Charge and substituent effects on affinity and metabolism of benzbromarone-based CYP2C19 inhibitors

Article abstract of DOI:10.1021/jm049605m

Human cytochrome P450 (CYP) 2C19 is one of the most important CYP2C family members responsible for metabolizing commonly prescribed drugs. This research describes synthetic modifications to benzbromarone (Bzbr) to create the most potent CYP2C19 inhibitor

Full text of DOI:10.1021/jm049605m

6768
J. Med. Chem. 2004, 47, 6768-6776
Charge and Substituent Effects on Affinity and Metabolism of
Benzbromarone-Based CYP2C19 Inhibitors
,
Charles W. Locuson, II,† Hisashi Suzuki,‡ Allan E. Rettie,§ and Jeffrey P. Jones *
School of Molecular Biosciences, Washington State University, Pullman, Washington, 99164, Meiji Seika Kaisha, Ltd.,
Pharmaceutical Research Center, Pharmacokin, 760 Morooka-cho, Kohoku-ku, Yokohama, Japan 222-8567,
Department of Medicinal Chemistry, University of Washington, Box 357610, Seattle, Washington 98195-7610,
and Department of Chemistry, Washington State University, P.O. Box 644630, Pullman, Washington 99164-4630
Received May 27, 2004
Human cytochrome P450 (CYP) 2C19 is one of the most important CYP2C family members
responsible for metabolizing commonly prescribed drugs. This research describes synthetic
modifications to benzbromarone (Bzbr) to create the most potent CYP2C19 inhibitor ever
reported. The most important features enabling analogues of Bzbr to bind to CYP2C19 with
high affinity are low acidity (high pKa or nonionizability) and hydrophobic substituents adjacent
to the phenol moiety. Though CYP2C19 was known to prefer neutral substrates, the extent
was perhaps not realized until the anionic, parent compound Bzbr (Ki ) 3.7 µM) was compared
to a less acidic dimethyl analogue (Ki ) 0.033 µM). However, differences in affinity for anionic
and neutral Bzbr analogues did not appear to affect the regiospecificity of their metabolism by
CYP’s 2C19 and 2C9. In addition, some Bzbr analogues were metabolized both on the phenol
and benzofuran rings. By using a substrate with a methyl ether in place of the Bzbr phenol,
it was shown that some Bzbr analogues must be able to freely reposition after binding and
oxidize the more energetically favorable position. Normally, O-demethylation of this methyl
ether is favored over benzofuran hydroxylation based on ion current from LC/MS. Deuterium
substitution of the methyl ether results in an inverse isotope effect on benzofuran hydroxylation
(i.e. increased oxidation of this less favorable site). Likewise, Bzbr-based CoMFA models of
CYP2C19 demonstrated no clear preference for any one ligand alignment. This suggests results
from this modeling method must be interpreted carefully for each CYP isoform. In summary,
Bzbr analogues have demonstrated they can be adapted to other CYP2C enzymes in order to
probe isoform-specific properties.
Introduction
(NSAIDs) and hypoglycemics (first and second genera-
tion sulfonylureas)1 and is greatly inhibited by acidic
compounds such as sulfaphenazole (SPA)2,3 and ben-
zbromarone (Bzbr).4,5 CYP2C19, on the other hand, is
a major oxidizer of omeprazole and (S)-mephenytoin6
and is subject to inhibition by potent N-3-benzyl deriva-
tives of nirvanol and phenobarbital,7 all of which are
primarily neutral at physiological pH. Thus, it appears
2C19 lacks the strong preference for anions. The fact
that these two enzymes are 91% identical at the protein
level raises an interesting question regarding specificity.
How do the 43 differing amino acid residues between
the two enzymes work in concert to confer preference
for some substrates over others? This is a biochemical
question that has clinical implications. These enzymes’
variant alleles lead to altered pharmacokinetics for
proton pump inhibitors (PPIs), mephenytoin, phenytoin,
and tricyclic antidepressants and therefore, at some
dosage levels, efficacy and safety (reviewed in Desta et
al.8).
Recently, we have demonstrated that the potent 2C9
Bzbr inhibitors derive part of their affinity from existing
in their anionic phenolate form at physiological pH.
Remarkably, 2C19 appears to lack such a preference
even though many of the active site residues are
conserved between 2C19 and 2C9 as predicted by
Gotoh’s sequence alignments.9 Mutation of a few 2C9
residues and construction of 2C9 chimeras with larger
A superfamily of oxidative enzymes known as the
cytochromes P450 are expressed in several tissues
where they are utilized in both the biosynthesis of
physiologically important metabolites and the biotrans-
formation of foreign molecules, including drugs. Many
of the drug metabolism reactions are catalyzed by P450s
in the liver. By studying the specificity of these prima-
rily hepatic isoforms, the ability to characterize the
types of drug templates that possess high affinity for
each P450 should improve. It is these compounds that
could potentially interfere with the expected in vivo
clearance of a second drug metabolized by the same
isoform, thus leading to a drug interaction. Therefore,
with some knowledge of a P450’s pharmacophore, high-
affinity P450 inhibitors can be identified and eliminated
earlier in the drug design process to improve safety and
lower the cost of development.
The CYP2C subfamily of P450s is particularly impor-
tant in the metabolism of widely used drugs. For
instance, 2C9 is known to metabolize acidic substrates
such as the nonsteroidal antiinflammatory drugs
* To whom correspondence should be addressed. Tel: (509)-335-
5983. Fax: (509)-335-8867. E-mail: joswigjones@earthlink.net.
† School of Molecular Biosciences, Washington State University.
‡ Meiji Seika Kaisha, Ltd.
§ University of Washington.
Department of Chemistry, Washington State University.
10.1021/jm049605m CCC: $27.50 © 2004 American Chemical Society
Published on Web 11/26/2004

Benzbromarone Analogues as CYP2C19 Inhibitors
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27 6769
Table 1. Benzbromarone Analogs Screened for Inhibitory
Potency against 2C19, or Used as Substrates or Metabolite
Standards
Figure 1. Effect of pKa on 2C19 inhibition by parent com-
pound benzbromarone and a less acidic analogue. The num-
bering scheme of Bzbr follows De Vries et al..24 The number
in parentheses after the Ki value represents the error associ-
ated with the last significant figure.
