5-Deazafolate analogues with a rotationally restricted glutamate or ornithine side chain: Synthesis and binding interaction with folylpolyglutamate synthetase

Article abstract of DOI:10.1021/jm9807205

Rotationally restricted analogues of 5-deazapteroyl-L-glutamate and (6R,6S)-5-deaza-5,6,7,8-tetrahydropteroyl-L-glutamate with a one-carbon bridge between the amide nitrogen and the 6'-position of the p-aminobenzoyl moiety were synthesized and tested as substrates for folylpolyglutamate synthetase (FPGS), a key enzyme in folate metabolism and an important determinant of the therapeutic potency and selectivity of classical antifolates. The corresponding bridged analogues of 5-deazapteroyl-L- ornithine and (6R,6S)-5-deaza-5,6,7,8-tetrahydropteroyl-L-ornithine were also synthesized as potential inhibitors. Condensation of diethyl L-glutamate with methyl 2-bromomethyl-4-nitrobenzoate followed by catalytic reduction of the nitro group, reductive coupling with 2-acetamido-6-formylpyrido[2,3- d]pyrimidin-4(3H)-one in the presence of dimethylaminoborane, and acidolysis with HBr/AcOH yielded 2-L-[5-[N-(2-acetamido-4(3H)-oxopyrido[2,3-d]pyrimidin- 6-yl)methylamino]-2,3-dihydro-1-oxo-2(1H)-isoindolyl]glutaric acid (1). When acidolysis was preceded by catalytic hydrogenation, the final product was the corresponding (6R,6S)-tetrahydro derivative 2. A similar sequence starting from methyl N(δ)benzyloxycarbonyl-L-ornithine led to 2-L-[5-[N-(2-amino- 4(3H)-oxopyrido[2,3-d]pyrimidin-6-yl)methylamino]-2,3-dihydro-1-oxo-2(1H)- isoindolyl]-5-aminopentanoic acid (3) and the (6R,6S)tetrahydro derivative 4. Compounds 3 and 4 were powerful inhibitors of recombinant human FPGS, whereas 1 and 2 were exceptionally efficient FPGS substrates, with the reduced compound 2 giving a Km (0.018 μM) lower than that of any other substrate identified to date. (6R,6S)-5-Deazatetrahydrofolate, in which the side chain is free to rotate, was rapidly converted to long-chain polyglutamates. In contrast, the reaction of 1 and 2 was limited to the addition of a single molecule of glutamic a cid. Hence rotational restriction of the side chain did not interfere with the initial FPGS-catalyzed reaction and indeed seemed to facilitate it, but the ensuing γ-glutamyl adduct was no longer an efficient substrate for the enzyme.

Full text of DOI:10.1021/jm9807205

3510
J . Med. Chem. 1999, 42, 3510-3519
5-Dea za fola te An a logu es w ith a Rota tion a lly Restr icted Glu ta m a te or Or n ith in e
Sid e Ch a in : Syn th esis a n d Bin d in g In ter a ction w ith F olylp olyglu ta m a te
Syn th eta se
Andre Rosowsky,*,† Ronald A. Forsch,† Allison Null,‡ and Richard G. Moran‡
Dana-Farber Cancer Institute and Department of Biological Chemistry and Molecular Pharmacology,
Harvard Medical School, Boston, Massachusetts 02115, and Department of Pharmacology/ Toxicology and
Massey Cancer Center, Medical College of Virginia, Virginia Commonwealth University, Richmond, Virginia 23298
Received December 21, 1998
Rotationally restricted analogues of 5-deazapteroyl-L-glutamate and (6R,6S)-5-deaza-5,6,7,8-
tetrahydropteroyl-L-glutamate with a one-carbon bridge between the amide nitrogen and the
6′-position of the p-aminobenzoyl moiety were synthesized and tested as substrates for
folylpolyglutamate synthetase (FPGS), a key enzyme in folate metabolism and an important
determinant of the therapeutic potency and selectivity of classical antifolates. The corresponding
bridged analogues of 5-deazapteroyl-L-ornithine and (6R,6S)-5-deaza-5,6,7,8-tetrahydropteroyl-
L-ornithine were also synthesized as potential inhibitors. Condensation of diethyl L-glutamate
with methyl 2-bromomethyl-4-nitrobenzoate followed by catalytic reduction of the nitro group,
reductive coupling with 2-acetamido-6-formylpyrido[2,3-d]pyrimidin-4(3H)-one in the presence
of dimethylaminoborane, and acidolysis with HBr/AcOH yielded 2-L-[5-[N-(2-acetamido-4(3H)-
oxopyrido[2,3-d]pyrimidin-6-yl)methylamino]-2,3-dihydro-1-oxo-2(1H)-isoindolyl]glutaric acid
(1). When acidolysis was preceded by catalytic hydrogenation, the final product was the
corresponding (6R,6S)-tetrahydro derivative 2. A similar sequence starting from methyl Nδ-
benzyloxycarbonyl-L-ornithine led to 2-L-[5-[N-(2-amino-4(3H)-oxopyrido[2,3-d]pyrimidin-6-yl)-
methylamino]-2,3-dihydro-1-oxo-2(1H)-isoindolyl]-5-aminopentanoic acid (3) and the (6R,6S)-
tetrahydro derivative 4. Compounds 3 and 4 were powerful inhibitors of recombinant human
FPGS, whereas 1 and 2 were exceptionally efficient FPGS substrates, with the reduced
compound 2 giving a Km (0.018 µM) lower than that of any other substrate identified to date.
(6R,6S)-5-Deazatetrahydrofolate, in which the side chain is free to rotate, was rapidly converted
to long-chain polyglutamates. In contrast, the reaction of 1 and 2 was limited to the addition
of a single molecule of glutamic acid. Hence rotational restriction of the side chain did not
interfere with the initial FPGS-catalyzed reaction and indeed seemed to facilitate it, but the
ensuing γ-glutamyl adduct was no longer an efficient substrate for the enzyme.
In tr od u ction
on purine and thymidine auxotrophy in FPGS-deficient
Chinese hamster ovary cells,11 and this concept was
recently reinforced by the finding that antisense down-
regulation of cellular FPGS activity leads to an overall
loss of cellular reduced folates, resulting in decreased
thymidylate synthesis.12
Folylpolyglutamate synthetase (FPGS) plays a critical
role in endogenous folate metabolism as well as in the
cellular pharmacology of classical antifolates.1-6 Intra-
cellularly formed γ-oligoglutamyl conjugates of natural
reduced folates such as 5,10-methylenetetrahydrofolate
or 10-formyltetrahydrofolate generally bind better than
the parent monoglutamates to their cognate enzymes
and also are better retained in the cell. The chain length
of each of the various reduced folate species in a cell is
regulated by the concerted action of FPGS and a second
enzyme, folylpolyglutamate hydrolase (FPGH), which
resides in lysosomes.7-9 Thus folate homeostasis is
maintained via a combination of FPGS activity in the
cytoplasmic compartment and FPGH activity in the
lysosomal compartment. A similar critical role is played
by FPGS and FPGH in the pharmacology of “classical”
antifolates whose biochemical action and therapeutic
selectivity both depend on preferential metabolism to
long-chain oligoglutamates in tumor versus host tissues.
FPGS was suggested to be a potential therapeutic target
a number of years ago10 on the basis of a seminal paper
Because of the important role of FPGS in the phar-
macologic action of classical antifolates, and more spe-
cifically in the therapeutic selectivity of these drugs and
the development of resistance,4-6 several early studies
sought to broadly identify the structural requirements
for binding of folates and folate antagonists to the active
site of the enzyme.13 More focused studies by several
groups showed that isosteric modification of the basic
folyl/antifolyl structure depicted in Figure 1 is well-
tolerated at the 5- and 8-positions of the B-ring,14-17 in
the 9,10-bridge,14 and in the phenyl ring,17e as are a five-
membered B-ring18 and even elimination of the B-ring
by removal of the C7 and N8 ring atoms.19 In contrast,
bulky groups in the so-called “bay region” formed by C7,
C9, N10, and the ortho position of the phenyl ring are
detrimental for binding.20-22 With at least some of these
“bay region” analogues,20 inefficient polyglutamation
has been found to be due to an unusual effect involving
changes in Vmax rather than Km.
† Dana-Farber Cancer Institute.
‡ Medical College of Virginia.
10.1021/jm9807205 CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/13/1999

5-Deazafolates with a Glu or Orn Side Chain
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18 3511
negative charge on the phosphapeptide side chain
impeded transport across the cell membrane.35
As part of a broader program of structure-activity
correlation involving substrates and inhibitors of FPGS,
we have previously synthesized analogues modified in
the region comprising the R-carbon and its flanking
CONH and R-carboxyl groups.34 In one of these ana-
logues for example, the amide carbonyl group was
replaced by CH2.34a In another, the (CH2)2COOH side
chain was moved from its normal position on the
R-carbon so that it resided on the amide nitrogen;34d in
yet another, the amide nitrogen was linked to the
â-carbon of glutaric acid.13a Binding to FPGS was
drastically curtailed in all three instances, and we
proposed that this could be due to a disruption of the
ability of the CONH group to participate in hydrogen
bonding. In the present study we were interested in
addressing the question of how the binding of com-
pounds with either a glutamate or ornithine side chain
would be influenced by preventing free rotation of the
bond between the phenyl ring and the CONH group.
To this end, we synthesized the rotationally restricted
glutamate-type analogues 1 and 2 as putative FPGS
substrates and the ornithine-type analogues 3 and 4 as
putative inhibitors. The 5-deazapteroyl derivatives were
chosen because previous experience with FPGS inhibi-
tors containing this ring system indicated excellent
binding to the enzyme.24d Compounds 3 and 4 are the
first reported examples of such rotationally restricted
compounds in which the side chain is ornithine. While
a handful of glutamate analogues with restricted rota-
tion about the amide bond are known,36-38 the interac-
tion of these compounds with FPGS has been charac-
terized in only one instance.37
F igu r e 1. General structure of folyl and antifolyl substrates
(Z ) COOH) and inhibitors (Z ) non-COOH) of FPGS. The
B-ring can be five-membered six-membered and can be aro-
matic, reduced, or cleaved. The aryl moiety can be carbocyclic
or heterocyclic and can be substituted or unsubstituted. The
4-oxo group can be replaced by 4-NH2 and the 2-NH2 group
by 2-H or 2-Me.
