Engineering bacterial cytochrome P450 (P450) BM3 into a prototype with human P450 enzyme activity using indigo formation

Article abstract of DOI:10.1124/dmd.109.030759

Human cytochrome P450 (P450) enzymes metabolize a variety of endogenous and xenobiotic compounds, including steroids, drugs, and environmental chemicals. In this study, we examine the possibility that bacterial P450 BM3 (CYP102A1) mutants with indole oxidation activity have the catalytic activities of human P450 enzymes. Errorprone polymerase chain reaction was carried out on the heme domain-coding region of the wild-type gene to generate a CYP102A1 DNA library. The library was transformed into Escherichia coli for expression of the P450 mutants. A colorimetric colony-based method was adopted for primary screening of the mutants. When the P450 activities were measured at the whole-cell level, some of the blue colonies, but not the white colonies, possessed apparent oxidation activity toward coumarin and 7-ethoxycoumarin, which are typical human P450 substrates that produce fluorescent products. Coumarin is oxidized by the CYP102A1 mutants to produce two metabolites, 7-hydroxycoumarin and 3-hydroxycoumarin. In addition, 7-ethoxycoumarin is simultaneously oxidized to 7-hydroxycoumarin by O-deethylation reaction and to 3-hydroxy,7-ethoxycoumarin by 3-hydroxylation reactions. Highly active mutants are also able to metabolize several other human P450 substrates, including phenacetin, ethoxyresorufin, and chlorzoxazone. These results indicate that indigo formation provides a simple assay for identifying CYP102A1 mutants with a greater potential for human P450 activity. Furthermore, our computational findings suggest a correlation between the stabilization of the binding site and the catalytic efficiency of CYP102A1 mutants toward coumarin: the more stable the structure in the binding site, the lower the energy barrier and the higher the catalytic efficiency. Copyright

Full text of DOI:10.1124/dmd.109.030759

Supplemental Material can be found at:
http://dmd.aspetjournals.org/content/suppl/2010/01/25/dmd.109.030759.DC1.html
0090-9556/10/3805-732–739$20.00
DRUG METABOLISM AND DISPOSITION
Copyright © 2010 by The American Society for Pharmacology and Experimental Therapeutics
DMD 38:732–739, 2010
Vol. 38, No. 5
30759/3574565
Printed in U.S.A.
Short Communication
Engineering Bacterial Cytochrome P450 (P450) BM3 into a
Prototype with Human P450 Enzyme Activity Using
□
S
Indigo Formation
Received October 14, 2009; accepted January 25, 2010
ABSTRACT:
Human cytochrome P450 (P450) enzymes metabolize a variety of is oxidized by the CYP102A1 mutants to produce two metabolites,
endogenous and xenobiotic compounds, including steroids, drugs, 7-hydroxycoumarin and 3-hydroxycoumarin. In addition, 7-ethoxy-
and environmental chemicals. In this study, we examine the possibil- coumarin is simultaneously oxidized to 7-hydroxycoumarin by O-
ity that bacterial P450 BM3 (CYP102A1) mutants with indole oxidation deethylation reaction and to 3-hydroxy,7-ethoxycoumarin by 3-hy-
activity have the catalytic activities of human P450 enzymes. Error- droxylation reactions. Highly active mutants are also able to
prone polymerase chain reaction was carried out on the heme do- metabolize several other human P450 substrates, including phenac-
main-coding region of the wild-type gene to generate a CYP102A1 etin, ethoxyresorufin, and chlorzoxazone. These results indicate that
DNA library. The library was transformed into Escherichia coli for indigo formation provides a simple assay for identifying CYP102A1
expression of the P450 mutants. A colorimetric colony-based method mutants with a greater potential for human P450 activity. Further-
was adopted for primary screening of the mutants. When the P450 more, our computational findings suggest a correlation between the
activities were measured at the whole-cell level, some of the blue stabilization of the binding site and the catalytic efficiency of
colonies, but not the white colonies, possessed apparent oxidation CYP102A1 mutants toward coumarin: the more stable the structure in
activity toward coumarin and 7-ethoxycoumarin, which are typical the binding site, the lower the energy barrier and the higher the
human P450 substrates that produce fluorescent products. Coumarin catalytic efficiency.
Cytochrome P450 (P450, also known as P450 for specific isoforms) ities makes soluble CYP102A1 an ideal model for mammalian, par-
enzymes constitute a large family of enzymes that are remarkably
diverse oxygenation catalysts found in archaea, bacteria, fungi, plants,
and animals. P450s function primarily in the oxidation of various
xenobiotics and endogenous compounds, especially in mammals. Due
to the catalytic diversity and broad substrate range of P450s, they are
attractive biocatalyst candidates for the production of fine chemicals,
including pharmaceuticals. Despite the potential use of mammalian
P450s in various biotechnology fields, they are not suitable as bio-
catalysts because of their low stability and low catalytic activity (Yun
et al., 2007). P450 BM3 (CYP102A1) from Bacillus megaterium was
the first P450 discovered to be fused to its redox partner, a mamma-
lian-like diflavin reductase. The fusion of these two enzymatic activ-
ticularly human P450 enzymes (Munro et al., 2002). More recently,
through rational design or directed evolution, wild-type CYP102A1
has been engineered to oxidize compounds that show little or no
structural similarity to the natural fatty acid substrates of CYP102A1
(Yun et al., 2007; Kim et al., 2008, 2009).
In a previous study, a screening procedure was developed for the
directed evolution of human CYP2A6 by using indigo formation
(Gillam et al., 1999). Because Escherichia coli cultures produce
indigo in the absence of supplemental indole, the rich expression
media used in this procedure are likely to have provided sufficient
indole as a substrate via tryptophan degradation. Indigo formation
from indole oxidation was first used to screen CYP102A1 mutants by
saturation mutagenesis at specific sites. A set of selected CYP102A1
mutants (F87V, F87V/L188Q, and F87V/L188Q/A74G) could effi-
ciently hydroxylate indole, which leads to the production of indigo
and indirubin (Li et al., 2000). More recently, wild-type and mutant
forms of the bacterial CYP102A1 have been found to metabolize
various drugs (Yun et al., 2007; Damsten et al., 2008). 7-Ethoxycou-
marin, a substrate of CYP2E1, CYP2A6, and CYP1A2, was also
found to be oxidized through reactions similar to those catalyzed by
CYP102A1 mutants (Kim et al., 2005, 2006, 2008). Indole is a
substrate for CYP2A6 (Wu et al., 2005). CYP2A6 mutants generated
This work was supported in part by the 21C Frontier Microbial Genomics and
the Application Center Program of the Ministry of Education, Science and Tech-
nology of the Republic of Korea; the Korea Science and Engineering Foundation
[Grant R01-2008-000-21072-02008]; and the Second Stage BK21 Project from
the Ministry of Education, Science and Technology of the Republic of Korea.
Article, publication date, and citation information can be found at
http://dmd.aspetjournals.org.
doi:10.1124/dmd.109.030759.
□S The online version of this article (available at http://dmd.aspetjournals.org)
contains supplemental material.
ABBREVIATIONS: P450, cytochrome P450; CYP102A1, P450 BM3; PCR, polymerase chain reaction; 7-OH coumarin, 7-hydroxycoumarin; 3-OH
coumarin, 3-hydroxycoumarin; LB, Luria-Bertani; IPTG, isopropyl-␤-D-thiogalactopyranoside; PhOD, phenacetin O-deethylation; EROD,
7-ethoxyresorufin O-deethylation; HPLC, high-performance liquid chromatography; LC/MS/MS, liquid chromatography-tandem mass spectrom-
etry; MD, molecular dynamics; Tm, thermal stability.