2C19 substitutions begins to confer similar, but not
identical activity, to that of 2C19.10,11 To test the role
of substrate charge for 2C19, both neutral and acidic
analogues of the potent 2C9 inhibitor Bzbr were screened
for their relative affinity. As with 2C9, the Bzbr
analogues with two hydrophobic groups on the benzoyl
ring were greatly preferred by 2C19; however, in
contrast to Bzbr’s affinity for 2C9, the compounds with
these hydrophobic groups were better 2C19 inhibitors
when they were neutral or contained less acidic phenols
than Bzbr. This is surprising since a key Arg residue
(R108) in 2C9 implicated in binding anionic substrates
is also present in 2C19.12 The best 2C19 inhibitor
identified here most closely resembled Bzbr, but had a
much less acidic phenol (pKa of 8.5 versus 4.5). This
dimethylphenol analogue ((2-ethylbenzofuran-3-yl)-(4-
hydroxy-3,5-dimethylphenyl)methanone) is the most
potent 2C19 inhibitor thus far described (Ki ) 0.033 µM)
and demonstrates the importance of this class of com-
pounds in exploring the subtle structural differences
between 2C19 and 2C9.
a The number in parentheses after the Ki value represents the
error associated with the last significant figure.
Results
Inhibition of 3-O-Methylfluorescein Demethyl-
ation by 2C19. All Bzbr analogue inhibitors (Figure 1
and Table 1) were found to be competitive inhibitors
of methylfluorescein turnover by recombinant, recon-
stituted 2C19 according to nonlinear least-squares
fitting results from EnzFitter 2.0 shown in Figure 2
(BIOSOFT, Ferguson, MO). 2C19’s Vm and Km for
methylfluorescein demethylation was 1.5 nmol/min/
nmol and 5.9 µM, respectively.
Comparison of Relative Affinities. The best in-
hibitors of 3-O-methylfluorescein turnover by 2C19
were, by far, those that retained the closest structural
similarity to the parent compound Bzbr, but were less
acidic than Bzbr (Table 1). All five of these compounds
had Ki values in the nanomolar range. This includes the
dimethylphenol analogue 2 where the only difference
from Bzbr is the substitution of methyl groups for two
bromine atoms (Figure 1). By substituting the electron-
withdrawing bromines with methyl groups of nearly the
same van der Waals radii, the pKa of the phenol group
increases nearly 4 log units to 8.6. This ensures the
Figure 2. Dimethylphenol analogue 2 inhibition of 2C19-
mediated 3-O-methylfluorescein turnover shown with the
nonlinear fit to the competitive inhibition equation.
protonated form of this compound dominates under the
in vitro incubation conditions used (pH ) 7.4). This
dimethylphenol is the most potent 2C19 inhibitor
described to date. Likewise, the next four most effective
2C19 inhibitors either had higher phenol pKas relative
to Bzbr, or methylated phenols. When the Bzbr ketone
is reduced to the racemic alcohol 4 or the phenol ring is
isolated from the ketone with a methylene group (6),

6770 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27
Lucuson et al.
Figure 4. Selected ion mass chromatogram (substrate + 16)
and MS/MS data (inset) resulting from metabolism of the
dimethylphenol analogue 2 by 2C19. Two major hydroxylation
products were observed, but they reside on different ends of
the molecule.
Figure 3. Selected ion mass chromatogram at m/z 441.2
(substrate + 16) and MS/MS data (inset) resulting from Bzbr
metabolism by 2C19. An identical profile was observed for 2C9.
the pKa exceeds 8.5. Methylating compound 2 to obtain
3 and methylating Bzbr itself (5) also provides more
potent 2C19 inhibitors than any of the anionic deriva-
tives.
two fragments. The fragment with an m/z of 279.1
represents the unmodified benzoyl phenol ring. It also
clearly possesses the bromine isotope pattern when the
parent ion m/z window is increased. The second most
abundant fragment (m/z ) 189.1) represents 2-ethyl-
benzofuran plus 16 mass units, or a hydroxylated
benzofuran. Hydroxylation could occur at one of four
aromatic positions or the C-2 ethyl group. On the basis
of differences in chromatographic retention times and
fragmentation patterns of our synthesized standard
1′-OH Bzbr, hydroxylation does not appear to occur on
the C-2 ethyl group. The 1′-OH metabolite standard
eluted 0.8 min earlier under the chromatography method.
In positive mode, this standard also appeared to lose
water at the allylic position (m/z ) 422.9) and form an
adduct (m/z ) 454.5) where the product from enzyme
incubations gave only the expected M + 1 ion (m/z )
441.2). Both the 1′-OH metabolite standard and the
product from enzyme incubations gave the expected M
- 1 ion in negative polarity mode, but MS/MS results
showed major differences (data not shown). No evidence
was seen for multiple hydroxylations of the substrate.
The dimethylphenol analogue 2 formed two hydroxy-
lated metabolites (M + 16) of retention times 14.5 and
16.9 min with 2C19 and 2C9 (Figure 4). Analyte from
the first peak to elute fragmented at the ketone to give
the unmodified dimethylbenzoylphenol ring (m/z )
149.1) and a hydroxylated 2-ethyl benzofuran (m/z )
189.1) similar to the Bzbr results. The primary daughter
ion of the second hydroxylation product to elute had an
m/z corresponding to M - 18 suggesting water loss. At
lower m/z values, fragments for an unmodified benzo-
furan ring (m/z ) 173.1) and a hydroxylated dimeth-
ylphenol ring (m/z ) 137.1) were found. Together, these
results would place the site of hydroxylation at one of
two benzylic positions on the phenol ring.
The trend for 2C19 in preferring neutral Bzbr ana-
logues is opposite to the trend observed with the same
compounds that were screened for inhibition toward
2C913 because in every case, the anionic Bzbr analogues
have the highest affinity for 2C9 (e.g. 1, 7, 8, 13).
Conversely, the less acidic compounds 2-6 are only
slightly poorer inhibitors of 2C9 than the anionic
derivatives; however, with 2C19, 2-6 give Ki values at
least an order of magnitude lower than the anionic
compounds 1, 7, 8, 13 on average. In fact, the 100-fold
difference in Kis between Bzbr (1) and the less acidic
dimethylphenol 2 is one of the most striking effects of
charge observed among 2C19 ligands (Figure 1). Despite
this trend, all but two of the inhibitors (2, 3) are more
active toward inhibiting 2C9 relative to 2C19.
The remaining Bzbr analogues tested substituent
effects and their locations on the benzoyl ring. Mirroring
the 2C9 results, 2C19 preferred the phenol group at the
C-4" position (para to the ketone; analogue 9) over a
methoxy group at the same position (12) or a phenol at
the C-3" position (11). The poorest 2C9 inhibitor (14)
was also one of the poorest 2C19 inhibitors. This
substituted pyridine-containing structure contains two
halogens in the same positions as Bzbr, but has a
nitrogen in the ring where the Bzbr phenol is located.
