Ligand-based design of a potent and selective inhibitor of cytochrome P450 2C19

Article abstract of DOI:10.1021/jm201346g

A series of omeprazole-based analogues was synthesized and assessed for inhibitory activity against CYP2C19. The data was used to build a CYP2C19 inhibition pharmacophore model for the series. The model was employed to design additional analogues with inh

Full text of DOI:10.1021/jm201346g

Article
pubs.acs.org/jmc
Ligand-Based Design of a Potent and Selective Inhibitor of
Cytochrome P450 2C19
Robert S. Foti,† Dan A. Rock,† Xiaogang Han,‡ Robert A. Flowers, II,§ Larry C. Wienkers,†
and Jan L. Wahlstrom*,†
†Pharmacokinetics and Drug Metabolism, Amgen Inc., Seattle, Washington 98119, United States
‡Pharmacokinetics and Drug Metabolism Biomarkers, Pfizer, Inc., Groton, Connecticut 06340, United States
§Department of Chemistry, Lehigh University, Bethlehem, Pennsylvania 18015, United States
ABSTRACT: A series of omeprazole-based analogues was synthesized and
assessed for inhibitory activity against CYP2C19. The data was used to build
a CYP2C19 inhibition pharmacophore model for the series. The model was
employed to design additional analogues with inhibitory potency against
CYP2C19. Upon identifying inhibitors of CYP2C19, ligand-based design
shifted to attenuating the rapid clearance observed for many of the inhibitors.
While most analogues underwent metabolism on their aliphatic side chain,
metabolite identification indicated that for analogues such as compound
30 which contain a heterocycle adjacent to the sulfur moiety, metabolism
primarily occurred on the benzimidazole moiety. Compound 30 exhibited
improved metabolic stability (Clint = 12.4 mL/min/nmol) and was selective
in regard to inhibition of CYP2C19-catalyzed (S)-mephenytoin hydrox-
ylation in human liver microsomes. Finally, representative compounds were
docked into a homology model of CYP2C19 in an effort to understand the enzyme−ligand interactions that may lead to
favorable inhibition or metabolism properties.
playing an important role in the observed CYP2C19 activities and
drug interaction profiles.11 Drugs that are metabolized in vivo by
CYP2C19 include anticonvulsants such as phenytoin, psychotropic
drugs such as imipramine and diazepam, and the proton pump
inhibitors omeprazole, lansoprazole, and rabeprazole.12
INTRODUCTION
■
The cytochromes P450 (P450) are a superfamily of heme-
containing proteins that are capable of catalyzing the oxidative
metabolism of a wide range of substrates. Common metabolic
pathways that occur through P450 catalysis include hydroxylation,
heteroatom oxidation or dealkylation, and aromatic oxidation.1
Nearly 60 human P450s have been identified, with the major
isoforms involved in drug metabolism being CYP1A2, CYP2A6,
CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1,
CYP3A4, and CYP3A5.2 In addition to being involved in the
metabolism of a wide array of drugs, P450 enzymes can also be the
target of clinically relevant drug interactions or can form
bioreactive metabolites that can have toxicological consequences.3
Multiple examples exist within the literature describing the
synthesis of potent and selective P450 inhibitors in order to
invoke a pharmacological response or as diagnostic tools to assess
the role of a given P450 isoform in a test compound’s metabolism
or drug interaction potential. Recent medicinal chemistry efforts
have identified potent and selective inhibitors of CYP1B1,
CYP2A6, and CYP26 through a combination of traditional
synthesis and computational tools.4−6
Proton pump inhibitors such as omeprazole block the release
of gastric acid through inhibition of the H+/K+-ATPase pump in
gastric parietal cells.13 They are widely used to treat gastresophageal
reflux disease (GERD) and Zollinger−Ellison syndrome.14,15 The
substituted benzimidazole sulfoxides are metabolized by a number
of P450s, including CYP2C19, CYP3A4, and CYP2D6.16,17 The
various proton pump inhibitors have gained attention from a drug
metabolism standpoint because of their potential to cause drug
interactions both in vitro and in vivo. In human liver microsomes,
omeprazole and lansoprazole have Ki values of approximately 3 μM
against CYP2C19.18 In vivo, omeprazole has been shown to have
drug interactions with diazepam19,20 and phenprocoumon.21 The
combination of the CYP2C19 metabolism and inhibition profiles of
proton pump inhibitors such as omeprazole makes the class of
compounds a useful tool to study both the substrate selectivity and
active site characteristics of CYP2C19.
An inherent difficulty in optimizing the properties of P450
ligands is the simultaneous interplay of catalysis and inhibition.
Unlike the development of a series of inhibitors against a
therapeutic target where target potency and metabolic stability
CYP2C19 is a polymorphic enzyme that is localized in the
liver (approximately 5% of hepatic P450) and intestine (less
than 3% of intestinal P450).7,8 The enzyme is inducible by
rifampicin, phenobarbital, and dexamethasone.9,10 The typical
reactions used to monitor CYP2C19 activity include (S)-
mephenytoin 4′-hydroxylation, omeprazole 5-hydroxylation,
and fluoxetine O-demethylation with the choice of substrate
Received: October 7, 2011
Published: January 12, 2012
© 2012 American Chemical Society
1205
dx.doi.org/10.1021/jm201346g | J. Med. Chem. 2012, 55, 1205−1214

Journal of Medicinal Chemistry
Article
are assessed in discrete in vitro systems, compounds designed
to inhibit a given P450 enzyme may be subject to metabolism
by the same enzyme. In addition, the relationship between inhi-
bition and metabolism may be series dependent. For example, a
positive correlation between intrinsic clearance and inhibition
potency has been noted for a series of quinoline-4-carboxamide
compounds with CYP3A4, potentially due to their ability to
type II bind to the here iron.22,23 Conversely, for a series of
HIV protease inhibitors, compounds with a lower intrinsic
clearance were able to more potently inhibit CYP3A4.24 Few
examples exist of both catalytic and inhibitory models applied
to the same set of compounds, most likely due to the complex-
ities added by metabolism by multiple enzymes or at different
sites on a molecule.25
Table 1. Omeprazole-Like Inhibitors of CYP2C19
The aim of this study was to design and synthesize a potent
and selective inhibitor of CYP2C19 based on a series of
omeprazole-based analogues. An initial set of 26 diverse
structural omeprazole analogues was synthesized and assessed
for inhibition of CYP2C19 (Ki). The data was subsequently
used to populate pharmacophore-based models. Information
gained from the pharmacophore models was used to direct the
synthesis of additional analogues within the series. Upon
identification of potent inhibitors of CYP2C19, the aim of the
study shifted toward improving metabolic stability within the
series. To this end, compounds were assessed for rates and sites
of metabolism by CYP2C19. Both ligand- and active-site
modeling were incorporated in order to more fully understand
the chemical properties that confer both metabolism by and
inhibition of CYP2C19. Finally, the interplay of metabolism
and inhibition was examined in order to identify a highly potent
and metabolically stable inhibitor of CYP2C19.
RESULTS
■
Assessment of CYP2C19 Inhibition. A series of omepra-
zole analogues was synthesized as depicted in Figure 1. The
initial set of 26 omeprazole-based analogues was assessed
for inhibition of CYP2C19-catalyzed 3-O-methylfluorescein
metabolism. Prior to conducting the inhibition assay, a Km
was determined for the formation of fluorescein from 3-O-
methylfluorescein in recombinant CYP2C19 (1.18 0.06 μM)
using reaction conditions that were previously determined to
be linear with respect to time and protein concentration.
IC50 values were obtained for each compound and used to
determine the range of inhibitor concentrations for the Ki assay.
Following the 5 min incubation, Ki values ranged from 0.04 μM
(compound 19) to >50 μM (compound 11 and 20; Table 2).
A feature-based pharmacophore model was generated using
the inhibition data obtained from the Ki assay. The data was fit
to a three, four, five, or six feature model. In all cases, the
hydrophobic regions (H) of the compound set represented a
key feature of the model. In addition to hydrophobicity, the
aromatic rings (R) of the benzimidazole moiety were also taken
into account (Figure 2). For the three feature (HHR) and four
feature (HHRR) pharmacophore models, the best fit was
obtained when using three partial-least squared factors. When
generating the pharmacophore models with five (ADHHR) or
six (DHHHRR) features, the hydrogen bond accepting (A) and
hydrogen bond donating (D) characteristics of the benzimida-
zole nitrogens were taken into account. The best overall fit
from the pharmacophore models was obtained from the five
feature ADHHR model (r2 = 0.958, q2 = 0.607, RMSE =
0.603). The features for this model corresponded to the
hydrogen bond accepting (A) and donating (D) features of the
Figure 1. General synthetic route for the omeprazole-based analogues.
benzimidazole nitrogens, the hydrophobicity of the side chains off
of the sulfur atom (H), the side chain off of the benzimidazole
moiety (H), and the aromaticity of the benzimidazole rings (R).
Following pharmacophore analysis, the series of omeprazole
analogues was expanded to include a pyridine moiety linked to
the mercapto-benzimidazole group at the R2 position with
varying side-chains at R1. Upon synthesis of the heterocycle-
containing compounds (compounds 27−32), an increase in the
observed potency was noted (compound 30, Ki = 0.02 μM).
