Synthesis of classical and nonclassical, partially restricted, linear, tricyclic 5-deaza antifolates

Article abstract of DOI:10.1021/jm0202369

Seven novel 2,4-diamino-5-deaza-6,7,8,9-tetrahydropyrido[3,4-g]pteridine derivatives 3-9 with different benzyl and a benzoyl substitution at the N7 position were designed and synthesized, as classical and nonclassical, partially restricted, linear tricyclic 5-deaza antifolates. The purpose was to investigate the effect of conformational restriction of the C6-C9 (τ₁) and C9-N10 (τ₂) bonds via an ethyl bridge from the N10 to the C7 position of 5-deaza methotrexate (MTX) on the inhibitory potency against dihydrofolate reductase (DHFR) from different sources and on antitumor activity. The synthetic methodology for most of the target compounds was a concise five-step total synthesis to construct the tricyclic nucleus, 2,4-diamino-5-deaza-7H-6,7,8,9tetrahydropyrido[3,4-g]pteridine (23), followed by regioselective alkylation of the N7 nitrogen. Biological results indicated that this partial conformational modification for the classical analogue N-[4-[(2,4-diamino-5-deaza-6,7,8,9-tetrahydropyrido[3, 4-g]pteridin-7-yl)methyl]benzoyl]-L-glutamic acid 3 was detrimental to DHFR inhibitory activity as well as to antitumor activity compared to MTX or 5-deaza MTX. However, the classical analogue 3 was a better substrate for folypolyglutamate synthetase (FPGS) than MTX. These results show that a classical 5-deaza folate partially restricted via a bridge between the N10 and C7 positions retains FPGS substrate activity and that the antitumor activity of classical tricyclic analogues such as 3 would be influenced by FPGS levels in tumor systems. Interestingly, the nonclassical analogues 4-9 showed moderate to good selectivity against DHFR from pathogenic microbes compared to recombinant human DHFR. These results support the idea that removal of the 5-methyl group of piritrexim along with restriction of τ₁ and τ₂ can translate into selectivity for DHFR from pathogens.

Full text of DOI:10.1021/jm0202369

J . Med. Chem. 2002, 45, 5173-5181
5173
Syn th esis of Cla ssica l a n d Non cla ssica l, P a r tia lly Restr icted , Lin ea r , Tr icyclic
5-Dea za An tifola tes1
Aleem Gangjee,*,† Yibin Zeng,† J ohn J . McGuire,‡ and Roy L. Kisliuk§
Division of Medicinal Chemistry, Graduate School of Pharmaceutical Sciences, Duquesne University,
Pittsburgh, Pennsylvania 15282, Grace Cancer Drug Center, Roswell Park Cancer Institute, Elm and Carlton Streets,
Buffalo, New York 14263, and Department of Biochemistry, Tufts University School of Medicine, Boston, Massachusetts 02111
Received May 30, 2002
Seven novel 2,4-diamino-5-deaza-6,7,8,9-tetrahydropyrido[3,4-g]pteridine derivatives 3-9 with
different benzyl and a benzoyl substitution at the N7 position were designed and synthesized,
as classical and nonclassical, partially restricted, linear tricyclic 5-deaza antifolates. The purpose
was to investigate the effect of conformational restriction of the C6-C9 (τ1) and C9-N10 (τ2)
bonds via an ethyl bridge from the N10 to the C7 position of 5-deaza methotrexate (MTX) on
the inhibitory potency against dihydrofolate reductase (DHFR) from different sources and on
antitumor activity. The synthetic methodology for most of the target compounds was a concise
five-step total synthesis to construct the tricyclic nucleus, 2,4-diamino-5-deaza-7H-6,7,8,9-
tetrahydropyrido[3,4-g]pteridine (23), followed by regioselective alkylation of the N7 nitrogen.
Biological results indicated that this partial conformational modification for the classical
analogue N-[4-[(2,4-diamino-5-deaza-6,7,8,9-tetrahydropyrido[3,4-g]pteridin-7-yl)methyl]ben-
zoyl]-L-glutamic acid 3 was detrimental to DHFR inhibitory activity as well as to antitumor
activity compared to MTX or 5-deaza MTX. However, the classical analogue 3 was a better
substrate for folypolyglutamate synthetase (FPGS) than MTX. These results show that a
classical 5-deaza folate partially restricted via a bridge between the N10 and C7 positions retains
FPGS substrate activity and that the antitumor activity of classical tricyclic analogues such
as 3 would be influenced by FPGS levels in tumor systems. Interestingly, the nonclassical
analogues 4-9 showed moderate to good selectivity against DHFR from pathogenic microbes
compared to recombinant human DHFR. These results support the idea that removal of the
5-methyl group of piritrexim along with restriction of τ1 and τ2 can translate into selectivity
for DHFR from pathogens.
In tr od u ction
Studies of previous conformational restricted analogues
of MTX and 5-deaza MTX have concentrated on tether-
ing the bridge N10 to the N5 or C5 position, which
afforded angular tricyclic analogues.8-10 However, none
of these modified analogues showed better DHFR
inhibitory or antitumor activity than MTX. Linear,
conformationally restricted, tricyclic analogues of MTX
or its pyrido[2,3-d]pyrimidine analogues have not been
explored to any significant extent in the literature as
potential antifolate or antitumor agents. It was there-
fore of interest to evaluate the biological effects of
conformational restriction of the C6-C9 (τ1) and C9-
N10 (τ2) bonds of 5-deaza MTX via an ethyl bridge from
the N10 to the C7 position. The additional carbon
between the tricyclic heterocycle and the phenyl ring
was added to allow conformational flexibility of the
phenyl ring to adopt an optimal orientation for DHFR
inhibitory and antitumor activity. Thus, N-[4-[(2,4-
diamino-5-deaza-6,7,8,9-tetrahydropyrido[3,4-g]pteridin-
7-yl)methyl]benzoyl]-L-glutamic acid (3), a novel, clas-
sical, conformationally restricted, linear, tricyclic 5-deaza
analogue of MTX was designed as an antifolate.
Folate metabolism has been an attractive chemo-
therapeutic target because of its crucial role in the
biosynthesis of nucleic acid precursors.2 Tetrahydro-
folate (FH4), the key component of folate metabolism,
serves as a cofactor to carry one-carbon units for the
biosynthesis of purines and thymidylate (dTMP). During
the synthesis of dTMP, FH4 is oxidized to 7,8-dihydro-
folate (FH2) and must be converted back to the reduced
form by dihydrofolate reductase (DHFR) to continue
DNA synthesis.3 Inhibition of DHFR depletes the FH4
pool and indirectly inhibits DNA synthesis. Thus, DHFR
inhibitors have found clinical utility as antitumor,
antibacterial, and antiprotozoan agents.4,5
Methotrexate (MTX, Figure 1) is a well-known clini-
cally used DHFR inhibitor. Because of the frequent
occurrence of tumor resistance and ineffectiveness
against many solid tumors, extensive structural modi-
fications of MTX have been reported to improve its
antitumor spectrum of activity and to circumvent tumor
resistance.4-7 One area of structural modification that
may afford increased biological activity is conforma-
tional restriction of flexible parts of a lead molecule.
Polyglutamylation via folylpolyglutamate synthetase
(FPGS) is an important mechanism for trapping clas-
sical folates and antifolates within the cell, thus main-
taining high intracellular concentrations and in some
instances increasing binding affinity for folate-depend-
ent enzymes.11a,b However, reduced levels of FPGS in
* To whom correspondence should be addressed. Phone: 412 396-
6070. Fax: 412 396-5593. E-mail: gangjee@duq.edu.
† Duquesne University.
‡ Roswell Park Cancer Institute.
§ Tufts University School of Medicine.
10.1021/jm0202369 CCC: $22.00 © 2002 American Chemical Society
Published on Web 10/01/2002

5174 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 23
Gangjee et al.
F igu r e 1.
tumor cells is a well-defined mechanism of resistance
to classical antifolates such as MTX, which may be
circumvented by antifolates that are not substrates for
FPGS.12,13 Substitutions at the C7 position of deaza
folates with small alkyl groups such as methyl have
been reported to abolish FPGS substrate activity.14
Thus, it was anticipated that compound 3, which is a
7-substituted analogue, would not be an FPGS substrate
and might circumvent resistance in tumor cells that
arise as a result of decreased FPGS activity.
F igu r e 2.
tive, lipophilic antifolates such as TMQ can be selec-
tively circumvented by coadminstration of a reduced
folate, typically leucovorin, which is taken up only by
host cells and hence bypasses DHFR inhibition and
reverses toxicity selectively in host cells. The other
advantage of these lipophilic agents is their penetration
into the central nervous system where T. gondii infec-
tions usually occur.19 The host rescue with leucovorin
has disadvantages that include the fact that leucovorin
is expensive and the occurrence of severe side effects
resulting from ineffective rescue, which sometimes
requires cessation of therapy. Thus, the development
of selective and potent pcDHFR and tgDHFR inhibitors
is of interest.