One of the biggest differences between 2C19 and 2C9
was also noted with the cyclohexane-containing ana-
logue 15. Compound 15 was an equally poor inhibitor
of 2C19 as the pyridine derivative, whereas the same
compound was better than seven other compounds
screened against 2C9.
Metabolism of Bzbr (1) and Its Dimethyl (2) and
3′′-Methoxy (16, 17) Analogues. Both Bzbr and its
dimethyl analogue appear to be metabolized to the same
products by 2C19 and 2C9 as judged by retention times
and fragmentation patterns. Only one Bzbr metabolite
was detected with both enzymes, and it had a retention
time of 17.5 min (Figure 3). This metabolite is hydroxy-
lated on the benzofuran as evidenced by the m/z of the
Two readily distinguishable metabolites of 16 were
also observed in 2C19 and 2C9 incubations. At a RT of
17.6 min, MS/MS provided a fragment corresponding
to a hydroxylated benzofuran (m/z ) 175.1) and another
for the 3′′-methoxy ring (m/z ) 135.0). The second
product at 19.2 min was found to be an O-demethylation
product that eluted with the same RT as compound 11

Benzbromarone Analogues as CYP2C19 Inhibitors
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27 6771
and displayed the same fragmentation pattern. Its
fragments matched the m/z values of an unmodified
benzofuran (m/z ) 159.1) and a benzoyl phenol ring (m/z
) 121.0).
To probe the extent of substrate mobility after binding
to 2C19, 11 was deuterated to make analogue 17, which
differs from 16 only at the isotope-labled 3′′-methoxy
group. Analogue 17 gave the same two products as 16
with the addition of 3 Da to the methoxy-d3 ring. The
ratio of the two products, however, differed significantly
relative to the protio version (Table 1). A normal isotope
effect where more O-demethylation occurs with the
protio than the deutero analogue was observed for
O-demethylation, and a corresponding inverse isotope
effect was observed for benzofuran hydroxylation. The
presence of an inverse isotope effect infers two possibili-
ties. One is that the substrate retains high mobility after
binding, assuming limited dissociation. Alternatively,
the presence of another bound substrate may allow
branching to this second molecule if it can approach the
heme.14 All product ratios remained constant after
addition of the dimethylphenol 2 substrate to incuba-
tions containing the methoxy analogue, suggesting the
isotope effects do not result from multiple substrate
binding.
Comparative Molecular Field Analysis Modeling
(CoMFA). No validated models could be generated
using the 15 Bzbr analogues as evidenced by negative
q2 values. While this is a small data set, the same
compounds produced models with error below 0.5 ln
units using 2C9-derived Ki values.13 Since charged
analogues were among some of the poorest inhibitors
of 2C19, the atomic partial charges of those with pKas
lower than 5.2 (7, 8, 13) were recalculated with a charge
of -1 after the phenol proton was removed. The result-
ing models could not correlate ligand charges with their
associated Ki values.
To increase the diversity of the data set, the Bzbr
analogues were added to our recently described model
of 24 phenobarbital, nirvanol, and phenytoin derivatives
(in press).15 After this merging, cross-validated models
were readily generated, but were largely independent
of how Bzbr was aligned (Figure 5). Model statistics
remained mostly unchanged for all four alignments with
cross-validated q2 values and standard errors (ln units)
averaging 0.64 and 0.89, respectively.
Figure 5. Bzbr CoMFA model contours for 2C19 were similar
for multiple ligand alignments to phenobarbitals and not much
different from the phenobarbital model lacking Bzbr. Arrow
denotes site of metabolism of (-)-N-3-benzylphenobarbital and
proposed sites of metabolism of 2. (a) Alignment of the Bzbr
benzofuran to the metabolized position of (-)-N-3-benzylphe-
nobarbital. (b) Alignment of the dimethylphenol 2 metabolized
benzylic position to the metabolized position of (-)-N-3-
benzylphenobarbital. (c) Phenobarbital/nirvanol/phenytoin model
without Bzbr analogues.
compound Bzbr (2-6), but with decreased acidity,
possessed the best affinity for 2C19 as judged by Ki
values (Figure 1 and Table 1). Acidity was adjusted in
two ways. In some cases the ionizable phenol was simply
blocked with a methyl group as in 3 and 5. Alterna-
tively, the benzoyl ring substituents were altered to
diminish their electron-withdrawing capacity as in 2,
4, and 6. More than one hydrophobic group on the
phenol ring was just as essential for 2C19 affinity.
Shuffling of single benzoyl ring substituents to different
positions did not improve their inhibitory capability
even though they had low acidity. Therefore, all three
of Bzbr’s benzoyl ring substituents at C-3′′, -4′′, and -5′′
are required for optimal binding to 2C19 with a nona-
cidic phenol favored over methoxy at C-4′′. At the
benzofuran ring, substitutions were only made to the
length of the C-2 alkyl chain. No correlation in Ki values
was witnessed between chain lengths ranging from one
to four carbons in compounds 1, 7, and 8, suggesting
the higher negative charge of these acidic derivatives
makes them equally weaker binders.
Discussion
Five of the Bzbr analogues described here exhibit
extraordinary inhibitory potency for 2C19 with their
submicromolar Ki values. Therefore, as arguably some
of the most complementary ligands for the 2C19 active
site, Bzbr analogues should answer many questions
regarding 2C19 specificity. For example, this is the first
report of a series of closely related compounds that
demonstrate the charge state of a substrate could
determine whether it is a micromolar or a low nano-
molar inhibitor of 2C19 (Figure 1). Since Bzbr was
initially found to be a 2C9 inhibitor, this work also
shows the Bzbr-like compounds can be altered to exploit
other 2C enzymes as was done with the dimethyl
sulfaphenazole analogue DMZ.16
These findings would appear to agree with 2C9’s
favorite benzoyl ring substituents, but contrast 2C9’s
ability to bind both anionic and neutral Bzbrs with
nearly equal affinity. When the less acidic dimethylphe-
nol 2 was screened for its ability to inhibit (S)-warfarin
metabolism by 2C9, it was nearly as potent (Ki ) 40
nM) as anionic Bzbr (Ki ) 19 nM). This demonstrates
Features of the Best Bzbr 2C19 Inhibitors. Struc-
tures with the highest resemblance to the parent

6772 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27
Lucuson et al.
that given the same steric properties, requirements for
charge state in 2C9 substrates are not very stringent.
In fact, this turned out to be a general trend in most of
the Bzbr analogues screened against 2C9. Hence, the
same C-3′′, -4′′, and -5′′ substituents probably make
favorable interactions with both enzymes while 2C19
is much more intolerant of substrates with a formal
negative charge under physiological pH.
Comparison of Bzbr with Other 2C19 Sub-
strates. The higher-affinity 2C19 substrates that are
larger than (S)-mephenytoin all share similar features.