To assess the selectivity of the omeprazole-based analogues
for CYP2C19, inhibition assays were repeated in human liver
microsomes with selective probe substrates. Similar inhibition
parameters (r2 = 0.86) were observed for compounds that
were active in both the recombinant system utilizing 3-O-
methylfluorescein as the probe substrate or human liver micro-
somes which utilized (S)-mephenytoin as the probe substrate.
Two of the most potent CYP2C19 inhibitors, compounds
15 and 30, were 37- and 66-fold selective, respectively, for
CYP2C19 in human liver microsomes as compared to the next
most potently inhibited P450 isoform (CYP1A2 for both
1206
dx.doi.org/10.1021/jm201346g | J. Med. Chem. 2012, 55, 1205−1214

Journal of Medicinal Chemistry
Article
Assessment of CYP2C19 Metabolism. Intrinsic clearance
of each of the analogues was determined in both recombinant
CYP2C19 and human liver microsomes. In general, the longer
chain aliphatic analogues and substituted phenyl analogues
were cleared the most rapidly (Tables 2 and 3). Hydroxyl and
carboxylic acid analogues demonstrated the greatest degree of
stability in both systems. Heterocyclic analogues were mod-
erately stable in both systems (rCYP2C19 Clint, 0.43−12.3 mL/
min/nmol; HLM Clint, 57.8−579.9 μL/min/mg).
Table 2. LogD(pH=7.4), Ki, and Intrinsic Clearance Values for
Omeprazole-Based Analogues
CYP2C19 Ki
(μM)
CYP2C19 Clint
(mL/min/nmol P450)
compd
LogD (pH = 7.4)
1
2.92
3.45
3.98
4.51
5.05
5.58
3.8
42.5
4.79
0.75
0.06
0.07
0.05
0.06
1.00
1.94
0.14
50.0
40.8
1.10
0.13
0.04
0.20
0.10
0.07
0.04
50.0
0.07
0.07
0.06
0.85
0.05
0.31
2.43
10.9
0.05
0.02
3.20
9.90
0.44
2.77
2
3
8.00
4
23.39
29.13
24.07
8.61
5
A feature-based pharmacophore model was also generated
using the intrinsic clearance values for the series of compounds
(Figure 3). For the intrinsic clearance pharmacophore model, a
minimum of five features (AADHR) with three partial least
squared factors predicted intrinsic clearance values within an
acceptable range (r2 = 0.857; q2 = 0.603; RMSE = 0.378). Both
r2 and q2 values decreased significantly when the model was
generated using only one or two partial least squared factors.
For the metabolism pharmacophore model, the five features
corresponded to the hydrogen bond accepting properties of
one of the benzimidazole nitrogens (A) as well as the oxygen
atom of the benzimidazole methoxy moiety (A), the hydrogen
bond donating properties of the second benzimidazole nitrogen
(D), the hydrophobicity of the side chains off of the sulfur atom
(H), and the aromaticity of the benzimidazole rings (R).
Correlation of CYP2C19 Inhibitory Potency and
Metabolic Stability With Physicochemical Properties.
Physical properties for each of the 32 analogues were calculated
using ChemDraw Ultra 9.0 (CambridgeSoft, Cambridge, MA)
or a proprietary data analysis prediction software package.
Inhibition potency values showed the best correlation with
hydrophobic descriptors such as LogD7.4. Plotting inhibition
and intrinsic clearance values versus LogD7.4 gave r2 values of
0.78 and 0.64, respectively. The most potent inhibition values
were observed for compounds with a LogD7.4 greater than 4. In
general, the most potent inhibitors were also the most rapidly
cleared in recombinant CYP2C19 (Figure 4). Plotting Ki values
determined in rCYP2C19 against intrinsic clearance values in
the same system resulted in an r2 value of 0.87.
6
7
8
4.15
3.29
4.98
1.98
−0.34
4.17
4.63
5.51
5.86
4.76
4.94
5.2
5.57
9
2.19
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
22.55
<0.43
<0.43
<0.43
24.80
20.98
28.53
26.08
25.66
19.67
0.44
0.95
3.88
5.92
5.4
20.66
11.62
34.51
<0.43
19.02
15.96
1.69
4.03
6.16
5.74
2.67
3.21
2.76
3.22
2.85
1.47
1.55
6.02
12.29
2.39
<0.43
Metabolite Identification. To determine the primary sites
of metabolism for the omeprazole-based series, qualitative
metabolite identification experiments were conducted on
representative aliphatic/phenyl and heterocycle compounds.
For the compounds that had an aliphatic side chain or phenyl
moiety, the primary site of metabolism was on the hydrophobic
side chain adjacent to the sulfur of the 2-mercapto-
5-methoxybenzimidazole core structure. For compounds 3, 5,
and 6, monohydroxylation on the alkyl chain was the major
route of metabolism (Figure 5). Fragmentation patterns from
product ion scans indicated an increase of 16 amu to each of
the alkyl chains, although the fragmentation was not specific
enough to assign the hydroxyl group to a specific carbon on the
chain. Synthesis of metabolite standards further indicated that
metabolism was primarily occurring at the ω-1 carbon of the
alkyl chain (data not shown). The primary metabolites of
compound 7 (isopropyl) and compound 8 (t-butyl) were
determined to be hydroxylation of one of the terminal carbons
of the alkyl chain. Compounds 10 (cyclohexyl), 14 (p-tolyl),
and 15 (p-cumyl) all showed a monohydroxylation on the
respective cyclic ring as the major metabolite in recombinant
CYP2C19. As with the alkyl side chains, the mass spectral
fragmentation patterns did not allow for the specific site of
metabolism on the cyclohexyl or aromatic rings to be
elucidated. Minor routes of metabolism included hydroxylation
Figure 2. CYP2C19 inhibition pharmacophore model for the
omeprazole-based analogues. The features for this five-feature model
corresponded to the hydrogen bond accepting (blue) and donating
(pink) features of the benzimidazole nitrogens, the hydrophobicity of
the side chains off of the sulfur atom and the benzimidazole moiety
(green), and the aromaticity of the benzimidazole rings (orange).
compounds). In general, inhibitor selectivity for CYP2C19
decreased as less potent CYP2C19 inhibitors were evaluated
(Table 3).
1207
dx.doi.org/10.1021/jm201346g | J. Med. Chem. 2012, 55, 1205−1214

Journal of Medicinal Chemistry
Article
Table 3. IC50 and Intrinsic Clearance Values for Omeprazole-Based Analogues in Human Liver Microsomes
IC50, μM (HLM)
HLM
Clint, app
selectivity
compd
1A2
2B6
2C8
2C9
50
2C19
2D6
2E1
3A4mdz
T1/2
1
2.34
0.93
0.71
0.44
0.39
0.79
1.02
49.9
0.95
2.67
8.08
50
5.1
50.0
50.0
10.5
15.4
14.8
50.0
16.6
50.0
21.4
5.5
6.97
3.12
33.7
9.64
2.42
50
50
4.93
1.94
2.63
19.4
2.61
5.33
11.6
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
14.0
9.6
493.6
724.1
823.0
1852.9
1843.1
2665.4
650.1
116.9
538.0
744.4
185.3
<57.8
<57.8
1141.7
1194.8
420.5
1560.8
1669.9
80.9
2
35.8
39.9
50
27.1
50
3
50
8.4
4
9.18
50
26.3
5.33
15.7
50
3.7
5
2.92
4.57
6.25
0.21
50
3.8
6
2.6
3.8
1.2
CYP1A2
CYP2B6
7
31.9
50
32.7
10.7
59.3
12.9
9.3
8
21.2
22.1
14.1
50
50
9
50
29.4
50
16.4
6.23
5.49
50.0 >
0.08 >
8.86
8.34
12.3
4.86
4.23
4.91
50.0
0.72
1.87
3.11
2.81
2.49
3.56
4.01
0.4
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
5.11
5.65
50
4.41
5.91
50
50
4.17
50
37.4
120
120
6.1
28.8
50.0
17.7
14.4
50.0
3.6
50.0
50.0
7.8
3.7
0.04
7.2
7.1
0.78
0.12
50.0
50
ND
1.39
1.48
50
50
9.73
0.12
4.28
2.05
1.36
1.23
0.09
0.41
50
3.32
50
6.73
3.42
2.54
5.98
3.25
5.15
5.8
37.0
3.4
CYP1A2
CYP2D6
0.97
1.44
1.87
3.87
50
16.5
4.4
0.95
0.81
1.81
50
9.6
4.2
4.4
85.7
8.06
12.8
1.8
50.0
19.9
1.1
50.0
859.8
543.5
3893.3
608.4
1408.5
1353.5
424.4
261.2
343.1
<57.8
221.4
430.4
579.9
10.69
0.67
1.34
0.73
0.63
>50
8.56
50
3.35
0.25
0.51
3.49
50
2.32
0.82
2.27
6.18
0.96
0.88
0.55
9.12
0.02
2.73
10.5
1.5
2.2
50
12.6
15.5
2.0
11.4
4.92
5.12
16.3
26.53
20.2
120
31.3
16.1
11.95
50
0.07
50
9.0
CYP1A2
CYP1A2
50.0
50.0
9.9
0.68
50
0.12
13.15
50
50
50
24.2
7.6
50
12.94
3.29
50
24.0
50.0
21.8
8.4
4.78
50
50
8.79
50
3.75 >
3.61
10.7
0.52
50
0.05
28.5
50
65.8
50
22.3
50
50
0.99
6.97
2.17
Figure 3. CYP2C19 metabolism pharmacophore model for the
omeprazole-based analogues. The five features for the metabolism
pharmacophore model corresponded to the hydrogen bond accepting
properties of a benzimidazole nitrogen and the oxygen atom of the
benzimidazole methoxy moiety (pink), the hydrogen bond donating
properties of the second benzimidazole nitrogen (blue), the
hydrophobicity of the side chains off of the sulfur atom (green), and
the aromaticity of the benzimidazole rings (orange).