The rationale for the synthesis of the nonclassical,
partially restricted, linear, tricyclic 5-deaza antifolates,
compounds 4-9 (Figure 2), was based on a comparison
of the X-ray crystallographic structures of human DHFR
and pcDHFR complexed with PTX and TMP. Two major
differences were reported in the active site.20 In the
upper pocket, the amino acid residue responsible for
hydrophobic interaction with the 5-methyl group in PTX
is Val115 in hDHFR, while in pcDHFR it is Ile123.
Thus, the 5-methyl moiety interacts with both hDHFR
as well as pcDHFR. TMQ, which contains a similar
5-methyl moiety, perhaps also interacts with hDHFR
and pcDHFR in a similar fashion. Thus, TMQ and PTX
display much higher DHFR inhibitory activity (against
both hDHFR and pcDHFR) than TMP, which lacks the
5-methyl group. TMP, however, is selective for pcDHFR
but is not potent.
A disadvantage of classical antifolates as antitumor
agents is that they require an active transport mecha-
nism to enter cells, which, when impaired, causes
resistance. In addition, cells that lack these transport
mechanisms, such as bacterial and protozoan cells, are
not susceptible to the action of classical antifolates. In
an attempt to overcome these potential drawbacks,
nonclassical lipophilic antifolates have been developed
as antitumor agents that do not require the folate
transport system(s) and enter cells via diffusion.15
Several lipophilic DHFR inhibitors such as trimetho-
prim (TMP), trimetrexate (TMQ), and piritrexim (PTX)
have also been used in the treatment of opportunistic
infections caused by Pneumocystis carinii (pc) and
Toxoplasma gondii (tg).16-19 These infections are the
principal cause of death in patients with acquired
immunodeficiency syndrome (AIDS). TMP is a weak
inhibitor of pcDHFR and tgDHFR, whereas TMQ and
PTX are highly potent pcDHFR and tgDHFR inhibitors,
with IC50 of 20 nM (TMQ against pcDHFR),17 19.3 nM
(PTX against pcDHFR), and 17.0 nM (PTX against
tgDHFR).16 Unfortunately, both TMQ and PTX inhibit
mammalian DHFR to an equal or greater extent.
Therefore, TMQ and PTX have potential toxicity when
used as single agents to treat opportunistic infections.
The clinical use of nonclassical antifolates to treat P.
carinii and T. gondii infections capitalizes on the fact
that a carrier-mediated active transport system for the
uptake of classical folates and antifolates with polar
glutamate side chains is present in mammalian cells
but is absent in the opportunistic organisms. Lipophilic
nonclassical antifolates penetrate these pathogens by
passive diffusion.17 Thus, the toxic effects of nonselec-
Gangjee et al.21 reported a 5-desmethyl analogue of
PTX, 1, which caused a decrease in potency against both
bacterial DHFR and mammalian DHFR. The decrease
in potency against mammalian DHFR was somewhat
greater and translated into higher selectivity against
pcDHFR and tgDHFR.

5-Deaza Antifolates
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 23 5175
Sch em e 1a
a
Reagents and conditions: (a) Na, EtOH reflux; (b) 75% AcOH, reflux, N2, 5-6 h.
Sch em e 2a
a
Reagents and conditions: (a) benzene/p-TSOH, reflux; (b) THF, -20 °C, 2 h; (c) NH3/MeOH, overnight; (d) DMF, N2,100 °C; (e) 30%
KOH/MeOH, N2, reflux; (f) DMF, Et3N, DMAP, N2, 60-80 °C, 3-4 h; (g) DMF, piv2O, Et3N, DMAP, N2, 60-80 °C, 1.5 h; (h) Me3Sil/
CH3Cl.
In addition, small substitutents such as a methyl
group when placed at the N10 position, as in compound
2, caused some conformational restriction to the flex-
ibility of the C9-N10 bridge, thus decreasing the
number of accessible low-energy conformations com-
pared with 1. Such N10 methyl moieties also increase
hydrophobic interactions, which could increase the
DHFR inhibitory potency while maintaining the selec-
tivity against pcDHFR and tgDHFR.
Like the classical analogue, the nonclassical, partially
restricted linear, tricyclic antifolates have also not been
explored in the literature as potential antitumor or
antibacterial agents. It was thus important to determine
the effects of conformational restriction of τ1 and τ2 in
compounds 1 and 2 with respect to DHFR inhibitory
activity. The absence of the 5-methyl could afford
selectivity against pcDHFR and/or tgDHFR as observed
for 1, while the C7-N10 ethyl bridge provides confor-
mational restriction of the side chain, which may
increase DHFR inhibitory activity and/or selectivity.
Thus, nonclassical 7-substituted 2,4-diamino-5-deaza-
6,7,8,9-tetrahydropyrido[3,4-g]pteridine derivatives 4-9
with different benzyl and a benzoyl substitution at the
N7 position were designed and synthesized. Different
conformations of the phenyl ring in TMP have been
implicated in its selectivity for bacterial DHFR.22 Thus,
the additional carbon between the tricyclic heterocycle
and the phenyl ring was added to allow conformational
flexibility of this part of the molecule to adopt different
orientations.
Ch em istr y
The overall strategy was to synthesize the 2,4-
diamino-5-deaza-7H-6,7,8,9-tetrahydropyrido[3,4-g]pte-
ridine tricyclic system 23 (Scheme 2) from which all the
target compounds could be obtained by regiospecific N7
alkylation.
It was initially anticipated that the condensation of
2,4,6-triaminopyrimidine 12 (Scheme 1) with a biselec-
trophile such as 3-hydroxymethylene-1-benzyl-4-piperi-
done 11 (generated in situ from the 1-benzyl-4-piperi-
done 10 and ethyl formate in the presence of sodium
metal and ethyl ether as previously described23) could
afford the desired tricyclic nucleus in target compounds
such as 4. This, followed by catalytic hydrogenation,
would afford the common intermediate 23 (Scheme 2).
However, in model reactions it was found that purifica-
tion of the annulated product was problematic. Several
attempts to modify the reaction conditions resulted in
poor yields of compound 4 (e30%).
An alternative method (Scheme 2) using modifications
of procedures reported by DeGraw et al.24 and Taylor

5176 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 23
Gangjee et al.
Sch em e 3
regiospecific N7 alkylation became an attractive route.
Several different methods were attempted for removal
of the Cbz group, which included catalytic hydrogena-
tion, catalytic hydrogen transfer, TMSI, acid, and
base.29-33 The most effective method for deprotection
was 30% KOH (aqueous) in methanol. The crucial
common intermediate 23 was isolated in 82% yield.
Regiospecific alkylation of the N7 nitrogen with the
appropriately substituted benzyl halides was carried out
under optimized conditions that included 1.2 equiv of
the benzyl halide, 1.5 equiv of triethylamine, and DMF
in the presence of a catalytic amount of DMAP at 60-
80 °C for 1.5 h. The desired compounds 5-8 and the
precursor to the classical analogue 24 were isolated in
45-63% yield.
et al.,25 where a total synthesis of the tricyclic ring
system was involved, proved to be successful. Thus, the
4-pyrrolidinoenamines of 1-aryl-4-piperidones were con-
verted to 2-amino-3-cyano-6-(4-aryl)-5,6,7,8-tetrahydro-
pyrido[4,3-b]pyridines (19, 20), which proceeded by the
sequence of reactions described in Scheme 2. This
method involved alkylation of an enamine with (chloro-
methylene)malononitrile to generate the 5-substituted
2-amino-3-cyano-5,6,7,8-tetrahydropyrido[4,2-b]pyri-
dines. In the model reaction, the requisite enamine 15
was prepared from 1-benzyl-4-piperidone (10) and pyr-
rolidine via quantitative removal of water by a Dean-
Stark apparatus and evaporation of the solvent.26 The
crude enamine was dissolved in THF and directly
alkylated (without purification) with (chloromethylene)-
malononitrile in THF at -40 °C followed (without
isolation of the intermediate) by treatment with metha-
nolic ammonia to afford 2-amino-3-cyanopyrido[4,3-b]-
pyridine 19 in 45% overall yield (Ar ) unsubstituted
benzyl) starting from the piperidone. Annulation to
generate the 2,4-diaminopyrimidine ring was then car-
ried out with guanidine free base in DMF at 100 °C
under N2 and afforded 4 in 63% yield. Freshly prepared
(chloromethylene)malononitrile 16 was used and was
obtained27 as outlined in Scheme 3 in 40% overall yield.
Attempts at debenzylation28 of 4, which included
catalytic hydrogenation, catalytic hydrogen transfer,
and acid/base, were unsuccessful. Thus, the N-Cbz-
protected 4-piperidone 13 was used in place of benzyl
4-piperidone. Following the same synthetic sequence,
the 2-amino-3-cyanopyrido[4,3-b]pyridine 20 was iso-
lated in 75% yield. Subsequent annulation with guani-
dine afforded 9 in 78% yield.
It was anticipated that protection of the 2,4-diamino
groups of 9 prior to the removal of the N7-Cbz protecting
group would minimize side product formation in the
subsequent alkylation step. Thus, 7-Cbz-2,4-dipivaloyl-
amino-5-deaza-6,7,8,9-tetrahydropyrido[3,4-g]pteri-
dine 21 was prepared from 9 in 92% yield using 4 equiv
of pivaloyl anhydride in DMF, in the presence of a
catalytic amount of DMAP, at 80 °C under N2. The Cbz
protecting group was most effectively removed with
trimethylsilane iodide.29 During the workup process,
however, most of the intermediate was transformed to
the N7-benzyl byproduct (60%) and the desired inter-
mediate 22 was isolated in only 30% yield. Thus, the
direct deprotection of the Cbz group from 9 followed by
Hydrolysis of 24 afforded the corresponding acid 25
(Scheme 4), which was coupled, using a modification of
a previous report,34 to diethyl-L-glutamate followed by
saponification34 to give the target compound 3.