Bzbr, the N-benzylphenobarbitals, and sulfoxide-con-
taining proton pump inhibitors (e.g. omeprazole) all
have an aromatic or benzylic position that is metabo-
lized, separated from a second aromatic/hydrophobic
group by a polar, bridging group such as a ketone or
sulfoxide (Figure 6). On the basis of these findings, 2C19
would appear to make an H-bond or electrostatic
interaction with the polar, bridging group and a hydro-
phobic/aromatic interaction further beyond the heme.
Proton pump inhibitors suggest a direct enzyme inter-
action with the polar bridging group. The regiospecific
metabolism of omeprazole is primarily determined by
17
its chiral sulfoxide in 2C19
whereas lansoprazole
displays a change in affinity for 2C19 dependent upon
the chirality of its sulfoxide.18 Unfortunately, specific
residue(s), not including backbone interactions, cannot
19
be easily singled out from the recent 2C9 structures
because the two most apparent choices based on their
location in 2C9, R108 and H99, have outward facing side
chains. However, recent evidence for the flexibility of
12
the B′-C loop in 2C9
could make these residues
candidates for 2C19-substrate interactions either dur-
ing substrate recruitment or during catalysis. H99 is
intriguing because it is one of the closest residues to
the heme that is different from 2C9. This histidine is
already known to be a major source of 2C19’s ability to
metabolize omeprazole by presumably donating an
H-bond to its sulfoxide.10
20-22
Previous 2C19 homology models
and chimera
studies 11,22,23 have determined the general localities of
residues further from the heme that make 2C19 unique
based on mutagenesis. These residues presumably affect
enzyme motion during substrate binding or protein-
protein interactions. Further speculations regarding the
roles of all 2C19 residues that differ from 2C9 can be
found in the literature for the F-G loop and B′-C loop
Figure 6. Sites of oxidation of Bzbr analogues by 2C19
17
compared to the substrates omeprazole
and (-)-N-3-ben-
10
regions and the I- and G-helix regions.11,22,23
zylphenobarbital.15
CoMFA Models. Bzbr analogues were added to 24
phenobarbital analogues to carry out CoMFA because
the Bzbr analogues were too limited to derive models.
Unexpectedly, this combination of compounds generated
CoMFA models with moderate to high q2 values (0.5 <
q2 < 0.7) for all four ways in which Bzbr was aligned to
the N-substituted phenobarbitals (Figure 5). The mod-
els, then, appear inconclusive because there is no way
to determine which one is more correct. For instance,
the two most logical alignments are the overlap of either
the benzofuran C-6 metabolized position or phenol ring
with that of the phenobarbital metabolized position (C-5
phenyl ring),15 and alignment of the other two rings.
Yet the q2 values and standard errors are comparable
and a similar set of three-dimensional electrostatic and
steric interactions are found in the two contour plots
(Figure 5). Therefore, all models are either correct to
different extents, or provide almost no insight into
potential enzyme-ligand interactions.
The lack of more concrete CoMFA results can be
explained by two scenarios. Scenario number one is that
the data set is too limited. After reviewing the com-
pounds, it appears that the difficulty in modeling 2C19
using CoMFA could arise from the lack of sufficient
chemical diversity in the Bzbr compounds. Bzbr ana-
logues are clearly lacking in modifications to the ben-
zofuran group. Plus, the contour plot features for the
phenobarbital model do not change much after addition
of the Bzbrs. Both models contain prominent, favorable
electropositive (blue) and steric interaction (green) sites
near C-5 and the N-benzyl group of the phenobarbital

Benzbromarone Analogues as CYP2C19 Inhibitors
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27 6773
barbiturate ring, respectively (Figure 5). C-5 lies above
the aromatic rings that are metabolized in both classes
of compounds, and the hydrophobic N-benzyl group
aligns with whichever Bzbr ring is not positioned for
metabolism.
The second scenario is that there is no single sub-
strate conformation that is greatly favored over another
for Bzbr analogues. A single P450 isoform is often
capable of metabolizing multiple positions on a sub-
strate, but orienting substrates toward their primary
oxidation site in CoMFA modeling has proven viable in
the past for 2C9. Obviously, alignments used to generate
QSARs, even with the aid of active site docking, must
be carefully tested before any conclusions are drawn
based on statistically validated models.
Metabolism of Bzbr and Its Analogues. To test
whether the CoMFA data set does not fully explore all
enzyme-substrate interactions or if more than one
favorable binding conformation is populated, metabo-
lism studies were carried out on select Bzbr analogues.
The metabolism of Bzbr by 2C19 and 2C9 was studied
using LC/MS/MS. Previously, microsomal incubations
and human subject samples have allowed the charac-
terization of several Bzbr metabolites with the primary
hydroxylated products being 6-OH and 1′-OH Bzbr in a
ratio of 6:1, respectively.24 It is reported here that 2C19
and 2C9 are involved in forming the 6-OH Bzbr me-
tabolite.24 Though NMR was not used to assign the
6-OH position as done previously by purifying it from
human urine,24 the product possessed a different reten-
tion time (17.5 versus 16.7 min) and fragmentation
pattern than the synthesized 1′-OH metabolite, 18.
While these findings are largely qualitative, this is the
first report demonstrating the hydroxylation of Bzbr in
a nonmicrosomal, in vitro system.
The metabolism of Bzbr analogues 2 and 16 mirrors
that of omeprazole where either hydrophobic end can
approach the heme (Figure 6). Bzbr is metabolized at
the aromatic C-6 of the benzofuran by both 2C19 and
2C9, but due to the presence of two bromines, the phenol
ring is unavailable for oxidation. On the other hand,
metabolism of the dimethylphenol 2 and 16 demonstrate
their ability to bind in alternate conformations. In
addition to an unidentified benzofuran metabolite, the
dimethylphenol 2 is also converted to a benzyl alcohol
metabolite whereas 16 is O-demethylated to give 11.
Therefore, both of these analogues show how Bzbr
analogues can bind in more than one conformation
where the sites of metabolism are at each end as in
omeprazole.17
Figure 7. Isotope effects on metabolism of the 3′′-methoxy
analogue 16 suggest there is rapid switching of substrate
conformations after binding to 2C19. Both ring systems are
capable of colliding with the reactive oxygen species at the
heme; therefore, it is possible the binding site for Bzbr
analogues may have a hydrophobic or aromatic region that
interacts equally well with the substrate rings, or a H-bond
donor (e.g. H99) that allows it to pivot at the ketone.
