Stereoselective inhibition of glutamate carboxypeptidase by organophosphorus derivatives of glutamic acid

Article abstract of DOI:10.1016/j.bmc.2004.08.016

A series of phosphonyl, thiophosphonyl, and dithiophosphonyl derivatives of glutamic acid were prepared and examined for inhibitory potency against glutamate carboxypeptidase (carboxypeptidase G). Although varied in extent, stereoselective inhibition was dependent upon both carbon and phosphorus stereochemistry. A series of alkyl and aryl phosphonyl, thiophosphonyl, and dithiophosphonyl derivatives of (S)- and (R)-glutamic acid were prepared and examined for inhibitory potency against glutamate carboxypeptidase (carboxypeptidase G). The acquisition of the phosphonamidodithioic acids and the individual phosphonamidothioic acid diastereomers was achieved through a common phosphonamidothiolate precursor, which also allowed for the chromatographic resolution of the chiral phosphorus center of the phosphonamidothioic acids. The most potent inhibitor of the series was the n-butylphosphonamidate derivative of the natural isomer of glutamic acid. Although each diastereomeric pair of three phosphonamidothionates exhibited stereoselective inhibition consistent with the configuration of the chiral phosphorus center, this effect was generally not remarkable. More important, was the effect of carbon stereochemistry upon glutamate carboxypeptidase inhibition as exemplified by a limited series of enantiomeric pairs of phosphonamidate and phosphonamidodithionate derivatives of glutamic acid. The phosphonamidate analogs derived from the unnatural stereoisomer of glutamic acid were devoid of inhibitory potency in contrast to their enantiomers. Surprisingly, the phosphonamidodithionates derived from the unnatural stereoisomer of glutamic acid demonstrated greater inhibitory potency than their naturally-derived antipodes.

Full text of DOI:10.1016/j.bmc.2004.08.016

Bioorganic & Medicinal Chemistry 12 (2004) 6011–6020
Stereoselective inhibition of glutamate carboxypeptidase by
organophosphorus derivatives of glutamic acid
Jeremy P. Mallari,a Cindy J. Choy,a Ying Hu,a Alicia R. Martinez,a Mia Hosaka,a
Yoko Toriyabe,a Jack Maung,a Joseph E. Blecha,a Stephen F. Pavkovicb and
Clifford E. Berkmana,*
aDepartment of Chemistry and Biochemistry, San Francisco State University, 1600 Holloway Ave., San Francisco, CA 94132, USA
bDepartment of Chemistry, Loyola University of Chicago, 6525 N. Sheridan Rd, Chicago, IL 60626, USA
Received 8 January 2003; revised 6 August 2004; accepted 12 August 2004
Available online 11 September 2004
Abstract—A series of alkyl and aryl phosphonyl, thiophosphonyl, and dithiophosphonyl derivatives of (S)- and (R)-glutamic acid
were prepared and examined for inhibitory potency against glutamate carboxypeptidase (carboxypeptidase G). The acquisition of
the phosphonamidodithioic acids and the individual phosphonamidothioic acid diastereomers was achieved through a common
phosphonamidothiolate precursor, which also allowed for the chromatographic resolution of the chiral phosphorus center of the
phosphonamidothioic acids. The most potent inhibitor of the series was the n-butylphosphonamidate derivative of the natural iso-
mer of glutamic acid. Although each diastereomeric pair of three phosphonamidothionates exhibited stereoselective inhibition con-
sistent with the configuration of the chiral phosphorus center, this effect was generally not remarkable. More important, was the
effect of carbon stereochemistry upon glutamate carboxypeptidase inhibition as exemplified by a limited series of enantiomeric pairs
of phosphonamidate and phosphonamidodithionate derivatives of glutamic acid. The phosphonamidate analogs derived from the
unnatural stereoisomer of glutamic acid were devoid of inhibitory potency in contrast to their enantiomers. Surprisingly, the phos-
phonamidodithionates derived from the unnatural stereoisomer of glutamic acid demonstrated greater inhibitory potency than their
naturally-derived antipodes.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
been recently sought for use in inhibiting non-tumor-
localized enzymes in ADEPT strategies.5
Our research efforts have been recently aimed at devel-
oping potent competitive inhibitors for glutamate carb-
oxypeptidases. This class of enzymes encompasses
various metallopeptidases such as N-acetylated-alpha-
linked-acidic dipeptidase (NAALADase),1 prostate-spe-
cific membrane antigen (PSMA),2 pteroylpoly-gluta-
Based upon preliminary evidence, alkylphosphonyl and
alkylthiophosphonyl derivatives of glutamic and
2-hydroxyglutaric acid have exhibited promise as potent
tetrahedral-intermediate analog inhibitors of glutamate-
liberating metallopeptidases.6 Although the phosphonyl
motif often provides suitable inhibitory potency toward
metallopeptidases, thiophosphonyl analogs maintain the
unique potential for probing enzyme active-site architec-
ture with complementary chiral phosphorus centers. In
some cases, a single alkylthiophosphonyl stereoisomer
displayed inhibitory potency notably greater than either
a stereoisomer of antipodal phosphorus stereochemistry
or a respective alkylphosphonyl analog.6b The basis for
the enhanced inhibitory potency of such compounds,
especially against zinc-metallopeptidases, was presuma-
bly due to favorable zinc–sulfur interactions within en-
zyme active sites.7
mate hydrolase (PPH),3 and carboxypeptidase
G
(CPG).4 The acquisition of inhibitors for such enzymes
is expected to further the understanding of the biological
role of these metallocarboxypeptidases as well as to
serve in the elucidation of germane active-site features.
In addition, inhibitors of carboxypeptidase G2 have
Keywords: Phosphonamidothionate; Phosphonamidodithionate; Ste-
reoselective; Metallopeptidase.
*
Corresponding author. Tel.: +1 415 338 6495; fax: +1 415 338
2384; e-mail: cberkman@sfsu.edu
0968-0896/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.08.016

6012
J. P. Mallari et al. / Bioorg. Med. Chem. 12 (2004) 6011–6020
—
—
those designated as ÔbÕ possess antipodal configuration
at that center.
O
P
CO2
O
P
CO2
—
—
R
N
H
R
R
N
H
CO2
CO2
—O
—O
1 R = Et
2 R = n-Bu
3 R = Ph
2. Results and discussion
2' R = n-Bu
3' R = Ph
Although phosphonamidate 3, in which the natural iso-
mer of glutamic acid was incorporated, was available
from previous studies,6c we prepared additional material
along with analogs 1 and 2 as described in Scheme 1.
Starting with the appropriate phosphonic dichloride,
intermediate phosphonamidates 10–12 were prepared
and subsequently deprotected to provide the desired
phosphonamidoic acids 1–3. In previous studies, we
found the fluorenylmethoxy group to be a convenient
protecting ligand for the preparation of phosphonic acid
derivatives and one that can be concomitantly removed
with methyl esters under mild basic conditions.8,9 The
enantiomeric analogs of 1–3, namely 10–30, in which
the configuration of the a-carbon on glutamic acid was
(R), were prepared in the manner described by Scheme
1 with the exception that (R)-glutamic acid dimethyl
ester was used in the latter half of the first step. To con-
firm that inhibitors 1–3 would be sufficiently stable un-
der the conditions of the enzyme inhibition studies,
stability studies at pH7.3 were conducted. No decompo-
sition of phosphonamidates was observed under the
these conditions over a 15h period.
