γ-(Monophenyl)phosphono glutamate analogues as mechanism-based inhibitors of γ-glutamyl transpeptidase

Article abstract of DOI:10.1016/j.bmc.2006.05.008

γ-Glutamyl transpeptidase (GGT, EC 2.3.2.2) catalyzes the hydrolysis and transpeptidation of extracellular glutathione and plays a central role in glutathione homeostasis. We report here the synthesis and evaluation of a series of hydrolytically stable γ-

Full text of DOI:10.1016/j.bmc.2006.05.008

Bioorganic & Medicinal Chemistry 14 (2006) 6043–6054
c-(Monophenyl)phosphono glutamate analogues
as mechanism-based inhibitors of c-glutamyl transpeptidase
Liyou Han,a Jun Hiratake,a,* Norihito Tachi,a Hideyuki Suzuki,b
Hidehiko Kumagaib,ꢀ and Kanzo Sakataa
aInstitute for Chemical Research, Kyoto University, Uji, Kyoto 611-0011, Japan
bGraduate School of Biostudies, Kyoto University, Kitashirakawa, Sakyo-ku, Kyoto 606-8502, Japan
Received 26 March 2006; revised 2 May 2006; accepted 3 May 2006
Available online 23 May 2006
Abstract—c-Glutamyl transpeptidase (GGT, EC 2.3.2.2) catalyzes the hydrolysis and transpeptidation of extracellular glutathione
and plays a central role in glutathione homeostasis. We report here the synthesis and evaluation of a series of hydrolytically stable
c-(monophenyl)phosphono glutamate analogues with varying electron-withdrawing para substituents on the leaving group phenols
as mechanism-based and transition-state analogue inhibitors of Escherichia coli and human GGTs. The monophenyl phosphonates
caused time-dependent and irreversible inhibition of both the E. coli and human enzymes probably by phosphonylating the catalytic
Thr residue of the enzyme. The inactivation rate of E. coli GGT was highly dependent on the leaving group ability of phenols with
electron-withdrawing groups substantially accelerating the rate (Brønsted blg = ꢀ1.4), whereas the inactivation of human GGT was
rather slow and almost independent on the nature of the leaving group. The inhibition potency and profiles of the phosphonate
analogues were compared to those of acivicin, a classical inhibitor of GGT, suggesting that the phosphonate-based glutamate ana-
logues served as a promising candidate for potent and selective GGT inhibitors.
ꢀ 2006 Elsevier Ltd. All rights reserved.
1. Introduction
significantly in human tumors, and its roles in tumor
progression12 and the expression of malignant pheno-
types of cancer cells such as drug resistance10,11,13,14
and metastasis15–17 have been repeatedly suggested.18
GGT is also implicated in many physiological disorders
such as neurodegenerative disease,19,20 diabetes,21,22 and
cardiovascular disease23,24 in relation to oxidative stress
and GSH homeostasis.25 Therefore, the important roles
played by GGT in GSH-mediated detoxification and
cellular response to oxidative stress strongly suggest that
GGT is an attractive pharmaceutical target for cancer
chemotherapy and a vast array of physiological disor-
ders related to oxidative stress caused by reactive oxygen
species.
c-Glutamyl transpeptidase (GGT; EC 2.3.2.2) is a hete-
rodimeric enzyme found widely among organisms from
bacteria to mammals and catalyzes the first step in glu-
tathione (GSH) metabolism.1–4 In mammals, GGT is
bound to the external surface of plasma membrane
and is expressed in high concentrations in kidney
tubules, biliary epithelium, and brain capillaries.1,4,5 This
enzyme plays critical roles in GSH homeostasis by
breaking down extracellular GSH to provide cells with
secondary source of cysteine, the rate-limiting substrate
for intracellular GSH biosynthesis,6–9 and in detoxifica-
tion of electrophilic/oxidative xenobiotics through the
metabolism of GSH conjugates to confer resistance
against oxidative stress and anti-tumor drugs such as
cisplatin.10,11 The expression of GGT is often increased
GGT catalyzes the cleavage of the c-glutamyl bond of
GSH, its S-conjugates, and structurally diverse c-glut-
amyl amides to transfer the c-glutamyl group to water
(hydrolysis) or to a variety of amino acids and peptides
(transpeptidation) through a modified ping–pong mech-
anism involving a c-glutamyl enzyme intermediate
(Scheme 1).26,27 Although a number of inhibitors have
been reported to date, little compounds appear to serve
as potent and specific inhibitors of GGT. (aS,5S)-
Keywords: c-Glutamyl transpeptidase; c-(Monophenyl)phosphono
glutamate analogues; Mechanism-based inhibitor; Phosphonylation.
*
Corresponding author. Tel.: +81 774 38 3231; fax: +81 774 38
3229; e-mail: hiratake@scl.kyoto.u-ac.jp
ꢀ
Present address: Research Institute for Bioresources and Biotech-
nology, Ishikawa Prefectural University, Nonoichi-cho, Ishikawa
921-8836, Japan.
a-Amino-3-chloro-4,5-dihydro-5-isoxazoleacetic
acid
0968-0896/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2006.05.008

6044
L. Han et al. / Bioorg. Med. Chem. 14 (2006) 6043–6054
Scheme 2. Design of c-(monophenyl)phosphono glutamate analogues
2a–d as mechanism-based inhibitors of GGT.
Scheme 1. Proposed catalytic mechanism of GGT.
ity toward E. coli GGT. The dependence of the inactiva-
tion rates on the leaving group ability suggested
significant differences in the inactivation transition state
between the E. coli and human enzymes. The phospho-
nate-based glutamate analogues are compared with aci-
vicin in terms of the potency and the mechanism of
inactivation.
(acivicin, AT-125), a natural product derived from
Streptomyces sviceus,28 is a classical and most widely
used irreversible inhibitor of GGT29–31, but is a strong
inhibitor of glutamine-dependent amidotransferases32,33
and inactivates many physiologically important enzymes
involved in purine and pyrimidine biosynthesis to exert
potent cytotoxicity.34,35 Other glutamine antagonists
such as L-azaserine and 6-diazo-5-oxo-L-norleucine
(DON) inhibit GGT,36 but these compounds also serve
as potent inhibitors of glutamine-dependent amid-
otransferases.34 The reversible and weak inhibition of
GGT by L-serine–borate complex37 lead to the develop-
ment of L-2-amino-4-boronobutanoic acid (c-boroGlu),
a c-boronic acid analogue of glutamic acid, as a potent
GGT inhibitor with an inhibition constant (Ki) of a
nanomolar range.38,39 This compound is believed to
form a covalent bond with a hydroxy residue in the
active site to mimic the transition-state of GGT
catalysis, but the inhibition is reversible, and the
inactivated enzyme regained activity rapidly.38
2. Results and discussion
2.1. Synthesis and stability of monophenyl phosphonates
2a–d
The monophenyl phosphonates 2a–d were synthesized
as shown in Scheme 3. The amino group of 2-amino-4-
phosphonobutanoic acid (APBA) was protected with
4-nitrobenzyloxycarbonyl [p(NO2)Z] group, and the
a-carboxy group was esterified selectively under acidic
conditions to afford the methyl ester 4 or the ethyl ester
40 with a free c-phosphono group. The methyl ester 4,
however, was partially hydrolyzed (22%) to the acid 3
We previously reported that 2-amino-4-(fluorophospho-
no)butanoic acid (1), a c-monofluorophosphono deriva-
tive of glutamic acid, served as an extremely effective
mechanism-based affinity labeling agent that inactivated
Escherichia coli GGT with a second-order rate constant
for enzyme inactivation (kon) of 48,000 Mꢀ1 sꢀ1
40
.
This
compound was used successfully for the identification
of the catalytic nucleophile of E. coli GGT as Thr-
391,40 the N-terminal residue of the small subunit, sug-
gesting that GGT is a member of the N-terminal nucle-
ophile hydrolase family.41,42 Compound 1 is a promising
lead for potent and selective inhibitors of GGT, because
it reacts with the catalytic residue of GGT in a mecha-
nism-based manner to form a stable monophosphonate
bond with its catalytic Thr to inactivate the enzyme.
However, the monofluorophosphonate 1 is highly reac-
tive and is hydrolytically unstable in alkaline to neutral
media.40 Herein we report the synthesis and evaluation
of a series of hydrolytically stable c-(monophenyl)phos-
phono glutamate analogues 2a–d with varying electron-
withdrawing groups at the para-position of the leaving
group phenols as mechanism-based inhibitors of E. coli
and human GGTs (Scheme 2). The monophenyl phos-
phonates 2a–d served as irreversible inhibitors that
caused time-dependent inhibition of both the E. coli
and human enzymes, but with significantly higher activ-
Scheme 3. Synthesis of monophenyl phosphonates 2a–d. Reagents: (a)
p(NO2)ZCl, NaHCO3; (b) SOCl2, MeOH or EtOH, 75%; (c) SOCl2,
cat. DMF, 100%; (d) p-XC6H4OH, NaH, dry THF (X = H, CF3, Ac
and CN), 38–97%; (e) i—0.2 N NaOH; ii—H2/5% Pd–C, MeOH or
THF; 2a: 59%, 2b: 16%, 2c: 80%, 2d: 38%.