Table 2. Kinetic Isotope Effects on Metabolism of Analogue 16
by 2C19
b
producta
kH/kD obsd
O-demethylation
benzofuran hydroxylation
2.6 ( 0.3
0.34 ( 0.05
a Characterized by LC/MS/MS. b Selected ion current of protio
and deutero analogue products corrected to internal standard and
divided by each other. Enzyme incubations were performed in
triplicate.
and methoxy metabolized positions are easier to oxidize
than the aromatic position on the benzofuran.25
To test whether Bzbrs are able to switch conforma-
tions after binding, or if the way they enter the enzyme
determines their metabolism, an isotope effect experi-
ment was performed. Compound 11 was deuterated
with iodomethane-2H3 to give 17, incubated with 2C19,
and the results compared with nonlabeled 16. Though
hydrogen atom abstraction during P450 catalysis is not
rate limiting, isotope effects are observable when a
substrate is able to switch conformations so that a
different position on the substrate is metabolized.25-27
A simplified scheme for this branching effect as pro-
posed by Miwa et al.28 is represented in Figure 7.
O-Demethylation is made more difficult by substituting
deuteriums, so an increase in the benzofuran metabolite
would signify the equilibrium population of the enzyme-
substrate complex that gives the benzofuran metabolite
increases, and this is what was found. An observed
kinetic isotope effect (kH/kD obs) of 2.6 for O-demethyla-
tion was noted (Table 2). Not only did O-demethylation
decrease, but benzofuran hydroxylation increased to
give an observed inverse isotope effect of 0.34. If only
O-demethylation decreased relative to the internal
standard, then this may indicate that the enzyme only
branches to the nonproductive shunt pathways; how-
ever, the witnessed isotope effects suggest there are
This challenges the hypothesis that the bromine and
methyl group phenol ring substituents interact at a site
somewhat removed from the heme. Due to the large
effects bromination and methylation of the phenol ring
have on 2C19 affinity, a complementary hydrophobic
pocket in the active site seemed likely. Other evidence
for this is that 2C9 appears to prefer these hydrophobic
groups for interactions at specific enzyme sites and is
not just a result of poorer water solubility.13 Certain
scenarios involving favorable hydrophobic pockets both
near the heme and further from the heme near the
nonmetabolized position can be envisioned; however,
another consideration is that of energetics. In the cases
of the dimethylphenol 2 and 16, the phenol ring benzylic

6774 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27
Lucuson et al.
Table 3. Product Ratios of Substrates 16 and 2 Observed with
ing drug-drug interactions and understanding the
molecular effects of point mutations resulting from
variant P450 genes. For 2C19, ligands with two hydro-
phobic regions separated by a polar group are the most
complementary according to this study with benzbro-
marone and previous works on phenobarbital analogues
and proton pump inhibitors. Particularly intriguing is
that high-affinity binding of benzbromarone ligands to
2C19 appears to be achieved without constraining
substrate mobility within the enzyme. This mobility
allows each hydrophobic end of the substrate to ap-
proach the heme for oxidation and is likely the reason
QSAR modeling, although statistically successful, could
not be evaluated objectively.
2C19 Separately and with Each Other
product ratio of 16a
O-demethylation to
substrate(s) benzofuran hydroxylation
product ratio of 2 benzylic
hydroxylation to benzofuran
hydroxylation
16
2.8 ( 0.2
-
2.8 ( 0.2
-
2
7.9 ( 0.5
7.5 ( 0.2
16 + 2
a Product ratios from enzyme incubations performed in triplicate
based on selected ion current.
multiple enzyme-substrate intermediates that inter-
change more rapidly than bond scission occurs. The
intrinsic isotope effect, which can be thought of as the
maximal isotope effect, can be estimated assuming
29
product branching
and irreversible substrate bind-
Materials and Methods
ing.28 If the products’ kH/kD obs ratios are divided, a value
of 7.6 is obtained as the intrinsic isotope effect esti-
mate.29 This value is near what would be observed for
a substrate that has unhindered mobility within the
enzyme.30
Chemicals. Fluorescein, 3-O-methylfluorescein, dichloro-
fluorescein, regenerating system components, and all bio-
chemicals were obtained from Sigma-Aldrich (St. Louis, MO).
Solvents were purchased from J. T. Baker (Phillipsburg, NJ).
Fine chemicals were from Aldrich (Milwaukee, WI). Iodo-
methane was purchased from Cambridge Isotope Laboratories
(Andover, MA). Warfarin and 4′OH-warfarin were gifts from
Prof. W. F. Trager at the University of Washington (Depart-
ment of Medicinal Chemistry).
Enzyme Sources. Purified, full-length human 2C19 was
kindly provided by Prof. A. E. Rettie at the University of
Washington (Department of Medicinal Chemistry). Details of
cDNA subcloning, expression in Trichoplusia ni cells, and
purification can be found in Suzuki et al.7 Bacterial expression
and purification of cytochrome P450 reductase and cytochrome
b5 were carried out using affinity methods as before.5
Synthesis. Synthesis, characterization, and pKa determi-
nation of Bzbr analogues 2-16 was previously described.5,13
Compounds 17 and 18 were characterized using a 300 MHz
Varian Mercury NMR spectrometer (Palo Alto, CA) and LC/
MS/MS (LCQ Advantage, Thermo Electron Corp., San Jose,
CA). The purity of both compounds 17 and 18 were confirmed
by HPLC in two different systems. The conditions and HPLC
traces are provided as Supporting Information.
An alternative explanation is the simultaneous oc-
cupation of a P450 active site by multiple substrates
where each molecule has the opportunity to collide with
the heme oxidizing species. If one molecule binds with
the methoxy group near the heme and a second molecule
binds with the benzofuran near the heme, then deute-
rium labeling would still allow branching to the benzo-
furan metabolite. This was first observed by Rock et al.
in fatty acid metabolism by P450 BM-3.14 Therefore, the
similar dimethylphenol substrate 2, which is also me-
tabolized at two positions, was added to 2C19 incuba-
tions with the methoxy analogue. Product ratios for both
substrates did not change (Table 3) indicating, but not
definitively ruling out, that one substrate binds per
catalytic turnover.
In summary, metabolic studies of Bzbr analogues
show for the first time that 2C19 regiospecificity is not
necessarily dictated by the orientation in which sub-
strate enters the enzyme. Figure 7 illustrates how
enzyme-substrate complexes exist at equilibrium after
binding to oxidize positions that are several angstroms
removed. On the basis of these and other current
findings, two kinds of enzyme-substrate interactions
could be proposed. First is the stereochemically sensitive
metabolism of proton pump inhibitors that is driven by
sulfoxide chirality. Since the Bzbr analogues are also
metabolized on multiple rings separated by a ketone, a
polar group may act as a pivot point for substrate
branching. The second observation is that N-benzyl
substitution of phenobarbital increases its affinity into
the low nanomolar Ki range. Therefore, hydrophobic or
aromatic interactions with substrate could be made with
either ring system so that one of the rings is always
placed above the heme.
