Metal-organic compounds: A new approach for drug discovery: N1-(4-methyl-2-pyridyl)-2,3,6-trimethoxybenzamide copper(II) complex as an inhibitor of human immunodeficiency virus 1 protease

Article abstract of DOI:10.1016/S0006-2952(02)00918-8

The use of metal-organic complexes is a potentially fruitful approach for the development of novel enzyme inhibitors. They hold the attractive promise of forming stronger attachments with the target by combining the co-ordination ability of metals with th

Full text of DOI:10.1016/S0006-2952(02)00918-8

Biochemical Pharmacology 63 (2002) 1863±1873
Metal-organic compounds: a new approach for drug discovery
N1-(4-methyl-2-pyridyl)-2,3,6-trimethoxybenzamide copper(II) complex
as an inhibitor of human immunode®ciency virus 1 protease
Florence Lebona,1, Nicole Boggettob,1, Marie Ledecqa, FrancËois Duranta, Zohra Benatallahc,
c
Sames Sicsic , Rene Lapouyade , Olivier Kahn , Ange Mouithys-Mickalad ,
d,2
d,y
e
Â
Ginette Deby-Dupont , Michele Reboud-Ravaux
e
b,*
Á
a
 Â
Facultes Universitaires Notre-Dame de la Paix, Laboratoire de Chimie Moleculaire Structurale, 61rue de Bruxelles, 5000 Namur, Belgium
b
 Â
Laboratoire d'Enzymologie Moleculaire et Fonctionnelle, Departement de Biologie Cellulaire, Institut Jacques Monod,
Â
CNRS-Universites Paris 6 and 7, 2 Place Jussieu, F-75251Paris Cedex 05, France
c
 Â
Laboratoire de Reconnaissance Moleculaire et Cellulaire, Biocis (UPRES-A CNRS 8076), Faculte de Pharmacie,
 Ã
5 rue J.B. Clement, 92296 Chatenay-Malabry Cedex, France
d
 Á Â
Laboratoire des Sciences Moleculaires, Institut de Chimie de La Matiere Condensee de Bordeaux,
Avenue du Dr. Schweitzer, 33608 Pessac Cedex, France
Á
e
 Á  Á
Universite de Liege, CORD-Centre de l'Oxygene Recherche et Developpement B6A, Sart-Tilman, 4000 Liege, Belgium
Received 28 September 2001; accepted 22 February 2002
Abstract
The use of metal-organic complexes is a potentially fruitful approach for the development of novel enzyme inhibitors. They hold the
attractive promise of forming stronger attachments with the target by combining the co-ordination ability of metals with the unique
stereoelectronic properties of the ligand. We demonstrated that this approach can be successfully used to inhibit the protease of the human
immunode®ciency virus (type 1). Several ligands bearing substituents designed to interact with the catalytic site of the enzyme when
complexed to Cu2‡ were synthesised. The inhibition pattern of the resulting copper(II) complexes was analysed. We showed that the
copper(II) complexof N1-(4-methyl-2-pyridyl)-2,3,6-trimethoxybenzamide (C1) interacts with the active site of the enzyme leading to
competitive inhibition. On the other hand, N2-pyridine-amide ligands and oxazinane carboxamide ligand were found to be poor chelators
of the cupric ion under the enzymatic assay conditions. In these cases, the observed inhibition was attributed to released cupric ions which
react with cysteine residues on the surface of the protease. While unchelated metal cations are not likely to be useful agents, metal chelates
such as C1 should be considered as promising lead compounds for the development of targeted drugs. # 2002 Elsevier Science Inc. All
rights reserved.
Keywords: HIV-1 protease; Antiprotease; AIDS; Inhibition; Metal-organic complexes; Copper complexes
1. Introduction
syndrome (AIDS). This success has been largely attributed
to chemotherapies that abolish the infectivity of the pre-
dominant causative agent, the HIV-1, by inhibiting essen-
tial viral enzymes. One of these is the PR whose activity is
a prerequisite for viral replication. Two main sites have
been identi®ed as potential targets for the inhibition of
HIV-1 PR, the active site [1±4] and the interface [5,6], the
latter being largely responsible for the stabilisation of the
enzyme dimeric structure. The compounds that have
reached clinical application to date target the active site
of HIV-1 PR [7,8]. These molecules act as transition state
analogues and were designed by modifying the natural
peptidic scaffold into peptidomimetics. However, mutations
Recently, western countries have recorded a decrease in
the death rate imputed to acquired immunode®ciency
* Corresponding author. Tel.: ‡33-1-44275078; fax: ‡33-1-44275994.
E-mail address: reboud@ccr.jussieu.fr (M. Reboud-Ravaux).
1 These authors contributed equally to this work.
2 Present address: Laboratoire d'Analyse Chimique par Reconnaissance
Â
Â
Moleculaire(LACReM), Ecole Nationale Superieure de Chimie et de
Physique de Bordeaux(ENSCPB), Pessac, France.
y Deceased.
Abbreviations: HIV-1, human immunodeficiency virus type I; PR,
protease; EPR, electron paramagnetic resonance; MEP, molecular electro-
static potentials.
0006-2952/02/$ ± see front matter # 2002 Elsevier Science Inc. All rights reserved.