Of all the molecular specificity requirements for FPGS
binding by folyl substrates, the most critical for sub-
strate activity appears to be the distance between the
amide nitrogen and the γ-carboxyl group in the side
chain. With rare but important exceptions,13a catalysis
of the addition of L-glutamic acid is not observed unless
the terminal carboxyl in the folyl substrate is separated
from the R-carbon by precisely two CH2 groups, and
indeed, when the distance between the R-carbon and the
γ-carboxyl is sufficiently long, inhibition may be ob-
served.23 Not surprisingly, a number of folyl analogues
in which the γ-carboxyl group itself is modified are
potent inhibitors, in some cases with a Ki in the nano-
molar range. To date, the best inhibitors all contain an
ornithine side chain,13b,d,e,14d,24-26 though analogues con-
taining homocysteic acid13b,e,24b,d,27 or 2-amino-4-phospho-
nobutanoic acid13e,24b,d,28 are also fairly active. It has
been speculated that compounds with a sulfonate or
phosphonate group in the side chain bind to the folyl
binding site, whereas those containing ornithine may
act as bisubstrate analogues, in that the terminal amino
group occupies the binding site for the incoming glutam-
ic acid cosubstrate.13c Interestingly, recent evidence
indicates that the terminal amino group in ornithine
analogues needs to be protonated in order for strong
inhibition to occur.29 This would be consistent with a
model in which the active site contains a positive center
that can interact with the negatively charged γ-carboxyl
group of the folyl substrate and a negative center that
can interact with the positively charged R-amino group
of the incoming L-glutamate cosubstrate.
Analogues with other types of terminal modification,
such as a very weakly basic alkylsulfoximine30a or histi-
dine moiety30b or chain branching at the γ-position,31
are very weak inhibitors. On the other hand, binding is
moderately retained upon replacement of the γ-carboxyl
group by an acidic γ-tetrazole ring.32 Replacement of
glutamate by â,â-difluoroglutamate,33a,b but not other
fluorinated glutamic acids,33b,c actually enhances sub-
strate activity. Of note in the context of the present
work, elongation of â,â-difluoromethotrexate ceases
abruptly at the diglutamate stage.33a In a different
approach to the design of FPGS inhibitors, an interest-
ing phosphonate dipeptide mimic of the transition state
for the addition of glutamic acid to methotrexate was
found to be 26-fold more inhibitory than the correspond-
ing ornithine derivative, though unfortunately the high
Ch em istr y
The synthesis of compounds 1-4 is depicted in Scheme
1 and began with commercially available 2-methyl-4-
nitroaniline (5), which was converted in four steps to
the key intermediate methyl 2-bromomethyl-4-nitroben-
zoate (10) via the nitrile 6, the amide 7, the acid 8, and
the ester 9. The preparation of 6-9 was originally
described in a publication which is not widely avail-
able,39 whereas a more recent paper36 describing the use
of these compounds as intermediates for the synthesis
of rotationally restricted folate antagonists only referred
to the earlier work but gave no synthetic details. The

3512 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18
Rosowsky et al.
Sch em e 1a
a
Reagents: (a) HNO2/CuCN + KCN; (b) 80% H2SO4; (c) 6 N HCl; (d) SOCl2/MeOH; (e) NBS/Bz2O2; (f) H2NCH(COOEt)CH2CH2COOEt
(11)/K2CO3; (g) H2NCH(COOMe)CH2CH2NHCbz/K2CO3 (16); (h) H2/5% Pd-C; (i) SnCl2.
Sch em e 2a
a
Reagents: (a) Me2NH‚BH3/AcOH; (b) 5:1 HCOOH/Ac2O; (c) HBr/AcOH; (d) H2/PtO2/TFA.
Sandmeyer-type reaction of 5 with CuCN afforded 6 in
81% yield, but only when an excess of KCN was added
in order to solubilize the CuCN. Conversion of 6 to 7
was accomplished in 98% yield by heating with 80%
H2SO4 at 100 °C. An attempt to carry out this reaction
with alkaline H2O2 was unsuccessful. Hydrolysis of the
amide group to form 8 was accomplished in 88% yield
by heating in 6 N HCl under reflux. Unlike the reaction
of 7, direct hydrolysis of nitrile 6 to acid 8 with 6 N HCl
proceeded in low yield because of the facile tendency of
the nitrile to steam distill into, and block, the reflux
condenser. Esterification of 8 with SOCl2/MeOH yielded
9 (99% yield), and radical bromination of the 2-methyl
group with N-bromosuccinimide and a catalytic amount
of benzoyl peroxide yielded the bromide 10.36 Separation
of 10 from unreacted 9 was difficult, and we therefore
for 20 h (Scheme 2), followed by treatment with a 5:1
mixture of 95% formic acid and acetic anhydride for 1.5
h at room temperature, afforded a complex mixture from
which 15 could be isolated as a pale-yellow solid (37%)
by chromatography on silica gel. Simultaneous removal
of the ester and amide blocking groups by heating with
HBr in AcOH at 80 °C for 15 min then gave 1, isolated
as a colorless solid (68%) by preparative HPLC on C18
silica gel followed by ion-exchange chromatography on
DEAE-cellulose (HCO3- form). To obtain the tetrahydro
analogue 2, the protected intermediate 15 was subjected
to catalytic hydrogenation in TFA solution, and the
crude reduced product was heated directly with HBr in
AcOH (70 °C, 15 min). Preparative HPLC followed by
ion-exchange chromatography on DEAE-cellulose af-
forded 2 as a colorless solid in 30% yield. Formylation
of N10 serves a useful function in this sequence by
preventing unwanted reductive cleavage of the C9-N10
bond during the hydrogenation step.16b
1
used the oily mixture (93% yield, ca. 70% purity by H
NMR) directly in the next step. Treatment with excess
diethyl L-glutamate hydrochloride (11‚HCl) and K2CO3
in DMF at room temperature for 3 days afforded
isoindoline 12 in 51% yield as waxy solid. Reduction of
the nitro group in 12 (H2/5% Pd-C) yielded the amine
13 (92%) as an oil. Further reaction of 13 with 2-aceta-
mido-6-formylpyrido[2,3-d]pyrimidin-4(3H)-one (14)16b
and Me2NH-BH3 in glacial AcOH at room temperature
For the synthesis of the ornithine analogue 3, bromide
10 was treated with excess methyl Nδ-Cbz-L-ornithinate
(16)24c in the presence of K2CO3 in DMF solution at room
temperature for 4 days to obtain the isoindolinone 17
(72%). Since catalytic hydrogenation of the nitro group
was precluded by the presence of a Cbz group, reduction

5-Deazafolates with a Glu or Orn Side Chain
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18 3513
Ta ble 1. Binding of Rotationally Restricted 5-Deaza-5,6,7,8-tetrahydrofolate Analogues to Recombinant Human FPGSa
Km or Ki (µM)
1
2
3
4
22
Km
Ki
2.3 ( 0.6
-
0.018 ( 0.007
-
-
-
0.26 ( 0.10
-
110 ( 43
0.20 ( 0.07
a
Km values for 1, 2, and 22 as competitive substrates were determined in the presence of 10 µM (6S)-tetrahydrofolate, whereas Ki
values for 3 and 4 as competitive inhibitors were determined in the presence of 50 gmM aminopterin. Assays were performed as described
in ref 19. Each value is the mean ( standard deviation for three separate experiments done on different days.
was carried out with SnCl2 to form the amine 18 in 75%
yield after chromatography on silica gel. Reductive
coupling of 18 with aldehyde 14 (Scheme 2), followed
by N10-formylation with 5:1 HCOOH-Ac2O and chro-
matographic purification on silica gel, yielded the
protected intermediate 19 (43%). A faster-moving im-
purity was also isolated, which was formulated as the
N-formyl derivative of 18 on the basis that treatment
with HCl in MeOH converted it back to 18. Deprotection
of 19 with HBr/AcOH gave a 35% yield of 3 as a pale-
yellow powder after HPLC purification. Catalytic hy-
drogenation in TFA and removal of the blocking groups
with HBr in AcOH converted 19 to 4 in 50% yield after
ion-exchange chromatography followed by preparative
HPLC.
to be 2.3 and 0.018 µM. In contrast, (6R,6S)-5-deaza-
tetrahydrofolate (22) gave a Km of 0.26 µM. Thus,
rotational restriction of the amide bond in 2 appeared
to cause a 14-fold increase in binding to the human en-
zyme. It is of interest that the Km values reported earlier
for 22 and its nonreduced analogue 24 as substrates for
murine FPGS were 9.7 and 157 µM, respectively.16a To
our knowledge, 2 is the best FPGS substrate described
to date. Indeed, estimation of its Km required us to
employ a different assay method than the one typically
used with most known FPGS substrates. The Vmax for
the FPGS-catalyzed reaction of 2 was roughly equiva-
lent to that of aminopterin (data not shown).45
The major natural tetrahydrofolate cofactors are all
reported to have Km values in the micromolar range as
substrates for both murine FPGS (10-formyltetrahydro-
folate, 3.9 µM; 5,10-methylenetetrahydrofolate, 4.8 µM;
tetrahydrofolate, 7.0 µM; 5-methyltetrahydrofolate, 87
µM)40 and human FPGS (10-formyltetrahydrofolate, 3.7
µM; tetrahydrofolate, 4.4 µM; 5-methyltetrahydrofolate,
48 µM).42b Several antifolates are also reported to have
Km values in the micromolar range for this enzyme. For
example, the potent thymidylate synthase inhibitor
tomudex (ZD1694) has a Km of 1.37 µM, while the
glycinamide ribonucleotide formyltransferase inhibitor
(6R)-5,10-dideazatetrahydrofolate (lometrexol, 23) and
the multitargeted antifolate LY231514 (25), which are
conformationally unrestrained in the side chain, have
Km values of 9.3 and 0.8 µM, respectively.16e Another
excellent FPGS substrate is the tricyclic TS inhibitor
BW1843U89 (26), with a reported Km of 0.41 µM for the
hog liver enzyme.37a An important conclusion suggested
by our results is that the low Km of 26 is probably due
more to the conformational rigidity of the bridged amide
bond than to the tricyclic nature of the benzo[f]quinazo-
line moiety.
In ter a ction w ith F P GS
Compounds 1 and 2 were tested as FPGS substrates,
and compounds 3 and 4 were tested as FPGS inhibitors
according to a previously described assay which uses
aminopterin and [3H]glutamic acid as the cosub-
strates.40 The assay depends on the formation of a
charcoal-adsorbable radioactive product that can be
readily separated from unreacted [3H]glutamic acid.
