Role of hydrophobic interactions in binding S-(N-aryl/alkyl-N-hydroxycarbamoyl)glutathiones to the active site of the antitumor target enzyme glyoxalase I

Article abstract of DOI:10.1021/jm000160l

Hydrophobic interactions play an important role in binding S-(N-aryl/alkyl-N-hydroxycarbamoyl)glutathiones to the active sites of human, yeast, and Pseudomonas putida glyoxalase I, as the log K(i) values for these mechanism-based competitive inhibitors de

Full text of DOI:10.1021/jm000160l

J . Med. Chem. 2000, 43, 3981-3986
3981
Role of Hyd r op h obic In ter a ction s in Bin d in g S-(N-Ar yl/Alk yl-N-h yd r oxy-
ca r ba m oyl)glu ta th ion es to th e Active Site of th e An titu m or Ta r get En zym e
Glyoxa la se I
Avinash Kalsi, Malcolm J . Kavarana, Tianfen Lu, Dale L. Whalen, Diana S. Hamilton,* and
Donald J . Creighton*
Department of Chemistry and Biochemistry, University of Maryland, Baltimore County, 1000 Hilltop Circle,
Baltimore, Maryland 21250
Received April 6, 2000
Hydrophobic interactions play an important role in binding S-(N-aryl/alkyl-N-hydroxycarbam-
oyl)glutathiones to the active sites of human, yeast, and Pseudomonas putida glyoxalase I, as
the log Ki values for these mechanism-based competitive inhibitors decrease linearly with
increasing values of the hydrophobicity constants (π) of the N-aryl/alkyl substituents.
Hydrophobic interactions also help to optimize polar interactions between the enzyme and the
glutathione derivatives, given that the Ki value for S-(N-hydroxycarbamoyl)glutathione (π )
0) with the human enzyme is 35-fold larger than the interpolated value for this compound
obtained from the log Ki versus π plot. Computational studies, in combination with published
X-ray crystallographic measurements, indicate that human glyoxalase I binds the syn-conformer
of S-(N-aryl-N-hydroxycarbamoyl)glutathiones in which the N-aryl substituents are in their
lowest-energy conformations. These studies provide both an experimental and a conceptual
framework for developing better inhibitors of this antitumor target enzyme.
Ch a r t 1. Structures of
In tr od u ction
S-(N-Aryl-N-hydroxycarbamoyl)glutathiones 1a -d and
S-(N-Alkyl-N-hydroxycarbamoyl)glutathiones 2a -h
The glyoxalase enzyme system has been the focus of
much recent attention as a potential target for antitu-
mor drug development.1,2 This enzyme system, com-
posed of the isomerase glyoxalase I (GlxI) and the
thioester hydrolase glyoxalase II (GlxII), promotes the
glutathione (GSH)-dependent conversion of cytotoxic
methylglyoxal to D-lactate.3,4 The toxicity of methyl-
glyoxal, a byproduct of normal cellular metabolism,5
appears to arise from its ability to cross-link proteins
and to form adducts with DNA.6,7 We recently demon-
strated that selected S-(N-aryl-N-hydroxycarbamoyl)-
glutathiones 1a -c (Chart 1, Table 1) are powerful
competitive inhibitors of human GlxI8,9 and that the
[glycyl,glutamyl]diethyl ester prodrugs inhibit the growth
of both murine and human tumors in vitro10 and in
vivo.11 Tumor toxicity appears to be due to inhibition of
intracellular GlxI, which results in elevated levels of
methylglyoxal.10 Thus, GlxI appears to be a potentially
important antitumor target.
Understanding the nature of the binding interaction
between the N-aryl-N-hydroxycarbamoyl esters of GSH
and the active site of GlxI is clearly important, to
provide a better basis for inhibitor design. Binding
appears to result from two conceptually separate and
distinct phenomena. First, the N-hydroxycarbamoyl
ester function contributes to binding by modeling the
stereoelectronic features of the tightly bound enediol
intermediate and/or flanking transition states that form
from the GSH-methylglyoxal-thiohemiacetal sub-
strate, eq 1.9
bic pocket in the active site, as indicated by the increase
in binding affinity with increasing hydrophobicity of the
N-aryl substituent. The recently determined high-
resolution X-ray crystal structure of the homodimeric
human enzyme in complex with enediol analogue 1d
confirms the presence of a hydrophobic binding pocket
in the active site.12 In the structure, a catalytically
essential active site Zn2+ directly coordinates both
oxygen atoms of the syn-conformation of the N-hydroxy-
Second, the N-aryl substituent also makes a major
contribution to binding by interacting with a hydropho-
* To whom correspondence should be addressed. Tel: 410-455-2518.
Fax: 410-455-2608. E-mail: creighto@umbc7.umbc.edu.
10.1021/jm000160l CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/26/2000

3982 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 21
Kalsi et al.
Ta ble 1. Inhibition Constants (Ki’s) of
Sch em e 1. Synthesis of
S-(N-Alkyl-N-hydroxycarbamoyl)glutathiones 2c-h a
S-(N-Aryl-N-hydroxycarbamoyl)glutathiones with Human
Erythrocyte, Yeast, and P. putida GlxIa
Ki (µM)
compd
πb
humanc
yeastc
P. putida
2b
1a
1b
1c
0.54
2.13
3.04
3.25
1.7 ( 0.1
0.16 ( 0.1
0.046 ( 0.004
0.014 ( 0.001
68 ( 5
86 ( 9
28 ( 3
16 ( 1
10 ( 2
11 ( 1
3.6 ( 0.3
1.2 ( 0.2
a
b
Phosphate buffer (50 mM, pH 7), 25 °C. Hansch hydropho-
bicity constants for R-substituents obtained from ref 20. c From
ref 9.
a
Reagents and conditions: (i) (C2H5)3SiH; (ii) 4-chlorophenyl
chloroformate; (iii) GSH/EtOH-H2O, pH 9.
carbamoyl ester function, and the N-aryl substituent
occupies a hydrophobic pocket composed of Phe 67A, Phe
62A, Leu 69A, Cys 60A, Phe 71A, Ile 88A, Leu 92A, Leu
174B, Leu 160B, Phe 162B, and Met 157B.
