![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Vol. 59, Issue 3, 475-484, March 2001
Department of Pharmacology and Toxicology, University of Texas Medical Branch, Galveston, Texas (M.S., Q.W., Y.Q.H., J.R.H.); Department of Molecular and Experimental Medicine, The Scripps Research Institute, La Jolla, California (M.R.W., E.F.J.)
 |
Abstract |
|---|
 Top
 Abstract
 Introduction
Experimental Procedures
Results
Discussion
References
|
|---|
The molecular basis for reversible inhibition of rabbit CYP2B4 and CYP2B5 and rat CYP2B1 by phenylimidazoles was assessed with active-site mutants and new three-dimensional models based on the crystal structure of CYP2C5. 4-Phenylimidazole was 17- to 32-fold more potent toward CYP2B4 and CYP2B1 than CYP2B5. The 3D models, along with site-directed mutagenesis data, revealed the importance of residue 114 for sensitivity to inhibition of all three CYP2B enzymes. Besides Ile 114, Val 367 was also found to be critical for inhibition of CYP2B4 and CYP2B1. The most interesting new insights were obtained from analysis of the CYP2B5 model and the CYP2B5 active-site mutants. Simultaneous substitution of residues 114, 294, 363, and 367 with the corresponding residues of CYP2B4 decreased the IC50 value for inhibition by 4-phenylimidazole 12-fold. Docking 4-phenylimidazole into the models of CYP2B5 mutants demonstrated that the inhibitor-binding site is strongly influenced by residue-residue interactions, especially between residues 114 and 294. A chlorine substitution at position 4 of the phenyl moiety of 4- and 1-phenylimidazole resulted in IC50 values 95- and 130-fold lower for CYP2B4 than for CYP2B5, respectively, suggesting that these compounds are selective inhibitors of CYP2B4. Overall, the study revealed that differences in the determinants of inhibition between CYP2B4 and CYP2B5 are caused not only by single residue inhibitor contacts but also by residue-residue interactions. This new generation of CYP2B models may provide valuable information for the design of selective inhibitors of human CYP2B6 and for the development of drugs that avoid drug interactions due to P450 inhibition.
| |
Introduction |
|---|
|
Top
Abstract
Introduction
Experimental Procedures
Results
Discussion
References
|
|---|
A
major challenge in the drug discovery process is the ability to predict
catalytic specificities of individual P450 enzymes. Selective
inhibitors are crucial for the identification of P450 enzymes involved
in biotransformation (Halpert, 1995
). One approach to finding P450
inhibitors is in vitro screening of compounds using microsomes or
heterologous expression systems (Pelkonen et al., 1998). Another
strategy is developing active-site mutants and homology models of the
enzymes. This more theoretical approach has the advantage of being
applicable to virtual screening of compounds not yet synthesized.
Furthermore, this method provides detailed knowledge of the molecular
basis of inhibition and facilitates the drug discovery process through
prediction of suitable structural alterations that minimize potential
drug-drug interactions caused by P450 inhibition but maintain
pharmacological activity. In the absence of an experimental structure,
models of mammalian P450s have been constructed by homology with
bacterial P450s of known crystal structure (Szklarz et al., 1995;
Lewis, 1998; de Groot et al., 1999; Payne et al., 1999; Dai et al.,
2000). Models based on bacterial enzymes such as CYP102 have
been used to pinpoint key residues that may be important for substrate
specificity (Lewis and Lake, 1997). Furthermore, the analysis of
interactions between single active-site residues and the substrate has
facilitated the understanding of regio- and stereospecificity of
substrate oxidation as well as susceptibility to inhibition or
inactivation (Szklarz et al., 1995; Chang et al. 1997). However, the
low sequence identity of mammalian P450s compared with the bacterial
enzymes has limited progress.
Rat CYP2B1 and rabbit CYP2B4 and CYP2B5 provide an excellent basis to
study the structural determinants of inhibition. These 2B enzymes have
been extensively studied by molecular modeling (Szklarz et al., 1995;
Chang et al., 1997; Lewis and Lake, 1997; Dai et al., 1998) and
site-directed mutagenesis. Studies in our laboratory have revealed that
residues 114, 206, 209, 290, 294, 302, 363, 367, 477, 478, and 480, which are located in five different substrate recognition sites (SRSs),
control specificities toward various substrates such as
androstenedione, progesterone, pentoxyresorufin, benzyloxyresorufin,
7-ethoxycoumarin, and benzphetamine (He et al., 1992, 1994, 1995;
Szklarz et al., 1996; Kobayashi et al., 1998). CYP2B4 and CYP2B5 are
especially intriguing, because they differ in only 12 amino acid
residues but exhibit different catalytic selectivities. Four amino acid
residues [those at position 114 (SRS-1), 294 (SRS-4), and 363 and 367 (SRS-5)] have been identified as crucial for distinct activities of
CYP2B4 and CYP2B5 (He et al., 1996; Szklarz et al., 1996). Recently,
the structural basis for selective CYP2B4 and CYP2B5 inactivation was
assessed using active-site mutants and homology modeling.
2-Ethynylnaphthalene was identified as a selective inactivator of
CYP2B4 but not of CYP2B5. In molecular dynamic simulations,
2-ethynylnaphthalene was stable within the CYP2B4 model but exhibited
significant movement away from the heme moiety in the CYP2B5 model.
