Conformationally restricted tricyclic analogues of lipophilic pyrido[2,3-d]pyrimidine antifolates [1a,b]

Article abstract of DOI:10.1002/jhet.5570380132

The effect of conformational restriction of the C9-N10 bridge on inhibitory potency and selectivity of trimetrexate against dihydrofolate reductase, was studied. Specifically three nonclassical tricyclic 1,3-diamino-8-(3′,4′,5′-trimethoxybenzyl)-7,9-dihydro- pyrrolo[3,4-c]pyrido[2,3-d]pyrimidin-6(5H,8H)-one (4), 1,3-diamino-8-(3′,4′,5′-trimethoxybenzyl)-9-hydro-pyrrolo [3,4-c]pyrido[2,3-d]pyrimidin-6-(8H)-one (5) and 1,3-diamino-(8H)-(3′,4′,5′-trimethoxybenzyl)-7,9-dihydro- pyrrolo[3,4-c]pyrido[2,3-d]pyrimidine (7) antifolates were synthesized. The tricyclic analogues 4 and 5 were obtained via the regiospecific cyclo-condensation of the β-keto ester 17 with 2,4,6-triaminopyrimidine. The analogue 7 was obtained via reduction of the lactam 4 with borane in tetrahydrofuran. Compounds 4, 5 and 7 were evaluated as inhibitors of dihydrofolate reductase from Pneumocystis carinii, Toxoplasma gondii and rat liven All three compounds were more selective than trimetrexate against Pneumocystis carinii dihydrofolate reductase and significantly more selective than trimetrexate against Toxoplasma gondii dihydrofolate reductase compared with rat liver dihydrofolate reductase.

Full text of DOI:10.1002/jhet.5570380132

Jan-Feb 2001
Conformationally Restricted Tricyclic Analogues of
Lipophilic Pyrido[2,3-d]pyrimidine Antifolates [1a,b]
Aleem Gangjee* [a], Farahnaz Mavandadi [a,b] and Sherry F. Queener [c]
213
[a] Division of Medicinal Chemistry Graduate School of Pharmaceutical Sciences
Duquesne University, Pittsburgh, PA 15282
[b] FibroGen 225 Gateway Boulevard, South San Francisco, CA 94080
[c] Department of Pharmacology and Toxicology School of Medicine, Indiana University, Indianapolis, IN 46202
Received March 3, 2000
The effect of conformational restriction of the C9-N10 bridge on inhibitory potency and selectivity of
trimetrexate against dihydrofolate reductase, was studied. Specifically three nonclassical tricyclic
1,3-diamino-8-(3',4',5'-trimethoxybenzyl)-7,9-dihydro-pyrrolo[3,4-c]pyrido[2,3-d]pyrimidin-6(5H,8H)-one
(4), 1,3-diamino-8-(3',4',5'-trimethoxybenzyl)-9-hydro-pyrrolo[3,4-c]pyrido[2,3-d]pyrimidin-6-(8H)-one
(5) and 1,3-diamino-(8H)-(3',4',5'-trimethoxybenzyl)-7,9-dihydro-pyrrolo[3,4-c]pyrido[2,3-d]pyrimidine
(7) antifolates were synthesized. The tricyclic analogues 4 and 5 were obtained via the regiospecific cyclo-
condensation of the β-keto ester 17 with 2,4,6-triaminopyrimidine. The analogue 7 was obtained via
reduction of the lactam 4 with borane in tetrahydrofuran. Compounds 4, 5 and 7 were evaluated as inhibitors
of dihydrofolate reductase from Pneumocystis carinii, Toxoplasma gondii and rat liver. All three compounds
were more selective than trimetrexate against Pneumocystis carinii dihydrofolate reductase and significantly
more selective than trimetrexate against Toxoplasma gondii dihydrofolate reductase compared with rat liver
dihydrofolate reductase.
J. Heterocyclic Chem., 38, 213 (2001).
Currently available treatments for Pneumocystis carinii
adoption of different conformations of the trimethoxy
benzyl side chain of trimethoprim when bound to bacterial
and vertebrate dihydrofolate reductase as gleaned from
X-ray crystal structures of trimethoprim with dihydro-
folate reductase from Escherichia coli and chicken liver
[11]. Thus the side chain of trimethoprim adopts a "down"
conformation in E. coli and an "up" conformation in
chicken liver dihydrofolate reductase with respect to the
2,4-diaminopyrimidine moiety. X-ray crystal structure of
trimethoprim and P. carinii dihydrofolate reductase
indicate that the trimethoxybenzyl side chain of
trimethoprim is oriented in the "down" conformation
similar to that observed in the E. coli enzyme [12]. The
selectivity of trimethoprim for dihydrofolate reductase
from P. carinii compared to rat liver, though not as high as
that for the E. coli enzyme perhaps also stems, in part,
from the different side chain orientation of trimethoprim
when bound to P. carinii dihydrofolate reductase
compared to the rat liver enzyme.
Bicyclic 2,4-diamino nonclassical dihydrofolate
reductase inhibitors related to the pteridines [13],
pyrido[2,3-d]pyrimidines [14], pyrido[3,2-d]pyrimidines
[15], quinazolines [16], pyrrolo[2,3-d]pyrimidines [17],
furo[2,3-d]pyrimidines [18], and thieno[2,3-d]pyrimidines
[19] among others [20,21] have been synthesized as
potential potent and selective inhibitors of dihydrofolate
reductase from P. carinii and T. gondii. Though structure-
activity/selectivity correlations for each bicyclic system
have been developed no structure-activity relationship
across the different heterocyclic systems is possible at this
time and each heterocyclic system needs to be considered
separately.