As part of a research program aimed at developing
highly specific, tight-binding inhibitors of human GlxI,
we have probed the dimensions and properties of the
hydrophobic binding pocket using both experimental
and computational methods. Essential to the conclu-
sions of this study is a comparison of the inhibition
constants of the N-aryl-N-hydroxycarbamoyl esters of
GSH (1a -c) versus the homologous series of N-alkyl-
N-hydroxycarbamoyl esters of GSH (2a -h ) shown in
Chart 1.
F igu r e 1. Log plot of competitive inhibition constants (Ki’s)
versus the Hansch hydrophobicity constants of the R-substit-
uents for S-(N-methyl-N-hydroxycarbamoyl)glutathione (2b)
and different S-(N-aryl-N-hydroxycarbamoyl)glutathiones 1a-c
with human, yeast, and P. putida GlxI. The data for human
and yeast GlxI were taken from ref 9.
Ch em istr y a n d En zym ology
Resu lts a n d Discu ssion
The S-(N-aryl-N-hydroxycarbamoyl)glutathione de-
rivatives 1a -c were prepared by reaction of GSH with
the 4-chlorophenyl esters of the corresponding N-aryl-
N-hydroxycarbamates, as previously described by this
laboratory.9 S-(N-Hydroxycarbamoyl)glutathione (2a )
was prepared by an acyl-interchange reaction between
N-hydroxycarbamate 4-chlorophenyl ester (3a ) and
GSH. The acylating reagent (3a ) was prepared by
reacting hydroxylamine with 4-chlorophenyl chlorofor-
mate. S-(N-Hydroxy-N-methylcarbamoyl)glutathione (2b)
was prepared by a published method from this labora-
tory.8 The S-(N-alkyl-N-hydroxycarbamoyl)glutathiones
2c-h were synthesized as outlined in Scheme 1. The
4-chlorophenyl esters of the N-alkyl-N-hydroxycarbam-
ates 3c-h were prepared using a modification of the
general method of Wu and Sun.13 In a one-pot reaction
mixture, triethylsilane was first used to reduce the alkyl
oximes 5c-h to the corresponding hydroxylamines 4c-
h , followed by reaction with 4-chlorophenyl chlorofor-
mate to give 3c-h . Reaction of GSH with 3c-h gave
crude preparations of 2c-h , which were purified to
apparent homogeneity by differential precipitation or
reverse-phase HPLC. NMR spectral assignments were
based on comparisons with previously published NMR
studies of GSH and its derivatives.14
The presence of a hydrophobic binding pocket in the
active site of yeast GlxI was first suggested by Vince
and co-workers, on the basis of the progressive decrease
in the inhibition constants of simple S-aryl and S-alkyl
GSH derivatives with increasing hydrophobicity of the
S-substituent.18
Role of Hyd r op h obicity in In h ibitor Bin d in g.
Indeed, the binding affinities of the enediol analogues
2b and 1a -c for human, yeast, and P. putida GlxI
increase with increasing hydrophobicity of the N-sub-
stituent (Table 1), indicated by the inverse relationship
between log Ki and the hydrophobicity constants (π) of
the N-methyl and N-aryl functions, Figure 1.19 The π
constant is defined as the log of the n-octanol/water
partition ratio for the N-aryl function, calculated ac-
cording to the method of Hansch.20 The slopes of the
lines through the data measure the change in the free
energy of transfer of the enediol analogues from aqueous
buffer at pH 7 to the active sites of the enzymes relative
to the change in the free energy of transfer of the
analogues between buffer and n-octanol. Thus, the
slopes for the yeast and human enzymes can be inter-
preted to indicate that the hydrophobicities of the
respective binding pockets are about 60% and 70% that
of n-octanol. The shallower slope for P. putida GlxI
indicates a hydrophobicity about 30% that of n-octanol.
Polar/charged residues that are not present in the
hydrophobic pockets of the human and yeast enzymes
can explain the smaller apparent hydrophobicity of the
Human erythrocyte GlxI was purified to homogeneity
from outdated human blood by a published procedure.15
Pseudomonas putida GlxI was purified from Escherichia
coli BL21(DE3) transformed with pBTac1/GlxI,16 using
methods described elsewhere.17

Hydrophobic Interactions in GSH Binding to GlxI
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 21 3983
Ta ble 2. Inhibition Constants (Ki’s) of
S-(N-Alkyl-N-hydroxycarbamoyl)glutathiones with Human
Erythrocyte GlxIa
compd
πb
Ki (µM)
compd
πb
Ki (µM)
2a
2c
2d
2e
0
183 ( 50
2f
2g
2h
2.70
3.24
3.78
0.17 ( 0.08
0.016 ( 004
0.018 ( 0.011
1.08
1.62
2.16
1.18 ( 0.07
0.80 ( 0.5
0.18 ( 0.09
a
b
Phosphate buffer (100 mM, pH 7), 25 °C. Hansch hydropho-
bicity constants for R-substituents calculated according to ref 20.
binding pocket in the bacterial enzyme. While there is
no reported structural data for the yeast and bacterial
enzymes, sequence comparisons with the human en-
zyme suggest the presence of analogous hydrophobic
pockets in the active sites of the yeast and bacterial
enzymes. In the human enzyme, Leu 92A and Ile 88A
are 2 of 11 residues that compose the hydrophobic
pocket identified in the X-ray crystal structure.21 These
two residues are replaced, respectively, by Lys and His
residues in the sequence of amino acids that comprise
the apparent hydrophobic pocket in the bacterial en-
zyme, presumably resulting in a lower overall hydro-
phobicity. In contrast, Leu 92A and Ile 88A are conser-
vatively replaced by two Phe residues in the yeast
enzyme, consistent with the similar observed hydro-
phobicities for the yeast and human enzymes.
