![[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. 304, Issue 1, 477-487, January 2003
Department of Pharmacology and Toxicology, University of Texas Medical Branch, Galveston, Texas (M.S., H.L., Q.W., J.R.H.); and Department of Physiology and Pharmacology, Oregon Health and Science University, Portland, Oregon (A.D., D.R.K.)
 |
Abstract |
|---|
 Top
 Abstract
 Introduction
Materials and Methods
Results
Discussion
References
|
|---|
Human CYP2B6 and CYP2E1 were used to investigate the extent to which differential substrate selectivities between cytochrome P450 subfamilies reflect differences in active-site residues as opposed to distinct arrangement of the backbone of the enzymes. Reciprocal CYP2B6 and CYP2E1 mutants at active-site positions 103, 209, 294, 363, 367, and 477 (numbering according to CYP2B6) were characterized using the CYP2B6-selective substrate 7-ethoxy-4-trifluoromethylcoumarin, the CYP2E1-selective substrate p-nitrophenol, and the common substrates 7-ethoxycoumarin, 7-butoxycoumarin, and arachidonic acid. This report is the first to study the active site of CYP2E1 by systematic site-directed mutagenesis. One of the most intriguing findings was that substitution of CYP2E1 Phe-477 with valine from CYP2B6 resulted in significant 7-ethoxy-4-trifluoromethylcoumarin deethylation. Use of three-dimensional models of CYP2B6 and CYP2E1 based on the crystal structure of CYP2C5 suggested that deethylation of 7-ethoxy-4-trifluoromethylcoumarin by CYP2E1 is impeded by van der Waals overlaps with the side chain of Phe-477. Interestingly, none of the CYP2B6 mutants acquired enhanced ability to hydroxylate p-nitrophenol. Substitution of residue 363 in CYP2E1 and CYP2B6 resulted in significant alterations of the metabolite profile for the side chain hydroxylation of 7-butoxycoumarin. Probing of CYP2E1 mutants with arachidonic acid indicated that residues Leu-209 and Phe-477 are critical for substrate orientation in the active site. Overall, the study revealed that differences in the side chains of active-site residues are partially responsible for differential substrate selectivities across cytochrome P450 subfamilies. However, the relative importance of active-site residues appears to be dependent on the structural similarity of the compound to other substrates of the enzyme.
| |
Introduction |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
The
structural basis of substrate selectivity of individual cytochromes
P450 is an issue of crucial importance for drug development and
toxicology. Site-directed mutagenesis and homology modeling have become
essential tools in understanding differential cytochrome P450
functions. Over the past decade, a number of studies especially of P450
family 2 enzymes have elucidated the role of active-site residues in
substrate recognition (reviewed in Domanski and Halpert, 2001
). These
active-site residues have counterparts in the recently elucidated
crystal structure of CYP2C5 (Williams et al., 2000). However, the
architecture of the active site may not be the only factor to determine
substrate selectivity. Residues lining the substrate access channel may
be of equal importance. Several studies have identified residues in
CYP2B6, CYP2B4, and CYP2C9 that may be part of a substrate access
channel and contribute to substrate selectivity (He et al., 1996;
Ibeanu et al., 1996; Domanski et al., 1999). In addition, differential
substrate selectivities might be due to spatial divergence of helices
and 
-sheets, as has been shown for the crystal structures of CYP102
and CYP2C5 (Williams et al., 2000).
The role of active-site residues in substrate selectivity within one
subfamily has been studied very thoroughly, especially in the case of
CYP2A, 2B, and 2C enzymes (Goldstein et al., 1994; He et al., 1996;
Negishi et al., 1996). However, elucidation of differential substrate
selectivity across subfamilies has been neglected so far with the
exception of one study of progesterone hydroxylation by CYP2B1 and
CYP2A4 (Luo et al., 1994). Examination of the function of human CYP2B6
and CYP2E1 revealed distinct but partially overlapping substrate and
inhibitor selectivities, which provide an excellent opportunity for
further analysis.
Human CYP2B6 has proven to be increasingly important for drug
metabolism in the last 5 years. The enzyme metabolizes about 3% of the
drugs in clinical use (Rendic and Di Carlo, 1997), such as
cyclophosphamide, ifosfamide (Roy et al., 1999), nevirapine (Riska et
al., 1999), efavirenz (Christ et al., 1997), bupropion (Faucette et
al., 2000), and S-mephenytoin (Heyn et al., 1996). Furthermore, 20- to 250-fold individual variability (Stresser and
Kupfer, 1999) and ethnic differences (Kim et al., 1997) in CYP2B6
expression have been observed that might contribute to clinically
significant variability in response to and toxicity of CYP2B6-dependent
substrates. Compared with other members of the CYP2B subfamily, human
CYP2B6 hydroxylates testosterone and androstenedione at a much lower
rate (Domanski et al., 1999). A study of the structural basis for
functional differences between human CYP2B6 and rat CYP2B1 suggested
that residues Phe-107 and Leu-363 are critical for CYP2B6 activity
(Domanski et al., 1999). Recently, N-terminal truncation and
modification of CYP2B6 were found to substantially increase
heterologous expression, which greatly facilitates structure-function
studies (Scott et al., 2001).