732

BACTERIAL P450 BM3 MUTANTS WITH HUMAN P450 ENZYME ACTIVITY
733
were of analytical grade or high quality and were purchased from commercial
suppliers and used without further purification.
by random and site-directed mutagenesis were able to oxidize bulky
substituted indole compounds, which are not substrates for the wild-
type enzyme.
Generation of the CYP102A1 Libraries. Error-prone PCR was carried out
on the heme domain-coding regions (base pairs 428–888) of wild-type
CYP102A1 to generate the first CYP102A1 DNA library. We constructed the
pCW BM3-KX vector, which includes KpnI and XhoI restriction sites, by
using silent mutagenesis. Oligonucleotide primers were used to introduce KpnI
and XhoI restriction sites: KpnI forward, 5Ј-GAGCATATTGAGGTACCG-
GAAGAC-3Ј; XhoI reverse, 5Ј-TACTAGAACTCGAGCTGCTTCTTC-3Ј.
The P450 heme domain region of pCW BM3-KX was randomly mutated using
a GeneMorph II random mutagenesis kit (Stratagene, La Jolla, CA). The 50-␮l
reaction volume contained 50 pmol of each primer (forward primer 5Ј-GAG-
CATATTGAGGTACCGGAAGAC-3Ј and reverse primer 5Ј-TACTAGAAC-
TCGAGCTGCTTCTTC-3Ј), 70 ng of the template plasmid DNA (pCW BM3-
KX), 2.5 U of Mutazyme DNA polymerase, 0.2 ␮M of each deoxynucleotide
triphosphate, and 10ϫ Mutazyme (Stratagene) reaction buffer. The PCR
reaction was initiated at 95°C for 2 min and run through 30 thermocycles of
95°C for 30 s, 50°C for 30 s, and 72°C for 1 min. After completion of the 30
cycles, the reaction medium was held at 72°C for 10 min. The amplified PCR
library fragments were purified and cloned into the pCW BM3-KX vector
using the KpnI and XhoI restriction sites. In the same way, a second library
was constructed by using the F162I/M237I mutant, which had the highest
7-ethoxycoumarin oxidation activity among a set of mutants from the first
screen, which was used as the template.
In the present study, we used random mutagenesis of wild-type
CYP102A1 to generate mutants with catalytic activities similar to
human P450 enzymes. Error-prone polymerase chain reaction (PCR)
was carried out on the heme domain-coding regions of the wild-type
gene to generate a CYP102A1 DNA library. The library was trans-
formed into E. coli for expression of the CYP102A1 mutants. A
colorimetric colony-based method was adopted for primary screening
of the mutants. When the P450 activities were measured at the
whole-cell level, some of the blue colonies, but not the white colonies,
showed apparent oxidation activity toward typical human P450 sub-
strates, including coumarin, 7-ethoxycoumarin, phenacetin, and chlor-
zoxazone (Fig. 1). The enzymatic and physical properties of selected
CYP102A1 mutants that showed human P450 activity were studied
further.
Materials and Methods
Chemicals. 7-Ethoxycoumarin, coumarin, 7-hydroxycoumarin (7-OH cou-
marin), 3-hydroxycoumarin (3-OH coumarin), phenacetin, acetaminophen,
p-nitrophenol, and chlorzoxazone were purchased from Sigma-Aldrich (St.
Louis, MO). 7-Ethoxyresorufin and 7-hydroxyresorufin were obtained from
Screening of the CYP102A1 Libraries. Randomized plasmid libraries
Invitrogen (Carlsbad, CA). All other chemicals and solvents used in this study were transformed into E. coli DH5␣F’IQ. E. coli cells that expressed the
FIG. 1. HPLC analyses of the oxidation of human P450 substrates catalyzed by the CYP102A1 F162I/M185T/L188P/M237I mutant. The substrate and major product(s)
are indicated. The substrates included in the reaction mixtures are as follows: A, 7-ethoxycoumarin; B, coumarin; C, phenacetin; and D, chlorzoxazone. Oxidation of the
substrates were catalyzed by the F162I/M185T/L188P/M237I mutant (bold line) but not by wild-type CYP102A1 (dash line). Peaks were identified by comparing the
retention times of the product with the reaction time of standard controls. Unidentified metabolites (marked with an asterisk) were also produced by the F162I/M185T/
L188P/M237I mutant.

734
PARK ET AL.
CYP102A1 mutants were spread on Luria-Bertani (LB) broth agar expression
plates, incubated at 37°C for 16 h, and then stored at 4°C for approximately 2
weeks. LB agar expression plates contained 25 g/l LB broth (BD Biosciences,
San Jose, CA), 15 g/l Bacto agar (BD Biosciences), and 100 ␮g/ml ampicillin.
No additives, including indole, isopropyl-␤-D-1-thiogalactopyranoside (IPTG),
or ␦-aminolevulinic acid, were added to the plates. Approximately 103 blue-
colored colonies were transferred into 96-deep-well plates that contained 0.3
ml of LB media and 100 ␮g/ml ampicillin. The bacteria were initially cultured
for 8 h at 37°C and 450 rpm (HT-MegaGrow incubator; Bioneer Co., Daejeon,
Korea). Aliquots of the cell cultures (30 ␮l) were transferred to single wells of
96-deep-well plates filled with 0.5 ml of Terrific Broth expression media
containing additives (10 ␮M IPTG, 0.5 mM ␦-aminolevulinic acid) and am-
picillin. These aliquots were cultured for 17 h at 30°C and 450 rpm. For
primary screening, the whole-cell activities of the mutants were measured by
using 7-ethoxycoumarin and coumarin as substrates. The cells were collected
by centrifugation at 3500 rpm for 10 min at 4°C and frozen at Ϫ20°C. Cell
pellets were resuspended in 300 ␮l of phosphate buffer (0.1 M, pH 8.0) that
contained 10 mM MgCl2. After a 60-min incubation at 30°C and 400 rpm, 100
␮M of the substrate was added, and the reactions were initiated by the addition
of an NADPH-generating system (final concentration of 0.5 mM NADPϩ, 10
mM glucose 6-phosphate, and 1.0 IU glucose-6-phosphate dehydrogenase).
The plates were then incubated for 30 min at 37°C and 350 rpm. The reaction
was terminated by the addition of 100 ␮l of trichloroacetic acid (20%; w/v),
and the mixtures were centrifuged at 3000 rpm for 5 min at 4°C. Aliquots of
the supernatant (50 ␮l) were transferred into a new black 96-well plate
containing 150 ␮l of Tris-HCl (pH 9.0), and the fluorescence (excitation at 355
nm and emission at 460 nm) of the mixture was measured by using a
microplate reader (Infinite M200; Tecan, Madison, WI). Fifty mutants were
selected for a second round of screening based on their high 7-ethoxycoumarin
and coumarin oxidation activities.
Protein Expression and Purification. Wild-type and selected mutants of
CYP102A1 were expressed in E. coli and purified as described previously
(Kim et al., 2008). The CYP102A1 concentration was determined from the
CO-difference spectra as described by Omura and Sato (1964) using ␧ ϭ 91
mM/cm. For all of the wild-type and mutant enzymes, a typical culture yielded
between 300 and 700 nM P450. The expression level of wild-type and mutant
CYP102A1 was typically within the range of 1.0 to 2.0 nmol P450/mg
cytosolic protein.
P450 Catalytic Activity Assays. Several typical substrates of human P450s
were used to examine the catalytic activities of the wild-type and the mutant
forms of CYP102A1, as described below (Figs. 1 and 2). Typical steady-state
reactions for 7-ethoxycoumarin oxidation, coumarin oxidation, phenacetin
O-deethylation (PhOD), 7-ethoxyresorufin O-deethylation (EROD), chlor-
zoxazone 6-hydroxylation, and p-nitrophenol hydroxylation were composed of
50 pmol of CYP102A1 in 0.25 ml of potassium phosphate buffer (100 mM, pH
7.4) and the specified amount of substrate. To determine the turnover numbers
of wild-type CYP102A1 and all of the tested mutants, we used substrate
concentrations of 2.0 mM, 2.0 ␮M, 200 ␮M, and 800 ␮M for phenacetin,
7-ethoxyresorufin, chlorozoxazone, and p-nitrophenol, respectively.