PII: S 0 0 0 6 - 2 9 5 2 ( 0 2 ) 0 0 9 1 8 - 8

1864
F. Lebon et al. / Biochemical Pharmacology 63 (2002) 1863±1873
Fig. 1. Metal-organic inhibitors of HIV-1 PR reported in the literature: (a) porphyrin derivative (ic50 ˆ 0:975 mM) [17], (b) BCDS-Cu(I) (Ki ˆ 1 mM) [18],
(c) setcez (Ki ˆ 480 mM) [10].
in HIV-1 PR that bypass the action of the protease inhi-
bitors appear rapidly and thus there is still an urgent need
for novel families of inhibitors. We have used de novo drug
design to investigate the potency of metal-organic com-
pounds to act as inhibitors of the HIV-1 PR [9,10]. Enzyme
inhibitors used as therapeutic agents are usually organic
molecules that interact mainly with the proteins through
hydrogen bonding and van der Waals contacts. Metal-
organic complexes hold the attractive promise of forming
stronger attachments with the target by combining the co-
ordination ability of metals (covalent and ionic bonding)
with the stereoelectronic properties of the ligand (hydrogen
bonding and van der Waals contacts) [11,12]. Complexes
with transition metals have previously been used to inhibit
proteins [13±16]. Only a few such compounds have been
reported as inhibitors of HIV-1 PR (Fig. 1). Metal-bound
porphyrins which are non-peptide compounds with useful
pharmacological properties, were identi®ed as micromolar
inhibitors of HIV-1 PR [17]. Bathocuproine sulfonic acid-
Cu(I) [BCDS-Cu(I)] was found to be a competitive inhi-
bitor of HIV-1 PR [18]. BCDS alone did not inhibit the
enzyme and the inhibition could not be attributed to free
Cu(I) ions: the inhibition by the complexremained in the
presence of ethylenediaminetetraacetic acid (EDTA),
which is known to prevent PR inhibition by uncomplexed
metals. These two features indicated that the metal com-
plexwas the active inhibitor. However, these compounds
did not emerge from a rational design aimed at using the
unique structural properties of metal-organic complexes.
We based our design on the hypothesis that a metal ion
could bind HIV-1 PR structural catalytic water molecule
and further block the activity of the enzyme. We used the
distinct Cu2‡ co-ordination geometry to design ligands that
would further bind selectively HIV-1 PR binding pockets
(Fig. 2). Based on this design, we reported that compound
diaqua [bis(2-pyridylcarbonyl) amido] copper(II) nitrate
dihydrate (setcez) behaves as a poor competitive inhibitor
(Ki ˆ 480 Æ 120 mM) of the PR [10]. Docking of this
compound into the active site of HIV-1 PR shows that
the side chains of the chelated ligand only partially occupy
the enzyme subsites. Larger ligands that could ®t into the
HIV-1 PR binding site when complexed to Cu2‡ were
designed to improve the binding af®nity of the metal-
organic complex. Compounds N1-(4-methyl-2-pyridyl)-
2,3,6-trimethoxybenzamide (1), N2-(2-methoxybenzyl)-
2-quinolinecarboxamide (2), N2-(2-methoxybenzyl)-6-
phenyl-1,3-oxazinane-2-carboxamide (7), N2-benzyl-2-
quinolinecarboxamide (8), N2-(2-cyclohexyl)-2-quinoli-
necarboxamide (9) were synthesized and complexed to
Cu2‡ leading to complexes C1, C2, C7, C8, and C9
respectively (Fig. 3). Complexes C1, C2, and C9, obtained
by crystallography, adopt a tetragonally elongated octahe-
dral geometry. Two ligands symmetrically co-ordinate
copper(II) by aromatic nitrogen and amide oxygen atoms,
forming an equatorial square plane [9]. The apical posi-
tions are occupied in the crystal by methanol or perchlorate
molecules. No crystallographic data was obtained for
complexes C7 and C8. We describe herein the inhibition
pattern analysis of this new family of metal-organic com-
pounds targeting HIV-1 PR through kinetic experiments,
EPR and molecular modelling techniques.
Fig. 2. Pharmacophore for the inhibition of HIV-1 PR by copper(II) co-
ordination compounds. The enzyme residues are represented in bold style.
Two ligands adopt an octahedral geometry able to organise the interaction
elements into the shape of the active site of the PR. Y represents an H-bond
acceptor. S1/S01 and S2/S02 represent the enzyme subsites following
Schechter and Berger nomenclature [33].

F. Lebon et al. / Biochemical Pharmacology 63 (2002) 1863±1873
1865
Fig. 3. (a) Structure of the pyridine-amide and oxazinane carboxamide ligands studied, (b) crystallographic structure of complexes C1, C2, and C9.
2. Experimental procedures
The synthesis of compound 7 was achieved through
classical methods, requiring the preparation of intermedi-
ate compounds 4, 5, and 6 (Scheme 1).
2.1. Synthesis
2.1.2. N1-(2-Methoxybenzyl)-2,2-
di(isopropylsulfanyl)acetamide (4)
Caution! Perchlorate salts are potentially explosive.
Although no detonation tendencies have been observed,
caution is advised and handling of small quantities is
recommended.
Following the method described by Qasmi et al. [19], to a
solution of 3 (5 g, 24 mmol) in anhydrous CH2Cl2 (100 mL)
was added successively at 08 under an argon atmosphere
oxalyl chloride (6.28 mL, 72 eq) and dimethylformamide
(0.8 mL). The reaction mixture was stirred at 08 for 1 hr
and then concentrated in vacuo. The resulting concen-
trate was dissolved in CH2Cl2 (100 mL). To the mixture
cooled at 08, was added dropwise 2-methoxybenzylamine
2.1.1. N1-(4-Methyl-2-pyridyl)-2,3,6-
trimethoxybenzamide (1) and N2-(2-methoxybenzyl)-2-
quinolinecarboxamide (2)
The synthesis of compounds 1 and 2 was previously
reported [9].