Human cytosolic recombinant FPGS cloned from human
leukemic cells, expressed in insect cells, and purified
to homogeneity was used as the enzyme.41,42 Because
of their exceptionally high efficiency as substrates, the
Km values of 1 and 2 were determined by an alternate
assay method20,43 in which (6S)-5,6,7,8-tetrahydrofolate
is used as the competing substrate and the product is
converted to a 5,10-methylene derivative with formal-
dehyde, allowing it to be trapped as a covalent complex
with 5-fluoro-2′-deoxyuridylate (FdUMP) and thymidy-
late synthase.44
As shown in Table 1, the bridged ornithine analogues
3 and 4 were both inhibitors of FPGS, with Ki values of
110 and 0.20 µM, respectively. The >100-fold difference
in Ki between these two compounds was consistent with
our previous results on 5-deazapteroyl-L-ornithine (20)
versus its (6R,6S)-tetrahydro derivative 21, which gave
Ki values against mouse FPGS of 5.7 and 0.030 µM,
respectively.24d While the weaker binding we observed
for the bridged relative to the nonbridged analogues
could be due to the different species of origin of the two
enzymes, it is known from other studies that the Ki
values of ornithine-containing inhibitors of mouse and
human FPGS are not very different; in the case of N-(4-
amino-4-deoxy-N10-methylpteroyl)-L-ornithine, for ex-
ample, these values are reported to be 20.4 µM24a and
13.4 µM,24b a difference of less than 2-fold. Thus we
believe that the lower inhibitory activity of 4 versus 20
and 3 versus 21 is due to the bridge between the phenyl
ring and the amide nitrogen.
A significant feature of 26 is that, despite its low
substrate Km for FPGS, catalysis is efficient only for the
addition of the first glutamyl residue. Thus, in contrast
to nonbridged analogues such as 2347 and 25,48 the
predominant metabolite of 26 is the diglutamate.37c It
was therefore of interest to determine whether the
amide bridge in 2 similarly causes the FPGS reaction
to become arrested at the first step. An experiment was
carried out in which the distribution of radioactive
metabolites was examined by HPLC after incubating
1, 2, and 22 with [3H]L-glutamic acid and human FPGS
for 10 min and 2 h. As shown in Table 2, conversion of
22 to its long-chain polyglutamates was highly efficient,
so that by 2 h the only radioactive species detected by
HPLC were the penta- and hexaglutamate compounds.
In sharp contrast, the reaction of both 1 and 2 was
arrested after the addition of one glutamic acid residue,
although there was evidence for very slight conversion
of 2 to longer conjugates. Thus, the behavior of these
rotationally restricted analogues was similar to that of
When the glutamate-containing analogues 1 and 2
were tested as substrates, their Km values were found

3514 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18
Rosowsky et al.
Ta ble 2. Effect of Rotational Restriction on the Polyglutamate Chain Length of 5-Deazatetrahydrofolate Analoguesa
number of glutamates added
compd
time (min)
0
1
2
3
4
5
1
10
120
10
120
10
13.4
15.0
6.9
0
10
0
86.6
83.9
93.1
95.5
7.8
0
2.0
0
1.3
46.7
0
0
0
0
1.7
32.7
0
0
0
0
1.5
4.0
67.4
0
0
0
0
0
2
22
120
0
32.6
a
Compounds 1 (5 µM), 2 (3.5 µM), andr 22 (5 µM) were incubated for 10 or 120 min with 1 µg of recombinant human cytosolic FPGS
in 200 mM Tris buffer, pH 7.9, with 10 mM MgCl2, 5 mM ATP, 2.5 mM [3H]L-glutamic acid, 20 mM 2-mercaptoethanol, and 30 mM KCl
in a total volume of 500 µL. Radioactive polyglutamates were separated by paired-ion reverse-phase HPLC as described in ref 52. The
chain length of the products was determined by the ratio of 3H to UV absorbance, normalized with respect to the ratio for the diglutamate.
between human and L. casei FPGS suggests that the
two proteins will prove to have similar structures. Thus
the effects of conformational restriction on the binding
of folyl derivatives may be due to differences in the ease
with which the γ-carboxyl group of the side chain can
extend into the ATP binding domain. A better under-
standing of how rotational restriction about the amide
bond affects the binding of substrates such as 1 and 2
and inhibitors such as 3 and 4 to FPGS must await
three-dimensional structural analysis of crystalline
ternary complexes of the enzyme with ATP and a folyl
ligand.
Cell Gr ow th In h ibition
Assays of 1-4 as inhibitors of the growth of cultured
CCRF-CEM cells during 72 h of drug exposure showed
these compounds to be inactive at concentrations of up
to 100 µM (data not shown). The reported IC50 values
of the nonbridged glutamate 22 against CCRF-CEM
cells is reported to be 0.01 µM,16b and that of the
nonreduced analogue 26 (against L1210 cells) is re-
ported to be 8.8 µM.51 Thus, except for the important
information provided by bridged compounds 1-4 with
regard to the binding preferences of FPGS ligands,
further assessment of the interaction of these weakly
cytotoxic compounds with other enzymes of folate
metabolism, such as GARFT, TS, or DHFR, was not
considered worthwhile.
Exp er im en ta l Section
UV spectra were obtained on a Varian model 210 instru-
ment, and 1H NMR spectra were obtained with a Varian
EM360L instrument using Me4Si as the reference. IR spectra
(data omitted for brevity) were routinely obtained on all
compounds and were consistent with assigned structures. TLC
analyses were done on fluorescent Whatman MKSF silica gel-
coated glass slides (60 Å layer), using 254-nm UV illumination
to visualize spots. Column chromatography was on Baker 3405
(60-200 mesh) silica gel or Whatman DE-52 preswollen
DEAE-cellulose (HCO3- form). Solvents in moisture-sensitive
reactions were of Sure-Seal grade (Aldrich, Milwaukee, WI)
or were dried over Linde 4A molecular sieves (Fisher, Boston,
MA). HPLC purification of the final products was on Waters
26, in that they were rapidly metabolized by FPGS but
were only converted in significant amount to the di-
glutamate.
A hypothetical model for the binding of folate and
antifolates substrates to FPGS was recently proposed
on the basis of a crystallographic analysis of recombi-
nant enzyme cloned from Lactobacillus casei.49 Although
the structure was only solved for the ATP-protein
complex (i.e. without a folyl ligand), it was found that
the enzyme consists of two distinct domains, one of
which resembles the folate binding region of DHFR,
whereas the other has the typical features of an ATP
binding motif. Not surprisingly in view of the demon-
strated role of ATP in activating the γ-carboxyl group
during the FPGS reaction,50 the glutamate tail of the
folyl substrate is believed to extend across the interdo-
main cleft and into the ATP binding site. While the
three-dimensional structure of the mammalian enzyme
has yet to be solved, the high sequence homology
C
18 radial compression cartridges (analytical: 5-µm particles,
5 × 100 mm; preparative: 15-µm particles, 25 × 100 mm). In
those instances where HPLC was followed by ion-exchange
chromatography, the latter was performed in order to ensure
that the sample was not contaminated with silica gel, which
can occur from C18 columns on repeated use. Diethyl 2-L-[N-
(2,3-dihydro-5-nitro-1-oxo-2(1H)-isoindolyl)amino]glutat-
arate (12),36 2-acetamido-6-formylpyrido[2,3-d]pyrimidin-4(3H)-
one (14),16b and methyl Nδ-Cbz-L-ornithinate (16)24c,d were
synthesized according to published methods. A minor modi-
fication of the synthesis of 12 was that 1,2-dichloroethane was

5-Deazafolates with a Glu or Orn Side Chain
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18 3515
used instead of carbon tetrachloride as the solvent for the
radical bromination of 9. Melting points were determined in
Pyrex capillary tubes in a Mel-Temp apparatus (Cambridge
Laboratory Devices, Cambridge, MA) and are not corrected.
Microanalyses were performed by Quantitative Technologies,
Inc., Whitehouse, NJ , and were within (0.4% of calculated
values unless otherwise indicated.
Aminopterin and other chemicals for the FPGS assays were
purchased from Sigma (St. Louis, MO). [3,4-3H]L-Glutamic acid
was obtained from DuPont-New England Nuclear (Wilming-
ton, DE). The scintillation cocktail was Safety-Solve from
Research Products International (Mount Prospect, IL). (6S)-
Tetrahydrofolate was prepared enzymatically.44 Lactobacillus
casei thymidylate synthase was expressed in an Escherichia
coli transfectant kindly provided by Dr. Daniel Santi (Uni-
versity of California, San Francisco) and was purified by
phosphocellulose chromatography. Cloning, expression, and
purification to electrophoretic homogeneity of FPGS from
CCRF-CEM human leukemic lymphoblasts were carried out
as reported earlier.41
2.68 (s, ca. 0.9H, 2-CH3 of 9), 4.02 (s, 3H, OCH3 in 9 + 10),
4.97 (s, ca. 1.4H, CH2Br), 8.18 (m, 3H, aromatic protons in 9
+ 10). The product was estimated from the 1H NMR spectrum
to be a 7:3 mixture of 10 and unreacted 9. Direct reaction of
this mixture with diethyl L-glutamate36 afforded nitroindoli-
none 12 in 51% yield based on the estimated amount of 10 in
the reaction: mp 78-79 °C (lit.36 mp 73-74 °C). Catalytic
hydrogenation of 12 over 5% Pd-C as described36 then gave
the amino isoindolinone 13 as an oil (0.84 g, 92%): 1H NMR
(CDCl3) δ 1.2 (m, 6H, two CH2CH3), 2.3 (m, 4H, two glutaryl
CH2), 4.1-4.6 (complex m, 6H, two CH2CH3 and isoindolinyl
CH2), 5.0 (m, 1H, glutaryl CH), 6.7 (m, 2H, isoindolinyl 4-H
and 6-H), 7.7 (d, J ) 8 Hz, isoindolinyl 7-H).
A solution of 13 (344 mg, 1.03 mmol) and 14 (232 mg, 1.00
mmol) in glacial AcOH (5 mL) was stirred at room temperature
for 20 h, then treated with Me2NH‚BH3 (41 mg, 0.70 mmol).
Stirring was continued at room temperature for 1 h and at
60-65 °C (internal) for 10 min. The solvent was evaporated
and the residue dissolved in a ice-cold mixture of Ac2O (1 mL)
and 95% HCOOH (5 mL) which had been made up in advance.