Oxidation of coumarin and 7-ethoxycoumarin. Assays of 7-ethoxycoumarin
and coumarin oxidation were performed by using the wild-type CYP102A1
and its mutants. The reaction mixture contained 50 pmol of CYP102A1, 100
mM potassium phosphate buffer (pH 7.4), and 2 mM of the substrate in a total
volume of 0.25 ml. Both substrate stocks (200 mM) were prepared in CH3CN
and diluted into the enzyme reaction such that the final organic solvent
concentration was always less than 1% (v/v). The sample was preincubated for
10 min, and the reaction was initiated by the addition of the NADPH-
generating system (0.5 mM NADPϩ, 10 mM glucose 6-phosphate, and 1.0 IU
glucose 6-phosphate dehydrogenase per milliliter) and continued in a water
bath at 37°C. After a 30-min incubation, the reaction was stopped by the
addition of 500 ␮l of ice-cold CH2Cl2. After centrifugation (3000 rpm for 10
FIG. 2. Catalytic activity of the CYP102A1 mutants toward human P450 substrates.
The following human P450 enzyme activities of the purified CYP102A1 mutant
enzymes were characterized: A, phenacetin O-deethylation; B, 7-ethoxyresolrufin
O-deethylation; C, chlorzoxazone 6-hydroxylation; D, p-nitrophenol hydroxylation.
Data are shown as the mean Ϯ S.E.M.
CYP102A1 enzyme, 100 mM potassium phosphate buffer (pH 7.4), an
NADPH-generating system, and 2 ␮M 7-ethoxyresorufin in a total volume of
0.25 ml. Preincubation with the reaction mixture was generally carried out for
5 min at 37°C. The reaction was initiated by the addition of the NADPH-
generating system, incubated for 5 min at 37°C, and terminated with 0.5 ml of
CH3OH. Metabolites were measured by using fluorescence and a resorufin
standard (excitation at 535 nm and emission at 585 nm).
PhOD. The reaction mixture consisted of 50 pmol of CYP102A1, 100 mM
potassium phosphate buffer (pH 7.4), an NADPH-generating system, and 2
mM phenacetin in a total volume of 0.25 ml. PhOD activity was determined by
HPLC, as described previously (Yun et al., 2000). Incubations with the
reaction mixture were performed for 20 min at 37°C, terminated with 50 ␮l
of HClO4 (17%), and centrifuged (103 g for 10 min). Then, 0.5 ml of a
min), 300-␮l aliquots of the organic layer from each incubation were trans- mixture of CHCl3 and 2-propanol (6:4, v/v) was added to the supernatant
ferred to a clean glass tube, and the CH2Cl2 was removed under a N2 stream. to extract the products, followed by centrifugation (twice at 103 g). The
The metabolites of 7-ethoxycoumarin and coumarin were analyzed by high-
performance liquid chromatography (HPLC) and liquid chromatography-tan- The products were analyzed by HPLC using a Gemini C18 column (4.6 ϫ
dem mass spectrometry (LC/MS/MS), as described below.
150 mm, 5 ␮m; Phenomenex, Torrance, CA) with a mobile phase of
EROD. Activities of EROD were measured by using a fluorescence assay H2O/CH3OH/CH3CO2H (65:35:0.1, v/v/v; flow rate of 1.0 ml/min) and
(Burke and Mayer, 1983). The reaction mixture contained 50 pmol of the monitored at A254
organic layers were combined, and the solvent was removed under N2 gas.
.

BACTERIAL P450 BM3 MUTANTS WITH HUMAN P450 ENZYME ACTIVITY
735
Chlorzoxazone 6-hydroxylation. The reaction mixture consisted of 50 pmol
of CYP102A1, 100 mM potassium phosphate buffer (pH 7.4), an NADPH- on the Merck molecular (Halgren, 1999) force field in the Discovery Studio
generating system, and 200 ␮M chlorzoxazone in a total volume of 0.25 ml.
environment (Accelrys Inc., San Diego, CA). The active site residues around the
Chlorzoxazone 6-hydroxylation activity was determined by HPLC, as de- heme pocket of the CYP102A1 mutants and coumarin were allowed to flex during
scribed previously (Guengerich et al., 1991). Incubation with the reaction
the docking process. The binding site pockets for automated docking were as-
mixture was performed for 30 min at 37°C, terminated with 25 ␮l of H3PO4 signed via the interaction of the CYP102A1 mutant with a set of 10,000 random
(43%) and 500 ␮l of ice-cold CH2Cl2, and centrifuged at 103 g for 10 min. The
coumarin conformers. The most energy-favorable coumarin conformer and its
organic layers were combined, and the solvent was removed under N2 gas. The binding receptor were selected for further analysis. To better represent the overall
products were analyzed by HPLC using a Luna C8 column (4.6 ϫ 150 mm, 5
strength of the hydrogen bonding between coumarin and the active site residues of
␮m; Phenomenex) with a mobile phase of H2O/CH3CN/H3PO4 (72.5:27:0.5, the binding cavity in a docked complex, the hydrogen-bonding energy associated
Docking of coumarin was carried out using the flexible docking module based
v/v/v; flow rate of 1.25 ml/min) and monitored at A287
.
with each simulated HꢀꢀꢀO/HꢀꢀꢀN distance was estimated using a module imple-
mented in the AutoDock 4.0 program suite (Scripps Research Institute, La Jolla,
CA) (Huey et al., 2007).
p-Nitrophenol hydroxylation. The reaction mixture consisted of 50 pmol of
CYP102A1, 100 mM potassium phosphate buffer (pH 7.4), an NADPH-
generating system, and 800 ␮M p-nitrophenol in a total volume of 0.25 ml.
p-Nitrophenol hydroxylation activity was determined by the spectrophotomet-
ric method described previously (Chang et al., 2006), with a minor modifica-
tion. Incubation with the reaction mixture was performed for 5 min at 37°C,
Molecular Dynamics Simulation. The molecular dynamics (MD) simula-
tions were performed with the AMBER 10 program (Pearlman et al., 1995; Case
et al., 2005). For each of the four dockings of coumarin with the CYP102A1
mutants (F162I/H236R and F162I/M185T/L188P/M237I), in vacuo MD trajecto-
ries were generated. The docking of coumarin to the CYP102A1 mutants (F162I/
H236R and F162I/M185T/L188P/M237I) was initially performed according to
two predicted topological binding sites (I or II) by several algorithms (Huang and
Schroeder, 2008). The in vacuo MD runs were then performed at a constant
volume and temperature with no periodic condition applied, because we were
concerned with only studying the docking of the coumarin molecule to the mutated
CYP102A1 enzymes. Each trajectory was equilibrated for 500 ps with a time step
of 0.001 ps. Then a 2500 ps run was performed, and the trajectory data were
collected. The average conformational-binding energy values were computed by
averaging the total (nonpolar ϩ polar) energy values over the 2500-ps run and then
using this average to calculate the interaction energy values. A 10-Å cutoff for
coulombic and long-range forces were adopted in each simulation. A relative
dielectric constant value of 1.0 was used in all of the simulations.
terminated with 100 ␮l of trichloroacetic acid (20%), and centrifuged (103
for 10 min). After 0.1 ml of NaOH (2 M) was added to the supernatant (0.2 ml),
the absorbance at 535 nm was measured. The amount of product was deter-
mined using a p-nitrocatechol standard.
g
HPLC Analysis of 7-Ethoxycomarin and Coumarin Metabolites. The
7-ethoxycoumarin metabolites were analyzed as described previously (Kim et al.,
2008). Coumarin products were dissolved in 50% CH3CN and analyzed by gradient
HPLC using the Germini C18 column. The mobile phase consisted of 10% CH3CN
containing 0.1% formic acid (buffer A) and 100% CH3CN containing 0.1% formic
acid (buffer B). The column was eluted at a flow rate of 1 ml/min by a gradient pump
(LC-20AD; Shimadzu, Kyoto, Japan) with the following gradient: 0 to 2 min, isocratic
with 30% buffer B; 2 to 8 min, 30 to 35% buffer B gradient; 8 to 10 min, 35 to 45%
buffer B gradient; 10 to 15 min, isocratic with 45% buffer B; 15 to 17 min, 45 to 30%
buffer B gradient; 17 to 20 min, isocratic with 30% buffer B (all v/v). The absorbance
at 330 nm was monitored.