The binding affinities to human GlxI of a homologous
series of S-(N-alkyl-N-hydroxycarbamoyl)glutathiones
2a -h (Chart 1) were determined for comparison with
those of the N-aryl derivatives, Table 2. Initially, we
anticipated that the slope of the log Ki versus π plot
would be shallower than that observed with the N-aryl
derivatives (Figure 1), because of the increasing con-
formational flexibility of the N-alkyl function proceeding
from N-methyl to N-heptyl. Conformational flexibility
(entropy) should negatively impact binding affinity, if
the active site binds only a limited subset of N-alkyl
conformers that exist in bulk solvent. By one estimate,
for each degree of rotational freedom lost upon binding
of a ligand to a protein, there is roughly a 9-fold decrease
in binding affinity.22 This effect will be less important
proceeding from N-methyl to the N-aryl derivatives, as
there is no systematic change in conformational flex-
ibility of the N-substituent. However, contrary to ex-
pectation, the slopes for the two classes of compounds
are identical within experimental error, Figure 2. This
suggests either that there is little change in conforma-
tional flexibility of the N-alkyl derivatives upon binding
to the enzyme or that there are compensating favorable
enthalpic and/or entropic terms (due to solvent/protein
reorganization), which are not reflected in the binding
affinities of the N-aryl derivatives to the active site. In
either case, conformational flexibility appears not to
adversely affect the binding affinity of the N-alkyl enediol
analogues with the active site of GlxI.
F igu r e 2. Log plot of competitive inhibition constants (Ki’s)
versus the Hansch hydrophobicity constants of the R-substit-
uents for S-(N-alkyl-N-hydroxycarbamoyl)glutathiones 2b-h
(0) and S-(N-aryl-N-hydroxycarbamoyl)glutathiones 1a -c (2)
with human GlxI. The solid line is the result of linear
regression analysis of the data for S-(N-alkyl-N-hydroxycar-
bamoyl)glutathiones. The dashed lines demarcate the 95%
confidence interval. Also shown is the experimentally deter-
mined inhibition constant for S-(N-hydroxycarbamoyl)glu-
tathione (2a ).
affinity for the human enzyme than for the yeast and
bacterial enzymes. Sequence comparisons indicate that
one of the ligands (Gln 33) to the active site Zn2+ of the
human enzyme is replaced by His in the yeast and
bacterial enzymes.21 Conceivably, this might indirectly
influence the stability of the direct coordination interac-
tion between the Zn2+ and the N-hydroxycarbamoyl
ester function of the bound enediol analogues. The
difference in binding affinities is less easily explained
by differential interactions with the glutathionyl moiety,
as the active site residues that interact with the
glutathionyl backbone of the bound inhibitors are
strictly conserved among the enzymes from the three
different biological sources.21
Of substantial interest is the observation that the
experimentally measured value of the inhibition con-
stant for S-(N-hydroxycarbamoyl)glutathione (2a , Ki )
183 µM) is 35-fold larger than the extrapolated value
of 5.2 µM for this compound obtained from the intercept
on the log Ki axis of the hydrophobicity plot of Figure
2. This strongly indicates that occupancy of the hydro-
phobic binding pocket somehow optimizes polar interac-
tions between the enzyme and the bound ligand, per-
haps by promoting alignments to the active site zinc ion.
This is the first evidence for a cooperative relationship
between the polar and hydrophobic interactions associ-
ated with binding an enediol analogue to the active site
of GlxI.
Com p u ta tion a l Stu d ies. The binding of the N-aryl
enediol analogues to the active site could be adversely
affected by the presence of steric restrictions in the
hydrophobic binding pocket. For example, the X-ray
crystal structure of the binary GlxI‚1d complex shows
an orthogonal relationship between the plane of the
phenyl ring and the plane of the N-hydroxycarbamoyl
ester function.12 The question of whether this might
reflect steric bumping between the phenyl ring and
active site residues was addressed using computational
methods.
Role of P ola r In ter a ction s in In h ibitor Bin d in g.
The intercept values on the log Ki axis of Figure 1 reflect
the contribution of polar interactions to binding affinity.
These values are the hypothetical inhibition constants
for an enediol analogue in which the N-aryl function is
replaced by N-H (π ) 0), a functionality that cannot
interact with the enzyme hydrophobically. The differ-
ence in the intercept values indicates that polar interac-
tions make a 25-fold greater contribution to binding

3984 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 21
Kalsi et al.
F igu r e 3. Geometry-optimized structures of the syn- and anti-conformations of S-(N-4-chlorophenyl-N-hydroxycarbamoyl)-
thioethane. Critical dihedral angles are given in Table 3.
Ta ble 3. Critical Dihedral Angles (deg) for the
Geometry-Optimized Syn- and Anti-Conformers of
Exp er im en ta l a n d Com p u ta tion a l Meth od s
Synthetic methods are outlined in Scheme 1. NMR spectra
were taken on a GE QE-300 NMR spectrometer. Mass spectral
data were obtained at the Midwest Center for Mass Spectrom-
etry, University of Nebraska-Lincoln, and the Washington
University Mass Spectrometry Resource, Washington Univer-
sity-Saint Louis. Elemental analyses were obtained at At-
lantic Microlabs, Inc., Norcross, GA, and are within 0.4% of
the calculated values unless otherwise indicated. Oximes 5c-h
were synthesized from the corresponding aldehydes by stan-
dard methods.23 NMR and IR analyses of the oximes matched
published standard spectra. All other reagents were purchased
from Aldrich.