(3-Methoxyphenyl)-(2-methylbenzofuran-3-yl)metha-
none-2H3 (17). Analogue 11 (1.0 mmol) 13 was added to 9 mL
of anhydrous THF containing 1 mL of DMF at 4 °C. Sodium
hydride dispersion (1.2 equiv) that had been previously washed
with hexanes and kept under argon was then added. Iodo-
methane-2H3 (99.5% enriched, 10 equiv) was then introduced
slowly by syringe and stirred for 1 h at which point the reaction
was quenched with 1 mL of methanol. After removal of solvent
in vacuo, the sample was run over a bed of silica with 5% ethyl
acetate in hexanes and eluted as a yellow oil. Isolated yield:
1
39%. H NMR (CDCl3): δ ) 2.55 (s, 3H), 7.20 (m, 3H, ArH),
7.37 (m, 3H, ArH), 7.45 (m, 2H, ArH). 13C NMR (CDCl3): δ )
15.0, 21.9, 111.1, 113.4, 117.2, 119.4, 121.6, 122.0, 123.8, 124.7,
127.1, 129.8, 140.9, 153.9, 160.0, 162.2, 192.0. ESI-MS: m/z
) 270.4 (M + H+), 159.1 (M - 110, 100), 138.1 (M - 131, 39)
at 40% ionization energy.
(3,5-Dibromo-4-hydroxyphenyl)-[2-(1-hydroxyethyl)-
benzofuran-3-yl]methanone (18). First, 0.25 mmol of Bzbr
protected as an acetate at the phenol was added to 1.2 equiv
of N-bromosuccinimide and 0.1 equiv of AIBN in 20 mL of CCl4
and heated to reflux for 4 h to form the 1′-bromo compound.
Bromine displacement was next carried out in 10 mL of DMF
with 5 equiv of cesium acetate at 23 °C for 24 h after which
the solvent was removed by vacuum. These first two steps
essentially went to completion based on TLC. The diester
compound was then hydrolyzed with anhydrous potassium
carbonate (5 equiv) in methanol for 12 h at 23 °C and
concentrated. Finally, the sample was subjected to preparative
TLC using 5% methanol in methylene chloride to develop the
plate and give the C-1′ alcohol of Bzbr. Isolated yield: 17%
1H NMR (CDCl3): δ ) 1.71 (d, 3H), 5.14 (q, 1H), 7.23 (m, 3H,
ArH), 7.37 (m, 1H, ArH), 7.58 (d, 1H, ArH), 8.04 (s, 2H, ArH).
Conclusions
The findings presented here demonstrate the in-
creased utility of benzbromarone analogues since they
have now been adapted to make the best 2C19 inhibitor
reported. Because P450s accommodate many substrates,
having high-affinity ligands should prove useful in
describing the most favorable enzyme-substrate inter-
actions based on their complementary features. This is
important for drug design from the standpoint of avoid-

Benzbromarone Analogues as CYP2C19 Inhibitors
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27 6775
13C NMR (CDCl3): δ ) 20.3, 63.7, 110.4, 112.2, 116.4, 121.4,
124.4, 125.7, 125.9, 133.0, 134.1, 153.7, 153.9, 166.9, 188.9.
ESI-MS: m/z ) 439.2 (M - H+), 251.0 (M - 189, 42) at 40%
ionization energy.
modes and a range of collision energies. Final collision energies
were 40% for Bzbr, the dimethylphenol, and the 3′′-methoxy
analogue in positive mode and 60% for the dimethyl analogue
in negative mode.
Enzyme Reconstitution and 3-O-Methylfluorescein
(MFL) Assay. Purified, recombinant 2C19 was reconstituted
with purified cytochrome P450 reductase and cytochrome b5
into unilamellar liposomes of dilauroylphosphatidylcholine
(DLPC) exactly as before.5 Reactions were carried out in a 1.5
mL polypropylene microcentrifuge tube with different concen-
trations of MFL at 3-6 fixed amounts of inhibitor (in metha-
nol) with 2.5 pmol of the reconstituted 2C19, an NADPH
regenerating system (2 mM MgCl2, 10 mM glucose-6-phos-
phate, 1 mM NADP, 1 U of glucose-6-phosphate dehydroge-
nase), and 1000 U of catalase in 50 mM potassium phosphate
buffer (pH 7.4) at a final volume of 0.2 mL. The stock solution
of MFL was prepared to be 0.2 mM by sonicating the MFL
with 1 equiv of potassium hydroxide in water. The reactions
were allowed to continue for 20 min in a shaking water bath
maintained at 37 °C. Acetonitrile (0.2 mL) containing 100 pM
dichlorofluorescein (DCFL, internal standard) was then used
to quench the incubations. The sample tubes were subse-
quently centrifuged at 12 000g for 20 min so the precipitated
protein components could be discarded.
Separation of fluorescein was carried out using HPLC
(Agilent Zorbax XDB-C8, 4.6 × 150 mm, 5 µm) using potas-
sium phosphate at pH 8.0 (A) and acetonitrile (B). The samples
were injected onto the column equilibrated at 95% A/5% B
whereupon a gradient was immediately initiated to reach 65%
A/35% B over 9 min and held at this ratio for an additional 4
min. A Hitachi (San Jose, CA) L-7480 fluorescence detector
was used to quantitate fluorescein based upon a standard
curve with the excitation and emission wavelengths set to 495
and 525 nm, respectively. Retention times were as follows: FL,
9.0 min, DCFL, 10.2 min, and MFL, 12.8 min.
Metabolism of Bzbr (1), and Its Dimethylphenol (2)
and 3′′-Methoxy (16, 17) Analogues. Reconstituted 2C19
and 2C9 were incubated with 25 µM Bzbr (1) and its dimethyl
analogue 2 for 30 min in 50 mM potassium phosphate, pH 7.4,
after initiating the reactions with 1 mM NADPH. Substrates
were dissolved in methanol to make 5 or 10 mM stock
solutions. Hydrochloric acid (3 M, 0.2 mL) was used to quench
the reactions, and 50 pmol of warfarin was added as internal
standard before the products were extracted into 6 mL of
chloroform. After being dried over MgSO4, the samples were
concentrated under nitrogen and dissolved in 200 µL of
methanol for LC/MS analysis. The HPLC method used a C-18
column (Agilent Hypersil BDS, 2.0 × 125 mm) and a solvent
system of 0.1% acetic acid (C) and methanol with 0.1% acetic
acid (D) with a flow rate of 0.2 mL/min. Initial conditions
consisting of 60% C/40% D were developed to 20% C/80% D
from t ) 0 to t ) 15 min and held at this solvent ratio for 10
additional minutes to elute the remaining substrate. A PDA
detector was used to monitor the UV traces at 282 nm.