Results and Discussion
LC/MS/MS Analysis of Coumarin Metabolites. LC/MS/MS analysis was
performed on a Waters Synapt HDMS system (Waters, Milford, MA) in the
Korea Basic Science Institute facility (Gwangju, Korea). Analysis was per-
formed in the ESI negative ion mode using an Acquity UPLC BEH C18
column (1.7 ␮m, 2.1 ϫ 100 mm; Waters). All analyses were performed using
a gradient from buffer A (10% CH3CN containing 0.1% formic acid) to buffer
B (100% CH3CN containing 0.1% formic acid). The following gradient
program was used with a flow rate of 0.4 ␮l/min, and the sample was injected
onto the column using the following A:B solvent mixtures: (70:30, v/v) for 0
to 1 min; (60:40, v/v) for 1 to 2 min; (50:50, v/v) for 2 to 2.5 min; (30:70, v/v)
for 2.5 to 3.5 min; (10:90, v/v) for 3.5 to 4 min; (0:100, v/v) for 4 to 5.5 min;
and (70:30, v/v) for 5.5 to 6 min. The temperature of the column was
maintained at 30°C. ESI conditions were as follows: capillary 3.1 kV, cone
voltage 40 V, source temperature of 80°C, desolvation temperature of 300°C,
and desolvation gas flow of 450 l/h.
Steady-State Kinetic Assays for 7-Ethoxycoumarin and Coumarin Ox-
idation Reactions. To determine the kinetic parameters of the CYP102A1
mutants, we used 0.01 to 2 mM 7-ethoxycoumarin and coumarin. The products
were analyzed by HPLC as described above. The kinetic parameters (kcat and Km)
were determined by using nonlinear regression analysis with GraphPad Prism
software (GraphPad Software Inc., San Diego, CA). The data were fit to the
standard Michaelis-Menten equation: v ϭ kcat[E][S]/([S] ϩ Km). The equation is
the standard Michaelis-Menten equation, where the velocity of the reaction is a
function of the rate-limiting step in turnover (kcat), the enzyme concentration ([E]),
substrate concentration ([S]), and the Michaelis constant (Km).
Mutant Modeling and Automated Ligand Docking. The 3-dimensional
structure models of two mutants of CYP102A1, F162I/H236R and F162I/
M185T/L188P/M237I, were constructed based on the X-ray structure of wild-
type CYP102A1 (Protein Data Bank code 1BU7) (Sevrioukova et al., 1999)
using MODELLER (University of California San Francisco, San Francisco,
CA) (Sali and Blundell, 1993) and were optimized using FoldX (Centre for
Genomic Regulation, Barcelona, Spain) (Schymkowitz et al., 2005). The
structure of the coumarin molecule was taken from the Cambridge Structural
Database. The partial atomic charges were fit to the electrostatic potential
computed by ab initio GAMESS calculations performed at the 6-31 G accuracy
level (Schmidt et al., 1993).
Construction of the CYP102A1 Libraries and Screening the
Mutants with Indigo Formation. Error-prone PCR was used to con-
struct randomized libraries, and a colorimetric colony-based method was
used in the screening of the mutants (Nakamura et al., 2001; Whitehouse
et al., 2008; Zhang et al., 2009). First, error-prone PCR was used to
generate a randomized library using wild-type CYP102A1 as the tem-
plate. The heme domain-coding region between the KpnI and XhoI
restriction sites was randomly mutated. A colorimetric colony-based
method was adopted for primary screening of the mutants. Approxi-
mately 300 blue colonies with indigo formation were transferred to
96-deep-well plates and grown as described previously (Li et al., 2008)
(Supplemental Fig. S1). Because 7-ethoxycoumarin and coumarin are
well recognized substrates of human P450 enzymes, whole-cell oxidation
activity of the mutants on the conversion of 7-ethoxycoumarin and
coumarin to 7-OH coumarin, a fluorescent product, was measured
(Hirano et al., 2002). Colonies that showed these catalytic activities were
selected, and each mutated enzyme was expressed in E. coli for purifi-
cation. The catalytic activities of the mutants on 7-ethoxycoumarin and
coumarin were compared (Supplemental Fig. S2). The F162I/M237I
mutant was found to have the highest activity toward both of the sub-
strates and was selected as the template for the generation of a second
library. The second screening was performed in the same way as the first
screening. After error-prone PCR, approximately 103 blue colonies were
screened, and fluorescent-based, whole-cell assays were performed. The
catalytic activities of the selected candidates (50 colonies) were con-
firmed by HPLC analysis. Through the iterative screenings, nine mutants
with high activity for both 7-ethoxycoumarin and coumarin oxidation
were finally chosen and subsequently characterized by nucleotide se-
quence analysis.
Oxidation of 7-Ethoxycoumarin and Coumarin by the CYP102A1
Mutants. Wild-type CYP102A1, F162I/M237I, and nine other promis-
ing mutants were purified to measure the rate of 7-ethoxycoumarin

736
PARK ET AL.
oxidation at a fixed substrate concentration (2 mM). Two major metab- tivity of the CYP102A1 mutants for coumarin 7-hydroxylation is com-
olites, 7-OH coumarin and 3-OH 7-ethoxycoumarin, were analyzed by parable to that of CYP2A6 purified from a human liver (Yun et al., 1991).
HPLC (Kim et al., 2006). All of the mutants had apparent catalytic
Next, we determined the total turnover numbers of the CYP102A1 mu-
activity toward 7-ethoxycoumarin (Supplemental Fig. S3A), whereas tants for the coumarin and 7-ethoxycoumarin oxidation reactions. The overall
wild-type CYP102A1 did not possess any apparent catalytic activity for range of 7-ethoxycoumarin O-deethylation was between 500 and 2000. The
this substrate (Ͻ0.1 minϪ1 for O-deethylation and 3-hydroxylation). The overall range of 7-ethoxycoumarin 3-hydroxylation was between 1500 and
F162I/H236R and F162I/M185V/M237I mutants had high rates of O- 3500. For 7-ethoxycoumarin oxidation, most of the mutants, except for
deethylation (19.7 minϪ1 and 17.8 minϪ1, respectively), whereas the F162I/E228K, showed an increase in product formation with increasing
F162I/K187E/M237I and F162I/M185T/L188P/M237I mutants had high incubation time (Supplemental Fig. S5, A and B). In terms of the coumarin
3-hydroxylation activity (38.0 minϪ1 and 41.6 minϪ1, respectively). The oxidation reactions, all of the tested mutants showed an increase in product
F162I single mutant and the L148I/F162I double-mutant showed little formation with increased incubation time (Supplemental Fig. S5, C and D).
effect on O-deethylation and 3-hydroxylation activity compared with the In general, the CYP102A1 mutants seem to prefer the hydroxylation reaction
other mutants, and these mutants had low O-deethylation and 3-hydroxy- over the O-deethylation reaction.
lation activity. Like the human P450 enzymes (Kim et al., 2006) and the
Oxidation of Other Human P450 Substrates Is Catalyzed by the
recently reported mutants of CYP102A1 (Kim et al., 2008), all of the CYP102A1 Mutants. The catalytic activities of the CYP102A1 mutant
mutants examined in the present study showed a higher rate of 3-hydroxy- enzymes toward four substrates that are known to be metabolized by
lation compared with their O-deethylation activity.
human P450 enzymes were investigated. First, the ability of wild-type
The catalytic activity of wild-type CYP102A1 and its mutants was also CYP102A1 and a set of CYP102A1 mutants to catalyze human P450
examined using coumarin, a typical P450 substrate. Coumarin is oxidized substrates was measured at fixed substrate concentrations (2.0 mM phen-
by the CYP102A1 mutants to produce two major metabolites (7-OH acetin, 2.0 ␮M 7-ethoxyresorufin, 200 ␮M chlorozoxazone, and 800 ␮M
coumarin and 3-OH coumarin) and two unknown minor products (Fig. p-nitrophenol) (Fig. 2).