Figure 4. Three-dimensional correlation between inhibition potency (Ki),
intrinsic clearance (Clint), and LogD7.4. Calculated r2 values were
0.78 (Ki versus LogD7.4), 0.64 (Clint versus LogD7.4), and 0.87 (Ki versus
Clint).
of the benzimidazole moiety and O-demethylation at the
5-methoxy position.
When heterocyclic pyridine analogues attached to either a
2-mercapto-5-methoxybenzimidazole or 2-mercapto-5-methyl-
1208
dx.doi.org/10.1021/jm201346g | J. Med. Chem. 2012, 55, 1205−1214

Journal of Medicinal Chemistry
Article
Figure 5. Identification of the primary sites of metabolism for this
series of compounds. Compounds with a methoxy group at the
5-position of the benzimidazole ring and an aliphatic chain off of the
sulfur were oxidatively metabolized at the ω-1 carbon of the aliphatic
side chain, while compounds lacking the methoxy group and
containing a heterocyclic ring off of the sulfur atom were primarily
metabolized on the substituent at the 5-position of the benzimidazole
ring.
Figure 6. Active-site docking of compounds 6 (green) and 30
(orange) into a homology model of CYP2C19 (I-helix shown in red).
Primary sites of metabolism were located approximately 3 Å from the
heme iron, with the 5-methoxy group of compound 6 and the pyridine
nitrogen of compound 30 occupying a similar space in the active site
within approximately 2 Å of Ala477.
benzimidazole core were assessed for metabolite formation in
recombinant CYP2C19, the major routes of metabolism were
observed to be a combination of hydroxylation on the
benzimidazole moiety and O-demethylation (compound 27)
or hydroxylation on the 5-methyl substituent of the
benzimidazole core (compound 30). Only minimal amounts
of sulfoxide or sulfone formation were observed for any of the
compounds noted above.
CYP2C19 Active Site Docking. To evaluate the active site
docking poses for CYP2C19 ligand interactions, a homology
model of CYP2C19 was built and compounds 6 and 30 docked
in the active site. Compound 6 docked in the active site of
CYP2C19 with the heptyl side chain oriented toward the heme.
The ω-1 carbon was located approximately 3.11 Å from the
heme iron. Compound 30 docked in the opposite orientation,
with the 5-methyl substituent of the benzimidazole core
oriented toward the heme at a distance of 3.07 Å. The pyridine
nitrogen of compound 30 occupied a similar space in the active
site of CYP2C19 as the 5-methoxy moiety of compound 6. In
addition, both compounds were located within 2.00 Å of the
CYP2C19 active site residues Val113, Val208, Glu300, Leu366,
and Ala477. Additional interactions (within 3.00 Å) included
Phe114, Ile205, Ser209, Thr301, Thr304, Ile362, Phe476, and
Val479 (Figure 6).
for (+)-N-3-benzylnirvanol.26,29 Thus, the primary aim of this
study was achieved in that the current set of compounds equals
or surpasses these known inhibitors with respect to inhibitory
potency against CYP2C19, with 14 of 32 compounds having
Ki values equal to or less than 100 nM. Compound 30, the
2-mercapto-5-methylbenzimidazole pyridine analogue, was the
most potent compound synthesized, with a Ki value of 20 nM.
While the range of inhibition values for the omeprazole-based
analogues is comparable to what has been observed for the
benzylnirvanol, benzylphenobarbital, and benzbromarone data
sets, a striking difference between the compounds is the meta-
bolic lability of the omeprazole-based compounds as compared
to the other chemotypes.
While a significant amount of attention has been placed on
using reaction conditions that avoid the depletion of the probe
substrate in enzyme activity assays,30 the stability of the test
compound being assessed for inhibition is often neglected,
except under conditions where inhibitory metabolites are sus-
pected. For example, N-desmethyldiltiazem and N,N-didesme-
thyldiltiazem are both more potent inhibitors of CYP3A4 than
diltiazem.31 Conversely, O-desmethyldiltiazem was shown to
have no inhibitory potential against CYP3A4. The compounds
in this data set appear to be rapidly converted by CYP2C19 to
metabolites with a decreased inhibitory affinity for the enzyme.
This was observed by a decrease in the observed Ki values when
enzyme activity was evaluated after a 5 min coincubation of
probe substrate and inhibitor as compared to a 20 min coin-
cubation (data not shown). In addition, compounds such as 11,
12, and 20 that contained either hydroxyl or carboxylic acid
moieties showed no inhibitory potential against CYP2C19,
suggesting a detrimental effect of polar groups on the affinity
of the compounds for CYP2C19 and corroborating the
importance of the hydrophobicity incorporated into both
pharmacophore models. In light of this finding, it is quite likely
that the hydroxylated metabolites of this series of analogues
also have a decreased inhibitory potential for CYP2C19. Thus,
for those compounds which were rapidly metabolized, cal-
culating a Ki in human liver microsomes on a time scale
consistent with the ability to measure an adequate amount of
DISCUSSION
■
Computational modeling of P450−ligand interactions has
gained an increased amount of attention in recent years.
Within this area of research, CYP2C19 has been the focus of a
number of computational models aimed at predicting
metabolism, structure−activity relationships, and the drug
interaction potential for a given series of compounds. Examples
include a series of benzbromarone-based analogues that was
used to assess the effects of substrate charge and functional
groups on affinity and metabolism,26 the use of a GRID/CPCA
(consensus principal component analysis) approach to compare
the selectivity of CYP2C19 to CYP2C8, CYP2C9, and
CYP2C18,27 and a homology model/active site docking study
to evaluate the metabolism of the sulfonylurea compound
gliclazide.28
The most potent CYP2C19 inhibitors identified to date
include (+)-N-3-benzylnirvanol, (−)-N-3-benzylphenobarbital,
and 3,5″-dimethyl-4″-hydroxybenzbromarone, with Ki values rang-
ing from 33 nM for the benzbromarone analogue to 250 nM
1209
dx.doi.org/10.1021/jm201346g | J. Med. Chem. 2012, 55, 1205−1214

Journal of Medicinal Chemistry
Article
(S)-mephenytoin hydroxylation may be difficult, as the actual
concentration of inhibitor in the incubation will be much
less than the original concentration added. Furthermore, if
the actual amount of inhibitor is depleted to a point where
it is equal to or less than the concentration of enzyme in
the incubation, a percentage of the inhibitor will exist in
equilibrium as the enzyme−inhibitor complex. The resulting
deviation from starting conditions can result in an incorrect
assessment of the inhibitor’s potency.32 A similar under-
estimation of the observed Ki or Kd has been reported for assays
carried out using small volumes with an excess amount of
enzyme.33 In the previous case, however, the depletion of
inhibitory ligand was caused by binding to a receptor that is in
excess compared to the ligand.
When the metabolism and inhibition data were probed for
physicochemical trends, the greatest correlation for both sets
was versus the LogD7.4 values for the compound set. Other
descriptors of lipophilicity that showed a good correlation with
the metabolism and inhibition values were cLogP, LogD4.0,
LogD6.5 as well as the number of hydrogen bond donors. It has
been previously reported that lipophilicity is an important
factor in determining the potency of a set of CYP2C19
inhibitors.34 More specifically, the lipophilic properties at the
N-3 position of a series of phenytoin, nirvanol, or phenobarbital
analogues were shown to drive the inhibitory potency of the
compounds against both CYP2C9 and CYP2C19.35
To explore the structure−activity relationships for a given set
of compounds, pharmacophore and docking/homology model-
ing are often utilized. The premise behind pharmacophore
modeling is the superimposition of structural analogues with
structural features including hydrogen bond acceptors or
donors and π-stacking aromatic groups that may interact with
the enzyme active site. Pharmacophore models can then be
more fully incorporated as a useful descriptor in quantitative
structure−activity modeling.36−38 While less prevalent than
inhibition modeling, recent attempts have also been made to
model the metabolic clearance of compounds by P450
enzymes.39 The development of both inhibition and metabo-
lism pharmacophore models was achieved for the current set of
compounds. We have successfully developed both an inhibition
(r2 = 0.958) as well as a metabolism pharmacophore model
(r2 = 0.857) for the omeprazole-based series of compounds. As
would be expected from the correlation of our data with
lipophilic properties, both models took into account the aroma-
ticity of the benzimidazole rings as well as multiple hydrophobic
binding interactions. Interestingly, the metabolism model
appeared to be more dependent on the hydrogen bond donating
and accepting properties of the compounds than the inhibition
model was, possibly indicating more rigid criteria for a com-
pound to be a substrate of CYP2C19 as compared to an inhibitor
of CYP2C19. In addition, substitution off of the sulfur atom
appeared to affect both the metabolism and inhibition properties
of the compounds in a manner more amenable to structure−
activity relationships, whereas substitution at the 5-position of
the benzimidazole moiety appeared to exert more influence on
the observed site of metabolism.
compounds 6 and 30, as representatives of analogues containing
an aliphatic moiety or heterocycle off of the sulfur linker.