—
CO2
—
CO2
S
P
S
P
—
CO2
—
CO2
—O
N
H
N
H
R
—O
4a R = Et
5a R = n-Bu
6a R = Ph
4b R = Et
5b R = n-Bu
6b R = Ph
—
CO2
—
CO2
S
P
S
P
—
CO2
—
CO2
R
N
H
R
N
H
—S
—S
7 R = Et
8 R = n-Bu
9 R = Ph
8' R = n-Bu
9' R = Ph
Figure 1.
The focus of the present study was threefold. First, we
were interested in exploring the significance of phospho-
rus stereochemistry of thiophosphonyl derivatives of
glutamic acid upon the inhibition of glutamate carb-
oxypeptidase. Second, we wanted to determine if dithio-
phosphonyl analogs could serve as reasonably potent
metalloprotease inhibitors. Lastly, we wanted to assess
the significance of glutamic acidÕs a-carbon stereochem-
istry upon the inhibitory potency of glutamate carboxy-
peptidases by phosphonyl and dithiophosphonyl
derivatives of glutamic acid.
The individual phosphonamidothioic acid stereoisomers
5a and 5b as well as 6a and 6b were available from pre-
vious work in our lab.8 The stereoisomers of the ethyl
analog (4a and 4b) were prepared as outlined in Scheme
2. Following the chromatographic resolution of phos-
phonamidothionate intermediates 13a and 13b, depro-
tection of the sulfur ligand was accomplished
stereospecifically with methylbenzylamine. Initially it
was envisioned that typical deprotection of the b-(acet-
ylmercapto)ethyl protecting group on sulfur by a hin-
The compounds for which the synthesis and inhibitory
potency against glutamate carboxypeptidase are
described herein are shown in Figure 1. Prime designa-
tions for both phosphonamidates
2
and
3
and
dered base such as quinine would yield a salt
phosphonamidodithionates 8 and 9 identify these com-
pounds as possessing unnatural or (R)-stereochemistry
at the alpha carbon of gluatmic acid. All other com-
pounds listed without prime designations are derivatives
of (S)-glutamic acid. Compounds designated as ÔaÕ iso-
mers maintain a (R)-configuration at phosphorus while
amenable to purification by crystallization. Though this
technique has been successful for other analogs,8 the
reaction of one isomer (13a) with quinine or strychnine
resulted in an impure salt that was recalcitrant to at-
tempts at recrystallization. However, it was found that
reaction of 13a and 13b each with L-(ꢀ)-a-methylbenzyl-
RP(O)Cl2
1. 9-fluorenylmethanol
DIPEA, tetrazole
2. (S)-H-Glu(OMe)OMe
O
P
CO2 Me
R
N
H
CO2 Me
CO2 — Li+
O
O
P
CO2 — Li+
LiOH
R
N
H
Li+ —
O
10 R = Et
11 R = n-Bu
12 R = Ph
1 R = Et
2 R = n-Bu
3 R = Ph
Scheme 1.

J. P. Mallari et al. / Bioorg. Med. Chem. 12 (2004) 6011–6020
6013
RP(O)Cl2
1. CH3C(O)S(CH2 )2SH
2. H-Glu(OMe)OMe
O
P
S
CO2 Me
S
P
S
CO2 Me
R
N
H
CO2 Me
R
N
H
CO2 Me
Lawesson’s
Reagent
O
O
13 R = Et
14 R = n-Bu
15 R = Ph
16 R = Et
17 R = n-Bu
18 R = Ph
S
S
1. NH3 /EtOH
2. LiOH
Chromatographic
resolution
CO2— Li+
S
O
P
S
CO2 Me
O
P
CO2 Me
S
P
CO2— Li+
N
H
CO2 Me
S
N
H
CO2 Me
R
N
H
Li+ —
S
+
O
O
7 R = Et
8 R = n-Bu
9 R = Ph
13a
13b
S
1. L-(-)-a-Methylbenzylamine
2. LiOH
CO2 — Li+
CO2— Li+
S
S
P
P
CO2— Li+
Li+ —
O
CO2— Li+
N
H
N
H
Li+ —
O
4a
4b
Scheme 2.
amine10 was stereospecific leading to pure deprotected
products. Subsequent hydrolysis of the methyl esters
with LiOH afforded 4a and 4b, respectively.
of this investigation, a crystal of phosphonamidate 11a
suitable for X-ray crystallography was identified. This
crystal was available from our previous work aimed at
resolving the individual stereoisomers of 11a through
fractional crystallization.8 The resulting X-ray crystallo-
graphic analysis established the stereochemical configu-
ration of the phosphorus center in 11a as illustrated in
Figure 2.11 With that known, as well as the stereospecif-
icity of the subsequent chemical reactions in Scheme 3
performed previously,8 we were able to unambiguously
determine the absolute configuration of the phospho-
rus center in inhibitor 5a. Briefly, phosphonamidate 11a
was first stereospecifically thionated with LawessonÕs
reagent.12 The final deprotection steps with quinine
We had previously reported that through fractional
crystallization, the individual stereoisomers of interme-
diate 11 possessing antipodal phosphorus stereochemis-
try (11a and 11b) could be obtained.8 Furthermore, the
individual stereoisomers 5a and 5b were obtained
through a sequence of stereospecific thionation and
deprotection steps not involving the chiral phosphorus
center (Scheme 3).8 Diastereomer 11a was crucial in
establishing the absolute configuration of the central
phosphorus atom of inhibitors 4–6. During the course
O
P
CO2Me
S
P
CO2Me
O
N
H
CO2Me
O
N
H
CO2Me
Lawesson's
Reagent
11a
19a
1. Quinine
2. LiOH
S
P
CO2–Li+
Li+–
O
N
H
CO2–Li+
5a
Scheme 3.

6014
J. P. Mallari et al. / Bioorg. Med. Chem. 12 (2004) 6011–6020
occurred through acid-catalyzed hydrolysis via a cyclic
mixed phosphonyl–carbonyl anhydride intermediate,
resulting from direct attack of the a-carboxyl group on
phosphorus.13 Indeed, desulfurization of phosphonami-
dothionates 4–6 at pH7.3 to the corresponding phos-
phonamidates 1–3 was observed by 31P NMR (Table
1). With respect to the individual stereoisomers of phos-
phonamidothionates 4–6, it was noted that for each dia-
stereomeric pair, the isomers exhibiting a greater 31P
NMR chemical shift was consistently less stable to des-
ulfurization at pH7.3 than the respective upfield iso-
mers. We hypothesized that the stereospecific
formation of the purported cyclic mixed phosphonyl–
carbonyl anhydride intermediates13 during desulfuriza-
tion was more energetically favorable for one diastereo-
mer. Accordingly, it would be expected that the two
diastereomers would exhibit, respectively, different rates
of desulfurization. These results indicated that the more
stable stereoisomers of each diastereomeric pair of 4–6
more likely than not possessed the same stereochemical
configuration at phosphorus while those less stable ste-
reoisomers maintained the opposite stereochemical con-
figuration. With the phosphorus stereochemistry of 5a
and established as described above, the analogous
downfield stereoisomers 4a and 6a were thus presumed
to also possess an (R)-configuration at the chiral phos-
phorus center while the respective diastereomers 4b,
5b, and 6b were presumed to maintain the opposite con-
figuration (SP)14 at phosphorus (Table 1).