L. Han et al. / Bioorg. Med. Chem. 14 (2006) 6043–6054
6045
during purification on an ODS column (MeOH–H2O),
probably because the proximal acidic c-phosphonyl
group intramolecularly assisted the transesterification
or hydrolysis of the a-carboxy group. The same transe-
sterification was also observed with the N-protected
monophenyl phosphonate derived from 6b during chro-
matography on an ODS column (MeOH–H2O) to give a
small amount (ca. 10%) of the corresponding methyl
ester (data not shown). This side reaction was complete-
ly prevented by using the ethyl ester 40.
of 2a–d were examined for inhibition of E. coli and
human GGTs as model enzymes and were compared
to those of monofluorophosphonate 1 and acivicin.
2.2. Inactivation of E. coli and human GGT by 2a–d
For evaluating the inhibitory activities of 2a–d, we used
the hydrolase, rather than the transpeptidase, activity of
GGT, because the inhibitors 2a–d have a para-substitut-
ed phenol moiety that may bind to the acceptor site of
GGT.46 Thus, a standard transpeptidase assay using a
high concentration of an acceptor molecule (e.g.,
10 mM Gly-Gly47) was thought to affect the binding of
the inhibitors to underestimate the inhibitory activities,
particularly when a continuous inhibition assay method
is employed (see Section 3). Another reason is that we
wanted to compare the inhibitory activities of 2a–d with
the activity of the enzyme, on the premise that the phos-
phonates 2a–d serve as mechanism-based inhibitors that
exploit the catalytic activity for enzyme inactivation.
Transpeptidation is not a good measure of the intrinsic
enzyme activity, because the apparent transpeptidase
activity is highly dependent on the pKa of the amino
group of the acceptor molecules that affects the effective
concentration of the deprotonated form of the acceptor
amino acid at a given pH.48 Therefore, prior to the inhi-
bition experiment, we measured the pH-dependence of
the hydrolase activities of E. coli and human GGTs
(Fig. 1). The pH-rate profiles indicated that the hydro-
lase activity of E. coli GGT was almost constant from
pH 5.5 to 9.0,49 but the activity of the human enzyme
was highly dependent on pH: the apparent activity at
pH 8 was about eight times higher than that at pH
5.5. The increase in the apparent activity at higher pH
was probably due to the increase in the catalytic activity
of the enzyme, because the activity was measured at
sufficiently low substrate concentration (c-Glu-AMC,
4.0 lM) to prevent autotranspeptidation,47 and the
observed convex upward curve is not likely to reflect
The phosphonic acids 4 and 40 were converted to the
phosphonic diesters 6a–d via the corresponding phos-
phono dichloridates 5 and 50 followed by treatment with
two equivalents of a series of p-substituted phenols and
triethylamine.43 Alkaline hydrolysis of 6a–d followed by
hydrogenolysis afforded the corresponding monophenyl
phosphonates 2a–d. The conditions for selective removal
of the p(NO2)Z group are worth noting: the acetyl and
the cyano groups attached to the phenyl ring are
reduced under a standard condition for catalytic hydrog-
enolysis. In fact, the hydrogenolysis of the mono(4-acet-
ylphenyl) phosphonate derived from 6c in an acidic
medium (MeOH/AcOH/H2O) resulted in complete
reduction of the acetyl group, along with the removal
of the p(NO2)Z group. However, the use of a protic,
but a neutral solvent system (MeOH–H2O) combined
with a less active catalyst (5% Pd–C) successfully cleaved
the p(NO2)Z group without reducing the acetyl group.
Since the cyano group is more susceptible to reduction
than the acetyl group, the use of an aprotic and less
polar solvent such as THF was required to prevent the
reduction of the cyano group of 2d. The use of less polar
solvent, however, was problematic in product recovery,
because the phosphonate monoesters 2b–d were sparing-
ly soluble in such solvents and were precipitated and
deposited on the Pd–C as the reaction proceeded. There-
fore, the products 2b–d were recovered from the reaction
mixture by extracting the Pd–C with hot water contain-
ing trifluoroacetic acid, followed by purification by crys-
tallization or column chromatography (see Section 3).
The hydrolytic stability was examined with the least sta-
ble 2d in D2O at 23 ꢁC. No hydrolysis was observed in
12 h, suggesting that the monophenyl phosphonates
2a–d were highly stable even with an electron-withdraw-
ing group attached to the para position of the phenyl
ring. The hydrolytic stability of 2a–d was also examined
at higher temperature (60 ꢁC) at pH 5.5 for 12 h. Com-
pound 2a was very stable under these conditions, but the
monophenyl phosphonates 2b–d with an electron-with-
drawing group were hydrolyzed partially to APBA
according to the extent of leaving group ability of the
phenols: 11%, 40%, and 65% hydrolysis was observed
for 2b, 2c, and 2d, respectively (31P NMR). Since the
half-life (t1/2) of the monofluorophosphonate 1 is
21.6 h (pH 5.5, 37 ꢁC)40 and 20 min (pH 7.5, 37 ꢁC),44
the hydrolytic stability of the monophenyl phosphonates
2a–d was significantly improved by replacing the highly
acidic leaving group (HF, pKa = 3.17) with less acidic
substituted phenols.45 However, less reactive electrophil-
ic phosphonates may in turn lead to less potent inhibito-
ry activities toward the enzyme. Therefore, the activities
Figure 1. pH-rate profiles of the hydrolysis of human GGT (d) and
E. coli GGT (h). The initial rates of hydrolysis were measured by using
0.2 and 4.0 lM c-Glu-AMC as substrate for E. coli and human GGT,
respectively, at 25 ꢁC (see Section 3). The hydrolytic activities are
shown relative to the rate of hydrolysis at pH 5.5 for each enzyme. The
following buffer systems were used: pH 5.5, 100 mM succinate–NaOH;
pH 6.1, 100 mM PIPES; pH 7.0, 100 mM MOPS; pH 8.1, 100 mM
tricine–NaOH; pH 9.1, 100 mM NaHCO3–Na2CO3 buffer.

6046
L. Han et al. / Bioorg. Med. Chem. 14 (2006) 6043–6054
the acid-base equilibrium of the amino group of the sub-
strate.48,50 We therefore measured the inhibitory activi-
ties of 2a–d at pH 5.5 for E. coli GGT, and at pH 5.5
and 8.0 for human GGT, under the assay conditions
for hydrolase activity.
calculated according to the Eq. (2) or (4). For compari-
son, the inactivation rates of E. coli and human GGTs
by the monofluorophosphonate 1 and acivicin were also
determined. The inhibition of human GGT by 1 was
irreversible, because no regain of enzyme activity was
observed after prolonged dialysis of the inactivated
enzyme (data not shown). The inactivation of human
GGT was probably caused by the phosphonylation of
the catalytic nucleophile as confirmed with E. coli
GGT inactivated by 1.40
Compounds 2b–d inhibited both E. coli and human
GGTs in a time-dependent manner as depicted by a ser-
ies of typical progress curves that show the rate of en-
zyme
concentration of the inhibitor (Fig. 2).
inactivation
increased
with
increasing
The values of kon and the pKa of the leaving groups for 1,
2a–d, and acivicin are listed in Table 1. The monophenyl
phosphonates 2a–d were far more potent inhibitors of
E. coli GGT than toward the human enzyme: the inacti-
vation rates of E. coli GGT were at least two orders of
magnitude larger than those of human GGT at pH 5.5.
This also holds for the inhibition by acivicin, where
human GGT was inactivated much more slowly than
the E. coli enzyme. These results can be explained in part
by the fact that E. coli GGT has a rather broad substrate
specificity and is more specific for aromatic amino
acids49 as compared to the mammalian GGT.26 Thus,
the phosphonates 2a–d with an aromatic moiety might
be better accommodated by E. coli GGT than by the
human enzyme to facilitate the enzyme inactivation.
Compound 2a, the least active monophenyl phospho-
nate, slowly inactivated E. coli GGT, but not the human
enzyme either at pH 5.5 or 8.0. The time-dependent inhi-
bition curves for 2b–d were fit to the first-order rate Eq.
(1) to determine the pseudo-first-order rate constant for
enzyme inactivation (kobsd) at each inhibitor concentra-
tion. The inhibition of E. coli GGT by 2a was so slow
that the values of kobsd were determined by measuring
the remaining activity periodically by incubating the en-
zyme with varying concentrations of the inhibitor with-
out substrate (discontinuous assay, see Section 3). A
plot of kobsd against the inhibitor concentration was
found to be linear for each inhibitor, and the second-or-
der rate constant for enzyme inactivation (kon) was
Interestingly, the inhibition potency of the phosphonates
1 and 2a–d toward E. coli GGT was highly dependent on
the pKa of the leaving group, whereas the inactivation
rate of human GGT was almost independent on the leav-
ing group ability at either pH 5.5 or 8.0. As a result, the
most reactive monofluorophosphonate 1 was by far
the most potent inhibitor toward E. coli GGT, whereas
the less reactive monophenyl phosphonates 2b–d served
as fairly good inhibitors of human GGT as compared
to 1 and acivicin at pH 5.5. The difference became even
smaller at pH 8, where the inhibitory activity of 2c al-
most rivaled that of 1 and was more than three times
higher than that of acivicin towards the human enzyme.
Due to the large difference in the dependence of inhibito-
ry activity on the leaving group ability, the selectivity of
the phosphonates 2a–d for E. coli GGT was significantly
increased as the leaving group ability was increased: the
phosphonate 2d, for example, was ca. 2300 times more
active toward E. coli GGT than toward the human en-
zyme at pH 5.5. These results indicate that there are sig-
nificant differences between the E. coli and the human
enzymes in the structure of the active site region and
the catalytic mechanism.
It should be noted that the kon values for human GGT
by the phosphonates 1 and 2a–d increased by 6 to 19
times when the pH was raised from 5.5 to 8.0. The
increase in the kon values corresponds approximately
to the increase in the enzyme activity between pH 5.5
and 8.0 (Fig. 1), suggesting that the phosphonates 2b–
d served as mechanism-based inhibitors of GGT that
reacted with the catalytic nucleophile by the assistance
of the catalytic power of the enzyme. In contrast, the
kon value of acivicin at pH 8.0 was only three times larg-
er than that at pH 5.5, arguing against the correlation of
the inhibition by acivicin with the catalytic mechanism
Figure 2. Typical time-dependent inhibition of GGT by the monophe-
nyl phosphonate 2. (A) Progress curves of E. coli GGT with 0.2 lM
c-Glu-AMC at pH 5.5 in the presence of varying concentrations of 2d:
(a) 0, (b) 7.08, (c) 14.2, (d) 35.4, (e) 70.8 lM. (B) Progress curves of
human GGT with 4.0 lM c-Glu-AMC at pH 8.0 in the presence of
varying concentrations of 2c: (a) 0, (b) 0.41, (c) 0.81, (d) 1.22 mM.