(P. carinii) and Toxoplasma gondii (T. gondii), which are
often fatal opportunistic infections in AIDS patients, are
usually a combination of agents [2-4]. The use of
trimethoprim and sulfamethoxazole is considered to be the
most effective combination for the treatment and prophyl-
axis of P. carnii infections [5]. T. gondii infections are
treated by the first line combination of pyrimethamine and
sulfadiazine [6]. Both these combinations utilize a selective
dihydrofolate reductase inhibitor trimethoprim or
pyrimethamine along with a dihydropteroate synthase
inhibitor that is the sulfa drug. Though these combinations
are effective, a significant number of patients suffer side
effects primarily attributed to the sulfonamide drug which
limits the dose and in many cases forces a discontinuation
of therapy [7,8]. The dihydrofolate reductase inhibitors in
these combinations are weak ineffective agents when used
as monotherapy and require the sulfonamide to afford a
synergistic, clinically effective combination [9,10]. In an
attempt to circumvent the use of the sulfonamide in these
combinations, due to its toxicity which necessitates
discontinuation of treatment, considerable effort has been
expended to design and synthesize pathogen selective dihy-
drofolate reductase inhibitors that also possess increased
potencies against P. carinii dihydrofolate reductase and
T. gondii dihydrofolate reductase. Such a potent and selec-
tive dihydrofolate reductase inhibitor would preclude the
necessity of the sulfonamide and hence the toxicity
associated with the combination and could be clinically
viable monotherapy for P. carinii and T. gondii infections.
The 3,000 to 10,000 fold selectivity of trimethoprim
(TMP) for bacterial dihydrofolate reductase over the
mammalian enzyme has been attributed, in part, to the

214
A. Gangjee, F. Mavandadi and S. F. Queener
Vol. 38
OMe
OMe
OMe
NH2
N
NH2 CH3
9
10
OMe
CH2NH
N
N
H2N
OMe
H2N
N
OMe
TMQ
TMP
NH2 R1
NH2 CH3
OCH3
OCH3
R2
X
Y
τ1 τ2
N
N
H2N
N
N
H2N
N
N
1 R1 = CH3, X = CH2, Y = NH or NCH3
2 R1 = H, X = NH or NCH3, Y = CH2
PTX
R1
R1
NH2
N
N
NH2
N
N
O
N
N
H2N
N
R2
H2N
N
3
3a R1 = 3,4,5-triOMePh
The bicyclic 2,4-diamino-6-substituted quinazoline,
trimetrexate (TMQ) and the pyrido[2,3-d]pyrimidine
piritrexim (PTX) were both found to be highly potent against
dihydrofolate reductase from P. carinii and T. gondii,
however these bicyclic agents were devoid of selectivity
against the pathogenic enzymes [10,22]. Trimetrexate, which
is approved as a second line agent against P. carinii
infections must be used along with the folate cofactor,
leucovorin, to rescue host cells [23]. Gangjee et al. [24] have
reported a generalization in the structure-activity/selectivity
of the pyrido[2,3-d]pyrimidine ring system which indicates
that selected analogues of the 5,10-dimethyl substituted
series of general structure 1 as well as selected analogues
containing a 9-methyl substitution of general structure 2 [14]
provide significant potency and selectivity for P. carinii
and/or T. gondii dihydrofolate reductase.
Molecular modeling of 2,4-diamino-5,10-desmethyl
6-substituted pyrido[2,3-d]pyrimidines compared with the
5,10-dimethyl congeners using SYBYL 6.2 [25] and its
SEARCH and MAXIMIN options indicated that mono and
dimethyl substitution at the 5- and/or 10-positions
significantly decreases the number of conformations
compared to the unsubstituted analogues. Thus partial
conformational restriction of the 9-10 bridge of
2,4-diamino-6-substituted pyrido[2,3-d]pyrimidines with
methyl moieties on the 5-and 9- or 10-positions provides
increased selectivity and/or potency against P. carinii
and/or T. gondii dihydrofolate reductase [14,24]. It was
therefore of interest to further conformationally restrict the
9-10 bridge of pyrido[2,3-d]pyrimidines in an attempt to
study the effect on potency and selectivity against
P. carinii and T. gondii dihydrofolate reductase.
Conformational restriction of the 9-10 bridge beyond
that obtained by methyl substitutions on the 5- and
10-positions could be achieved by tethering the 5- and
10-positions via a methyl or ethyl link to afford tricyclic
pyrrolo or pyrido annulated 2,4-diamino pyrido-
[2,3-d]pyrimidines. Gangjee et al. [26] recently reported
the tricyclic pyrido annulated analogues in which the C5
and N10 were tethered via an ethyl link effectively
restricting the 9-10 bridge. Some members of this series of
conformationally restricted tricyclic analogues represented
by general structure 3 were potent inhibitors of rat liver
dihydrofolate reductase with inhibitory constants (IC )
50
which were highly potent (IC 86 nM, for 3 R = 3,4,5-
50
1
OMePh, R = OH) but the series lacked selectivity for
2
P. carinii or T. gondii dihydrofolate reductase and were in
fact much more inhibitory against mammalian
dihydrofolate reductase.