Figure 2. X-ray crystal structure of phosphonamidate 11a.
and LiOH, reactions which did not involve the chiral
phosphorus center, provided 5a. By comparing the 31P
NMR chemical shift of this material with that of isomers
5a and 5b prepared previously via a route analogous to
that outlined in Scheme 2, the phosphorus stereochemis-
try of both 5a and 5b was established.8 Numerous
unsuccessful attempts aimed at preparing suitable crys-
tals of the remaining diastereomeric phosphonamidoth-
ionate pairs (4a,b and 6a,b) as well as their synthetic
intermediates were made. Therefore, the absolute con-
figurational assignments at the phosphorus atom in
these diastereomeric pairs were surmised by comparing
trends in the 31P NMR chemical shifts and rates of des-
ulfurization in buffer with those of 5a and 5b.
Phosphonamidodithionates 7–9 were prepared as out-
lined in Scheme 2. Following the preparation of the fully
protected phosphonamidothionates 13–15,8 which were
each obtained as mixture of diastereomers unresolved
at phosphorus, thionation with LawessonÕs reagent pro-
vided 16–18 as immediate precursors to the target phos-
phonamidodithionates 7–9. Removal of the protecting
group from the sulfur ligand to phosphorus with meth-
anolic ammonia followed by deprotection of the methyl
esters with LiOH provided phosphonamidodithionates
Stability studies were conducted to confirm that inhibi-
tors 4–6 would be sufficiently stable under the condi-
tions of the enzyme inhibition studies. The impetus for
such studies was based upon earlier reports that demon-
strated desulfurization of phosphonamidothionates
Table 1. The effect of phosphorus ligands and stereochemistry upon the inhibition of glutamate carboxypeptidase
–
X
P
CO2
–
R
N
H
CO2
–Y
Entry
R
X
O
Y
P-configuration
31P NMRa
t1/2 (min)b
Ki (lM)c
1
Et
E
O
S
––
RO
SO
––
––
R
29.12
71.01
69.56
90.54
27.87
69.31
69.01
90.11
15.24
57.28
56.29
81.36
––
63 (1.1)
4a
4b
7
t
t
24.5
49.9
––
2.2 (0.13)
20 (1.6)
25 (1.3)
E
Et
S
S
O
S
S
S
O
S
S
S
S
2
n-Bu
n-Bu
n-Bu
n-Bu
Ph
O
O
O
S
––
0.63 (0.15)
1.2 (0.11)
3.7 (0.17)
22 (1.4)
5a
5b
8
36.5
66.9
––
S
––
––
R
3
O
O
O
S
––
24 (1.2)
6a
6b
9
Ph
Ph
23.0
40.5
––
13 (0.65)
46 (2.0)
530 (71)
S
Ph
––
a Chemical shifts externally referenced to 85% H3PO4 (d = 0.00ppm).
b No detectable degradation was observed for 1–3 and 7–9.
c Standard deviation in parentheses.

J. P. Mallari et al. / Bioorg. Med. Chem. 12 (2004) 6011–6020
6015
7–9 in good yield. It was noted that using a higher con-
centration of methanolic ammonia (7N vs 2M) and in
larger excess (30equiv vs 10equiv) in the first step of
the deprotection resulted in a more complete reaction
and minimized the appearance of side products. The
enantiomeric analogs of 8 and 9, namely 80 and 90, in
which the configuration of the a-carbon on glutamic
acid was (R), were also prepared in the manner de-
scribed by Scheme 2 with the exception that (R)-gluta-
mic acid dimethyl ester was used in the latter half of
the first step for the formation of 140 and 150. In stability
studies at pH7.3, no notable decomposition of the phos-
phonamidodithionates 7–9 was observed.
nothioic acid stereochemistry has the potential to be sig-
nificant with respect to metallopeptidase inhibition, in
some cases (5a/b and 6a/b) we observed that it was rela-
tively independent with Ki values being only two- to
three-fold different for diastereomeric pairs.
The phosphonamidodithionates 7–9 were consistently
less potent inhibitors than either of their respective dia-
stereomeric analogs 4a–6b, with the enzyme apparently
discriminating the least for the smallest alkyl analogs
(4a,b, and 7). It was hypothesized that the loss of inhib-
itory activity against glutamate carboxypeptidase as a
result of the replacement of the oxygen ligand of the
phosphonamidothionates 4–6 with sulfur (7–9), as had
been similarly postulated previously for inhibitors of
carboxypeptidase A, was due to the considerably greater
1
In addition to characterization by H and 13C NMR as
well as high-resolution mass spectrometry, each target
compound (1–9) displayed a single resonance in the
31P NMR spectrum indicating complete deprotection
in the last step and formation of clean product. Once
obtained in sufficient quantity, the individual inhibitors
1–9 were examined for inhibitory potency against gluta-
mate carboxypeptidase employing an HPLC-based
assay to conventionally measure enzymatic hydrolysis
of methotrexate. With Km and VMAX previously deter-
mined (1.6lM and 25lmolminꢀ1 mg-proteinꢀ1, respec-
tively),15 Dixon analyses were performed to obtain Ki
values the inhibitors (Tables 1 and 2). In all analyses,
the initial substrate (methotrexate) concentration was
10lM and triplicate determinations were made and
averaged for each inhibitor concentration, which gener-
ally varied within a range from 5 to 200lM, depending
upon the potency of the individual inhibitors. The corre-
lation coefficient of the Dixon analyses for each inhibi-
tor was greater than 0.99.16 For each analogous set
(R = Et, n-Bu, Ph) of organophosphorus derivatives of
the natural isomer of glutamic acid (Table 1) the most
consistently observed trend in the inhibitory potency
against glutamate carboxypeptidase was held by the dia-
stereomeric pairs of phosphonothionates 4–6. In each
case the (RP)-isomer was a more potent inhibitor than
the (SP)-isomer. The relatively unique inhibitory po-
tency of the 5a and 5b suggests that butyl ligand may
achieve greatest binding to the enzyme, presumably
through hydrophobic interactions, which are not as
complete nor as accessible for the ethyl or phenyl ligands
of 4a,b and 6a,b, respectively. Although the phospho-
˚
length of the P–S bond (1.85A) compared to that of the
17
˚
P–O bond (1.39A). Hence, the difference in bond
length between P–S and P–O could cause sufficient steric
congestion in the vicinity of an active-site zinc ion in a
metalloprotease resulting in weaker enzyme–inhibitor
interactions. Not surprising, in the case where the steric
bulk contributions by the alkyl ligand to phosphorus
was less significant, specifically when R = Et, the effect
of such increased steric congestion due to the two sulfur
ligands of the phosphonamidodithionates was observed
to be less dramatic.
