Identification and characterization of novel benzil (diphenylethane-1,2- dione) analogues as inhibitors of mammalian carboxylesterases

Article abstract of DOI:10.1021/jm049011j

Carboxylesterases (CE) are ubiquitous enzymes responsible for the metabolism of xenobiotics. Because the structural and amino acid homology among esterases of different classes, the identification of selective inhibitors of these proteins has proved problematic. Using Telik's target-related affinity profiling (TRAP) technology, we have identified a class of compounds based on benzil (1,2-diphenylethane-1,2-dione) that are potent CE inhibitors, with K _i values in the low nanomolar range. Benzil and 30 analogues demonstrated selective inhibition of CEs, with no inhibitory activity toward human acetylcholinesterase or butyrylcholinesterase. Analysis of structurally related compounds indicated that the ethane-1,2-dione moiety was essential for enzyme inhibition and that potency was dependent on the presence of, and substitution within, the benzene ring. 3D-QSAR analyses of these benzil analogues for three different mammalian CEs demonstrated excellent correlations of observed versus predicted K_i(r² > 0.91), with cross-validation coefficients (q²) of 0.9. Overall, these results suggest that selective inhibitors of CEs with potential for use in clinical applications can be designed.

Full text of DOI:10.1021/jm049011j

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
2906
J. Med. Chem. 2005, 48, 2906-2915
Identification and Characterization of Novel Benzil (Diphenylethane-1,2-dione)
Analogues as Inhibitors of Mammalian Carboxylesterases
Randy M. Wadkins,† Janice L. Hyatt,‡ Xin Wei,‡ Kyoung Jin P. Yoon,‡ Monika Wierdl,‡ Carol C. Edwards,‡
Christopher L. Morton,‡ John C. Obenauer,§ Komath Damodaran,# Paul Beroza,# Mary K. Danks,‡ and
Philip M. Potter*,‡
Department of Chemistry and Biochemistry, University of Mississippi, University, Mississippi 38677, Department of Molecular
Pharmacology and Hartwell Center for Bioinformatics and Biotechnology, St. Jude Children’s Research Hospital,
Memphis, Tennessee 38105, and Telik Inc., Palo Alto, California 94304
Received December 3, 2004
Carboxylesterases (CE) are ubiquitous enzymes responsible for the metabolism of xenobiotics.
Because the structural and amino acid homology among esterases of different classes, the
identification of selective inhibitors of these proteins has proved problematic. Using Telik’s
target-related affinity profiling (TRAP) technology, we have identified a class of compounds
based on benzil (1,2-diphenylethane-1,2-dione) that are potent CE inhibitors, with Ki values
in the low nanomolar range. Benzil and 30 analogues demonstrated selective inhibition of CEs,
with no inhibitory activity toward human acetylcholinesterase or butyrylcholinesterase. Analysis
of structurally related compounds indicated that the ethane-1,2-dione moiety was essential
for enzyme inhibition and that potency was dependent on the presence of, and substitution
within, the benzene ring. 3D-QSAR analyses of these benzil analogues for three different
mammalian CEs demonstrated excellent correlations of observed versus predicted Ki (r2 > 0.91),
with cross-validation coefficients (q2) of 0.9. Overall, these results suggest that selective
inhibitors of CEs with potential for use in clinical applications can be designed.
Introduction
to SN-38 (7-ethyl-10-hydroxycamptothecin), whereas
hCE1 is approximately 100- to 1000-fold less efficient
at this reaction.9,10
The enzymatic detoxification of xenobiotics in both
prokaryotes and eukaryotes is achieved in part by CEs.1
These enzymes are ubiquitous and are expressed in
tissues such as the lung, liver, kidney, and intestine that
are exposed to chemicals or toxins. CEs metabolize
many ester-containing compounds including heroin,
cocaine, procaine, CPT-11 (irinotecan, 7-ethyl-10-[4-
(1-piperidino)-1-piperidino]carbonyloxycamptothecin),
capecitabine, and mivacurium (see refs 1-4 and
references therein). The enzymes can also act as suicide
substrates for organophosphate poisons, providing pro-
tection from agents such as sarin, soman, and tabun.5
Recently, the X-ray crystal structures of a rabbit CE
(rCE) and a human liver CE (hCE1) have been deter-
mined,6-8 and the proteins demonstrate significant
structural homology to other classes of esterases, par-
ticularly acetylcholinesterases (AcChE) and lipases.