R
R
8
9
N
N
NH2
1
NH2
N
7
2N
N
6
3
H2N
N
N5
H
O
H2N
N
O
4
4
5
R
R
N
N
NH2
NH2
N
N
N
n-Bu
N
N
H
H2N
N
6
7

Jan-Feb 2001
Tricyclic Analogues of Lipophilic Pyrido[2,3-d]pyrimidine Antifolates
215
The lack of potent inhibitory effects against P. carinii
and T. gondii dihydrofolate reductase of analogues 3 was
surprising, particularly since the 5,10-dimethyl analogue 1
analogues 3. Molecular modeling [25] of analogues 4-7
indicated that the triOMePh side chain was oriented in
different conformations to that observed for analogues 3 and
were in close proximity to that observed for trimetrexate
bound to P. carinii dihydrofolate reductase [12]. Molecular
modeling further indicated that a methylene spacer between
the tricyclic pyrrolo annulated system and the 3,4,5-
trimethoxyphenyl side chain allowed for flexibility in the
side chain and that the trimethoxybenzyl analogues 4-7
provided a better overlap with dihydrofolate reductase
bound trimetrexate. Thus the trimethoxybenzyl moiety was
included as the side chain in preference to the
trimethoxyphenyl. Though compound 7 was the primary
target, we were very much interested in the partially reduced
analogues 4 and 6 as well as the essentially planar analogue
5. The varying levels of unsaturation in rings B and C of 4-7
provide subtle variations in the tricyclic ring conformation
which provides increased potency and/or selectivity as we
have previously reported for the pyrido annulated analogues
[26]. In fact, in the tricyclic series 3, the most selective
(R = CH , X = CH , Y = NCH , R = 3,4,5 triOMePh)
1
3
2
3
2
was significantly potent against all three dihydrofolate
reductases, and displayed a 9-fold selectivity for T. gondii
dihydrofolate reductase as compared to rat liver
dihydrofolate reductase (Table 1). Clearly the ethyl bridge
between the 5- and 10-positions was detrimental to
potency against P. carinii and T. gondii dihydrofolate
reductase. The reason for this was attributed, in part, to the
inappropriate orientation of the trimethoxyphenyl ring,
which is predicated on the conformation of the C-ring, and
was not conducive to P. carinii or T. gondii dihydrofolate
reductase binding.
In an attempt to further restrict the 9-10 bond, tricyclic
analogues 4-7 in which the 5- and 10-positions were linked
via a methylene rather than an ethyl link were designed.
This afforded a 5-membered C-ring in which the 9-10
bridge was now constrained with less flexibility than in
Scheme I
A.
NH2
OCH3
OCH3
a
CH3
H3C
N
+
CH3
H3C
O
O
OCH3
OCH3
OCH3
8
9
10
b, c, d
OCH3
CHO
CH3
H3C
N
OCH3
OCH3
OCH3
11
a: Toluene/reflux; b: DMF/ POCl3/15 °C; c: 10 in DMF/30minutes/ 0 °C; d: 90 °C/2 hours
B.
H3C
OCH3
NH2
NH2
N
e
N
N
OCH3
+
11
CH3
H2N
N
NN CH
H
OCH3
H2N
N
NHNH2
12
13
OCH3
OCH3
H3C
NH2
N
f
CH3
N
OCH3
H2N
N
N
14
e: EtOH/Glacial AcOH/90 °C/48 hours; f: Fischer indole conditions.

216
A. Gangjee, F. Mavandadi and S. F. Queener
Vol. 38
analogue was the essentially planar compound 3a thus
supporting the synthesis and evaluation of compound 5 as a
target analogue.
Gangjee et al. [28] reported a novel strategy for the
synthesis of the tricyclic pyrido[2,3-d]pyrimidines via a
Fischer indole type cyclization of a 2-amino-4-oxo
afforded the 3-formyl pyrrole, 11, as a tan solid in 91%
yield. The hydrazone 13 was synthesized by condensing an
equimolar quantity of the hydrazine 12 and aldehyde 11 in a
mixture of ethanol and glacial acetic acid. Attempted
Fischer indole cyclization of 13 under a variety of
conditions, however, failed to yield the desired product.
Scheme 2
a
COOC2H5
+
9
R-H2C-HN
COOC2H5
O
15
16
b
c
R
R-H2C-N
COOC2H5
COOC2H5
N
COOC2H5
17
18
NH2
N
d
e
+
[ 6 ]
air
oxidation / f
+ 7
18
4
H2N
g
N
NH2
19
OCH3
OCH3
OCH3
R =
5
a: THF/ 60 °C/3 days; b: BrCH CO C H /DMF/60 °C; c: NaOEt/ Toluene; d: Glacial AcOH/120 °C/12 hours; e: BH /THF; f: MeOH/reflux; g: Dowtherm-
2
2
2
5
3
A/240 °C/2 hours.
hydrazone. A similar strategy was initially attempted for the
synthesis of tricyclic analogues 4-7 as shown in Scheme I.
The synthesis of 2,4-diamino-6-hydrazinopyrimidine (12)
involved the reaction of 2,4-diamino-6-chloropyrimidine
with excess hydrazine in butanol. The product precipitated
from the reaction mixture as a white, methanol insoluble
solid and was isolated as a powder in 45% yield. The insol-
ubility of 12 in deuterated chloroform and its instability in
An alternate methodology, which could afford the
desired tricyclic analogues, was the condensation of 2,4,6-
triaminopyrimidine with a suitable biselectrophile.
Gangjee et al. [27] had reported the synthesis of similar
compounds using this method and hence this was a logical
choice. Syntheses of the target compounds 4-7 were
accomplished via the regiospecific cyclocondensation of
the β-keto ester 18 (as the biselectrophile) with 2,4,6-
triaminopyrimidine (Scheme 2). Hurlbert et al. [29] as
well as reports from Gangjee et al. [30,31] and DeGraw et
al. [32] have confirmed that β-keto esters condense with
appropriate 6-aminopyrimidines to afford regiospecifically
angular, 5,6-disubstituted pyrido[2,3-d]pyrimidines rather
than the linear, 6,7-disubstituted isomers. Compound 16
was synthesized by the reaction of 3,4,5-trimethoxy-
benzylamine 9 with ethyl acrylate (15) followed by
alkylation with ethylbromoacetate to afford 17 which on
Dieckmann cyclization gave the desired β-ketoester 18.