Biologica l Eva lu a tion a n d Discu ssion
Compounds 3-9 were evaluated as inhibitors of
recombinant human (rh), Pneumocystis carinii, Toxo-
plasma gondii, and Escherichia coli (ec) DHFRs.35-38
The inhibitory potencies (IC50) are listed in Table 1 and
compared with MTX, TMP, and TMQ. The classical
compound 3 only marginally inhibited rhDHFR. Against
pcDHFR and ecDHFR, 3 was also much less potent than
MTX. In contrast to MTX, which was equipotent against
tgDHFR and rhDHFR, compound 3 was 6-fold more
potent against tgDHFR compared to rhDHFR. In the
nonclassical series 4-9, except for the ortho-substituted
analogue 7, the other analogues showed some selectivity
against pathogenic DHFR compared to rhDHFR. This
validates the idea that the removal of the 5-methyl
group of PTX and conformational restriction about τ1
and τ2 does afford selectivity against DHFR from
pathogenic microbes. The most potent and selective
tgDHFR inhibitor in this nonclassical series was the
unsubstituted benzyl analogue 4 with an IC50 of 1.4 µM
and a rhDHFR/tgDHFR ratio of 25.6. Substitution on
the phenyl ring by electron-withdrawing Cl or electron-
donating OMe groups resulted in decreased tgDHFR
inhibitory potency as well as selectivity. The most
selective pcDHFR inhibitor in this series was the 3,4-
diCl-benzyl analogue 8 with a rhDHFR/pcDHFR ratio
of >5.2. For ecDHFR, the most potent and selective
inhibitor was the N7-Cbz analogue 9. For rhDHFR
inhibition, both the nature and position(s) of substitu-
tion were important. Thus, the 2,4-diCl analogue 7 was
the most potent. Compared to the benzyl substitution
(4), the benzoyl substituent (9) led to an increase in
Sch em e 4

5-Deaza Antifolates
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 23 5177
Ta ble 1. Inhibitory Concentration (IC50, µM) and Selectivity Ratios against ecDHFR, pcDHFR, and tgDHFR vs rhDHFRa
compd
Ar
ec
rh
rh/ec
6
pc
rh/pc
tg
0.9
1.4
2.9
2.7
2.1
5.1
22
3.4
0.022
0.0051
rh/tg
3
4
5
6
7
8
9
4-benzoyl-L-glu
Bn
0.9
5.4
1.8
3
2
nd
1.48
0.32
>5.2
nd
22.4
20
6
25.6
9.3
8.2
0.62
>5
1.3
100
1
10
8
3
3
3
3
32
27
22
3.2
3.4
7.3
0.43
16
4-OMe-Bn
3,5-diOMe-Bn
2,4-diCl-Bn
3,4-diCl-Bn
Cbz
>15 (35%)
14
4
5
1.3
>26 (33%)
>8.6
10
17000
30
340
>15 (35%)
TMP
MTX
TMQ
0.02
0.006
0.007
15
0.022
0.018
3.7
2.6
0.0011
0.01
1.8
3.6
a
All enzymes were assayed spectroscopically in a solution containing 50 µM dihydrofolate, 80 µM NADPH, 0.05 µM Tris HCl, 0.001 M
2-mercaptoethanol, and 0.001 M EDTA at pH 7.4 and 30 °C. The reaction was initiated with an amount of enzyme yielding a change in
OD at 340 nM of 0.015/min.
Ta ble 2. Activity of 3 as a Substrate for Human FPGSa
Ta ble 3. Growth Inhibition of CCRF-CEM Human Leukemia
Cells and Its Methotrexate-Resistant Sublines (EC50, nM)
during Continuous Exposure (0-120 h)a
substrate
Km (µM)
Vmax, rel
Vmax/Km
n
aminopterin
3
MTX
2.15 ( 0.55
27.30 ( 0.8
49b
1.00 ( 0.00
1.07 ( 0.03
0.82c
0.50 ( 0.13
0.04 ( 0.00
0.02c
2
2
cell line
resistance mechanism
MTX
3
CCRF-CEM sensitive
12.0 ( 3.1 771 ( 111
(n ) 4)
(n ) 3)
>10000
(n ) 2)
>10000
a
1000 units of FPGS activity; one unit of FPGS catalyzes the
incorporation of 1 pmol of [3H]Glu/h. Values presented are average
( range. Vmax values are calculated relative to aminopterin within
R1
increased DHFR
595 ( 85
(n ) 2)
decreased in transport 1700 ( 0
(n ) 2)
R2
b
the same experiment. Data derived from ref 12. c Data derived
(n ) 2)
1900 ( 0
(n ) 2)
from ref 13.
R30dm
decreased FPGS
16 ( 0
(n ) 2)
inhibitory potency and selectivity against ecDHFR. In
contrast, against tgDHFR, the benzyl analogue 4 was
more potent and selective. These data suggest that the
basicity of the N7 nitrogen, along with the conforma-
tional flexibility of the phenyl side chain, may play a
role in the inhibitory potency against DHFR from
different pathogenic microbes and that the effects are
quite different against different DHFR. Whether this
difference in activity and selectivity of the benzoyl-
substituted analogue compared to the benzyl analogue
is due to the basicity of N7 or to the presence of the
amide carbonyl and/or its ability to restrict the confor-
mation of the side chain or other factors is not known
at present.
a
EC50 is the concentration of drug required to decrease cell
growth by 50% relative to untreated control after 5 days of
treatment.
Ta ble 4. Protection of CCRF-CEM Cells from the Growth
Inhibitory Effects of MTX and 3 by Leucovorin (LV)a
relative growth (% of control)
drug conc [LV] ) 0 [LV] ) 0.1 µM [LV] ) 1 µM [LV] ) 10 µM
MTX 30 nM
40 nM
2.8
3.7
9.5
4.3
95
74
101
109
92
75
107
112
91
77
101
105
3
2 µM
5 µM
a
LV up to 10 µM itself had no effect on growth of CCRF-CEM
cells. Values are a range of closely agreeing duplicates. The entire
experiment was repeated with similar results.
Since polyglutamylation plays a role in the mecha-
nism of action of some classical antifolates,11a,b it was
of interest to evaluate the classical analogue 3 as a
substrate for human FPGS (Table 2). Compound 3 was
a reasonable substrate for FPGS, albeit less so than
aminopterin, primarily because of a high Km. This
suggests that catalysis by FPGS is relatively insensitive
to conformational restriction in the bridge region. Since
compound 3 is substituted at the C7 position, its ability
to function as a substrate for hFPGS was surprising,
since 7-CH3-substituted folate analogues are generally
not substrates for FPGS.14,39a,b
Compound 3 was also evaluated as an inhibitor of the
growth of CCRF-CEM human leukemia cells and its
MTX-resistant sublines in culture during continuous
exposure (Table 3).40-43 Compound 3 was a weak
inhibitor of CCRF-CEM growth in continuous (120 h)
exposure; it was 60-fold less potent than MTX. Against
R30dm, a subline expressing decreased FPGS activity,
compound 3 was 3-fold cross-resistant in continuous
exposure, which suggests that polyglutamylation plays
a role in its mechanism of action. The data are consis-
tent with the data in Table 2 showing that 3 is a
substrate for FPGS. Although DHFR is often insensitive
to the polyglutamylation status of substrates and in-
hibitors,44 enhanced potency of polyglutamates of a few
novel antifolates against DHFR has been reported.45
A
DHFR-overexpressing subline (R1) and a subline with
reduced MTX transport (R2) were >14-fold cross-
resistant to compound 3. Despite its relatively lower
potency (compared to MTX) as an inhibitor of purified
rhDHFR (Table 1), the data are consistent with some
need for polyglutamylation, transport via the reduced
folate carrier, and inhibition of DHFR.
Leucovorin protection studies were carried out with
3 in order to further elucidate its mechanism of action.
Both 3 (at 2 or 5 µM) and MTX (at 30 and 40 nM), at
drug levels that inhibited growth by 92-96%, were fully
protected by as little as 0.1 µM leucovorin, indicating
that 3 acts as an antifolate (Table 4), which is consistent
with the cell culture results and its DHFR inhibitory
activity.
Compounds 3, 4, and 9 were evaluated as inhibitors
of the growth of tumor cells in culture in the National
Cancer Institute in vitro preclinical antitumor screening
program.46 In the leukemia cell lines, these analogues
showed inhibitory activities with GI50 values of 10-6
-
10-7 M, while in all the other tumor cell lines the
compounds did not exhibit significant cytotoxic effect
at 10-5 M.