Human CYP2E1 is responsible for the biotransformation of a large number
of low molecular weight compounds of toxicological importance, such as
chloroform and vinyl chloride (Koop, 1992). In addition, ethanol and
halogenated anesthetics such as enflurane and halothane are primarily
metabolized by CYP2E1 (Garton et al., 1995). Analysis of the structural
requirements of the enzyme has suggested that it has a smaller active
site than other P450 family 2 enzymes (Wang et al., 1995). However,
despite the toxicological and clinical importance of CYP2E1, no
systematic mutagenesis studies have been performed until now. The only
amino acid residue investigated for its role in substrate interaction
is Thr-303 (Moreno et al., 2001). In addition to substances of low
molecular weight, this enzyme is also able to catalyze rather large and
negatively charged fatty acids, such as lauric and arachidonic acid
(Laethem et al., 1993; Adas et al., 1999). It has been suggested that
the hydrocarbon end of the fatty acid binds to the substrate binding
site, leaving the carboxylic acid group outside of a proposed substrate
access channel (Wang et al., 1995).
This report is the first to investigate systematically differential substrate selectivity between two P450 subfamilies. The active sites of both P450 enzymes were compared on the basis of the crystal structure of CYP2C5 using common substrates and specific substrates to determine the extent to which active-site residues are responsible for the distinct catalytic activities of CYP2B6 and CYP2E1. The results have important ramifications for prediction of differential substrate selectivities using models based on the same template.
| |
Materials and Methods |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
Materials.
Oligonucleotide primers were obtained
from the University of Texas Medical Branch Molecular Biology Core
Laboratory (Galveston, TX) or from Sigma Genosys (Woodlands, TX). The
Expand PCR kit was purchased from Roche Diagnostics (Indianapolis, IN).
T4 DNA ligase was purchased from Invitrogen (Carlsbad, CA). Restriction enzymes were obtained from New England Biolabs (Beverly, MA). The
GeneClean kit was purchased from Qbiogene Inc. (Carlsbad, CA). The
Escherichia coli strains JM109 and TOPP3 were obtained from
Stratagene (La Jolla, CA). Luria-Bertani and Terrific Broth media were
purchased from Invitrogen. Ni2+-NTA affinity
resin was obtained from QIAGEN (Valencia, CA). NADPH, NADP (grade III),
-ALA, isopropyl -D-thiogalactoside,
glucose 6-phosphate, glucose-6-phosphate dehydrogenase (type XII),
p-nitrophenol (pNP), and 4-nitrocatechol were obtained from
Sigma-Aldrich (St. Louis, MO). 7-Ethoxy-4-trifluoromethyl-coumarin
(7-EFC) and 7-hydroxy-4-trifluoromethylcoumarin (7-HFC) were purchased
from Molecular Probes (Eugene, OR). 7-Butoxycoumarin, 7-(2-hydroxy)butoxycoumarin, 7-(3-hydroxy)butoxycoumarin, and 7-(4-hydroxy)butoxycoumarin were synthesized by Dr. E. Mash (Department of Chemistry, University of Arizona, Tucson, AZ). 7-Ethoxycoumarin (7-EC) and 7-hydroxycoumarin (7-HC) were purchased from Aldrich Chemical Co. (Milwaukee, WI).
[1-14C]Arachidonic acid (AA) was purchased from
American Radiolabeled Chemicals (St. Louis, MO). HEPES was purchased
from Calbiochem (San Diego, CA). Rat NADPH-cytochrome P450 reductase
and cytochrome b5 from rat liver were
prepared as described earlier (Harlow and Halpert, 1997). All other
chemicals and supplies used were from standard sources.
N-Terminal Truncation/Modification and Addition of C-Terminal His
Tag to CYP2E1 and Construction of Reciprocal CYP2B6 and CYP2E1
Mutants.
CYP2B6 with N-terminal truncation/modification and a
C-terminal 4xHis tag (CYP2B6dH) was constructed previously (Scott et al., 2001). Analogous to CYP2B6dH, the coding region for human CYP2E1dH
with N-terminal truncation and a C-terminal 4xHis tag was generated by
a single PCR reaction. The plasmid pGEM 4-h2E1, kindly provided by Dr.
M. Ingelman-Sundberg, was used as a template. The forward (sense)
primer contained the 5' truncation including an NcoI site
and 21 bases of native sequence starting at the PPGP-region. The
reverse (antisense) primer contained 21 bases of the C-terminal native
sequence followed by 4 histidine codons, the stop codon, and a unique
HindIII site. The resultant PCR product was run on an
agarose gel, and the appropriately sized DNA band was excised and
purified using GeneClean. The CYP2E1dH product was digested with
NcoI and HindIII and ligated into a similarly cut
pKK233-2 plasmid (Pharmacia, Peapack, NJ). The reciprocal mutants of
CYP2B6dH and CYP2E1dH were constructed by overlap extension PCR.
Changes of active-site residues are shown in Table
1.2
CYP2B6dH or CYP2E1dH were used as the template with primers A and C in
one reaction and primers B and D in a second reaction. Primers B and C
contained the desired mutation and generally overlapped by 18 to 20 bp.
