7.40 (m, 8H, aromatic protons), 8.60
(s, 1H, C7-H). Anal. (C19H15N7S‚0.7CH3OH) C, H, N.
N-[(2,4-Diam in opter idin -6-yl)m eth yl]car bazole (4f). The
crude product after evaporation of the THF was purified flash
chromatography (95:5 CHCl3-MeOH) to obtain 4f as a bright-
yellow powder in 59% yield: mp 267 °C dec; IR (KBr) ν 3440,
3340, 3200, 1610, 1555, 1480, 1445, 1450, 1320, 1205 cm-1
;
1H NMR (DMSO-d6) δ 5.75 (s, 2H, bridge CH2), 7.10-8.25 (m,
8H, aromatic protons), 8.45 (s, 1H, C7-H); (MS FAB) m/z (M
+ 1) ) 342. Anal. (C19H15N7‚0.5H2O) C, H, N.
N-[(2,4-Dia m in op t er id in -6-yl)m et h yl]d ib en z[b,f]a ze-
p in e (4a ). A stirred solution of dibenz[b,f]azepine (78 mg,
0.404 mmol) in dry THF (10 mL) was purged at 0 °C with a
gentle stream of dry N2 (5 min) and treated with a single
portion of powdered NaH (95%, 51 mg, 2.13 mmol). After
another 10 min of stirring, 5‚HBr (100 mg, 0.392 mmol) was
added and the flask was fitted with a balloon of N2, allowed
to come to room temperature, and left to stir for another 12 h,
during which the color changed from greenish-yellow to
brilliant yellow. After decomposition of the excess NaH by
dropwise addition of MeOH (2 mL), the mixture was concen-
trated to dryness by rotary evaporation at 45 °C (bath
temperature). Flash chromatography on silica gel (95:5 CHCl3-
MeOH) gave 4a as a yellow powder which was dried in vacuo
at 70 °C overnight: yield 113 mg (78%); mp 251 °C dec; MS
(FAB) m/z (M + 1) ) 368; IR (KBr) ν 3430, 3290, 3180, 3100,
N-[(2,4-Dia m in op t er id in -6-yl)m et h yl]-N,N-d ip h en yl-
a m in e (4g). The crude product after evaporation of the THF
was purified by flash chromatography (95:5 CHCl3-MeOH)
to obtain 4g as a bright-yellow powder in 43% yield: mp >250
°C dec; MS (FAB) m/z (M + 1) ) 344; IR (KBr) ν 3450, 3340,
3170, 1630, 1590, 1550, 1490, 1450, 1360, 1220 cm-1; 1H NMR
(DMSO-d6) δ 5.10 (s, 2H, CH2), 6.70 (m, 10H, aromatic
protons), 8.60 (s, 1H, C7-H). Anal. (C19H17N7‚0.8H2O) C, H,
N.
DHF R Assa ys. The previously described spectrophotomet-
ric assay was used to measure the ability of compounds 4a -g
to decrease the rate of enzymic reduction of dihydrofolate to
tetrahydrofolate in the presence of NADPH.16 The standard
assay mixture contained 0.092 mM dihydrofolate, 0.117 mM
NADPH, 8.9 mM 2-mercaptoethanol, and 3.7 IU of enzyme (1
IU ) 0.005 optical density units/min) in 40.7 mM sodium
phosphate buffer, pH 7.4, in a total volume of 1.0 mL. The
assay mixture in the experiments using maDHFR also con-
tained 150 mM KCl. The enzyme was obtained from a clinical
isolate of M. avium (‘serovar 4’) as described earlier26 and
contained both DHFR and dihydropteroate synthase activity.
1620, 1600, 1575, 1550, 1470, 1440, 1350 cm-1 1H NMR
;
(DMSO-d6) δ 5.20 (s, 2H, 9-CH2), 6.80-7.40 (complex m, 1H,
CHdCH and aryl ring protons), 8.75 (s, 1H, C7-H). Anal.
(C21H17N7‚0.5CH3OH) C, H, N.