1). The identities of the major metabolites and the substrate were verified
The CYP102A1 mutant enzymes converted phenacetin into one major
by comparing the results of HPLC and LC/MS/MS with standard com- metabolite. HPLC analysis identified the metabolite as acetaminophen
pounds (Supplemental Fig. S4). CYP2A6, the major enzyme for the (Fig. 1C), and LC-MS comparison (data not shown) with the authentic
hydroxylation reaction of coumarin in the human liver, catalyzes only the standard compound confirmed that the metabolite was acetaminophen.
7-hydroxylation reaction and not the 3-hydroxylation reaction (Kim et al., Acetaminophen and acetol are known oxidation products of phenacetin
2005; Yun et al., 2005). However, unlike the CYP2A6 enzyme, all of the that are catalyzed by human CYP1A2 (Yun et al., 2000). Acetol was not
CYP102A1 mutants showed a preference for the 3-hydroxylation reac- formed by any of the CYP102A1 mutant enzymes. The turnover numbers
tion over the 7-hydroxylation. 3-OH coumarin is reported to be produced for the entire set of 10 mutants for O-deethylation of phenacetin varied
as a major metabolite of coumarin by CYP3A and CYP1A in humans, over a wide range. We found that wild-type CYP102A1 showed very low
whereas CYP2A6 does not catalyze 3-OH coumarin formation (Born et activity (0.40 minϪ1 for PhOD) under the test conditions (Fig. 2A),
al., 2002). Our results obtained for the CYP102A1 mutants for coumarin whereas the F162I/E228K mutant showed the highest activity (42 minϪ1
3-hydroxylation are similar to those for 7-ethoxycoumarin 3-hydroxyla- among the mutants.
)
tion (Supplemental Fig. S3). Wild-type CYP102A1 showed no catalytic
The EROD activity of the CYP102A1 mutants was also examined
activity for coumarin. The F162I/M185T/L188P/M237I mutant, which (Fig. 2B). Spectrofluorometric analysis identified the metabolite as
showed high activity for O-deethylation and 3-hydroxylation of resorufin, and the identification was compared with an authentic resorufin
7-ethoxycoumarin, also had high activity for coumarin hydroxylation. standard compound (data not shown). Although the F162I/E228K mutant
The F162I and L148I/F162I mutants, which possessed low 7-ethoxycou- showed the highest EROD reaction rate (0.57 minϪ1) among all of the
marin oxidation activity, also showed low coumarin oxidation activity.
mutants, its activity was relatively low compared with the other substrates
Seven high-activity mutants were chosen and used to measure the tested.The F162I/F165L/M177T/M237I mutant did not show any appar-
kinetic parameters of the coumarin and 7-ethoxycoumarin oxidation ent activity toward 7-ethoxyresorufin, but it did show high hydroxylation
reactions. The mutants displayed a more than 3-fold variation in kcat for activity toward chlorzoxazone (9.5 minϪ1). All of the CYP102A1 mu-
both 7-hydroxylation and 3-hydroxylation reactions of coumarin. The tants tested here could catalyze the hydroxylation reaction of chlorzoxa-
F162I/K187E/M237I and F162I/M185T/L188P/M237I mutants were the zone and p-nitrophenol, which are typical human CYP2E1 substrates.
most effective with 20 minϪ1 and 4 minϪ1 for 3-hydroxylation and The wild-type CYP102A1 did not possess any apparent activity toward
7-hydroxylation of coumarin, respectively (Table 1). For 7-ethoxycou- these substrates (Fig. 2, C and D).
marin, the F162I/M185T/L188P/M237I and L148I/F162I mutants were
Stability of the CYP102A1 Mutants. To estimate the stability of the
the most effective with 41 minϪ1 and 14 minϪ1 for 3-hydroxylation and mutants, we determined the thermal stability (Tm, oC) of wild-type
O-deethylation, respectively (Supplemental Table S1). The catalytic ac- CYP102A1 and the mutants (Supplemental Fig. S6). The Tm value of
TABLE 1
Kinetic parameters of the CYP102A1 mutants for 7-hydroxylation and 3-hydroxylation of coumarin
7-OH Coumarin
3-OH Coumarin
Mutant
kcat
Km
kcat/Km
kcat
Km
kcat/Km
minϪ1
mM
minϪ1/mM
minϪ1
mM
minϪ1/mM
F162I/M237I
L148I/F162I
F162I/K187E
F162I/E228K
F162I/H236R
F162I/K187E/M237I
F162I/M185T/L188P/M237I
1.2 Ϯ 0.2
2.2 Ϯ 0.6
1.2 Ϯ 0.1
1.4 Ϯ 0.1
1.9 Ϯ 0.1
2.5 Ϯ 0.2
4.0 Ϯ 0.3
3.2 Ϯ 0.8
13 Ϯ 1
0.39 Ϯ 0.10
0.17 Ϯ 0.05
0.18 Ϯ 0.09
0.86 Ϯ 0.09
1.1 Ϯ 0.1
7.5 Ϯ 2.0
12 Ϯ 2
5.2 Ϯ 0.2
11 Ϯ 2
10 Ϯ 1
20 Ϯ 3
6.3 Ϯ 2.1
15 Ϯ 3
1.2 Ϯ 0.4
0.77 Ϯ 0.16
7.0 Ϯ 0.4
2.0 Ϯ 0.4
3.9 Ϯ 0.4
4.4 Ϯ 0.9
13 Ϯ 2
0.64 Ϯ 0.07
1.7 Ϯ 0.3
1.7 Ϯ 0.2
1.7 Ϯ 0.2
1.1 Ϯ 0.2
0.76 Ϯ 0.08
5.3 Ϯ 1.0
2.6 Ϯ 0.3
4.6 Ϯ 1.0
1.2 Ϯ 0.2
1.5 Ϯ 0.2
3.8 Ϯ 0.3
16 Ϯ 1

BACTERIAL P450 BM3 MUTANTS WITH HUMAN P450 ENZYME ACTIVITY
737
wild-type CYP102A1 was determined to be 52.3 Ϯ 0.22°C, whereas the (Fig. 3B). Therefore, we docked coumarin to models of these two
Tm value of the mutants varied. The F162I/M185T/L188P/M237I mu- CYP102A1 mutants (Fig. 3). Several computational investigations by
tant, which had the highest 7-ethoxycoumarin and coumarin oxidation ligand docking and MD simulations were performed to study the rela-
activities among the tested mutants, showed the lowest stability (47.5 Ϯ tionship between the protein structure of CYP2A6 and the catalytic
0.1°C). It was recently reported that the L188P mutation is associated activity toward coumarin (Lewis et al., 2006; Sansen et al., 2007; Li et al.,
with a destabilizing effect and causes a large conformational change in 2009). Using the docking model complexes, previous studies assuredly
the enzyme during catalysis (Fasan et al., 2008). This conformational gave close insight into the best pose orientation of coumarin docked in the
change seems to have a significant affect on the specificity and activity of desired orientation for hydroxylation in the topological aspect. The key
the enzyme. Therefore, the mutant with the highest activity showed the residues involved in the formation of the active-site pocket of P450 can
lowest thermal stability. The mutants that contained substitutions of form a hydrogen bond with the carbonyl oxygen of coumarin, which
residues near L188 (F162I/K187E, F162I/K187E/M237I, and F162I/ orientates the molecule for hydroxylation, because the relevant oxidizable
K187E/L188P/M237I) also showed a relatively low stability. On the carbon in coumarin lies directly above the heme iron in the model
other hand, the F162I/M237I and F162I/E228K mutants showed in- complex. The actual position of coumarin hydroxylation depends on the
creased stability compared with the wild-type CYP102A1 (54.4 Ϯ 0.4 facility of approximation to the heme iron and the disposition of key
and 55.9 Ϯ 0.3°C, respectively; Supplemental Fig. S6).