When the inhibitory potencies of phosphonamidates 1–3
were compared to their respective phosphonamidoth-
ionate analogs 4–6, a different relationship was observed
in each case. While the n-butylphosphonamidate 2 was
more potent than both phosphonamidothionate diaste-
reomers 5a and 5b and the contrary was true for the
ethyl analogs 1, 4a, and 4b. The phenylphosphonami-
date 3 exhibited intermediate inhibitory potency when
compared to both phosphonamidothionates 6a and 6b.
As such, no simple trend could be identified for the com-
parison of phosphonamidates with the corresponding
phosphonamidothionates.
In order to determine if the stereochemistry of the a-car-
bon on glutamic acid could attenuate the inhibitory po-
tency of the phosphorus moiety, the inhibition of
glutamate carboxypeptidase by phosphonamidates 2
and 3 and phosphonamidodithionates 8 and 9 were
Table 2. The effects of a-carbon stereochemistry upon the inhibition of glutamate carboxypeptidase
–
X
P
CO2
–
R
N
H
CO2
*
–X
Entry
R
X
Y
C-configuration
Ki (lM)a
2
n-Bu
n-Bu
n-Bu
n-Bu
Ph
O
O
S
O
O
S
S
R
S
R
S
R
S
R
63 (1.1)
ND
20
8
22 (1.4)
14 (3.2)
24 (1.2)
ND
80
3
S
S
O
O
S
O
O
S
30
Ph
Ph
9
90
530 (71)
77 (5.0)
Ph
S
S
a ND = no detectable inhibition at 500lM inhibitor. Standard deviation in parentheses.

6016
J. P. Mallari et al. / Bioorg. Med. Chem. 12 (2004) 6011–6020
compared with their enantiomers 20, 30, 80, and 90,
respectively. When the configuration of glutamate was
inverted as in the phosphonamidates 20 and 30, a com-
plete loss of inhibitory potency was observed. Assuming
that the glutamate carboxylates served to anchor the
inhibitor into the binding site, these results suggested
that inverting stereochemistry at the a-carbon could pre-
vent the inhibitors 20 and 30 from adopting a conforma-
tion in which the phosphonyl oxygens could interact
sufficiently with an active-site zinc ion. Conversely and
surprisingly, the unnatural glutamate-containing phos-
phonamidodithionates 80 and 90 proved to be more
potent inhibitors than their respective naturally-derived
enantiomers 8 and 9. For the phosphonamidodithio-
nates 8 and 9, steric crowding in the active site due to
the two sulfur ligands was hypothesized to be responsi-
ble for the exhibited reduced inhibitory potency when
compared to analogous phosphonamidothionates (5
and 6) and phosphonamidates 2 and 3. Thus it was pos-
tulated that by inverting the stereochemistry of the
glutamate a-carbon of phosphonamidodithionates 8
and 9, their enantiomers 80 and 90 were able to adopt
conformations in which the dithiophosphonyl moiety
imparted less steric congestion in the active site than 80
and 90 and that inhibitory potency could be recovered.
tive to TMS (d = 0.00ppm), CDCl3 (d = 7.26ppm), or
CD3OD (d = 4.87 and 3.31ppm). 13C NMR chemical
shifts are relative to CD3OD (d = 49.15ppm) or CDCl3
(d = 77.23ppm). 31P NMR chemical shifts in CDCl3,
CD3OD, or D2O were externally referenced to 85%
H3PO4 (d = 0.00ppm) in CDCl3, CD3OD, and D2O,
respectively. Combustion analyses were performed by
Quantitative Technologies Inc., Whitehouse, NJ. High-
resolution mass spectra (FAB) were performed by the
University of Notre Dame Mass Spectrometry Facility,
Notre Dame, IN 46556–5670. X-ray crystal analysis was
performed by Dr. Stephen Pavkovic, Loyola University
of Chicago, Chicago, IL and the crystal structure of 11a
was deposited at the Cambridge Crystallographic Data
Centre (deposition number CCDC 200063). Crystallo-
graphic data (excluding structure factors) for the struc-
tures in this paper have been deposited with the
Cambridge Crystallographic Data Centre as supplemen-
tary publication nos. CCDC 200063. Copies if the data
can be obtained, free of charge, on application to
CCDC, 12 Union Road, Cambridge, CB2 1EZ, UK
(fax: +44 (0)1223 336033 or e-mail: deposit@ccdc.
cam.ac.uk).
4.1.2. General procedure for phosphonamidates 10–12.
A solution of 9-flourenemethanol (2.38mmol, 0.467g)
and DIPEA (2.38mmol, 0.308g) in benzene (8.0mL)
was added dropwise to a stirred solution of the appro-
priate alkyl or arylphosphonic dichloride (2.61mmol),
1H-tetrazole (0.47mmol, 0.033g), and benzene (8.0mL)
under an Ar(g) atmosphere at 4ꢁC. The solution was stir-
red for 3h during which it warmed to room temperature.
A solution of neutralized L-glutamic acid dimethyl ester
(2.85mmol, 0.499g), DIPEA (2.85mmol, 0.368g), and
benzene (3.5mL) was added and the reaction mixture
was stirred for an additional 3h, filtered, and concen-
trated in vacuo to a yellow oil. The products were puri-
fied by flash chromatography, to give white solids.
3. Conclusion
In summary, the most potent inhibitor of glutamate
carboxypeptidase identified in this study, as determined
by Ki values, was phosphonamidate 2 in which the phos-
phonyl moiety possessed the natural isomer of glutamic
acid and the n-butyl group as the remaining ligands to
phosphorus. For the diastereomeric pairs of phosphon-
amidothionates 4–6, the least stable inhibitor of each
pair (4a, 5a, 6a), presumably possessing R-configuration
at phosphorus, consistently exhibited greater stereose-
lective inhibition than its diastereomer. Although the
natural stereochemistry of glutamic acid was presumed
to be required for effective inhibitor binding to the active
site of glutamate carboxypeptidase, analogs derived
from the unnatural amino acid exhibited a surprisingly
improved inhibitory potency only when they possessed
a phosphorus moiety of greater steric bulk as was the
case with phosphonamidodithionates 70–90. It is
expected that these results will be of value in the consid-
eration of inhibitor designs for other glutamate carboxy-
peptidases of medical importance such as PSMA and
NAALADase.