Though diphenyl ether had been previously employed as
the solvent [27], it was of interest to explore different
solvent systems since the extent of unsaturation in the
product could be influenced by the solvent in the
cyclocondensation reaction [26,30,31]. Cyclocondensation
of 18 with 2,4,6-triaminopyrimidine (19) in glacial acetic
acid at 120 °C for 12 hours afforded exclusively the
1
deuterated dimethylsulfoxide made it difficult to obtain a H
nmr spectrum. For model studies, the 2,5-dimethyl pyrrole-
3-carboxaldehyde, 11, was selected to avoid any potential
α-polymerization. The N-substituted pyrrole-3-carboxal-
dehyde, 11, was obtained using the Paal-Knorr synthesis
(Scheme 1). The pyrrole 10 was synthesized first by
refluxing equimolar qauntities of acetonyl acetone (8) and
3,4,5-trimethoxybenzylamine (9), in a flask fitted with a
Dean-Stark trap, until separation of water ceased (4 hours).
Upon cooling, a tan solid was obtained in 70% yield. The
N-trimethoxybenzyl-2,5-dimethylpyrrole (10) was
formylated via a Vilsmeier Haack formylation using
dimethylformamide and phosphorus oxychloride. A
solution of 10 in dimethylformamide was slowly added to a
freshly prepared mixture of dimethylformamide and
phosphorus oxychloride at 0 °C. The reaction mixture was
then heated briefly (2 hours) at 90 °C and after work up

Jan-Feb 2001
Tricyclic Analogues of Lipophilic Pyrido[2,3-d]pyrimidine Antifolates
217
angular lactam 4 in 55% yield, while similar cyclocon-
densation in Dowtherm-A at 240 °C for 2 hours gave the
dehydrogenated angular lactam 5 as the sole product in
suggests that partial planarity of the C-ring with the
pyrido[2,3-d]pyrimidine system was important for potency
and selectivity in this series. Thus, conformationally
restricting τ and τ of 1 with partial planarity of the C-ring
1
40% yield. H nmr of 4 showed an exchangeable (D O)
2
1
2
lactam NH at δ 8.98 and three sets of methylene protons at
as in compound 5 significantly increases selectivity
against both P. carinii and T. gondii dihydrofolate
reductase compared to 1 and trimetrexate. A comparison
of 4 and 7 suggests that in the absence of planarity in the
C-ring the lactam is somewhat detrimental to potency as
well as selectivity (except for T. gondii dihydrofolate
reductase). Since compound 5 was the most potent and
selective analogue against both P. carinii and T. gondii
dihydrofolate reductase in this series and compound 5 also
contains the nitrogen least likely to be protonated in the
C-ring, due to conjugation, it is also possible that the
increased potency and selectivity of 5 could arise due to its
inability to be protonated at the nitrogen atom. Thus the
selectivity and potency of 5 could be attributed to the
planarity of the tricyclic systems and/or the inability of 5 to
be protonated compared to 4 and 7.
δ 3.17, 3.78 and 4.43, as was previously reported for
1
similar tricyclic compounds [30-32]. In contrast the H
nmr of 5 indicated a nonexchangeable olefinic proton at δ
8.82 as we have observed earlier, and only two sets of
methylene protons at δ 4.38 and 4.65. The lactam 4 was
reduced using borane in tetrahydrofuran to afford a
1
mixture of 6 and 7 in a 1:1 ratio as indicated by the H nmr
and the mass spectrum of the mixture. All attempts to
separate the mixture resulted in the oxidation of 6 to 7.
Thus only 7 was isolated as the pure reduced lactam. The
oxidation also occurs when the mixture is allowed to stand
at room temperature for an extended period. Conversion of
the mixture to pure 7 was also accomplished by reflux in
methanol. Compound 7 has been claimed in a patent [33]
via a different synthetic route [34], however, no spectral or
biological data was reported.
Table I
Inhibitory Concentrations (IC ) in µM against Dihydrofolate Reductase and Selectivity Ratios of 4, 5 and 7.
50
Selectivity Ratio
rl/pc
Selectivity Ratio
Compound
P. carinii [a]
Rat liver [b]
T. gondii [c]
rl/tg
1
0.013
0.042
94.5
3.10
12.6
0.007
0.003
42.9
20.3
11.7
0.58
0.07
0.45
6.50
0.90
0.0008
0.0010.30
9.00
0.62
5.30
8.90
Trimetrexate
4
5
7
4.80
32.7
2.20
[a] Pneumocystis carinii dihydrofolate reductase. [b] Rat liver dihydrofolate reductase. [c] Toxoplasma gondii dihydrofolate reductase.