hydrogen bond donor residues fairly close to the preferred sites of
Models of the Docking of Coumarin to CYP102A1 and Its metabolism. As expected, coumarin binds to the active site of the
Mutants. The findings presented in the wet experiments above suggest CYP102A1 mutants with a similar pattern. All docked coumarin com-
that the structural basis for coumarin metabolism can be elucidated by the plexes have a similar orientation to the X-ray structure of the human
docking of coumarin to its probable binding site (either I or II) in the microsomal CYP2A6 complexed with coumarin (Protein Data Bank code
CYP102A1 mutants, F162I/H236R and F162I/M185T/L188P/M237I 1Z10) (Yano et al., 2005). Two threonine residues (Thr260 and Thr268)
FIG. 3. Docking of coumarin to CYP102A1
mutant models and analysis of binding com-
plex of the mutants with coumarin through
molecular dynamics simulation. A, the resi-
dues in red represent the residues mutated in
the CYP102A1 F162I/H236R mutant, and the
residues in blue represent the residues mu-
tated in the F162I/M185T/L188P/M237I mu-
tant. B, the two pockets for binding site I (for
coumarin 7-hydroxylation) and binding site II
(for coumarin 3-hydroxylation) are indicated by
the cavity holes. The active sites of wild-type
CYP102A1 (Protein Data Bank code 1BU7,
gray) and its mutants (purple), F162I/H236R (C
and D) and F162I/M185T/L188P/M237I (E and
F), are shown. The green and salmon colors
represent coumarin and heme, respectively. Al-
though similar active-site topologies have been
obtained for wild-type CYP102A1 and the
F162I/H236R and F162I/M185T/L188P/M237I
mutants when the active sites are superimposed,
the spatial orientation of the active sites of the
two mutants is somewhat different from that
of the wild-type. The average distances between
the carbonyl oxygen and the Thr260 OG1 atom
in the simulated coumarin complex, which ori-
entates the molecule for 7-hydroxylation, were
3.57 and 2.52 Å for F162I/H236R (C) and
F162I/M185T/L188P/M237I (E), respectively.
The average distance between the carbonyl ox-
ygen, which serves to orientate the coumarin for
3-hydroxylation, and the OG1 atoms of Thr268
was 2.92 Å and 2.57 Å for F162I/H236R (D)
and F162I/M185T/L188P/M237I (F), respec-
tively. For the simulation in binding site I, the
average distance from coumarin C-7 to the Fe of
heme in the orientation of 7-hydroxylation is
3.67 and 2.94 Å for F162I/H236R (C) and
F162I/M185T/L188P/M237I (E), respectively.
For the simulation in binding site II, the average
distance from coumarin C-3 to the Fe of heme
in the orientation of 3-hydroxylation is 3.01 and
2.45
Å for F162I/H236R (D) and F162I/
M185T/L188P/M237I (F), respectively.

738
PARK ET AL.
in the two CYP102A1 mutants form a hydrogen bond with the carbonyl marin and the heme pocket in each simulated complex structure is
oxygen of the coumarin ester group, which orientates the molecule for 7- considered to be the average EHB value of all snapshots taken from the
and 3-hydroxylation. The relevant hydroxylation carbon atom in couma- stable MD trajectory. The total hydrogen-bonding energies (Ϫ5.5 and
rin lies directly above the heme iron in the modeled complexes of the Ϫ6.2 kcal/mol for the F162I/H236R and F162I/M185T/L188P/M237I
F162I/H236R and F162I/M185T/L188P/M237I mutants with coumarin. mutants, respectively) for coumarin 3-hydroxylation estimated in this
There also tends to be a ␲-␲ stacking interaction between coumarin and way are systematically higher (i.e., more negative) than the corre-
Phe87 in the two CYP102A1 mutants.
sponding total hydrogen-bonding energies (Ϫ4.6 and Ϫ4.9 kcal/mol
Analysis of the Binding Complex of CYP102A1 with Coumarin for the F162I/H236R and F162I/M185T/L188P/M237I mutants, re-
through MD Stimulation. To more reliably explain why the two spectively) estimated in a way that is suitable for the catalysis of
CYP102A1 mutants show a possibly higher catalytic efficiency against 7-hydroxylation. Moreover, the two sets of total EHB values are
coumarin than wild-type CYP102A1, we performed MD simulations on qualitatively consistent with each other in terms of the relative hy-
coumarin binding with the two mutants. A detailed analysis of the drogen-bonding strengths in the four simulated CYP102A1 com-
MD-simulated structures of the two CYP102A1 mutants in complex with plexes. In particular, the two sets of total EHB values consistently
coumarin revealed that Thr260 and Thr268 are the key residues in the reveal that the overall strength of the hydrogen bonds between the
structural change required for 7- and 3-hydroxylation, respectively. Cou- carbonyl oxygen of coumarin and the heme pocket in the simulated
marin binding to the active site of the CYP102A1 mutants caused coumarin-binding complex structure of the F162I/M185T/L188P/
considerable changes in the positions of the Thr260, Ala264, Thr268, and M237I mutant is significantly stronger than that of the F162I/H236R
Leu437 residues compared with an adjacent region of the heme pocket in mutant. Hydrogen bonding between the substrate and the active site
the wild-type CYP102A1. These changes followed hydrogen bond break- residue was suggested to be important for CYP2A6-catalyzed reac-
age between Ala364 N and Thr268 OG1, together with a ␲-␲ stacking tions (Wu et al., 2005). These results suggest a clear correlation
interaction between the aromatic ring of coumarin and the side-chain of between stabilization in binding site II and the catalytic efficiency of
Phe87. Compared with the simulated complexes of F162I/H236R or the F162I/M185T/L188P/M237I mutant for coumarin hydroxylation:
F162I/M185T/L188P/M237I and coumarin, the average distances be- the more stable the structure of binding site II, the lower the energy
tween the coumarin carbonyl oxygen and Thr260/Thr268 OG1s in the barrier, and the higher the catalytic efficiency.