The following compounds were obtained by essentially the
As an illustration of the reproducibility of the DHFR assays,
the IC50 values (mean ( standard error) obtained over a recent
5-year period in Dr. Queener’s laboratory with pyrimethamine
against pcDHFR, tgDHFR, and rlDHFR have been 2.50 ( 0.23,
1.52 ( 0.32, and 2.39 ( 0.42 µM, respectively.
Mod elin g Stu d ies of DHF R Bin d in g. Molecular docking
simulations for compounds 4a -g into the active site of
pcDHFR and hDHFR were performed with the help of the
builder function of the Sybyl program22 using the standard
minimizer function and were similar to others we recently
reported.23 Models were made with the SETOR program.24
Recently analyzed crystal structures of the complexes of MTX
with pcDHFR and hDHFR12c were used in fitting the analogues
into the active site. The preliminary fit was made by super-
imposing the diaminopteridine ring of the analogue onto the
diaminopteridine ring of MTX.
same method as described for the synthesis of 4a .
N-[(2,4-Dia m in op ter id in -6-yl)m eth yl]-9,10-d ih yd r od i-
ben z[b,f]a zep in e (4b). Excess NaH in the reaction was
quenched by dropwise addition of a mixture of MeOH (1 mL)
and glacial AcOH (0.2 mL). After the solvents were removed
by rotary evaporation at 45 °C (bath temperature), the residue
was purified by preparative TLC (4:1 CHCl3-MeOH) to obtain
4b as a yellow powder in 39% yield: mp 275 °C dec; MS (FAB)
m/z (M + 1) ) 370; IR (KBr) ν 3450, 3300, 3220, 1610, 1555,
1
1480, 1440, 1360, 1220 cm-1; H NMR (DMSO-d6) δ 3.35 (s,
4H, CH2CH2), 5.15 (s, 2H, bridge CH2), 6.80-7.40 (m, 8H,
aromatic protons), 8.80 (s, 1H, C7-H). Anal. (C21H19N7‚0.5CH3-
OH) C, H, N.
N-[(2,4-Diam in opter idin -6-yl)m eth yl]-9,10-dih ydr oacr i-
d in e (4c). The crude product after evaporation of the THF
was taken up in 9:1 CHCl3-MeOH (40 mL), preadsorbed onto
silica gel (2 g), and applied to the top of a flash chromatography
column, which was eluted with 95:5 followed by 9:1 CHCl3-
MeOH. Appropriately pooled fractions were repurified on a
standard gravity column (95:5 CHCl3-MeOH) to obtain 4c as
a yellow solid in 50% yield: mp 222-224 °C dec; IR (KBr) ν
3450, 3370, 3320, 3200, 1640, 1615, 1590, 1560, 1480, 1440,
1360 cm-1; 1H NMR (DMSO-d6) δ 4.21 (s, 2H, ring CH2), 5.30
(s, 2H, bridge CH2), 6.50-7.50 (m, 8H, aromatic protons), 8.37
(s, 1H, C7-H); MS (FAB) m/z (M + 1) ) 370. Anal. (C20H17N7‚
0.5H2O) C, H, N.
Ack n ow led gm en t. This work was supported in part
by Grants RO1-AI29904 (A.R.) and RO1-GM51670
(V.C.) and by Contract NO1-AI35171 (S.F.Q.) from the
National Institutes of Health, DHHS.
Su p p or tin g In for m a tion Ava ila ble: Predicted contact
distances for the side-chain tricyclic moiety of compounds
4a -g with several key residues in pcDHFR and hDHFR. This
material is available free of charge via the Internet at http://
pubs.acs.org.
N-[(2,4-Diam in opter idin -6-yl)m eth yl]ph en oxazin e (4d).
The crude product after evaporation of the THF was purified
by flash chromatography (9:1 CHCl3-MeOH) followed by
preparative TLC (4:1 CHCl3-MeOH) to obtain 4d as a
brownish yellow solid in 54% yield: mp 250 °C dec; IR (KBr)
ν 3440, 3320, 3200, 3060, 1620, 1590, 1560, 1490, 1450, 1370,
Refer en ces
(1) (a) Falloon, 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:
Philadelphia, 1992. (b) Klepser, M. E.; Klepser, T. B. Drug
treatment of HIV-related opportunistic infections. Drugs 1997,
53, 40-73.