The reaction mixture was kept at room temperature for 1.5 h
before being evaporated under reduced pressure. Analysis by
TLC (silica gel, 2:1 acetone-hexane) showed a spot at the
origin and three mobile spots with Rf values of 0.1 (coupling
product 15), 0.3, and 0.6. With 10:1 CHCl3-MeOH, the Rf of
15 was 0.5. The crude mixture was chromatographed twice
on silica gel (20 g, 2 × 14 cm), first with 2:1 acetone-hexane
as the eluent and then with 20:1 acetone-MeOH. Pooled TLC
homogeneous fractions were evaporated and the residue was
dried in vacuo at 70 °C over P2O5 to obtain a pale-yellow solid
Meth yl 2-Meth yl-4-n itr oben zoa te (9). Crushed ice (100
g) was added to a suspension of 2-methyl-4-nitroaniline (5)
(15.2 g, 0.1 mol) in 12 N HCl (25 mL) and the mixture was
stirred at 0-5 °C while adding a solution of NaNO2 (7.0 g, 0.1
mol) in H2O (20 mL) over a period of 15 min. Despite the
addition of an extra 10% of the NaNO2 solution, some starting
material remained undissolved. After careful neutralization
at 0-5 °C with aqueous Na2CO3 (ca u tion : CO2 evolution),
the neutralized mixture was added dropwise with cooling and
stirring to a two-phase mixture of EtOAc (100 mL) and a
solution of CuCN (8.95 g, 0.1 mol) and KCN (15 g, 0.23 mol)
in H2O (50 mL). After addition was complete the mixture was
allowed to warm to room temperature and left to stand
overnight. The mixture was filtered, the organic layer was
separated, the aqueous layer was extracted with EtOAc, and
the combined EtOAc layers were evaporated. The residue was
stirred in warm hexane, enough acetone was added to dissolve
almost all the solid, the warm mixture was filtered, and the
filtrate was cooled until crystals of nitrile 6 were obtained:
1
(215 mg, 37%): mp 118-121 °C; H NMR (CDCl3) δ 1.2 (m,
6H two CH2CH3), 2.4 (m, 4H, â- and γ-CH2), 2.5 (s, 3H, CH3-
CO), 4.1-4.6 (complex m, 6H, two CH2CH3, isoindolinyl CH2),
5.2 (m, 3H, 9-CH2 and R-CH), 7.2 (m, 2H, isoindolinyl 4-H and
6-H), 7.9 (d, J ) 8 Hz, 1H, isoindolinyl 7-H), 8.4 (d, J ) 2 Hz,
1H, pyridyl 5-H), 8.7 (s, 1H, CHdO), 8.93 (d, J ) 2 Hz, 1H,
pyridyl 7-H). Anal. (C28H30N6O8‚0.5H2O) C, H, N.
2-L-[5-[N-(2-Am in o-4(3H)-oxop yr id o[2,3-d ]p yr im id in -6-
yl)m eth yla m in o]-1-oxo-2(1H)-isoin d olyl]-L-glu ta r ic Acid
(1). A solution of 15 (134 mg, 0.232 mmol) in glacial AcOH (4
mL) was treated with 48% aqueous HBr (4 mL) and heated to
80 °C (internal) for 15 min. After rotary evaporation the
residue was taken up in dilute ammonia and purified by
preparative HPLC (C18 silica gel, 4% MeCN in 0.1 M NH4OAc,
pH 6.9, 1.0 mL/min, 280 nm) followed by ion-exchange chro-
1
yield 13.2 g (81%); mp 99-100 °C; H NMR (CDCl3) δ 2.7 (s,
3H, Me), 7.8 (d, J ) 10 Hz, 1H, 6-H), 8.2 (s, 1H, overlapping
dd, 1H, 3-H and 5-H).
A solution of 6 (11.5 g, 0.0707 mol) in 80% H2SO4 (75 mL)
was heated in an oil bath at 100 °C for 1.25 h, then cooled,
and poured onto ice. The precipitate was collected and dried
on a lyophilizer to obtain amide 7 as a light-yellow solid (12.5
-
matography on DEAE-cellulose (HCO3 form) using 0.4 M
1
g, 98%): mp 167-168 °C (lit.39 mp 167 °C); H NMR (CDCl3)
NH4HCO3 as the initial eluent, then 0.4 M NH4HCO3 adjusted
to pH 10 with 28% NH4OH. Pooled pure fractions were
lyophilized to dryness to obtain a white powder (80 mg, 68%):
mp >240 °C dec; UV (0.1 N NaOH) λmax 219 nm (ꢀ 30 200),
240 infl (24 600), 297 (20 400), 280 infl (19 400); 1H NMR
(DMSO-d6 + 1 drop each of TFA and D2O) δ 2.2 (m, 4H, â-
and γ-CH2), 4.2-4.9 (m, 5H, 9-CH2, R-CH, and indolinyl CH2),
6.8 (m, 2H isoindolinyl 4-H and 6-H), 7.4 (d, J ) 9 Hz, 1H,
isoindolinyl 7-H), 8.7 (m, 2H, pyridyl 5-H and 7-H). Anal.
(C21H20N6O8‚3H2O) C, H, N.
2-L -[2,3-Dih yd r o-5-[N -[(6R ,6S )-2-a m in o-4(3H )-oxo-
5,6,7,8-tetr ah ydr opyr ido[2,3-d]pyr im idin -6-yl]m eth ylam i-
n o]-1-oxo-2(1H)-isoin d olyl]glu ta r ic Acid (2). A solution of
15 (109 mg, 0.188 mmol) in TFA (8 mL) in a Parr apparatus
was shaken with PtO2 (15 mg) under H2 (3 atm) for 1 h. The
catalyst was filtered and washed with glacial AcOH, and the
combined filtrate and washings were evaporated under re-
duced pressure. The residue was redissolved in a glacial AcOH
(4 mL) and 48% aqueous HBr (4 mL) and the mixture was
heated at 70 °C (internal) for 15 min. The solution was re-
evaporated to dryness, and the residue was stirred in H2O
while enough 28% NH4OH was added to dissolve nearly all of
the solid. The insoluble portion was removed and the product
was purified by preparative HPLC followed by ion-exchange
chromatography as described in the preceding experiment.
Lyophilization of appropriately combined fractions from the
ion-exchange column yielded 2 as a white powder (49 mg, 30%)
whose elemental analysis indicated that it was a partial
δ 2.6 (s, 3H, Me), 7.6 (d, J ) 8 Hz, 1H, 5-H), 8.1 (m, 2H, 3-H
and 6-H).
A stirred mixture of 7 (12.5 g, 0.0693 mol) and 6 N HCl (500
mL) was refluxed for 20 h, and the black solid which remained
undissolved was removed by filtration and repeatedly ex-
tracted with hot water until no more of it dissolved. Cooling
of the combined filtrate and extracts caused precipitation of a
solid, which was filtered, washed with H2O, and dried on a
lyophilizer to obtain acid 8 as a yellow solid: yield 11.0 g (88%);
mp 154-155 °C (lit.39 mp 153-154 °C); 1H NMR (DMSO-d6) δ
2.6 (s, 3H, Me), 8.1 (m, 3H, aryl).
Thionyl chloride (18 g, 11 mL, 0.15 mol) was added dropwise
to a stirred solution of 8 (12.4 g, 0.0683 mol) in MeOH (100
mL) while maintaining the internal temperature below 12 °C.
When addition was complete the mixture was left to stand at
room temperature for 21 h to obtain needles of 9 (6.80 g): mp
74 °C (lit.39 mp 68-69 °C). Evaporation of the filtrate yielded
another 6.35 g of solid with mp 68-69 °C: total yield 13.2 g
1
(99%); H NMR (CDCl3) δ 2.7 (s, 3H, Me), 4.0 (s, 3H, OMe),
8.1 (m, 3H, aryl).
Dieth yl 2-L-[5-[N-(2-Aceta m id o-4(3H)-oxop yr id o[2,3-d ]-
pyr im idin -6-yl)m eth ylfor m am ido]-1-oxo-2(1H)-isoin dolyl]-
glu ta r a te (15). A solution of 9 (6.35 g, 0.0326 mol) in 1,2-
dichloroethane (50 mL) was treated with N-bromosuccinimide
(5.80 g, 0.0326 mol) and benzoyl peroxide (50 mg), and the
mixture was refluxed for 2 days, then cooled, washed with H2O,
and evaporated to an oil (8.32 g, 93%): 1H NMR (CDCl3) δ

3516 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18
Rosowsky et al.
ammonium salt: mp >240 °C dec; UV (0.1 N NaOH) λmax 279
nm (ꢀ 24 100), 292-302 (21 500); 1H NMR (DMSO-d6 + 1 drop
of D2O) δ 2.2 (m, 5H, â- and γ-CH2, tetrahydropyridyl 6-CH),
3.1 (m, 4H, 7- and 9-CH2), 4.4 (m, 3H, isoindolinyl CH2 and
glutaric acid R-CH), 6.7 (m, 2H, isoindolinyl 4-H and 6-H), 7.4
(1H, d, J ) 8 Hz, isoindolinyl 7-H). The tetrahydropyridyl
4-CH2 group, assumed to be at δ 3.6, was obscured by a large
H2O peak. Anal. (C21H24N6O6‚0.5NH3‚2H2O) C, H, N.
mg, 43%) which was pure enough to be used in the next step:
softening 125-130 °C; H NMR (CHCl3) δ 1.3-2.4 (m, >4H,
1
â- and γ-CH2, CH3 from residual acetone in the sample), 2.5
(s, 3H, CH3CO), 3.3 (m, 2H, δ-CH2), 3.7 (s, 3H, CH3), 4.5 (d, J
) 7 Hz, isoindolinyl CH2), 5.2 (m, 5H, 9-CH2, benzylic CH2,
and pentanoic acid R-CH), 7.4 (s, 7H, phenyl, isoindolinyl 4-H
and 6-H), 7.9 (d, J ) 10 Hz, 1H, isoindolinyl 7-H), 8.5 (d, J )
2 Hz, 1H, pyridyl 5-H), 8.7 (s, 1H, CHdO), 8.9 (d, J ) 2 Hz,
1H, pyridyl 7-H).
Meth yl 2-L-(2,3-Dih ydr o-5-n itr o-1-oxo-2(1H)-isoin dolyl)-
5-[N-(ben zyloxycar bon yl)am in o]pen tan oate (17). A stirred
The protected derivative 19 (187 mg, 0.285 mmol) was
dissolved in 6 mL of 15% HBr in AcOH with the aid of an
ultrasonic bath, kept at room temperature for 30 min, treated
with 16% aqueous HBr (3 mL), heated at 70 °C (internal) for
15 min, and evaporated to dryness. Analytical HPLC (C18 silica
gel, 1% AcOH, 5% EtOH, 1 mL/min) showed a major peak at
9.0 min corresponding to the product 3, along with minor peaks
at 4.0, 6.2, and 19.0 min. The solid was stirred in H2O while
adding NaOH dropwise until a clear solution formed (pH >10).