Indeed, the rCE and hCE1 structures are highly ho-
mologous, with only ∼1.0 Å rmsd variation over 455
residues of the R-carbon trace,7,8 consistent with the
amino acid sequences that are greater than 81% identi-
cal. However, despite this apparent structural similar-
ity, significant differences in substrate specificity exist.
For example, rCE catalyzes the hydrolysis of CPT-11
Because of the high degree of similarity among the
esterases, reports detailing the identification of selective
inhibitors of CEs have been limited. Of those that have
been identified, the structures are typically based on
either organophosphate compounds such as bomin or
trifluoromethyl ketone (TFK) analogues as described by
Wheelock et al.11,12 It would be expected that both
classes of compounds would be readily hydrolyzed by
water. In fact, the TFK-containing compounds were
reported to be hygroscopic, thereby hampering efforts
to develop these compounds for use in clinical applica-
tions.
Our goal, therefore, was to identify nontoxic selective
CE inhibitors that might have applications in either the
prevention of drug toxicity (e.g., by inhibiting morphine
production from heroin) or increasing the half-life of
drugs that are inactivated by CEs. Such drugs include
flestolol, a â-adrenergic blocking agent, which has a
dramatically reduced half-life in vivo due to CE-medi-
ated hydrolysis.13 By administration of a specific inhibi-
tor in combination with the drug of choice, prolonged
plasma concentrations and hence greater efficacy might
be achieved. To identify such inhibitors, we applied
Telik’s lead technology TRAP (target-related affinity
profiling.14-16 Using this approach, we identified an
analogue of benzil that selectively inhibited a variety
of CEs. Following analysis of over 31 benzil analogues,
we have developed QSAR models for the inhibition of
rCE, hCE1, and a human intestinal CE (hiCE, hCE2).
* To whom correspondence should be addressed. Phone: 901-495-
3440. Fax: 901-521-1668. E-mail: phil.potter@stjude.org.
† University of Mississippi.
‡ Department of Molecular Pharmacology, St. Jude Children’s
Research Hospital.
§ Hartwell Center for Bioinformatics and Biotechnology, St. Jude
Children’s Research Hospital.
# Telik Inc.
10.1021/jm049011j CCC: $30.25 © 2005 American Chemical Society
Published on Web 03/29/2005

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Benzil Analogues as Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 8 2907
to 100 µM, and data were fitted to the following equation to
These analogues do not inhibit human acetylcholinest-
erase (hAcChE) or butyrylcholinesterase (hBuChE) but
are potent inhibitors of mammalian CEs, with Ki values
in the low nanomolar range.
determine the mode of enzyme inhibition:22
[I]{[s](1 - â) + Ks(R - â)}
i )
[I]{[s] + RKs} + Ki{R[s] + RKs}
Experimental Section
where i is the fractional inhibition, [I] is the inhibitor
concentration, [s] is the substrate concentration, R is the
change in affinity of substrate for enzyme, â is the change in
the rate of enzyme substrate complex decomposition, Ks is the
dissociation constant for the enzyme substrate complex, and
Ki is the inhibitor constant. Examination of the curve fits
where R ranged from 0 to ∞ and â ranged from 0 to 1 was
performed using GraphPad Prism software and Perl data
language. The curves generating the highest r2 values were
analyzed using Akaike’s information criteria23,24 to identify the
best model for enzyme inhibition. Ki values were then calcu-
lated from the equation predicted by Prism to be the best fit
for the experimental data.
Enzyme Inhibition Assays Using 4-Methylumbellifer-
one Acetate (4-MUA) as a Substrate. Metabolism of 4-MUA
was assessed by a fluorimetric assay using either a Hitachi
F2000 fluorescence spectrophotometer or a Millipore Cytofluor
2350 with an excitation wavelength of 365 nm and an emission
wavelength of 460 nm. Samples were incubated with 0.75 mM
of 4-MUA in 50 mM Hepes, pH7.4, and the release of
4-methylumbelliferone was measured over a 1 min time
interval. The rate of substrate catalysis was linear over this
time period. Enzyme inhibition by selected compounds and
determination of Ki values were performed as described above.