Metabolism of 16 and 17 was carried with the following
changes. Substrate concentration was increased to 50 µM,
incubations lasted 60 min, 4′OH-warfarin was used as an
internal standard, and mobile phase solvent D was increased
to only 65% over the same time course used for Bzbr.
Incubations containing both 2 and 16 had these substrate
concentrations at 25 and 50 µM, respectively. Retention times
under the described conditions were: 4′OH-warfarin, 12.8 min,
warfarin, 14 min, dimethylphenol 2, 20 min, Bzbr (1), 25 min,
and 3′′-methoxy analogue 16, 27.4 min. Details of metabolites
are described in the Results section.
Comparative Molecular Field Analysis (CoMFA) Mod-
eling. CoMFA is a ligand-based predictive software with the
ability to derive three-dimensional quantitative structure-
activity relationships (3D-QSARs). Using regression analysis
on the charges for each atom of each molecule in a 3-D aligned
library, CoMFA can correlate electrostatic and steric properties
with activities (e.g. Ki values) obtained by experiment. Two
utilities of CoMFA analysis are its use as a predictive tool to
predict Kis (or other properties) of other ligands and contour
plots that allow the visualization of predicted favorable and
unfavorable properties of the ligands in 3-D. Generally, these
plots can be directly translated to interactions with the
enzyme, allowing a more directed route of ligand design that
exploits the desired properties.
All methods required to use Bzbr analogues in CoMFA
analysis were previously described in greater detail for the
2C9 model.13 This includes Bzbr ligand alignment, minimiza-
tion, and calculation of MNDO partial charges using the
MOPAC module within SYBYL v6.6.31 The Bzbr analogues
were then added to a new molecular spreadsheet containing
the reciprocal ln(-Ki) values determined for 2C19 experimen-
tally. All CoMFA analysis used the default parameters of 30
kcal/mol electrostatic and steric cutoffs, and leave-one-out
cross-validation with 2.0 kcal/mol column filtering was used
in all cases.
Bzbr ligands were aligned to the phenobarbital, nirvanol,
and phenytion analogues used in our previous 2C19 CoMFA
15
models
in four distinct ways. The first two methods
fashioned the two sites of metabolism for analogue 2 to overlie
the metabolized position of the most potent phenobarbital
analogue, (-)-N-3-benzylphenobarbital, whose structure is in
Figure 6. The remaining aromatic ring was then aligned with
the benzyl group of (-)-N-3-benzylphenobarbital. Figure 5a
and 5b illustrate these two alignments. The third and fourth
alignments that are not presented placed the Bzbr benzofuran
in the plane of the barbiturate ring or rotated the template
benzyl ring 180°. For the latter, this placed the nonmetabolized
ring position for both compounds at the other face of the
barbiturate ring.
Acknowledgment. We thank Jan Wahlstrom for
detailing the synthetic scheme for 1′-OH Bzbr and
suggesting target Bzbr analogues, and Tamara Dowers,
Greg Crouch, and Josh Alfaro for technical help.
This research was supported by NIH research grants
GM032165 and ES009122.
Supporting Information Available: HPLC traces of
compounds 17 and 18 are available. This material is free of
charge via the Internet at http://pubs.acs.org.
References
(1) Miners, J.; Birkett, D. Cytochrome P4502C9: an enzyme of major
importance in human drug metabolism. Br. J. Clin. Pharmacol.
1998, 45, 525-538.
(2) Rettie, A.; Korzekwa, K.; Kunze, K.; Lawrence, R.; Eddy, A.;
Aoyama, T.; Gelboin, H.; Gonzalez, F.; Trager, W. Hydroxylation
of warfarin by human cDNA-expressed cytochrome P-450: a role
for P-4502C9 in the etiology of (S)-warfarin-drug interactions.
Chem Res Toxicol. 1992, 5, p 54-59.
(3) Mancy, A.; Dijols, S.; Poli, S.; Guengerich, P.; Mansuy, D.
Interaction of sulfaphenazole derivatives with human liver
cytochromes P450 2C: molecular origin of the specific inhibitory
effects of sulfaphenazole on CYP 2C9 and consequences for the
substrate binding site topology of CYP 2C9. Biochemistry 1996,
35 (50), p 16205-16212.
(4) Takahashi, H.; Sato, T.; Shimoyama, Y.; Shioda, N.; Shimizu,
T.; Kubo, S.; Tamura, N.; Tainaka, H.; Yasumori, T.; Echizen,
H. Potentiation of anticoagulant effect of warfarin caused by
enantioselective metabolic inhibition by the uricosuric agent
benzbromarone. Clin Pharmacol Ther. 1999, 66 (6), 569-581.
All mass spectrometry was carried out with a Thermo
Finnigan LCQ Advantage (Thermo Electron Corp., San Jose,
CA) first in full scan mode and then in selected ion monitoring
mode to see if any minor hydroxylated metabolites (M + 16)
could be detected. No dihydroxylated products were detected
in full scan mode, but SIM mode was not used to rule out the
existence of such metabolites. After selecting the metabolite
parent ions ((M + 1) or (M - 1)) with m/z widths of 1 amu,
MS/MS experiments were then conducted under both polarity

6776 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 27
Lucuson et al.
(5) Locuson, C. W., II.; Wahlstrom, J. L.; Rock, D. A.; Rock, D. A.;
Jones, J. P. A new class of CYP2C9 inhibitors: probing 2C9
specificity with high-affinity benzbromarone derivatives. Drug
Metab Dispos. 2003, 31 (7), 967-971.
(19) 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 (6947), 464-
468.
(6) Goldstein, J. A.; Faletto, M. B.; Romkes-Sparks, M.; Sullivan,
T.; Kitareewan, S.; Raucy, J. L.; Lasker, J. M.; Ghanayem, B. I.
Evidence that CYP2C19 is the major (S)-mephenytoin 4′-
hydroxylase in humans. Biochemistry 1994, 33 (7), 1743-1752.
(7) Suzuki, H.; Kneller, M. B.; Haining, R. L.; Trager, W. F.; Rettie,
A. E. (+)-N-3-benzyl-nirvanol and (-)-N-3-benzyl-phenobar-
bital: new potent and selective in vitro inhibitors of CYP2C19.
Drug Metab. Dispos. 2002, 30, 235-239.
(20) Payne, V. A.; Chang, Y. T.; Loew, G. H. Homology modeling and
substrate binding study of human CYP2C18 and CYP2C19
enzymes. Proteins 1999, 37, 204-217.