Scheme 1. Pathway used for the synthesis of compound 7.

1866
F. Lebon et al. / Biochemical Pharmacology 63 (2002) 1863±1873
(3.13 mL, 1 eq). The resulting mixture was stirred at 08 for
1 hr and at room temperature for 12 hr. After concentration,
the residue was treated with aqueous Na2CO3 (10%), and
the mixture was extracted with CH2Cl2 (3 Â 100 mL). The
organic fractions were pooled, washed with saturated aqu-
eous NaCl, dried over MgSO4 and concentrated in vacuo.
The residue was applied to a column packed with silica gel,
and the column was eluted (CH2Cl2:isopropanol, 95:5). The
fractions containing the product with an Rf ˆ 0:66 were
pooled and concentrated in vacuo to an oil (7 g, 90%): 1H-
NMR (CDCl3) d 7.25 (m, 3H), 6.85 (m, 2H), 4.45 (d, 2H,
J ˆ 6 Hz), 4.3 (s, 1H), 3.85 (s, 3H), 3 (m, 2H), 1.2 (m, 12H).
Anal. (C16H25NO2S2) C, H, N.
2.1.6. N2-Benzyl-2-quinolinecarboxamide (8) and N2-
(2-cyclohexyl)-2-quinolinecarboxamide (9)
Compounds 8 [20±22] and 9 [23] previously described,
were classically synthesised through the chloride deriva-
tive of 2-quinolinecarboxylic acid.
2.1.7. Copper(II) complexation of N2-(2-methoxybenzyl)-
6-phenyl-1,3-oxazinane-2-carboxamide (C7)
A solution of Cu(ClO4)2Á6H2O (123 mg in 15 mL of
methanol) was added to 109 mg of compound 7 previously
dissolved in 15 mL of methanol to give a transparent blue
solution. After several days, dark blue and light blue
crystals were formed. The dark blue crystals were dis-
solved in acetonitrile and the solution was placed into a
closed area saturated with diethyl ether vapour. Large blue
crystals that desegregate in air deposit.
2.1.3. N1-(2-Methoxybenzyl)-2-oxoacetamide (5)
To a solution of 4 (5 g, 15 mmol) in water:acetonitrile
20:80 (100 mL) cooled at 08 was added dropwise a
solution of N-bromosuccinimide (3.2 g, 18 mmol) in
acetonitrile (20 mL). The mixture was stirred at 08 for
15 min and poured in a mixture of water:CH2Cl2 1:1
(100 mL). The mixture was shacked and washed with
CH2Cl2 (2 Â 100 mL). The organic layers were pooled,
dried over MgSO4 and evaporated. Column chromato-
graphy through silica gel (ethyl acetate) gave 5 as an oily
residue (1.8 g, 63%): 1H-NMR (CDCl3) d 9.3 (s, 1H), 7.5
(t, 1H, J ˆ 7 Hz), 7.25 (m, 2H), 6.85 (m, 2H), 4.45 (d, 2H,
J ˆ 7 Hz), 3.8 (s, 3H).
2.1.8. Copper(II) complexation of N2-benzyl-2-
quinolinecarboxamide (C8)
A solution of Cu(ClO4)2Á6H2O (350 mg in 5 mL H2O)
was added to 262 mg of compound 8 previously dissolved
in 20 mL of methanol. Yellow-green crystals were formed
which became opaque in air.
2.1.9. Copper(II) complexation of N2-(2-cyclohexyl)-2-
quinolinecarboxamide (C9)
A solution of Cu(ClO4)2Á6H2O (370 mg in 5 mL of
methanol) was added to 202 mg of compound 9 previously
dissolved in 20 mL of methanol to form, after a few days,
green crystals.
2.1.4. 3-Hydroxy-3-phenylpropylamine (6)
To a mixture of LiAlH4 (2 g, 55 mmol) in anhydrous
tetrahydrofuran:ether 1:1 (40 mL) at 08 was added drop-
wise a solution of benzoylacetonitrile (2 g, 13.8 mmol) in
tetrahydrofuran:ether 1:1 (25 mL). The mixture was then
stirred and heated at re¯uxfor 12 hr. After cooling (0 8),
water (2.28 mL) was slowly added and the mixture was
stirred for 30 min, then aqueous NaOH 15% (2.28 mL) and
water (6.8 mL) were successively added. The mixture was
®ltrated and the ®ltrate was dried (MgSO4) and evaporated.
The residue was distilled (E0:2 ˆ 1208) giving rise to 6 as a
white solid (1 g, 50%): mp 608; 1H-NMR (CDCl3) d 7.37±
7.22 (m, 5H), 4.9±4.8 (m, 1H), 3±2.8 (m, 5H), 1.8±1.74 (m,
2H). Anal. (C9H13ON) C, H, N.
2.2. Biological evaluation
HIV-1 protease was kindly supplied by H.J. Schramm,
MaxPlanck Institut fu Èr Biochimie (Martinsried, Ger-
many). It was expressed using the plasmid pET9c-PR,
isolated and puri®ed as described by Billich et al. [24].