Interconversion of 2-ethynylnaphthalene susceptibility was achieved for
CYP2B4 and CYP2B5 by a single alteration at position 363 (Strobel et al., 1999). However, the mechanism-based inactivation process is
complex, and elucidation of the molecular basis of reversible inhibition would be of great interest.
Some of the most potent reversible inhibitors of P450 enzymes are the
nitrogen heterocycles, including imidazoles and quinolines. P450
inhibition by these compounds results from direct interaction between
the aromatic nitrogen of the heterocycle and the heme moiety of the
enzyme (Murray, 1987). X-ray crystallographic studies of bacterial
P450cam complexed with 1-, 2- or
4-phenylimidazole have demonstrated that a sterically accessible lone
electron pair provided by the heterocyclic nitrogen atom is required
for heme iron coordination (Poulos and Howard, 1987). Furthermore, type II binding studies of phenylpyridines, phenylimidazoles, and
pyridylimidazoles to cytochrome P450 in hepatic microsomes from
phenobarbital-induced rats showed the highest binding affinity for
compounds in which steric hindrance near the nitrogen was minimal
(Murray and Wilkinson, 1984). Therefore, phenylimidazoles promised to
be valuable probes for studying the molecular basis of differential
P450 inhibition.
This study identifies specific active-site residues critical for
differential sensitivity to inhibition of CYP2B4, CYP2B5, and CYP2B1 by
phenylimidazoles. The recent elucidation of the crystal structure of
rabbit CYP2C5 (Williams et al., 2000) provided an invaluable template
for construction of a new generation of homology models of the CYP2B
enzymes. These models and inhibition studies of active-site mutants
with 4-phenylimidazole revealed that residues 114, 294, 363, and 367 are critical for CYP2B4, CYP2B5, and CYP2B1 inhibition. The results
provide novel insight into residue-residue interactions in the active
site that confer differential inhibition sensitivity and catalytic
selectivity of the CYP2B enzymes.
| |
Experimental Procedures |
|---|
|
Top
Abstract
Introduction
Experimental Procedures
Results
Discussion
References
|
|---|
Materials.
Restriction endonucleases, Luria-Bertani Broth
and Terrific Broth media for bacterial growth were purchased from Life
Technologies (Grand Island, NY). The Escherichia coli strain
Topp3 was obtained from Stratagene (La Jolla, CA). Androstenedione,
resorufin, 7-benzyloxyresorufin, 7-pentoxyresorufin, NADPH, dimethyl
sulfoxide, dilauroyl-L-3-phosphatidylcholine, CHAPS, 
-aminolevulinic acid, isopropyl
-d-thiogalactopyranoside, 1-(2-trifluoromethylphenyl)imidazole and BioMax MR-1 film were purchased from Sigma Chemical Co. (St. Louis, MO). HEPES was obtained from Calbiochem Co. (La Jolla, CA).
[14C]Androstenedione was purchased from
DuPont-NEN (Boston, MA). Thin-layer chromatography plates [silica gel,
250 µm, Si 250PA (19C)] were obtained from J.T. Baker, Inc.
(Phillipsburg, NJ). Rat NADPH cytochrome P450 reductase was expressed
in E. coli as described previously (Harlow and Halpert,
1997). 1-Phenylimidazole, 4-phenylimidazole, 1-benzylimidazole,
2-methylimidazole, 2-phenylimidazole, and
1-(2,3,5,6-tetrafluorophenyl)imidazole were obtained from Aldrich
Chemical Co. (Milwaukee, WI). 4-(4-Chlorophenyl)imidazole was purchased
from Maybridge Chemical Company (Tintagel Cornwall, UK), and
1-(4-chlorophenyl)imidazole was obtained from Lancaster (Pelham, NH).
All other chemicals and supplies used were from standard sources.
Subcloning of P450 2B1 Mutants and Heterologous Expression.
cDNA encoding CYP2B1 V367L was previously subcloned into the pBC vector
yielding pBC2B1 V367L (He et al., 1994). For expression in E. coli, the cDNA 2B1 V367L was subcloned into pKK2B1 (He et al.,
1995). pBC2B1 V367L was digested with the unique restriction enzymes
HindIII and PstI. The appropriate fragment was
purified by GeneClean II (Bio 101, Vista, CA) and ligated into purified pKK2B1, which was also digested with HindIII and
PstI. To construct pKK2B1 L58F/I114F, plasmid
pSE2B2FF (L58F-I114F) (Strobel and Halpert, 1997)
was digested with PstI and KpnI, and the
resulting fragment was subsequently subcloned into pKK2B1. The triple
mutant CYP2B1 L58F-I114F-S294T was constructed by digesting pKK2B1
L58F-I114F and pKK2B1 S294T with BamHI. The fragment
containing the mutations of L58F-I114F was ligated into
phosphatase-treated pKK2B1 S294T and transformed into E. coli DH5
cells. All plasmids were identified by restriction
mapping and/or sequencing (373 XL ABI DNA sequencer; ABI, Norwalk, CT).
All other mutants tested in this study were constructed previously (He
et al., 1995, 1996; Szklarz et al., 1996; Strobel and Halpert, 1997).
E. coli Topp3 cells were transformed with pKK2B4, pKK2B5,
pKK2B1, and pSE2B2 wild-type and mutant plasmids. Conditions for
expression were as described earlier (John et al., 1994).
CHAPS-solubilized membranes were prepared as described previously (John
et al., 1994). Total P450 concentration was measured by reduced CO
difference spectra (Omura and Sato, 1964).