Consistent with the observed primary sites of metabolism on each
molecule, the ω-1 position of compound 6 and the 5-methyl
group of compound 30 were nearly superimposed approximately
3 Å from the heme iron. As both the physicochemical properties
and the pharmacophore models indicated the importance of
lipophilicity in predicting both the metabolism and inhibition for
these compounds, it is not surprising that the closest amino acid
residues to both compounds all have the potential for hydrophobic
interactions. The 5-methoxy moiety of compound 6 and the
pyridine nitrogen of compound 30, which also occupied similar
positions in the active site of CYP2C19, were both located within
2 Å of Ala477, indicating a potentially important role of hydrogen
binding interactions with this residue in the positioning of these
analogues in a catalytically favorable position within the active site
of CYP2C19. Previous studies have also implicated Phe114,
Ile362, and Leu 366 in the binding of hydrophobic molecules
within the CYP2C19 active site.35
CONCLUSION
■
In summary, a highly potent set of CYP2C19 inhibitors that
are also metabolized by CYP2C19 have been synthesized and
used for pharmacophore and homology modeling studies with
CYP2C19. Compound 30 represents the most potent
CYP2C19 inhibitor currently reported in the literature.
Assessing the metabolism of the compound set indicates that
the primary site of metabolism is on the hydrophobic side chain
off of the sulfur atom for those compounds with an aliphatic or
phenyl group off of the sulfur linker and on the benzimidazole
portion of the molecule for those compounds with a hetero-
cycle adjacent to the sulfur linker. Finally, the data points to the
dynamic interplay of binding affinity and metabolic stability in
properly defining a Ki value for enzyme inhibitors that may also
be metabolized by the same enzyme that they inhibit.
EXPERIMENTAL SECTION
■
Materials. Recombinant CYP2C19 Supersomes and 3-O-methyl-
fluorescein were purchased from Becton Dickinson (San Jose, CA).
NADPH was obtained from EMD Chemicals (Gibbstown, NJ).
Fluorescein, magnesium sulfate, and all fine chemicals were purchased
from Sigma-Aldrich (St. Louis, MO). All solvents were from J. T.
Baker (Phillipsburg, NJ). Deuterated DMSO was purchased from
Cambridge Isotope Laboratories (Andover, MA).
Synthesis. The appropriate mercapto-benzimidazole analogue
(unsubstituted, methoxy, ethoxy, amino, methyl, or difluoromethoxy;
1.1 equiv) was dissolved in 15 mL of methanol with 3 mL of 2N
sodium hydroxide and heated to 70 °C (Figure 1). Upon complete
dissolution, 1 equiv of a bromo- or chloro-alkylating agent was added
to the reaction and stirred at 70 °C for 12 h, at which point the
reactions were checked for completion using thin layer chromatog-
raphy and subsequently evaporated to dryness under a stream of
nitrogen. Reactions were reconstituted in 2.5% aqueous sodium
hydroxide and extracted twice with methylene chloride. The organic
layers were combined, dried over magnesium sulfate, and evaporated
to dryness prior to LC-MS/MS and NMR characterization. Preparative
thin layer chromatography was carried out as warranted to ensure the
purity of the final products (Table 1). All final products were ≥95%
pure as determined by HPLC/UV/mass spectrometry.
As P450 crystal structures become more readily available, the
use of docking and homology modeling has also increased and
been used to predict and explore select enzyme−ligand
interactions.40−44 As the crystal structure of CYP2C19 has
not been solved yet, a homology model based on the CYP2C9
crystal structure was used for active site docking studies. Two
compounds were docked into the active site of CYP2C19,
NMR Analysis. Compounds were characterized using a 500 MHz
Varian NMR spectrometer (Varian, Palo Alto, CA) using DMSO-d6 as
the solvent. Additional confirmation was obtained using electrospray
ionization mass spectrometry (ESI-MS) on a ThermoFisher LTQ ion
trap mass spectrometer (ThermoFisher Scientific, Waltham, MA).
1210
dx.doi.org/10.1021/jm201346g | J. Med. Chem. 2012, 55, 1205−1214

Journal of Medicinal Chemistry
Article
Compound 1 (115.4 mg, 47.4% yield). δ 1.34 (t, J = 7.8, 3H), 3.23
(q, J = 5.8, 2H), 3.76 (s, 3H), 6.73 (d, J = 9.8, 1H), 6.97 (d, J = 7.8,
1H), 7.30 (dd, J = 5.8, 7.8, 1H), 12.34 (s, 1H). ESI-MS (M + H)+
found, 209.0615; calculated for C10H12N2OS, 209.0670.
Compound 2 (207.5 mg, 70.5% yield). δ 0.99 (t, J = 7.8, 3H), 1.70
(t, J = 7.8, 2H), 3.20 (m, J = 7.8, 2H) 3.76 (s, 3H), 6.72 (d, J = 7.8,
1H), 6.96 (d, J = 7.8, 1H), 7.31 (dd, J = 5.8, 7.8, 1H), 12.34 (s, 1H).
ESI-MS (M + H)+ found, 223.0800; calculated for C11H14N2OS,
223.0827.
Compound 3 (184.1 mg, 70.4% yield). δ 0.90 (t, J = 7.8, 3H), 1.42
(m, J = 7.8, 2H), 1.67 (m, J = 7.8, 2H), 3.23 (t, J = 5.8, 2H), 3.76
(s, 3H), 6.72 (d, J = 7.8, 1H), 6.96 (d, J = 7.8, 1H), 7.32 (dd, J = 5.8,
7.8, 1H), 12.34 (s, 1H). ESI-MS (M + H)+ found, 237.0984; calculated
for C12H16N2OS, 237.0983.
7.30 (d, J = 9.8, 1H), 12.33 (s, 1H). ESI-MS (M + H)+ found,
313.1234; calculated for C18H20N2OS, 313.1296.
Compound 16 (275.6 mg, 64.1% yield). δ 1.33 (s, 9H), 3.76
(s, 3H), 4.19 (s, 2H), 6.72 (d, J = 9.8, 1H), 6.94 (s, 1H), 6.98 (m, J =
5.8, 2H), 7.19 (d, J = 5.8, 2H), 7.30 (d, J = 9.8, 1H), 12.33 (s, 1H). ESI-
MS (M + H)+ found, 327.1401; calculated for C19H22N2OS, 327.1453.
Compound 17 (156.3 mg, 42.9% yield). δ 3.76 (s, 3H), 4.51
(s, 2H), 6.76 (d, J = 5.8, 1H), 6.88 (m, J = 5.8, 1H), 7.10 (m, J = 5.8,
1H), 7.36 (d, J = 7.8, 2H), 7.45 (d, J = 9.8, 2H), 12.43 (s, 1H). ESI-MS
(M + H)+ found, 305.0476; calculated for C15H13ClN2OS, 305.0437.
Compound 18 (210.3 mg, 46.6% yield). δ 3.76 (s, 3H), 4.51
(s, 2H), 6.74 (d, J = 5.8, 1H), 6.86 (m, J = 5.8, 1H), 7.12 (m, J = 5.8,
1H), 7.38 (d, J = 7.8, 2H), 7.47 (d, J = 9.8, 2H), 12.34 (s, 1H). ESI-MS
(M + H)+ found, 349.0092; calculated for C15H13BrN2OS, 348.9932.
Compound 19 (365.9 mg, 78.5% yield). δ 3.76 (s, 3H), 4.47
(s, 2H), 6.65 (d, J = 11.7, 1H), 6.95 (m, J = 7.8, 1H), 7.23 (d, J = 7.8,
2H), 7.33 (m, J = 7.8, 1H), 7.66 (d, J = 7.8, 2H), 12.43 (s, 1H). ESI-
MS (M + H)+ found, 396.9782; calculated for C15H13IN2OS,
396.9793.
Compound 4 (255.7 mg, 77.2% yield). δ 0.86 (t, J = 7.8, 3H), 1.31
(m, J = 7.8, 2H), 1.38 (m, J = 7.8, 2H), 1.68 (m, J = 7.8, 2H), 3.22
(t, J = 7.8, 2H), 3.76 (s, 3H), 6.72 (d, J = 5.8, 1H), 6.94 (s, 1H), 7.30
(d, J = 9.8, 1H), 12.34 (s, 1H). ESI-MS (M + H)+ found, 251.1196;
calculated for C13H18N2OS, 251.1140.
