Simultaneous measurement of CYP1A2 activity, regioselectivity, and coupling: Implications for environmental sensitivity of enzyme-substrate binding

Article abstract of DOI:10.1016/j.abb.2010.10.002

The cytochrome P450 (CYP) reaction mechanism often yields a broad array of coupled and uncoupled products from a single substrate. While it is well known that reaction conditions can drastically affect the rate of P450 catalysis, their effects on regioselectivity and coupling are not well characterized. To investigate such effects, the CYP1A2 oxidation of 7-ethoxymethoxy-3- cyanocoumarin (EOMCC) was examined as a function of buffer type, buffer concentration, pH, and temperature. A high-throughput, optical method was developed to simultaneously measure the rate of substrate depletion, NADPH depletion, and generation of the O-dealkylated product. Increasing the phosphate buffer concentration and temperature increased both the NADPH and EOMCC depletion rates by 6-fold, whereas coupling was constant at 7.9% and the regioselectivity of O-dealkylation to other coupled pathways was constant at 21.7%. Varying the buffer type and pH increased NADPH depletion by 2.5-fold and EOMCC depletion by 3.5-fold; however, neither coupling nor regioselectivity was constant, with variations of 14.4% and 21.6%, respectively. Because the enzyme-substrate binding interaction is a primary determinant of both coupling and regioselectivity, it is reasonable to conclude that ionic strength, as varied by the buffer concentration, and temperature alter the rate without affecting binding while buffer type and pH alter both.

Full text of DOI:10.1016/j.abb.2010.10.002

Archives of Biochemistry and Biophysics 505 (2011) 186–193
Contents lists available at ScienceDirect
Archives of Biochemistry and Biophysics
journal homepage: www.elsevier.com/locate/yabbi
Simultaneous measurement of CYP1A2 activity, regioselectivity, and coupling:
Implications for environmental sensitivity of enzyme–substrate binding
⇑
Matthew J. Traylor, Jack Chai, Douglas S. Clark
Department of Chemical and Biomolecular Engineering, University of California, Berkeley, CA 94720, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
The cytochrome P450 (CYP) reaction mechanism often yields a broad array of coupled and uncoupled
products from a single substrate. While it is well known that reaction conditions can drastically affect
the rate of P450 catalysis, their effects on regioselectivity and coupling are not well characterized. To
investigate such effects, the CYP1A2 oxidation of 7-ethoxymethoxy-3-cyanocoumarin (EOMCC) was
examined as a function of buffer type, buffer concentration, pH, and temperature. A high-throughput,
optical method was developed to simultaneously measure the rate of substrate depletion, NADPH
depletion, and generation of the O-dealkylated product. Increasing the phosphate buffer concentration
and temperature increased both the NADPH and EOMCC depletion rates by 6-fold, whereas coupling
was constant at 7.9% and the regioselectivity of O-dealkylation to other coupled pathways was constant
at 21.7%. Varying the buffer type and pH increased NADPH depletion by 2.5-fold and EOMCC depletion
by 3.5-fold; however, neither coupling nor regioselectivity was constant, with variations of 14.4% and
Received 25 August 2010
and in revised form 4 October 2010
Available online 8 October 2010
Keywords:
P450 regioselectivity
P450 coupling
CYP1A2
Cyanocoumarin
21.6%, respectively. Because the enzyme–substrate binding interaction is a primary determinant of both
coupling and regioselectivity, it is reasonable to conclude that ionic strength, as varied by the buffer
concentration, and temperature alter the rate without affecting binding while buffer type and pH alter
both.
Ó 2010 Elsevier Inc. All rights reserved.