Spectral Binding Studies.
Difference spectra were recorded
on a Shimadzu UV-2600 spectrophotometer at 37°C. A solution of 0.4 µM CYP2B1 WT and CYP2B1 mutants in a buffer containing 100 mM MOPS
and 10% glycerol, pH 7.3, was prepared and divided into two 0.5-ml
quartz cuvets (1-cm path length), and a baseline was recorded between
350 and 500 nm. An aliquot of inhibitor in methanol was then added to
the sample cuvette, and the same amount of methanol was added to the reference cuvette. The difference spectra were obtained after the
system reached equilibrium (3 to 5 min). The spectral dissociation constants (KD) were obtained by fitting the
data to the equation for "tight binding" 
A = ((KD + [I0] + [E0]) 
((KD + [I0] + [E0]))2 4[E0][I0])1/2)
/ 2 when KD 
4 µM or to the
conventional equation A = [I][E0] /
(KD+[I]) (Copeland, 2000).
Inhibition Studies.
Inhibition studies were carried out with
7-benzyloxyresorufin, 7-pentoxyresorufin, or androstenedione as
substrates. Some of the mutants of CYP2B4 and CYP2B5 showed very low
activity for one or two of these substrates. To achieve the most exact
IC50 values, the substrate yielding the highest
activity was chosen in each case. Benzyloxyresorufin
O-debenzylase (BROD) activities were measured for CYP2B1 and
all CYP2B1 mutants, for CYP2B4, the CYP2B4 single mutants, and the
CYP2B5 F114I-T294S-V363I-A367V quadruple mutant. Activities of CYP2B5,
the CYP2B5 single mutants, and the CYP2B4 I114F-S294T-I363V-V367A
quadruple mutant were determined by the pentoxyresorufin
O-dealkylase (PROD) assay. Androstenedione was used to
investigate the inhibition of all double and triple mutants of CYP2B4
and CYP2B5. BROD, PROD, and androstenedione hydroxylase activities were
determined as described previously (He et al., 1995). Inhibitors were
added from a 100× methanol stock solution (final methanol
concentration was 0.7%). Control reactions without inhibitor
were performed by adding the same amount of methanol. The final
500-µl reaction mixture contained 20 pmol of P450 to determine BROD
activity (linear from 5-25 pmol) and 15 pmol of P450 to determine PROD
activity (linear from 5-17.5 pmol). Both substrates were used at a
concentration of 10 µM. Preliminary studies showed linearity
of resorufin formation up to 15 min incubation time at 37°C for BROD
activity and up to 7.5 min for PROD activity; thus the reactions were
stopped with methanol after 10 and 5 min, respectively. For the
androstenedione hydroxylase assay, 10 pmol of P450 were used in a final
100-µl reaction mixture. The reaction was stopped with
tetrahydrofuran after 15-min incubation and metabolites were resolved
on TLC plates (He et al., 1995). The half-maximal inhibitor
concentrations (IC50) were obtained from the mean
inhibition observed at five or more different inhibitor concentrations
in two separate determinations performed in duplicate.
IC50 values were calculated by linear regression
analysis of the degree of inhibition as a function of inhibitor
concentration (Augustinsson, 1948). The degree of inhibition was
expressed as the ratio of the reaction rates of uninhibited and
inhibited enzyme. The graph directly gives the IC50 value as the molar concentration of the
inhibitor when
V0/Vinhibitor = 2 (V0 = uninhibited reaction rate;
Vinhibitor = inhibited reaction rate). A
difference between IC50 values of at least 2-fold
was considered significant.
Computer Modeling.
Molecular models were constructed using
the InsightII software package (Homology/InsightII,
Discover_3/InsightII, Biopolymer/InsightII, and Docking/InsightII from
Molecular Simulations Inc., San Diego, CA). The CYP2B models were
constructed based on the crystal structure of CYP2C5 (pdb accession
number: 1dt6 on hold) (Cosme and Johnson, 2000). The sequences of
CYP2B1, CYP2B4, and CYP2B5 were obtained from SwissProt (accession
numbers P00176, P00178, and P12789, respectively). The sequence
alignment of CYP2C5, CYP2B1, CYP2B4, and CYP2B5 was done by GCG
(Wisconsin Package Version 10.0; PileUp; Genetics Computer Group,
Madison, WI). In the crystal structure of CYP2C5, the coordinates for
the N-terminal residues 1 to 30 and the F-G loop residues 212 to 222 were missing. Therefore, the models were constructed from residues 31 to 491. The segment between residues 212 and 222 was modeled based on the coordinates of CYP2C5 containing one of two alternative models for
density corresponding to the F-G loop (E.F.J., unpublished observations). The coordinates of residues 276 to 278, the only segment
not considered to be conserved, were generated using the random tweak
method in Homology/InsightII (Shenkin et al., 1987) (Fig. 2). The
coordinates of the conserved residues were assigned based on the
corresponding residues of CYP2C5 by Homology/InsightII. The heme group
was copied from CYP2C5 into the CYP2B models.
, 1992). The cvff force field encoded in InsightII was
used for the other part of the enzyme. First, splices between residues
211 and 212, 222 and 223, 275 and 276, and 278 and 279 were repaired by
Homology/InsightII to avoid steric hindrance in these junction regions.