Compound 5 (259.8 mg, 78.3% yield). δ 0.86 (t, J = 7.3, 3H), 1.27
(m, J = 4.9, 4H), 1.40 (m, J = 7.3, 2H), 1.67 (m, J = 7.3, 2H), 3.22
(t, J = 7.3, 2H), 3.76 (s, 3H), 6.73 (d, J = 7.3, 1H), 6.94 (s, 1H), 7.30
(d, J = 9.8, 1H), 12.34 (s, 1H). ESI-MS (M + H)+ found, 265.1283;
calculated for C14H20N2OS, 265.1296.
Compound 6 (221.7 mg, 79.1% yield). δ 0.85 (m, J = 12.2, 3H),
1.25 (m, J = 12.2, 7H), 1.39 (m, J = 7.3, 2H), 1.68 (m, J = 7.3, 2H),
3.22 (t, J = 7.3, 2H), 3.75 (s, 3H), 6.72 (d, J = 7.3, 1H), 6.94 (s, 1H),
7.31 (d, J = 7.3, 1H), 12.34 (s, 1H). ESI-MS (M + H)+ found,
279.1466; calculated for C15H22N2OS, 279.1453.
Compound 7 (126.1 mg, 47.3% yield). δ 1.00 (d, J = 7.8, 6H), 1.93
(m, J = 5.8, 1H), 3.15 (d, J = 3.9, 2H), 3.76 (s, 3H), 6.73 (d, J = 7.8,
1H), 6.92 (d, J = 5.8, 1H), 7.30 (dd, J = 7.8, 5.8, 1H), 12.50 (s, 1H).
ESI-MS (M + H)+ found, 237.0911; calculated for C12H16N2OS,
237.0983.
Compound 8 (17.1 mg, 5.7% yield). δ 1.06 (s, 9H), 2.85 (s, 2H),
3.76 (s, 3H), 6.72 (d, J = 7.3, 1H), 6.94 (s, 1H), 7.31 (d, J = 7.3, 1H),
12.50 (s, 1H). ESI-MS (M + H)+ found, 251.1127; calculated for
C13H18N2OS, 251.1140.
Compound 9 (194.2 mg, 74.4% yield). δ 0.31 (t, J = 5.8, 2H), 0.53
(t, J = 7.8, 2H), 1.18 (m, J = 3.9, 1H), 3.51 (d, J = 5.8, 2H), 3.76
(s, 3H), 6.73 (d, J = 7.8, 1H), 6.92 (d, J = 7.8, 1H), 7.30 (dd, J = 7.8,
5.8, 1H), 12.50 (s, 1H). ESI-MS (M + H)+ found, 235.0868; calculated
for C12H14N2OS, 235.0827.
Compound 10 (54.7 mg, 16.4% yield). δ 1.01 (m, J = 11.7, 4H),
1.15 (m, J = 11.7, 2H), 1.60 (d, J = 7.8, 4H), 1.67 (m, J = 11.7, 1H),
1.83 (d, J = 11.7, 2H), 3.75 (s, 3H), 6.72 (d, J = 9.8, 1H), 6.94 (s, 1H),
7.30 (d, J = 9.8, 1H), 12.33 (s, 1H). ESI-MS (M + H)+ found,
277.1288; calculated for C15H20N2OS, 277.1296.
Compound 20 (36.4 mg, 12.5% yield). δ 3.87 (s, 3H), 4.19
(s, 2H), 6.65 (d, J = 11.7, 1H), 6.95 (m, J = 7.8, 1H), 7.23 (d, J = 7.8,
2H), 7.33 (m, J = 7.8, 1H), 7.66 (d, J = 7.8, 2H), 12.58 (s, 1H), 12.78
(s, 1H). ESI-MS (M + H)+ found, 315.0752; calculated for
C16H14N2O3S, 315.0725.
Compound 21 (230.7 mg, 59.5% yield). δ 3.63 (s, 6H), 3.76
(s, 3H), 4.19 (s, 2H), 6.13 (s, 2H), 6.18 (s, 1H) 6.72 (d, J = 9.8, 1H),
6.94 (s, 1H), 7.30 (d, J = 9.8, 1H), 12.34 (s, 1H). ESI-MS (M + H)+
found, 331.0888; calculated for C17H18N2O3S, 331.1038.
Compound 22 (296.2 mg, 70.8% yield). δ 3.75 (s, 3H), 4.45
(s, 2H), 6.73 (d, J = 5.8, 1H), 6.93 (m, J = 5.8, 1H), 7.25 (d, J = 5.8,
1H), 7.30 (m, J = 9.8, 1H), 7.37 (m, J = 5.8, 2H), 7.42 (m, J = 5.8, 2H),
7.46 (m, J = 5.8, 2H), 7.60 (d, J = 5.8, 1H), 12.36 (s, 1H). ESI-MS
(M + H)+ found, 347.0996; calculated for C21H18N2OS, 347.1140.
Compound 23 (194.1 mg, 57.0% yield). δ 3.76 (s, 3H), 4.50
(s, 2H), 6.94 (d, J = 7.8, 1H), 7.12 (s, 1H), 7.18 (d, 1H), 7.55 (m, J =
7.8, 6H), 8.02 (d, J = 5.8, 1H), 12.33 (s, 1H). ESI-MS (M + H)+
found, 321.0957; calculated for C19H16N2OS, 321.0983.
Compound 24 (175.4 mg, 67.2% yield). δ 3.76 (s, 3H), 4.56
(s, 2H), 6.72 (d, J = 9.8, 1H), 6.94 (s, 1H), 7.30 (m, J = 9.8, 2H), 7.54
(t, J = 7.8, 1H), 7.73 (t, J = 7.8, 1H), 7.94 (m, J = 5.8, 2H), 8.02 (d, J =
5.8, 1H), 12.33 (s, 1H). ESI-MS (M + H)+ found, 322.1072; calculated
for C18H15N3OS, 322.0936.
Compound 25 (428.6 mg, 91.0% yield). δ 3.76 (s, 3H), 4.50 (d, J =
11.7, 2H), 6.74 (d, J = 5.8, 1H), 6.87 (d, J = 7.8, 1H), 6.95 (d, J = 7.8,
2H), 7.07 (d, J = 9.8, 1H), 7.12 (t, J = 7.7, 1H), 7.19 (d, J = 5.8, 1H),
7.23 (d, J = 7.8, 1H), 7.31 (m, J = 7.8, 1H), 7.38 (d, J = 7.8, 3H), 12.34
(s, 1H). ESI-MS (M + H)+ found, 363.1063; calculated for
C21H18N2O2S, 363.1089.
Compound 11 (28.0 mg, 8.8% yield). δ 1.03 (t, J = 5.8, 2H), 3.27
(m, J = 5.8, 2H), 3.51 (t, J = 5.8, 2H), 3.76 (s, 3H), 4.69 (t, J = 5.8,
1H), 6.72 (d, J = 9.8, 1H), 6.94 (s, 1H), 7.30 (d, J = 9.8, 1H), 12.33
(s, 1H). ESI-MS (M + H)+ found, 239.0786; calculated for C11H14-
N2O2S, 239.0776.
Compound 12 (32.1 mg, 10.2% yield). δ 1.03 (t, J = 5.86, 2H),
3.27 (m, J = 5.8, 2H), 3.51 (t, J = 5.8, 2H), 3.76 (s, 3H), 6.72 (d, J =
9.8, 1H), 6.94 (s, 1H), 7.30 (d, J = 9.8, 1H), 10.94 (s, 1H), 12.33
(s, 1H). ESI-MS (M + H)+ found, 267.0824; calculated for C12H14-
N2O3S, 267.0725.
Compound 26 (403.1 mg, 80.5% yield). δ 3.76 (s, 3H), 4.29
(s, 2H), 5.14 (s, 2H), 6.65 (d, J = 7.8, 2H), 6.87 (d, J = 7.8, 2H), 6.72
(d, J = 9.8, 1H), 6.94 (s, 1H), 7.30 (d, J = 9.8, 1H), 7.37 (t, J = 7.8,
1H), 7.41 (t, J = 5.8, 2H), 7.48 (d, J = 5.8, 2H), 12.43 (s, 1H). ESI-MS
(M + H)+ found, 376.9986; calculated for C22H20N2O2S, 377.1245.
Compound 27 (114.2 mg, 57.6% yield). δ 3.76 (s, 3H), 4.51
(s, 2H), 6.93 (d, J = 7.8, 1H), 7.12 (s, 1H), 7.15 (t, J = 7.8, 1H), 7.30
(d, J = 9.8, 1H), 7.56 (d, J = 7.8, 1H), 7.74 (t, J = 7.8, 1H), 8.67 (d, J =
7.8, 1H), 12.43 (s, 1H). ESI-MS (M + H)+ found, 272.0748; calculated
for C14H13N3OS, 272.0779.
Compound 13 (115.2 mg, 55.3% yield). δ 3.76 (s, 3H), 4.19
(s, 2H), 6.93 (d, J = 7.8, 1H), 7.10 (d, J = 7.8, 2H), 7.20 (s, 1H), 7.30
(m, J = 7.1, 2H), 7.51 (d, J = 7.8, 1H), 7.80 (m, J=7.8, 1H), 12.33
(s, 1H). ESI-MS (M + H)+ found, 271.0865; calculated for C15H14N2-
OS, 271.0827.