The products of the AC and BD PCR were run on an agarose gel. The
correct size DNA was excised, purified using GeneClean, and used as the
template in a second PCR with primers A and D. The full-length product
was run on and excised from an agarose gel and purified using
GeneClean. The constructs of the CYP2B6dH mutants except V477F were
digested with AatII and HindIII. CYP2B6dH V477F
was digested with AatII and EcoRV. The correct size fragments were identified on an agarose gel, excised, purified using GeneClean, and ligated into the
AatII/HindIII or
AatII/EcoRV pKK2B6dH construct. The constructs of
the CYP2E1dH single mutants were digested with NcoI and
HindIII and ligated into similarly cut pKK233-2. For the
construction of the double mutant CYP2E1dH L367V-F477V, 2E1dH F477V was
used as a template. The same primers and restriction enzymes were used
as for CYP2E1dH L367V. All coding regions were fully sequenced to
verify the intended mutations and to detect any unintended mutations
(Protein Chemistry Laboratory, University of Texas Medical Branch,
Galveston, TX). The primers for each construct are shown in Table
2.
|
|
Expression and Purification of CYP2B6dH, CYP2E1dH, and
mutants.
CYP2B6dH and mutants and CYP2E1dH and mutants were
expressed in JM109 and TOPP3 cells, respectively. The cultures were
induced as described previously (Scott et al., 2001). Cultures were
grown for 48 h (CYP2B6dH and mutants) or 72 h (CYP2E1dH and
mutants) at 30°C and harvested by spinning at 3840g for 15 min at 4°C. Preparation of solubilized membranes and P450
purification were performed with modifications as described previously
(Scott et al., 2001). The cell pellets were resuspended in 4% of the
original volume, 20 mM potassium phosphate, 20% glycerol, 10 mM ME,
pH 7.4, and lysozyme was added to 0.2 mg/ml. The suspension was stirred for 60 min at 4°C. After the addition of an equal volume of cold water the stirring was continued for 10 min. The suspension was ultracentrifuged at 113,000g for 30 min at 4°C, and the
resulting cell pellet was resuspended in 0.5 M potassium phosphate and
sonicated on ice. The samples were incubated on ice with gentle shaking for 60 min and ultracentrifuged at 113,000g for 30 min at
4°C. The supernatant contained the truncated P450s and was used for further purification. Sodium cholate (0.5%) and 20 mM ME were added
to the supernatant. The sample was loaded on a
Ni2+-NTA column equilibrated with 500 mM buffer A
(potassium phosphate, 20% glycerol, 20 mM ME, 0.5% sodium cholate,
7.5 mM imidazole, pH 8). For CYP2B6dH and mutants the column was washed
as follows: 1) 500 mM buffer A and 7.5 mM imidazole; 2) 20 mM buffer A
and 15 mM imidazole; 3) 20 mM buffer A and 25 mM imidazole; and 4) 20 mM buffer A and 40 mM imidazole. All washing steps were performed with
0.1% sodium cholate and 20% glycerol at pH 8. For CYP2E1dH and
mutants, the column was washed similar to CYP2B6dH and mutants, but no
sodium cholate was added at steps 2, 3, and 4. CYP2B6dH and mutants
were eluted with 20 mM buffer A, 20% glycerol, 0.1% sodium cholate,
500 mM imidazole, and 0.5 M NaCl at pH 7.4 and dialyzed twice against 4 liters of 10 mM potassium phosphate, 20% glycerol, pH 7.4. CYP2E1dH
and mutants were eluted with 100 mM buffer A, 20% glycerol, 500 mM
imidazole, and 1 M NaCl at pH 7.4 and dialyzed twice against 4 liters
of 100 mM potassium phosphate, 20% glycerol, pH 7.4.
Enzymatic Assays.
Assay conditions were adapted for the
truncated enzymes to achieve high activity (Scott et al., 2001).
Purified P450 enzymes were reconstituted with rat NADPH-P450 reductase
and cytochrome b5 in a 1:4:2 ratio
without the addition of dilauroylphosphatidylcholine. 7-EFC
deethylation, 7-EC deethylation, and 7-butoxycoumarin oxidation were
carried out in HEPES buffer (50 mM HEPES, pH 7.6, 15 mM
MgCl2, 0.1 mM EDTA) and started by the addition
of NADPH (1 mM final concentration). The 7-EFC and 7-EC deethylation
assays were performed with a protein concentration of 2.5 pmol or 10 pmol in a final volume of 100 µl. The reaction was stopped after 10 min of incubation at 37°C (linear up to 20 min) with 50 µl of 20%
trichloroacetic acid, and product formation was determined
fluorometrically (Domanski et al., 1999).
. Reconstitutions were
carried out as described for 7-EC and 7-EFC, except they contained 20 pmol of P450. Substrate was added at a final concentration of 300 µM
and the reaction performed in a final volume of 1 ml.
pNP hydroxylation to 4-nitrocatechol was performed with a protein
concentration of 30 pmol in 1 ml final volume in 100 mM potassium
phosphate, pH 7.4. The reaction was stopped after 10 min of incubation
at 37°C (linear up to 15 min) by the addition of 300 µl of 20%
trichloroacetic acid. After the addition of 100 µl of 10 M NaOH
4-nitrocatechol was determined spectrophotometrically at 520 nm (Dicker
et al., 1990).