S-(N-4-Chlorophenyl-N-hydroxycarbamoyl)thioethane (Figure 3)
Obtained at the HF/6-31G*Level of Theory
bonds
syn-conformer
anti-conformer
C(2)-C(4)-N(8)-C(13)
C(2)-C(4)-N(8)-O(14)
C(7)-C(4)-N(8)-C(13)
C(7)-C(4)-N(8)-O(14)
C(4)-N(8)-C(13)-O(16)
C(4)-N(8)-C(13)-S(17)
O(14)-N(8)-C(13)-O(16)
O(14)-N(8)-C(13)-S(17)
-90.75
122.76
90.78
-55.71
-160.16
23.01
-56.27
156.48
125.49
-21.77
21.18
-161.26
168.32
-14.13
-14.29
168.89
Competitive inhibition constants were calculated from the
variation in the apparent Km of GlxI with GSH-methylgly-
oxal-thiohemiacetal substrate in the presence of different
concentrations of enediol analogue.9 Apparent Km values were
obtained from computer fits of the initial rate data to the
Michaelis-Menten equation.
The lowest-energy gas-phase conformation of S-(N-4-chlo-
rophenyl-N-hydroxycarbamoyl)thioethane was calculated us-
ing the molecular modeling program Spartan (Wavefunction,
Inc., Irvine, CA).
N-Hyd r oxy-N-h ep tylca r ba m a te 4-Ch lor op h en yl Ester
(3h ). This compound was prepared by a modification of a
published method.23 Heptanal oxime 5h (1.0 g, 0.008 mol) was
dissolved in 10 mL chloroform in a dry, nitrogen-flushed vessel
on ice. Triethylsilane (0.90 g, 0.008 mol) was added dropwise,
followed by dropwise addition of 4-chlorophenyl chloroformate
(1.47 g, 0.008 mol). The mixture was stirred overnight under
relative energy (kcal/mol)
0
-3.4
The energy-minimized syn- and anti-conformers of the
model compound, S-(N-4-chlorophenyl-N-hydroxycar-
bamoyl)thioethane, were calculated by both semiem-
pirical and ab initio methods, Figure 3. The conforma-
tion of the S-substituent of the less stable syn-conformer
is nearly identical to that of the S-substituent of the
syn-conformer of 1d observed in the X-ray crystal
structure. At the Hartree-Fock/6-31G* level of theory,
the dihedral angle between the plane of phenyl ring and
that of the N-hydroxycarbamoyl ester function is 91°,
which is close to that observed in the X-ray structure,
Table 3. Apparently, the geometry of the ground-state
structure of the model compound in the gas phase is
determined largely by steric interactions between the
ortho-protons of the ring and the N-OH and S-ethyl
functions. By inference, human GlxI likely binds a low-
energy conformer of syn-1d in which the N-aryl function
is orthogonal to the plane of the N-hydroxycarbamoyl
ester function. Thus, computational methods indicate
that the conformation of the S-substituent of the enzyme-
bound enediol analogue is a property of the ligand and
not the result of steric interactions with the enzyme
protein.
a
nitrogen atmosphere at room temperature. The crude
mixture was purified by flash gel chromatography, eluting with
chloroform-methanol (40:1 to 30:1). The solvent was removed
in vacuo to give a colorless oil, which crystallized to white
needles on standing. The product was recrystallized from
ether-hexane. Yield: 40%. Mp: 54 °C. 300 MHz 1H NMR
(CDCl3, TMS): δ 0.86-0.90 (m, 3H), 1.27-1.32 (m, 8H), 1.67-
1.72 (m, 2H), 1.78 (s, OH), 3.65 (t, J ) 7.2 Hz, 2H), 7.06 (d, J
) 8.7 Hz, aromatic-2H), 7.33 (d, J ) 8.7 Hz, aromatic-2H). IR
(KBr): 3250, 2920, 2840, 1670, 1480, 1420, 1220, 1160, 1080,
1010, 870, 810, 750 cm-1. Anal. (C14H20NClO3) C, H, N.
N-Hyd r oxy-N-h exylca r ba m a te 4-Ch lor op h en yl Ester
(3g). This compound was prepared by the same general
method used to prepare 3h . The product was recrystallized
from ether-hexane as colorless plates. Yield: 28%. Mp: 61-
62 °C. 300 MHz 1H NMR (CDCl3, TMS): δ 0.86-0.91 (m, 3H),
1.2-1.4 (m, 8H), 1.6-1.8 (m, 2H), 3.66 (t, J ) 7.2 Hz, 2H),
7.06 (d, J ) 8.7 Hz, aromatic-2H), 7.33 (d, J ) 8.7 Hz,
aromatic-2H). IR (KBr): 3250, 2900, 2840, 1660, 1480, 1420,
1220, 1160, 1080,1010,870, 810, 750 cm-1. Anal. (C13H18NClO3)
C, H; N: calcd, 5.15; found, 5.08.
Im p lica tion s for Ca ta lysis a n d In h ibitor Design .