The compounds 4-5 and 7 were evaluated as inhibitors
of dihydrofolate reductase from P. carinii [35], T. gondii
EXPERIMENTAL
All evaporations were carried out in vacuo with a rotary
evaporator. Analytical samples were dried in vacuo (0.2 mm
mercury) in an Abderhalden drying apparatus over phosphorus
pentoxide and refluxing ethanol or toluene. Melting points were
determined on a Fisher-Johns melting point apparatus and are
uncorrected. Nuclear Magnetic Resonance spectra for proton
[36] and rat liver. Selectivity ratios (IC rat liver
50
dihydrofolate reductase/IC P. carinii dihydrofolate
50
reductase or T. gondii dihydrofolate reductase) were
determined using rat liver dihydrofolate reductase as the
mammalian source. These IC values along with
50
selectivity ratios are shown in Table I. The inhibitory
1
( H nmr) were recorded on a Bruker WH-300 (300 MHz) spec-
values for 1 (Y = NCH , R = 3,4,5-triOMe) and
3
2
trometer. Data was accumulated by 16 K size with a 0.5 s delay
time and 70° tip angle. The chemical shift values are expressed in
ppm (parts per million) relative to tetramethylsilane as internal
standard; s = singlet, d = doublet, dd = doublet of doublet, t =
triplet, q = quartet, m = multiplet, bs = broad singlet. The relative
integrals of peak areas agreed with those expected for the
assigned structures. Thin layer chromatography was performed
trimetrexate are also included for comparison. The most
potent of the three target compounds against P. carinii and
T. gondii dihydrofolate reductase was the dehydrogenated
lactam 5. More significantly this compound was 93-fold
more selective against P. carinii dihydrofolate reductase
and about 110-fold more selective against T. gondii
dihydrofolate reductase than trimetrexate. All three
compounds, 4, 5 and 7, were more selective than
trimetrexate against both P. carinii and T. gondii
dihydrofolate reductase. The significant increase in both
potency and selectivity of 5 compared to 4 strongly
on POLYGRAM Sil G/UV
silica gel plates with fluorescent
254
indicator and the spots were visualized under 254 nm and 366 nm
illumination. Proportions of solvents used for thin layer
chromatography are by volume. Elemental analysis were
performed by Atlantic Microlabs Inc., Norcoss, GA. Analytical
results indicated by element symbols are within ± 0.4% of the

218
A. Gangjee, F. Mavandadi and S. F. Queener
Vol. 38
calculated values. Fractional moles of water or organic solvents
frequently found in some analytical samples of antifolates could
not be removed in spite of 24-48 hours of drying in vacuo and
1,3-Diamino-(8H)-(3',4',5'-trimethoxybenzyl)-7,9-dihydro-
pyrrolo[3,4-c]pyrido[2,3-d]pyrimidine (7).
A 1.0 M borane-tetrahydrofuran complex 8.00 ml,
(8.00 mmoles) was added to a cold suspension of the dihydrol-
actam 3 0.40 g, (1.00 mmole) in tetrahydrofuran (25 ml). On
completion of the addition, a clear brown solution formed that was
stirred at room temperature for 16-18 hours. The mixture was then
cooled to 0 °C, acidified carefully with 6 N hydrochloric acid to pH
2 and stirred at 0 °C for 1 hour and evaporated under reduced
pressure. Cold distilled water (25 ml) was added to the residue to
afford a thick white suspension, which was neutralized at 0 ° C
with 1 N sodium hydroxide and stirred at 0 °C for 3 hours. The
white suspension was filtered, washed with water and dried under
vacuum. Tlc of this showed the presence of the starting lactam 3,
hence, the mixture was retreated with another portion of 1.0 M
borane-tetrahydrofuran complex 6.40 ml, (6.40 mmole) for 16
hours. The product obtained on acidification, evaporation, addition
of water, neutralization and filtration did not show the presence of
the lactam (tlc) and was triturated in refluxing methanol (20 ml) for
5 hours. Filtration and evaporation of the filtrate under reduced
pressure afforded a residue which was dissolved in glacial acetic
acid (5 ml), stirred at room temperature for 18 hours and
concentrated under reduced pressure. The concentrate was tritu-
rated with anhydrous diethyl ether for 3-4 hours and the precipitate
obtained was filtered to yield 0.12 g (31%) of 7 as a light tan solid.
1
were confirmed, where possible, by their presence in the H nmr
spectrum. All solvents and chemicals were purchased from
Aldrich Chemical Co. and Fisher Scientific and were used as
received.
1,3-Diamino-8-(3',4',5'-trimethoxybenzyl)-7,9-dihydropyrrolo-
[3,4-c]pyrido[2,3-d]pyrimidin-6(5H,8H)-one (4).
A mixture of triaminopyrimidine (19) 0.113 g, (0.90
mmole), 18 0.30 g, (0.90 mmole) and glacial acetic acid
(10 ml) in a three necked flask fitted with a Dean-Stark trap
was immersed in an oil bath heated to 120 °C. The suspension
was stirred at this temperature under nitrogen for 12 hours.
The reaction mixture was then cooled to room temperature
and ether (50 ml) was added to the yellow suspension which
was stirred for 30 minutes, filtered and washed with ether to
yield crude lactam 4. The crude lactam was then dissolved in
methanol and silica gel (1.00 g) was added to this solution and
the solvent evaporated under reduced pressure to form a
uniformly coated silica gel plug which was dried under
vacuum and loaded onto a dry silica gel column (2.4 x 20 cm)
and eluted with a 10%-50% methanol in chloroform (1%
ammonium hydroxide) gradient. Fractions containing the
product (tlc) were pooled, concentrated under reduced
pressure and neutralized with glacial acetic acid to yield
mp > 300 °C, tlc R 0.4 (chloroform/methanol/ammonium
f
+
1
hydroxide 9:4:0.1, silica gel) MS m/z 383 (M ), H nmr
0.19 g (55%) of 4 as a tan powder: mp > 300 °C, tlc R 0.60
(DMSO-d ): δ 3.55 (s, 2H, CH ), 3.60 (s, 2H, CH ), 3.68 (s, 3H,
f
6
2
2
(chloroform/methanol.ammonium hydroxide 9:4:0.1, silica
OCH ), 3.72 (s, 6H, (OCH ) ), 5.51 (s, 2H, NH ), 5.68 (s, 2H,
3
3 2
2
1
+
gel), H nmr (DMSO-d ): δ 3.17 (s, 2H, CH ), 3.41 (s, 2H,
NH ), 6.30 (s, 1H, NH ), 6.57 (m, 3H, 2',6'-CH, 5-CH).