All hydrogen atoms were minimized by fixing the heavy atoms of the 2B
enzymes. The side chains were minimized by fixing the backbone. A 25-Å
sphere and 3-Å surface layer of water were soaked into and around the
minimized protein with the SOAK function in Viewer/InsightII. Energy
minimization was performed again on the whole soaked enzyme. For the
2B5 mutants, the coordinates of the corresponding residues were changed
in the 2B5 3D model by Biopolymer/InsightII, and the resulting 2B5 mutants were minimized.
The quality of the models was checked by Prostat/InsightII, which
allows protein specific bond lengths, angles, and torsions to be
checked against the corresponding reference values. The cutoff used,
which represents the significant difference for bond length, bond
angle, and torsion from the reference value, is 5 S.D. For the 2B1
model, none of the bond distances, one bond angle, and five dihedral
angles were identified to have more than 5 S.D. For the 2B4 model, none
of the bond distances, three bond angles, and seven torsions had more
than 5 S.D. For the 2B5 model, none of the bond distances, five bond
angles, and 11 dihedral angles have more than 5 S.D.
4-Phenylimidazole was manually docked into the 3D models of CYP2B1,
CYP2B4, CYP2B5, and CYP2B5 mutants. The initial position of
4-phenylimidazole in the active site of these enzymes was determined from the position of 4-phenylimidazole in the active site of
P450cam after superimposing the heme groups. The
crystal structure of the
4-phenylimidazole-P450cam complex was obtained
from the Brookheaven Protein Databank (1PHF; Poulos and Howard, 1987),
where the distance between the nitrogen atom of 4-phenylimidazole (N)
and the heme iron (Fe) was 2.2 Å. Energy minimizations were performed on the 4-phenylimidazole CYP2B complexes. During the minimization process, the Fe-N distance was fixed at 2.2 Å to maintain the coordinate bond. For each CYP2B enzyme, possible orientations of the
imidazole ring and of the phenyl ring of 4-phenylimidazole were
examined separately. A series of initial orientations of the imidazole
ring were constructed where the ring was fixed every 30° in a 360°
rotation. Regardless of the initial orientation, molecular mechanics
minimization of each of these structures invariably placed the
imidazole ring in one of two positions that were related to each other
by a rotation of approximately 180° degrees. In one of these
positions, the phenyl ring points to a portion of the active site
distant from the key residues 114, 294, and 367, which is inconsistent
with experimental results and was disregarded. The other
position is shown in Fig. 3, A, B, and D. A series of initial
orientations of the phenyl ring was also constructed. The ring was
oriented every 30° in a 180° rather than 360° rotation because of
the C2 symmetry of the phenyl ring. Only one position was obtained in
each CYP2B enzyme after molecular mechanics minimization on these
initial structures, as shown in Fig. 3, A, B, and D.
| |
Results |
|---|
|
Top
Abstract
Introduction
Experimental Procedures
Results
Discussion
References
|
|---|
Inhibitor Selection and Basic Experimental Approach.
An
initial screening of CYP2B4 and CYP2B5 inhibition using different
phenylimidazole compounds revealed 4-phenylimidazole as a selective
inhibitor for CYP2B4. Inhibition of CYP2B4 activity by
4-phenylimidazole was 18-fold more potent than that of CYP2B5 activity
(Fig. 1A). The IC50
values of 1-, 2-, and 4-phenylimidazole for CYP2B4 were 0.9 µM, 1.8 mM, and 0.49 µM, respectively. These data are consistent with other
reports (Murray and Wilkinson, 1984) and showed that the affinity of
phenylimidazoles is largely dependent on the accessibility of the
nitrogen atom of the heterocycle, which is impaired with
2-phenylimidazole. Because 4-phenylimidazole showed significantly
different inhibition effects on CYP2B4 and CYP2B5 activity and the
highest binding affinity, further investigations of the molecular
requirements for inhibition were performed with 4-phenylimidazole.
|
). This finding reflects the dominance of the strong
coordinate bond between the heme iron of cytochrome P450 and the
nitrogen of the imidazole ring of the phenylimidazole compound in
determining inhibitor potency. To ensure that neither different
substrates nor different substrate concentrations had a significant
effect on inhibitor potency, IC50 values of
4-phenylimidazole for CYP2B4 WT and CYP2B5 WT were determined under a
variety of conditions. The IC50 values of
4-phenylimidazole for CYP2B4 WT determined by three assays differed
only insignificantly (BROD, IC50 = 0.49 µM;
PROD, IC50 = 0.46 µM; androstenedione
hydroxylase activity: IC50 = 0.3 µM) (Fig. 1B).
Similar results were observed for CYP2B5 WT when 7-pentoxyresorufin and
androstenedione were used as substrates (PROD,
IC50 = 8.4 µM; androstenedione hydroxylase activity, IC50 = 10.6 µM). Moreover, in a
separate experiment, the IC50 for inhibition of
CYP2B4 WT activity by 4-phenylimidazole was constant over a 10-fold
range of 7-benzyloxyresorufin concentrations (data not shown). Finally,
inhibition of CYP2B5 by 4-phenylimidazole using androstenedione as a
substrate indicated virtually noncompetitive inhibition, as shown in
Fig. 1C. Based on all these findings, IC50 values
were used in subsequent experiments as a measure of inhibitor potency.
Differential Sensitivity of CYP2B4 and 2B5 to Inhibition by
4-Phenylimidazole and Analysis of Active-Site Mutants.