The solution was placed on top of a column of Dowex 50W-X2
(H+ form, 2 × 21 cm) which was eluted first with distilled H2O
until the eluate was neutral (desalting step) and then with
3% NH4OH. Fractions were monitored by analytical HPLC for
the appearance of a peak eluting at 9.0 min. Appropriate
fractions were pooled, reduced to a smaller volume, filtered
through a sintered glass filter to remove some turbidity, and
finally purified by preparative HPLC (C18 silica gel, 1% AcOH,
5% EtOH). The product eluting at 9.0 min was collected and
lyophilized, the resulting amorphous solid taken up in dilute
ammonia, and the solution relyophilized to obtain a pale-yellow
1
solution of bromide 10 (6.24 g, 0.0228 mol, based on H NMR)
in dry DMF (50 mL) was treated with 16 (7.22 g, 0.0228 mol)
and K2CO3 (6.9 g, 50 mmol), and after 4 days the mixture was
diluted with EtOAc and filtered. The filter cake was washed
with EtOAc, and the combined filtrate and washings were
rinsed with H2O. TLC (silica gel, 2:1 hexane-acetone) of the
organic layer showed 17 as a strong spot with Rf 0.3. The
organic layer was evaporated onto silica gel, and the dried
silica gel was applied to the top of a column of flash-grade silica
gel (50 g, 4 × 9 cm), which was eluted successively with 2:1
hexane-acetone and then with 3:2 hexane-acetone. Evapora-
tion of appropriately combined eluates yielded a yellow solid
(7.24 g, 72%): mp 83-85 °C. Recrystallization from a mixture
of hexane and acetone afforded the analytical sample as an
1
off-white solid: mp 95-96 °C with prior softening; H NMR
(CDCl3) δ 1.6 (m, 4H, pentanoic acid â- and γ-CH2), 3.3 (q, J
) 6 Hz, 2H, CH2NH), 3.7 (s, 3H, OCH3), 4.5-5.0 (m, 3H,
pentanoic acid R-CH, isoindolinyl CH2), 5.1 (s, 2H, benzylic
CH2), 7.3 (s, 5H, phenyl), 8.0 (d, J ) 10 Hz, 1H, isoindolinyl
6-H), 8.3 (m, 2H, isoindolinyl 4-H and 7-H). Anal. (C22H23N3O7)
C, H, N.
1
powder (49 mg, 35%): mp >240 °C; H NMR (DMSO-d6 + 2
drops TFA) δ 1.3-2.3 (m, 4H, â- and γ-CH2), 2.6-3.0 (m, 2H,
δ-CH2), 4.1-4.9 (complex m, 5H, R-CH, 9-CH2, and isoindolinyl
CH2), 6.8 (m, 2H, 2-NH2), 7.3-8.0 (m, 6H, δ-NH3+, isoindolinyl
4-H, 6-H, and 7-H), 8.4 (broad m, 1H, NH), 8.6 (d, J ) 2 Hz,
1H, pyridyl 5-H), 8.7 (d, J ) 2 Hz, 1H, pyridyl 7-H). Anal.
(C21H23N7O4‚3H2O) C, H, N.
Met h yl 2-L-(5-Am in o-2,3-d ih yd r o-1-oxo-2(1H )-isoin -
d olyl)-5-[N-(ben zyloxyca r bon yl)a m in o]p en ta n oa te (18).
A solution of the nitro isoindolinone 17 (0.93 g, 0.0211 mol) in
EtOAc (25 mL) was treated with SnCl2‚2H2O (2.37 g, 0.0105
mol), and after being heated to reflux for 1.5 h the mixture
was cooled and quenched with 5% aqueous NaHCO3. The
inorganic salts were filtered, the two layers of the filtrate were
separated, and the EtOAc layer was evaporated to a foam (0.70
g, 80% crude yield): Rf 0.4 (silica gel, EtOAc). Column
chromatography on silica gel (15 g, 2 × 11 cm) with EtOAc as
the eluent gave pure 18 as a hardened straw-colored foam (0.65
g, 75%): mp 49-52 °C; 1H NMR (CDCl3) δ 1.20 (m, >4H,
pentanoyl â- and γ-CH2 overlapping residual H2O in the
sample), 3.7 (s, 3H, OCH3), 3.9-4.4 (m, 4H, NH2, isoindolinyl
CH2), 5.0 (m, 4H, NH, benzylic CH2, pentanoic acid R-CH), 6.6
(m, 2H, isoindolinyl 4-H and 6-H), 7.3 (s, 5H, phenyl), 7.7 (d,
J ) 7 Hz, 1H, isoindolinyl 7-H). Anal. (C22H25N3O5‚0.5H2O)
C, H, N.
2-L-[5-[N-(2-Am in o-4(3H)-oxop yr id o[2,3-d ]p yr im id in -6-
yl)m et h yla m in o]-2,3-d ih yd r o-1-oxo-2(1H )-isoin d olyl]-5-
a m in op en ta n oic Acid (3). A mixture of 14 (360 mg, 155
mmol) and the aminoisoindolinone 18 (630 mg, 153 mmol) in
glacial AcOH (10 mL) was stirred at room temperature for 1
day before being treated with Me2NH‚BH3 (59 mg, 1.00 mmol).
Stirring was continued at room temperature for 1 h and then
at 60-65 °C (internal) for 10 min. The solvent was evaporated
and the residue dissolved in 95% HCOOH, cooled in an ice
bath, and treated with premixed ice-cold 95% HCOOH (5 mL)
and Ac2O (1 mL). After 1.5 h at room temperature the solution
was evaporated, and the residue was redissolved in a mixture
of acetone and MeOH and evaporated onto silica gel (2 g). The
dried silica gel was placed on top of a silica gel column (20 g,
2 × 15 cm) which was eluted first with 2:1 acetone-hexane to
remove a fast moving impurity with an Rf value of 0.4 (silica
gel, 2:1 acetone-hexane), which appeared to be the N-formyl
derivative of 18 (it was converted back to 18 upon reaction
with HCl in MeOH). After this impurity was removed the
column was eluted with 20:1 acetone-MeOH to recover a solid
(534 mg) whose TLC (silica gel, 10:1 CHCl3-MeOH) contained
a major spot with Rf 0.5 (coupling product 19) and a minor
contaminant with Rf 0.3. Rechromatography using the same
system, with acetone being used to rinse collection tubes from
the column, yielded 19 as a hardened straw-colored foam (489
2-L-[5-[N-[(6R,6S)-2-Am in o-4(3H )-oxo-5,6,7,8-t et r a h y-
d r op yr id o[2,3-d ]p yr im id in -6-yl]m eth yla m in o]-2,3-d ih y-
d r o-1-oxo-2(1H)-isoin d olyl]-5-a m in op en ta n oic Acid (4).
A solution of 19 (210 mg, 0.321 mmol) in TFA (8 mL) was
shaken in a Parr apparatus with PtO2 (20 mg) under 3 atm of
H2 for 1 h. The catalyst was filtered and washed with glacial
AcOH, the combined filtrate and washings were evaporated,
the residue was taken up in a mixture of glacial AcOH (4 mL)
and 48% aqueous HBr (4 mL), and the solution was heated at
80 °C (internal) for 15 min. After evaporation of the solvent
the residue was dissolved in H2O by addition of enough NaOH
to raise the pH above 10. A small amount of insoluble material
was removed by filtration; the filtrate was added to the top of
an ion-exchange column (Dowex 50W-X2, H+ form, 2 × 25 cm).
The column was eluted with distilled H2O until the pH of the
eluate was neutral (desalting step) and then with 3% NH4OH
while monitoring fractions by HPLC (C18 silica gel, 1% AcOH,
5% EtOH, 1.0 mL/min). Fractions showing a peak at 12 min
were pooled, concentrated to ca. 70 mL, treated with glacial
AcOH (0.7 mL), and chilled. A very small amount of solid
which precipitated was collected, the filtrate was repurified
by HPLC, and the pooled eluates showing a single peak at 12
min were freeze-dried to an off-white solid (90 mg, 50%): mp
>240 °C; 1H NMR (DMSO-d6 + 1 drop TFA) δ 1.4-3.5
(complex m, 16H, CH3COOH, â-, γ-, and δ-CH2, 9-CH2, and
tetrahydropyridyl 5-CH2, 6-CH, and 7-CH2), 4.4 (m, 2H,
isoindolinyl CH2), 4.8 (m, 1H, pentanoic acid R-CH), 6.3 (s, 1H,
NH), 6.8 (m, 2H, 2-NH2), 7.1 (s, 1H, isoindolinyl 4-H), 7.3 (m,
6H, isoindolinyl 6-H and 7-H, δ-NH3+, NH). Anal. (C21H27N7O4‚
CH3COOH‚2H2O) C, H, N.
F P GS Assa ys. The activity of 1, 2, and 22 as FPGS
substrates was determined according to a two-stage microas-
say method43 which is more satisfactory than the standard
charcoal adsorption procedure when the test compound (e.g.
2) has a low Km. The practical utility of this coupled assay
has been discussed in detail elsewhere.20 In the first step,
different concentrations of the test compound (10-600 µM)
were incubated at 37 °C for 30 min in a mixture containing

5-Deazafolates with a Glu or Orn Side Chain
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18 3517
FPGS, (6S)-tetrahydrofolate (10 µM), ATP (5 mM), [3H]L-
glutamic acid (1 mM, 4 mCi/mmol), KCl (30 mM), MgCl2 (10
mM), and Tris‚HCl, pH 8.5 (200 mM), in a total volume of 10
µL. In the second step, a mixture of purified recombinant E.
coli TS (1 µM), Na2HPO4, pH 7.2 (30 mM), 2-mercaptoethanol
(8 mM), formaldehyde (15 mM), and 5-fluoro-2′-deoxyuridylate
(2 µM) in a total volume of 100 µL was added, and incubation
at 37 °C was continued for another 30 min. The reaction
mixture was then passed through a Sephadex G-50 spin
column and the eluate analyzed for total 3H by scintillation
counting. A Dixon plot was used to estimate the Km of the
competitive substrate from the following equation: x-intercept
) -Km/(1 + [S]/Km,tetrahydrofolate), where [S] is the substrate
concentration. It is assumed in this calculation that the
reactions of (6S)-tetrahydrofolate and the test compound are
mutually exclusive. The Km of (6S)-tetrahydrofolate was taken
to be 1.0 µM, based on a separate kinetic determination of this
value.
The Ki of 3 and 4 as competitive inhibitors of the FPGS-
catalyzed reaction of aminopterin with [3H]L-glutamic acid was
determined by the standard charcoal adsorption method as
previously described.13a Briefly, different concentrations (0-
1000 µM) of the test compound were incubated at 37 °C for 1
h in the presence of aminopterin (50 µM), [3H]L-glutamic acid
(1 mM, 4 mCi/mmol), ATP (5 mM), MgCl2 (10 mM), KCl, (30
mM), and R-thioglycerol (20 mM) in Tris‚HCl, pH 8.6 (200
mM), in a total volume of 0.25 mL. In some experiments the
products of the FPGS reaction were separated by paired-ion
reverse-phase HPLC.52
turnover of methotrexate polyglutamates in lysosomes derived
from S180 cells: definition of a two-step process limited by
mediated lysosomal permeation of polyglutamates and activating
reduced sulfhydryl compounds. J . Biol. Chem. 1992, 267, 15356-
15361.