The recombinant protease mutant C67A, C95A was a gift
from D.A. Davis (National Institutes of Health, Bethesda,
USA). The ¯uorogenic substrate Dabcyl-g-Abu-S-Q-N-Y-
P-I-V-Q-Edans (Dabcyl, 4-(4-dimethylaminophenylazo)-
benzoic acid; Edans, 5-[(2-aminoethyl) amino]naphtha-
lene sulfonic acid; g-Abu, aminobutyric acid) was
purchased from Bachem. The ¯uorescence-based assays
were performed using a LS 50B Perkin Elmer lumines-
cence spectro¯uorometer equipped with a thermostated
cell holder. The spectral evolution were followed using
an Uvikon 941 spectrophotometer.
2.1.5. N2-(2-Methoxybenzyl)-6-phenyl-1,3-oxazinane-2-
carboxamide (7)
In a round bottomed ¯ask equipped with a Dean±Stark
apparatus containing a solution of 5 (0.7 g, 3.6 mmol) in
benzene (12 mL) were added pyridinium p-toluene sulfo-
nate (15 mg) and 6 (1 g, 6.6 mmol). The mixture was
re¯uxed for 4 hr and evaporated. Column chromatography
through silica gel (CH2Cl2:isopropanol, 96:4) of the resi-
due gave 7 as a white solid (0.45 mg, 38%): mp 108±1118;
1H-NMR (CDCl3) d 7.36±7.2 (m, 8H), 6.9±6.8 (m, 2H),
4.77 (s, 1H), 4.69±4.62 (m, 1H), 4.5 (d, 2H, J ˆ 6 Hz),
3.78 (s, 3H), 3.4±3.3 (m, 1H), 3.2±3 (m, 1H), 2 (s, 1H), 1.8±
1.6 (m, 2H). Anal. (C19H22O3N2) C, H, N.
2.3. Enzyme inhibition
Enzymatic assays were performed either in 100 mM
sodium acetate, 100 mM NaCl, 3% Me2SO (v/v) at pH
5.5, or in 25 mM sodium phosphate, 100 mM NaCl, 3%
Me2SO (v/v) at pH 6.0 or 6.5 and 308. The inhibitory effect
of metal-organic compounds C1, C2, C7, C8, C9 were

F. Lebon et al. / Biochemical Pharmacology 63 (2002) 1863±1873
1867
compared with that of CuCl2 at pH 5.5 (mutant protease)
and pH 6.0 or 6.5 (wild-type protease) in the following
conditions: 0.52 mL of a 3 mM solution of substrate (®nal
concentration, 5.2 mM) was added to 8.5 mL of various
concentrations of inhibitor or CuCl2 (0.5±4.7 mM) in a ®nal
volume of 300 mL. The assays were started with the
addition of HIV-1 PR (®nal concentration, 22.7 nM) or
mutant protease (®nal concentration, 97.8 nM) prediluted
in buffer containing 1 mg/mL bovine serum albumin. The
change in ¯uorescence at 490 nm (lexc ˆ 340 nm) was
monitored over a period of 5 min, 30 s after the addition of
enzyme. The dependence of the inhibitory effect of C1
towards wild-type protease on time was analysed at pH 5.5
and 308 using the progress curve method as previously
described [25]. The ®rst derivatives of the progress curve
(or velocities v) for the variation with time of the ¯uores-
cence at 490 nm (lexc ˆ 340 nm) were computer-calcu-
lated using Kaleidagraph version 3.0.1 from Abelbeck
Software. The parameter p was obtained by ®tting the
data in the equation: v ˆ v0eÀpt. The inactivation constants
ki and KI were obtained by non-linear regression ®tting of
the experimental data to the equation p ˆ ki[I] (1 À a)/
[KI ‡ ‰IŠꢀ1 À a†], with a ˆ ‰SŠ/Km ‡ ‰SŠ, [S] is the initial
substrate concentration, [I] the inhibitor concentration and
Km the Michaelis constant for the substrate with
Km ˆ 103 mM [26]. Enzyme, substrate and inhibitor C1
concentrations were 7.5 nM, 5.2 mM, 0.55±22 mM, respec-
tively.
Bruker ESP 300 E spectrometer equipped with a TM110
cavity. The samples (compounds C1, C2, C7, C8, C9, and
Cu(OAc)2, 5 mM) were prepared in a 100 mM acetate
buffer at pH 5.5 or in methanol and immediately transferred
into a quartz ¯at cell. The experimental settings were as
follow: 100 kHz modulation frequency; microwave fre-
quency 9.55 GHz; microwave power 10 mW; sweep width
1000 G; centre ®eld 3120 G; time conversion 81.92 ms;
time constant 163.84 ms; receiver gain 2 Â 104; and
number of scan 4.
2.5. Molecular modelling
All molecular mechanic calculations were performed
with the DISCOVER module in the INSIGHT II software
(Biosym/Molecular Simulations Inc., San Diego, version
97.0) on MSI Silicon Graphics Indigo 2 workstations. The
initial HIV-1 PR structure was the X-ray crystal form of
HIV-1 PR solved by Erickson et al. [27]. The previously
obtained crystal co-ordinates of compound C1 [9] were
used for docking into the active site of HIV-1 PR. Four
orientations were considered allowing the pyridyl groups
to bind either to the S1/S01 or S2/S02 pockets. The resulting
complexes formed between the enzyme and compound C1
(where the crystal apical methanol co-ordinates were sub-
stituted by more relevant physiological water molecules)
were submitted to a minimisation procedure, using the esff
metal adapted force®eld [28]. At the end of the procedure,
the energy of interaction between C1 and PR was computed
for the various orientations [29]. In the most favourable
orientation, the pyridyl group occupies the S2/S02 pockets.