AA metabolism was monitored with modifications as described previously
(Laethem et al., 1993). The assay was performed with a cytochrome P450
concentration of 50 pmol in 0.5 ml final volume in 50 mM HEPES buffer,
pH 7.6, and 10 mM magnesium chloride. The substrate concentration was
7.2 µM (0.20 µCi). The NADPH-generating system consisted of 0.5 mM
NADP and 10 mM glucose 6-phosphate. After preincubation for 3 min at
37°C, the reactions were initiated by the addition of 3 units of
glucose-6-phosphate dehydrogenase, and were quenched after 10 min by
acidification with 10 µl of 6.7% (v/v) formic acid. Each reaction
was transferred to a new vial, the substrate concentration was
confirmed by counting the radioactivity of a 10-µl aliquot, and the
mixture was extracted twice with 2 volumes of ethyl acetate. The
solvent was removed in vacuo and the metabolites separated as
previously described (Laethem et al., 1993). The dried extracts were
dissolved in 90 µl of 50% acetonitrile containing 0.1% acetic acid.
The reaction products were separated on a Beckman model 344 HPLC
equipped with a C18-microsorb column (4.6 × 250 mm, 5 µm; Rainin, Woburn, MA) and a Radiomatic FLO-ONE A250
radiochemical detector (Tampa, FL), using a linear gradient (1.25% per
min) from 50% acetonitrile in water containing 0.1% acetic acid to
0.1% acetic acid in acetonitrile at a flow rate of 1 ml/min.
Other Methods.
Protein concentration was determined by the
bicinchoninic acid method with bovine serum albumin as standard
(Pierce, Rockford, IL). Cytochrome P450 was determined using the
reduced CO difference spectrum (Omura and Sato, 1964).
Computer Modeling.
Molecular models were constructed using
the InsightII software package (Homology/InsightII,
Discover_3/InsightII, Biopolymer/InsightII, Builder/InsightII, and
Docking/InsightII from Molecular Simulations Inc., San Diego, CA). The
sequences of CYP2B6 and 2E1 were obtained from SwissProt (accession
numbers P20813 and P05181, respectively). The sequence alignment of
CYP2C5, CYP2B6, and CYP2E1 was done by GCG (Wisconsin Package version
10.0; PileUp; Genetics Computer Group, Madison, WI) (Fig.
1). The model of CYP2B6 was developed
recently (Wang and Halpert, 2002). The CYP2E1 model was constructed
based on the crystal structure of CYP2C5 recovered from the Brookhaven
Protein Data Bank (http://www.rcsb.org/pdb/) (pdb accession number:
1dt6 on hold) (Williams et al., 2000). 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 model was
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
(Spatzenegger et al., 2001). The coordinates of residues 276 to 278, the only segment not considered to be conserved, were generated by
searching the Protein Data Bank to find the regions of proteins that
meet the geometric criteria of the loop. 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 CYP2B6 and CYP2E1 models.
|
, 1992). The consistent valence 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 CYP2E1. 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 CYP2E1 mutants, the coordinates of the
corresponding residues were changed in the CYP2E1 three-dimensional
model by Biopolymer/InsightII, and the resulting CYP2E1 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 2E1
model, none of the bond distances, nine bond angle, and nine dihedral
angles were identified to have more than 5 S.D.
The structures of 7-EC and 7-EFC were constructed using the Builder
module of the InsightII modeling package. Energy minimization and
molecular dynamics simulations were carried out with the Discover_3 program, using the consistent valence force field. 7-EC and 7-EFC were
manually docked into the three-dimensional models of CYP2B6, CYP2E1,
and CYP2E1 mutants in a reactive binding orientation, leading to
deethylation. The initial oxidation step involved hydrogen abstraction
at the side chain carbon bonded to the oxygen of the ethoxy group.
Therefore, the C1 atom of the ethyl chain was placed 3.7 Å from ferryl
oxygen, with one of the hydrogen atoms bonded to C1 directed toward
ferryl oxygen (C-H-ferryl O angle of 180°) to promote hydrogen bond
formation. Energy minimization was performed on the substrate enzyme
complexes. During the minimization process, the carbon atom to be
oxidized and the hydrogen atom to be abstracted were fixed, while the
rest of the substrate molecule, along with the side chains of protein
residues within 5 Å of the substrate, were allowed to move. For each
of the enzymes, possible orientations of the benzofuran ring were
examined separately. The substrate position was obtained in each of the
enzymes after molecular mechanics minimization on these initial structures.
Data Analysis.
Estimates of
Km or
S50 and
Vmax values were obtained by fitting
the data to the Michaelis-Menten or Hill equation, using SigmaPlot (Jandel Scientific, San Rafael, CA). Data are the average of two experiments performed in duplicate. The coefficient of variation for
all data were 
20%.
| |
Results |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
Protein Expression, Purification, and Initial Characterization. The wild-type enzymes and reciprocal single mutants were generated and expressed in E. coli. Of these, CYP2B6dH M103A and CYP2E1dH D294S did not produce hemeprotein. CYP2B6dH and mutants were expressed in E. coli JM109 cells, because expression levels were more consistent than in E. coli TOPP3 and MV1304 cells. Typical expression levels of CYP2B6dH and mutants ranged from 40 to 120 nmol/l. The overall yield after purification was about 40%, and the average specific content of CYP2B6dH and mutants was about 7 nmol of P450/mg of protein.