(8) (a) Samuels, L. L.; Goutas, L. J .; Priest, D. G.; Piper, J . R.;
Sirotnak, F. M. Hydrolytic cleavage of methotrexate γ-poly-
glutamates by folylpolyglutamyl hydrolase derived from various
tumors and normal tissues of the mouse. Cancer Res. 1986, 46,
2230-2235.
(9) (a) Yao, R.; Rhee, M. S.; Galivan, J . Effects of γ-glutamyl
hydrolase on folyl and antifolylpolyglutamates in cultured H35
hepatoma cells. Mol. Pharmacol. 1995, 48, 505-511. (b) Yao,
R.; Nimec, Z.; Tyan, T. J .; Galivan, J . Identification, cloning, and
sequencing of a cDNA coding for rat γ-glutamyl hydrolase. J .
Biol. Chem. 1996, 271, 8525-8528.
(10) Moran, R. Characterization of the function of mammalian
folylpolyglutamate synthetase (FPGS). Adv. Exp. Med. Biol.
1983, 163, 327-339.
(11) McBurney, M. W.; Whitmore, G. F. Isolation and biochemical
characterization of folate deficient mutants of Chinese hamster
cells. Cell 1974, 2, 173-182.
(12) Liu, Y.; Raghunathan, K.; Hill, C.; He, Y.; Bunni, M. A.; Barredo,
J .; Priest, D. G. Effect of antisense-based folylpoly-γ-glutamate
synthetase down-regulation on reduced folates and cellular
proliferation in CCRF-CEM cells. Biochem. Pharmacol. 1998,
55, 2031-2037.
(13) (a) Moran, R. G.; Colman, P. D.; Rosowsky, A.; Forsch, R. A.;
Chan, K. K. Structural features of 4-amino antifolates required
for substrate activity with mammalian folylpolyglutamate syn-
thetase. Mol. Pharmacol. 1985, 27, 156-166. (b) George, S.;
Cichowicz, D. J .; Shane, B. Mammalian folylpoly-γ-glutamate
synthetase. 3. Specificity for folate analogues. Biochemistry 1987,
26, 522-529. (c) Moran, R. G.; Colman, P. D.; Rosowsky, A.
Structural requirements for the activity of antifolates as sub-
strates for mammalian folylpolyglutamate synthetase. NCI
Monogr. 1987, 5, 133-138. (d) McGuire, J . J .; Piper, J . R.;
Coward, J . K.; Galivan, J . Folate analogue nonsubstrates and
inhibitors of folylpolyglutamate synthetase as potential cancer
chemotherapy drugs. NCI Monogr. 1987, 5, 139-144. (e)
Rosowsky, A.; Moran, R. G.; Freisheim, J . H.; Bader, H.; Forsch,
R. A.; Solan, V. C. Synthesis and biologic activity of new side-
chain-altered methotrexate and aminopterin analogues with
dual inhibitory activity against dihydrofolate reductase and
folylpolyglutamate synthetase. NCI Monogr. 1987, 5, 145-152.
(14) (a) Cichowicz, D. J .; Hynes, J . B.; Shane, B. Substrate specificity
of mammalian folylpoly-γ-glutamate synthetase for 5,8-dide-
azafolates and 5,8-dideaza analogues of aminopterin. Biochim.
Biophys. Acta 1988, 957, 363-369. (b) Coll, R. J .; Cesar, D.;
Hynes, J . B.; Shane, B. In vitro metabolism of 5,8-dideazafolates
and 5,8-dideazaisofolates by mammalian folypoly-γ-glutamate
synthetase. Biochem. Pharmacol. 1991, 42, 833-838. (c) Singh,
S. K.; Dev, I. K.; Duch, D. S.; Ferone, R.; Smith, G. K.; Freisheim,
J . H.; Hynes, J . B. Synthesis and biological evaluation of
5-deazaisofolic acid, 5-deaza-5,6,7,8-tetrahydroisofolic acid, and
their N9-substituted analogues. J . Med. Chem. 1991, 34, 606-
610. (d) Patil, S. A.; Shane, B.; Freisheim, J . H.; Singh, S. K.;
Hynes, J . B. Inhibition of mammalian folylpolyglutamate syn-
thetase and human dihydrofolate reductase by 5,8-dideaza
analogues of folic acid and aminopterin bearing a terminal
L-ornithine. J . Med. Chem. 1989, 32, 1559-1565.
Ack n ow led gm en t. This work was supported in part
by Grant RO1-CA63064 (A.R.) and Grant RO1-CA39687
(R.G.M.) from the National Cancer Institute, DHHS.
The authors are indebted to Dr. Peter Beardsley and
Ms. Barbara Moroson, Yale University School of Medi-
cine, for carrying out growth inhibition assays with
CCRF-CEM cells.
Refer en ces
(1) McGuire, J . J .; Coward, J . K. In Folates and Pterins; Blakley,
R. L., Benkovic, S. J ., Eds.; J ohn Wiley & Sons: New York, 1984;
Vol. 1, pp 135-190.
(2) Shane, B. Folylpolyglutamate synthesis and role in the regula-
tion of one-carbon metabolism. Vitam. Horm. 1989, 45, 263-
335.
(3) Fabre, I.; Fabre, G.; Goldman, I. D. Polyglutamylation, an
important element in methotrexate cytotoxicity and selectivity
in tumor versus murine granulocyte progenitor cells in vitro.
Cancer Res. 1984, 44, 3190-3195.
(4) Chabner, B. A.; Allegra, C. J .; Curt, G. A.; Clendeninn, N. J .;
Baram, J .; Koizumi, S.; Drake, J . C.; J olivet, J . Polyglutamation
of methotrexate. Is methotrexate a prodrug? J . Clin. Invest.
1985, 76, 907-912.
(5) (a) Pizzorno, G.; Mini, E.; Coronnello, M.; McGuire, J . J .;
Moroson, B. A.; Cashmore, A. R.; Dreyer, R. N.; Lin, J . T.;
Mazzeri, T.; Periti, P.; Bertino, J . R. Impaired polyglutamylation
of methotrexate as a cause of resistance in CCRF-CEM cells
after short-term, high-dose treatment with this drug. Cancer Res.
1988, 48, 2149-2155. (b) Pizzorno, G.; Chang, Y.-M.; McGuire,
J . J .; Bertino, J . R. Inherent resistance of human squamous
carcinoma cell lines to methotrexate as a result of decreased
polyglutamylation of this drug. Cancer Res. 1989, 49, 5275--
5280. (c) Li, W.-W.; Lin, J . T.; Tong, W. P.; Trippett, T. M.;
Brennan, M. F.; Bertino, J . R. Mechanisms of natural resistance
to antifolates in human soft tissue sarcomas. Cancer Res. 1992,
52, 1434-1438.
(6) (a) Rumberger, B. G.; Barrueco, J . R.; Sirotnak, F. M. Differing
specificities for 4-aminofolate analogues of folylpolyglutamate
synthetase from tumors and proliferative intestinal epithelium
of the mouse with significance for selective antitumor action.
Cancer Res. 1990, 50, 4639-4643. (b) Rumberger, B. G.; Schmid,
F. A.; Otter, G. M.; Sirotnak, F. M. Preferential selection during
therapy in vivo by edatrexate compared to methotrexate of
resistant L1210 variants with decreased folylpolyglutamate
synthetase activity. Cancer Commun. 1990, 2, 305-310.
(7) (a) Barrueco, J . R.; Sirotnak, F. M. Evidence for the facilitated
transport of methotrexate polyglutamates into lysosomes derived
from S180 cells. J . Biol. Chem. 1991, 266, 11732-11737. (b)
Barrueco, J . R.; O′Leary, D. F.; Sirotnak, F. M. Metabolic
(15) Moran, R. G.; Colman, P. D.; J ones, T. R. Relative substrate
activities of structurally related pteridine, quinazoline, and
pyrimidine analogues for mouse liver folylpolyglutamate syn-
thetase. Mol. Pharmacol. 1989, 36, 736-743.
(16) (a) Taylor, E. C.; Harrington, P. J .; Fletcher, S. R.; Beardsley,
G. P.; Moran, R. G. Synthesis of the antileukemic agents 5,10-
dideazaaminopterin and 5,10-dideaza-5,6,7,8-tetrahydroamino-
pterin. J . Med. Chem. 1985, 28, 914-921. (b) Taylor, E. C.;
Hamby, J . M.; Shih, C.; Grindey, G. B.; Rinzel, S. M.; Beardsley,
G. P.; Moran, R. M. Synthesis and antitumor activity of 5-deaza-
5,6,7,8-tetrahydrofolic acid and its N10-substituted analogues.
J . Med. Chem. 1989, 32, 1517-1522. (c) Beardsley, G. P.;
Moroson, B. A.; Taylor, E. C.; Moran, R. G. A new folate
antimetabolite, 5,10-dideaza-5,6,7,8-tetrahydrofolate, is a potent
inhibitor of de novo purine synthesis. J . Biol. Chem. 1989, 264,
328-333. (d) Moran, R. G.; Baldwin, S. W.; Taylor, E. C.; Chih,
C. The 6S- and 6R-diastereomers of 5,10-dideaza-5,6,7,8-tet-
rahydrofolate are equiactive inhibitors of de novo purine syn-
thesis. J . Biol. Chem. 1989, 264, 21047-21051. (e) Habeck, L.
L.; Mendelsohn, L. R.; Shih, C.; Taylor, E. C.; Colman, P. D.;
Gossett; Leitner, T. A.; Schultz, R. M.; Andis, S. L.; Moran, R.
G. Substrate specificity of mammalian folylpolyglutamate syn-
thetase for 5,10-dideazatetrahydrofolate analogues. Mol. Phar-
macol. 1995, 48, 326-333.
(17) Varney, M. D.; Palmer, C. L.; Romines III, W. H.; Boritzky, T.;
Margosiak, S. A.; Almassy, R.; J anson, C. A.; Bartlett, C.;
Howland, E. J .; Ferre, R. Protein structure-based design,

3518 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18
Rosowsky et al.
synthesis, and biological evaluation of 5-thia-2,6-diamino-4(3H)-
oxopyrimidines: potent inhibitors of glycinamide ribonucleotide
transformylase with potent cell growth inhibition. J . Med. Chem.
1997, 40, 2502-2524.
(25) (a) McGuire, J . J .; Hsieh, P.; Franco, C. T.; Piper, J . R.
Folylpolyglutmate synthetase inhibition and cytotoxic effects of
methotrexate analogues containing 2,ω-diaminoalkanoic acids.
Biochem. Pharmacol. 1986, 35, 2607-2613. (b) McGuire, J . J .;
Bolanowska, W. E.; Piper, J . R. Structural specificity of inhibi-
tion of human folylpolyglutamate synthetase by ornithine-
containing folate analogues. Biochem. Pharmacol. 1988, 37,
3931-3939.