Ab initio calculations on ligands 1, 2, 8 and 9 were per-
formed with an optimisation procedure in the framework of
2.4. Electron paramagnetic resonance spectroscopy
Electron paramagnetic resonance (EPR) study was per-
formed at room temperature in the X-band region with a
Fig. 4. Inactivation of HIV-1 PR by C1 at pH 5.5 and 308. The enzyme (7.5 nM) was incubated in 100 mM sodium acetate, 100 mM NaCl, 3% (v/v) Me2SO
with the fluorogenic substrate Dabcyl-g-Abu-S-Q-N-Y-P-I-V-Q-Edans (5.2 mM) in the presence of various concentrations of C1 (0.55±22 mM).

1868
F. Lebon et al. / Biochemical Pharmacology 63 (2002) 1863±1873
the HF-LCAO-MO-SCF formalism. The 3±21 g basis set
was used and all the computations were carried out using the
Gaussian 98 program. The molecular electrostatic potentials
(MEP) were calculatedattheHF/6±31 gà levelonthecrystal
conformation.
instead of acting at the catalytic site of the enzyme [30,31].
It was shown that a mutant protease where cysteine 67 and
95 residues were mutated to alanine was not inhibited by
cupric ions [30]. When tested, complex C1 was found to
behave as a time-dependent inhibitor of wild-type HIV-1
PR. The kinetic parameters for the inactivation process
were determined using the progress curve method in which
the substrate and the inhibitor compete (Fig. 4). At 308 and
pH 5.5, the inactivation ef®ciency index ki/KI was equal to
7.730 MÀ1 sÀ1, and ki and KI to 8:5 Â 10À3 sÀ1 and
1.1 mM, respectively. To discriminate between a putative
effect of the decomplexed cupric ion (known to be an
inactivator) from the metal-organic complexand the effect
of the complexitself (supposed to act as a reversible
inhibitor), the inhibitory ef®ciencies of compound C1
and CuCl2 were compared in a phosphate buffer using
equal concentrations of the two molecules at pH 6.0 and
6.5. The compounds were incubated for 30 s with the wild-
type enzyme before determination of the remaining activ-
ity. As shown in Fig. 5a, inhibition of HIV-1 PR by C1 and
CuCl2 led to two different sigmoid curves demonstrating
that the inhibition due to C1 was dual: part of the inhibition
could be attributed to free Cu2‡ (released from the com-
plex), the difference between the two curves corresponding
to the true effect of the metal-complex. An estimation of
the Ki values for the interaction between complex C1 and
HIV-1 PR, and CuCl2 and HIV-1 PR, was obtained by
3. Results and discussion
3.1. Enzyme inhibition
Ligands 1, 2, 7, 8, and 9 were synthesised and com-
plexed to copper(II). The resulting complexes were tested
for their ability to inhibit HIV-1 PR in a puri®ed enzyme
system. The inhibition obtained was compared to the effect
of cupric ions alone. In fact, it was previously described
that cupric chloride (CuCl2) can also act as an inhibitor of
HIV-1 PR by targeting the sulfhydryl group of cysteines
Fig. 5. Inhibition of HIV-PR (22.7 nM) by C1 (*, a), C2 (*, b) and
CuCl2 (&, a and b) at pH 6.5 and 308. The enzyme was added in 0.025 M
Na phosphate, 0.1 M NaCl, 3% Me2SO (v/v), containing the fluorogenic
substrate Dabcyl-g-Abu-S-Q-N-Y-P-I-V-Q-Edans (5.2 mM) and different
concentrations of C1 (0±2.2 mM, a), C2 (0±4.7 mM, b) and CuCl2 (0±
4.7 mM, a and b).
Fig. 6. EPR spectra at room temperature of (curve a) free copper(II),
(curve b) C2, (curve c) C7, (curve d) C8, and (curve e) C9 recorded in the
acetate buffer (pH 5.5).

F. Lebon et al. / Biochemical Pharmacology 63 (2002) 1863±1873
1869
Fig. 7. Comparison of the spectrum of free copper(II) (solid line) and complex C1 (dashed line) at room temperature. Both spectra were normalised to the
same integrated intensity. Concentration of C1: 1 mM, acetate buffer (pH 5.5).
plotting logꢀv=vi† as a function of 1 ‡ ‰IŠ/Ki for a compe-
titive type of inhibition (v and vi represent the initial rate in
the absence and in the presence of the inhibitor, respec-
tively with S0 ! Km). A similar inhibitory constant Ki was
upon30 sincubationwiththeenzyme. Thiscon®rmsthatthe
corresponding complexes are not stable enough in the assay
buffer to inhibit the enzyme. In these cases, the inhibitory
effect was fully attributed to free cupric ion decomplexed
from the metal-compounds. To con®rm the observed inhibi-
tion of HIV-1 PR, the ability of metal compound C1 to
inhibit a recombinant protease devoid of cysteine residue
(C67A, C95A)[18]wasexamined. We found that themutant
protease was inhibited (>20%) in the presence of 1.7±2 mM
of compound C1 whereas no inhibitory effect was observed
found for the metal-complexand CuCl (ꢁ1 mM). The
2
value obtained for CuCl2 con®rms the global KI value
determined by the progress curve method for the inactiva-
tion process. For compounds C2 (Fig. 5b), C7, C8, and C9,
the inhibition curves were superimposable to that obtained
for CuCl2 using the same concentration of reagents and
Fig. 8. Comparison of the EPR spectrum of complex C1 (solid line) in the acetate buffer and calculated EPR spectrum (dashed line). The latter was obtained
by subtracting the EPR signal of copper(II) to the EPR spectrum of complex C1 in the acetate buffer.