Introduction
substrate oxidation, and uncoupled products where the supplied
electrons are lost to reduced oxygen species [8]. Thus, there are
Human cytochrome P4501 enzymes are important to the phar-
maceutical and biotechnology industries due to their central role
in drug metabolism, ability to generate and optimize lead com-
pounds, and specialized oxidative abilities [1]. The use of P450 en-
zymes in these capacities is complicated, however, by the wide
variability of reaction rates observed in reconstituted in vitro sys-
tems [2,3]. This variability has been linked to the composition of
the reaction medium [4–7]. Because the P450 reaction mechanism
is capable of generating multiple products, it is not known whether
the variability in rates as a function of medium composition is
accompanied by significant changes in selectivity.
two relevant reaction selectivities in P450 systems: the electron
selectivity, commonly referred to as the coupling efficiency, which
is the ratio of coupled to uncoupled product-generation rates; and
the regioselectivity, which is the ratio of two different possible
coupled product-generation rates. While it is well known that
reaction conditions can affect P450 activity over orders of magni-
tude [6], it is currently unclear how P450 enzymes interact with
environmental parameters such as pH, ionic strength, temperature,
and small molecule additives to modulate these two selectivities.
An improved understanding of these interactions could lead to
more accurate in vitro – in vivo metabolic correlations facilitating
drug development [3], and assist in the development and optimiza-
tion of synthetic schemes in the biotechnology sector [2,9].
Multiple P450 reaction products comprise two classes: coupled
products where the electrons supplied by NADPH are coupled to
P450 regioselectivity is thought to be controlled by a combina-
tion of substrate positioning with respect to the active oxidant and
the relative reactivity of different potential oxidation sites on the
substrate [10]. Regioselectivity of a very mobile substrate, one able
to bind in many orientations, is likely governed primarily by the
relative reactivity of the different chemical moieties in proximity
to the heme. In contrast, regioselectivity of a tightly-bound sub-
strate, one able to bind in only a few orientations, is presumably
determined primarily by the relative occupancy of the different
binding orientations [11,12]. Thus, an experimental parameter that
⇑
Corresponding author. Fax: +1 510 643 1228.
E-mail address: clark@berkeley.edu (D.S. Clark).
Abbreviations used: 7HCC, 7-hydroxy-3-cyanocoumarin; CYP, cytochrome P450;
1
P450, cytochrome P450; EOMCC, 7-ethoxymethoxy-3-cyanocoumarin; PBO, pipero-
nyl butoxide; NADPH, nicotinamide adenine dinucleotide phosphate; LCMS, liquid
chromatography tandem mass spectrometry; MS/MS, tandem mass spectrometry;
7HCC, initial rate of 7-hydroxy-3-cyanocoumarin generation; -rEOMCC, initial rate of 7-
ethoxymethoxy-3-cyanocoumarin depletion; -rNADPH initial rate of nicotinamide
r
,
adenine dinucleotide phosphate depletion; HEPES, 4-(2-hydroxyethyl)-1-piperazi-
neethanesulfonic acid; RFU, relative fluorescence units.
0
003-9861/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.abb.2010.10.002

M.J. Traylor et al. / Archives of Biochemistry and Biophysics 505 (2011) 186–193
187
alters regioselectivity would be expected to modify the occupancy
Materials and methods
or orientation of different binding modes of the substrate or
differentially modify the reactivity of the substrate. The latter has
been clearly demonstrated in kinetic isotope effect studies where
the substitution of a heavy atom at one oxidation site of a mobile
substrate can alter the regioselectivity toward reaction at other
sites [13]. Although previous studies have investigated the effect
of buffer concentration, pH, and magnesium concentration on the
regioselectivity of P450 reactions, the regulation of P450 selectivity
by reaction conditions remains poorly understood. Buffer and mag-
nesium concentrations have been found to alter catalytic activity
but not regioselectivity, while pH has been shown to alter both
Materials
Piperonyl butoxide (PBO), potassium phosphate, nicotinamide
adenine dinucleotide phosphate (NADPH), and acetonitrile were
purchased from Sigma (St. Louis, MO). Microsomes from baculovi-
rus-infected insect cells (Baculosomes) expressing human CYP1A2
and rabbit cytochrome P450 reductase (Lot 484429C), EOMCC,
and 7HCC were purchased from Invitrogen (Carlsbad, Ca). All
chemicals were used as received. All fluorescence data were ac-
quired on a Spectramax M2 plate reader (Molecular Devices, Sun-
nyvale, CA).