Inhibition of Acetylcholinesterase and Butyrylcho-
linesterase. The ability of compounds to inhibit hAcChE and
hBuChE was performed as previously described using either
1 mM acetylthiocholine (AcTCh) or 1 mM butyrylthiocholine
(BuTCh), respectively, as substrates.25-28
Chemicals. Compounds were purchased from the following
sources. Benzil (1), 3-7, 14, 19, 20, 22-26, 28, 31, 36, 38-
40, 53, 57-59, and 61 were all purchased from Sigma Aldrich
Biochemicals (St. Louis, MO). 2, 18, 21, 48, and 49 were
obtained from Pfaltz and Bauer (Waterbury, CT). 8, 11, 15,
43-47, 50-52, and 56 were purchased from Alfa Aesar (Ward
Hill, MA). 10, 41, and 60 were obtained from VWR Scientific
Products (West Chester, PA). 16 was purchased from Ryan
Scientific (Isle of Palms, SC). 17 was obtained from Toronto
Research Chemicals (Toronto, Canada). 27, 32, and 35 were
purchased from Maybridge (Tintagel, U.K.). 29, 30, 33, and
34 were obtained from Industrial Research Ltd., (Auckland,
New Zealand). 37 and 54 were purchased from TCI America
(Portland, Oregon), and 42 and 55 were obtained from ACROS
(Fisher Scientific, Pittsburgh, PA). The synthesis of compounds
9, 12, and 13 is described in this article.
CPT-11 was kindly provided by Dr. J. P. McGovren (Pfizer,
Cambridge, MA).
Synthetic Reagents. 4-Chlorobenzaldehyde, 3,5-difluoro-
benzaldehyde, 3,4,5-trifluorobenzaldehyde, thiamine, copper
acetate, and ammonium nitrate were all purchased from
Sigma Aldrich.
Enzymes. Pure rCE and hCE1 were prepared as described
previously.17 hiCE was prepared by concentration of baculo-
virus media from Sf9 cells expressing a secreted form of the
protein. While not homogeneous, the preparation was at least
60% pure. Since no CE activity is expressed or secreted from
uninfected Sf9 cells, the only CE present in the culture media
was the recombinant protein. The Genbank accession numbers
for the cDNAs used for CE production were as follows: rCE,
AF036930;18 hCE1, M73499;19 hiCE, Y09616.20
Synthesis of Benzil Analogues. Benzil analogues were
synthesized by condensation of the substituted benzaldehydes
(0.04 mol) in the presence of thiamine hydrochloride (0.002
mol) and alcoholic sodium hydroxide (0.005 mol) to yield the
benzoin,29 with subsequent oxidation using copper acetate
(0.001 mol) and ammonium nitrate (0.006 mol) in 80% acetic
acid.30 Products were purified by recrystallization, melting
points were determined using a Mel-temp (Barnstead Inter-
national, Dubuque, IA), and purity and structures were
assessed by TLC, NMR, and total C, H, N analysis.
Human AcChE and BuChE were obtained from Sigma
Aldrich.
Library of Compounds and TRAP Screening. Screening
of compounds was performed using a library of small molecules
obtained from Telik using their target-related affinity profiling
(TRAP) technology.15 Briefly, TRAP defines an “affinity finger-
print” for a particular molecule by determining its binding
affinities to a panel of proteins. In an iterative screening
process (∼70 compounds for each iteration), computational
analysis of fingerprints is performed and compared to the
biological activity obtained for each compound. The first series
of compounds to be assayed was chosen because their finger-
prints encompassed the diversity of fingerprints present within
the entire library. Subsequent analyses are based on compu-
tational analysis of the fingerprints of the compounds that
inhibited CEs. Routinely three to four iterations were per-
formed using approximately 300 compounds. The advantage
of this methodology is that molecules of multiple chemotypes
demonstrating inhibitory activity can be identified by assaying
as few as 200 compounds.14,15
Enzyme Inhibition Assays and Determination of Ki
Values Using Nitrophenyl Acetate (o-NPA) as a Sub-
strate. CE inhibition was determined using a spectrophoto-
metric multiwell plate assay with 3mM o-NPA as a sub-
strate.18,21 Briefly, test compound (100 µM) and substrate were
aliquoted into wells and enzyme was added using a multiwell
pipettor. The rate of change in absorbance at 420 nm was
measured at 15 s intervals for 5 min and compared to wells
containing no inhibitor. Compounds that demonstrated 50%
reduction in CE activity were subsequently evaluated in detail.