L. Han et al. / Bioorg. Med. Chem. 14 (2006) 6043–6054
Table 1. Inhibitory activities of 1, 2a–d and acivicin toward E. coli and human GGT
6047
kon (Mꢀ1 sꢀ1
)
a
Inhibitor
Leaving group
pKa
E. coli GGT (pH 5.5)
Human GGT (pH 5.5)
Human GGT (pH 8.0)
2a
2b
C6H5OH
9.98
0.26
18
NIb
NIb
p-CF3AC6H4OH
p-AcAC6H4OH
p-CNAC6H4OH
HF
8.51
8.12
1
0.031 0.001
0.34 0.02
0.12 0.01
0.82 0.02
0.40 0.02
0.30 0.1
4.9 0.3
2.3 0.1
5.0 0.3
1.3 0.1
2c
2d
130 20
270 10
49100 800c
4210 10
7.95
3.17
1
Acivicin
HCl
ꢀ6.18
a Second-order rate constant for enzyme inactivation.
b No inhibition.
c Ref. 40.
of GGT (see later). In fact, acivicin was shown to bind
covalently to Ser406 of human GGT, but this residue
is not essential for catalysis.31 Acivicin is a potent inhib-
itor of E. coli GGT (kon = 4210 sꢀ1 Mꢀ1, Table 1), but
its inactivation rate was much smaller than expected
from its leaving group ability (Clꢀ), suggesting again
that the mechanism of enzyme inactivation with acivicin
was totally different from that with the phosphonates 1
and 2a–d.
For quantitative analysis of the substituent effect on the
inhibitory activity, Brønsted plots have been constructed
for the inactivation of E. coli and human GGTs by the
phosphonates 1 and 2a–d. A large negative slope (blg
ꢀ1.4) with excellent linearity was obtained for the inac-
tivation of E. coli GGT by the monophenyl phospho-
nates 2a–d (Fig. 3A). Such a large leaving group
dependence in phosphoryl transfer reactions51 suggests
that the reaction of the catalytic nucleophile of E. coli
GGT with the electrophilic phosphorus proceeded via
an anionic and dissociative transition state with substan-
tial bond cleavage between the phosphorus and the leav-
ing group as shown below. Interestingly, the plot of the
monofluorophosphonate 1 was located well below the
line, which is indicative of significantly lower activity
of 1 than expected from its excellent leaving group abil-
ity of Fꢀ. However, this can be attributed to the lower
affinity of 1 rather than its reactivity, because the bimo-
lecular reaction constant shown here (kon) contains the
kinetic parameter (Ki) with respect to the reversible
binding step of the inhibitor before the formation of a
covalent bond with the enzyme. Since E. coli GGT has
a preference for aromatic amino acids as a c-glutamyl
acceptor,49 the fluorine atom in 1 may be too small to
be accommodated effectively in the acceptor site of the
E. coli enzyme. If this is the case, then the excellent
linearity obtained with 2a–d cogently suggests that
E. coli GGT accommodates these monophenyl
phosphonates almost equally regardless of the substitu-
ent on the para position. This is also consistent with the
broad substrate specificity of E. coli GGT.
Figure 3. Brønsted plots relating the second-order rate constants for
enzyme inactivation (kon) by 1 and 2a–d with the leaving group ability.
(A) E. coli GGT at pH 5.5. (B) Human GGT at pH 5.5 (filled symbols)
and pH 8.0 (open symbols).
In contrast, the Brønsted plot for the inactivation of
human GGT gave a slope near zero with much poor
linearity (Fig. 3B). A small value of blg is consistent with
a transition state that is primarily associative in charac-
ter with little bond cleavage between the phosphorus
and the leaving group as shown in the following figures
(A), but is also consistent with general acid catalysis,
where an acidic group in the active site provides a pro-
ton that offsets the development of negative charge on
the leaving oxygen as the P–O bond breaks (B). Keillor
et al. reported that the reaction catalyzed by rat kidney
GGT proceeded via the general-acid-catalyzed acylation
with an ammonium ion acting as a general acid, derived
from the results of free energy relationships using a ser-

6048
L. Han et al. / Bioorg. Med. Chem. 14 (2006) 6043–6054
ies of para-substituted L-glutamyl c-anilide as sub-
strate.52 Although the mechanism for the enzyme inacti-
vation is different from that for the enzyme catalysis, the
evidence for the general-acid-catalyzed acylation of
GGT favors the participation of a general acid catalysis
also in the phosphonylation of human GGT by the
monophenyl phosphonates 2b–d.
also found to be accelerated by these compounds.29,55,56
From these observations, Meister and Tate et al. con-
cluded that maleate and the hippurate analogues were
bound at the Cys-Gly binding site and that these accep-
tor site-directed binders modulated the enzyme activity
by cooperative interactions between the c-glutamyl
donor and the acceptor binding domains of the GGT
active site through conformational change.56 It is inter-
esting to test these hypotheses by using the fluor-
ophosphonate 1 as a chemical probe, because the
phosphonate 1 served as a mechanism-based inhibitor
that reacted with the c-glutamyl donor site, but does
not have a large substituent that may bind to the accep-
tor site. We therefore measured the effect of 3-BPA on
the inactivation of E. coli and human GGTs by 1 and
acivicin (Fig. 4). Contrary to our initial expectations,
the effect of 3-BPA was inhibitory, rather than accelera-
tive, to the inactivation of both E. coli and human
GGTs by 1 (Fig. 4A and C): the inactivation rate was
decreased by 39% and 47% for E. coli and human
GGT, respectively, in the presence of 40 mM 3-BPA.
The same protective effect was observed in the inactiva-
tion of E. coli GGT by acivicin (Fig. 4B), where the
addition of 3-BPA decreased the inactivation rate by
70%. However, the inactivation of human GGT by aci-
vicin was significantly accelerated (ca. 30-fold) by the
addition of 3-BPA (Fig. 4D). Of particular interest is
that the acceleration of enzyme inactivation was
observed solely for human GGT by acivicin. These re-
sults indicate some important facets regarding the differ-
ences between mammalian and bacterial GGTs and the
properties of acivicin. First, the modulation of enzyme
activity by acceptor site-directed molecules reported
2.3. Effect of 3-benzoylpropionic acid (3-BPA) on the
inactivation of E. coli and human GGT by 1
The glutaminase activity of rat GGT was reported to be
stimulated by maleate,53,54 hippurate, and its analogues
including 3-BPA,55,56 but at the expense of the transpep-
tidase activity. The inactivation by acivicin or DON was
Figure 4. Effect of 3-benzoylpropionic acid (3-BPA, 40 mM) on the inactivation of E. coli GGT (A, B) and human GGT (C, D) by acivicin and
monofluorophosphonate 1 in 100 mM succinate buffer (pH 5.5). Inhibitors: (A) 8.37 nM of 1, (B) 0.17 lM acivicin, (C) 0.15 mM of 1, (D) 0.20 mM
acivicin.