6
2
2
CH ), 3.67 (s, 3H, 4'-OCH ), 3.78 (s, 6H, 3',5'-OCH ), 4.43
Anal. Calcd. for C H N O •0.5CH OH•1.0H O: C, 49.60;
2
3
3
19 21
6
3
3
2
(s, 2H, N-CH ), 5.60 (s, 2H, NH ), 5.98 (s, 2H, NH ), 6.56
H, 6.08; N, 17.80. Found: C, 49.92; H, 6.11; N, 17.63.
2
2
2
(s, 2H, 2',5'-CH), 8.98 (s, 1H, 5-NH).
Anal. Calcd. for C N O •0.3H O•0.5CH COOH: C,
N-(3',4',5'-Trimethoxybenzyl)-2,5-dimethylpyrrole (10).
H
19 22
6
4
2
3
55.37; H, 5.72; N, 19.37. Found: C, 55.02; H, 5.42; N, 19.70.
Acetonyl acetone (8) (1.97 ml, 10.0 mmole) and 3,4,5-
trimethoxybenzylamine (9) 1.97 g, (10.0 mmole) were dissolved
in 5 ml of toluene and heated under reflux in a flask fitted with a
Dean-Stark trap until separation of water ceased (4 hours). The
reaction was cooled to room temperature and allowed to stand
overnight. The sandy precipitate obtained was filtered and
1,3-Diamino-8-(3',4',5'-trimethoxybenzyl)-9-hydropyrrolo-
[3,4-c]pyrido[2,3-d]pyrimidin-6-(8H)-one (5).
A mixture of triaminopyrimidine (19) 0.75 g, (6.0 mmoles),
18 2.00 g, (6.0 mmoles) and Dowtherm-A (20 ml) in a three
necked flask fitted with a Dean-Stark trap was immersed in an
oil bath preheated to 240 °C. The suspension was stirred at
this temperature under nitrogen for 2 hours until no increase
was observed in the ethanol/water level in the Dean-Stark
trap. The reaction mixture was cooled to room temperature.
Ether (10 ml) was added to the mixture, which was further
cooled to 0 °C and filtered to yield the crude lactam. This was
evaporated under reduced pressure to form a uniformly coated
silica gel plug which was dried under vacuum and loaded onto
a dry silica gel column (2.4 x 20 cm) which was eluted with a
10-50% chloroform:methanol gradient. Fractions containing
the product were pooled and evaporated to yield 0.80 g (40%)
washed with methanol to yield 1.92 g (70%) of 10 as a tan solid:
1
mp 104-106 °C; tlc R 0.72 (ethyl acetate, silica gel); H nmr
f
(60 MHz, DMSO-d ): δ 2.10 (s, 6H, 2,5-CH ), 3.65 (s, 9H,
6
3
OCH ), 5.00 (s, 2H, N-CH ), 5.7 (s, 2H, 3,4-CH), 6.2 (s, 2H,
3
2
2',6'-CH).
N-(3',4',5'-Trimethoxybenzyl)-2,5-dimethyl-3-formylpyrrole (11).
To dimethylformamide (1.85 ml) in a three neck flask under
nitrogen at 0 °C was added phosphorus oxychloride (POCl )
3
(0.56 ml) dropwise with stirring over a period of 15 minutes. To
this mixture was added a solution of 10 1.0 g, (3.60 mmoles) in
dimethylformamide (7 ml) over a period of 30 minutes at 0 °C.
The reaction mixture was then heated at 90 °C for 2 hours, cooled
to room temperature and poured into ice. The mixture was then
made basic to pH 10 with 30% sodium hydroxide. The precipitate
formed was filtered and air dried to yield 0.96 g (91%) of 11 as a
of 5 as a yellow solid. mp > 300 °C, tlc R 0.62 (chloroform/-
f
1
methanol/ammonium hydroxide 9:4:0.1, silica gel), H nmr
(DMSO-d ): δ 3.64 (s, 3H, 4'-OCH ), 3.75 (s, 6H,
6
3
3',5'-OCH ), 4.38 (s, 2H, CH ), 4.65 (s, 2H, CH ), 6.62 (s, 2H,
3
2
2
o
tan solid. mp 110 C; tlc R 0.59 (ethyl acetate , silica gel);
2',5'-CH), 6.76 (bs, 2H, NH ), 7.73 (bs, 2H, NH ), 8.82 (s, 1H,
f
2
2
1
H nmr (60 MHz, DMSO-d ): δ 2.10 (s, 3H, 5-CH ), 2.45 (s, 3H,
7-CH).
Anal. Calcd. for C H N O •0.5H O: C, 56.29; H, 5.22; N,
6
3
2-CH ), 3.60 (d, 9H, OCH ), 5.10 (s, 2H, N-CH ), 6.25 (s, 3H,
19 20
6
4
2
3
3
2
20.73. Found: C, 56.33; H, 5.30; N, 20.40.
4-CH, 2',6'-CH), 9.75 (s, 1H, CHO).

Jan-Feb 2001
Tricyclic Analogues of Lipophilic Pyrido[2,3-d]pyrimidine Antifolates
219
3-[2-(2,4-Diaminopyrimidin-6-yl)hydrazo]methenyl-1-(3',4',5'-
trimethoxybenzyl)-2,5-dimethylpyrrole (13).
yield a viscous orange-yellow oil. A solid separated from the oil
on refrigeration which upon filtration afforded 21.7 g (61.5%) of
the ketoester 18 as an off-white solid: mp 115-120 °C; tlc R 0.38
(ethyl acetate/1 drop acetic acid, silica gel).
f
A mixture of 11 0.29 g, (1.25 mmoles) and 12 0.195 g,
(1.25 mmoles) in ethanol (10 ml) containing glacial acetic acid
(0.35 ml) was heated at 90 °C for 48 hours. The reaction mixture
was cooled to room temperature and poured into crushed ice. The
pH of the mixture was adjusted to 10 with 60% sodium
hydroxide. The solid obtained was filtered to yield 0.22 g (54%)
Anal. Calcd. for C H O N•0.2H O: C, 60.06; H, 6.69; N,
17 22
6
2
4.12. Found: C, 59.97; H, 6.92; N, 4.17.
Acknowledgement.