1870
F. Lebon et al. / Biochemical Pharmacology 63 (2002) 1863±1873
Fig. 9. Comparison of the EPR spectrum of complex C1 in methanol (solid line) and EPR spectrum of free copper in methanol (dashed line).
using the same concentrations of CuCl2. In contrast, com-
pounds C2, C7, C8, and C9 failed to inhibit the mutant
protease at 0.5±5 mM. These results are in agreement with
3.2. Electron paramagnetic resonance characterisation
An EPR analysis was performed in order to determine
whether the studied complexes remained stable in the
enzymatic assay buffer (acetate buffer, pH 5.5). The com-
parison of the spectrum of the free Cu(II) sample with the
È
the work of Karlstrom and Levine [30] who demonstrated
that only copper chelates (and not cupric ions) act as
inhibitors of the mutant protease.
Fig. 10. MEP at À30 kcal/mol calculated for ligands 1, 2, 8, and 9 (HF/6±31 gÃ). Carbon atoms are depicted in green, nitrogen and oxygen atoms in violet
and red respectively and hydrogen atoms are represented in white.

Fig. 11. Interaction between HIV-1 PR and compound C1 (carbon atoms in gray) as predicted by molecular mechanics simulations and minimisation. For clarity only residues 8, 25, and 50 in both monomers are
Ê
represented (carbon atoms in green). Nitrogen and oxygen atoms are represented in violet and red, respectively. Hydrogen bonds (A) are formed between C1 and residues Asp25, Asp125, Ile50, Arg8, and Arg108.

1872
F. Lebon et al. / Biochemical Pharmacology 63 (2002) 1863±1873
spectra of complexes C2, C7, C8, and C9 shows a similar
pro®le characterised by a single broad line centred at
around 3300 G (Fig. 6, curves a±e). This ®rst observation
seems to indicate that the metal ion was totally released in
the medium, thus leading us to assume the lack of stability
of the studied complexes in the enzymatic assay buffer. In
contrast to the previously studied complexes, the EPR
spectrum of C1, recorded under the same conditions,
exhibits a different shape pro®le as illustrated in Fig. 7.
This difference allows us to assume that a second com-
pound exists in solution in addition to the free copper(II)
component. In order to better understand the behaviour of
C1 in the conditions of the enzymatic assay buffer, a
comparison was made between the EPR spectrum of C1
found in the acetate buffer (Fig. 8, solid line) and the
calculated spectrum (Fig. 8, dashed line) resulting from the
subtraction of the free copper(II) from the total spectrum of
C1 in acetate buffer. The resulting spectrum is in agree-
ment with literature data in which similar Cu(II) complexes
exhibit two main lines located between 3000 and 3500 G
[32]. To con®rm this hypothesis, the EPR spectrum of
complex C1 was recorded in methanol and compared to
that of free copper(II) in the same solvent. Indeed, we
previously showed that C1 exists as a complex in MeOH
[9]. It appears that the spectrum of C1 exhibits a more
resolved structure in comparison to the spectrum obtained
in the acetate buffer. The spectrum of C1 in MeOH has two
main lines located around 3200 and 3350 G (Fig. 9, solid
line) vs. the main line for the free copper(II) (Fig. 9, dashed
line). This pro®le is mostly identical to the one obtained for
the previously calculated spectrum of C1 in the acetate
buffer (Fig. 8, dashed line). Therefore, one can state that
C1 exists in the enzymatic assay buffer but releases
copper(II) in a slow process. Although qualitative, the
EPR analysis suggests that compound 1 is the only ligand
that remains chelated to copper(II) under the conditions of
the enzymatic assay and could be considered as a more
stable ligand relative to the other ones studied.
of compound C1 [9] were used for docking into the active
site of HIV-1 PR (Fig. 3b). Molecular modelling studies
were performed on the complexformed between HIV-1 PR
and C1 using the metal adapted force®eld esff [28]. The
results suggest that C1 is able to co-ordinate the catalytic
water molecule which remains hydrogen bonded to the
essential active site residues Asp25 and Asp125. A solvent
water molecule is a candidate for the co-ordination of the
remaining apical position of the complexand interacts with
the ¯ap residue Ile50. In the most favourable orientation,
the pyridyl group occupies the S2/S02 pockets while the
trimethoxybenzyl group occupies the larger S1/S01 subsite
(nomenclature of Schechter and Berger [33]). The oxygen
atom of the methoxy in the ortho and/or meta position of
the two trimethoxyphenyl groups interacts with the gua-
nidinium group of Arg8 and Arg108 residues (Fig. 11).