Compound 14 (291.7 mg, 85.3% yield). δ 2.25 (s, 3H), 3.76
(s, 3H), 4.47 (s, 2H), 6.74 (d, J = 5.8, 1H), 6.96 (s, 1H), 7.11 (d, J =
7.8, 2H), 7.30 (m, J = 5.8, 3H), 12.34 (s, 1H). ESI-MS (M + H)+
found, 285.0949; calculated for C16H16N2OS, 285.0983.
Compound 15 (340.8 mg, 85.8% yield). δ 1.19 (d, J = 5.8, 6H),
2.86 (m, J = 5.8, 1H), 3.76 (s, 3H), 4.19 (s, 2H), 6.72 (d, J = 9.8, 1H),
6.94 (s, 1H), 6.98 (m, J = 5.8, 2H), 7.19 (d, J = 5.8, 2H),
Compound 28 (132.2 mg, 64.2% yield). δ 1.33 (t, J = 7.8, 3H),
4.01 (m, J = 7.8, 2H), 4.63 (d, J = 11.7, 2H), 6.74 (t, J = 7.8, 1H), 6.87
(s, 1H), 7.27 (m, J = 3.8, 1H), 7.39 (d, J = 7.8, 1H), 7.51 (t, J = 7.8,
1H), 7.78 (t, J = 7.8, 1H), 8.51 (d, J = 3.9, 1H), 12.47 (s, 1H). ESI-MS
(M + H)+ found, 286.0978; calculated for C15H15N3OS, 286.0936.
Compound 29 (146.3 mg, 66.4% yield). δ 4.52 (s, 2H), 7.15 (t, J =
9.8, 1H), 7.21 (t, J = 7.8, 2H), 7.27 (d, J = 9.8, 1H), 7.45 (d, J = 7.8,
2H), 7.74 (t, J = 7.8, 1H), 8.67 (d, J = 7.8, 1H), 12.43 (s, 1H). ESI-MS
(M + H)+ found, 242.0643; calculated for C13H11N3S, 242.0674.
Compound 30 (283.9 mg, 86.7% yield). δ 2.38 (s, 3H), 4.65
(s, 2H), 6.91 (s, 1H), 6.96 (d, J = 8.8, 1H), 7.29 (m, J = 5.8, 2H),
7.51 (d, J = 8.8, 1H), 7.75 (t, J = 8.8, 1H), 8.52 (d, J = 5.8, 1H),
1211
dx.doi.org/10.1021/jm201346g | J. Med. Chem. 2012, 55, 1205−1214

Journal of Medicinal Chemistry
Article
12.33 (s, 1H). ESI-MS (M + H)+ found, 256.0803; calculated for
C14H13N3S, 256.0830.
Compound 31 (292.3 mg, 74.4% yield). δ 4.67 (s, 2H), 6.97 (d, J =
9.8, 1H), 7.02 (s, 1H), 7.17 (s, 1H), 7.24 (t, J = 72.6, 1H), 7.30 (m, J =
3.9, 1H), 7.40 (d, J = 7.8, 1H), 7.53 (d, J = 7.8, 1H), 8.52 (d, J = 5.8,
1H), 12.33 (s, 1H). ESI-MS (M + H)+ found, 308.0608; calculated for
C14H11F2N3OS, 308.0591.
Compound 32 (75.6 mg, 32.5% yield). δ 4.52 (s, 2H), 5.85
(s, 2H), 6.72 (d, J = 9.8, 1H), 6.94 (s, 1H), 7.15 (t, J = 7.8, 1H), 7.27
(d, J = 7.8, 1H), 7.30 (d, J = 9.8, 1H), 7.74 (t, J = 7.8, 1H), 8.67 (d, J =
7.8, 1H), 12.43 (s, 1H). ESI-MS (M + H)+ found, 257.0792; calculated
for C13H12N4S, 257.0783.
equipped with a dual-solvent self-washing system (CTC Analytics,
Carrboro, NC).
Statistical Analysis. Inhibition data (IC50 and Ki) was calculated
using Graphpad Prism (version 5.01; Graphpad Software Inc., San
Diego, CA). For fluorescence assays, all parameters were calculated
after first subtracting out background fluorescence from 3-O-
methylfluorescein. Inhibition curves were fit to a standard four-site
eq 1 to estimate an IC50 value.
(max − min)
(1 + 10
%activity = min +
(log[I]−logIC
)
50
)
(1)
P450 Inhibition Assay. Inhibition of recombinant CYP2C19
microsomes was determined for each analogue using 3-O-methyl-
fluorescein as a probe substrate. Initial experiments were performed to
ensure reaction linearity with respect to protein and time. Each
incubation (n = 2) consisted of 2 pmol recombinant protein in 100
mM potassium phosphate buffer (pH 7.4) containing 3 mM MgCl2.
Following a 3 min preincubation period, reactions were initiated with
the addition of 1 mM NADPH (final concentration). Organic solvent
concentrations were the same for all incubations and did not exceed
0.5% (v/v). IC50 values for each analogue were determined for
CYP2C19 using a final test compound concentration range of 0−50
μM. 3-O-methylfluorescein was run at 1 μM (approximate Km). All
incubations were allowed to proceed at 37 °C for 5 min and quenched
with 50 μL of 2N NaOH. Samples were vortex-mixed for 5 min prior
to analysis.
Ki estimates were determined using nonlinear regression analysis. The
mechanism of inhibition and inhibitory model selection was
determined through visual inspection of the Dixon ([I] vs 1/v) and
Lineweaver−Burke (1/S vs 1/v) plots and through comparison of the
Akaike information criteria. All data were subsequently fit to a
competitive inhibition model (2), where [S] is the probe substrate
concentration, Km is equal to the probe substrate concentration at half-
maximum reaction velocity, [I] is the test compound concentration in
the incubation, and Ki is the dissociation constant for the enzyme−
inhibitor complex. For Ki determinations, Km, Ki, and Vmax were
treated as global parameters. Goodness of fit was determined using
global r2 values.
V
·[S]
max
v =
⎛
⎞
[I]
K ⎜1 +
⎟ + [S]
m
Once IC50 values were estimated, a Ki was determined for each test
compound against CYP2C19 catalyzed 3-O-methylfluorescein metab-
olism. Ki experiments were run under similar conditions as the IC50
determination except that the probe substrate was run at 0.5, 1, 2, and
4 μM corresponding to concentrations equal to 0.5 Km, Km, 2Km, and
4Km, respectively. Each test compound was run at 0IC50, 0.5IC50, IC50,
2IC50, and 4IC50. Incubations were allowed to proceed at 37 °C for
5 min and quenched with 50 μL of 2N NaOH. All Ki determinations
were performed in duplicate.
K
⎝
⎠
(2)
i
Intrinsic Clearance Determination. To support the generation
of a pharmacophore model and assess the stability of the omeprazole
analogues, intrinsic clearance determinations were carried out in
recombinant CYP2C19 and human liver microsomes. Incubations in
recombinant CYP2C19 included 2 pmol CYP2C19, 2 mM MgCl2, and
0.5 μM substrate in 100 mM potassium phosphate buffer (pH 7.4).
Incubations were initiated with the addition of 1 mM NADPH (final
concentration) and aliquots removed at 0, 2.5, 5, 10, 20, and 40 min.
Reactions were quenched with 1 volume (v/v) of ice-cold acetonitrile
containing 1 μM tolbutamide as an internal standard. Human liver
microsomal clearance determinations were carried out in a similar
fashion and contained 0.1 mg/mL human liver microsomes, 3 mM
MgCl2, and 0.5 μM substrate in 100 mM potassium phosphate buffer
(pH 7.4). LC-MS/MS analysis of the omeprazole analogues was
performed using a similar instrumental protocol to that described for
the CYP inhibition assays.
Pharmacophore Model Generation. Ligands (compounds 1
through 26) were prepared using LigPrep (Ligprep 2.0 (2006)
Schrodinger, LLC, New York, NY), which attached hydrogen atoms
and converted two-dimensional .sdf files to three-dimensional
structures with proper stereocenters. The ligands were generated as
a neutral series. Ligand conformations were submitted to a pre-energy
minimization using the OPLS_2005 force field with a distance
dependent dielectric model. A maximum of 200 initial conformers
were examined. Minimized structures within 1 kcal energy and rmsd of
1 Å or less were eliminated. Each ligand was examined against the six
default pharmacophore features: hydrogen bond acceptor (A),
hydrogen bond donor (D), hydrophobic (H), negative charge (N),
positive ionization (P), and aromatic ring (R). The pharmacophore
model was generated using feature-based pharmacophore model. Four
ligands were chosen as active ligands for the model. Active compounds
were chosen based upon two criteria: (1) active compounds must
reside within the top 15% with respect to inhibitor potency and (2)
must contain some sense of structural diversity from one another. The
number of pharmacophore features was varied from a minimum of
three to a maximum of six features during the alignment process with a
limit of three hydrophobic sites allowed. Common pharmacophores
features were required to match at least 3 of the 4 active compounds.