To assess
the structural basis of differential sensitivity of CYP2B4 and CYP2B5,
inhibition studies were performed with active-site mutants. Based on
recent findings of the importance of residues 114, 294, 363, and 367 for substrate specificities of CYP2B4 and CYP2B5, mutants containing
alterations at these positions were tested. These mutants had been
previously constructed (He et al., 1996; Szklarz et al., 1996). Because
of low basal activities, inhibition of the CYP2B4 I114F and CYP2B5
F114I enzymes could not be determined.
|
|
Sensitivity of CYP2B1 to Inhibition by 4-Phenylimidazole and
Analysis of the Active Site Mutants.
It was of great interest to
ascertain whether the findings regarding the molecular requirements for
CYP2B4 and CYP2B5 inhibition could be extrapolated to rat CYP2B1.
Furthermore, the investigation of inhibitor sensitivity of CYP2B1
allowed us to analyze the importance of certain active-site residues
more carefully. With an additional substitution of the nonactive site
Leu 58 with Phe, the CYP2B1 I114F mutant expressed sufficiently to
obtain reasonable activity. The inhibition of CYP2B1 WT activity by
4-phenylimidazole resulted in a low IC50 value,
similar to that observed for CYP2B4 WT (Table 3). Spectral studies of 4-phenylimidazole
binding to CYP2B1 WT and selected mutants indicated a strong
correlation between KD values and
IC50 values, confirming the pivotal role of the
Fe-N bond for inhibition (Table 3). The tight type II binding of
4-phenylimidazole to CYP2B1 WT (KD = 0.3 ± 0.1 µM) was in strong contrast to the weak type I binding
of 2-phenylimidazole (KD= 1.5 ± 0.2 mM). Substitution of Val 363 with Ala or of Ser 294 with Ala or Thr
caused almost no change in sensitivity to inhibition. However, a Val
363-to-Leu substitution led to a decrease of almost 3-fold in the
IC50 value. Similar to the findings with CYP2B4,
substitution of Val 367 with Ala increased the
IC50 value 23-fold, whereas the mutation of this
residue to Leu decreased the IC50 value 6-fold
compared with the WT. Furthermore, the substitution of Leu 58 and Ile
114 with Phe increased the IC50 value 130-fold
compared with the
WT.1 Strikingly,
the replacement of Ser 294 with Thr in 2B1 L58F-I114F resulted in an
85-fold decrease in the IC50 value back to
wild-type levels. In conclusion, interesting similarities between the
results of CYP2B1 and CYP2B4 inhibition were observed. Besides, the
importance of residue 114 for sensitivity to inhibition the
construction of the mutants CYP2B1 L58F-I114F and CYP2B1
L58F-I114F-S294T suggested an interaction between residues 114 and 294, as noted in CYP2B5.
|
Modeling of CYP2B1, CYP2B4, CYP2B5, and Mutants and Docking of
4-Phenylimidazole into the Active Site.
The docking of
4-phenylimidazole into models constructed based on the crystal
structure of bacterial P450s (Szklarz et al., 1996) was uninformative.
Therefore, 2B models were constructed based on the coordinates of the
recently solved CYP2C5 crystal structure (Williams et al., 2000). The
sequence alignment used for modeling CYP2B1, CYP2B4, and CYP2B5 is
summarized in Fig. 2. This sequence
alignment was performed by GCG except for the insertion between 274Q
and 280N (QEN... N), which was changed to QE... NN after the
analysis of the CYP2C5 structure. This alteration allows room for the
insertion without disrupting the structure.
|
|
-stacking of the phenyl ring of Phe 114 with the phenyl moiety of
4-phenylimidazole (Fig. 4A). The additional substitution of Phe 114 to
Ile resulted in more space for 4-phenylimidazole and in movement of the
phenyl ring of 4-phenylimidazole toward residues 363 and 367 (Fig. 4B). Substitutions of Val 363 and Ala 367 with the larger residues Ile and
Val, as shown in Fig. 4C, decreased the distances of these residues to
4-phenylimidazole. Because of the additional replacement of Thr 294 with the smaller Ser, the phenyl ring of Phe 114 moved closer to Ser
294, and the phenyl moiety of 4-phenylimidazole moved closer to residue
363 (Fig. 4D). However, a significant alteration in the position of the
phenyl ring of 4-phenylimidazole was only observed when Phe 114, Val
363, and Ala 367 were simultaneously substituted with Ile, Ile, and
Val, respectively (Fig. 4E). Consequently, residues 363 and 367 are
located within 5 Å of 4-phenylimidazole. When all four residues were
simultaneously replaced by the corresponding residues of CYP2B4, the
distances between the residues from 4-phenylimidazole were similar to
those in the CYP2B4 WT model with the exception of Val 367, which was
1.2 Å farther away from 4-phenylimidazole than in the CYP2B4 WT model
(Fig. 4F). These detailed observations of CYP2B5 mutants with
4-phenylimidazole docked into the active site reveal an intriguing
interplay between active-site residues, which is consistent with the
complex experimental data, especially for residues 114 and 294. These
residue-residue interactions might strongly influence the potency of
inhibition of CYP2B5.
|
Identification of Phenylimidazole Derivatives with Enhanced
Differential Inhibition of CYP2B4 and CYP2B5.