4.1.3. N-[9-Fluorenylmethoxy(ethyl)phosphinyl]-L-gluta-
mic acid dimethyl ester (10). Chromatography condi-
tions: ethyl acetate/hexane 3:1 v:v, Rf = 0.14. Yield:
27%. 1H NMR (acetone-d6): d 1.05 (dt, J = 11.1,
4.5Hz, 3H), 1.63–1.90 (m, 4H), 2.39–2.46 (m, 2H),
3.55 and 3.57 (s, 3H), 3.63 and 3.64 (s, 3H), 3.94–4.31
(m, 4H), 7.27–7.41 (m, 4H), 7.69 (d, J = 7.2Hz, 2H),
7.83 (d, J = 7.2Hz, 2H). 31P NMR (CDCl3): d 36.85,
37.49. Anal. Calcd for C23H28NO6P: C, 62.0; H, 6.3;
N, 3.1. Found: C, 61.67; H, 6.49; N, 3.33.
4. Experimental
4.1. Synthesis
4.1.4.
N-[9-Fluorenylmethoxy(n-butyl)phosphinyl]-L-
glutamic acid dimethyl ester (11). Chromatography con-
ditions: ethyl acetate/hexane 4:1 v:v, Rf = 0.28. Yield:
33%. H NMR (CD3OD): d 0.89 (t, J = 7.2Hz, 3H),
1.30–1.51 (m, 4H), 1.65–1.88 (m, 3H), 2.04–2.11 (m,
1H), 2.38–2.44 (m, 2H), 3.62 and 3.66 (s, 3H), 3.68 (s,
3H), 3.83–3.92 (m, 1H), 4.18–4.39 (m, 3H), 7.30–7.43
(m, 4H), 7.67 (t, J = 6.9Hz, 2H), 7.81 (d, J = 7.5Hz,
2H). 31P NMR (CD3OD): d 38.31, 39.08. Anal. Calcd
for C25H32NO6P: C, 63.3; H, 6.8; N, 3.14. Found: C,
63.37; H, 6.83; N, 2.95.
1
4.1.1. General. All solvents used in reactions and diiso-
propylethylamine (DIPEA) were either obtained anhy-
drous or freshly distilled prior to use. All other
reagents were used as supplied unless otherwise stated.
Liquid (flash) chromatography was carried out using
silica gel 60 (230–400mesh). 1H, 13C, and 31P NMR
spectra were recorded on a Bruker DRX 300MHz
1
NMR spectrometer. H NMR chemical shifts are rela-

J. P. Mallari et al. / Bioorg. Med. Chem. 12 (2004) 6011–6020
6017
4.1.5. N-[9-Fluorenylmethoxy(n-butyl)phosphinyl]-D-glut-
amic acid dimethyl ester (110). Compound 110 was pre-
pared from D-glutamic acid dimethyl ester. The NMR
spectra of 110 were identical to that of 11 above.
4.1.13. N-[Hydroxy(phenyl)phosphinyl]-D-glutamic acid
trilithium salt (30). Compound 30 was prepared from
1
31
120. The H, 13C, P NMR spectra of 30 were identical
to that of 3 above.
4.1.14. General procedure for phosphonamidothiolates
13–15.
4.1.6. N-[9-Fluorenylmethoxy(phenyl)phosphinyl]-L-glut-
amic acid dimethyl ester (12). Chromatography condi-
tions: ethyl acetate/hexane 2:1 v:v, Rf = 0.23. Yield:
A solution of ethanedithiol monoacetate
(6.0mmol, 0.816g) and DIPEA (6.0mmol, 0.775g) in
benzene (22mL) was added dropwise to a stirring solu-
tion containing alkylphosphonic dichloride (6.6mmol),
1H-tetrazole (1.2mmol, 0.084g), and benzene (22mL)
under Ar(g) at 4ꢁC. The solution was stirred for 3h dur-
ing which it warmed to room temperature. A solution of
neutralized L-glutamic acid dimethyl ester (7.2mmol,
1.261g), DIPEA (7.2mmol, 0.931g), and benzene
(15.0mL) was added and stirred for an additional 3h.
The reaction mixture was filtered, concentrated in vacuo
to a yellow oil, and then purified by flash chromatogra-
phy. Phosphonamidothiolate 13 was further separated
chromatographically (ethyl acetate/hexane 10:1 v:v,
Rf = 0.25 and 0.34) into the individual stereoisomers
13a and 13b to ultimately provide 4a and 4b,
respectively.
1
36%. H NMR (CD3OD): d 1.78–1.88 (m, 1H), 1.98–
2.05 (m, 1H), 2.29–2.36 (m, 2H), 3.55 and 3.59 (s, 3H),
3.60 and 3.61 (s, 3H), 3.74–3.88 (m, 1H), 4.24–4.47 (m,
3H), 7.27–7.74 (m, 11H), 7.82 (d, J = 7.5Hz, 2H). 31P
NMR (CD3OD): d 24.597, 25.299. Anal. Calcd for
C27H28NO6P: C, 65.7; H, 5.7; N, 2.8. Found: C, 65.57;
H, 5.73; N, 2.80.
4.1.7. N-[9-Fluorenylmethoxy(phenyl)phosphinyl]-L-glut-
amic acid dimethyl ester (120). Compound 120 was pre-
pared from D-glutamic acid dimethyl ester. The NMR
spectra of 120 were identical to that of 12 above.
4.1.8. General procedure for phosphonamidoic acids 1–3.
Phosphonamidates 10–12 (0.202mmol) were each dis-
solved in methanol (1.06mL) to which an aqueous solu-
tion of LiOH (1M, 1.013mmol) was added. The
reaction mixture was stirred 14h, filtered, and concen-
trated in vacuo to dryness. The resulting solid was sus-
pended in anhydrous methanol, filtered (0.2lm Teflon
membrane), and concentrated in vacuo to yield the
respective trilithium salts.
4.1.15. (RP)-N-[b-(Acetylmercapto)ethylthio(ethyl)phos-
phinyl]-L-glutamic acid dimethyl ester (13a). Yield:
36%. 1H NMR (CDCl3): d 1.16–1.28 (m, 3H), 1.92–
2.20 (m, 4H), 2.34 (s, 3H), 2.40–2.51 (m, 2H), 2.92–
2.99 (m, 2H), 3.14–3.20 (m, 2H), 3.68 (s, 3H), 3.76 (s,
3H), 4.00–4.13 (m, 1H). 31P NMR (CDCl3): d 53.58.
Anal. Calcd for C13H24NO6PS2: C, 40.5; H, 6.3; N,
3.6. Found: C, 41.35; H, 6.32; N, 3.60.
4.1.9. N-[Hydroxy(ethyl)phosphinyl]-L-glutamic acid tri-
1
lithium salt (1). Yield: 60%. H NMR (CD3OD): d 1.03
4.1.16. (SP)-N-[b-(Acetylmercapto)ethylthio(ethyl)phos-
phinyl]-L-glutamic acid dimethyl ester (13b). Yield:
36%. 1H NMR (CDCl3): d 1.17–1.30 (m, 3H), 1.92–
2.19 (m, 4H), 2.33 (s, 3H), 2.44–2.49 (m, 2H), 2.93–
2.98 (m, 2H), 3.10–3.15 (m, 2H), 3.65 (s, 3H), 3.76 (s,
3H), 4.07–4.18 (m, 1H). 31P NMR (CDCl3): d 52.68.