Upon the basis of the size of pharmacophore set, a maximum of three
principal features were allowed. During pharmacophore hypothesis
testing, 50% of ligands were randomly assigned to the training set,
Sample analysis was performed on a Tecan Safire2 microplate reader
(Mannedorf, Switzerland). Fluorescein formation was monitored using
̈
an excitation wavelength of 485 nm (10 nm bandwidth) and an
emission wavelength of 530 nm (20 nm bandwidth). A manual gain of
40 (arbitrary units) was utilized with no lag time and an integration
time of 40 μs. Each sample was read twice and the average response
used in IC50 and Ki calculations.
To determine the selectivity of the omeprazole analogues versus
other P450 isoforms, IC50 determinations were carried out in
human liver microsomes with P450-selective substrate probes. Briefly,
0.1 mg/mL human liver microsomes, 3 mM MgCl2, 100 mM potas-
sium phosphate buffer (pH 7.4), and the P450-selective substrate
probe at previously determined Km value (CYP1A2, 85 μM phenacetin;
CYP2B6, 75 μM bupropion; CYP2C8, 7 μM paclitaxel; CYP2C9,
17 μM diclofenac; CYP2C19, 10 μM (S)-mephenytoin; CYP2D6, 5 μM
dextromethorphan; CYP2E1, 75 μM chlorzoxazone; CYP3A4, 2.5 μM
midazolam) were combined with each omeprazole analogue (0−50 μM,
final concentration) and preincubated for 5 min at 37 °C. Reactions were
initiated via addition of 1 mM NADPH (final concentration) and
quenched after 10 min (5 min for midazolam) with 2 volumes of ice cold
acetonitrile containing 1 μM tolbutamide as an internal standard.
LC-MS/MS analysis for the inhibition assay was carried out using
multiple reaction monitoring on an Applied Biosystems API 4000 Q-
Trap (operated in triple quadrupole mode) equipped with an
electrospray ionization source (Applied Biosystems, Foster City,
CA). Prior to analysis, mass spectrometry parameters were optimized
for each test compound. Global mass spectrometry settings included
the curtain gas (40 arbitrary units), ion spray voltage (4700 V), source
temperature (550 °C), and collision-assisted dissociation gas (set to
medium). Chromatographic separation was achieved using a rapid
gradient with a Gemini C18 30 mm × 2.0 mm (5 μm) column
(Phenomenex, Torrace, CA) coupled to two LC-20AD pumps with an
in-line CBM-20A controller and DGU-20A5 solvent degasser
(Shimadzu, Columbia, MD) and a LEAP CTC HTS PAL autosampler
1212
dx.doi.org/10.1021/jm201346g | J. Med. Chem. 2012, 55, 1205−1214

Journal of Medicinal Chemistry
Article
while the remainder was included as the test set. The remaining model
criteria were set to PHASE default parameters.
(3) Guengerich, F. P.; MacDonald, J. S. Applying mechanisms of
chemical toxicity to predict drug safety. Chem. Res. Toxicol. 2007, 20,
344−369.
Metabolite Identification in rCYP2C19. To elucidate the sites of
metabolism of the omeprazole analogues, metabolite identification
studies were performed in recombinant CYP2C19. Briefly, incubations
contained 20 pmol CYP2C19, 3 mM MgCl2, and 20 μM substrate in
100 mM potassium phosphate buffer (pH 7.4). Incubations were
allowed to proceed for 20 min before being quenched with 2 volumes
of ice-cold acetonitrile containing 0.1% formic acid (v/v). Samples
were centrifuged, and the resulting supernatant collected and
evaporated under nitrogen. Prior to mass spectral/UV analysis,
samples were reconstituted in water:acetonitrile:formic acid
(50:50:0.1). Analysis of metabolite formation was performed on a
ThermoFisher LTQ ion trap mass spectrometer. Chromatographic
separation was achieved using a Luna C18 150 μm × 2.0 5 μm column
(Phenomenex, Torrance, CA) and a 20 min linear gradient from
95:5 A:B to 5:95 A:B (A, water containing 0.1% formic acid; B,
acetonitrile containing 0.1% formic acid) at a flow rate of 250 μL/min.
CYP2C19 Homology Model Studies. A three-dimensional
homology model of CYP2C19 was built using Prime (Schrodinger,
LLC, New York, NY) and was based on the crystal structure of
CYP2C9 with flurbiprofen bound (PDB file 1R9O). The crystal
structure of CYP2C9 was chosen due to its sequence homology with
CYP2C19 and because it did not contain any amino acid
modifications. The amino acid sequence of CYP2C19 was obtained
from GenBank (CAH73444.1). Ramachandran plots confirmed the
structural plausibility of the homology model. Visual evaluation of the
model detected no odd bond angles or bond lengths.
SiteMap (Schrodinger, LLC) located and defined the P450 2C19
active site using a series of adjacent dummy atoms that estimated the
size and boundary of the active site. The SiteMap-defined active site
was employed to generate the docking grid for subsequent docking
experiments. The docking grid was defined by a 14 × 14 × 14 Å3 box
whereby the mass center of the each docked ligand was centered. The
remaining algorithm parameters were set to the default settings.
Ligands were docked using the Ligand Docking algorithm within Glide
(Schrodinger, LLC). Individual ligand poses were examined either by
GlideScore or eModel. Prior to docking, ligands were prepared in a
similar fashion to that described for the pharmacophore model.
(4) Denton, T. T.; Zhang, X.; Cashman, J. R. 5-Substituted,
6-Substituted, and Unsubstituted 3-Heteroaromatic Pyridine Analogues
of Nicotine as Selective Inhibitors of Cytochrome P-450 2A6. J. Med.
Chem. 2004, 48, 224−239.
(5) Gomaa, M. S.; Bridgens, C. E.; Aboraia, A. S.; Veal, G. J.; Redfern,
C. P.; Brancale, A.; Armstrong, J. L.; Simons, C. Small Molecule
Inhibitors of Retinoic Acid 4-Hydroxylase (CYP26): Synthesis and
Biological Evaluation of Imidazole Methyl 3-(4-(Aryl-2-ylamino)-
phenyl)propanoates. J. Med. Chem. 2011, 54, 6803−6811.
(6) Kim, S.; Ko, H.; Park, J. E.; Jung, S.; Lee, S. K.; Chun, Y.-J.
Design, Synthesis, and Discovery of Novel trans-Stilbene Analogues as
Potent and Selective Human Cytochrome P450 1B1 Inhibitors. J. Med.
Chem. 2001, 45, 160−164.
(7) Paine, M. F.; Hart, H. L.; Ludington, S. S.; Haining, R. L.; Rettie,
A. E.; Zeldin, D. C. The human intestinal cytochrome P450 “pie”.
Drug Metab. Dispos. 2006, 34, 880−886.
(8) Ring, B. J.; Eckstein, J. A.; Gillespie, J. S.; Binkley, S. N.;
VandenBranden, M.; Wrighton, S. A. Identification of the human
cytochromes p450 responsible for in vitro formation of R- and
S-norfluoxetine. J. Pharmacol. Exp. Ther. 2001, 297, 1044−1050.
(9) Feng, H. J.; Huang, S. L.; Wang, W.; Zhou, H. H. The induction
effect of rifampicin on activity of mephenytoin 4′-hydroxylase related
to M1 mutation of CYP2C19 and gene dose. Br. J. Clin. Pharmacol.
1998, 45, 27−29.
(10) Raucy, J. L.; Mueller, L.; Duan, K.; Allen, S. W.; Strom, S.;
Lasker, J. M. Expression and induction of CYP2C P450 enzymes in
primary cultures of human hepatocytes. J. Pharmacol. Exp. Ther. 2002,
302, 475−482.
(11) Foti, R. S.; Wahlstrom, J. L. CYP2C19 inhibition: the impact of
substrate probe selection on in vitro inhibition profiles. Drug Metab.
Dispos. 2008, 36, 523−528.
(12) Wang, J. H.; Li, P. Q.; Fu, Q. Y.; Li, Q. X.; Cai, W. W. Cyp2c19
genotype and omeprazole hydroxylation phenotype in Chinese Li
population. Clin. Exp. Pharmacol. Physiol. 2007, 34, 421−424.
(13) Holt, S.; Howden, C. W. Omeprazole. Overview and opinion.
Digestive Dis. Sci. 1991, 36, 385−393.
(14) Howden, C. W. Clinical pharmacology of omeprazole. Clin.
Pharm. 1991, 20, 38−49.
(15) Spencer, C. M.; Faulds, D. Lansoprazole. A reappraisal of its
pharmacodynamic and pharmacokinetic properties, and its therapeutic
efficacy in acid-related disorders. Drugs 1994, 48, 404−430.
(16) Meyer, U. A. Metabolic interactions of the proton-pump
inhibitors lansoprazole, omeprazole and pantoprazole with other
drugs. Eur. J. Gastroenterol. Hepatol. 1996, 8 (Suppl 1), S21−S25.
(17) Meyer, U. A. Interaction of proton pump inhibitors with
cytochromes P450: consequences for drug interactions. Yale J. biol.
Med. 1996, 69, 203−209.
AUTHOR INFORMATION
■
Corresponding Author
*Phone: 206-265-7423. E-mail: janw@amgen.com.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We acknowledge Dr. Brooke M. VandenBrink for valuable
assistance in preparing the pharmacophore and homology
model figures.
(18) Ko, J. W.; Sukhova, N.; Thacker, D.; Chen, P.; Flockhart, D. A.