(2) DeClercq, E. Toward improved anti-HIV chemotherapy: thera-
peutic strategies for intervention with HIV infections. J . Med.
Chem. 1995, 38, 2491-1517.
(3) (a) Queener, S. F. New drug developments for opportunistic
infections in immunocompromised patients: Pneumocystis cari-
nii. J . Med. Chem. 1995, 38, 4739-4759. (b) Gangjee, A.; Elzein,
E.; Kothare, M.; Vasudevan, A. Classical and nonclassical
antifolates as potential antipneumocystis and antitoxoplasma
agents. Curr. Pharm. Des. 1996, 2, 263-280.
1270 cm-1 1H NMR (DMSO-d6) δ 5.02 (s, 2H, bridge CH2),
;
6.60-6.90 (m, 8H, aromatic protons), 8.70 (s, 1H, C7-H). Anal.
(C19H15N7O‚0.5CH3OH) C, H, N.
N-[(2,4-Dia m in op t er id in -6-yl)m et h yl]p h en ot h ia zin e
(4e). The crude product after evaporation of the THF was
purified by flash chromatography on silica gel (9:1 CHCl3-
MeOH) followed by preparative TLC (4:1 CHCl3-MeOH) to
obtain 4e as a greenish-yellow powder in 44% yield: mp 232-
234 °C dec; IR (KBr) ν 3450, 3380, 3310, 3200, 1630, 1590,
1560, 1540, 1450, 1360, 1220 cm-1; 1H NMR (DMSO-d6) δ 5.30

Synthesis, Antiparasitic Activity of Diaminopteridines
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 23 4859
(4) (a) Allegra, C. J .; Chabner, B. A.; Tuazon, C. U.; Ogata-Arakaki,
D.; Baird, B.; Drake, J . C.; Simmons, J . T.; Lack, E. E.;
Shelhamer, J . H.; Balis, F.; Walker, R.; Kovacs, J . A.; Lane, H.
C.; Masur. H. Trimetrexate for the treatment of Pneumocystis
carinii pneumonia in patients with the acquired immunodefi-
ciency syndrome. N. Engl. J . Med. 1987, 317, 978-985. (b)
Sattler, F. R.; Frame, P.; Davis, R.; Nichols, L.; Shelton, B.; Akil,
B.; Baughman, R.; Hughlett, C.; Weiss, W.; Boylen, C. T.; van
der Horst, C.; Black, J .; Powderly, W.; Steigbigel, R. T.; Leedom,
J . M.; Masur, H.; Feinberg, J .; Benoit, S.; Eyste, E.; Gocke, D.;
Beck, K.; Lederman, M.; Phari, J .; Reichman, R.; Sacks, H. S.;
Soiero, R. Trimetrexate with leucovorin versus trimethoprim-
sulfamethoxazole for moderate to severe episodes of Pneumocys-
tis carinii pneumonia in patients with AIDS - - a prospective
controlled multicenter investigation of the AIDS clinical trials
group Protocol 029/031. J . Infect. Dis. 1994, 170, 165-172.
(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) (a) Fischl, M. A.; Dickinson, G. M.; La Voie, L. Safety and efficacy
of sulfamethoxazole and trimethoprim chemoprophylaxis for
Pneumocystis carinii pneumonia in AIDS. J . Am. Med. Assoc.
1988, 105, 45-48. (b) Medina, I.; Mills, J .; Leoung, G.; Hopewell,
P. C.; Modin, G.; Benowitz, N.; Wofsy, C. B. Oral therapy for
Pneumocystis carinii pneumonia in the acquired immunodefi-
ciency syndrome. A controlled trial of trimethoprim-sulfamethox-
azole versus trimethoprim-dapsone. N. Engl. J . Med. 1990, 323,
776-782. (c) Podzamczer, D.; Salazar, A.; J imenes, J .; Santin,
M.; Consiglio, E.; Casanova, A.; Rufi, G.; Gudiol, F. Intermittent
trimethoprim-sulfamethoxazole compared with dapsone-py-
rimethamine for the simultaneous prophylaxis of Pneumocystis
pneumonia and toxoplasmosis in patients infected with HIV.