Anal. Calcd for C13H24NO6PS2: C, 40.5; H, 6.3; N,
3.6. Found: C, 40.04; H, 6.28; N, 3.60.
(dt, J = 18.0, 7.5Hz, 3H), 1.40–1.53 (m, 2H), 1.80–1.98
(m, 2H), 2.19–2.31 (m, 2H), 3.47–3.54 (m, 1H). 13C
NMR (CD3OD): d 8.02, 8.11, 22.11, 23.74, 33.54,
33.59, 35.08, 57.72, 183.17, 184.05. 31P NMR (CD3OD):
d 29.12.
4.1.10. N-[Hydroxy(n-butyl)phosphinyl]-L-glutamic acid
trilithium salt (2). Yield: 75%. 1H NMR (D2O): d 0.79 (t,
J = 7.2Hz, 3H), 1.23–1.45 (m, 6H), 1.70–1.75 (m, 2H),
2.11–2.16 (m, 2H), 3.35–3.38 (m, 1H). 13C NMR
(D2O): d 24.00, 25.48, 25.54, 28.26, 29.88, 32.72, 32.78,
56.89, 182.35, 183.43. 31P NMR (CD3OD): d 27.87.
FAB-HRMS (M ꢀ Li)ꢀ calcd 278.0955, found
278.0958 for C9H14Li3NO6P.
4.1.17. N-[b-(Acetylmercapto)ethylthio(n-butyl)phosphin-
yl]-L-glutamic acid dimethyl ester (14). Chromatography
conditions: ethyl acetate/hexane 10:1 v:v, Rf = 0.26.
1
Yield: 29%. H NMR (CDCl3): d 0.91–0.97 (m, 3H),
1.38–1.50 (m, 2H), 1.59–1.70 (m, 2H), 1.91–2.04 (m,
3H), 2.10–2.22 (m, 1H), 2.34 (s, 3H), 2.39–2.46 (m,
2H), 2.89–3.03 (m, 2H), 3.12–3.20 (m, 2H), 3.48–3.61
(m, 1H), 3.67 and 3.69 (s, 3H), 3.76 and 3.77 (s, 3H),
4.03–4.15 (m, 1H). 31P NMR (CDCl3): d 50.99,
51.36.
4.1.11. N-[Hydroxy(n-butyl)phosphinyl]-D-glutamic acid
trilithium salt (20). Compound 20 was prepared from 110.
The H, 13C, P NMR spectra of 20 were identical to
that of 2 above.
1
31
4.1.18. N-[b-(Acetylmercapto)ethylthio(n-butyl)phosphin-
yl]-D-glutamic acid dimethyl ester (140). Compound 140
was prepared from D-glutamic acid dimethyl ester. The
NMR spectra of 140 were identical to that of 14
above.
4.1.12. N-[Hydroxy(phenyl)phosphinyl]-L-glutamic acid
trilithium salt (3). Yield: 75%. H NMR (D2O): d 1.77–
1
1.86 (m, 2H), 2.09–2.19 (m, 2H), 3.39–3.46 (m, 1H),
7.27–7.28 (m, 3H), 7.67–7.70 (m, 2H). 13C NMR
(D2O): d 32.48, 34.01, 56.72, 128.30, 128.47, 130.38,
130.77, 130.89, 181.58, 183.35. 31P NMR (CD3OD): d
12.04. FAB-HRMS (M ꢀ Li)ꢀ calcd 298.0642, found
298.0645 for C11H11Li3NO6P.
4.1.19. N-[b-(Acetylmercapto)ethylthio(phenyl)phosphin-
yl]-L-glutamic acid dimethyl ester (15). Chromatography
conditions: ethyl acetate/hexane 4:1 v:v, Rf = 0.22.
1
Yield: 19%. H NMR (CDCl3): d 1.92–2.04 (m, 1H),

6018
J. P. Mallari et al. / Bioorg. Med. Chem. 12 (2004) 6011–6020
2.09–2.21 (m, 1H), 2.29 (s, 3H), 2.43–2.50 (m, 2H), 2.88–
3.01 (m, 2H), 3.05–3.11 (m, 2H), 3.62 and 3.67 (s, 3H),
3.70 and 3.77 (s, 3H), 3.99–4.19 (m, 1H), 7.45–7.59 (m,
3H), 7.83–7.93 (m, 2H). 31P NMR (CDCl3): d 32.027,
32.942.
4.1.26. N-[b-(Acetylmercapto)ethylthio(n-butyl)phosphin-
othioyl]-L-glutamic acid dimethyl ester (17). Chromato-
graphy conditions: CH2Cl2/ethyl acetate 50:1 v:v,
Rf = 0.24. Yield: 76%. 1H NMR (CDCl3): d 0.58 (t,
J = 7.2Hz, 3H), 1.40–1.50 (m, 2H), 1.62–1.72 (m, 2H),
1.89–2.20 (m, 4H), 2.34 (s, 3H), 2.39–2.49 (m, 2H),
2.94–3.16 (m, 4H), 3.35–3.51 (m, 1H), 3.68 and 3.68 (s,
3H), 3.76 (s, 3H), 4.06–4.26 (m, 1H). 31P NMR (CDCl3):
d 81.05, 82.45. Anal. Calcd for C15H28NO5PS3: C, 41.9;
H, 6.6; N, 3.3. Found: C, 41.97; H, 6.54; N, 3.27.
4.1.20. N-[b-(Acetylmercapto)ethylthio(phenyl)phosphin-
yl]-D-glutamic acid dimethyl ester (150). Compound 150
was prepared from D-glutamic acid dimethyl ester.
The NMR spectra of 150 were identical to that of 15
above.
4.1.27.
N-[b-(Acetylmercapto)ethylthio(n-butyl)phos-
phinothioyl]-D-glutamic acid dimethyl ester (170). Com-
pound 170 was prepared from 140. The NMR spectra
of 170 were identical to that of 17 above.
4.1.21. General procedure for phosphonamidothioic acids
4a and 4b. A single isomer of phosphonamidothiolate 13
(0.155mmol, 0.060g) and L-(ꢀ)-a-methylbenzylamine
(0.326mmol, 0.040g) was refluxed in methanol
(3.0mL) 13h followed by concentration in vacuo to give
a yellow oil. The resulting crude product was dissolved
in water and washed with a diethyl ether/hexane mixture
(1:1, v:v). The aqueous layer was concentrated in vacuo
to yield the intermediate methylbenzylammonium phos-
phonamidothioate (0.141mmol, 0.057g), which was
subsequently dissolved in methanol (0.71mL) and stir-
red 14h with an aqueous solution of LiOH (1M,
0.71mL). The reaction mixture was concentrated
in vacuo to dryness, suspended in anhydrous methanol,
filtered (0.2lm Teflon membrane), and concentrated
in vacuo to yield the trilithium salts 4a and 4b.