An initial screen of
derivatives with substituents at various positions of the phenyl moiety
of the phenylimidazole compounds showed that all phenylimidazoles
tested were more potent inhibitors of CYP2B4 than of CYP2B5 activity. A
chlorine substituent at position 4 of the phenyl moiety of the
phenylimidazoles resulted in significantly increased potency of
inhibition of CYP2B4 (Table 4). CYP2B4
was 7.5-fold more sensitive to 1-(4-chlorophenyl)imidazole and
12.3-fold more sensitive to 4-(4-chlorophenyl)imidazole than to 1- and
4-phenylimidazole, respectively. Interestingly,
1-(4-chlorophenyl)imidazole was 3.7-fold less potent as an inhibitor of
CYP2B5 compared with 1-phenylimidazole. However,
4-(4-chlorophenyl)imidazole was twice as potent for CYP2B5 as the
parent compound 4-phenylimidazole. Whereas the
IC50 values of 1- and 4-phenylimidazole were 5- and 17-fold lower for CYP2B4 than for CYP2B5, the values for 1- and
4-(4-chlorophenyl)imidazole were 130- and 95-fold lower for CYP2B4 than
for CYP2B5. CYP2B1 showed not only similar active-site residues
involved in inhibition, but all tested phenylimidazole compounds
yielded IC50 values comparable with CYP2B4.
Interestingly, inhibition of the CYP2B4 quadruple mutant and the CYP2B5
quadruple mutant by 1-(4-chlorophenyl)imidazole exhibited the same
interconversion of sensitivity as observed for 4-phenylimidazole (data
not shown). Inhibition of the CYP2B4 quadruple mutant and the CYP2B5
quadruple mutant by 1-(4-chlorophenyl)imidazole yielded
IC50 values of 5.2 µM and 0.16 µM,
respectively. Studies of inhibition by 4-(4-chlorophenyl)imidazole of
the androstenedione hydroxylase activities in phenobarbital-induced rat
liver microsomes showed 50% inhibition of CYP2B1 activity at a 200 nM
inhibitor concentration, whereas CYP2A1, CYP3A1/2, and CYP2C activities were inhibited only in the micromolar range of
4-(4-chlorophenyl)imidazole (data not shown).
|
| |
Discussion |
|---|
|
Top
Abstract
Introduction
Experimental Procedures
Results
Discussion
References
|
|---|
This study investigated the structural basis of differential
inhibition by phenylimidazoles of CYP2B4, CYP2B5, and CYP2B1 using
active-site mutants and molecular models. Recent studies have shown
that the SRS residues 114, 294, 363, and 367 are critical for regio-
and stereospecificity of androstenedione hydroxylation and for the
oxidation of several other substrates by CYP2B4 and CYP2B5 (Szklarz et
al., 1996). In the present investigation, we have found that these
active-site residues are also responsible for differential sensitivity
to inhibition by 4-phenylimidazole. Moreover, the simultaneous
substitution of the four residues 114, 294, 363, and 367 in 2B5 by the
residues of CYP2B4 confers the sensitivity to inhibition of CYP2B4 on
CYP2B5 and vice versa. The same interconversion was observed for
androstenedione hydroxylase activity (He et al., 1996).
To elucidate the complex role of active-site residues in
inhibitor sensitivity, molecular models were used. Initially,
4-phenylimidazole was docked into the active site of CYP2B models based
on the bacterial P450s P450cam, P450 BM-3, and
P450terp (Szklarz et al., 1995). However, these models did not aid the
interpretation of our experimental data, especially those on CYP2B5.
Very recently, the first mammalian P450, CYP2C5, was crystallized
(Williams et al., 2000). The 2B enzymes belong to the same P450 family
as CYP2C5, and the sequence identity is high (51.5% for CYP2B1; 50.7%
for CYP2B4 and CYP2B5), as opposed to 15 to 20% with the bacterial
P450s. Therefore, the homology modeling of CYP2B1, CYP2B4, and CYP2B5
based on the crystal structure of CYP2C5 provides a significant
refinement compared with 2B models based on the bacterial P450s.
The first novel insight gained with the new 2B models was provided by
docking 4-phenylimidazole into CYP2B4 and from the analysis of
active-site mutants of CYP2B4. Both inhibition studies and molecular
modeling indicate that CYP2B4 residues 114 and 367 play a critical role
in sensitivity to 4-phenylimidazole (Table 1, Fig. 3A). A single Val
367-to-Ala substitution causes a 10-fold decrease in sensitivity to a
level comparable with that of CYP2B5. The Ile residue at position 114 may be of similar importance for the sensitivity to inhibition of 2B4.
The 3D model shows that residue 114 is as near as residue 367 to the
phenyl moiety of 4-phenylimidazole. Simultaneous substitution of Ile
114 with the bulkier aromatic residue Phe and of Ser 294 with Thr leads
to a significant decrease in sensitivity to inhibition, suggesting that
the side chain of phenylalanine may cause steric hindrance. The side
chain of position 114 was also observed to be crucial for the selective
androstenedione hydroxylation in 2B4 and 2B5 (Szklarz et al., 1996).
However, the interaction of the substrate with this residue could not
be explained by molecular modeling based on bacterial P450s. Residue
114 is located in the B'-C-loop, a segment known to differ
significantly between the bacterial P450s and the mammalian CYP2
family. Therefore, for the first time the crucial position of this
residue for inhibitor and substrate interactions could be explained.
Another remarkable and intriguing new feature was observed in the model
of CYP2B5 and the inhibition results with the active-site mutants.