CYP2E1dH and mutants were expressed in E. coli TOPP3 cells because of higher and more consistent expression levels than in E. coli JM109, MV1304, and DH5
cells. Expression levels
between 200 and 1000 nmol of P450/liter were achieved. Yields after
purification of CYP2E1dH and mutants were between 45 and 90% with an
average specific content of 9 nmol of P450/mg of protein. CYP2B6dH,
CYP2E1dH, and mutants were expressed over different time courses
(12-72 h), and conditions of purification were modified to improve
yield and purity. CYP2B6dH and CYP2E1dH and mutants showed the highest yield and purity when the protein was applied to the column in 0.5%
sodium cholate. To avoid precipitation of CYP2E1dH and mutants, elution
and subsequent dialysis were performed in 100 mM potassium phosphate.
Differential Substrate Selectivities of CYP2B6dH and CYP2E1dH.
Initial characterization of CYP2B6dH and CYP2E1dH was performed with
three substrates at a single substrate concentration. In recent
studies, extensive comparisons of 7-EFC deethylation by CYP2B6 versus
CYP2B6dH (Scott et al., 2001) and of testosterone, 7-EFC, and
7-benzyloxyresorufin oxidation by CYP2B1 versus CYP2B1dH (EE Scott, YQ
He, and J Halpert, submitted for publication) showed very similar
Km values for the wild-type and
truncated enzymes. Therefore, truncation does not seem to affect the
affinity of the enzymes for substrates. Furthermore, initial
experiments in this study with preparations of the same enzyme with
different specific contents showed no differences in overall enzymatic activity.
; Code et al., 1997). 7-EC was used to characterize the
activity of both enzymes. Kinetic studies of 7-EC deethylation revealed
similar Vmax values for CYP2B6dH and
CYP2E1dH. Both enzymes showed similar sigmoidal kinetic behavior with a
S50 value about 2-fold higher for
CYP2B6dH than for CYP2E1dH (Table 3). For
7-EFC deethylation, the Km and
Vmax values for CYP2B6dH were found to
be 8.5 µM and 14 nmol/min/nmol of P450, respectively. Because of low
activity, the kinetic parameters for 7-EFC deethylation by CYP2E1dH
could not be determined. pNP hydroxylation by CYP2E1dH showed
Km and
Vmax values of 70 µM and 33 nmol/min/nmol of P450, respectively. In this case, the kinetic parameters for CYP2B6dH could not be determined because of very low pNP
hydroxylation activity.
|
7-EFC Deethylation Assays.
Each CYP2B6dH and CYP2E1dH mutant
was tested for 7-EFC deethylation activity. All CYP2B6dH mutants showed
significantly different kinetic behavior from the wild-type enzyme
(Table 4). I209L and S294D displayed a
4-fold higher Km value, and the
Vmax value for S294D was decreased
about 13-fold. V477F exhibited a 2-fold increase in the
Vmax value and a similar
Km value compared to CYP2B6dH. Interestingly, calculation of
Vmax/Km
indicated that V477F also had an enhanced efficiency compared with the
wild-type enzyme. 7-EFC deethylation by L363V showed positive
cooperativity, whereas deethylation by V367L displayed negative
cooperativity. These data suggested that all the residues investigated
might play a role in CYP2B6-selective 7-EFC deethylation.
|
|
|
7-EC Deethylation Assays.
Figure
4 illustrates the 7-EC deethylation
activity of the CYP2B6dH and the CYP2E1dH mutants at a single
saturating concentration of 300 µM. Of the CYP2B6dH mutants (Fig.
4A), V477F showed the highest activity. V367L displayed about the same
activity as CYP2B6dH, and I209L, S294D, and L363V all exhibited a
substantial decrease in activity.
|
Docking of 7-EC and 7-EFC into the Active Site of CYP2B6, CYP2E1,
and CYP2E1 F477V.
Three-dimensional models of CYP2B6 and CYP2E1
were used to analyze the differential substrate selectivity with 7-EC
and 7-EFC. Figure 5, A and C, shows 7-EC
docked in the active site of the wild-type CYP2B6 and CYP2E1 models,
respectively, in an orientation leading to deethylation. 7-EC fits well
into the CYP2B6 and CYP2E1 active sites with no van der Waals overlaps.
The position of 7-EC in the two enzymes is very similar with residues
103, 209, 363, 367, and 477 lying within 5 Å of the substrate.
Residues 209, 367, and 477 are the closest to 7-EC. The docking results
are consistent with the experimental results showing similar affinity of both wild-type enzymes for 7-EC.
|
7-Butoxycoumarin Oxidation Assays.
In several studies,
7-butoxycoumarin was found to be very informative as an active-site
probe for CYP2B1, as this substrate can undergo regioselective
hydroxylation on the side chain (Kobayashi et al., 1998; Domanski et
al., 2001). Therefore, 7-butoxycoumarin allowed us to analyze the
architecture of the active sites of CYP2B6 and CYP2E1 more carefully
and was used at a concentration according to Domanski (2001). As shown
in Table 5, CYP2B1dH (used as a positive
control) produced 7-(3-hydroxybutoxy)coumarin as the major metabolite
along with about two-thirds as much 7-hydroxycoumarin, about half as
much 7-(2-hydroxybutoxy)coumarin, and a very small amount of
7-(4-hydroxybutoxy)coumarin. The metabolic profile and overall activity
displayed by CYP2B1dH were similar to the full-length form (Domanski et
al., 2001). Total activity for CYP2B6dH was the same as for CYP2B1dH.