L. Han et al. / Bioorg. Med. Chem. 14 (2006) 6043–6054
6049
with rat GGT55,56 does not always hold for all the
GGTs. At least, the E. coli enzyme does not appear to
have such properties. This might be related to the differ-
ences in the physiological functions of bacterial and
mammalian GGTs: the E. coli enzyme is a soluble pro-
tein located in the periplasmic space49 and is probably
responsible for utilization of exogenous c-glutamyl pep-
tides as an amino acid source.57 In contrast, the mam-
malian GGT is anchored to the extracellular surface of
the plasma membrane of cells with secretory or absorp-
tive functions and is highly specific for glutathione and
its analogues.1 The in vivo role of hippurate was also
implicated in the regulation of the activity of renal
GGT.55 Second, the protective effect of 3-BPA on the
inactivation of human GGT by 1 (Fig. 4C) is indicative
of binding of 3-BPA not only in the acceptor site, but
also in the c-glutamyl donor site. This notion is support-
ed by the protective effect of 3-BPA on the inactivation
of E. coli GGT either by compound 1 and acivicin.
Finally, the inhibition of GGT by acivicin is a general
observation, but its inactivation mechanism seems total-
ly different from that by the mechanism-based inhibitors
1 and 2a–d shown here. It is still unclear why 3-BPA sig-
nificantly accelerated the inactivation of human GGT by
acivicin, but it seems likely that the inhibition of GGT
by acivicin is a rather fortuitous event that is not related
to the enzyme catalysis, and its acceleration by 3-BPA
may also be unrelated to the acceleration of the catalytic
activity of the enzyme.
[sodium 3-(trimethylsilyl)propanesulfonate for 1H in
D2O and 85% H3PO4 for 31P]. Infrared spectra were
recorded on a Hitachi U-215 infrared spectrophotome-
ter. Mass spectra were obtained with a JEOL JMS 700
spectrometer. Elemental analyses were performed on a
Yanaco MT-5. Thin-layer chromatography was carried
out using silica gel plates (Merck 5715, 0.25 mm). Com-
pounds were purified by flash column chromatography
on silica gel 60 (Merck 9385, 40–63 lm) or by medi-
um-pressure reversed-phase column chromatography
using a Yamazen YFLC (Yamazen Co., Osaka, Japan).
3.2. Synthesis
3.2.1. 2-(N-4-Nitrobenzyloxycarbonylamino)-4-phospho-
nobutanoic acid (3). 4-Nitrobenzyl chloroformate
(12.1 g, 56.1 mmol) and NaHCO3 (4.71 g, 56.1 mmol)
were added portion wise over 2 h to a vigorously stir-
red solution of APBA (6.85 g, 37.4 mmol) and NaOH
(4.48 g, 112 mmol) in a mixture of H2O (110 mL)–
Et2O (40 mL) at 0 ꢁC. The mixture was stirred
vigorously at ambient temperature for 26 h. After
completion of the reaction (TLC, BuOH/AcOH/
H2O = 5:2:2, ninhydrin for consumption of APBA),
the reaction mixture was washed with Et2O (3·
50 mL), acidified with 1 N HCl to pH 1, and was
evaporated to dryness. Acetone (150 mL) was added
to the residue, and the insoluble salt (NaCl) was re-
moved by filtration. The residual oil was purified by
medium-pressure reversed-phase column chromatogra-
phy on a Daisogel SP-120-40/60-ODS-B (Daiso Co.
Ltd., Osaka, Japan). The column was eluted with a
linear gradient of MeOH/H2O (40–60%), and the
fractions containing the product (TLC, BuOH/
AcOH/H2O = 5:2:2, UV) were lyophilized to afford
3 as a highly hygroscopic solid (10.1 g, 75%): 1H
NMR (300 MHz, D2O) dH 1.6–1.8 (m, 2H,
CH2CH2P), 1.8–2.2 (2 · m, 2H, CH2CH2P), 4.2 (m,
1H, a-CH), 5.2 (s, 2H, benzyl), 7.6 (d, J = 8.4 Hz,
2H) and 8.2 (d, J = 8.3 Hz, 2H) (C6H4NO2); 31P
NMR (121 MHz, D2O) dp 27.71; Anal. Calcd for
C12H15N2O9P: C, 39.79; H, 4.17; N, 7.73. Found:
C, 39.62; H, 4.35; N, 7.63.
3. Experimental
3.1. Materials and methods
All chemicals were obtained commercially and used
without further purification unless otherwise noted.
7-(c-L-Glutamylamino)-4-methylcoumarin (c-Glu-AMC)
and (aS,5S)-a-amino-3-chloro-4,5-dihydro-5-isoxazole-
acetic acid (acivicin) were purchased from Sigma. Race-
mic 2-amino-4-phosphonobutanoic acid (APBA) was
synthesized by the reported procedure.58,59 Racemic
2-amino-4-(fluorophosphono)butanoic acid (1) was syn-
thesized previously.40,44 Dry toluene and dry CH2Cl2
were prepared by distillation from CaH2 and stored over
3.2.2. Methyl 2-(N-4-nitrobenzyloxycarbonylamino)-4-
phosphonobutanoate (4). Thionyl chloride (3.22 g,
27.1 mmol) was added dropwise to MeOH (100 mL) at
ꢀ5 ꢁC, and the mixture was stirred for 15 min. Com-
pound 3 (6.54 g, 18.1 mmol) was added to the solution,
and the mixture was stirred at ambient temperature for
76 h. Volatiles were removed in vacuo, and the residual
oil was purified by medium-pressure reversed-phase col-
umn chromatography (ODS, Daisogel) eluted with a lin-
ear gradient of MeOH/H2O (50–99%). The fractions
containing the product (TLC, BuOH/AcOH/
H2O = 5:2:2, UV) were collected and lyophilized to af-
ford 4 as a highly hygroscopic amorphous solid
(4.96 g, 73%, containing 22% of 3 derived from 4 by
hydrolysis during chromatography). Analytical sample
was obtained from the fractions containing pure 4:
1H-NMR (300 MHz, acetone-d6) dH 1.8 (m, 2H,
PCH2CH2), 1.9–2.2 (2 · m, 2H, PCH2CH2), 3.7 (s, 3H,
CH3O), 4.3 (m, 1H, a-CH), 5.3 (s, 2H, benzyl), 7.08
˚
molecular sieves 4 A. Dry tetrahydrofuran (THF) was
purchased from Kanto Kagaku (Tokyo, Japan). E. coli
c-glutamyltranspeptidase was purified from the periplas-
mic fraction of a recombinant strain of E. coli K-12
(SH642) by lysozyme treatment, ammonium sulfate pre-
cipitation, and chromatofocusing as described previous-
ly.49 Human c-glutamyltranspeptidase HC-GTP (T-72)
was a generous gift from Asahi Kasei Corporation (Osa-
ka, Japan). The latter enzyme preparation contained a
large amount of bovine serum albumin (BSA) as enzyme
stabilizer (GGT content <1%).
The protein concentration was determined by Brad-
ford’s method.60 1H, 13C, and 31P NMR spectra were
recorded on a JEOL JNM-AL 300 (300 MHz for H)
1
or a JEOL JNM-AL 400 (400 MHz). Chemical shifts
were recorded relative to the internal standard (tetra-
methylsilane for 1H and 13C) or to the external standard