F162I/M185T/L188P/M237I mutant/coumarin complex are all shorter
Analysis of the Distance between the Heme Iron and Oxidizable
than in the F162I/H236R mutant/coumarin complex. Furthermore, cou- Carbons of Coumarin. Verifying the binding mode of coumarin that
marin binding to the F162I/M185T/L188P/M237I mutant is very similar is most likely to bind the two CYP102A1 mutants tightly depends on
to the binding to the F162I/H236R mutant, but the positional change of the analysis of the distance between possible oxidizable carbons (C-3
the Ala264/Leu437 residues and the spatial alteration of the nonpolar- or C-7) of coumarin and the heme iron with its ferric resting state in
binding cavity caused by the coumarin binding are significant only in the the two CYP102A1 mutants. The hydrogen bond interactions of the
F162I/M185T/L188P/M237I mutant. Thus, these data suggest that cou- carbonyl oxygen from coumarin with the surrounding residues at the
marin binds more strongly to the F162I/M185T/L188P/M237I mutant active site, as described above, are also very important for verifying
than to the F162I/H236R mutant. The energy barrier for coumarin hy- which binding mode would be most likely to bind the two mutants
droxylation in the F162I/M185T/L188P/M237I mutant might be slightly tightly.
lower than that in the F162I/H236R mutant. Because the mutated resi-
Because the interaction between oxidizable carbons and the Fe
dues in the F162I/H236R and the F162I/M185T/L188P/M237I mutants of heme plays an important role as coumarin binds to the F162I/
stay outside of the active sites and do not directly contact the coumarin H236R mutant, the distances from coumarin C-7 and C-3 to the Fe
molecule, these residues were initially expected to further increase or of heme were examined along all of the MD simulations (Supple-
decrease the free space of the active site pocket so that coumarin binding mental Fig. S7, A and B). For the simulation in binding site I, the
would be easier for the CYP102A1 mutants than for wild-type distance from coumarin C-7 to the Fe of heme in the orientation of
CYP102A1.
7-hydroxylation stays below 3.7 Å, and the average is 3.67 Å. For
The binding energy calculated via MD simulation suggests that the the simulation in binding site II, the distance from coumarin C-3 to
orientation of coumarin in binding site II for 3-hydroxylation is the Fe of heme in the orientation of 3-hydroxylation is between 2.9
significantly more stable than the orientation of coumarin in binding and 3.1 Å, with an average of 3.01 Å. Therefore, the binding mode
site I for 7-hydroxylation. The stability of the coumarin orientation is that leads to 7-hydroxylation contributes much less to the binding
due to a significant increase in the hydrophobic interaction between of coumarin toward the F162I/H236R mutant than to that of
coumarin and the binding surface of the mutant enzyme. These results 3-hydroxylation. The difference in the interaction energy of the
suggest that the coumarin-binding structure in binding sites I and II of coumarin complex with the F162I/H236R mutant at the two bind-
the F162I/M185T/L188P/M237I mutant should also be more stable ing sites is Ϫ12.24 kcal/mol. These data indicate that the binding
than that of the F162I/H236R mutant (Supplemental Table S2). The mode that leads to 3-hydroxylation in binding site II is more
average distance between the carbonyl oxygen and the Thr260 OG1 favorable than 7-hydroxylation in binding site I of the F162I/
atom in the simulated coumarin complex, which orientates the mol- H236R mutant (Supplemental Table S2).
ecule for 7-hydroxylation, was 3.57 and 2.52 Å for the F162I/H236R
The time dependence of the distances between oxidizable carbons
and the F162I/M185T/L188P/M237I mutants, respectively (Fig. 3, C of coumarin and the heme iron in the F162I/M185T/L188P/M237I
and E). The average distances between the carbonyl oxygen, which mutant was also examined (Supplemental Fig. S7, C and D). In
serves to orientate the coumarin for 3-hydroxylation, and the OG1 binding site I, the distance from coumarin C-7 to the Fe of heme in the
atoms of Thr268 were 2.92 and 2.57 Å for the F162I/H236R and orientation of 7-hydroxylation is shorter than 3.0 Å, and the average
F162I/M185T/L188P/M237I mutants, respectively (Fig. 3, D and F). is 2.94 Å. In binding site II, distance from coumarin C-3 to the Fe of
A hydrogen-bonding energy (EHB) value can be evaluated for each heme in the orientation of 3-hydroxylation is always shorter than 2.5
hydrogen bond with the carbonyl oxygen of coumarin for each snap- Å, and the average is 2.45 Å. When the coumarin is docked in a
shot of the MD-simulated structure. The estimated total EHB of the reactive binding orientation that leads to 3-hydroxylation, the inter-
MD-simulated hydrogen bonds between the carbonyl oxygen of cou- action between coumarin C-3 and the Fe of heme would be much

BACTERIAL P450 BM3 MUTANTS WITH HUMAN P450 ENZYME ACTIVITY
739
Damsten MC, van Vugt-Lussenburg BM, Zeldenthuis T, de Vlieger JS, Commandeur JN, and
Vermeulen NP (2008) Application of drug metabolising mutants of cytochrome P450 BM3
(CYP102A1) as biocatalysts for the generation of reactive metabolites. Chem Biol Interact
171:96–107.
stronger than that between coumarin C-7 and the Fe of heme. In
comparison to the binding interaction energy calculation in the com-
plexes with the F162I/M185T/L188P/M237I mutant, the binding
mode that leads to 3-hydroxylation in the binding site II is favored at
approximately 9.26 kcal/mol, in contrast to 7-hydroxylation in binding
site I (Supplemental Table S2).
Fasan R, Meharenna YT, Snow CD, Poulos TL, and Arnold FH (2008) Evolutionary history of
a specialized P450 propane monooxygenase. J Mol Biol 383:1069–1080.
Gillam EM, Aguinaldo AM, Notley LM, Kim D, Mundkowski RG, Volkov AA, Arnold FH,
Soucek P, DeVoss JJ, and Guengerich FP (1999) Formation of indigo by recombinant
mammalian cytochrome P450. Biochem Biophys Res Commun 265:469–472.
Guengerich FP, Kim DH, and Iwasaki M (1991) Role of human cytochrome P-450 IIE1 in the
oxidation of many low molecular weight cancer suspects. Chem Res Toxicol 4:168–179.
Halgren TA (1999) MMFF VII. Characterization of MMFF94, MMFF94s, and other widely
available force fields for conformational energies and for intermolecular-interaction energies
and geometries. J Comput Chem 20:730–748.
Hirano Y, Uehara M, Saeki K, Kato T, Takahashi K, and Mizutani T (2002) The influence of
quinolines on coumarin 7-hydroxylation in bovine liver microsomes and human CYP2A6.
J Health Sci 48:118–125.
Huang B and Schroeder M (2008) Using protein binding site prediction to improve protein
docking. Gene 422:14–21.
The difference of interaction energies for the two CYP102A1
mutants at binding sites I and II average Ϫ15.48 and Ϫ12.50 kcal/mol,
respectively (Supplemental Table S2). These data indicate that there is
a stronger interaction between coumarin and the F162I/M185T/
L188P/M237I mutant. The interaction between oxidizable carbons of
coumarin and the heme iron may contribute greatly to tight binding of
coumarin toward the CYP102A1 mutants, as well as hydrogen-
bonding interaction with the surrounding residues at the active site.
In conclusion, we examined the possibility that CYP102A1
mutants with indole oxidation activity can have the catalytic ac-
tivities of human P450 enzymes. Error-prone PCR was carried out
on the heme domain-coding region of the wild-type gene to gen-
erate a CYP102A1 DNA library. The library was transformed into
E. coli to express the P450 mutants. A colorimetric colony-based
method was used for primary screening of the mutants, and the
P450 activities were measured at the whole-cell level. Some of the
blue colonies, but not the white colonies, showed apparent oxida-
tion activity toward typical P450 substrates, including coumarin,
7-ethoxycoumarin, 7-ethoxyresorufin, phenacetin, chlorzoxazone,
and p-nitrophenol. These results indicate that indigo formation
provides a simple assay for identifying CYP102A1 mutants that
have a greater potential for human P450 activity. Our computa-
tional findings suggest a correlation between the stabilization of
the binding site and the catalytic efficiency of the CYP102A1
mutants toward coumarin: the more stable the structure in the
binding site, the lower the energy barrier, and the higher the
catalytic efficiency. Taken together, these data suggest that
CYP102A1 mutants engineered by random mutagenesis can be
developed as biocatalysts for industrial applications of human
P450 activities.
Huey R, Morris GM, Olson AJ, and Goodsell DS (2007) A semiempirical free energy force field
with charge-based desolvation. J Comput Chem 28:1145–1152.