This work was supported in part by a grant from the National
Institutes of Health NIH GM 52811 (A.G.), AI41743 (A.G.), and
NIH Contracts N01-AI-87240 (S.F.Q.) and N01-AI-3171
(S.F.Q.).
of 13 as a tan solid; mp 170-180 °C; tlc R 0.45 (chloroform/-
f
methanol/glacial acetic acid 9:4:0.1, silica gel); MS m/z 426
+
1
(M ); H nmr (DMSO-d ): δ 2.11 (s, 3H, CH ), 2.24
6
3
(s, 3H, CH ), 3.61 (s, 3H, 4"-OCH ), 3.66 (s, 6H, 3", 5"-OCH ),
3
3
3
5.01 (s, 2H, N-CH ), 5.56 (s, 2H, NH ), 5.74 (s, 2H, NH ), 6.10
(s, 2H, 2'-CH, 5-CH), 6.21 (s, 2H, 2",6"-CH), 7.96 (s, 1H,
N=CH), 9.87 (s, 1H, NH).
2
2
2
REFERENCES AND NOTES
[1a] Taken in part from the dissertation submitted by F.M. to the
Graduate School of Pharmaceutical Sciences, Duquesne University, in
partial fulfillment of the requirements for the Doctor of Philosophy
degree, July 1995. [b] A. Gangjee, F. Mavandadi and S. F. Queener.
Advances in Experimental Medicine and Biology, J. E. Ayling, M. G.Nair
and C. M.Baugh, Eds., Plenum Press, 1993, 338, 441.
[2] M. E. Klepser and T. B. Klepser, Drugs, 53(1), 40 (1997).
[3] C. Carmichael, Primary Care, 24(3), 561 (1997).
[4] A. K. Freedberg, J. A. Scharfstein, G. R. Seage III, E. Losina,
M. C. Weinstein, D. E. Craven and A. D. Paltiel, JAMA, 279(2), 130
(1998).
Ethyl N-(3,4,5-trimethoxybenzyl)-β-amino Propionate (16).
A mixture of 3,4,5-trimethoxybenzylamine (9) 39.4 g,
(200 mmoles) and ethyl acrylate (15) 21.6 ml, (199.6 mmoles) in
tetrahydrofuran (THF) (50 ml) was stirred at room temperature
for 2 days. The mixture was concentrated in vacuo and the
residue was flash distilled to obtain 35.0 g (59%) of 16 as a
viscous colorless liquid; bp 145-150 °C at 0.05 mm mercury, tlc
R 0.32 (ethyl acetate/1 drop acetic acid, silica gel). The
f
compound was used without further purification in the next step.
[5a] E. Warren, S. George, J. You and P. Kazanjian,
Pharmacotherapy, 17(5), 900 (1997); [b] S. S. Morris-Jones and P. J.
Easterbrook, Antimicrobial Chemother., 40, 315 (1997).
[6a] V. St. Georgiev. Drugs, 48(2), 179 (1994); [b] R. Behbahani,
M. Moshfeghi and J. D. Baxter, Ann. Pharmacother, 29, 760 (1995); [c]
D. Podzamczer, A. Salazar, J. Jiménez, E. Consiglio, M. Santin, A.
Casanova, G. Rufi and F. Gudiol.; Ann. Int. Med., 122(10), 755 (1995);
[d] D. A. Bouboulis, A. Rubinstein, J. Shliozberg, J. Madden and M.
Frieri, Ann. All. Ast. Immu., 74, 491 (1995).
[7] F. M. Gordon, G. L. Simin, C. B. Wosby and J. Mills, Ann.
Inter. Med., 100, 495 (1984).
[8] S. Safrin, D. M. Finkelstein, J. Feinberg, P. Frame, G.
Simpson., A. Wu, T. Cheung, R. Soeiro, P. Hojczk and J. R. Black, Ann.
Intern. Med., 124, 792 (1996).
N-[β-(Ethoxycarbonyl)ethyl]-N-(3,4,5-trimethoxybenzyl)-
glycinate (17).
Ethylbromoacetate 6.75, (40.44 mmole) was added
dropwise over a period of 2 hours at room temperature to a
mixture of 16 10.0 g, (33.7 mmoles) and anhydrous potassium
carbonate (K CO ) 6.28 g, (45.5 mmoles) in 25 ml of
2
3
dimethylformamide. The mixture was stirred at 60 °C for
5 hours after which it was cooled to room temperature and
poured in a thin stream into 20 ml of ice-cold 0.1 N sodium
hydroxide solution. This was then stirred at room temperature
for 4-5 hours (until the characteristic odor of ethyl-
bromoacetate dissipated). The aqueous solution was extracted
with ether (4 x 25 ml) and the combined extracts were washed
with water (4 x 20 ml), dried (magnesium sulfate), filtered and
the ether evaporated under reduced pressure and the resulting
oily product was purified by flash distillation which afforded
11.6 g (90%) of 17 as a pale yellow oil; bp 185-195 °C/0.05
[9] P. D. Walzer, J. Foy, P. Steele and M. White, Antimicrob.
Agents Chemother., 37, 1436 (1993).
[10] C. J. Allegra, J. A. Kovacs, J. C. Drake, J. C. Swan, B. A.
Chabner and H. Masur, J. Exp. Med., 165, 926 (1987).
[11] D. A. Mathews, J. T. Bolin, J. M. Burdge, K. W. Filman, B. T.
Kaufman, C. R. Bedell, J. N. Champness, D. N. Stammers and J. Kraut, J.