6050
L. Han et al. / Bioorg. Med. Chem. 14 (2006) 6043–6054
(br d, J = 8.1 Hz, 1H, NH), 7.6 (d, J = 8.7 Hz, 2H) and
8.2 (d, J = 8.7 Hz, 2H) (C6H4NO2); 31P NMR
(121 MHz, acetone-d6) dp 31.44; HRMS (FAB, glycerol)
calcd for C13H18N2O9P (MH+) 377.0750, found
377.0751; Anal. Calcd for C13H16N2O9 PÆ0.68 H2O: C,
40.19; H, 4.76; N, 7.21. Found: C, 40.17; H, 4.77; N, 7.41.
3.2.5. Ethyl 2-(N-4-nitrobenzyloxycarbonylamino)-4-
[bis(4-trifluoromethylphenyl)phosphono]butanoate (6b).
The phosphonic acid 40 (3.11 g, 7.97 mmol) was dis-
solved in thionyl chloride (5 mL) and a catalytic amount
of DMF and was heated at 60 ꢁC for 1 h. The resulting
solution was evaporated to give the phosphonodichlori-
date 50 as an yellow oil: 1H-NMR (300 MHz, CDCl3) dH
1.3 (t, J = 7.1 Hz, 3H, OCH2CH3), 2.1–2.8 (m, 4H,
PCH2CH2), 4.3 (q, J = 7.1 Hz, 2H, OCH2CH3), 4.5
(m, 1H, a-CH), 5.20 and 5.25 (2 · d, J = 17.9 Hz, 2H,
benzyl), 5.7 (br d, J = 6.8 Hz, 1H, NH), 7.5 (d,
J = 8.6 Hz, 2H) and 8.2 (d, J = 8.6 Hz, 2H
(C6H4NO2); 31P NMR (121 MHz, CDCl3) dp 49.07.
An oil suspension of 60% NaH (0.73 g, 18.3 mmol)
was washed twice with hexane and was suspended in
dry THF (15 mL) under an argon atmosphere. A solu-
tion of 4-trifluoromethylphenol (3.23 g, 19.9 mmol) in
dry THF (15 mL) was added to this suspension at
0 ꢁC under an argon atmosphere to give an yellowish
emulsion. To this mixture, a solution of the phospho-
nodichloridate 50 (7.97 mmol) in dry THF (15 mL) was
added dropwise over 10 min at 0 ꢁC. After the addition
was complete, the mixture gave a clear yellowish solu-
tion. The mixture was stirred at 0 ꢁC for 1 h, and at
ambient temperature for further 1 h under an argon
atmosphere. After the reaction was completed (31P
NMR), the reaction mixture was concentrated and dilut-
ed with EtOAc (100 mL), washed successively with 1 N
HCl (100 mL), and sat. NaCl (100 mL). The organic
layer was dried over anhydrous Na2SO4 and was evapo-
3.2.3. Ethyl 2-(N-4-Nitrobenzyloxycarbonylamino)-4-
phosphonobutanoate (40). The ethyl ester 40 was prepared
from 3 and EtOH–HCl by the same procedure as
described for the synthesis of the methyl ester 4. The
product was purified by medium-pressure reversed-
phase column chromatography (ODS-S-50D, Yamazen
Co.) eluted with a linear gradient of MeOH/H2O
(50–65%) to give pure 40 as a colorless amorphous solid
(75%): 1H-NMR (300 MHz, acetone-d6) dH 1.2
(t, J = 7.1 Hz, 3H, OCH2CH3), 1.8 (m, 2H, PCH2CH2),
1.9–2.3 (2 · m, 2H, PCH2CH2), 4.2 (2 · q, J = 7.1 Hz,
2H, OCH2CH3), 4.3 (m, 1H, a-CH), 5.3 (s, 2H, benzyl),
7.04 (br d, J = 8.2 Hz, 1H, NH), 7.6 (d, J = 8.8 Hz, 2H)
and 8.2 (d, J = 8.8 Hz, 2H) (C6H4NO2), 9.14 [br s, 3H,
COOH and P(O)(OH)2]; 31P NMR (121 MHz, ace-
tone-d6) dp 32.67; Anal. Calcd for C14H19N2O9P: C,
43.09; H, 4.91; N, 7.18. Found: C, 42.99; H, 4.91; N,
7.17.
3.2.4. Methyl 2-(N-4-nitrobenzyloxycarbonylamino)-4-
(diphenylphosphono)butanoate (6a). A solution of 4
(0.87 g, 2.31 mmol) in dry CH2Cl2 was gently refluxed
in the presence of a catalytic amount of DMF under
an argon atmosphere. To this solution was added oxalyl
chloride (0.73 g, 5.75 mmol), and the mixture was
refluxed for 2 h. During the addition, copious gas was
evolved, and the heating was adjusted to keep gentle
refluxing. The reaction mixture was evaporated in vacuo
to give the corresponding phosphonodichloridate 5 as
1
rated to give 6b as an yellowish oil (5.87 g): H-NMR
(300 MHz, CDCl3) dH 1.2 (t, J = 6.6 Hz, 3H,
OCH2CH3), 2.0–2.5 (m, 4H, PCH2CH2), 4.2 (m,
J = 6.8 Hz, 2H, OCH2CH3), 4.4 (m, 1H, a-CH), 5.08
and 5.17 (2 · d, J = 13.0 Hz, 2H, benzyl), 5.9 (br d,
J = 6.7 Hz, 1H, NH), 7.2–7.5 (m, 10H, CF3C6H4 and
C6H4NO2), 8.1 (d, J = 7.7 Hz, 2H, C6H4NO2); 31P
NMR (121 MHz, CDCl3) dp 25.51. Compound 6b was
used for the next reaction without further purification.
1
an yellow oil: H-NMR (300 MHz, CDCl3) dH 2.2–2.8
(m, 4H, PCH2CH2), 3.8 (s, 3H, CH3O), 4.5 (m, 1H,
a-CH), 5.2 (m, 2H, benzyl), 5.8 (br, 1H, NH), 7.5 (d,
J = 8.7 Hz, 2H) and 8.2 (d, J = 8.7 Hz, 2H)
(C6H4NO2); 31P NMR (121 MHz, CDCl3) dp 53.15. A
mixture of phenol (0.46 g, 4.91 mmol) and NEt3
(0.70 g, 6.93 mmol) in dry CH2Cl2 (20 mL) was added
dropwise over 10 min to a solution of 5 (2.31 mmol) in
dry CH2Cl2 (10 mL) at 0 ꢁC. After the addition was
complete, the mixture was stirred at 0 ꢁC for 1.5 h.
The reaction mixture was concentrated, diluted with
EtOAc (50 mL) and was washed successively with 1 N
HCl (50 mL), H2O (50 mL), satd NaHCO3 (50 mL)
and sat. NaCl (50 mL). The organic layer was dried over
anhydrous Na2SO4. Solvent was removed, and the resid-
ual oil was purified by flash column chromatography on
silica gel 60 (EtOAc/hexane, 3:2) to give 6a as a colorless
3.2.6. Methyl 2-(N-4-nitrobenzyloxycarbonylamino)-4-
[bis(4-acetylphenyl)phosphono]butanoate (6c). The bis(4-
acetylphenyl) phosphonate 6c was synthesized from
freshly prepared 5 (2.82 mmol) and 4-hydroxyacetophe-
none (7.0 mmol) according to the same procedure as
described for the synthesis of 6b and was purified by
flash column chromatography on silica gel 60 (EtOAc/
hexane, 4:1) to give 6c as an yellow oil (1.47 g, 85%):
1H-NMR (300 MHz, CDCl3) dH 2.1–2.5 (m, 4H,
PCH2CH2), 2.6 (s, 6H, 2 · CH3CO), 3.7 (s, 3H,
CO2CH3), 4.5 (m, 1H, a-CH), 5.17 and 5.24 (2 · d,
J = 13.8 Hz, 2H, benzyl), 5.7 (d, J = 7.8 Hz, 1H, NH),
7.2 (d, J = 8.4 Hz, 4H) and 7.9 (d, J = 8.4 Hz, 4H)
(2 · C6H4COCH3), 7.5 (d, J = 8.4 Hz, 2H) and 8.2 (d,
J = 8.4 Hz, 2H) (C6H4NO2); 31P NMR (121 MHz,
CDCl3) dp 25.04; HRMS (FAB, glycerol) calcd for
C29H30N2O11P (MH+) 613.1588, found 613.1594; Anal.
Calcd. for C29H29N2O11P: C, 56.87; H, 4.77; N, 4.57.
Found: C, 56.60; H, 4.82; N, 4.38.
1
oil (0.79 g, 65%): H-NMR (300 MHz, CDCl3) dH 2.0–
2.5 (m, 4H, PCH2CH2), 3.77 (s, 3H, CH3O), 4.48 (m,
1H, a-CH), 5.18 and 5.22 (2 · d, J = 15 Hz, 2H, benzyl),
5.8 (br d, J = 7.8 Hz, 1H, NH), 7.1–7.4 (m, 10H,
2 · C6H5), 7.5 (d, J = 8.7 Hz, 2H) and 8.2 (d,
J = 8.7 Hz, 2H) (C6H4NO2); 31P NMR (121 MHz,
CDCl3) dp 28.69; IR (NaCl) mmax 1715, 1590, 1510,
1490, 1340, 1260–1160 (br), 1060–1000 (br), 925, and
760 cmꢀ1; Anal. Calcd for C25H25N2O9P: C, 56.80; H,
4.77; N, 5.30. Found: C, 56.74; H, 4.77; N, 5.23.
3.2.7. Ethyl 2-(N-4-nitrobenzyloxycarbonylamino)-4-
[bis(4-cyanophenyl)phosphono]butanoate (6d). The bis(4-
cyanophenyl) phosphonate 6d was synthesized from