5178 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 23
Gangjee et al.
perature and was stirred under N2 for 6 h. The reaction
mixture was then subjected to another cycle of activation and
coupling using half the quantities listed above. The reaction
mixture was slowly allowed to warm to room temperature, was
stirred under nitrogen for 24 h, and then was subjected to a
third round of activation and coupling using the same the
quantities as the second round and was stirred for an ad-
ditional 24 h. TLC showed the formation of one major spot at
Rf ) 0.38 (MeOH/CHCl3, 1:5). The reaction mixture was
evaporated to dryness under reduced pressure. The residue
was dissolved in a minimum amount of CH3Cl/MeOH, 4:1, and
chromatographed on a silica gel column (2 cm × 15 cm) and
with 4% MeOH in CHCl3 as the eluent. Fractions that showed
the desired single spot at Rf ) 0.38 were pooled and evaporated
to afford 26 (120 mg) as a syrup. Attempts at crystallization
of this syrup were unsuccessful, and it was used directly for
the next step.
To a solution of the diester 26 (120 mg, 0.2 mmol) in MeOH
(10 mL) was added 1 N NaOH (6 mL), and the mixture was
stirred under nitrogen at room temperature for 16 h. TLC
showed the disappearance of the starting material (Rf ) 0.38)
and one major spot at the origin (MeOH/CHCl3, 2:5). The
reaction mixture was evaporated to dryness under reduced
pressure. The residue was dissolved in water (10 mL), the
resulting solution was cooled in an ice bath, and the pH was
adjusted to 3 with dropwise addition of 1 N HCl. The resulting
suspension was frozen in a dry ice/acetone bath, thawed in
the refrigerator at 4-5 °C, and filtered. The residue was
washed with a small amount of cold water and ethyl acetate
and was dried in vacuo over P2O5 to afford 100 mg (56%) of 3
as a brown powder: mp >300 °C (dec); TLC Rf ) 0.22 (MeOH/
CHCl3, 2:5); 1H NMR (DMSO-d6) δ 2.11-2.16 (2 sets of t, 2 H,
Glu â-CH2), 2.32-2.37 (t, 2 H, Glu γ-CH2), 2.81-2.83 (t, 2 H,
C9-CH2), 2.88-2.90 (t, 2 H, C8-CH2), 3.54 (s, 2 H, C6-CH2),
3.76 (s, 2 H, benzylic-CH2), 4.41 (m, 1 H, Glu R-CH), 6.17 (s,
2 H, 2-NH2), 7.29 (s, 2 H, 4-NH2), 7.52-7.55 (d, 2 H, C6H4),
7.93-7.96 (d, 2 H, C6H4), 8.00 (s, 1 H, 5-CH), 8.59-8.61 (d, 1
H, -CONH-), 12.14 (br, 2 H, 2 × COOH). Anal. (C23H25N7O5‚
1.5HCl‚0.3CH3COOEt) C, H, N, Cl.
In summary, seven novel 2,4-diamino-5-deaza-6,7,8,9-
tetrahydropyrido[3,4-g]pteridine derivatives with a var-
ied benzyl and a benzoyl substitution at the N7 position
(3-9) were designed and synthesized as classical and
nonclassical, conformationally restricted, linear, tricyclic
5-deaza antifolates to investigate the effect of the
conformational restriction, of C6-C9 (τ1) and C9-N10
(τ2) bonds via an ethyl bridge from the N10 to the C7
position of 5-deaza MTX, with respect to the inhibitory
potency against different pathogenic DHFR and anti-
tumor activity. Biological studies indicated that this
conformational modification was detrimental to DHFR
inhibitory activity as well as to antitumor activity
compared to MTX or 5-deaza MTX. Surprisingly, the
classical compound 3 was a better substrate than MTX
for FPGS. These results indicate that the classical
5-deaza folate, tethered at the N10 and C7 positions,
retains FPGS substrate activity and that the antitumor
activity of tricyclic analogues such as 3 could be
influenced by FPGS levels in tumor systems. Interest-
ingly, the nonclassical analogues 4-9 showed moderate
to good selectivity against pathogenic DHFR compared
to rhDHFR. These results further support the idea that
removal of the 5-methyl group of PTX along with τ1 and
τ2 restriction can afford selectivity against pathogenic
DHFR vs rhDHFR.
Exp er im en ta l Section
All evaporations were carried out in vacuo with a rotary
evaporator. Analytical samples were dried in vacuo (0.2
mmHg) in an Abderhalden drying apparatus over P2O5 and
refluxing ethanol. Melting points were determined on a MEL-
TEMP II melting point apparatus with a FLUKE 51 K/J
electronic thermometer and are uncorrected. Nuclear magnetic
resonance spectra for proton NMR were recorded on a Bruker
WH-300 (300 MHz) spectrometer. The chemical shift values
are expressed in ppm (parts per million) relative to tetra-
methylsilane as internal standard; s ) singlet, d ) doublet, t
) triplet, q ) quartet, m ) mutiplet, br ) broad singlet. The
relative integrals of peak areas agreed with those expected
for the assigned structures. Thin-layer chromatography (TLC)
was performed on POLYGRAM Sil G/UV254 silica gel plates
with fluorescent indicator, and the spots were visualized under
254 and 366 nm illumination. Proportions of solvents used for
TLC are by volume. Column chromatography was performed
on 230-400 mesh silica gel purchased from Aldrich, Milwau-
kee, WI. Elemental analyses were performed by Atlantic
Microlab, 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 could not be prevented despite 24-48 h of drying
in vacuo and were confirmed, where possible, by their presence
in the 1H NMR spectra. All solvents and chemicals were
purchased from Aldrich Chemical Co. and Fisher Scientific and
were used as received.
N-[4-[(2,4-Dia m in o-5-d ea za -6,7,8,9-t et r a h yd r op yr id o-
[3,4-g]p ter id in -7-yl)m eth yl]ben zoyl]-L-glu ta m ic Acid (3).
To a solution of 4-(2,4-diamino-5-deaza-6,7,8,9-tetrahydropy-
rido[3,4-g]pteridin-7-yl)methyl benzoic acid (25) (130 mg, 0.37
mmol) in anhydrous DMF (9 mL) was added triethylamine
(130 µL), and the mixture was stirred under N2 at room
temperature for 5 min. The resulting solution was cooled to 0
°C, isobutyl chloroformate (130 µL, 0.96 mmol) was added, and
the mixture was stirred at 0 °C for 30 min. At this time TLC
(MeOH/CHCl3, 2:5) indicated the formation of the activated
intermediate at Rf ) 0.40 and the disappearance of the starting
acid at Rf ) 0.28. Diethyl-L-glutamate hydrochloride (240 mg,
0.96 mmol) was added to the reaction mixture followed
immediately by triethylamine (130 µL, 0.96 mmol). The
reaction mixture was slowly allowed to warm to room tem-
2,4-Diam in o-7-ben zyl-5-deaza-6,7,8,9-tetr ah ydr opyr ido-
[3,4-g]p ter id in e.4 Meth od A. Guanidine hydrochloride (1.50
g, 12 mmol) was added in one portion to a slurry of 1.5 g (13.6
mmol) of potassium tert-butoxide in 6 mL of anhydrous DMF
under N2, and the mixture was stirred at room temperature
for 1 h. To this solution was added 2-amino-3-cyano-6-4-
benzyl-5,6,7,8-terahydropyrido[4,3-b]pyridine 19 (1.4 g, 4 mmol)
in one portion, and the resulting mixture was heated at 100
°C under N2 with stirring overnight. The reaction mixture was
cooled and filtered, the yellow residue was washed with hot
water (20 mL) followed by methanol (10 mL × 2), acetone (20
mL), and ethyl acetate (20 mL) and was dried in vacuo to give
1 g (63%) of 4 as a light-yellow powder: mp 290.5-292 °C; Rf
1
) 0.30 (MeOH/CHCl3, 2:5); H NMR (DMSO-d6) δ 2.80-2.82
(t, 2 H, C8-CH2), 2.87-2.89 (t, 2 H, C9-CH2), 3.51 (s, 2 H,
C6-CH2), 3.68 (s, 2 H, benzylic-CH2), 6.15 (s, 2H, 2-NH2), 7.29
(s, 2H, 4-NH2), 7.30-7.37 (m, 5 H, C6H5), 7.99 (s, 1 H, C5-
CH). Anal. (C18H18N6O2‚0.4H2O) C, H, N.
Met h od B. In a 250 mL three-neck flask (freshly dried)
equipped with a drying tube was placed sodium metal (0.56
g, 25.5 mmol) that had been cut into small pieces (<1 cm3).
Anhydrous ether (50 mL) was added followed by ethyl formate
(2.8 g, 38 mmol). To this mixture was added 1-benzyl-4-
piperidone 10 (4.8 g, 25.5 mmol). The mixture was then cooled
to 5 °C using an ice bath. The reaction was initiated by the
addition of 0.2 mL of absolute ethanol, and the mixture was
stirred for 6 h. The yellow solid formed was filtered and
washed with ethyl ether (100 mL) and dried in vacuo under
P2O5 overnight to give 5.2 g (88%) of 1123 as a yellow powder.