Biol. Chem., 260, 381 (1985).
[12] J. N. Champness, A. Achari, S. P. Ballantine, P. K. Bryant, C.
J. Deleves and D. K. Stammers, Structure, 915 (1994).
[13] J. R. Piper, C. A. Johnson, C. A. Krauth, R. L. Carter, C. A.
Hosmer, S. F. Queener, S. E. Borotz and E. R. Pfefferkorn. J. Med.
Chem., 39, 1271 (1996).
mm mercury; tlc R 0.61 (ethyl acetate/1 drop acetic acid,
f
1
silica gel); H nmr (CDCl ): δ 1.26 (t, 6H, CH ), 2.52 (t, 2H,
3
3
N-CH CH ), 3.07 (t, 2H, CH CH COO), 3.77 (s, 2H,
2
2
2
2
NCH CO), 3,85 (m, 9H, OCH ), 4.15 (m, 6H, (OCH ) ,
2
3
2 2
N-CH ), 6.58 (s, 2H, 2',6'-CH).
2
4-(Ethoxycarbonyl)-N-(3',4',5'-trimethoxybenzyl)pyrrolidin-
3-one (18).
[14] A. Gangjee, A. Vasudevan, S. F. Queener and R. L. Kisliuk, J.
Med. Chem., 39, 1438 (1996).
[15] A. Gangjee, Y. Zhu and S. F. Queener, J. Med. Chem., 41,
4533 (1998).
Compound 17 40.0 g, (107.82 mmoles) was added to an ice-
cold slurry of sodium ethoxide 11.32 g, (155.07 mmoles) in
freshly distilled toluene (80 ml) containing 4 Å molecular sieves.
The mixture was stirred at room temperature for 5 hours after
which it was extracted with ice-cold water (4 x 50 ml). The
aqueous layer was washed with ether (4 x 20 ml) and then
neutralized to pH 7 with concentrated hydrochloric acid. The
neutral solution was extracted with ether (4 x 25 ml). The organic
layer was dried (magnesium sulfate) and the ether evaporated to
[16a] A. Rosowsky, C. E. Mota, J. E. Wright and S. F. Queener, J.
Med. Chem., 37, 4522 (1994); [b] A. Gangjee, A. P. Vidwans, A.
Vasudevan, S. F. Queener, R. L. Kisliuk, V. Cody, R. Li, N. Galitsky, J. R.
Luft and W. Pangborn, J. Med. Chem., 41, 3426 (1998).
[17] A. Gangjee, F. Mavandadi and S. F. Queener, J. Med. Chem.,
40, 1173 (1997).
[18] A. Gangjee, X. Guo, S. F. Queener, V. Cody, N. Galitsky, J. R.
Luft and W. Pangborn, J. Med. Chem., 41, 1263 (1998).

220
A. Gangjee, F. Mavandadi and S. F. Queener
Vol. 38
[19] A. Rosowsky, A. T. Papoulis and S. F. Queener, J. Med.
Chem., 40, 3694 (1997).
[28] A. Gangjee, J. Patel, R. L. Kisliuk and Y. Gaumont, J. Med.
Chem., 35, 3678 (1992).
[20] S. F. Queener, J. Med. Chem., 38, 4739 (1995).
[21] A. Gangjee, E. Elzein, M. Kothare and A. Vasudevan, Current
Pharmaceutical Design, 2, 263 (1996).
[22] J. A. Kovacs, C. A. Allegra, J. C. Swan, J. C. Drake, J. E.
Parillo, B. A. Chabner and H. Masur, Antimicrob. Agents Chemother., 32,
430 (1988).
[23] H. Masur, M. A. Polis, C. U. Tuazon, D. Ogata-Arakaki, J. A.
Kovacs, D. Katz, D. Hilt, T. Simmons, I. Feuerstein, B. Lundgren, H. C.
Lane, B. A. Chabner and C. J. Allegra. J. Infect. Dis., 167, 1422 (1992).
[24] A. Gangjee, J. Shi, S. F. Queener, L. R. Barrows and R. L.
Kisliuk, J. Med. Chem., 36, 3437 (1993).
[29] B. S. Hurlbert, K. W. Ledig, P. Stenbuck, B. F. Valenti and G.
H. Hitchings, J. Med. Chem., 11, 703 (1968).
[30] A. Gangjee, I. O. Donkor, R. L. Kisliuk, Y. Gaumont and J.
Thorndike, J. Med. Chem., 34, 611 (1991).
[31] A. Gangjee, J. K. O'Donnell, T. J. Bardos and T. I. Kalman,
J. Heterocyclic Chem., 21, 873 (1984).
[32] J. I. DeGraw, P. H. Christie, W. T. Colwell and F. M. Sirotank,
J. Med. Chem., 35, 320 (1992).
[33] K. A. Watanabe. U.S. Patent 4,925,939, (1990); Chem. Abstr.
113, 152462.
[34] T-L. Su and K. A. Watanabe, J. Org. Chem., 54, 220 (1989).
[35a] M. C. Broughton and S. F. Queener, Antimicrob. Agents
Chemother., 35, 1348 (1991); [b] L-C. Chio and S. F. Queener,
Antimicrob. Agents Chemother., 37, 1914 (1993).
[36] M. C. Broughton and S. F. Queener, Antimicrob. Agents
Chemother., 35, 1348 (1991).
[25] Tripos Associate, Inc., 1699 S. Hanely Road, Suite 303, St.
Louis, MO 63144.
[26] A. Gangjee, J. Shi and S. F. Queener, J. Med. Chem., 40, 1930
(1997).
[27] A. Gangjee, J. Patel and F-T. Lin, J. Heterocyclic Chem., 25,
1597 (1988).