L. Han et al. / Bioorg. Med. Chem. 14 (2006) 6043–6054
6051
freshly prepared 50 (8.45 mmol) and 4-cyanophenol
(21.2 mmol) according to the same procedure as
described for the synthesis of 6b. The crude product
(5.97 g) was used in the next step without purification:
1H-NMR (300 MHz, CDCl3) dH 1.3 (t, J = 7.2 Hz, 3H,
OCH2CH3), 2.0–2.5 (m, 4H, PCH2CH2), 4.2 (q,
J = 6.8 Hz, 2H, OCH2CH3), 4.5 (m, 1H, a-CH), 5.17
and 5.26 (2 · d, J = 13.4 Hz, 2H, benzyl), 5.6 (br d,
J = 7.3 Hz, 1H, NH), 7.3 (d, J = 9.1 Hz, 4H) and 7.7
(d, J = 8.4 Hz, 4H) (2 · C6H4CN), 7.5 (d, J = 8.4 Hz,
2H) and 8.2 (d, J = 8.7 Hz, 2H) (2 · C6H4NO2); 31P
NMR (121 MHz, CDCl3) dp 25.78.
by medium-pressure reversed-phase column chromatog-
raphy (ODS-S-50B, Yamazen Co.). The column was
eluted with a linear gradient of MeOH/H2O (50–90%),
and the fractions containing the product (TLC,
BuOH/AcOH/H2O, 5:2:2, UV) were collected and evap-
orated to afford the N-protected mono(4-trifluorometh-
ylphenyl) phosphonate as an yellowish solid (2.07 g,
1
51%): H-NMR (300 MHz, acetone-d6) dH 2.0–2.4 (m,
4H, PCH2CH2), 4.4 (m, 1H, a-CH), 5.2 (2 · d,
J = 13.9 Hz, 2H, benzyl), 7.4 (d, J = 8.3 Hz, 2H,
C6H4CF3), 7.6 (m, 4H, C6H4CF3 and C6H4NO2), 8.2
(d, J = 8.7 Hz, 2H, C6H4NO2); 31P NMR (121 MHz,
acetone-d6) dp 25.76; HRMS (FAB, glycerol) calcd for
C19H19F3N2O9P (MH+) 507.0781, found 507.0767. The
mono(4-trifluoromethylphenyl) phosphonate (0.79 g,
1.56 mmol) was dissolved in THF (45 mL), and hydro-
gen gas was passed through the solution in the presence
of 5% Pd–C (50 mg) at ambient temperature for 2 h. The
reaction was monitored with TLC (BuOH/AcOH/H2O,
5:2:2, ninhydrin). The precipitated 2b and the Pd–C
were collected by filtration, washed with THF, and
extracted with hot H2O (30 mL). The extract was evap-
orated to give an yellow oil (200 mg). This residue was
dissolved in trifluoroacetic acid (50 lL) and was diluted
with H2O (1.9 mL). This solution was kept at ambient
temperature for 3 days to give crystals. The crystals were
collected by filtration and was washed with EtOH and
Et2O successively to afford 2b as light yellow crystals
(68 mg, 16%): 1H-NMR (300 MHz, D2O) dH 1.7–2.0
(m, 2H, PCH2CH2), 2.1–2.3 (m, 2H, PCH2), 4.0 (m,
1H, a-CH), 7.3 (d, J = 8.3 Hz, 2H) and 7.7 (d,
J = 8.3 Hz, 2H) (C6H4CF3); 31P NMR (121 MHz,
D2O) dp 23.66; HRMS (FAB, glycerol) calcd for
C11H14F3NO5P (MH+) 328.0562, found 328.0562; Anal.
Calcd. for C11H13F3NO5P: C, 40.38; H, 4.00; N, 4.28.
Found: C, 40.34; H, 4.06; N, 4.30.
3.2.8. 2-Amino-4-(monophenylphosphono)butanoic acid
(2a). An aqueous solution (50 mL) of NaOH (0.40 g,
10.0 mmol) was added to a solution of 6a (1.01 g,
1.91 mmol) in THF (20 mL). The mixture was stirred
at ambient temperature for 25 h. After completion of
the reaction (TLC, hexane/EtOAc, 1:4, UV), the reac-
tion mixture was acidified (pH 2) with 2 N HCl and
was evaporated to dryness. Acetone (100 mL) was add-
ed to the residue, and the resulting precipitate (NaCl)
was removed by filtration through Celite. The filtrate
was evaporated to dryness to give the N-protected mon-
ophenyl phosphonate as a deep red oil: 1H-NMR
(300 MHz, acetone-d6) dH 1.9–2.2 (m, 2H, PCH2CH2),
2.2–2.4 (m, 2H, PCH2), 4.4 (m, 1H, a-CH), 5.3 (s, 2H,
benzyl), 7.2–7.4 (m, 5H, C6H5), 7.7 (d, J = 8.1 Hz, 2H)
and 8.2 (d, J = 8.7 Hz, 2H) (C6H4NO2). The crude mon-
ophenyl phosphonate (1.91 mmol) was dissolved in
MeOH/AcOH/H2O (8:2:1, 44 mL). Hydrogen gas was
passed through the solution in the presence of 10%
Pd–C (100 mg) for 1.5 h at ambient temperature. The
Pd–C was removed by filtration through Celite, and
the filtrate was evaporated. The residual oil was purified
by medium-pressure reversed-phase column chromato-
graphy on an adsorption resin DIAION HP20SS (Mits-
ubishi Chemical Corp., Tokyo, Japan). The column was
eluted with H2O, and the fractions containing the prod-
uct (TLC, BuOH/AcOH/H2O, 5:2:2, ninhydrin) were
collected and lyophilized to afford 2a as a colorless solid
(0.29 g, 59%): mp 173–175 ꢁC; IR (KBr) mmax 3600–2300
3.2.10. 2-Amino-4-[mono(4-acetylphenyl)phosphono]buta-
noic acid (2c). Compound 6c (1.47 g, 2.40 mmol) was
hydrolyzed by NaOH (0.65 g, 15.6 mmol) in a mixture
of H2O (75 mL) and THF (45 mL) at ambient tempera-
ture for 15 h. After the reaction was complete (TLC,
hexane/EtOAc, 1:4, UV), the reaction mixture was
worked up by the same procedure as described for the
hydrolysis of 6a to give the crude N-protected
mono(4-acetylphenyl) phosphonate as an yellow oil
(1.52 g): 1H-NMR (300 MHz, acetone-d6) dH 2.0–2.4
(m, 4H, PCH2CH2), 2.6 (s, 3H, CH3CO), 4.4 (m, 1H,
a-CH), 5.3 (s, 2H, benzyl), 7.3 (d, J = 8.9 Hz, 2H,
C6H4Ac) and 7.7 (d, J = 8.9 Hz, 2H, C6H4NO2), 8.0
(d, J = 8.5 Hz, 2H, C6H4Ac), 8.2 (d, J = 8.7 Hz, 2H,
C6H4NO2). The crude mono(4-acetylphenyl) phospho-
nate (2.41 mmol) was dissolved in MeOH/H2O (4:1,
25 mL), and hydrogen gas was passed through the solu-
tion in the presence of 5% Pd–C (100 mg) at ambient
temperature for 5.5 h. The reaction was monitored with
TLC (BuOH/AcOH/H2O, 5:2:2, ninhydrin). The precip-
itated 2c and the Pd–C were collected by filtration,
washed with THF, and extracted with hot 10% TFA
in water (5 mL). The extract was evaporated to give
crude 2c as a deep red oil (0.67 g). The oil was washed
with EtOH and dissolved in 33% HBr/AcOH (4 mL).
The volatile was removed in vacuo to give the HBr salt
(br), 1700, 1590, 1485, 1160, 1040, 880, and 750 cmꢀ1
;
1H-NMR (300 MHz, D2O) dH 1.7–1.9 (m, 2H,
PCH2CH2), 2.1-2.3 (m, 2H, PCH2CH2), 4.1 (t,
J = 6.0 Hz, 1H, a-CH), 7.1–7.2 and 7.3–7.4 (m, 5H,
C6H5); 31P NMR (121 MHz, D2O) dp 22.66; 13C NMR
1
(75.4 MHz, D2O) dC 25.1 (d, JC-P = 137.4 Hz, PCH2),
3
27.0 (PCH2CH2), 55.9 (d, JC-P = 17.9 Hz, a-CH),
123.5, 127.1, 132.5 and 153.8 (d, JC-P = 7.4 Hz)
2
(C6H5), 174.3 (C@O); HRMS (FAB, glycerol) calcd
for C10H15NO5P (MH+) 260.0689, found 260.0684;
Anal. Calcd. for C10H14NO5PÆ0.6 H2O: C, 44.48; H,
5.67; N, 5.19. Found: C, 44.45; H, 5.51; N, 5.10.
3.2.9.
2-Amino-4-[mono(4-trifluoromethylphenyl)phos-
phono]butanoic acid (2b). The crude 6b (7.97 mmol)
was hydrolyzed by NaOH (1.93 g, 47.6 mmol) in a mix-
ture of H2O (30 mL) and THF (45 mL) at ambient tem-
perature for 36 h. After the reaction was complete (TLC,
BuOH/AcOH/H2O, 5:2:2, UV), the reaction mixture
was worked up by the same procedure as described for
the hydrolysis of 6a. The crude product was purified