[
14–17].
Electron selectivity, or coupling, has also been linked to sub-
strate–enzyme binding interactions. The structural complementar-
ity of the substrate and P450 active site has been shown to be an
important determinant of coupling [18,19], where noncomplemen-
tarity results in improperly positioned water in the active site. These
water molecules are thought to disrupt the proton delivery machin-
ery required for the successive protonation of the ferric-peroxo
intermediate, leading to a decrease in coupling efficiency [20].
Guengerich points out that substrate dissociation rates can be large
compared to catalytic rates, and coupling efficiency can be
decreased by substrate dissociation after initiation of catalysis but
Liquid chromatography mass spectrometry
EOMCC oxidation reactions were conducted in 120 mM pH 8
potassium phosphate buffer with 25 nM CYP1A2 Baculosomes
with or without the inhibitor PBO. Reactions were initiated with
the simultaneous addition of NADPH and EOMCC to a concentra-
tion of 450
200 L. Reactions were prepared and conducted at ambient tem-
perature for 1 h and were quenched with 100 L ice cold acetoni-
trile then centrifuged at 4 °C for 10 min at 14,000g. Products were
resolved using reverse-phase Alltech Prevail C18
lM and 100 lM, respectively, to a final volume of
l
l
before oxidation [21]. Thus,
a
parameter that modifies the
a
3 lm
substrate–enzyme interaction and alters the off-rate or the packing
of the active site is expected to affect coupling. Notably, an increase
in buffer concentration was shown to increase coupling in the
benzphetamine N-demethylation activity of CYP2B4enriched rabbit
liver microsomes [22]; however, there are few studies in the
literature examining the effect of reaction conditions on
uncoupling.
3.0 ꢀ 150 mm column with a 0.4 ml/min flow rate and a gradient
of 10–80% acetonitrile over 40 min with a constant 0.05% concen-
tration of formic acid. LCMS data were collected on an Agilent
1100 LC/MSD equipped with an electrospray ion source. Injec-
tions of inhibited, uninhibited, and no-protein control reactions
were conducted separately in positive and negative mode. The
quadrupole ion analyzer MS was operated with a fragmentor set-
ting of 70 V, 3 kV capillary, 35 psig nebulizer pressure, and with
In general, very little has been published on the relationship
between environment and P450 selectivity. In particular, no previ-
ous studies have investigated the effect of reaction conditions on
CYP1A2 regioselectivity or coupling. This void in the literature
can be partially attributed to the large number of measurements
needed to calculate these selectivities. To address this limitation,
we developed a ‘‘one pot” method utilizing a single experimental
technique, fluorescence intensity measurements, to simulta-
neously quantify activity, regioselectivity, and coupling. Depletion
rates of NADPH and of an optically active substrate are analyzed in
concert with the generation rate of at least one product. This
method should be generally applicable to any substrate that is
optically distinguishable from NADPH and one or more of its prod-
ucts. Here the method is demonstrated with the CYP1A2 oxidation
of EOMCC to 7-hydroxy-3-cyanocoumarin (7HCC), which is a par-
ticularly challenging case study because the substrate’s fluores-
cence spectrum overlaps with the spectrum of NADPH. In
addition, due to the broad reactivity of cyanocoumarins [23], this
model reaction can be used for many human drug metabolizing
P450 enzymes allowing for a more direct comparison of selectivity
data between isozymes.
2
N drying gas set to 300 °C and 9.5 L/min. LCMS/MS spectra were
acquired using a ThermoFisher Scietific LTQ Orbitrap XL mass
spectrometer in positive and negative modes. The flow rate was
0.2 mL/min and the mass spectrometer source parameters were
used as follows: spray voltage, 3.0 kV; capillary temperature,
275 °C; capillary voltage, 44 V (positive mode) and ꢁ50 V (nega-
tive mode); tube lens, 125 V (positive mode) and ꢁ120 V (nega-
tive mode).