4. Conclusion
The interaction pattern derived by molecular modelling
is consistent with a competitive inhibition involving two
types of interactions: (1) ``Metal-based''Ðthe copper ion
present in complex C1 is properly positioned to chelate the
structural catalytic water molecule that initiates the pro-
teolytic reaction. (2) ``Organic-based''Ðthe chelating
ligand is involved in several hydrogen bonds and van
der Waals contacts with the enzyme. It contributes to
interaction speci®city and to the global stabilisation of
the inhibitor within the enzyme binding cleft. Ongoing
crystallographic experiments will help con®rming the dual
aspect of the interaction of metal-organic complexes with
HIV-1 PR. Optimising the interaction of complexsetcez
with the hydrophobic subsites of HIV-1 PR active site led
to the design of a novel co-ordination compound C1, with a
binding af®nity increased by a factor 103. Based on the
molecular modelling of compound C1 into the active site
of HIV-1 PR, ligands that are better structural and chemical
complements of the HIV-1 PR binding site are being
designed. Complex C1 which possesses trimethoxybenzyl
groups to interact into the S1/S01 subsites is rather large for
HIV-1 PR active site. However, those methoxy groups are
partially responsible for the higher stability of C1 in the
assay buffer, compared to the other N2-pyridine-amide
ligands studied. Therefore, when designing the new
ligands, we have to keep in mind that the stability of
the ®nal complexin the assay buffer is crucial to the true
inhibition observed. One way to ensure stability is to
consider more chelating polydentate ligands. The proper
understanding of the unique structural paradigm of metal-
organic complexes should eventually lead to the develop-
ment of potent and highly selective inhibitors of HIV-1 PR.
The expertise and comprehension gained in the framework
of HIV-1 PR will prove useful for the speci®c inhibition of
enzymes containing sites likely to bind a metal ion [15].
While metal cations are not likely to be useful agents,
3.3. Molecular modelling
We investigated the reason for the difference in stability
between the C1, C2, C8, and C9 complexes by calculating
the MEP on the respective N2-pyridine-amide ligands 1, 2,
8, and 9 (Fig. 10). A large attractive area was observed on
the pyridyl nitrogen atom of compound 1 involved in the
N±Cu bond. In contrast, a much smaller attractive area was
present on the corresponding nitrogen atom in compounds
2, 8, and 9. In those compounds, the carbonyl linked at the
ortho position of the aromatic ring acts as a strong electron
attractor, withdrawing electron density from the nitrogen
involved in the N±Cu bond. Moreover, the attractive area
located on the carbonyl group of ligand 1 is reinforced by
the trimethoxyphenyl donor group. The enhanced stability
of the N±Cu and O±Cu bonds allows C1 to exist in the
assay buffer. The previously obtained crystal co-ordinates

F. Lebon et al. / Biochemical Pharmacology 63 (2002) 1863±1873
1873
[15] Katz BA, Clark JM, Finer-Moore JS, Jenkins TE, Johnson CR, Ross
MJ, Luong C, Moore WR, Stroud RM. Design of potent selective
zinc-mediated serine protease inhibitors. Nature 1998;391:608±12.
[16] Janc JW, Clark JM, Warne RL, Elrod KC, Katz BA, Moore WR. A
novel approach to serine protease inhibition: kinetic characterisation
of inhibitors whose potencies and selectivities are dramatically
enhanced by zinc(II). Biochemistry 2000;39:4792±800.
metal chelates such as C1 should be considered as promis-
ing lead compounds for the development of targeted drugs.
Acknowledgments
The authors wish to thank H.J. Schramm (MaxPlanck
Institut fuÈr Biochemie, Martinsried, Germany), D.A. Davis
(NationalInstitutes of Health, Bethesda, USA) forproviding
us with the wild-type protease and the recombinant protease
mutant C67A, C95A, respectively and X. Damoiseau for
helpfuldiscussions. ThisworkwassupportedbyBIOMED 2
Network (BMH4-CT96-0823). F. Lebon and M. Ledecq
thank the National Belgian Foundation for Scienti®c
Research(FNRS)andtheIndustryandAgriculture Research
Foundation (FRIA) for ®nancial support as well as IBM-
[17] DeCamp DL, Babe LM, Salto R, Lucich JL, Koo M-S, Kahl SB,
Craik CS. Specific inhibition of HIV-1 protease by boronated
porphyrins. J Med Chem 1992;35:3426±8.
[18] Davis DA, Branca AA, Pallenberg AJ, Marschner TM, Patt LM,
Chatlynne LG, Humphrey RW, Yarchoan R, Levine RL. Inhibition of
the human immunodeficiency virus-1 protease and human immuno-
deficiency virus-1 replication by bathocuproine disulfonic acid Cu1‡
.
Arch Biochem Biophys 1995;322:127±34.
Â
[19] Qasmi D, de Rosny E, Rene L, Badet B, Vergely I, Boggetto N,
Reboud-RavauxM. Synthesis of N-glyoxylyl peptides and their in
vitro evaluation as HIV-1 protease inhibitors. Bioorg Med Chem
1997;5:707±14.
Â
Belgium and the Facultes Universitaires Notre-Dame de la
[20] Benincori T, Brenna E, Sanniedo F. Studies on Wallach's imidazole
synthesis. J Chem Soc, Perkin Trans 1 1993;675±80.
Paix(FUNDP) for the use of the Namur Scienti®c Comput-
ing Facility. Z. Benatallah thanks ``Ensemble contre le
Sida'', SIDACTION for ®nancial support.
[21] Gracheva IN, Ioffina DI, Tochilkin AI, Gorkin VZ. Monoamine
oxidase based on 2-, 4-, and 8-substituted quinolines. Pharm Chem J
1991:160±5 [English translation].
[22] Piechaczek B-D. Studies on monoamine oxidase inhibitors. VII.
Derivatives of quinolinecarboxylic acids. Acta Pol Pharm 1966;9±14.