Kim D, Wu ZL, and Guengerich FP (2005) Analysis of coumarin 7-hydroxylation activity of
cytochrome P450 2A6 using random mutagenesis. J Biol Chem 280:40319–40327.
Kim DH, Ahn T, Jung HC, Pan JG, and Yun CH (2009) Generation of the human metabolite
piceatannol from the anticancer-preventive agent resveratrol by bacterial cytochrome P450
BM3. Drug Metab Dispos 37:932–936.
Kim DH, Kim KH, Kim DH, Liu KH, Jung HC, Pan JG, and Yun CH (2008) Generation of
human metabolites of 7-ethoxycoumarin by bacterial cytochrome P450 BM3. Drug Metab
Dispos 36:2166–2170.
Kim KH, Isin EM, Yun CH, Kim DH, and Guengerich FP (2006) Kinetic deuterium isotope
effects for 7-alkoxycoumarin O-dealkylation reactions catalyzed by human cytochromes P450
and in liver microsomes. Rate-limiting C-H bond breaking in cytochrome P450 1A2 substrate
oxidation. FEBS J273:2223–2231.
Lewis DF, Ito Y, and Lake BG (2006) Metabolism of coumarin by human P450s: a molecular
modelling study. Toxicol In Vitro 20:256–264.
Li HM, Mei LH, Urlacher VB, and Schmid RD (2008) Cytochrome P450 BM-3 evolved by
random and saturation mutagenesis as an effective indole-hydroxylating catalyst. Appl Bio-
chem Biotechnol 144:27–36.
Li QS, Schwaneberg U, Fischer P, and Schmid RD (2000) Directed evolution of the fatty-acid
hydroxylase P450 BM-3 into an indole-hydroxylating catalyst. Chemistry 6:1531–1536.
Li W, Ode H, Hoshino T, Liu H, Tang Y, and Jiang HL (2009) Reduced catalytic activity of P450
2A6 mutants with coumarin: a computational investigation. J Chem Theory Comput 5:1411–1420.
Munro AW, Leys DG, McLean KJ, Marshall KR, Ost TW, Daff S, Miles CS, Chapman SK,
Lysek DA, Moser CC, et al. (2002) P450 BM3: the very model of a modern flavocytochrome.
Trends Biochem Sci 27:250–257.
Nakamura K, Martin MV, and Guengerich FP (2001) Random mutagenesis of human cyto-
chrome P450 2A6 and screening with indole oxidation products. Arch Biochem Biophys
395:25–31.
Omura T and Sato R (1964) The carbon monoxide-binding pigment of liver microsomes. II.
Solubilization, purification, and properties. J Biol Chem 239:2379–2385.
Pearlman DA, Case DA, Caldwell JW, Ross WS, Cheatham TE, Debolt S, Ferguson D, Seibel
G, and Kollman P (1995) Amber, a package of computer-programs for applying molecular
mechanics, normal-mode analysis, molecular-dynamics and free-energy calculations to sim-
ulate the structural and energetic properties of molecules. Comput Phys Commun 91:1–41.
Sali A and Blundell TL (1993) Comparative protein modelling by satisfaction of spatial
restraints. J Mol Biol 234:779–815.
Sansen S, Hsu MH, Stout CD, and Johnson EF (2007) Structural insight into the altered substrate
specificity of human cytochrome P450 2A6 mutants. Arch Biochem Biophys 464:197–206.
Schmidt MW, Baldridge KK, Boatz JA, Elbert ST, Gordon MS, Jensen JH, Koseki S, Matsunaga
N, Nguyen KA, Su S, et al. (1993) General atomic and molecular electronic structure system.
J Comput Chem 14:1347–1363.
School of Biological Sciences and
Technology, Chonnam National
University, Gwangju, Republic of Korea
(S.-H.P., Do.-H.K., Da.-H.K., C.-H.Y.);
Systems Microbiology Research Center,
Korea Research Institute of Bioscience
and Biotechnology, Daejeon, Republic
of Korea (Doo.K., H.-C.J., J.-G.P.);
Department of Biochemistry, College
of Veterinary Medicine, Chonnam
National University, Gwangju,
SUN-HA PARK
DONG-HYUN KIM
DOOIL KIM
DAE-HWAN KIM
HEUNG-CHAE JUNG
JAE-GU PAN
Schymkowitz J, Borg J, Stricher F, Nys R, Rousseau F, and Serrano L (2005) The FoldX web
server: an online force field. Nucleic Acids Res 33:W382–388.
TAEHO AHN
DONGHAK KIM
CHUL-HO YUN
Sevrioukova IF, Li H, Zhang H, Peterson JA, and Poulos TL (1999) Structure of a cytochrome
P450-redox partner electron-transfer complex. Proc Natl Acad Sci USA 96:1863–1868.
Whitehouse CJ, Bell SG, Tufton HG, Kenny RJ, Ogilvie LC, and Wong LL (2008) Evolved
CYP102A1 (P450BM3) variants oxidise a range of non-natural substrates and offer new
selectivity options. Chem Commun (Camb) 966–968.
Wu ZL, Podust LM, and Guengerich FP (2005) Expansion of substrate specificity of cytochrome
P450 2A6 by random and site-directed mutagenesis. J Biol Chem 280:41090–41100.
Yano JK, Hsu MH, Griffin KJ, Stout CD, and Johnson EF (2005) Structures of human
microsomal cytochrome P450 2A6 complexed with coumarin and methoxsalen. Nat Struct Mol
Biol 12:822–823.
Yun CH, Kim KH, Calcutt MW, and Guengerich FP (2005) Kinetic analysis of oxidation of
coumarins by human cytochrome P450 2A6. J Biol Chem 280:12279–12291.
Yun CH, Kim KH, Kim DH, Jung HC, and Pan JG (2007) The bacterial P450 BM3: a prototype
for a biocatalyst with human P450 activities. Trends Biotechnol 25:289–298.
Yun CH, Miller GP, and Guengerich FP (2000) Rate-determining steps in phenacetin oxidations
by human cytochrome P450 1A2 and selected mutants. Biochemistry 39:11319–11329.
Yun CH, Shimada T, and Guengerich FP (1991) Purification and characterization of human liver
microsomal cytochrome P-450 2A6. Mol Pharmacol 40:679–685.
Republic of Korea (T.A.); and
Department of Biological Sciences,
Konkuk University, Seoul,
Republic of Korea (Don.K.)
References
Born SL, Caudill D, Fliter KL, and Purdon MP (2002) Identification of the cytochromes P450
that catalyze coumarin 3,4-epoxidation and 3-hydroxylation. Drug Metab Dispos 30:483–487.
Burke MD and Mayer RT (1983) Differential effects of phenobarbitone and 3-methylcholan-
threne induction on the hepatic microsomal metabolism and cytochrome P-450-binding of
phenoxazone and a homologous series of its n-alkyl ethers (alkoxyresorufins). Chem Biol
Interact 45:243–258.
Zhang ZG, Liu Y, Guengerich FP, Matse JH, Chen J, and Wu ZL (2009) Identification of amino
acid residues involved in 4-chloroindole 3-hydroxylation by cytochrome P450 2A6 using
screening of random libraries. J Biotechnol 139:12–18.
Case DA, Cheatham TE 3rd, Darden T, Gohlke H, Luo R, Merz KM Jr, Onufriev A, Simmerling
C, Wang B, and Woods RJ (2005) The Amber biomolecular simulation programs. J Comput
Chem 26:1668–1688.
Address correspondence to: Prof. Chul-Ho Yun, School of Biological Sci-
ences and Technology, Chonnam National University, Gwangju 500-757, Repub-
lic of Korea. E-mail: chyun@jnu.ac.kr
Chang TK, Crespi CL, and Waxman DJ (2006) Spectrophotometric analysis of human CYP2E1-
catalyzed p-nitrophenol hydroxylation. Methods Mol Biol 320:127–131.