The inhibitor binding studies reported here indicate
the presence of a large, diffuse hydrophobic binding
pocket in the active site of human GlxI that plays a
major role in binding the enediol analogues to the
active site. Occupancy of this pocket is necessary in
order to maximize polar interactions between the en-
zyme and bound enediol analogue. By inference, the
methyl group of GSH-methylglyoxal thiohemiacetal
substrate might participate in catalysis by helping
to optimize interactions between active site residues
and the enediol intermediate and flanking transition
states. Each of these factors must be given careful
consideration in the design of future mechanism-based
competitive inhibitors of GlxI as potential antitumor
agents.
N-Hyd r oxy-N-p en tylca r ba m a te 4-Ch lor op h en yl Ester
(3f). This compound was prepared by the same general method
used to prepare 3h . The product was recrystallized from
ether-hexane as white rhombahedral crystals. Yield: 46%.
1
Mp: 47-48 °C. 300 MHz H NMR (CDCl3, TMS): δ 0.91 (t, J
) 6.6 Hz, 3H), 1.28-1.37 (m, 4H), 1.72 (m, 2H), 3.67 (t, J )
6.9 Hz, 2H), 6.66 (bs, OH), 7.07 (d, J ) 8.7 Hz, aromatic-2H),
7.04 (d, J ) 8.7 Hz, aromatic-2H). IR (KBr): 3250, 2920, 2845,
1660, 1480, 1420, 1210, 1160, 1080, 1005,865, 810, 745 cm-1
Anal. (C12H16NClO3) C, H; N: calcd, 5.43; found, 5.37.
.

Hydrophobic Interactions in GSH Binding to GlxI
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 21 3985
N-Hyd r oxy-N-bu tylca r ba m a te 4-Ch lor op h en yl Ester
(3e). This compound was prepared by the same general
method used to prepare 3h , using chloroform-ethyl acetate
(100:1) as a chromatographic solvent. The product was recrys-
tallized from ether-hexane as white flat needles. Yield: 23%.
Glu-CγH2), 3.04 (q, J ) 9.0, 14.4 Hz, Cys-CâHb), 3.24 (q, J )
4.2, 14.4 Hz, Cys-CâHa), 3.27 (t, J ) 6.6 Hz, Glu-CR H), 3.57
(t, J ) 6.9 Hz, N-CH2), 3.72 (d, J ) 17.1 Hz, Gly-CRHa), 3.80
(d, J ) 17.1 Hz, Gly-CRHb), 4.50 (q, J ) 4.2, 9.0 Hz, Cys-CRH).
FAB MS consistent with C17H30N4SO8. Anal. (C17H30N4SO8)
C, H; N: calcd, 12.44; found, 12.28.
1
Mp: 42-43 °C. 300 MHz H NMR (CDCl3, TMS): δ 0.93 (t, J
) 7.2 Hz, 3H), 1.37 (m, 2H), 1.68 (m, 2H), 3.66 (t, J ) 6.9 Hz,
2H), 7.2 (bs, OH), 7.06 (d, J ) 8.7 Hz, aromatic-2H), 7.333 (d,
J ) 8.7 Hz, aromatic-2H). IR (KBr): 3250, 2920, 2850, 1640,
1460, 1400, 1260, 1190, 1130, 1050, 1020, 980, 845, 790, 725
cm-1. Anal. (C11H14NClO3) C; H: calcd, 5.79; found, 5.73; N:
calcd, 5.75; found, 5.83.
S-(N-P en tyl-N-h ydr oxycar bam oyl)glu tath ion e (2f). This
compound was prepared by the same general method used to
prepare 2h . Yield: 53%. 300 MHz 1H NMR (D2O, pD 10.4,
DSS): δ 0.87 (methyl-3H), 1.28 (m, alkyl-4H), 1.61 (m, alkyl-
2H), 1.91 (m, Glu-CâH2), 2.39 (m, Glu-CγH2), 3.04 (q, J ) 9.0,
14.4 Hz, Cys-CâHb), 3.26 (q, J ) 4.2, 14.4 Hz, Cys-CâHa), 3.35
(t, J ) 6.3 Hz, Glu-CRH), 3.57 (t, J ) 6.9 Hz, N-CH2), 3.73 (d,
J ) 17.1 Hz, Gly-CRHa), 3.80 (d, J ) 17.1 Hz, Gly-CRHb), 4.52
(q, J ) 4.2, 9.0 Hz, Cys-CRH). FAB MS consistent with
N-Hyd r oxy-N-p r op ylca r ba m a te 4-Ch lor op h en yl Ester
(3d ). This compound was prepared by the same general
method used to prepare 3h , using chloroform-ethyl acetate
(100:1) as a chromatographic solvent. The product was crystal-
lized from ether-hexane as colorless flat needles. Yield: 21%.
C
16H28N4SO8. Anal. (C16H28N4SO8) C; H: calcd, 6.47; found,
6.38; N: calcd, 12.84; found, 12.71.
1
Mp: 58-59 °C. 300 MHz H NMR (CDCl3, TMS): δ 0.97 (t, J
S-(N-Bu tyl-N-h yd r oxyca r ba m oyl)glu ta th ion e (2e). This
compound was prepared by the same general method used to
prepare 2h , using a reaction time of 115 h. The crude, acidified
product was purified by flash chromatography on a silica gel
column, using n-propanol:acetic acid:water (10:1:5) as an
eluting solvent. The fractions containing the crude product
were pooled, brought to dryness and further purified by
reverse-phase column chromatography (Whatman µBondapak
C18, 0.78 × 30 cm), using 0.25% acetic acid and 35% methanol
in water as an eluting solvent (retention volume: ∼26 mL).