However, in contrast to CYP2B1dH, the metabolite profile was dominated
by the production of 7-(3-hydroxybutoxy)coumarin, which represented
83% of total product for CYP2B6dH but only 43% for CYP2B1dH. The
formation of 7-hydroxycoumarin and 7-(2-hydroxybutoxy)coumarin was
dramatically decreased in CYP2B6dH compared with CYP2B1dH. Of the
CYP2B6dH mutants, V477F showed a metabolic profile most similar to
CYP2B6dH. However, the total activity was increased about 3-fold. The
CYP2B6dH L363V mutant showed a striking metabolic switching, with an
overall metabolite profile very similar to that of CYP2B1dH, suggesting
that Leu-363 is a crucial residue for the unique CYP2B6 profile.
|
pNP Hydroxylation Assays.
All CYP2E1dH mutants showed
different kinetic parameters from the wild-type (Table
6). Interestingly, the F477V mutant lost almost all pNP hydroxylation activity. For all mutants except A103M,
Km and
Vmax values were decreased compared
with CYP2E1dH. A103M displayed the highest
Km value, which was 2.5-fold higher than that for the wild-type enzyme. Like A103M, L209I exhibited a
Vmax value about 3-fold lower than the
wild-type enzyme. However, in L209I the
Km value was decreased about 3-fold
compared with the wild-type enzyme. V363L displayed a pNP hydroxylation
activity most similar to CYP2E1dH. However, calculation of
Vmax/Km
indicated a significantly higher efficiency of V363L compared with the
wild-type enzyme. L367V kinetics showed positive cooperativity. The
striking loss of pNP activity of the F477V mutant supported the
importance of residue Phe-477 for the unique substrate selectivity of
CYP2E1 inferred from the studies with 7-EFC. CYP2B6dH mutants
hydroxylated pNP at a rate less than or equal to that of CYP2B6dH (Fig.
6). Thus, none of the mutants gained
CYP2E1dH-like pNP hydroxylation activity. Preliminary spectral studies
showed that pNP binds to the active site of both CYP2E1dH and CYP2B6dH
with a similar KS value of ~15 µM.
However, the difference spectra were distinct. Whereas the binding of
pNP to CYP2B6dH produced a typical type I spectrum, CYP2E1dH exhibited
a reverse type I and a type II spectrum, which will require further
investigation.
|
|
AA Metabolism.
Several studies have used fatty acids as a
substrate for CYP2E1 to investigate the architecture of the active site
and the substrate access channel of this enzyme (Wang et al., 1995;
Adas et al., 1999; Smith et al., 2000). Therefore, additional probing of CYP2E1 and mutants was performed with AA. CYP2E1dH produced 
-1
hydroxylated AA as the major product (60%), and smaller amounts of
-2 and -hydroxylated AA, and of 14,15-epoxyeicosatrienoic acid
(EET) and 11,12-, 8,9-, and 5,6-EET (Fig.
7). The data are in good agreement with
prior studies reported on 2E1 metabolism of AA (Rifkind et al., 1995).
A103M showed a metabolite profile similar to CYP2E1dH. V363L showed a
metabolic switching in favor of -2 and -hydroxylated AA. The most
dramatic changes in the metabolite pattern were observed for the
mutants L209I and F477V, which produced about equal amounts of -1
hydroxylated AA and 14,15-EET as major metabolites. Metabolism by L367V
resulted in a shift toward more 11,12-, 8,9-, and 5,6-EET production
than CYP2E1dH and the other mutants. CYP2B6dH produced only small
amounts of 11,12-, 8,9-, and 5,6-EET (data not shown). The data
revealed that residues Leu-209 and Phe-477 and to a smaller extent
residue Leu-367 are important for the orientation of AA in the active site of CYP2E1.
|
| |
Discussion |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
A number of recent studies have successfully elucidated
active-site residues responsible for differential substrate
selectivities among P450 enzymes from a single subfamily (reviewed in
Domanski and Halpert, 2001). In the present study, we have found that
site-directed mutants and molecular models are also invaluable tools
for gaining insight into the structural basis of differential substrate
selectivities across subfamilies. Furthermore, the use of various
substrates for CYP2B6 and CYP2E1 such as 7-EFC, pNP, 7-EC,
7-butoxycoumarin, and AA enabled us to explore the relationship between
the shape of the active site and the chemical structure of the
substrate. In addition, this report is the first to study the active
site of CYP2E1 by systematic site-directed mutagenesis.
One of the major findings of this study was that substitution of
individual active-site residues across P450 subfamilies can significantly enhance activity toward a substrate if it is similar to a
substrate usually metabolized by the enzyme. This interesting insight
was gained by site-directed mutagenesis and molecular modeling of
CYP2E1 and comparison of 7-EC and 7-EFC deethylation activity.