(21) Lewis, D. F. V. Modelling human cytochromes P450 involved in
drug metabolism from the CYP2C5 crystallographic template.
J. Inorg. Biochem. 2002, 91, 502-514.
(22) Jung, F.; Griffin, K. J.; Song, W.; Richardson, T. H.; Yang, M.;
Johnson, Identification of amino acid substitutions that confer
a high affinity for sulfaphenazole binding, E. F.; a high catalytic
efficiency for warfarin metabolism to P450 2C19. Biochemistry
1998, 37, 16270-16279.
(23) Niwa, T.; Kageyama, A.; Kishimoto, K.; Yabusaki, Y.; Ishibashi,
F.; Katagiri, Amino acid residues affecting the activities of
human cytochrome P450 2C9, M.; 2C19. Drug Metab. Dispos.
2002, 30 (8), 931-936.
(24) De Vries, J.; Walter-Sack, I.; Voss, A.; Forster, W.; Ilisistegui
Pons, P.; Stoetzer, F.; Spraul, M.; Ackermann, M.; Moyna, G.
Metabolism of benzbromarone in man: structures of new oxida-
tive metabolites, 6-hydroxy- and 1′-oxo-benzbromarone, and the
enantioselective formation and elimination of 1′-hydroxyben-
zbromarone. Xenobiotica 1993, 23 (12), 1435-1450.
(25) Higgins, L.; Korzekwa, K. R.; Rao, S.; Shou, M.; Jones, J. P. An
assessment of the reaction energetics for cytochrome P450-
mediated reactions. Arch. Biochem. Biophys. 2001, 385 (1), 220-
30.
(26) Jones, J. P.; Korzekwa, K. R.; Rettie, A. E.; Trager, Isotopically
sensitive branching, W. F.; its effect on the observed intramo-
lecular isotope effects in cytochrome P-450 catalyzed reactions:
a new method for the estimation of instrinsic isotope effects. J.
Am. Chem. Soc. 1986, 108, 7074-7078.
(27) Rock, D. A.; Boitano, A. E.; Wahlstrom, J. L.; Jones, J. P. Use of
kinetic isotope effects to delineate the role of phenylalanine 87
in P450(BM-3). Bioorg. Chem. 2002, 30 (2), 107-18.
(28) Miwa, G. T.; Garland, W. A.; Hodshon, B. J.; Lu, A. Y. H.;
Northrop, D. B. Kinetic isotope effects in cytochrome P-450-
catalyzed oxidation reactions. Intermolecular and intramolecular
deuterium isotope effects during the N-demethylation of N,N-
dimethylphentermine. J. Biol. Chem. 1980, 255, 6049-6054.
(29) Korzekwa, K. R.; Gillette, J. R.; Trager, W. F. Isotope effect
studies on the cytochrome P450 enzymes. Drug Metab. Rev.
1995, 27 (1-2), 45-59.
(8) Desta, Z.; Zhao, X.; Shin, J.-G.; Flockhart, D. A. Clinical
significance of the cytochrome P450 2C19 genetic polymorphism.
Clin Pharmacokinet, 2002, 41 (12), 913-958.
(9) Gotoh, O. Substrate recognition sites in cytochrome P450 family
2 (CYP2) proteins inferred from comparative analyses of amino
acid and coding nucleotide sequences. J. Biol. Chem. 1992, 267
(1), 83-90.
(10) Ibeanu, G. C.; Ghanayem, B. I.; Linko, P.; Li, L.; Pedersen, L.
G.; Goldstein, Identification of residues 99, J. A.;220, and 221
of human cytochrome P450 2C19 as key determinants of
omeprazole hydroxylase activity. J. Biol. Chem. 1996, 271 (21),
12496-12501.
(11) Tsao, C.; Wester, M.; Ghanayem, B.; Coulter, S.; Chanas, B.;
Johnson, E.; Goldstein, J. Identification of human CYP2C19
residues that confer S-mephenytoin 4′-hydroxylation activity to
CYP2C9. Biochemistry 2001, 40, 1937-1944.
(12) Dickmann, L. J.; Locuson, C. W., II.; Jones, J. P.; Rettie,
Differential roles of R97, A. E.; D293, and R108 in enzyme
stability and substrate specificity of CYP2C9. Mol. Pharmacol.
2004, 65, 842-850.
(13) Locuson, C. W., II.; Rock, D. A.; Jones, J. P. Quantitative binding
models for CYP2C9 based on benzbromarone analogues. Bio-
chemistry 2004, 43, 6948-6958.
(14) Rock, D. A.; Perkins, B. N.; Wahlstrom, J.; Jones, J. P. A method
for determining two substrates binding in the same active site
of cytochrome P450 µB3: an explanation of high energy omega
product formation. Arch Biochem Biophys, 2003, 416 (1), 9-16.
(15) Suzuki, H.; Kneller, M. B.; Rock, D. A.; Jones, J. P.; Trager, W.
F.; Rettie, A. E. Active-site characteristics of CYP2C19 probed
with novel inhibitors: development of an integrated CoMFA-
homology model. Arch. Biochem. Biophys. 2004, in press.
(16) Marques-Soares, C.; Dijols, S.; Macherey, A. C.; Wester, M. R.;
Johnson, E. F.; Dansette, P. M.; Mansuy, D. Sulfaphenazole
derivatives as tools for comparing cytochrome P450 2C5 and
human cytochromes P450 2Cs: identification of a new high
affinity substrate common to those enzymes. Biochemistry 2003,
42 (21), 6363-6369.
(30) Iyer, K. R.; Jones, J. P.; Darbyshire, J. F.; Trager, W. F.
Intramolecular isotope effects for benzylic hydroxylation of
isomeric xylenes and 4,4′-dimethylbiphenyl by cytochrome
P450: Relationship between distance of methyl groups and
masking of the intrinsic isotope effect. Biochemistry 1997, 36
(23), 7136-7143.
(17) Abelo, A.; Andersson, T. B.; Antonsson, M.; Naudot, A. K.;
Skanberg, I.; Weidolf, L. Stereoselective metabolism of omepra-
zole by human cytochrome P450 enzymes. Drug Metab. Dispos.
2000, 28 (8), 966-972.
(18) Kim, K. A.; Kim, M. J.; Park, J. Y.; Shon, J. H.; Yoon, Y. R.;
Lee, S. S.; Liu, K. H.; Chun, J. H.; Hyun, M. H.; Shin, J. G.
Stereoselective metabolism of lansoprazole by human liver
cytochrome p450 enzymes. Drug Metab. Dispos. 2003, 31 (10),
1227-1234.
(31) SYBYL, Tripos Inc., v6.6: St. Louis, MO.
JM049605M