The peak fractions were lyophilized to dryness to give the final
product as a white powder. Yield: 35%. 300 MHz 1H NMR
(D2O, pD 3.3, DSS): δ 0.89 (t, J ) 7.3 Hz methyl-3H), 1.28
(m, alkyl-2H), 1.59 (m, alkyl-2H), 2.15 (m, Glu-CâH2), 2.51 (m,
Glu-CγH2), 3.14 (q, J ) 8.4, 14.3 Hz, Cys-CâHb), 3.37 (q, J )
4.8, 14.3 Hz, Cys-CâHa), 3.65 (t, J ) 7.0 Hz, N-CH2), 3.79 (t, J
) 6.6 Hz, Glu-CRH), 3.94 (s, Gly-CRH2), 4.62 (q, J ) 4.8, 8.4
Hz, Cys-CRH). FAB MS consistent with C15H26N4SO8. Anal.
(C15H26N4SO8‚H2O) C: calcd, 37.86; found, 37.46; H: calcd,
5.87; found, 5.76; N: calcd, 13.58; found, 13.26.
) 7.0 Hz, 3H), 1.76 (m, 2H), 3.64 (t, J ) 7.0 Hz, 2H), 6.84 (bs,
OH), 7.07 (d, J ) 8.8 Hz, aromatic-2H), 7.33 (d, J ) 8.8 Hz,
aromatic-2H). IR (KBr): 3230, 2940, 2920, 2860, 1655, 1480,
1460, 1420, 1285, 1240, 1220, 1170, 1080, 1040,1005, 865, 810,
745 cm-1. Anal. (C10H12NClO3) C; H: calcd, 5.27; found, 5.32;
N: calcd, 6.10; found, 6.05.
N-Hyd r oxy-N-et h ylca r b a m a t e 4-Ch lor op h en yl E st er
(3c). This compound was prepared by the same general
method used to prepare 3h , using chloroform-ethyl acetate
(100:1) as a chromatographic solvent. The product was crystal-
lized from ether-hexane as white needles. Yield: 16%. Mp:
67-68 °C. 300 MHz 1H NMR (CDCl3, TMS): δ 1.31 (t, J ) 7.0
Hz, 3H), 3.73 (t, J ) 7.0 Hz, 2H), 6.40 (bs, OH), 7.08 (d, J )
8.8 Hz, aromatic-2H), 7.34 (d, J ) 8.8 Hz, aromatic-2H). IR
(KBr): 3250, 2970, 2910, 2880, 2860, 1665, 1480, 1425, 1270,
1220, 1170, 1080, 1030, 1005, 960, 865, 810, 745 cm-1. Anal.
(C9H10NClO3) C; H: calcd, 4.67; found, 4.79; N: calcd, 6.50;
found, 6.38.
N-Hyd r oxyca r ba m a te 4-Ch lor op h en yl Ester (3a ). This
compound was prepared by reacting hydroxylamine with
4-chlorophenyl chloroformate. The product was crystallized
from ether-hexane as colorless plates. Yield: 34%. Mp: 137-
139 °C. 300 MHz 1H NMR (CDCl3, TMS): δ 5.83 (bs, 1H), 7.11
(d, J ) 8.7 Hz, 2H), 7.38 (d, J ) 8.7 Hz, 2 H), 7.47 (bs, 1H). IR
S-(N-P r opyl-N-h ydr oxycar bam oyl)glu tath ion e (2d). This
compound was prepared by the same general method used to
prepare 2e. The flash chromatography column was eluted with
n-propanol:acetic acid:water (70:25:30). The reverse-phase
column was eluted with 0.25% acetic acid and 25% methanol
in water (retention volume: ∼23 mL). The peak fractions were
lyophilized to dryness to give the final product as a white
powder. Yield: 35%. 300 MHz 1H NMR (D2O, pD 3.3, DSS): δ
0.86 (t, J ) 7.3 Hz, methyl-3H), 1.63 (m, J ) 7.0, J ) 7.3 alkyl-
2H), 2.14 (m, Glu-CâH2), 2.51 (m, Glu-CγH2), 3.12 (q, J ) 8.4,
14.3 Hz, Cys-CâHb), 3.38 (q, J ) 4.8, 14.3 Hz, Cys-CâHa), 3.61
(t, J ) 7.0 Hz, N-CH2), 3.79 (t, J ) 6.2 Hz, Glu-CRH), 3.94 (s,
Gly-CRH2), 4.62 (q, J ) 4.8, 8.4 Hz, Cys-CRH). FAB MS
consistent with C14H24N4SO8. Anal. (C14H24N4SO8‚H2O) C; H:
calcd, 6.15; found, 5.90; N: calcd, 6.50; found, 6.38.
(KBr): 3300, 1725, 1465, 1260, 1210, 1080, 1005, 840 cm-1
.
Anal. (C7H6NClO3) C; H: calcd, 3.22; found, 3.34; N: calcd,
7.47; found, 7.39.
S-(N-Heptyl-N-h ydr oxycar bam oyl)glu tath ion e (2h ). Into
a stirring solution of 6.3 mL degassed, nitrogen-saturated
ethanol:water (2:1) and 3h (53 mg, 0.185 mmol) was placed a
6.3-fold excess of glutathione (563 mg, 1.16 mmol). The slurry
was slowly brought to pH 9 by the dropwise addition of 6 N
NaOH, during which the mixture became a homogeneous
solution. The solution was placed under nitrogen and allowed
to stand at room temperature for 24 h. The solution was
brought to pH 3.5 with 6 N HCl and the solvent removed in
vacuo. The white residue was suspended in 1 mL water, stirred
for 9 h at room temperature, and the precipitate collected by
centrifgation. This digestion procedure was repeated 4 more
times. The dried white solid was then triturated 3 times with
0.8 mL portions of diethyl ether to remove unreacted 3h and
unreacted 4-chlorophenol. Yield: 35%. 300 MHz 1H NMR (D2O,
pD 10.4, HOD reference): δ 0.88 (methyl-3H), 1.2-1.4 (m,
alkyl-8H), 1.63 (m, alkyl-2H), 1.91 (m, Glu-CâH2), 2.40 (m, Glu-
CγH2), 3.07 (q, J ) 9.0, 14.4 Hz, Cys-CâHa), 3.28 (q, J ) 4.2,
14.4 Hz, Cys-CâHb), 3.32 (t, J ) 6.0 Hz, Glu-CRH), 3.59 (t, J )
6.6 Hz, N-CH2), 3.75 (d, J ) 17.1 Hz, Gly-CRHa), 3.83 (d, J )
17.1 Hz, Gly-CRHb), 4.53 (q, J ) 4.2, 9.0 Hz, Cys-CRH). FAB
MS consistent with C18H32N4SO8. Anal. (C18H32N4SO8) C, H;
N: calcd, 12.06; found, 11.92.