Ann. Intern. Med. 1995, 122, 755-761.
(7) (a) Caumes, E.; Roudier, C.; Rogeaux, O.; Bricaire, F.; Gentilini,
M. Effect of corticosteroids on the incidence of adverse cutaneous
reactions to trimethoprim-sulfamethoxazole associated Pneu-
mocystis carinii pneumonia. Clin. Infect. Dis. 1994, 17, 319-
323. (b) Roudier, C.; Caumes, E.; Rogeauz, O.; Bricaire, F.;
Gentilini, M. Adverse cutaneous reactions to trimethoprim-
sulfamethoxazole in patients with the acquired immunodefi-
ciency syndrome and Pneumocystis carinii pneumonia. Arch.
Dermatol. 1994, 30, 1383-1386.
(8) (a) Gangjee, A.; Devraj, R.; Queener, S. F. Synthesis and
dihydrofolate reductase activities of 2,4-diamino-5-deaza and 2,4-
diamino-5,10-dideaza lipophilic antifolates. J . Med. Chem. 1997,
40, 470-478. [For earlier papers from this group in this area,
see ref 3a; for recent papers from other groups, see refs 9 and
10.] (b) Gangjee, A.; Vasudevan, A.; Queener, S. F. Synthesis
and biological evaluation of nonclassical 2,4-diamino-5-meth-
ylpyrido[2,3-d]pyrimidines with novel side chain substituents as
potential inhibitors of dihydrofolate reductases. J . Med. Chem.
1997, 40, 479-485. (c) Gangjee, A.; Madvandadi, F.; Queener,
S. F. Effect of N9-methylation and bridge atom variation on the
activity of 5-substituted 2,4-diaminopyrrolo[2,3-d]pyrimidines
against dihydrofolate reductases from Pneumocystis carinii and
Toxoplasma gondii. J . Med. Chem. 1997, 40, 1173-1177. (d)
Gangjee, A.; Shi, J .; Queener, S. F. Synthesis and biological
activities of conformationally restricted, tricyclic nonclassical
antifolates as inhibitors of dihydrofolate reductases. J . Med.
Chem. 1997, 40, 1930-1936. (e) Gangjee, A.; Vasudevan, A.;
Queener, S. F. Conformationally restricted analogues of trimeth-
oprim: 2,6-diamino-8-substituted purines as potential dihydro-
folate reductase inhibitors from Pneumocystis carinii and Toxo-
plasma gondii. J . Med. Chem. 1997, 40, 3032-3039.
F. 2,4-Diaminothieno[2,3-d]pyrimidine lipophilic antifolates as
inhibitors of Pneumocystis carinii and Toxoplasma gondii dihy-
drofolate reductase. J . Med. Chem. 1997, 40, 3694-3699. (c)
Rosowsky, A.; Forsch, R. A.; Queener, S. F. Synthesis of 2,4-
diaminopteridines with bulky lipophilic groups at the 6-position
as inhibitors of Pneumocystis carinii, Toxoplasma gondii, and
mammalian dihydrofolate reductase. Pteridines 1997, 8, 173-
187. (d) 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 Toxo-
plasma gondii dihydrofolate reductase. J . Med. Chem. 1998, 41,
913-918. (e) Rosowsky, A.; Papoulis, A. T.; Forsch, R. A.;
Queener, S. F. Synthesis and antiparasitic and antitumor
activity of 2,4-diamino-6-(arylmethyl)-5,6,7,8-tetrahydroquinazo-
line analogues of piritrexim. J . Med. Chem. 1999, 42, 1007-
1017. (f) Rosowsky, A.; Papoulis, A. T.; Queener, S. F. One-step
synthesis of novel 2,4-diaminopyrimidine antifolates from bridged
alicyclic ketones and cyanoguanidine. J . Heterocycl. Chem. 1999,
36, 723-728.
(11) (a) 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. (b) Sirawaraporn, W.; Edman, J .
C.; Santi, D. V. Purification and properties of recombinant
Pneumocystis carinii dihydrofolate reductase. Protein Express.