Routinely, assays were performed in duplicate and included
both positive (50 µM bis(4-nitrophenyl) phosphate (BNPP)) and
negative controls (DMSO, no enzyme).
4,4′-Dichlorobenzil (1,2-Bis(4-chlorophenyl)ethane-1,2-
dione, 9). 4,4′-Dichlorobenzil was synthesized from 1,2-bis-
(4-chlorophenyl)-2-hydroxyethanone to yield a yellow solid
(yield 62%). Physical and NMR parameters of 9 were as
1
follows: mp 196-198 °C; H NMR (500 MHz, CDCl3) δ 7.53
(d, 2H, J ) 8.5 Hz, Ar-H), 7.94 (d, 2H, J ) 8.5 Hz, Ar-H).
1,2-Bis(3,5-difluorophenyl)ethane-1,2-dione (12). 1,2-
Bis(3,5-difluorophenyl)ethane-1,2-dione was synthesized from
1,2-bis(3,5-difluorophenyl)-2-hydroxyethanone to yield a yellow
solid (yield 74%). Physical and NMR parameters of 12 were
as follows: mp 135-137 °C; 1H NMR (500 MHz, CDCl3) δ 7.15
(dd, 2H, J ) 2.4 Hz, Ar-H), 7.51 (dt, 4H, J ) 2.2 Hz, Ar-H).
1,2-Bis(3,4,5-trifluorophenyl)ethane-1,2-dione (13). 1,2-
Bis(3,4,5-trifluorophenyl)ethane-1,2-dione was synthesized from
2-hydroxy-1,2-bis(3,4,5-trifluorophenyl)ethanone to yield a yel-
low solid (yield 96%). Physical and NMR parameters of 13 were
as follows: mp 102-103 °C; 1H NMR (500 MHz, CDCl3) δ 7.68
(d, 4H, J ) 6.6 Hz, Ar-H).
Irreversible Enzyme Inhibition Assays. Irreversible
inhibition of CEs by the inhibitors was assessed by preincu-
bating the enzymes with the compounds at 1 µM on ice for
either 1 or 24 h. After incubation, the samples were extensively
diluted in 50 mM Hepes, pH7.4, and CE activity was deter-
mined using 3mM o-NPA as a substrate. As a positive control,
50 µM BNPP was included in these assays.
Molecular Modeling of Benzil-Based Carboxylesterase
Inhibitors. 3D-QSAR analyses were performed using Quasar
4.031,32 running on a Macintosh G4. Structures for each
analogue were initially constructed with ChemBats3D for
Further analysis of inhibitors was performed in a similar
fashion except that inhibitor concentrations ranged from 1 pM

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
2908 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 8
Wadkins et al.
Table 1. Ki Values for the Inhibition of CEs by Benzil and Related Compoundsa
a Values for CEs were determined using o-NPA as a substrate, and those for hAcChE were determined using AcTCh as a substrate.
ogy.14,15,37,38 Briefly, 70-80 compounds were screened
Macintosh. Partial atomic charges from the bond charge
correction method33 and AMBER atom types were assigned
for inhibition of o-NPA metabolism by each CE, and the
using the antechamber module of AMBER7 (University of
initial set of compounds was chosen from the library
California, San Francisco, CA). Structures for QSAR analysis
because they represented the diversity of the entire
were performed by initial energy minimization of benzil using
the AM1 Hamiltonian of MOPAC.34 The final structure was
in excellent agreement with the structure obtained from higher
level molecular orbital calculations and experimental data.35
All benzil derivatives were then modeled from this minimized
structure and individually minimized to remove steric conflicts.