6052
L. Han et al. / Bioorg. Med. Chem. 14 (2006) 6043–6054
as a deep red oil. The oil was dissolved in EtOH (5 mL)
and filtered through Celite. To this filtrate was added
propylene oxide (2.5 mL), and the mixture was kept at
4 ꢁC overnight. The precipitate was filtered and washed
with EtOH and Et2O, successively, to afford 2c as light
yellow powder (0.58 g, 80%): 1H-NMR (300 MHz,
D2O) dH 1.7–2.0 (m, 2H, PCH2CH2), 2.1-2.3 (m, 2H,
PCH2), 2.6 (s, 3H, CH3CO), 4.0 (t, J = 5.8 Hz, 1H,
a-CH), 7.2 (d, J = 8.1 Hz, 2H) and 8.0 (d, J = 8.4 Hz,
2H) (C6H4Ac); 31P NMR (121 MHz, D2O) dp 23.44;
HRMS (FAB, glycerol) calcd for C12H17NO6P
(MH+) 302.0794, found 302.0800; Anal. Calcd for
C12H16NO6PÆ0.12 H2O: C, 47.50; H, 5.39; N, 4.62.
Found: C, 47.51; H, 5.39; N, 4.52.
3.3. Hydrolytic stability of 2a–d
The hydrolysis of 2d was monitored by 31P NMR
(121 MHz) by incubating 29.3 mM of 2d in D2O at
23ꢁ. The ratio of 2d and the hydrolyzed product, APBA,
was calculated from the integration of the 31P NMR
peak areas (dP 23.67 for 2d, 24.32 for APBA).
3.4. Enzyme assay
The hydrolytic activity of E. coli GGT was measured
fluorimetrically by using 0.2 lM c-Glu-AMC as sub-
strate61 at pH 5.5 at 25 ꢁC. An aliquot (10 lL) of an
enzyme stock solution was added to 100 mM succi-
nate–NaOH buffer (pH 5.5) in a total volume of
1 mL containing 100 lL c-Glu-AMC stock solution
(2 lM in water) at 25 ꢁC. The release of 7-amino-4-
methylcoumarin (AMC) was monitored continuously
with a Hitachi F-2000 spectrophotometer (350 nm
excitation, 440 nm emission). AMC concentrations
were calculated from a standard calibration curve
of fluorescence intensity (F) versus AMC concentra-
tion (C): DF/DC = 0.11 nMꢀ1. The fluorescence inten-
sity was proportional to the concentration of AMC
up to 2.0 lM. The Michaelis constant (Km) for
c-Glu-AMC was determined as 0.2 lM under these
conditions. Protein was determined by the Bio-Rad
protein assay dye, using bovine serum albumin as
standard.
3.2.11. 2-Amino-4-[mono(4-cyanophenyl)phosphono]buta-
noic acid (2d). Crude 6d (8.45 mmol) was hydrolyzed by
NaOH (0.81 g, 20.3 mmol) in a mixture of H2O (50 mL)
and THF (40 mL) at ambient temperature for 24 h.
After the reaction was complete (TLC, BuOH/AcOH/
H2O, 5:2:2, UV), the reaction mixture was worked up
by the same procedure as described for the hydrolysis
of 6a. The crude product was purified by medium-pres-
sure reversed-phase column chromatography (ODS-
S-50B, Yamazen Co.). The column was eluted with a
linear gradient of CH3CN/H2O (20–50%), and the frac-
tions containing the product (TLC, BuOH/AcOH/H2O,
5:2:2, UV) were collected and evaporated to afford the
N-protected mono(4-cyanophenyl) phosphonate as an
yellow solid (1.08 g, 56%): 1H-NMR (300 MHz, ace-
tone-d6) dH 2.1–2.4 (m, 4H, PCH2CH2), 4.4 (m, 1H,
a-CH), 5.3 (s, 2H, benzyl), 7.0 (br s, 1H, NH), 7.4 (d,
J = 8.6 Hz, 2H, C6H4CN), 7.6 (d, J = 8.9 Hz, 2H,
C6H4NO2), 7.8 (d, J = 8.8 Hz, 2H, C6H4CN), 8.2 (d,
J = 8.7 Hz, 2H, C6H4NO2); 31P NMR (121 MHz, ace-
tone-d6) dp 25.76; HRMS (FAB, glycerol) calcd for
C19H19N3O9P (MH+) 464.0860, found 464.0856; Anal.
Calcd for C19H18N3O9P0.3 H2O: C, 48.69; H, 4.00; N,
8.96. Found: C, 48.67; H, 3.99; N, 8.99. Hydrogen gas
was passed through a mixture of the mono(4-cyanophe-
nyl) phosphonate (0.73 g, 1.58 mmol) and 5% Pd–C
(200 mg) in THF (30 mL) at ambient temperature for
4 h. The reaction was monitored with TLC (BuOH/
AcOH/H2O, 5:2:2, ninhydrin). The precipitated2d and
the Pd–C were collected by filtration, washed with
THF, and was extracted with TFA (2 mL). The extract
was evaporated to give crude 2d as a deep red oil
(442 mg). The oil was dissolved in 33% HBr/AcOH
(10 mL) and was evaporated to give the hydrogen bro-
mide salt as an oil. The oil was dissolved in EtOH
(30 mL) and filtered through Celite. To this filtrate
was added propylene oxide (10 mL), and diethyl ether
(10 mL) was added to promote the precipitation of the
product. The mixture was kept at 4 ꢁC for 12 h to give
2d as yellow powder (170 mg, 38%): 1H-NMR
(300 MHz, D2O) dH 1.7–2.0 (m, 2H, PCH2CH2), 2.1–
2.3 (m, 2H, PCH2), 4.0 (t, J = 6.0 Hz, 1H, a-CH), 7.3
(d, J = 8.6 Hz, 2H) and 7.8 (d, J = 8.7 Hz, 2H)
(C6H4CN); 31P NMR (121 MHz, D2O) dp 23.67; HRMS
(FAB, glycerol) calcd for C11H14N2O5P (MH+)
285.0641, found 285.0647; Anal. Calcd for
C11H13N2O5P: C, 46.49; H, 4.61; N, 9.86. Found: C,
46.28; H, 5.25; N, 9.43.
The hydrolytic activity of human GGT was measured
under the same conditions, except that the final sub-
strate concentration was 4.0 lM in 100 mM succinate–
NaOH buffer (pH 5.5) or 100 mM tricine–HCl (pH
8.0). The Km values for c-Glu-AMC were determined
as 12.6 lM (pH 5.5) and 4.0 lM (pH 8.0), respectively,
under these conditions.
3.5. Time-dependent inhibition assay
The inhibition of GGT was measured under pseudo-
first order rate conditions for both continuous and
discontinuous inhibition assay methods. In the contin-
uous inhibition assay (for 1, 2b–d with E. coli and
human GGT), the enzyme was added to a preincubat-
ed mixture of varying concentrations of the inhibitor
and the substrate (0.2 or 4.0 lM c-Glu-AMC for
E. coli and human GGT, respectively) in 100 mM
sodium succinate buffer (pH 5.5) or 100 mM tricine–
HCl (pH 8.0) at 25 ꢁC. Time-dependent inhibition of
the enzyme was measured by monitoring the product
release (AMC) continuously for 10 min. Pseudo-first-
order rate constants for enzyme inactivation (kobsd
were determined by fitting each curve to a first-order
rate equation:62
)
½Pꢁ ¼ ½Pꢁ1½1 ꢀ expðꢀkobsdtÞꢁ
where [P] and [P] are the concentrations of product
ð1Þ
1
formed at time t and at the time approaching infinity,
respectively. Since a plot of kobsd versus inhibitor con-
centration ([I]) exhibited no saturation until the highest
possible inhibitor concentration, the second-order rate