This powder was used directly for the next step. 1H NMR (D2O)
δ 2.42-2.45 (t, 2 H, C-CH2-N), 2.57-2.60 (t, 2 H, OdC-CH2),
3.20 (s, 2 H, CdC-CH2), 3.58 (s, 2 H, benzylic-CH2), 7.47 (br,
5 H, C6H5), 9.17 (s, 1 H, dCH-ONa).
2,4,6-Triaminopyrimidine 12 (2.5 g, 20 mmol) in 100 mL of
75% AcOH (aqueous, w/w) was heated to 100 °C, the resulting

5-Deaza Antifolates
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 23 5179
clear solution was cooled to 70 °C, and 11 (4.8 g, 20 mmol)
was added in portions over 30 min. The resulting mixture was
refluxed at 130 °C (oil bath) for 5 h. Distilled water (50 mL)
was added, and the resulting mixture was cooled overnight in
a refrigerator. The pH was adjusted to 7 using an ice bath
with concentrated NH4OH. The precipitate formed was col-
lected and washed with water and MeOH and was dried in
vacuo. The solid obtained was flash-chromatographed through
a silica gel column (4 cm × 20 cm) using 10% MeOH (saturated
with ammonia) in CHCl3 as the eluent. The desired fractions
(TLC) was pooled and evaporated. The residue was triturated
with MeOH and filtered to afford 1 g (17%) of 4 as a yellow
power that was identical in every respect to the sample
prepared in method A.
Gen er a l P r oced u r e for th e Syn th esis of 2,4-Dia m in o-
7-(su bstitu ted ben zyl)-5-d ea za -6,7,8,9-tetr a h yd r op yr id o-
[3,4-g]p ter id in es (5-8 a n d 24). Freshly dried 2,4-diamino-
7H-5-deaza-6,7,8,9-tetrahydropyrido[3,4-g]pteridine 23 (110
mg, 0.5 mmol), DMF (5 mL), 4-N,N-dimethylaminopyridine
(DMAP) (20 mg), triethylamine (0.2 mL), and the correspond-
ing benzyl halide (0.6 mmol) were added in one portion to a
100 mL flask. The resulting mixture was stirred under N2 at
60-80 °C (oil bath) for 2-3 h until the starting material
disappeared (TLC). After evaporation of the solvent under
reduced pressure, silica gel (5 g) and methanol (20 mL) were
added and the resulting suspension was evaporated to dryness
under reduced pressure. This silica gel plug was loaded on a
dry silica gel column (3 cm × 15 cm) and was flash-chromato-
graphed initially with CHCl3 (100 mL), then sequentially with
200 mL of 5% MeOH (saturated with ammonia) in CHCl3 and
200 mL of 10% and 15% MeOH (saturated with ammonia) in
CHCl3. Fractions that showed the desired product spot were
pooled and evaporated to dryness. The residue was recrystal-
lized from ethyl acetate and MeOH to afford the desired 2,4-
diamino-7-(substituted benzyl)-5-deaza-6,7,8,9-tetrahydropyrido-
[3,4-g]pteridines (5-8, 24) in 45-64% yield.
7.59-7.64 (m, 2 H, C6H3), 8.02 (s, 1 H, 5-CH). Anal. (C17H16N6-
Cl2‚0.3H2O) C, H, N, Cl).
2,4-Dia m in o-7-ca r boben zyloxy-5-d ea za -6,7,8,9-tetr a h y-
d r op yr id o[3,4-g]p t er id in e (9). Guanidine hydrochloride
(1.50 g, 12 mmol) was added in one portion to a slurry of
potassium tert-butoxide (1.5 g, 13.6 mmol) in anhydrous DMF
(6 mL) under N2, and the mixture was stirred at room
temperature for 1 h. To this solution was added 2-amino-3-
cyano-6-(4-Cbz)-5,6,7,8-terahydropyrido[4,3-b]pyridine 20 (1.4
g, 4 mmol) in one portion, and the resulting mixture was
heated at 100 °C under N2 with stirring overnight. The
reaction mixture was cooled and filtered, and the yellow
residue was washed with hot water (20 mL) followed by
methanol (10 mL × 2), acetone (20 mL), and ethyl acetate (20
mL) to give 1.1 g (78%) of 9 as a light-yellow powder: mp
290.5-292 °C; TLC Rf ) 0.28 (MeOH/CHCl3, 1:7); 1H NMR
(DMSO-d6) δ 2.80-2.82 (br, 2 H, C9-CH2), 3.74 (br, 2 H, C8-
CH2), 4.60 (s, 2 H, C6-CH2), 5.14 (s, 2 H, benzylic-CH2), 6.25
(s, 2 H, 4-NH2), 7.32-7.39 (m, 7 H, C6H5 and 2-NH2), 8.16 (s,
1 H, 5-CH). Anal. (C18H18N6O2) C, H, N.
2,4-Diam in o-7-(4-car bom eth oxyben zyl)-5-deaza-6,7,8,9-
tetr a h yd r op yr id o[3,4-g]p ter id in e (24). Compound 23 (110
mg, 0.5 mmol) and 4-carbomethoxybenzyl bromide (135 mg,
0.6 mmol) following the general procedure described earlier
afforded 110 mg (60%) of 24 as a yellow solid: mp 278-281
°C (dec); TLC Rf ) 0.42 (CHCl3/MeOH, 5:2); 1H NMR (DMSO-
d6) δ 2.81-2.83 (t, 2 H, C9-CH2), 2.88-2.90 (t, 2 H, C8-CH2),
3.54 (s, 2 H, C6-CH2), 3.76 (s, 2 H, benzylic-CH2), 3.85 (s, 3
H, OCH3), 6.17 (s, 2 H, 2-NH2), 7.29 (s, 2 H, 4-NH2), 7.52-
7.55 (d, 2 H, C6H4), 7.93-7.96 (d, 2 H, C6H4), 8.00 (s, 1 H,
5-CH). Anal. (C19H20N6O2‚0.2EtOAc) C, H, N.
2-Am in o-3-cya n o-6-(4-ben zyl)-5,6,7,8-ter a h yd r op yr id o-
[4,3-b]p yr id in e (19). To 1-benzyl-4-piperidone (10) (3.78 g,
20 mmol) and pyrrolidine (5 mL, 60 mmol) in 20 mL of
anhydrous benzene was added 30 mg of p-toluenesulfonic acid,
and the resulting mixture was stirred under reflux overnight
using a Dean-Stark apparatus to remove the water (∼0.4
mL).26 After evaporation of the solution, the residue, 1-benzyl-
4-(N-pyrrolidino)-1,2,5,6-tetrahydropyridine 14, was dissolved
in 100 mL of THF and was used directly for the next step. To
this solution was added triethylamine (3.5 mL), and the
reaction mixture cooled to -40 °C in a dry ice/methanol bath.
To this solution was added, dropwise, a solution of freshly
prepared 1-chloro-2,2-dicyanoethylene (16)27 (2.82 g, 25 mmol)
in THF (5 mL) over a period of 2 min. The reaction mixture
rapidly turned deep-red. The mixture was stirred at -40 °C
for 30 min and then at room temperature for 1 h and was
filtered through Celite, and the residue washed with copious
amounts of THF and methylene chloride. Evaporation of the
filtrate gave a deep-red residue (17), which was suspended in
100 mL of saturated methanolic ammonia. The mixture was
stirred at room temperature overnight to give a pink suspen-
sion that was filtered, and to the filtrate was added silica gel
(5 g). The solvent was evaporated to afford a plug. This plug
was loaded onto a silica gel column (3 cm × 5 cm) and was
eluted with hexane/ethyl acetate (2:1). The desired fractions
(TLC) were pooled and evaporated and the residue was
recrystallized from ethyl acetate to afford 2.4 g (45%) of 19 as
a yellow powder: mp 211.3-213.2 °C; Rf ) 0.33 (hexane/ethyl
acetate, 1:1); MS m/z 264 (M+); 1H NMR (DMSO-d6) δ 2.70 (t,
2 H, C8-CH2), 3.31-3.36 (m, 4 H, C6-CH2, C5-CH2), 3.62
(s, 2 H, benzylic-CH2), 6.63 (s, 2H, 2-NH2), 7.27-7.34 (m, 5 H,
C6H5), 7.56 (s, 1 H, 4-CH). Anal. (C16H16N4) C, H, N.
2,4-Dia m in o-7-(4-m et h oxyb en zyl)-5-d ea za -6,7,8,9-t et -
r a h yd r op yr id o[3,4-g]p ter id in e (5). Compound 23 (110 mg,
0.5 mmol) and 4-methoxylbenzyl chloride (95 mg, 0.6 mmol)
afforded 80 mg (48%) of 5 as a light-yellowish solid: mp 257.2-
259.6 °C; TLC Rf ) 0.32 (CHCl3/MeOH, 5:2); 1H NMR (DMSO-
d6) δ 2.80 (t, 2 H, C9-CH2), 2.88 (t, 2 H, C8-CH2), 3.48 (s, 2
H, C6-CH2), 3.60 (s, 2 H, benzylic-CH2), 3.74 (s, 3 H, OCH3),
6.25 (s, 2 H, 2-NH2), 6.88-6.91 (d, 2 H, C6H4), 7.26-7.29 (d, 2
H, C6H4), 7.34 (s, 2 H, 4-NH2), 8.00 (s, 1 H, 5-CH). Anal.
(C18H20N6O‚0.3H2O) C, H, N.