[23] Coppa F, Fontana F, Lazzarini E, Minisci F. A facile convenient and
selective homolytic carbamoylation of heteroaromatic bases. Hetero-
cycles 1993;2687±96.
References
[1] West ML, Fairlie D. Targeting HIV-1 protease: a test of drug-design
methodologies. TiPS 1995;16:67±75.
[24] Billich A, Hammerschmid F, Winkler G. Purification, assay and
kinetic features of HIV-1 protease. Biol Chem Hoppe-Seyler 1990;371:
265±72.
[2] Kempf DJ, Sham HL. HIV protease inhibitors. Curr Pharm Des
1996;2:225±46.
[3] Deeks SG, Smith M, Holodniy M, Kahn O. HIV-1 protease inhibitors.
A review for clinicians. J Am Med Assoc 1997;277:145±53.
[4] Lebon F, Ledecq M. Approaches to the design of effective HIV-1
protease inhibitors. Curr Med Chem 2000;7:455±77.
[5] Schramm HJ, de Rosny E, Reboud-RavauxM, BuÈttner J, Dick A,
Schramm W. Lipopeptides as dimerisation inhibitors. Biol Chem
1999;380:593±6.
[25] Doucet C, Pochet L, Thierry N, Pirotte B, Delarge J, Reboud-Ravaux
M. 6-Substituted 2-oxo-2H-1-benzopyran-3-carboxylic acid as a core
structure for specific inhibitors of human leukocyte elastase. J Med
Chem 1999;42(20):4161±71.
[26] Matayoshi ED, Wang GT, Krafft GA, Erickson J. Novel fluorogenic
substrates for assaying retroviral proteases by resonance energy
transfer. Science 1990;247:954±8.
[6] Bouras A, Boggetto N, Benatalah Z, de Rosny E, Sicsic S, Reboud-
RavauxM. Design, synthesis and evaluation of conformationaly
constrained tongs, new inhibitors of HIV-1 protease dimerisation. J
Med Chem 1999;42:957±62.
[27] Erickson J, Neidhart DJ, VanDrie J, Kempf DJ, Wang XC, Norbeck
DW, Plattner JJ, Rittenhouse JW, Turon M, Wideburg N, Kohlbrenner
WE, Simmer R, Helfrich R, Paul DA, Knigge M. Design, activity,
Ê
and 2.8 A crystal structure of a C2 symmetric inhibitor complexed to
[7] Flexner C. HIV-protease inhibitors. New Engl J Med 1998;338:1281±
92.
HIV-1 protease. Science 1990;249:527±33.
[28] Biosym/MSI, Discover user guide, vol. 1. San Diego, 1995.
[29] Holloway MK, Wai JM, Halgren TA, Fitzgerald PMD, Vacca JP,
Dorsey BD, Levin RB, Thompson WJ, Chen LJ, Desolms SJ, Gaffin
N, Ghosh AK, Giuliani EA, Graham SL, Guare JP, Hungate RW, Lyle
TA, Sanders WM, Tucker TJ, Wiggins M, Wiscount CM, Woltersdorf
OW, Young SD, Darke PL, Zugay JA. A priori prediction of activity
for HIV-1 protease inhibitors employing energy minimisation in the
active site. J Med Chem 1995;38:305±17.
[8] Molla A, Granneman GR, Sun E, Kempf DJ. Recent developments in
HIV protease inhibitor therapy. Antivir Res 1998;39:1±23.
[9] Lebon F, Ledecq M, Benatallah Z, Sicsic S, Lapouyade R, Kahn O,
GarcËon A, Reboud-RavauxM, Durant F. Synthesis and structural
analysis of copper(II) pyridineamide complexes as HIV-1 protease
inhibitors. J Chem Soc, Perkin Trans 2 1999;4:795±800.
[10] Lebon F, de Rosny E, Reboud-RavauxM, Durant F. De novo drug
design of a new copper chelate molecule acting as HIV-1 protease
inhibitor. Eur J Med Chem 1998;33:733±7.
È
[30] Karlstrom ARK, Levine RL. Copper inhibits the protease from
human immunodeficiency virus 1 by both cysteine-dependant and
cysteine-independent mechanisms. Proc Natl Acad Sci USA 1991;88:
5552±6.
[11] Sadler PJ, Inorganic chemistry and drug design. In: Advances in
inorganic chemistry, vol. 36. New York: Academic Press, 1991. p. 1±
49.
È
[31] Karlstrom ARK, Shames BD, Levine RL. Reactivity of cysteine
[12] Louie AY, Meade TJ. Metal complexes as enzyme inhibitors. Chem
Rev 1999;99:2711±34.
residues in the protease from human immunodeficiency virus:
identification of a surface-exposed region which affects enzyme
function. Arch Biochem Biophys 1993;304:163±9.
[13] Lippard SJ. In: Bertini I, Gray HB, Lippard SJ, Valentine JS, editors.
Bioinorganic chemistry. Mill Valley, CA: University Science, 1994.
p. 505±83.
[32] Nagane R, Koshigoe T, Chikira M, Long EC. The DNA-bound
orientation of Cu(II)-Xaa-Gly-His metallopeptides. J Inorg Biochem
2001;83:17±23.
[14] Berners-Price SJ, Sadler PJ. Co-ordination chemistry of metallo-
drugs: insights into biological speciation from NMR spectroscopy.
Coord Chem Rev 1996;151:1±40.
[33] Schechter I, Berger A. On the size of the active site in proteases I.
Papain Biochem Biophys Res Commun 1967;27:157±62.