Similar to CYP2B4, the side chain of residue 114 is also crucial for
the inhibition of CYP2B5 by 4-phenylimidazole. Docking of
4-phenylimidazole into the active site of CYP2B5 indicates that binding
of the compound may be impeded by steric hindrance from the phenyl ring
of Phe 114 (Fig. 3B). This finding may explain the significantly higher
IC50 value of 4-phenylimidazole for CYP2B5 compared with that of CYP2B4 (Table 2). In contrast to CYP2B4, none of
the single mutations in CYP2B5 could confer sensitivity to inhibition
similar to CYP2B4 WT. Interestingly, substitution of Thr 294 with Ser
causes a 10-fold increase in the IC50 value. Upon
alteration of residue 294, the phenyl ring of Phe 114 changes its
position, as indicated by comparison of the 3D models of this mutant
and CYP2B5 WT (Figs. 4A and 3B). Consequently, the - interaction
of the two parallel phenyl rings in CYP2B5 WT disappears. The
substitution of Phe 114 with Ile favors a higher sensitivity to
inhibition by removing the steric hindrance for 4-phenylimidazole. However, the IC50 value of CYP2B5 F114I-T294S
remains higher than that of the WT, perhaps because of the loss of the
- interaction (Table 2; Fig. 4B). The substitution of Val 363 and
Ala 367 with the larger residues Ile and Val, respectively, may enhance
the hydrophobic interactions with 4-phenylimidazole (Fig. 4C). However, in CYP2B5 mutants V363I-A367V and T294S-V363I-A367V, steric hindrance from Phe 114 could prevent the phenylimidazole compound from tighter binding. Subsequently, no significantly higher inhibition sensitivity for these two mutants was observed compared with the WT, despite the
increased hydrophobic interactions of Ile 363 and Val 367 with the
compound (Fig. 4, C and D; Table 2). Both the free movement of the
compound in the active site achieved by the substitution of Phe 114 with Ile and the hydrophobic interactions of the Ile 363 and Val 367 with the compound are necessary to significantly increase the
sensitivity to 4-phenylimidazole (Fig. 4E; Table 2). However,
substitution of all four residues by the corresponding residues of
CYP2B4 is required to achieve an inhibition sensitivity similar to that
of CYP2B4 (Fig. 4F; Table 2). The experimental and modeling results
with CYP2B5 and its mutants demonstrated that the exact architecture of
the inhibitor binding site is determined not only by individual
contributions from key inhibitor contact residues, but also by
residue-residue interactions. This intriguing interplay of active-site
residues, especially involving residues 114 and 294, was previously
inferred from a study of androstenedione hydroxylase specificity of
CYP2B4 and CYP2B5 (He et al., 1996). As the model available at that
time was based on bacterial P450s, the mutagenesis data could not fully
be explained.
A further interesting point is the extrapolation of the findings above
to rat CYP2B1. Previous investigations of the CYP2B enzymes have
already shown that residues important for activities of CYP2B4 and
CYP2B5 are also essential for other members of the 2B family (He et
al., 1994; Szklarz et al., 1995; Lewis and Lake, 1997). Inhibition
studies and the 3D models demonstrated striking similarities between
CYP2B4 and CYP2B1 (Tables 1, 3, and 4). Val 367 plays a major role for
sensitivity to inhibition by 4-phenylimidazole in CYP2B1 as well as in
CYP2B4. In CYP2B1 and CYP2B4, the sensitivity to inhibition decreases
significantly upon replacement of Val 367 with the smaller Ala residue.
To further investigate the importance of residue 367, we tested CYP2B1
V367L, which exhibited a significant lower IC50
value than the wild-type enzyme, presumably because of tighter
hydrophobic binding with the longer side chain of Leu. These findings
were supported by the 3D model of CYP2B1, which showed that Val 367 is
substantially closer to 4-phenylimidazole than any other of the four
investigated residues (Fig. 3D). As observed for CYP2B4 and CYP2B5, the
side chain of residue 114 may be also of great importance for the
sensitivity to inhibition in CYP2B1. Mutant I114F mutant did not
express. To achieve reasonable expression and activity, CYP2B1
L58F-I114F was constructed. Consistent with the results obtained for
CYP2B4 and CYP2B5, this mutant showed a dramatic decrease in
sensitivity. However, the additional Ser 294-to-Thr substitution
restored sensitivity to inhibition. This finding is consistent with the
114-294 residue interaction observed in CYP2B5.
Interestingly, substantially higher selectivity for CYP2B4 over CYP2B5 was achieved with a chlorine substituent on the phenyl moiety. These findings suggest 1- and 4-(4-chlorophenyl)imidazole as useful inhibitors to distinguish between CYP2B4 and CYP2B5. As already observed for 4-phenylimidazole, CYP2B1 yielded about the same IC50 values as CYP2B4 for all tested phenylimidazole compounds. Although the orientation of the phenyl moiety of 4-phenylimidazole in the 3D models of CYP2B4 and CYP2B1 is different, the similar sensitivity to inhibition of these two enzymes might be caused by the same residues at positions 114 and 367.
In conclusion, the results of the present investigation reveal that the
active-site residues 114, 294, 363, and 367 in CYP2B4, 2B5, and 2B1 are
crucial for reversible inhibition, as already observed for substrate
specificity and mechanism-based inactivation (He et al., 1992, 1994,
1995; Strobel et al., 1999). Furthermore, molecular models based on the
crystal structure of CYP2C5 are of great value to explain the complex
molecular basis of reversible inhibition by phenylimidazoles. These
findings may contribute to a considerably better understanding of the
CYP2B enzymes and to the design of a selective human CYP2B6 inhibitor.