2,4-Dia m in o-7-(3,5-d im eth oxyben zyl)-5-d ea za -6,7,8,9-
tetr a h yd r op yr id o[3,4-g]p ter id in e (6). Compound 23 (110
mg, 0.5 mmol) and 3,5-dimethoxylbenzyl chloride (110 mg, 0.6
mmol) afforded 105 mg (57%) of 6 as a light-yellow solid: mp
1
263.2-265.7 °C; TLC Rf ) 0.35 (CHCl3/MeOH, 5:2); H NMR
(DMSO-d6) δ 2.79-2.81 (t, 2 H, C9-CH2), 2.88-2.90 (t, 2 H,
C8-CH2), 3.53 (s, 2 H, C6-CH2), 3.60 (s, 2 H, benzylic-CH2),
3.73 (s, 6 H, OCH3), 6.23 (s, 2 H, 2-NH2), 6.40 (s, 2 H, C6H3),
6.54 (s, 2 H, C6H3), 7.33 (s, 2 H, 4-NH2), 8.02 (s, 1 H, 5-CH).
Anal. (C19H22N6O2‚0.3 MeOH) C, H, N.
2,4-Dia m in o-7-(2,4-d ich lor oben zyl)-5-d ea za -6,7,8,9-tet-
r a h yd r op yr id o[3,4-g]p ter id in e (7). Compound 23 (110 mg,
0.5 mmol) and 2,4-dichlorobenzyl chloride (115 mg, 0.6 mmol)
afforded 120 mg (64%) of 7 as a yellow solid: mp 268.1-270
°C (dec); TLC Rf ) 0.39 (CHCl3/MeOH, 5:2); 1H NMR (DMSO-
d6) δ 2.84 (t, 2 H, C9-CH2), 2.90 (t, 2 H, C8-CH2), 3.61 (s, 2
H, C6-CH2), 3.76 (s, 2 H, benzylic-CH2), 6.24 (s, 2 H, 2-NH2),
7.37 (s, 2 H, 4-NH2), 7.43-7.46 (m, 1 H, C6H3), 7.57-7.63 (m,
2 H, C6H3), 8.03 (s, 1 H, 4-CH). Anal. (C17H16N6Cl2‚0.2H2O) C,
H, N, Cl).
2,4-Dia m in o-7-(3,4-d ich lor oben zyl)-5-d ea za -6,7,8,9-tet-
r a h yd r op yr id o[3,4-g]p ter id in e (8). Compound 23 (110 mg,
0.5 mmol) and 3,4-dichlorobenzyl chloride (115 mg, 0.6 mmol)
afforded 85 mg (45%) of 8 as a light-yellow solid: mp 258.1-
261 °C (dec); TLC Rf ) 0.40 (CHCl3/MeOH, 5:2); 1H NMR
(DMSO-d6) δ 2.83 (t, 2 H, C9-CH2), 2.90 (t, 2 H, C8-CH2),
3.53 (s, 2 H, C6-CH2), 3.69 (s, 2 H, benzylic-CH2), 6.22 (s, 2
H, 2-NH2), 7.33 (s, 2 H, 4-NH2), 7.37-7.40 (m, 1 H, C6H3),
2-Am in o-3-cya n o-6-(4-ca r boben zyloxy)-5,6,7,8-ter a h y-
d r op yr id o[4,3-b]p yr id in e (20). To benzyl 4-oxo-1-piperidi-
necarboxylate 13 (6 g, 20 mmol) and pyrrolidine (5 mL, 60
mmol) in 20 mL of anhydrous benzene was added 30 mg of
p-toluenesulfonic acid, and the resulting mixture was stirred
under reflux overnight using a Dean-Stark apparatus to
remove the water (∼0.4 mL). After evaporation of the solution,
the residue 15 was dissolve in 100 mL of THF and was directly
used for the next step. To this solution was added triethyl-
amine (3.5 mL), and the reaction mixture was cooled to -40
°C in a dry ice/methanol bath. To this solution was added,

5180 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 23
Gangjee et al.
dropwise, a solution of freshly prepared 16 (2.82 g, 25 mmol)
in THF (5 mL) over a period of 2 min, followed by the same
process as described for the synthesis of 19. Evaporation of
the filtrate gave a deep-red residue (18) that was suspended
in 100 mL of saturated methanolic ammonia. The mixture was
stirred at room temperature overnight to give a pink suspen-
sion. The solid collected by filtration was washed with a small
amount of methanol followed by ethyl acetate to afford 4.6 g
(75%) of 20 as a light-yellow powder: mp 167.3-169.1 °C; TLC
Rf ) 0.68 (MeOH/CHCl3, 1:5); 1H NMR (DMSO-d6) δ 2.69-
2.73 (t, 2 H, C8-CH2), 3.67 (t, 2 H, C6-CH2), 4.48 (s, 2 H,
C5-CH2), 5.11 (s, 2 H, benzylic-CH2), 6.76 (s, 2 H, 2-NH2),
7.30-7.38 (m, 5 H, C6H5), 7.75 (s, 1 H, 4-CH). Anal.
(C17H16N4O2‚0.3H2O) C, H, N.
2,4-Dia m in o-7H-5-d ea za -6,7,8,9-tetr a h yd r op yr id o[3,4-
g]p ter id in e (23). A suspension of 9 (540 mg, 1.5 mmol) and
40% KOH (20 mL) in MeOH (30 mL) was refluxed under N2
for 2 h to give a clear solution. TLC indicated the disappear-
ance of the starting material, Rf ) 0.28 (MeOH/CHCl3, 1:7),
and a major spot near the origin. After evaporation of half the
amount of solvent, the resulting precipitate was collected
through filtration and washed with small amounts of ice water,
MeOH, and ethyl acetate and was dried in vacuo over P2O5 to
afford 270 mg (82%) of 23 as a white powder: mp 271.5-273.1
°C (dec); TLC Rf ) 0.03 (MeOH/CHCl3, 2:5) δ 2.76 (t, 2 H, C9-
CH2), 3.02 (t, 2 H, C8-CH2), 3.84 (s, 2 H, C6-CH2), 6.16 (s, 2
H, 2-NH2), 7.38 (s, 2 H, 4-NH2), 7.80 (s, 1 H, 7-NH), 8.00 (s, 1
H, 4-CH). Anal. (C10H12N6‚0.1H2O) C, H, N.
4-(2,4-Dia m in o-5-d ea za -6,7,8,9-tetr a h yd r op yr id o[3,4-g]-
p ter id in -7-yl)m eth ylben zoic Acid (25). To a solution of 24
(180 mg, 0.5 mmol) in MeOH/DMSO (2:1, 25 mL) was added
2 N NaOH (5 mL), and the mixture was stirred under N2 at
room temperature for 16 h. TLC showed the disappearance of
the starting material (Rf ) 0.42) and one major spot at the
origin (MeOH/CHCl3, 2:5). The reaction mixture was evapo-
rated to dryness under reduced pressure. The residue was
dissolved in distilled water (8 mL), and the solution was
filtered through Celite and washed with water (5 mL). The
filtrate was cooled in an ice bath and the pH adjusted to 4 by
dropwise addition of 2 N HCl. The resulting suspension was
frozen in a dry ice/acetone bath, thawed in the refrigerator at
4-5 °C, and filtered. The residue was washed with a small
amount of cold water and ethyl acetate and was dried in vacuo
over P2O5 to afford 150 mg (85%) of 25 as a yellow powder:
mp >300 °C (dec); TLC Rf ) 0.28 (MeOH/CHCl3, 2:5); 1H NMR
(DMSO-d6) δ 2.82-2.84 (t, 2 H, C9-CH2), 2.89-2.91 (t, 2 H,
C8-CH2), 3.56 (s, 2 H, C6-CH2), 3.79 (s, 2 H, benzylic-CH2),
6.18 (s, 2 H, 2-NH2), 7.25 (s, 2 H, 4-NH2), 7.54-7.57 (d, 2 H,
C6H4), 7.95-7.98 (d, 2 H, C6H4), 8.01 (s, 1 H, 5-CH), 12.03 (br,
1 H, COOH). Anal. (C18H18N6O2‚0.2H2O) C, H, N.
F olylp olyglu ta m a te Syn th eta se (F P GS) P u r ifica tion
a n d Assa y. Recombinant human FPGS was expressed in E.
coli J M109 (λDE3) cells from a pET3A expression vector
containing the entire open reading frame of the human
cytosolic FPGS.48 This expression vector was constructed in,
and was a generous gift of, the laboratory of Dr. Barry Shane,
Department of Nutrition, University of California, Berkeley.