Purif. 1991, 2, 313-316. (c) Delves, C. J .; Ballantine, C. P.;
Tansik, R. L.; Baccanari, D. P. Stammers, D. K. Refolding of
recombinant Pneumocystis carinii dihydrofolate reductase and
characterization of the enzyme. Protein Express. Purif. 1993, 4,
16-23. (d) Margosiak, S. A.; Appleman, J . R.; Santi, D. V.
Blakley, R. L. Dihydrofolate reductase from the pathogenic
fungus Pneumocystis carinii: catalytic properties and interaction
with antifolates. Arch. Biochem. Biophys. 1993, 305, 499-508.
(12) (a) 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. (b) 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 Crys-
tallogr. D 1997, D53, 638-649. (c) Cody, V.; Galitsky, N.; Rak,
D.; Luft, J . R.; Pangborn, W.; Queener, S. F. Ligand-induced
conformational changes in the crystal structures of Pneumocystis
carinii dihydrofolate reductase complexes with folate and NADP+.
Biochemistry 1999, 38, 4304-4312.
(13) (a) Southerland, W. M. A molecular model of the folate binding
site of Pneumocystis carinii dihydrofolate reductase. J . Comput.-
Aided Mol. Des. 1994, 8, 113-122. (b) Geschwend, D. A.; Good,
A. C.; Kuntz, I. D. Molecular docking towards drug discovery.
J . Mol. Recognit. 1996, 9, 175-186. (c) Geschwend, D. A.;
Sirawaraporn, W.; Santi, D. V.; Kuntz, I. D. Specificity in
structure-based design: identification of
a novel, selective
inhibitor of Pneumocystis carinii dihydrofolate reductase. Pro-
teins 1997, 29, 59-67.
(14) (a) Gangjee, A.; Guo, X.; Queener, S. F.; Cody, V.; Galitsky, N.;
Luft, J . R.; Pangborn, W. Selective Pneumocystis carinii dihy-
drofolate reductase inhibitors: design, synthesis, and biological
evaluation of new 2,4-diamino-5-substituted-furo[2,3-d]pyrim-
idines. J . Med. Chem. 1998, 44, 1263-1271. (b) Gangjee, A.;
Elzein, E.; Queener, S. F.; McGuire, J . J . Synthesis and biological
activities of tricyclic conformationally restricted tetrahydropyrido
annulated furo[2,3-d]pyrimidines as inhibitors of dihydrofolate
reductases. J . Med. Chem. 1998, 41, 1409-1416. (c) Gangjee,
A.; Vidwans, A. P.; Vasudevan, A.; Queener, S. F.; Kisliuk, R.
L.; Cody, V.; Li, R.; Galitsky, N.; Luft, J . R.; Pangborn, W.
Structure-based design and synthesis of lipophilic 2,4-diamino-
6-substituted quinazolines and their evaluation as inhibitors of
dihydrofolate reductases and potential antitumor agents. J . Med.
Chem. 1998, 41, 3426-3434.
(9) (a) 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 evalu-
ations. J . Med. Chem. 1996, 39, 1271-1280. (b) J ackson, H. C.;
Biggadike, K.; McKillingin, E.; Kinsman, O. S.; Queener, S. F.;
Lane, A.; Smith, J . E. 6,7-Disubstituted 2,4-diaminopteridines:
novel inhibitors of Pneumocystis carinii and Toxoplasma gondii
dihydrofolate reductase. Antimcirob. Agents Chemother. 1996,
40, 1371-1374. (c) Robson, C.; Meek, M. A.; Grundwaldt, J .-D.;
Lambert, P. A.; Queener, S. F.; Schmidt, D.; Griffin, R. J .
Nonclassical 2,4-diamino-5-aryl-6-ethylpyrimidine antifolates:
activity as inhibitors of dihydrofolate reductase from Pneu-
mocystis carinii and Toxoplasma gondii and as antitumor agents.
J . Med. Chem. 1997, 40, 3040-3048.
(15) Piper, J . R.; Montgomery, J . A. Preparation of 6-(bromomethyl)-
2,4-pteridinediamine hydrobromide and its use in improved
syntheses of methotrexate and related compounds. J . Org. Chem.
1977, 42, 208-210.