S-(N-Eth yl-N-h yd r oxyca r ba m oyl)glu ta th ion e (2c). This
compound was prepared by the same general method used to
prepare 2e. The flash chromatography column was eluted with
n-propanol:acetic acid:water (130:1:70). The reverse-phase
column was eluted with 0.25% acetic acid and 5% methanol
in water (retention volume: ∼24 mL). The peak fractions were
lyophilized to dryness to give the final product as a white
powder. Yield: 48%. 300 MHz 1H NMR (D2O, pD 3.3, DSS): δ
1.16 (t, J ) 7.0 Hz, methyl-3H), 2.15 (m, Glu-CâH2), 2.51 (m,
Glu-CγH2), 3.12 (q, J ) 8.4, 14.3 Hz, Cys-CâHb), 3.38 (q, J )
4.8, 14.3 Hz, Cys-CâHa), 3.66 (t, J ) 7.0 Hz, N-CH2), 3.79 (t, J
) 6.2 Hz, Glu-CRH), 3.94 (s, Gly-CRH2), 4.63 (q, J ) 4.8, 8.4
Hz, Cys-CRH). FAB MS consistent with C13H22N4SO8. Anal.
(C13H22N4SO8‚H2O) C: calcd, 37.86; found, 37.46; H: calcd,
5.87; found, 5.76; N: calcd, 13.58; found, 13.26.
S-(N-Hyd r oxyca r ba m oyl)glu ta th ion e (2a ). This com-
pound was prepared by the same general method used to
prepare 2h , using a 3.5-fold excess of GSH and a reaction time
of 4 h. The crude reaction mixture was treated with 4-pyridine
disulfide in order derivatize any unreacted GSH to the mixed
S-(N-Hexyl-N-h ydr oxycar bam oyl)glu tath ion e (2g). This
compound was prepared by the same general method used to
prepare 2h , using a reaction time of 70 h. Yield: 66%. 300
MHz H NMR (D2O, pD 10.4, DSS): δ 0.86 (methyl-3H), 1.28
(m, alkyl-6H), 1.60 (m, alkyl-2H), 1.87 (m, Glu-CâH2), 2.36 (m,
1

3986 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 21
Kalsi et al.
(11) Sharkey, E. M.; O’Neill, H. B.; Kavarana, M. J .; Wang, H.;
Creighton, D. J .; Sentz, D. L.; Eiseman, J . L. Pharmacokinetics
and Antitumor Properties in Tumor-bearing Mice of an Enediol
Analogue Inhibitor of Glyoxalase I. Cancer Chemother. Phar-
macol. 2000, 46, 156-166.
(12) Cameron, A. D.; Ridderstrom, M.; Olin, B.; Kavarana, M. J .;
Creighton, D. J .; Mannervik, B. Reaction Mechanism of Glyox-
alase I Explored by an X-ray Crystallographic Analysis of the
Human Enzyme in Complex with a Transition State Analogue.
Biochemistry 1999, 38, 13480-13490.
disulfide. This allowed a clean separation of the desired
product by reverse-phase column chromatography using 0.25%
acetic acid in water (retention volume: ∼16 mL). The peak
fractions were lyophilized to dryness to give the final product
1
as a white powder. Yield: 30%. 300 MHz H NMR (D2O, pD
3.2, DSS): δ 2.15 (m, Glu-CâH2), 2.51 (m, Glu-CγH2), 3.20 (q,
J ) 8.1, 14.3 Hz, Cys-CâHb), 3.43 (q, J ) 4.8, 14.3 Hz, Cys-
CâHa), 3.81 (t, J ) 6.2 Hz, Glu-CRH), 3.95 (s, Gly-CRH2), 4.66
(q, J ) 4.8, 8.1 Hz, Cys-CRH). FAB MS consistent with
(13) Wu. P.-L.; Sun, C.-J . A Simple Method for the Preparation of
N-Substituted Hydroxamic Acids by Reductive Acylation of
Oximes. Tetrahedron Lett. 1991, 32, 4137-4138.
C
11H18N4SO8. Anal. (C11H18N4SO8‚H2O) C; H: calcd, 5.24;
found, 5.17; N: calcd, 14.58; found, 14.36.
(14) Rabenstein, D. L.; Keire, D. A. The Glyoxalase System. In
Coenzymes and Cofactors: Glutathione; Dolphin, D., Poulson,
R., Avramovic, O., Eds.; J ohn Wiley and Sons: New York, 1989;
Vol. 3 (Part A), pp 67-101.
(15) Aronsson, A.-C.; Tibbelin, G.; Mannervik, B. Purification of
Glyoxalase I from Human Erythrocytes by the Use of Affinity
Chromatography and Separation of Three Isozymes. Anal.
Biochem. 1979, 92, 390-393.
Ack n ow led gm en t. This work was supported by
grants from the National Cancer Institute (CA 59612)
and the U.S. Army Medical Research and Materiel
Command.