Activity and selectivity assays
Six replicates were prepared with 25 nM CYP1A2 Baculosomes
at the buffer condition noted in Fig. 4 with and without 1 mM
PBO. NADPH solutions were prepared fresh daily and the concen-
tration was checked via absorbance using a molar extinction coef-
ꢁ1
ꢁ1
ficient of 6.22 mM cm . Reactions were incubated for 20 min at
the required temperature before initiation with the simultaneous
addition of NADPH and EOMCC to 250
to a final volume of 100 L.
lM and 10 lM, respectively
l
The reaction was followed via fluorescence intensity measure-
ments in a Spectramax M2 plate reader. Fluorescence intensity
was simultaneously recorded at three excitation/emission wave-
lengths: 340 nm/460 nm, 340 nm/520 nm, 409 nm/460 nm. Mod-
els correlating fluorescence intensity with the concentration of
EOMCC, NADPH, and 7HCC (Eqs. (1)–(3)) were generated in JMP
5.1 using least-squares regression. The coefficients of Eqs. (1)–
(3) were determined using on-plate standards at each reaction
condition tested. The on-plate standards included a three-level,
full factorial array of mixtures of NADPH and EOMCC to quantify
fluorescence interactions. Significant model parameters were
identified with the Backward Elimination method using a thresh-
old P-value of 5%. Correlations typically took the form shown
below.
To explore the relationship between environment and selec-
tivity, the rates of NADPH depletion, EOMCC depletion, and
7
HCC generation were measured as a function of buffer type, buf-
fer concentration, pH, and temperature in the CYP1A2 oxidation
of EOMCC, and the regioselectivity and coupling were compared
under these conditions. The results indicate that both regioselec-
tivity and coupling remain constant when temperature and buf-
fer concentration are varied but are altered when buffer type
and pH are varied. This indicates that temperature and ionic
strength, as varied by buffer concentration, affect the rate of
catalysis independent of the substrate–enzyme binding interac-
tion while pH and buffer type affect both the rates and the bind-
ing interaction.

1
88
M.J. Traylor et al. / Archives of Biochemistry and Biophysics 505 (2011) 186–193
RFU340=460 ¼ a
1
½EOMCCꢂ þ b
½7HCCꢂ
1
2
½EOMCCꢂ½NADPHꢂ þ c
1
½NADPHꢂ
for the generation of models correlating fluorescence intensity
with fluorophore concentration. Correlation parameters were
determined using on-plate standards at each condition tested
and were validated by comparing the predicted and measured
fluorescence intensities for mixtures of known concentrations of
þ d
1
ð1Þ
2
RFU340=520 ¼ a
2
½NADPHꢂ þ b
½NADPHꢂ þ c
2
½EOMCCꢂ
2
þ d
2
½NADPHꢂ½7HCCꢂ þ e
2
½7HCCꢂ
ð2Þ
ð3Þ
fluorophores, each yielding R values greater than 0.98 indicating
accurate fluorophore quantification (data not shown).
RFU409=460 ¼ a
3
½7HCCꢂ
These correlation equations were solved to yield each fluoro-
phore concentration as a function of time. Representative time tra-
jectories of the three fluorophore concentrations are shown in
Fig. 2 for reactions with and without the P450 suicide inhibitor,
PBO. Inhibition by PBO is hypothesized to occur via the formation
of a stable carbene linkage to the heme iron blocking further catal-
ysis by the P450 [24]. Degradation of EOMCC and NADPH occurs in
the presence of PBO, which can be attributed to both the instability
of these molecules and to reactions with other components of the
heterogeneous baculosome mixture. The background response of
the inhibited system, which contains all components except func-
tional P450, is thus crucial to isolate the contribution of only
CYP1A2 from other reactions occurring in the system, such as the
background oxidation of NADPH by the reductase. The absence of
any 7HCC generation in the presence of PBO also indicates that,
within the reaction mixture, only functional CYP1A2 can catalyze
the conversion of EOMCC to 7HCC. Monoexponential fits were ap-
plied to the corrected time trace data and initial rates were calcu-
lated for the degradation of EOMCC, the generation of 7HCC, and
the degradation of NADPH. The exponential fit does not imply a
mechanistic interpretation and was used because it closely repro-
duces the observed data (Fig. 2). Coupling is reported as the ratio of
EOMCC depletion to NADPH depletion, and regioselectivity is indi-
cated by the ratio of 7HCC generation to EOMCC depletion. A table
of all measured rates and calculated selectivities can be found in
the Supplemental materials.