The Quasar program produces a 3D receptor-surface model
containing information about the molecular properties of both
the ligand and the docking site. Essentially the software
constructs pseudoreceptor models by positioning points of
charge at vertexes around the pharmacophore under examina-
tion. These vertexes had charges of +0.1e, +0.5e, 0.0e, -0.1e,
or -0.5e. Initially, a population of 200 models was generated
for each enzyme inhibition data set and evaluated via a genetic
algorithm to produce approximately 7000 new pseudoreceptor
models. Evaluation of these models was then performed until
the cross-validated correlation coefficient (q2) reached g0.9 for
predicted vs experimental Ki values for each CE. Under these
conditions, correlation coefficients (r2) for the predicted vs
experimental Ki data were linear (>0.91) for each enzyme
model.
compound collection. The screening process is itera-
tive: the initial set of compounds to screen is chosen
for the diversity of its general protein binding charac-
teristics; subsequent sets are chosen on the basis of
computational analysis of hits identified in earlier
iterations.14 This process allowed the identification of
bona fide potent CE inhibitors with as few as 230
assays. These data indicated that benzil was a potent
inhibitor of hiCE, hCE1, and rCE but demonstrated no
inhibition of hAcChE.
Having identified benzil (see Table 1) as a general
inhibitor of mammalian CEs, we assessed the ability of
analogues having similar structures to this compound
to inhibit catalysis of o-NPA by CEs. Table 1 indicates
the Ki values for benzil and a series of structurally
related agents. Replacement of either or both of the
carbonyl chemotypes with hydroxyl moieties resulted in
loss of enzyme inhibition. Similarly, removal of one
carbonyl group as in benzophenone (4) or increasing the
spacing between the diones as in 1,3-diphenylpropane-
1,3-dione (5) resulted in compounds that were unable
to inhibit CE-mediated catalysis of o-NPA. Finally,
replacement of the benzene rings with methyl groups
increased the Ki values of the inhibitors from 47- to
6800-fold. For example, with hCE1 the Ki values for
benzil (1), 1-phenylpropane-1,2-dione (2), and butan-2,3-
Quantitation of CPT-11 and SN-38. CPT-11 and SN-38
concentrations were determined using an HPLC-based assay
with a fluorescent detector as previously described.9,36
Results
Identification of Benzil as a General Carboxyl-
esterases Inhibitor. The identification of the benzil
chemotype as a general inhibitor of CEs was the result
of a directed screen using Telik’s TRAP technol-

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Benzil Analogues as Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 8 2909
Table 2. Structure of Benzil-Based Carboxylesterase Inhibitorsa
compd
name
R1 R2
R3
R4
R5 R6
R7
R8
R9
R10
1
8
9
benzil (diphenylethane-1,2-dione)
4,4′-difluorobenzil (1,2-bis(4-fluorophenyl)ethane-1,2-dione)
4,4′-dichlorobenzil (1,2-bis(4-chlorophenyl)ethane-1,2-dione)
F
Cl
F
Cl
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
2,2′-dichlorobenzil (1,2-bis(2-chlorophenyl)ethane-1,2-dione) Cl
4,4′-dibromobenzil (1,2-bis(4-bromophenyl)ethane-1,2-dione)
1,2-bis(3,5-difluorophenyl)ethane-1,2-dione
1,2-bis(3,4,5-trifluorophenyl)ethane-1,2-dione
4-[oxo(phenyl)acetyl]benzoic acid
1-(4-chlorophenyl)-2-phenylethane-1,2-dione
1-(4-chlorophenyl)-2-(4-methylphenyl)ethane-1,2-dione
1-(4-methylphenyl)-2-phenylethane-1,2-dione
1,2-bis(4-methylphenyl)ethane-1,2-dione
1-(3,4-dimethylphenyl)-2-phenylethane-1,2-dione
1-(4-methoxyphenyl)-2-phenylethane-1,2-dione
1,2-bis(4-methoxyphenyl)ethane-1,2-dione
Cl
Br
F
Br
F
F
F
F
F
F
F
F
F
COOH
Cl
Cl
CH3
CH3
CH3
OCH3
OCH3
CH3
CH3
CH3
OCH3
Br
1,2-bis(3-methoxyphenyl)ethane-1,2-dione
OCH3
NO2
OCH3
1-(2-chlorophenyl)-2-(3,4-dimethoxyphenyl)ethane-1,2-dione Cl
1-[4-(bromomethyl)phenyl]-2-phenylethane-1,2-dione