Knowledge of the residues responsible for inhibition and of
residue-residue interactions may also help to modulate chemical
structures to avoid inhibition and resulting drug-drug interactions.
Moreover, because the five main CYP2 subfamilies possess overlapping
structure-function and substrate specificities (Lewis, 1998), the
findings presented here may apply to other members of the CYP2 family.
| |
Acknowledgments |
|---|
We thank Dr. Tammy Domanski for kindly providing several CYP2B1 mutants.
| |
Footnotes |
|---|
Received August 18, 2000; Accepted November 2, 2000
1 A very similar effect was observed in CYP2B2 when Leu 58 and Ile 114 were replaced simultaneously with Phe residues. The IC50 value of 4-phenylimidazole increased from 0.8 µM for the wild-type enzyme to 44 µM for CYP2B2 L58F-I114F.
2 The distances reported under Results are those between the nearest atoms of the residue and 4-phenylimidazole.
This work was supported by AstraZeneca and National Institutes of Health Grants ES03619 (J.R.H.), GM31001 (E.F.J.), and Center Grant ES06676. An abstract of this work appeared in FASEB J 14:158.
Send reprint requests to: Dr. Margit Spatzenegger, Department of Pharmacology and Toxicology, University of Texas Medical Branch at Galveston, 301 University Blvd., Galveston, TX 77555-1031. E-mail: maspatze{at}utmb.edu
| |
Abbreviations |
|---|
P450, cytochrome P450; SRS, substrate recognition site; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid; BROD, benzyloxyresorufin O-debenzylase; PROD, pentoxyresorufin O-dealkylase; WT, wild-type.
| |
References |
|---|
|
Top
Abstract
Introduction
Experimental Procedures
Results
Discussion
References
|
|---|
This article has been cited by other articles:
|
|
 |
|
  R. Stadel, J. Yang, J. W. Nalwalk, J. G. Phillips, and L. B. Hough High-Affinity Binding of [3H]Cimetidine to a Heme-Containing Protein in Rat Brain Drug Metab. Dispos., March 1, 2008; 36(3): 614 - 621. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. K. Muralidhara, S. Negi, C. C. Chin, W. Braun, and J. R. Halpert Conformational Flexibility of Mammalian Cytochrome P450 2B4 in Binding Imidazole Inhibitors with Different Ring Chemistry and Side Chains: SOLUTION THERMODYNAMICS AND MOLECULAR MODELING J. Biol. Chem., March 24, 2006; 281(12): 8051 - 8061. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Kumar, C. S. Chen, D. J. Waxman, and J. R. Halpert Directed Evolution of Mammalian Cytochrome P450 2B1: MUTATIONS OUTSIDE OF THE ACTIVE SITE ENHANCE THE METABOLISM OF SEVERAL SUBSTRATES, INCLUDING THE ANTICANCER PRODRUGS CYCLOPHOSPHAMIDE AND IFOSFAMIDE J. Biol. Chem., May 20, 2005; 280(20): 19569 - 19575. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H.-L. Lin, U. M. Kent, H. Zhang, L. Waskell, and P. F. Hollenberg The Functional Role of Threonine-205 in the Mechanism-Based Inactivation of P450 2B1 by Two Ethynyl Substrates: The Importance of the F Helix in Catalysis J. Pharmacol. Exp. Ther., December 1, 2004; 311(3): 855 - 863. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. E. Scott, M. A. White, Y. A. He, E. F. Johnson, C. D. Stout, and J. R. Halpert Structure of Mammalian Cytochrome P450 2B4 Complexed with 4-(4-Chlorophenyl)imidazole at 1.9-A Resolution: INSIGHT INTO THE RANGE OF P450 CONFORMATIONS AND THE COORDINATION OF REDOX PARTNER BINDING J. Biol. Chem., June 25, 2004; 279(26): 27294 - 27301. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. E. Scott, Y. A. He, M. R. Wester, M. A. White, C. C. Chin, J. R. Halpert, E. F. Johnson, and C. D. Stout From The Cover: An open conformation of mammalian cytochrome P450 2B4 at 1.6-A resolution PNAS, November 11, 2003; 100(23): 13196 - 13201. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H.-L. Lin, H. Zhang, L. Waskell, and P. F. Hollenberg Threonine-205 in the F Helix of P450 2B1 Contributes to Androgen 16{beta}-Hydroxylation Activity and Mechanism-Based Inactivation J. Pharmacol. Exp. Ther., August 1, 2003; 306(2): 744 - 751. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Kumar, E. E. Scott, H. Liu, and J. R. Halpert A Rational Approach to Re-engineer Cytochrome P450 2B1 Regioselectivity Based on the Crystal Structure of Cytochrome P450 2C5 J. Biol. Chem., May 2, 2003; 278(19): 17178 - 17184. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
) and CYP2B5 (
) by
4-phenylimidazole. B, inhibition of CYP2B4 determined with three
substrates. The IC50 value was determined by the inhibition
of BROD (
)
activity as described under Experimental Procedures. The
lines shown were generated by linear regression (r2 > 0.96). The IC50 values are given as the micromolar
concentration of 4-phenylimidazole when
V0/V4-P = 2. C, Lineweaver-Burk plot of
the inhibition of the androstenedione 16
, 20 µM
4-phenylimidazole.