The expressed protein was purified by Affi-Gel Blue and size
exclusion chromatography.48 Substrate activity of 3 was
determined as previously described49-51 except that the DEAE
cellulose columns used to separate 3H-radiolabeled poly-
glutamate products from unreacted [3H]glutamate were eluted
with 10 mM Tris-HCl (pH 7.5)/25 mM 2-mercaptoethanol
containing 95 mM NaCl (instead of the standard 110 mM
NaCl). This change was required to ensure the quantitative
retention of all polyglutamates of 3. Kinetic constants were
determined for 3 and aminopterin in the same experiment by
nonlinear curve-fitting to the rectangular hyperbola function
(SigmaPlot); initial estimates of parameters were based on
visual inspection of kinetic data.
nm (45 500), λmax-2 ) 322 nm (11 600); pH 7, λmax ) 227 nm
(37 600); pH 13, λmax-1 ) 229 nm (37 500), λmax-2 ) 243 nm
(38 500)). Extinction coefficients for methotrexate (MTX), a
generous gift of Immunex (Seattle, WA), were from the
literature.47
All cell lines were verified to be negative for Mycoplasma
contamination using the GenProbe test kit. The human
T-lymphoblastic leukemia cell line CCRF-CEM40 and its
methotrexate-resistant sublines R1,41 R2 (Bos),42 and R30dm43
used in these studies were cultured as described.43 R1 ex-
presses 20-fold elevated levels of dihydrofolate reductase
(DHRF), the target enzyme of MTX. R2 has dramatically
reduced MTX uptake. R30dm expresses only 1% of the FPGS
activity of CCRF-CEM and is resistant to short-term, but not
continuous, MTX exposure; however, R30dm is generally cross-
resistant in continuous exposure to antifolates requiring
polyglutamylation to form potent inhibitors. EC50 values were
interpolated from plots of percent control growth versus
logarithm of drug concentration.
Protection against growth inhibition of CCRF-CEM cells was
assayed by including leucovorin (0.1-10 µM) simultaneously
with a concentration of drug previously determined to inhibit
growth by 90-95%; the remainder of the assay was as
described. Growth inhibition was measured relative to the
appropriate metabolite-treated control; metabolites caused no
growth inhibition in the absence of drug, however.
Ack n ow led gm en t. This work was supported in part
by NIH Grants AI 41743 (A.G.) and AI 44661 (A.G.)
from the National Institute of Allergy and Infectious
Diseases and in part by Grants CA89300 (A.G.), CA43500
(J .J .M.), CA 16056 (J .J .M.), and CA10914 (R.L.K.) from
the National Cancer Institute. The authors thank Dr.
Barry Shane (University of California, Berkeley) for
providing the expression vector for human FPGS.
Refer en ces
(1) Presented in part at the 219th National Meeting of the American
Chemical Society, San Francisco, CA, April 2000; Paper MEDI
89. Presented in part at the 92nd National Meeting of the
American Association for Cancer Research, New Orleans, LA,
April 2001; Abstract 1568.
(2) Berman, E. M.; Werbel, L. M. The Renewed Potential for Folate
Antagonists in Contemporary Cancer Chemotherapy. J . Med.
Chem. 1991, 34, 479-485.
(3) Kisliuk, R. L. The Biochemistry of Folates. In Folate Antagonists
as Therapeutic Agents; Sirotnak, F. M., Burchall, J . J ., Ens-
minger, W. D., Montgomery, J . A., Eds.; Academic Press: New
York, 1984; pp 1-68.
(4) Gangjee, A.; Elzein, E.; Kothare, M.; Vasudevan, A. Classical
and Nonclassical Antifolates as Potential Antitumor, Antipneu-
mocystis and Antitoxoplasma Agents. Curr. Pharm. Des. 1996,
2, 263-280.
(5) Kisliuk, R. L.; Gaumont, Y.; Powers, J . F.; Thorndike, J .; Nair,
M. G.; Piper, J . R. Synergistic Growth Inhibition by Combination
of Antifolates. In Evaluation of Folate Metabolism in Health and
Disease; Picciano, M. F., Stokstad, E. L. R., Gregory, J . F., III,
Eds.; Alan R. Liss: New York, 1990; pp 79-89.
(6) Rosowsky, A. Chemistry and Biological Activity of Antifolates.
In Progress in Medicinal Chemistry; Ellis, G. P., West, G. B.,
Eds.; Elsevier Science Publishers: Amsterdam, 1989; pp 1-252.
(7) Slavik, K.; Slavikova, V.; Kolma, Z. Metabolism of Folic Acid.
VI. Preparation of Intermediary Antimetabolites of Folic Acid.
Collect. Czech. Chem. Commun. 1960, 25, 1929-1937.
(8) DeGraw, J . I.; Colwell, W. T.; Kisliuk, R. L.; Gaumont, Y.;
Sirotnak, F. M. Synthesis and Antifolate Properties of 5,10-
Methylenetetrahydro-8,10-dideazaaminopterin. J . Med. Chem.
1986, 29, 1786-1789.
(9) Su, T. L.; Yang, Y. K.; Huang, J . T.; Ren, W. Y.; Watanabe, K.
A.; Chou, T. C. Synthesis of 4-[(1,3-Diaminopyrrolo[3′,4′:4′,5′]-
pyrido[2,3-d]pyrimidin-8-yl)-benzoyl]-L-glutamic Acid as Poten-
tial Antifolates. J . Heterocycl. Chem. 1993, 30, 1437-1443.
(10) DeGraw, J . I.; Christie, P. H.; Colwell, W. T.; Sirotnak, F. M.
Synthesis and Antifolate Properties of 5,10-Ethano-5,10-didea-
zaaminopterin. J . Med. Chem. 1992, 35, 320-324.
(11) (a) Kisliuk, R. L.; Gaumont, Y.; Kumar, P.; Coutts, M.; Nair, M.
G.; Nanavati, N. T.; Kalman, T. I. The Effect of Polyglutamyl-
ation on the Inhibitory Activity of Folate Analogs. In Proceedings
of the Second Workshop on Folyl and Antifolyl Polyglutamates;
Goldman, I. D., Ed.; Praeger: New York, 1985; pp 319-328. (b)
Cell Lin es a n d Meth od s for Mea su r in g Gr ow th In -
h ibitor y P oten cy (Ta ble 3). Solutions used in cell culture
studies were standardized using extinction coefficients. Ex-
tinction coefficients were determined for (pH 1, λmax-1 ) 224

5-Deaza Antifolates
McGuire, J . J . Antifolate Polyglutamylation in Preclinical and
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 23 5181
(33) Ihara, M.; Taniguchi, N.; Noguchi, K.; Fukumoto, K. Total
Synthesis of Hydrocinchonidine and Hydrocinchonine via Photo-
oxygenation of an Indole Derivative. J . Chem. Soc., Perkin Trans.
1 1988, 1277-1281.
(34) Gangjee, A.; Zeng, Y.; McGuire, J . J .; Kisliuk, R. L. Effect of
C9-Methyl Substitution and C8-C9 Conformational Restriction
on Antifolate and Antitumor Activity of Classical 5-Substituted
2,4-Diaminofuro[2,3-d]pyrimidines. J . Med. Chem. 2000, 43,
3125-3133.
(35) . E. coli DHFR was kindly provided by Dr. R. L. Blakley, St.
J ude Children’s Research Hospital, Memphis, TN.
(36) P. cariniii DHFR was kindly supplied by Dr. D. Borhani,
Southern Research Institute, Birmingham, AL.
Clinical Antifolate Resistance. In Antifolate Drugs: Basic Re-
search and Clinical Practice; J ackman, A. L., Ed.; Humana
Press: Totowa, NJ , 1999; pp 339-363.
(12) Tsukamoto, T.; Haile, W. H.; McGuire, J . J .; Coward, J . K.
Mechanism-Based Inhibition of Human Folypolyglutamate Syn-
thetase: Design, Synthesis, and Biomedical Characterization of
Phosphopeptide Mimic of the Tetrahedral Intermediate. Arch.
Biochem. Biophys. 1998, 355, 109-118.
(13) McGuire, J . J .; Hsieh, P.; Franco, C. T.; Piper, J . R. Folypoly-
glutamate Synthesis Inhibition and Cytotoxic Effects of Meth-
otrexate Analogs Containing 2,omega-Diaminoalkanoic Acids.
Biochem. Pharmacol. 1986, 35, 2607-2613.
(14) Sanghani, P. C.; J ackman, A.; Evans, V. R.; Thornton, T.;
Hughes, L.; Calvert, A. H.; Moran, R. A Strategy for the Design
of Membrane-Permeable Folypoly-γ-glutamate Synthetase In-
hibitors: “Bay-Region”-Substituted 2-Desamino-2-methyl-5,8-
dideazafolate Analogs. Mol. Pharmacol. 1994, 45, 341-351.
(15) Werbel, L. M. In Folate Antagonists as Therapeutic Agents;
Sirotnak, F. M., Burchall, J . J ., Ensminger, W., Montgomery,
J . A., Eds.; Academic Press: Orlando, FL, 1984; Vol. 1, pp 261.
(16) Allegra, C. J .; Kovcas, J . A.; Drake, J . C.; Swan, J . C.; Chabner,
B. A.; Masur, H. Activity of Antifolates against the Pneumocystis
carinii Dihydrofolate Reductase and Identification of a Potent
New Agent. J . Exp. Med. 1987, 165, 926-927.
(37) T. gondii DHFR was kindly supplied by Dr. D. V. Santi,
University of California, San Francisco CA.
(38) Recombinant human DHFR was kindly supplied by Dr. J . H.
Freisheim, Medical College of Ohio, Toledo, OH.