1,2-bis(4-bromo-3-nitrophenyl)ethane-1,2-dione
1-(4-methyl-3-nitrophenyl)-2-phenylethane-1,2-dione
1-(4-nitrophenyl)-2-phenylethane-1,2-dione
1-(2,4-dinitrophenyl)-2-phenylethane-1,2-dione
1,2-bis(3-nitrophenyl)ethane-1,2-dione
OCH3 OCH3
CH2Br
NO2
NO2
Br
CH3
NO2
NO2
NO2
NO2
NO2
1,2-bis(4-hydroxyphenyl)ethane-1,2-dione
OH
OH
NO2 OH
NO2 OCH3
OH
1,2-bis(5-bromo-2-hydroxyphenyl)ethane-1,2-dione
1,2-bis(2,4-dihydroxyphenyl)ethane-1,2-dione
1,2-bis(4-hydroxy-3-nitrophenyl)ethane-1,2-dione
1,2-bis(4-methoxy-3-nitrophenyl)ethane-1,2-dione
1,2-bis[4(dimethylamino)phenyl]ethane-1,2-dione
OH
OH
Br
Cl
OH
OH
Br
F
OH
OH
OCH3
(CH3)2N
F
NO2
NO2
(CH3)2N
Cl
1-(pentachlorophenyl)-2-(pentafluorophenyl)ethane-1,2-dione Cl Cl
1-{4-[oxo(phenyl)acetyl]phenyl}-2-phenylethane-1,2-dione
Cl
F
F
F
COCOC6H5
a R ) H unless otherwise indicated.
dione (3) were 45, 5270, and greater than 100 000 nM,
respectively. Overall, these results suggest that the 1,2-
dione structure is essential for enzyme inhibition and
that increased potency toward CEs is provided by the
phenyl rings.
Analysis of CE Inhibition by Benzil Analogues.
The TRAP screen identified the key structural compo-
nents that were required for CE inhibition (e.g., 1,2-
dione, aromatic moieties). We then tested commercially
available agents having these characteristics (see Tables
2, 3, 4, and 5) for their ability to inhibit the CEs, hiCE,
hCE1, and rCE, as well as hAcChE and hBuChE. As
indicated in Table 4, a wide range of Ki values for the
inhibition of CE-mediated catalysis of o-NPA was
observed for various benzil analogues. None of these
agents inhibited hAcChE or hBuChE at concentrations
up to 100 µM.
In general, the inhibitors fell into two distinct
classes: those that inhibited all three CEs (e.g., benzil
(1), 15, and 37) and those that only inhibited hiCE and
rCE (e.g., compounds 21, 28, and 31). A general cor-
relation between the size of the inhibitor and the ability
to inhibit CEs was observed such that the larger
compounds inhibited both rCE and hiCE and the
smallest molecules inhibited all three CEs.
Assessment of the Mode of CE Inhibition by
Benzil. Analysis of the curve fits obtained using the
equation described by Webb22 suggested that the mode
of enzyme inhibition for benzil and its analogues was
partially competitive. Hence, while benzil and its ana-
logues acted in a competitive fashion, at infinite con-
centrations of inhibitor the reaction velocity was not
reduced to zero. This observation suggests that the
compounds act as inhibitors because they demonstrate
similarity to bona fide enzyme substrates. To determine
whether benzil and several selected analogues could
produce irreversible inhibition of CEs, we determined
the ability of the enzymes to metabolize o-NPA after
preincubation with the inhibitors. As indicated in Figure
1, no irreversible inhibition was observed with benzil
or any of the analogues tested. In contrast, greater than
95% inhibition was observed using BNPP, an inhibitor
known to irreversibly bind to and inactivate esterases.
QSAR Analysis of Benzil-Mediated CE Inhibi-
tion. 3D-QSAR analysis of the inhibition data was
performed using Quasar 4.0 (Figure 2), and Table 6
indicates the correlation coefficients (r2 values) for the
experimental versus predicted Ki values. As indicated,
these coefficients are close to unity (0.919-0.925),
suggesting that the models provide an accurate estima-
tion of the inhibition constants. In addition, the q2
values (the cross-correlation coefficients) for all of the
models were 0.90. Since q2 correlations of greater than
0.4 are considered adequate for QSAR modeling of
biological data,39 these results indicate that these
models should be suitable for future inhibitor design.
Graphic representations of the 3D-QSAR models for
hiCE, hCE1, and rCE are depicted in Figure 3. Benzil