4.1.28.
phinothioyl]-L-glutamic acid dimethyl ester (18). Chroma-
N-[b-(Acetylmercapto)ethylthio(phenyl)phos-
tography conditions: ethyl acetate/hexane 4:1 v:v,
1
Rf = 0.31. Yield: 74%. H NMR (CDCl3): d 1.99–2.03
(m, 1H), 2.08–2.12 (m, 1H), 2.30 (s, 3H), 2.38–2.47 (m,
2H), 2.99–3.10 (m, 4H), 3.61 and 3.67 (s, 3H), 3.70
and 3.75 (s, 3H), 4.09–4.12 (m, 1H), 7.45–7.56 (m,
3H), 8.01 (q, J = 7.5Hz, 2H). 31P NMR (CDCl3): d
72.51, 74.15. Anal. Calcd for C17H24NO5PS3: C, 45.4;
H, 5.4; N, 3.1. Found: C, 45.34; H, 5.28; N, 3.11.
4.1.29.
N-[b-(Acetylmercapto)ethylthio(phenyl)phos-
phinothioyl]-D-glutamic acid dimethyl ester (180). Com-
pound 180 was prepared from 150. The NMR spectra
of 180 were identical to that of 18 above.
4.1.22. (RP)-N-[Hydroxy(ethyl)phosphinothioyl]-L-gluta-
mic acid trilithium salt (4a). Yield: 86%. 1H NMR
(D2O): d 0.93–1.07 (m, 3H), 1.64–1.92 (m, 4H), 2.11–
2.17 (m, 2H), 3.42–3.57 (m, 1H). 13C NMR (D2O): d
7.75, 7.81, 30.09, 31.35, 32.32, 32.38, 34.17, 57.14,
182.04, 183.30. 31P NMR (D2O): d 71.01.
4.1.30. General procedure for phosphonamidodithioic
acids
7–9.
Phosphonamidodithionates
16–18
(0.114mmol) were each dissolved in methanol
(0.188mL) and reacted with a 7N methanolic solution
of NH3 (0.482mL) without stirring under an Ar(g)
atmosphere for 45min. The product mixtures were
immediately concentrated in vacuo, washed with diethyl
ether, and filtered to yield white solids. Each solid was
dissolved in methanol (0.385mL) and stirred 14h with
an aqueous 1M LiOH solution (0.385mL). The reaction
mixture was concentrated in vacuo to dryness, sus-
pended in anhydrous methanol, and filtered (0.2lm
Teflon membrane) to, respectively, give phospho-
namidodithioic acids 7–9 as the trilithium salts.
4.1.23. (SP)-N-[Hydroxy(ethyl)phosphinothioyl]-L-gluta-
mic acid trilithium salt (4b). Yield, 86%. 1H NMR
(D2O): d 0.92–1.04 (m, 3H), 1.63–1.94 (m, 4H), 2.11–
2.17 (m, 2H), 3.48–3.61 (m, 1H). 13C NMR (D2O): d
7.92, 7.97, 29.25, 30.51, 32.45, 32.51, 34.28, 57.08,
182.09, 183.29. 31P NMR (D2O): d 69.56.
4.1.24. General procedure for phosphonamidodithionates
16–18. Diastereomeric mixtures of phosphonamidothio-
lates 13–15 (0.451mmol) each were refluxed 2.5h in tol-
uene (7.0mL) with LawessonÕs reagent (0.248mmol,
0.100g). The reaction mixture was concentrated in va-
cuo to a yellow oil, which was purified by flash
chromatography.
4.1.31. N-[Hydroxy(ethyl)phosphinodithioyl]-L-glutamic
acid trilithium salt (7). Yield: 31%. H NMR (CD3OD):
1
d 0.71–0.78 (m, 3H), 1.42–1.57 (m, 4H), 1.71–1.92 (m,
2H), 3.35–3.44 (m, 1H). 13C NMR (CD3OD): d 9.43,
33.97, 35.51, 39.19, 40.16, 59.07, 183.13, 183.46. 31P
NMR (D2O): d 90.54. FAB-HRMS (M ꢀ Li)ꢀ calcd
282.0185, found: 282.0187 for C7H11Li3NO4PS2.
4.1.25. N-[b-(Acetylmercapto)ethylthio(ethyl)phosphinoth-
ioyl]-L-glutamic acid dimethyl ester (16). Chromatogra-
phy conditions: CH2Cl2/ethyl acetate 100:3 v:v,
1
Rf = 0.35. Yield: 70%. H NMR (CDCl3): d 1.24–1.32
4.1.32. N-[Hydroxy(n-butyl)phosphinodithioyl]-L-gluta-
mic acid trilithium salt (8). Yield: 36%. 1H NMR
(CD3OD): d 0.96 (t, J = 7.5Hz, 3H), 1.41–1.50 (m,
2H), 1.61–1.72 (m, 2H), 1.88–1.97 (m, 2H), 2.05–2.13
(m, 2H), 2.27–2.32 (m, 2H), 3.72–3.81 (m, 1H). 13C
NMR (D2O): d 13.17, 13.38, 23.00, 23.25, 26.39, 32.09,
34.10, 43.49, 44.42, 57.63, 181.75, 181.16. 31P NMR
(m, 3H), 1.90–2.20 (m, 4H), 2.34 (s, 3H), 2.40–2.47 (m,
2H), 3.00–3.22 (m, 4H), 3.37–3.50 (m, 1H), 3.68 and
3.69 (s, 3H), 3.76 (s, 3H), 3.87–4.27 (m, 1H). 31P
NMR (CDCl3): d 89.43, 90.82. Anal. Calcd for
C13H24NO5PS3: C, 38.9; H, 6.0; N, 3.5. Found: C,
39.14; H, 5.91; N, 3.37.

J. P. Mallari et al. / Bioorg. Med. Chem. 12 (2004) 6011–6020
6019
(CD3OD): d 86.17. FAB-HRMS (M ꢀ Li)ꢀ calcd
with a mobile phase of CH3OH/(potassium phosphate,
50mM, pH6.8) (22:78, v:v). At a flow rate of 0.9mL/
min, methotrexate and its hydrolytic product were de-
tected at 304nm with retention times of 3.8 and
7.7min, respectively. Under the assay conditions de-
scribed above, it was noted that the initial substrate con-
centration was not substantially depleted during the
time course of the incubation (e.g., approximately 10%
conversion to product was observed for incubations
with the lowest substrate concentration, 1lM).
310.0498, found: 310.0501 for C9H15Li3NO4PS2.
4.1.33. N-[Hydroxy(n-butyl)phosphinodithioyl]-D-gluta-
mic acid trilithium salt (80). Compound 80 was prepared
from 170. The NMR spectra of 80 were identical to that
of 8 above.
4.1.34. N-[Hydroxy(phenyl)phosphinodithioyl]-L-glutamic
acid trilithium salt (9). Yield: 67%. H NMR (CD3OD):
1
d 1.78–1.90 (m, 4H), 3.69–3.77 (m, 1H), 7.21–7.27 (m,
3H), 8.08–8.15 (m, 2H). 13C NMR (CD3OD): d 33.55,
35.35, 59.12, 128.72, 128.35, 130.12, 131.42, 131.57,
181.86, 183.62. 31P NMR (CD3OD): d 74.43. FAB-
HRMS (M ꢀ Li)ꢀ calcd 330.0185, found: 330.0188 for
C11H11Li3NO4PS2.