Substitution of Phe-477 in CYP2E1 with the corresponding residue of
CYP2B6 resulted in about 5-fold higher 7-EFC deethylation activity than
CYP2E1 while preserving activity with 7-EC. Docking of 7-EFC into the
active site of CYP2E1 and CYP2E1 F477V showed that 7-EFC deethylation
might be impeded by van der Waals overlaps between 7-EFC and the side
chain of Phe-477. Likewise, significant van der Waals overlap between
the substrate and Phe-477 was observed when a trifluoromethyl group was
added at position 4 of the coumarin structure in the 7-EC CYP2E1
complex. The crucial role of residue Phe-477 was also supported by the
results observed with the marker substrate of CYP2E1, pNP. CYP2E1 F477V
lost almost all pNP hydroxylation activity, suggesting that a 
-
interaction between the phenyl ring of Phe-477 and the phenyl ring of
pNP might be required for hydroxylation. All other CYP2E1 mutants
showed parallel activities for pNP, 7-EC, and 7-butoxycoumarin, with
A103M and L209I exhibiting lower activity than wild-type and V363L and
L367V exhibiting about the same activity as CYP2E1. Interestingly, the
substitution of Leu-367 with the smaller amino acid residue valine
conferred positive cooperativity for both 7-EFC and pNP kinetics. These
results suggest that a smaller residue at position 367 in CYP2E1 might
enhance the volume of the active site and enable two molecules of the compounds to bind to the active site.
These observations provide new insight into the active site of CYP2E1.
Until now, only one CYP2E1 residue, Thr-303, has been investigated for
a role in substrate interactions (Moreno et al., 2001). Previous
studies using AA to probe the substrate binding site (Wang et al.,
1994, Smith et al., 2000) suggested a CYP2E1 model consisting of a long
hydrophobic pocket with the hydrocarbon end of the AA binding to the
active site (Wang et al., 1994). Our results suggest that several
active-site residues are responsible for the orientation of AA
characteristic for CYP2E1. The most striking alterations in the
metabolite profile were observed for the mutants L209I and F477V. They
produced equal amounts of -1 hydroxylated AA and 14,15-EET as major
metabolites, whereas CYP2E1 produced only -1 hydroxylated AA as the
major metabolite. In addition, substitution of Leu-367 by valine
favored production of 11,12-, 8,9-, and 5,6-EET. The alteration toward
more epoxidation activity suggests that at least residues 209 and 477 are critical for shifting metabolism toward the carboxy-terminal end of
AA. This metabolic switching in favor of epoxygenase activity might provide a hint for the differential substrate activities of CYP2E1 and
CYP2B enzymes. CYP2B1, 2B2, 2B4, and 2B5 produce 14,15-, 11,12-, 8,9-, and 5,6-EET with CYP2B1 predominantly producing 14,15-EET and CYP2B2,
2B4, and 2B5 producing equal amounts of all four epoxides (Laethem et
al., 1994). In our study, CYP2B6dH produced small amounts of 11,12-, 8,9-, and 5,6-EET, suggesting some similarity with the other members of
the CYP2B family.
Investigation of the active site of CYP2B6 addressed a further question: can individual substitution of active-site residues also confer higher catalytic activity toward a substrate that is very different in structure from substrates the enzyme usually oxidizes? pNP is an example of small aromatic compounds, which are often specific CYP2E1 but not CYP2B6 substrates. Binding studies of pNP with CYP2E1 and CYP2B6 suggested that pNP reaches the active site in both enzymes. However, in contrast to CYP2E1 the binding of pNP to the active site of CYP2B6 is nonproductive, which is consistent with the distinct binding spectra of the enzymes with pNP. As none of the CYP2B6 mutants acquired enhanced ability to hydroxylate pNP, a different overall arrangement of the backbone of the enzymes might be involved in the different substrate selectivities of CYP2B6 and CYP2E1.
In contrast to pNP, substrates with a chemical structure similar to
that of 7-EC revealed the involvement of active-site residues for
specific catalysis by CYP2B6. Substitution of Leu-363 with valine in
CYP2B6 reduced the -1 hydroxylation of 7-butoxycoumarin significantly and increased O-dealkylation. The resulting
metabolite profile was remarkably similar to that of CYP2B1. Kobayashi
et al. (1998) observed that substitution of Val-363 with leucine in
CYP2B1 almost abolished the production of 7-hydroxycoumarin and favored
the production of 7-(3-hydroxybutoxy)coumarin. Our data revealed that
the special metabolite ratio observed for CYP2B1 V363L is
characteristic for CYP2B6. These intriguing results suggest that
residue 363 is a crucial residue across the 2B and 2E subfamilies and
across species, in the case of the 2B subfamily.
It was expected that substitution of the small residue valine at
position 477 in CYP2B6 with the aromatic residue phenylalanine might
restrict space in the active site of and/or reduce the affinity for
substrates. Surprisingly, the overall activity of V477F with 7-EFC,
7-EC, and 7-butoxycoumarin increased 2- to 3-fold and the metabolite
profile was unchanged, indicating that a larger aromatic residue might
restrict mobility of the substrate and favor a productive binding
orientation. Indeed, Szklarz et al. (1995) have shown that substitution
of Ile-477 by the smaller residues valine and alanine in CYP2B1 reduced
androstenedione and progesterone hydroxylation significantly,
suggesting that this residue holds both substrates in the proper
orientation relative to the active oxygen.