L. Han et al. / Bioorg. Med. Chem. 14 (2006) 6043–6054
6053
constant for enzyme inactivation (kon) was determined
according to the Eq. (2) derived from the kinetic mech-
anism described below:
Acknowledgments
We thank Asahi Kasei Corporation for a generous gift
of human c-glutamyl transpeptidase HC-GTP (T-72).
This study was supported in part by a Grant-in-Aid
for Scientific Research from Japan for the Promotion
of Science for J.H. (contract No. 16310152).
Km
E þ S ꢀ ES ! E þ P
# kon½Iꢁ
E-I
References and notes
kobsd ¼ kon½Iꢁ=ð1 þ ½Sꢁ=KmÞ
ð2Þ
1. Taniguchi, N.; Ikeda, Y. Adv. Enzymol. Rel. Areas Mol.
Biol. 1998, 72, 239.
2. Hiratake, J.; Suzuki, H.; Kumagai, H. In Handbook of
Proteolytic Enzymes; Barrett, A. J., Rawlings, N. D.,
Woessner, J. F., Eds., 2nd ed.; Elsevier: Amsterdam, 2004;
pp 2090–2094.
where S is c-Glu-AMC; [S] and Km are 0.2 and 0.2 lM
(E. coli GGT), and 4.0 and 12.6 lM (human GGT).
Nonlinear regression analyses of kinetic data were per-
formed using the KaleidaGraph program (Synergy
Software).
3. Zhang, H.; Forman, H. J.; Choi, J. Methods Enzymol.
2005, 401, 468.
4. Whitfield, J. B. Crit. Rev. Clin. Lab. Sci. 2001, 38, 263.
5. Hanigan, M. H.; Frierson, H. F. J., Jr. Histochem.
Cytochem. 1996, 44, 1101.
6. Hanigan, M. H.; Ricketts, W. A. Biochemistry 1993, 32,
6302.
7. Vincent, B. R.; Mousset, S.; Jacquemin-Sablon, A. Eur. J.
Biochem. 1999, 262, 873.
8. Ruoso, P.; Hedley, D. W. Cancer Chemother. Pharmacol.
2004, 54, 49.
9. Anderson, M. E.; Meister, A. Proc. Natl. Acad. Sci.
U.S.A. 1983, 80, 707.
10. Godwin, A. K.; Meister, A.; O’Dwyer, P. J.; Huang, C.-S.;
Hamilton, T. C.; Anderson, M. E. Proc. Natl. Acad. Sci.
U.S.A. 1992, 89, 3070.
11. Hanigan, M. R.; Gallagher, B. C.; Townsend, D. M.;
Gabarra, V. Carcinogenesis 1999, 20, 553.
12. Hanigan, M. H. Carcinogenesis 1995, 16, 181.
13. Lewis, A. D.; Hayes, J. D.; Wolf, C. R. Carcinogenesis
1988, 9, 1283.
In the discontinuous inhibition assay (for 2a with
E. coli and human GGT), the enzyme (final concd
of 0.5 lg/mL of crude human GGT) was incubated
at 25 ꢁC in 100 mM succinate–NaOH (pH 5.5) or
100 mM tricine–HCl (pH 8.0) containing varying con-
centrations of inhibitor. An aliquot (10 lL) of the
solution was withdrawn at various times and was
assayed for remaining activity. The values for kobsd
were obtained by fitting the remaining activities to
the following equation:
at=a0 ¼ expðꢀkobsdtÞ
ð3Þ
where at is the remaining activity at time t, and a0 is
the initial activity (t = 0) before inactivation. Nonlin-
ear regression analyses of kinetic data were also per-
formed using the KaleidaGraph program. The values
for kon were calculated according to the following
equation:
kobsd ¼ kon½Iꢁ
ð4Þ
14. Hanigan, M. H.; Pitot, H. C. Carcinogenesis 1985, 6, 165.
15. Benlloch, M.; Ortega, A.; Ferrer, P.; Segarra, R.; Obrador,
E.; Asensi, M.; Carretero, J.; Estrela, J. M. J. Biol. Chem.
2005, 280, 6950.
16. Ortega, A. L.; Carretero, J.; Obrador, E.; Gambini, J.;
Asensi, M.; Rodilla, V.; Estrela, J. M. J. Biol. Chem. 2003,
278, 13888.
17. Obrador, E.; Carretero, J.; Ortega, A.; Medina, I.;
Rodilla, V.; Pellicer, J. A.; Estrela, J. M. Hepatology
2002, 35, 74.
18. Pompella, A.; De Tata, V.; Paolicchi, A.; Zunino, F.
Biochem. Pharmacol. 2006, 71, 231, and references cited
therein.
3.6. Effect of 3-benzoylpropionic acid (3-BPA) on the
inhibition of GGT
In a discontinuous assay, the enzyme was incubated at
25 ꢁC in a solution of 100 mM succinate–NaOH (pH
5.5) or 100 mM tricine–HCl (pH 8.0) containing the
inhibitor and 3-BPA (40 mM). An aliquot (20 lL) was
withdrawn at various times and was assayed for remain-
ing activity.
19. Owen, A. D.; Schapira, A. H.; Jenner, P.; Marsden, C. D.
Ann. N.Y. Acad. Sci. 1996, 786, 217.
3.7. Regain of activity of inactivated human GGT
20. Sian, J.; Dexter, D. T.; Lees, A. J.; Daniel, S.; Jenner, P.;
Marsden, S. D. Ann. Neurol. 1994, 36, 356.
21. Lee, D.-H.; Ha, M.-H.; Kim, J.-H.; Christiani, D. C.;
Gross, M. D.; Steffes, M.; Blomhoff, R.; Jacobs, D. R.
Diabetologia 2003, 46, 359.
22. Meisinger, C.; Lo¨wel, H.; Heier, M.; Schneider, A.;
Thorand, B. J. Internal Med. 2005, 258, 527.
23. Paolicchi, A.; Minotti, G.; Tonarelli, P.; De Cesare, D.;
Mezzetti, A.; Dominici, S.; Comporti, M.; Pompella, A. J.
Invest. Med. 1999, 47, 151.
The enzyme was incubated with 1 mM of 1 in
100 mM succinate–NaOH (pH 5.5) at 25 ꢁC for
20 min to inactivate the enzyme completely. The
excess inhibitor was removed by rapid gel filtration
at 4 ꢁC through a Bio-Spin column (Bio-Rad) equili-
brated with 100 mM 3,3-dimethylglutaric acid–NaOH
buffer (pH 6.0). The inactivated enzyme was dialyzed
against 100 mM 3,3-dimethylglutaric acid–NaOH
buffer (pH 6.0) at 4 ꢁC for 1 week. The regain of
hydrolytic activity was measured using c-Glu-AMC
as substrate. The control experiment was conducted
under the same conditions without using the
inhibitor.
24. Ruttmann, E.; Brant, L. J.; Concin, H.; Diem, G.; Rap,
K.; Ulmer, H. Circulation 2005, 112, 2130.
25. Townsend, D. M.; Tew, K. D.; Tapiero, H. Biomed.
Pharmacother. 2003, 57, 145.
26. Allison, R. D. Methods Enzymol. 1985, 113, 419.

6054
L. Han et al. / Bioorg. Med. Chem. 14 (2006) 6043–6054
´
27. Keillor, J. W.; Menard, A.; Castonguay, R.; Lherbet, C.;
Rivard, C. J. Phys. Org. Chem. 2004, 17, 529.
28. Hanka, L. J.; Diez, A. Antimicrob. Agents Chemother.
1973, 3, 425.
44. Inoue, M.; Hiratake, J.; Sakata, K. Biosci. Biotechnol.
Biochem. 1999, 63, 2248.
45. Bartlett, P. A.; Lamden, L. A. Bioorg. Chem. 1986, 14,
356.
29. Gardell, S. J.; Tate, S. S. FEBS Lett. 1980, 122, 171.
30. Stole, E.; Seddon, A. P.; Wellner, D.; Meister, A. Proc.
Natl. Acad. Sci. U.S.A. 1990, 87, 1706.
31. Smith, T. K.; Ikeda, Y.; Fujii, J.; Taniguchi, N.; Meister,
A. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 2360.
32. Jayaram, H. N.; Cooney, D. A.; RRyan, J. A.; Neil, G. L.;
Dion, R. L.; Bono, V. H. Cancer Chemother. Rep. 1975,
59, 481.
46. Thompson, G. A.; Meister, A. J. Biol. Chem. 1977, 252,
6792.
47. Tate, S. S.; Meister, A. Methods Enzymol. 1985, 113, 400.
48. Cook, N. D.; Peters, T. J. Biochim. Biophys. Acta 1985,
832, 142.
49. Suzuki, H.; Kumagai, H.; Tochikura, T. J. Bacteriol. 1986,
168, 1325.
50. McIntyre, T. M.; Curthoys, N. P. J. Biol. Chem. 1979, 254,
6499.
51. Hengge, A. C.; Onyido, I. Curr. Org. Chem. 2005, 9, 61.
33. Zalkin, H. Adv. Enzymol. Rel. Areas Mol. Biol. 1998, 72, 87.
34. Earhart, R. H.; Neil, G. L. Adv. Enzyme Regul. 1985, 24,
179.
´
52. Menard, A.; Castonguay, R.; Lherbet, C.; Rivard, C.;
35. Chittur, S. V.; Klem, T. J.; Shafer, C. M.; Jo Davisson, V.
Biochemistry 2001, 40, 876.
36. Tate, S. S.; Meister, A. Proc. Natl. Acad. Sci. U.S.A. 1977,
74, 931.
37. Tate, S. S.; Meister, A. Proc. Natl. Acad. Sci. U.S.A. 1978,
75, 4806.
38. Stein, R. L.; DeCicco, C.; Nelson, D.; Thomas, B.
Biochemistry 2001, 40, 5804.
Roupioz, Y.; Keillor, J. W. Biochemistry 2001, 40, 12678.
53. Tate, S. S.; Meister, A. Proc. Natl. Acad. Sci. U.S.A. 1974,
71, 3329.
54. Thompson, G. A.; Meister, A. J. Biol. Chem. 1979, 254,
2956.
55. Thompson, G. A.; Meister, A. J. Biol. Chem. 1980, 255,
2109.
56. Gardell, S. J.; Tate, S. S. J. Biol. Chem. 1983, 258, 6198.
57. Suzuki, H.; Hashimoto, W.; Kumagai, H. J. Bacteriol.
1993, 175, 6038.
39. London, R. E.; Gabel, S. A. Arch. Biochem. Biophys. 2001,
385, 250.
40. Inoue, M.; Hiratake, J.; Suzuki, H.; Kumagai, H.; Sakata,
K. Biochemistry 2000, 39, 7764.
41. Brannigan, J. A.; Dodson, G.; Duggleby, H. J.; Moody, P.
C. E.; Smith, J. L.; Tomchick, D. R.; Murzin, A. G.
Nature 1995, 378, 416.
58. Kosolapoff, G. M. J. Am. Chem. Soc. 1948, 70, 1971.
59. Chambers, J. R.; Isbell, A. F. J. Org. Chem. 1964, 29, 832.
60. Bradford, M. M. Anal. Biochem. 1976, 72, 248.
61. Smith, G. D.; Ding, J. L.; Peters, T. J. Anal. Biochem.
1979, 100, 136–139.
42. Suzuki, H.; Kumagai, H. Biochemistry 2002, 277, 43536.
43. Hirschmann, R.; Yager, K. M.; Taylor, C. M.; Withe-
rington, J.; Sprengeler, P. A.; Phillips, B. W.; Moore, W.;
Smith, A. B., III J. Am. Chem. Soc. 1997, 119, 8177.
62. Stein, R. L.; Strimpler, A. M.; Edwards, P. D.; Lewis, J.
J.; Mauger, R. C.; Schwartz, J. A.; Stein, M. M.;
Trainor, D. A.; Wildonger, R. A.; Zottola, M. A.
Biochemistry 1987, 26, 2682.