(16) (a) Broughton, M. C.; Queener, S. F. Pneumocystis carinii
dihydrofolate reductase used to screen potential antipneumocys-
tis drugs. Antimicrob. Agents Chemother. 1991, 35, 1348-1355.
(b) 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) (a) Queener, S. F.; Bartlett, M. S.; J ay, M. A.; Durkin, M. M.;
Smith, J . W. Activity of lipid-soluble inhibitors of dihydrofolate
reductase against Pneumocystis carinii in culture and in a rat
model of infection. Antimicrob. Agents Chemother. 1987, 31,
1323-1327. (b) Kovacs, J . A.; Allegra, C. J .; Kennedy, S.; Swan,
J . C.; Drake, J .; Parrillo, J . E.; Chabner, B.; Masur, H. Efficacy
of trimetrexate, a potent lipid-soluble antifolate, in the treatment
of rodent Pneumocystis carinii pneumonia. Am. J . Trop. Med.
(10) (a) Rosowsky, A.; Mota, C. E.; Queener, S. F. Brominated
trimetrexate analogues as inhibitors of Pneumocystis carinii and
Toxoplasma gondii dihydrofolate reductase. J . Heterocycl. Chem.
1996, 33, 1959-1966. [For earlier papers from this group in this
area, see ref 3a.] (b) Rosowsky, A.; Papoulis, A. T.; Queener, S.

4860 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 23
Rosowsky et al.
Hyg. 1988, 39, 491-496. (c) Bartlett, M. S.; Shaw, M.; Navaran,
P.; Smith, J . W.; Queener, S. F. Evaluation of potent inhibitors
of dihydrofolate reductase in a culture model for growth of
Pneumocystis carinii. Antimicrob. Agents Chemother. 1995, 39,
2436-2441.
de novo folate synthesis in Pneumocystis carinii and Toxoplasma
gondii: potential for screening therapeutic agents. J . Infect. Dis.
1989, 160, 312-320.
(22) Tripos, Inc., 1699 S. Handley Rd., Suite 303, St. Louis, MO
63144.
(18) (a) J ohnson, S. J .; Gupta, S. V.; Stevenson, K. J .; Freisheim, J .
H. Purification and characterization of dihydrofolate reductase
from Walker 256 carcinosarcoma. Can. J . Biochem. 1982, 60,
1132-1142. (b) Delcamp, T. J .; Susten, S. S.; Blankenship, D.
T.; Freisheim, J . H. Purification and characterization of dihy-
drofolate reductase from methotrexate-resistant human lym-
phoblastoid cells. Biochemistry 1983, 22, 633-639.
(23) Rosowsky, A.; Queener, S. F.; Cody V. Inhibition of dihydrofolate
reductases from Toxoplasma gondii, Pneumocystis carinii, and
rat liver by rotationally restricted analogues of pyrimethamine
and metoprine. Drug Des. Discovery 1999, 16, 25-40.
(24) Evans, S. V. SETOR. J . Mol. Graph. 1993, 11, 148-153.
(25) Tarnoky, A. L. Some properties of acridane and 2-chloro-7-
methoxyacridane: their possible relationships to excretion
product of mepacrine. Biochem. J . 1950, 46, 297-300.
(26) Chio, L.-C.; Boylard, L. A.; Nasr, M.; Queener, S. F. Identification
of a class of sulfonamides highly active against dihydropteroate
synthase from Toxoplasma gondii, Pneumocystis carinii, and
Mycobacterium avium. Antimicrob. Agents Chemother. 1996, 40,
727-733.
(19) Burchall, J . J .; Hitchings, G. H. Inhibitor binding analysis of
dihydrofolate reductases from various species. Mol. Pharmacol.
1965, 1, 126-136.
(20) Shoen, C. M.; Choromanska, O.; Reynolds, R. C.; Piper, J . R.;
J ohnson, C. A.; Cynamon, M. H. In vitro activities of several
diaminomethylpyridopyrimidines against Mycobacterium avium
complex. Antimicrob. Agents Chemother. 1998, 42, 3315-3316.
(21) Kovacs, J . A.; Allegra, C. J .; Beaver, J .; Boarman, D.; Lewis,
M.; Parrillo, J . E.; Chabner, B.; Masur, H. Characterization of
J M990331Q