Evaluation of omeprazole and lansoprazole as inhibitors of cytochrome
P450 isoforms. Drug Metab. Dispos. 1997, 25, 853−862.
(19) Andersson, T.; Andren, K.; Cederberg, C.; Edvardsson, G.;
Heggelund, A.; Lundborg, P. Effect of omeprazole and cimetidine on
plasma diazepam levels. Eur. J. Clin. Pharmacol. 1990, 39, 51−54.
(20) Andersson, T.; Cederberg, C.; Edvardsson, G.; Heggelund, A.;
Lundborg, P. Effect of omeprazole treatment on diazepam plasma
levels in slow versus normal rapid metabolizers of omeprazole. Clin.
Pharmacol. Ther. 1990, 47, 79−85.
ABBREVIATIONS USED
■
P450, cytochrome P450; QSAR, quantitative structure−activity
relationship; Ki, inhibition constant; Clint, intrinsic clearance;
LC-MS/MS, liquid chromatography-tandem mass spectrometry
REFERENCES
■
(1) Ortiz de Montellano, P. R.; De Voss, J. J. Substrate Oxidation by
Cytochrome P450 Enzymes. In Cytochrome P450: Structure,
Mechanism, and Biochemistry, 3rd ed.; Ortiz de Montellano, P. R.,
Ed.; Kluwer Academic/Plenum Publishers: New York, 2005;
pp 183−245.
(21) Enderle, C.; Muller, W.; Grass, U. Drug interaction: omeprazole
and phenprocoumon. BMC Gastroenterol. 2001, 1, 2.
(22) Pearson, J.; Dahal, U. P.; Rock, D.; Peng, C. C.; Schenk, J. O.;
Joswig-Jones, C.; Jones, J. P. The kinetic mechanism for cytochrome
P450 metabolism of type II binding compounds: evidence supporting
direct reduction. Arch. Biochem. Biophys. 2011, 511, 69−79.
(23) Peng, C. C.; Pearson, J. T.; Rock, D. A.; Joswig-Jones, C. A.;
Jones, J. P. The effects of type II binding on metabolic stability and
(2) Guengerich, F. P. Human Cytochrome P450 Enzymes. In
Cytochrome P450: Structure, Mechanism, and Biochemistry, 3rd ed.;
Ortiz de Montellano, P. R., Ed.; Kluwer Academic/Plenum Publishers:
New York, 2005; pp 377−530.
1213
dx.doi.org/10.1021/jm201346g | J. Med. Chem. 2012, 55, 1205−1214

Journal of Medicinal Chemistry
Article
binding affinity in cytochrome P450 CYP3A4. Arch. Biochem. Biophys.
2010, 497, 68−81.
P450 2C8. Evidence for a peripheral fatty acid binding site. J. Biol.
Chem. 2004, 279, 9497−9503.
(42) Wester, M. R.; Yano, J. K.; Schoch, G. A.; Yang, C.; Griffin, K. J.;
Stout, C. D.; Johnson, E. F. The structure of human cytochrome P450
2C9 complexed with flurbiprofen at 2.0-A resolution. J. Biol. Chem.
2004, 279, 35630−35637.
(43) Williams, P. A.; Cosme, J.; Ward, A.; Angove, H. C.; Matak
Vinkovic, D.; Jhoti, H. Crystal structure of human cytochrome P450
2C9 with bound warfarin. Nature 2003, 424, 464−468.
(44) Yano, J. K.; Wester, M. R.; Schoch, G. A.; Griffin, K. J.; Stout,
C. D.; Johnson, E. F. The structure of human microsomal cytochrome
P450 3A4 determined by X-ray crystallography to 2.05-A resolution.
J. Biol. Chem. 2004, 279, 38091−38094.
(24) Chiba, M.; Jin, L.; Neway, W.; Vacca, J. P.; Tata, J. R.; Chapman,
K.; Lin, J. H. P450 interaction with HIV protease inhibitors:
relationship between metabolic stability, inhibitory potency, and
P450 binding spectra. Drug Metab. Dispos. 2001, 29, 1−3.
(25) Li, H.; Sun, J.; Fan, X.; Sui, X.; Zhang, L.; Wang, Y.; He, Z.
Considerations and recent advances in QSAR models for cytochrome
P450-mediated drug metabolism prediction. J. Comput.-Aided Mol. Des.
2008, 22, 843−855.
(26) Locuson, C. W. II; Suzuki, H.; Rettie, A. E.; Jones, J. P. Charge
and substituent effects on affinity and metabolism of benzbromarone-
based CYP2C19 inhibitors. J. Med. Chem. 2004, 47, 6768−6776.
(27) Ridderstrom, M.; Zamora, I.; Fjellstrom, O.; Andersson, T. B.
Analysis of selective regions in the active sites of human cytochromes
P450, 2C8, 2C9, 2C18, and 2C19 homology models using GRID/
CPCA. J. Med. Chem. 2001, 44, 4072−4081.
(28) Yao, Y.; Han, W. W.; Zhou, Y. H.; Li, Z. S.; Li, Q.; Chen, X. Y.;
Zhong, D. F. The metabolism of CYP2C9 and CYP2C19 for gliclazide
by homology modeling and docking study. Eur. J. Med. Chem. 2009,
44, 854−861.
(29) Suzuki, H.; Kneller, M. B.; Haining, R. L.; Trager, W. F.; Rettie,
A. E. (+)-N-3-Benzyl-nirvanol and (−)-N-3-Benzyl-phenobarbital: new
potent and selective in vitro inhibitors of CYP2C19. Drug Metab.
Dispos. 2002, 30, 235−239.
(30) Houston, J. B., Kenworthy, K. E. Galetin, A. Typical and
Atypical Enzyme Kinetics. In Drug Metabolizing Enzymes: Cytochrome
P450 and Other Enzymes in Drug Discovery and Development; Lee, J. S.,
Obach, R. S., Fisher, M. B., Eds.; Marcel Dekker: New York, 2003; pp
211−254.
(31) Sutton, D.; Butler, A. M.; Nadin, L.; Murray, M. Role of
CYP3A4 in human hepatic diltiazem N-demethylation: inhibition of
CYP3A4 activity by oxidized diltiazem metabolites. J. Pharmacol. Exp.
Ther. 1997, 282, 294−300.
(32) Fukushima, Y.; Ushimaru, M.; Takahara, S. On the error of the
dixon plot for estimating the inhibition constant between enzyme and
inhibitor. Biochem. Mol. Biol. Ed. 2002, 30, 90−92.
(33) Carter, C. M.; Leighton-Davies, J. R.; Charlton, S. J.
Miniaturized receptor binding assays: complications arising from
ligand depletion. J. Biomol. Screening 2007, 12, 255−266.
(34) Lewis, D. F.; Lake, B. G.; Ito, Y.; Dickins, M. Lipophilicity
relationships in inhibitors of CYP2C9 and CYP2C19 enzymes.
J. Enzyme Inhib. Med. Chem. 2006, 21, 385−389.
(35) Suzuki, H.; Kneller, M. B.; Rock, D. A.; Jones, J. P.; Trager,
W. F.; Rettie, A. E. Active-site characteristics of CYP2C19 and
CYP2C9 probed with hydantoin and barbiturate inhibitors. Arch.
Biochem. Biophys. 2004, 429, 1−15.
(36) Ekins, S.; Bravi, G.; Binkley, S.; Gillespie, J. S.; Ring, B. J.; Wikel,
J. H.; Wrighton, S. A. Three and four dimensional-quantitative
structure−activity relationship (3D/4D-QSAR) analyses of CYP2D6
inhibitors. Pharmacogenetics 1999, 9, 477−489.
(37) Ekins, S.; Bravi, G.; Binkley, S.; Gillespie, J. S.; Ring, B. J.; Wikel,
J. H.; Wrighton, S. A. Three- and four-dimensional quantitative
structure−activity relationship analyses of cytochrome P-450 3A4
inhibitors. J. Pharmacol. Exp. Ther. 1999, 290, 429−438.
(38) Ekins, S.; de Groot, M. J.; Jones, J. P. Pharmacophore and three-
dimensional quantitative structure−activity relationship methods for
modeling cytochrome p450 active sites. Drug Metab. Dispos. 2001, 29,
936−944.
(39) Sciabola, S.; Morao, I.; de Groot, M. J. Pharmacophoric
fingerprint method (TOPP) for 3D-QSAR modeling: application to
CYP2D6 metabolic stability. J. Chem. Inf. Model. 2007, 47, 76−84.
(40) Rowland, P.; Blaney, F. E.; Smyth, M. G.; Jones, J. J.; Leydon,
V. R.; Oxbrow, A. K.; Lewis, C. J.; Tennant, M. G.; Modi, S.;
Eggleston, D. S.; Chenery, R. J.; Bridges, A. M. Crystal structure of
human cytochrome P450 2D6. J. Biol. Chem. 2006, 281, 7614−7622.
(41) Schoch, G. A.; Yano, J. K.; Wester, M. R.; Griffin, K. J.; Stout,
C. D.; Johnson, E. F. Structure of human microsomal cytochrome
1214
dx.doi.org/10.1021/jm201346g | J. Med. Chem. 2012, 55, 1205−1214