(26) (a) Patil, S. A.; Shane, B.; Freisheim, J . H.; Singh, S.; Hynes. J .
B. Inhibition of mammalian folylpolyglutamate synthetase and
human dihydrofolate reductase by 5,8-dideaza analogues of folic
acid and aminopterin bearing a terminal L-ornithine. J . Med.
Chem. 1989, 32, 1559-1565. (b) Singh, S. K.; Singer, S. C.;
Ferone, R.; Waters, W. K.; Mullin, R. J .; Hynes, J . B. Synthesis
and biological evaluation of NR-(5-deaza-5,6,7,8-tetrahydropteroyl)-
L-ornithine. J . Med. Chem. 1992, 35, 2002-2006. (c) Hynes, J .
B.; Singh, S. K.; Fetzer, O.; Shane, B. Inhibition of hog liver
folylpolyglutamate synthetase by 5-substituted 5,8-dideaza ana-
logues of folic acid bearing a terminal L-ornithine residue. J .
Med. Chem. 1992, 35, 4078-4085.
(27) (a) Rosowsky, A.; Moran, R. G.; Forsch, R.; Colman, P.; Wick,
M. Methotrexate analogues. 17. The antitumor activity of
4-amino-4-deoxy-N10-methylpteroyl-D,L-homocysteic acid and its
dual inhibition of dihydrofolate reductase and folyl poly-
glutamate synthetase. Biochem. Pharmacol. 1984, 33, 155-162.
(b) Rosowsky, A.; Forsch, R. A.; Freisheim, J . H.; Moran, R. G.;
Wick, M. Methotrexate analogues. 19. Replacement of the
glutamate side-chain in classical antifolates by L-homocysteic
and L-cysteic acids: effect on enzyme inhibition and antitumor
activity. J . Med. Chem. 1984, 27, 600-604.
(18) (a) Taylor, E. C.; Kuhnt, D.; Shih, C.; Rinzel, S. M.; Grindey, G.
B.; Barredo, J .; J annatipour, M.; Moran, R. G. A dideazatet-
rahydrofolate analogue lacking a chiral center at C-6, N-[4-[(2-
amino-3,4-dihydro-4-oxo-7H-pyrrolo[2,3-d]pyrimidinyl-5-ethyl]-
benzoyl]-L-glutamic acid, is an inhibitor of thymidylate synthase.
J . Med. Chem. 1992, 35, 4450-4454. (b) Itoh, F.; Russello, O.;
Akimoto, H.; Beardsley, G. P. Novel pyrrolo[2,3-d]pyrimidine
antifolate TNP-351: cytotoxic effect on methotrexate-resistant
CCRF-CEM cells and inhibition of transformylases of de novo
purine biosynthesis. Cancer Chemother. Pharmacol. 1994, 34,
273-279. (c) McGuire, J . J .; Bergoltz, V. V.; Heitzman, K. J .;
Haile, W. H.; Russell, C. A.; Bolanowska, W. A.; Kotake, Y.;
Haneda, T.; Nomura, H. Novel 6,5-fused ring heterocyclic
antifolates: biochemical and biological characterization. Cancer
Res. 1994, 54, 2673-2679. (d) Gangjee, A.; Devraj, R.; McGuire,
J . J .; Kisliuk, R. L.; Queener, S. F.; Barrows, L. R. Classical and
nonclassical furo[2,3-d]pyrimidines as novel antifolates: syn-
thesis and biological activities. J . Med. Chem. 1994, 37, 1169-
1176. (e) Gangjee, A.; Mavandadi, F.; Queener, S. F.; McGuire,
J . J . Novel 2,4-diamino-5-substituted-pyrrolo[2,3-d]pyrimidines
as classical and nonclassical antifolate inhibitors of dihydrofolate
reductases. J . Med. Chem. 1995, 38, 2158-2165. (f) Gangjee,
A.; Devraj, R.; McGuire, J . J .; Kisliuk, R. L. Effect of bridge
region variation on antifolate and antitumor activity of classical
5-substituted 2,4-diaminofuro[2,3-d]pyrimidines. J . Med. Chem.
1995, 38, 3798-3805.
(19) (a) Kelley, J . L.; McLean, E. W.; Cohn, N. K.; Edelstein, M. P.;
Duch, D. S.; Smith, G. K.; Hanlon, M. H.; Ferone, R. Synthesis
and biological activity of an acyclic analogue of 5,6,7,8-tetrahy-
drofolic acid, N-[4-[[3-(2,4-diamino-1,6-dihydro-6-oxo-5-pyrim-
idinyl)propyl]amino]benzoyl-L-glutamic acid. J . Med. Chem.
1990, 33, 561-567. (b) Bigham, E. C.; Hodson, S. J .; Mallory,
W. R.; Wilson, D.; Duch, D. S.; Smith, G. K.; Ferone, R. Synthesis
and biological activity of open-chain analogues of 5,6,7,8-
tetrahydrofolic acid - - potential antitumor agents. J . Med. Chem.
1992, 35, 1399-1410. (c) Mullen, R. J .; Keith, B. R.; Bigham, E.
C.; Duch, D. S.; Ferone, R.; Heath, L. S.; Singer, S.; Waters, K.
A.; Wilson, H. R. In vivo antiumor activity and metabolism of a
series of 5-deazaacyclotetrahydrofolate (5-DACTHF) analogues.
Biochem. Pharmacol. 1992, 1627-1634.
(20) Sanghani, P. C.; J ackman, A.; Evans, V. R.; Thornton, T.;
Hughes, L.; Calvert, A. H.; Moran, R. G. A strategy for the design
of membrane-permeable folylpoly-γ-glutamate synthetase inhibi-
tors: “bay-region”-substituted 2-desamino-2-methyl-5,8-dide-
azafolate analogues. Mol. Pharmacol. 1994, 45, 341-351.
(21) (a) Matsuoka, H.; Kato, N.; Tsuji, K.; Maruyama, N.; Suzuki,
H.; Mihara, M.; Takeda, Y.; Yano, K. Antirheumatic agents. I.
Novel methotrexate derivatives bearing an indoline moiety.
Chem. Pharm. Bull. 1996, 44, 1332-1337. (b) Matsuoka, H.; Ohi,
N.; Mihara, M.; Susuki, H.; Miyamoto, K.; Maruyama, N.; Tsuji,
K.; Kato, N.; Akimoto, T.; Takeda, Y.; Yano, K.; Kuroki, T.
Antirheumatic agents: novel methotrexate derivatives bearing
a benzoxazine ring or benzothiazine moiety. J . Med. Chem. 1997,
40, 105-111.
(22) Piper, J . R.; J ohnson, C. A.; Maddry, J . A.; Malik, N. D.; McGuire,
J . J .; Otter, G. M.; Sirotnak, F. M. Studies on analogues of
classical antifolates bearing the naphthoyl group in place of
benzoyl in the side chain. J . Med. Chem. 1993, 36, 4161-4171.
(23) Rosowsky, A.; Bader, H.; Kohler, W.; Freisheim, J . H.; Moran,
R. G. Methotrexate analogues. 34. Replacement of the glutamate
moiety in methotrexate and aminopterin by long-chain 2-ami-
noalkanedioic acids. J . Med. Chem. 1988, 31, 1338-1344.
(24) (a) Rosowsky, A.; Freisheim, J . H.; Moran, R. G.; Solan, V. C.;
Bader, H.; Wright, J . E.; Radike-Smith, M. Methotrexate
analogues. 26. Inhibition of dihydrofolate reductase and folylpoly-
glutamate synthetase activity and in vitro tumor cell growth by
methotrexate and aminopterin analogues containing a basic
amino acid side-chain. J . Med. Chem. 1986, 29, 655-660. (b)
Clarke, L.; Rosowsky, A.; Waxman, D. J . Inhibition of human
liver folylpolyglutamate synthetase by non-γ-glutamylatable
antifolate analogues. Mol. Pharmacol. 1987, 31, 122-127. (c)
Rosowsky, A.; Forsch, R. A.; Bader, H.; Freisheim, J . H.
Synthesis and in vitro biological activity of new deaza analogues
of folic acid, aminopterin, and methotrexate with an L-ornithine
side chain. J . Med. Chem. 1991, 34, 1447-1454. (d) Rosowsky,
A.; Forsch, R. A.; Reich, V. E.; Freisheim, J . H.; Moran, R. G.
Side chain modified 5-deazafolate and 5-deazatetrahydrofolate
analogues as mammalian folylpolyglutamate synthetase and
glycinamide ribonucleotide formyltransferase inhibitors: syn-
thesis and in vitro biological evaluation. J . Med. Chem. 1992,
35, 1578-1588. (e) Rosowsky, A.; Forsch, R. A.; Moran, R. G.
Inhibition of folylpolyglutamate synthetase by substrate ana-
logues with an ornithine side chain. J . Heterocycl. Chem. 1996,
33, 1355-1361.
(28) Rosowsky, A.; Moran, R. G.; Forsch, R. A.; Radike-Smith, M.;
Colman, P. D.; Wick, M. M.; Freisheim, J . H. Methotrexate
analogues. 27. Dual inhibition of dihydrofolate reductase and
folylpolyglutamate synthetase by methotrexate and aminopterin
analogues with
Biochem. Pharmacol. 1986, 35, 3327-3333.
a γ-phosphonate group in the side chain.
(29) Tsukamoto, T.; Haile, W. H.; McGuire, J . J .; Coward, J . K.
Synthesis and biological evaluation of NR-(4-amino-4-deoxy-10-
methylpteroyl)-DL-4,4-difluoroornithine. J . Med. Chem. 1996,
39, 2536-2540.
(30) (a) Moran, R. G.; Colman, P. D.; Harvison, P. J .; Kalman, T. I.
Evaluation of pteroyl-S-alkylhomocysteine sulfoximines as in-
hibitors of mammalian folylpolyglutamate synthetase. Biochem.
Pharmacol. 1988, 37, 1997-2003. (b) Mao, Z.; Pan, J .; Kalman,
T. I. Design and synthesis of histidine analogues of folic acid
and methotrexate as potential folylpolyglutamate synthetase
inhibitors. J . Med. Chem. 1996, 39, 4340-4344.
(31) (a) Rosowsky, A.; Forsch, R. A.; Moran, R. G.; Freisheim, J . H.
Synthesis and in vitro biological evaluation of â,γ-methano
analogues of methotrexate and aminopterin. Pteridines 1990,
2, 133-139. (b) Abraham, A.; McGuire, J . J .; Galivan, J .; Nimec,
Z.; Kisliuk, R. L.; Gaumont, Y.; Nair, M. G. Folate analogues.
34. Synthesis and antitumor activity of nonpolyglutamatable
inhibitors of dihydrofolate reductase. J . Med. Chem. 1991, 34,
222-227.
(32) McGuire, J . J .; Russell, C. A.; Bolanowska, W. E.; Freitag, C.