(39) (a) Marsham, P. R.; Wardleworth, J . M.; Boyl, F. T.; Hennequin,
L. F.; Kimbell, R.; Brown, M.; J ackman, A. L. Design and
Synthesis of Potent Non-Polyglutamatable Quinazoline Anti-
folate Thymidylate Synthase Inhibitors. J . Med. Chem. 1999,
42, 3809-3820. (b) Boyle, F. T.; Matusiak, Z. S.; Hughes, L. R.;
Slater, A. M.; Stephens, T. C.; Smith, M. N.; Kimbell, R.;
J ackman, A. L. Substituted 2-Desamino-2-methylquinazolines.
A Series of Novel Antitumour Agents. Chemistry and Biology of
Pteridines and Folates, Advances in Experimental Medicine and
Biology; Ayling, J . E., Nair, M. G., Baugh, C. M., Eds.; Plenum
Press: New York, 1995; pp 585-588.
(40) Foley, G. F.; Lazarus, H.; Farber, S.; Uzman, B. G.; Boone, B.
A.; McCarthy, R. E. Continuous Culture of Lymphoblasts from
Peripheral Blood of a Child with Acute Leukemia. Cancer 1965,
18, 522-529.
(41) Mini, E.; Srimatkandada, S.; Medina, W. D.; Moroson, B. A.;
Carman, M. D.; Bertino, J . R. Molecular and Karyological
Analysis of Methotrexate-Resistant and -Sensitive Human
Leukemic CCRF-CEM Cells. Cancer Res. 1985, 45, 317-325.
(42) Rosowsky, A.; Lazarus, H.; Yuan, G. C.; Beltz, W. R.; Mangini,
L.; Abelson, H. T.; Modest, E. J .; Frei, E., III. Effects of
Methotrexate Esters and Other Lipophilic Antifolates on Meth-
otrexate-Resistant Human Leukemic Lymphoblasts. Biochem.
Pharmacol. 1980, 29, 648-652.
(43) McCloskey, D. E.; McGuire, J . J .; Russell, C. A.; Rowan, B. G.;
Bertino, J . R.; Pizzorno, G.; Mini, E. Decreased Folylpoly-
glutamate Synthetase Activity as a Mechanism of Methotrexate
Resistance in CCRF-CEM Human Leukemia Sublines. J . Biol.
Chem. 1991, 266, 6181-6187.
(44) McGuire, J . J .; Coward, J . K. Pteroylpolyglutamates: Biosyn-
thesis, Degradation, and Function. In Folates and Pterins.
Chemistry and Biochemistry of Folates, 1st ed.; Blakley, R. L.,
Benkovic, S. J ., Eds.; Wiley: New York, 1984; Vol. 1, pp 135-
190.
(45) Rosowsky, A.; Galivan, J .; Beardsley, G. P.; Bader, H.; O’Connor,
B. M.; Russelo, O.; Moroson, B. A.; DeYarman, M. T.; Kerwar,
S. S.; Freisheim, J . H. Biochemical and Biological Studies on
2-Desamino-2-methylaminopterin, an Antifolate the Polyglutu-
mates of Which Are More Potent than the Monoglutamate
against Three Key Enzymes of Folate Metabolism. Cancer Res.
1992, 52, 2148-2155.
(46) We thank the Developmental Therapeutics Program of the
National Cancer Institute for performing the in vitro anticancer
evaluation.
(17) Kovcas, J . A.; Allegra, B. A.; Chabner, B. A.; Swan, J . C.; Drake,
M.; Lunde, M.; Parrillo, J . E.; Masur, H. Potent Effect of
Trimetrexate, a Lipid-Soluble Antifolate, against Toxoplasma
gondii. J . Infect. Dis. 1987, 155, 1027.
(18) Kovcas, J . A.; Allegra, J . C.; Swan, J . C.; Drake, J . C.; Parrillo,
J . E.; Chabner, B. A.; Masur, H. Potent Antipneumocystis and
Antitoxoplasm Activities of Piritrexim, a Lipid-Soluble Anti-
folate. Antimicrob. Agents Chemother. 1988, 32, 430-431.
(19) Masur, H.; Polis, M. A.; Tuazon, C. U.; Ogata-Arakaki, D.;
Kovcas, J . A.; Katz; Hilt, D.; Simmons, T.; Feuerstein, I.;
Lundgren, B.; Lane, H. C.; Chabner, B. A.; Allegra, J . C. Salvage
Trial of Trimetrexate-Leucovorin for the Treatment of Cerebral
Toxoplasmosis in Patients with AIDS. J . Infect. Dis. 1993, 167,
1422-1423.
(20) Champness, J . N.; Achari, A.; Ballantine, S. P.; Bryant, P. K.;
Delves, C. J .; Stammers, D. K. The Structure of Pneumocytis
carinii Dihydrofolate Reductase to 1.9 Å Resolution. Structure
1994, 2, 915-924.
(21) Gangjee, A.; Devraj, R.; Queener, S. Synthesis and Dihydrofolate
Reductase Inhibitory Activities of 2,4-Diamino-5-deaza and 2,4-
Diamino-5,10-dideaza Lipophilic Antifolates. J . Med. Chem.
1997, 40, 470-478.
(22) Mathews, D. A.; Bolin, J . T.; Burdge, J . M.; Filman, K. W.;
Kaufman, B. T.; Bedell, C. R.; Champness, J . N.; Stammers, D.
N.; Kraut, J . Dihydrofolate Reductase. The Stereochemistry of
Inhibitor Selectivity. J . Biol. Chem. 1985, 260, 392-396.
(23) Gangjee, A.; Donkor, I. O. Synthesis of Substituted Tetrahydro-
pyrimido[4,5-b][1,7]naphthyridines as Potential Antitumor Agents.
J . Heterocycl. Chem. 1989, 26, 705-708.
(24) DeGraw, J . I.; Christie, P. H.; Kisliuk, R. L.; Gaumont, Y.;
Sirotnak, F. M. Synthesis and Antifolate Properties of 10-Alkyl-
5,10-dideaza Analogs of Methotrexate and Tetrahydrofolic Acid.
J . Med. Chem. 1990, 33, 673-7.
(25) Taylor, E. C.; Skotnicki, J . S.; Fletcher, S. R. Synthesis of D,L-
7,10-Ethano-5-deazaaminopterin and L-Ethano-5-deazafolic Acid.
J . Org. Chem. 1985, 50, 1005-1010.
(26) McCabe, P. H.; Milne, N. J .; Sim, G. A. Conformational Study
of Bridgehead Lactams. Preparation and X-ray Structural
Analysis of 1-Azabicyclo[3.3.1]nonane-2,6-dione. J . Chem. Soc.,
Perkin Trans. 2 1989, 1459-1461.
(27) J osey, A. D.; Dickinson, C. L., J r.; Dewhirst, K. C.; McKusick,
B. C. Trifluoromethylmalononitrile. The Reaction of 1,1-Dichloro-
2,2-dicyanoethylene with Argentous Fluoride. J . Org. Chem.
1967, 32, 1941-1944.
(47) Blakley, R. L. The Biochemistry of Folic Acid and Related
Pteridines; Elsevier North-Holland: Amsterdam, 1969; pp 92-
94.
(48) Atkinson, I.; Garrow, T.; Brenner, A.; Shane, B. Human Cytosolic
Folylpoly-Gamma-Glutamate Synthase; Methods Enzymol. 1997,
281, 134-140.
(49) McGuire, J . J .; Hsieh, P.; Coward, J . K.; Bertino, J . R. In Vitro
Methotrexate Polyglutamate Synthesis by Rat Liver Folylpoly-
glutamate Synthetase and Its Inhibition by Bromosulfophtha-
lein. Adv. Exp. Med. Biol. 1983, 163, 199-214.
(50) McGuire, J . J .; Hsieh, P.; Coward, J . K.; Bertino, J . R. Enzymatic
Synthesis of Folylpolyglutamates. Characterization of the Reac-
tion and Its Products. J . Biol. Chem. 1980, 255, 5776-5788.
(51) Bolanowska, W. E.; Russell, C. A.; McGuire, J . J . Activation of
Mammalian Folylpolyglutamate Synthetase by Sodium Bicar-
bonate. Arch. Biochem. Biophys. 1990, 281, 198-203.
(28) . Protective Groups in Organic Synthesis; Greene, T. W., Wuts,
P. G. M., Eds.; J ohn Wiley & Sons: New York, 1999; pp 579-
580.
(29) Lott, R. S.; Chauhan, V. S.; Stammer, C. H. Trimethylsilyl Iodide
as a Peptide Deblocking Agent. J . Chem. Soc., Chem. Commun.
1979, 495-496.
(30) Angle, S. R.; Arnaiz, D. O. Stereoselective Synthesis of Substi-
tuted Pipecolic Acids. Tetrehedron Lett. 1989, 5, 515-518.
(31) Chelucci, G.; Falorni, M.; Giacomelli, G. Synthesis of 1-Substi-
tuted 2-[(2S)-2-Pyrrolidinyl]pyridine from L-Proline. Synthesis
1990, 1121-1122.
(32) Anantharamaiah, G. M.; Sivanandaiah, K. M. Transfer Hydro-
genation: A Convenient Method for Removal of Some Commonly
Used Protecting Group in Peptide Synthesis. J . Chem. Soc.,
Perkin Trans. 1 1977, 490-491.
J M0202369