Refer en ces
(16) Lu, T.; Creighton, D. J .; Antoine, M.; Fenselau, C.; Lovett, P. S.
The Gene Encoding Glyoxalase I from Pseudomonas putida:
Cloning, Overexpression, and Sequence Comparisons with Hu-
man Glyoxalase I. Gene 1994, 150, 93-96.
(17) Rhee, H.-I.; Murata, K.; Kimura, A. Purification and Charac-
terization of Glyoxalase I from Pseudomonas putida. Biochem.
Biophys. Res. Commun. 1986, 141, 993-999.
(18) Vince, R.; Daluge, S.; Wadd, W. Studies on the Inhibition of
Glyoxlase I by S-Substituted Glutathiones J . Med. Chem. 1971,
14, 402-405.
(1) Thornalley, P. J . Glutathione-Dependent Detoxification of Alpha-
Oxoaldehydes by the Glyoxalase System: Involvement in Dis-
ease Mechanisms and Antiproliferative Activity of Glyoxalase I
Inhibitors. Chem. Biol. Interact. 1998, 111-112, 137-151.
(2) Creighton, D. J .; Hamilton, D. S.; Kavarana, M. J .; Sharkey, E.
M.; Eiseman, J , L. Glyoxalase Enzyme System as a Potential
Target for Antitumor Drug Development. Drugs Future 2000,
25, 385-392.
(3) Vander J agt, D. L. The Glyoxalase System. In Coenzymes and
Cofactors: Glutathione; Dolphin, D., Poulson, R., Avramovic, O.,
Eds.; J ohn Wiley and Sons: New York, 1989; Vol. 3 (Part A), pp
597-641.
(4) Creighton, D. J .; Pourmotabbed, T. Glutathione-Dependent
Aldehyde Oxidation Reactions. In Molecular Structure and
Energetics: Principles of Enzyme Activity; Liebman, J . F.,
Greenberg, A., Eds.; VCH Publishers: New York, 1988; Vol. 9,
pp 353-386.
(5) Richard, J . P. Kinetic Parameters for the Elimination Reaction
Catalyzed by Triosephosphate Isomerase and an Estimation of
the Reaction’s Physiological Significance. Biochemistry 1991, 30,
4581-4585.
(6) White, J . S.; Rees, K. R. Inhibitory Effects of Methylglyoxal on
DNA, RNA and Protein Synthesis in Cultured Guinea Pig
Karatocytes. Chem. Biol. Interact. 1982, 38, 339-347.
(7) Papoulis, A.; Al-Abed, Y.; Bucala, R. Identification of N2-(1-
Carboxyethyl)guanine (CEG) as a Guanine Advanced Glycation
End Product. Biochemistry 1995, 34, 648-655.
(8) Hamilton, D. S.; Creighton, D. J . Inhibition of Glyoxalase I by
the Enediol Mimic S-(N-Hydroxy-N-methylcarbamoyl)gluta-
thione: the Possible Basis of a Tumor-Selective Anticancer
Strategy. J . Biol. Chem. 1992, 267, 24933-24936.
(9) Murthy, N. S. R. K.,; Bakeris, T.; Kavarana, M. J .; Hamilton,
D. S.; Lan, Y.; Creighton, D. J . S-(N-Aryl-N-hydroxycarbamoyl)-
glutathione Derivatives are Tight-Binding Inhibitors of Glyox-
alase I and Slow Substrates for Glyoxalase II. J . Med. Chem.
1994, 37, 2161-2166.
(10) Kavarana, M. J .; Kovaleva, E. G.; Creighton, D. J .; Wollman,
M. B.; Eiseman, J . L. Mechanism-Based Competitive Inhibitors
of Glyoxalase I: Intracellular Delivery, In Vitro Antitumor
Activities, and Stabilities in Human Serum and Mouse Serum.
J . Med. Chem. 1999, 42, 221-228.
(19) Electronic effects appear to have little influence on binding
affinity. The greater electron-withdrawing ability of the para-
substituted N-aryl substituents, relative to the N-methyl func-
tion, would be expected to decrease the stability of the direct
coordination interaction between the active site Zn2+ and the
N-hydroxycarbamoyl ester function of the bound ligands. If this
were a major factor in binding affinity, the N-aryl enediol
analogues should bind less tightly to the enzyme than the
N-methyl enediol analogue, contrary to observation. Electronic
effects on binding of the N-aryl enediol analogues to the enzyme
are probably small because the X-ray structure of the binary
GlxI‚1d complex shows that the plane of the N-aryl function is
orthogonal to the plane of the N-hydroxycarbamoyl function.
This would effectively eliminate resonance effects on the distri-
bution of charge in the N-hydroxycarbamoyl ester function.
(20) Hansch, C.; Kim, K. W.; Sarma, R. H. Structure-Activity
Relationship in Benzamides Inhibiting Alcohol Dehydrogenase.
J . Am. Chem. Soc. 1973, 95, 6447-6449.
(21) Cameron, A. D.; Olin, B.; Ridderstrom, M.; Mannervik, B.; J ones,
A. Crystal Structure of Human Glyoxalase I - Evidence for Gene
Duplication and 3D Domain Swapping. EMBO 1997, 16, 3386-
3395.
(22) Page, I. M.; J encks, W. P. Entropic Contributions to Rate
Accelerations in Enzymic and Intramolecular Reactions and the
Chelate Effect. Proc. Natl. Acad. Sci. U.S.A. 1971, 68, 1678-
1683.
(23) Sandler, S. R.; Karo, W. Organic Functional Group Preparations;
Academic Press: New York, 1972; Vol. III, Chapter 11.
J M000160L