M.; J ones, C. S.; Kalman, T. I. Biochemical and growth inhibition
studies of methotrexate and aminopterin analogues containing
a tetrazole ring in place of the γ-carboxyl group. Cancer Res.
1990, 50, 1726-1731.
(33) (a) McGuire, J . J .; Hart, B. P.; Haile, W. H.; Rhee, M.; Galivan,
J .; Coward, J . K. DL-â,â-Difluoroglutamic acid mediates position-
dependent enhancement or termination of pteroylpoyl(γ-glu-
tamate) synthesis catalyzed by folylpolyglutamate synthetase.
Arch. Biochem. Biophys. 1995, 321, 319-328. (b) McGuire, J .
J .; Haile, W. H.; Licato, N. J .; Bolanowska, W. E.; McGuire, J .
J .; Coward, J . K. Synthesis and biological activity of folic acid
and methotrexate analogues containing L-threo-(2S,4S)-4-fluo-
roglutamic acid and DL-3,3-difluoroglutamic acid. J . Med. Chem.
1996, 39, 56-65. (b) McGuire, J . J .; Hart, B. P.; Haile, W. H.;
Magee, K. J .; Rhee, M.; Bolanowska, W. E.; Russell, C.; Galivan,
J .; Paul, B.; Coward, J . K. Biological properties of fluoro-
glutamate- containing analogues of folates and methotrexate
with altered capacities to form poly(γ-glutamate) metabolites.
Biochem. Pharmacol. 1996, 52, 1295-1303. (c) Tsukamoto, T.;
Kitazume, T.; McGuire, J . J .; Coward, J . K. Synthesis and
biological evaluation of DL-4,4-difluoroglutamic acid and DL-
γ,γ-difluoromethotrexate. J . Med. Chem. 1996, 39, 66-72.
(34) (a) Rosowsky, A.; Forsch, R. Methotrexate analogues. 16.
Importance of the side-chain amide carbonyl group as a struc-
tural determinant of biological activity. J . Med. Chem. 1982, 25,
1454-1459. (b) Moran, R. G.; Colman, P. D.; Forsch, R. A.;
Rosowsky, A. A mechanism for the addition of multiple moles
of glutamate by folylpolyglutamate synthetase. J . Med. Chem.
1984, 27, 1263-1267. (c) Rosowsky, A.; Forsch, R. A.; Freisheim,
J . H.; Danenberg, P. V.; Moran, R. G.; Wick, M. M. Methotrexate
analogues. 29. The effect of γ-aminobutyric acid spacers between
the pteroyl and glutamate moieties on enzyme binding and cell

5-Deazafolates with a Glu or Orn Side Chain
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 18 3519
growth inhibition. J . Med. Chem. 1986, 29, 1872-1876. (d)
Rosowsky, A.; Bader, H.; Forsch, R. A.; Moran, R. G.; Freisheim,
J . H. Methotrexate analogues. 31. Meta and ortho isomers of
aminopterin, compounds with a double bond in the side chain,
and a novel analogue modified at the R-carbon: chemical and
in vitro biological studies. J . Med. Chem. 1988, 31, 763-768.
(35) Tsukamoto, T.; Haile, W. H.; McGuire, J . J .; Coward, J . K.
Mechanism-based inhibition of human folylpolyglutamate syn-
thetase: design, synthesis, and biochemical characterization of
a phosphapeptide mimic of the tetrahedral intermediate. Arch.
Biochem. Biophys. 1998, 355, 109-118.
(36) Marsham, P. R.; J ackman, A. L.; Hayter, A. J .; Daw, M. R.;
Snowden, J . L.; O′Connor, B. M.; Bishop, J . A. M.; Calvert, A.
H.; Hughes, L. R. Quinazoline antifolate thymidylate synthase
inhibitors: bridge modifications and conformationally restricted
analogues in the C2-methyl series. J . Med. Chem. 1991, 34,
2209-2218.
(37) (a) Duch, D. S.; Banks, S.; Dev, I. K.; Dickerson, S. H.; Ferone,
R.; Heath, S. S.; Humphreys, J .; Knick, V.; Pendergast, W.;
Singer, S.; Smith, G. K.; Waters, K.; Wilson, H. R. Cancer Res.
1993, 53, 810-818. (b) Pendergast, W.; Dickerson, S. H.; Dev,
I. K.; Ferone, R.; Duch, D. S.; Smith, G. K. Benzo[f]quinazoline
inhibitors of thymidylate synthase: methyleneamino-linked
aroylglutamate derivatives. J . Med. Chem. 1994, 37, 838-844.
(c) Hanlon, M. H.; Ferone, R. In vitro uptake, anabolism, and
cellular retention of 1843U89 and other benzoquinazoline inhibi-
tors of thymidylate synthase. Cancer Res. 1996, 56, 3301-3306.
(38) (a) Taylor, E. C.; Zhou, P.; J ennings, L. D.; Mao, Z.; Hu, B.; J un,
J .-G. Novel synthesis of a conformationally constrained analogue
of DDATHF. Tetrahedron Lett. 1997, 38, 521-524. (b) Taylor,
E. C.; J ennings, L. D.; Mao, Z.; Hu, B.; J un, J .-G.; Zhou, P.
Synthesis of conformationally constrained glutamate analogues
of the antitumor agents DDATHF, LY254155, and LY231514.
J . Org. Chem. 1997, 62, 5392-5403.
(39) Peltier, D. Interaction of substituents in the benzene ring.
Substituted o-toluic acids and their esters. Bull. Soc. Sci.
Bretagne 1956, 31, 7-92; Chem. Abstr. 1958, 52, 9016h.
(40) Moran, R. G.; Colman, P. D. Mammalian folyl polyglutamate
synthetase: partial purification and properties of the mouse liver
enzyme. Biochemistry 1984, 23, 4580-4589.
(41) (a) Sanghani, P.; Sanghani, S.; Freemantle, S. J .; Taylor, S. M.;
Moran, R. G. Eukaryotic expression of human folylpoly-γ-
glutamate synthetase. AACR Proc. 1995, 36, 379. (b) Free-
mantle, S. J .; Taylor, S. M.; Krystal, G.; Moran, R. G. Upstream
organization of and multiple transcripts from the human
folylpoly-γ-glutamate sythetase gene. J . Biol. Chem. 1995, 270,
9579-9584. (c) Taylor, S. M.; Freemantle, S. J .; Moran, R. G.
Structural organization of the human folylpoly-γ-glutamate
synthetase gene: evidence for a sigle genomic locus. Cancer Res.
1995, 55, 6030-6034.
J .; Garrow, T. A.; Choi, Y.-J .; Shane, B. Purification and
properties of human cytosolic folylpoly-γ-glutamate synthetase
and organization, localization, and differential splicing of its
gene. J . Biol. Chem. 1996, 271, 13077-13087.
(43) Antonsson, B.; Barredo, J .; Moran, R. G. A microassay for
mammalian folypolyglutamate synthetase. Anal. Biochem. 1990,
186, 8-13.
(44) Moran, R. G.; Spears, C. P.; Heidelberger, C. Biochemical
determinants of tumor sensitivity to 5-fluorouracil: ultrasensi-
tive methods for determination of 5-fluoro-2′-deoxyuridylate, 2′-
deoxyuridylate, and thymidylate synthetase. Proc. Natl. Acad.
Sci. U.S.A. 1979, 76, 1456-1460.
(45) It should be noted that the question whether the substrate
kinetics of binding of the individual diastereomers of this
compound (or of 4) to FPGS are different could not be addressed
in the present study because the 6R,6S mixtures were not
resolved. However it is known that the kinetic constants for the
6R (i.e. the one with the ‘natural′ configuration)46 and 6S
diastereomers of DDATHF (23) as FPGS substrates are almost
the same16d and that the transport kinetics, antipurine activity,
and cytotoxicity of the two diastereomers are likewise very
similar.16d,e Thus the likelihood of widely divergent biochemical
and biological activities for the individual diastereomers of 2 and
4 seems remote.
(46) Barnett, C. J .; Wilson, T. M.; Wendel, S. R.; Winningham, M.
J .; Deeter, J . B. Asymmetric synthesis of lometrexol ((6R)-5, 10-
dideaza-5,6,7,8-tetrahydrofolic acid). J . Org. Chem. 1994, 59,
7038-7045; see also earlier reference cited.
(47) Pizzorno, G.; Sokoloski, J . A.; Cashmore, A. R.; Moroson, B. A.;
Cross, A. D.; Beardsley, G. P. Intracellular metabolism of 5,10-
dideazatetrahydrofolic acid in human leukemic cell lines. Mol.
Pharmacol. 1991, 39, 85-89.
(48) Shih, C.; Chen, V. J .; Gossett, L. S.; Gates, S. B.; MacKellar, W.
C.; Habeck, L. L.; Shackelford, K. A.; Mendelsohn, L. G.; Soose,
D. J .; Patel, V. F.; Andis, L. L.; Bewley, J . R.; Rayl, E. A.;
Moroson, B. A.; Beardsley, G. P.; Kohler, W.; Ratnam, M.;
Schultz, R. M. LY231514,
a pyrrolo[2,3-d]pyrimidine-based
antifolate that inhibits multiple folate-requiring enzymes. Can-
cer Res. 1997, 57, 1116-1123.
(49) Sun, X.; Bognar, A. L.; Baker, E. N.; Smith, C. A. Structural
homologies with ATP- and folate-binding enzymes in the crystal
structure of folylpolyglutamate synthetase. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 6647-6652.
(50) Banerjee, R. V.; Shane, B.; McGuire, J . J .; Coward, J . K.
Dihydrofolate synthetase and folylpolyglutamate synthetase:
direct evidence for intervention of acyl phosphate intermediates.
Biochemistry 1988, 27, 9062-9070.
(51) Taylor, E. C.; Tseng, C.-P.; Harrington, P. J .; Beardsley, G. P.;
Rosowsky, A.; Wick, M. Synthesis and biological activity of
5-deazafolic acid and 5-deazaaminopterin. In Chemistry and
Biology of Pteridines; Blair, J . A., Ed.; Walter de Gruyter: Berlin,
1983; pp 115-119.
(52) Sanghani, S.; Moran, R. G. Tight binding of folate substrates
and inhibitors to recombinant mouse glycinamide ribonucleotide
formyltransferase. Biochemistry 1997, 36, 10506-10516. 38
(42) The human enzyme has also been expresssed in bacteria: (a)
Garrow, T. A.; Admon, A.; Shane, B. Expression cloning of a
human cDNA encoding folylpoly(γ-glutamate) synthetase and
determination of its primary structure. Proc. Natl. Acad. Sci.
U.S.A. 1992, 89, 9151-9155. (b) Chen, L.; Qi, H.; Korenberg,
J M9807205