Acknowledgements
This work was supported in part by grants from the Na-
tional Institutes of Health, MBRS SCOREProgram-
NIGMS (Grant No. S06 GM52588-04), and fellowships
from the Department of Defense Science Scholars Pro-
gram (Grant No. 6-93472), MBRS-RISEProgram-
NIGMS (Grant No. GM59298-02), and NIH-PREP
Scholarship Program (Grant No. 1R25GM64078-01).
4.1.35. N-[Hydroxy(phenyl)phosphinodithioyl]-D-gluta-
mic acid trilithium salt (90). Compound 90 was prepared
from 180. The NMR spectra of 90 were identical to that
of 9 above.
4.2. Stability studies of inhibitors
References and notes
Inhibitors 1–3, 4a–6a, 4b–6b, and 7–9 (10lmol) were
each dissolved in 1mL of Tris buffer (100mM, pH7.3)
to give a 10mM inhibitor solution and 0.5mL of the
resulting solution was transferred to a 5mm NMR tube.
31P NMR data were acquired every 10–20min for up to
10h from the time when the buffer was added.
1. Robinson, M. B.; Blakely, R. D.; Couto, R.; Coyle, J. T. J.
Biol. Chem. 1987, 262, 14498.
2. Carter, R. E.; Feldman, A. R.; Coyle, J. T. Proc. Natl.
Acad. Sci. U.S.A. 1996, 93, 749.
3. Reisenauer, A. M.; Krumdieck, C. L.; Halsted, C. H.
Science 1977, 198, 196.
4. (a) Goldman, P.; Levy, C. C. Proc. Natl. Acad. Sci. U.S.A.
1967, 58, 1299; (b) Levy, C. C.; Goldman, P. J. Biol.
Chem. 1967, 242, 2933; (c) McCullough, J. L.; Chabner, B.
A.; Bertino, J. R. J. Biol. Chem. 1971, 246, 7207; (d)
Sherwood, R. F.; Melton, R. G.; Alwan, S. M.; Hughes, P.
Eur. J. Biochem. 1985, 148, 447.
5. Kahn, T. H.; Eno-Amooquaye, E. A.; Searle, F.; Browne,
P. J.; Osborn, H. M. I.; Burke, P. J. J. Med. Chem. 1999,
42, 951.
6. (a) Rodriguez, C. E.; Lu, H.; Dinh, T. T.; Mlodnosky, K.
L.; Dastgah, A.; Lam, V. Q.; Nichols, C. B.; Berkman, C.
E. Bioorg. Med. Chem. Lett. 1999, 9, 1415; (b) Lu, H.;
Berkman, C. E. Bioorg. Med. Chem. 2001, 9, 395; (c)
Rodriguez, C. E.; Lu, H.; Martinez, A. R.; Brunelle, A.;
Berkman, C. E. J. Enzym. Inhib. 2001, 16, 359.
7. (a) Bell, C. F. In Principles and Applications of Metal
Chelation; Atkins, P. W., Holker, J. S. E., Holliday, A. K.,
Eds.; Oxford University Press: Oxford, 1977; (b) Prasad,
A. S. In Biochemistry of Zinc; Frieden, E., Ed.; Plenum:
New York, 1993.
8. Lu, H.; Hu, Y.; Choy, C. J.; Mallari, J. P.; Villanueva, A.
F.; Arrozal, A. F.; Berkman, C. E. Tetrahedron Lett. 2001,
42, 4313.
9. Watanabe, Y.; Nakatomi, M. Tetrahedron Lett. 1998, 39,
1583; Seeberger, P. H.; Yau, E.; Caruthers, M. H. J. Am.
Chem. Soc. 1995, 117, 1472.
10. Hall, C. R.; Inch, T. D. Phosphorus Sulfur 1979, 7, 171.
11. Compound information: exptl crystal color = colorless;
chemical melting point = 383–385 ꢁC; exptl crystal descrip-
tion = prismatic; exptl crystal preparation = acetone–hex-
ane. Unit cell information: cell length a = 9.729; cell length
b = 14.694; cell length c = 17.600; cell angle a = 90.00; cell
angle b = 90.00; cell angle c = 90.00; cell volume = 2516.0;
symmetry cell setting = orthorhombic; symmetry space
group name H-M = P 21 21 21; cell formula units Z = 4;
exptl crystal density = 1.250.
4.3. Enzyme inhibition
4.3.1. General. Glutamate carboxypeptidase (carboxy-
peptidase G from Pseudomonas sp. strain ATCC
25301, C 4053) and methotrexate were obtained from
Sigma Chemical Co. (St. Louis, MO). All other chemi-
cals were of the highest purity and purchased from com-
mercial sources.
4.3.2. Inhibition assay procedures. Inhibition of gluta-
mate carboxypeptidase by the synthetic inhibitors was
determined as described previously.6 Briefly, a typical
incubation mixture (final volume 0.25mL) was prepared
by the addition of 200lL Tris buffer (50mM, pH7.3) to
either a 25lL mixture of both methotrexate and inhibi-
tor in buffer or 25lL of a buffered solution of metho-
trexate alone. The enzymatic reaction was initiated by
the addition of 25lL of an enzyme solution (0.17–
0.21lgprotein/mL buffer). In all cases, the final concen-
tration of methotrexate was 10lM while the inhibitor
concentration varied from 7.5 to 120lM. The reaction
was allowed to proceed for 1min with constant shaking
at 30ꢁC and was terminated by the addition of 100lL
methanolic TFA (1% trifluoroacetic acid by volume in
methanol) followed by vortexing and centrifugation
(7000g). A 100lL aliquot of the resulting supernatant
was then quantified by subsequently quantified by
HPLC. Methotrexate and its hydrolytic product (4-(N-
[2,4-diamino-6-pteridinylmethyl]-N-methylamino)ben-
zoic acid) were separated and quantified with an
analytical reversed phase HPLC column (4.6 · 150mm,
Sphereclone 5u ODS(2), Phenomenex, Torrence, CA)

6020
J. P. Mallari et al. / Bioorg. Med. Chem. 12 (2004) 6011–6020
12. Jackson, J. A.; Berkman, C. E.; Thompson, C. M.
Tetrahedron Lett. 1992, 41, 6061.
13. Jacobsen, N. E.; Bartlett, P. A. J. Am. Chem. Soc. 1982,
105, 1619–1626.
15. Lu, H.; Ng, R.; Shieh, C. C.; Martinez, A. R.; Berkman,
C. E. Phosphorus, Sulfur Silicon 2003, 178, 17.
16. The correlation coefficient for the Dixon analysis of
compounds 4b and 8 were both 0.97.
14. (RP) and (SP) refer to the stereochemical configuration of
phosphorus.
17. Hill, J. M.; Lowe, G. (1995) J. Chem. Soc., Perkin Trans. 1
2001.