Topology studies comparing CYP2B and CYP2E1 enzymes suggest a common
architecture of the active sites of these enzymes. However, the
similarities between the active sites might decrease as one moves to
increasing distances from the heme plane (Mackman et al., 1996).
Comparison of our CYP2B6 and CYP2E1 models showed very similar overall
structures for both enzymes (Fig. 8A), as predicted when both are modeled on a single template. Likewise, the
architecture of the active sites of the CYP2B6 and CYP2E1 models varies
only in the differential side chains of the active-site residues (Fig.
8B). Several studies have shown a major improvement in predicting
substrate and inhibitor interactions with individual P450 enzymes or
with members of one subfamily by homology modeling based on CYP2C5
(Afzelius et al., 2001; Spatzenegger et al., 2001; Wang and Halpert,
2002). However, even these models can be successfully used only under
defined conditions such as including only inhibitors with defined types
of inhibition and a limited range of
Ki values (Afzelius et al., 2001).
Given the extremely diverse chemical features of compounds with which
cytochromes P450 can interact, complete understanding of substrate
selectivity might be possible only when representative P450 enzymes of
each subfamily are crystallized.
|
In conclusion, the results of the present and previous
investigations suggest that differences in the side chains of
individual active-site residues may be primarily responsible for
differential substrate selectivities between two cytochromes P450 only
under certain conditions. These include: 1) enzymes from the same
subfamily (Spatzenegger et al., 2001); 2) enzymes from different
subfamilies acting on the same substrate but with different
regioselectivity (Luo et al., 1994); and 3) enzymes from different
subfamilies acting on a substrate that is structurally similar to a
common substrate for both enzymes (present work). In most other cases we suspect that distinct substrate selectivities also reflect a
different arrangement of the backbones of the enzymes and/or differences in the substrate access channel. These latter differences will be very difficult, if not impossible, to predict by substrate docking into the active sites of P450 models built using the same template. At present, the combination of homology models based on
CYP2C5 and pharmacophore models (Afzelius et al., 2001; Wang et al.,
2002) appears to provide one of the more reliable and accurate means of
predicting substrate interactions with human cytochromes P450.
| |
Acknowledgments |
|---|
We thank Dr. Emily E. Scott for kindly providing the CYP2B6dH clone and Dr. Dimitri Davydov for comments on the manuscript.
| |
Footnotes |
|---|
Accepted for publication September 30, 2002.
Received for publication August 15, 2002.
1 Permanent address: Center for Drug Discovery and Design, The State Key Laboratory of New Drug Research, Shanghai Institute of Materia Medica, Shanghai Institute for Biological Sciences, Chinese Academy of Sciences, Shanghai 200031, P. R. China.
2 To simplify the comparison of CYP2B6 and CYP2E1, the CYP2B6 numbering will be used throughout the paper.
3 The assays of the CYP2E1dH mutants were performed with a different lot of 7-EFC, which resulted in lower specific activity than that used for CYP2B6dH mutants.
4 The double mutant CYP2E1dH L367V-F477V did not produce sufficient stable hemeprotein for further studies.
This work was supported by National Institutes of Health Grants ES03619 (J.R.H.) and AA08608 (D.R.K.) and Center Grant ES06676. This work was presented at the 2nd Southwest P450 Meeting, Camp Allen, TX, May 2002.
DOI: 10.1124/jpet.102.043323
Address correspondence to: Dr. James R. Halpert, Professor and Chairman, Department of Pharmacology and Toxicology, The University of Texas Medical Branch, 301 University Blvd., Galveston, TX 77555-1031. E-mail: jhalpert{at}utmb.edu
| |
Abbreviations |
|---|
P450, cytochrome P450;
pNP, p-nitrophenol;
7-EFC, 7-ethoxy-4-trifluoromethylcoumarin;
7-HFC, 7-hydroxy-4-trifluoromethylcoumarin;
7-EC, 7-ethoxycoumarin;
7-HC, 7-hydroxycoumarin;
AA, arachidonic acid;
PCR, polymerase chain
reaction;
ME, -methylmercaptoethanol;
Ni2+-NTA, nickel-nitrilotriacetic acid;
EET, epoxyeicosatrienoic acid.
| |
References |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
and (-1)-hydroxylations of fatty acids by human cytochromes P450 2E1 and 4A11.
J Lipid Res
40:
1990-1997-hydroxylase.
Arch Biochem Biophys
309:
52-57[CrossRef][Medline].This article has been cited by other articles:
|
|
 |
|
  T.-A. Nguyen, M. Tychopoulos, F. Bichat, C. Zimmermann, J.-P. Flinois, M. Diry, E. Ahlberg, M. Delaforge, L. Corcos, P. Beaune, et al. Improvement of Cyclophosphamide Activation by CYP2B6 Mutants: From in Silico to ex Vivo Mol. Pharmacol., April 1, 2008; 73(4): 1122 - 1 |
3'.
), CYP2E1dH L367V (
), and CYP2E1dH F477V (
). CYP2B6dH:
Km = 12 µM,
Vmax = 6.9 nmol/min/nmol of P450;
CYP2E1dH L367V: S50 = 54 µM,
n = 1.7, Vmax = 1.4 nmol/min/nmol of P450; CYP2E1dH F477V:
Km = 83 µM,
Vmax = 5.2 nmol/min/nmol of P450. The
results are the average of two experiments performed in duplicate.