The three coupled algebraic equations representing the depen-
dence of fluorescence intensity on fluorophore concentration at
each wavelength were solved for each data point in Matlab to yield
concentration trajectories with time. The background reactions
including PBO were subtracted out at each condition. The corrected
time trajectory data were fit with a monoexponential function
from which initial rates were calculated.
Results
Simultaneous in situ monitoring of EOMCC, 7HCC, and NADPH
The quantification of NADPH, EOMCC, and 7HCC via fluores-
cence intensity measurements is complicated in this system by
the overlap of the fluorescence spectra of EOMCC and NADPH.
While the emission of 7HCC is not affected by the presence of
the other fluorophores, the emission maximum of NADPH
(
460 nm) overlaps with a strong signal from EOMCC (Fig. 1). How-
ever, at longer emission wavelengths, the NADPH signal is more
distinct from that of the other fluorophores. The incorporation of
emission data from both 460 nm and 520 nm upon excitation at
3
40 nm into a standard least-squares regression algorithm allowed
Multiple coupled products in the CYP1A2 oxidation of EOMCC
To verify that the measured regioselectivity is governed by an
enzyme-mediated process and not simply the rearrangement of a
common intermediate into multiple products [25], the product
profile was investigated with mass spectrometry. The reaction
products of the CYP1A2 oxidation of EOMCC were analyzed with
LCMS in the presence and absence of PBO. Product peaks were
identifiable from background peaks as the product peaks disap-
peared in the presence of PBO. This was seen in three peaks eluting
at 18.9 min, 23.6 min, and 30.5 min, which yielded major ions at
m/z +236, ꢁ186, and ꢁ207, respectively (Fig. 3). Although the peak
at 30.5 min exhibited a base peak at m/z ꢁ207, another product ion
was present in this peak at a much lower abundance at m/z ꢁ234.
EOMCC eluted at 33.4 min and as expected exhibited a larger peak
with PBO than in its absence. The substrate peak co-eluted with an
EOMCC standard while the product at 23.6 min co-eluted with a
7
HCC standard (data not shown). The small peak at 33.4 min in
the m/z ꢁ186 chromatogram (Fig. 3b) arises from fragmentation
of the substrate EOMCC during ionization.
To determine whether CYP1A2 oxidizes EOMCC at multiple
sites, LCMS/MS analysis was conducted on EOMCC, 7HCC, and
the unknown m/z +236 product. Unfortunately, the m/z ꢁ234 prod-
uct could not be captured with LCMS/MS. Mass spectra, exact mass
data, chemical formula predictions, and spectra interpretation are
provided in the Supplemental materials. The loss of 30 amu fol-
lowed by 26 amu was identified in EOMCC as a fragmentation mo-
tif indicating the sequential loss of formaldehyde and ethylene
from the 7-ethoxymethoxy moiety of the substrate. This motif
was also seen in the fragmentation of the unknown m/z +236 prod-
uct, demonstrating an intact 7-ethoxymethoxy moiety. The O-deal-
kylation of this moiety generates 7HCC; therefore, it can be
Fig. 1. Emission spectra of 250
1
4
l
M NADPH, 10
00 mM potassium phosphate at pH 7.5 upon excitation at (a) 340 nm and (b)
09 nm.
lM EOMCC, and 1.2 lM 7HCC in

M.J. Traylor et al. / Archives of Biochemistry and Biophysics 505 (2011) 186–193
189
Fig. 2. Representative data for the depletion of EOMCC (a) and NADPH (b), and the generation of 7HCC (c), in the presence (s) and absence (d) of the inhibitor PBO. The data
with inhibition are subtracted from the data without inhibition to yield the corrected depletion of EOMCC (d), NADPH (e), and generation of 7HCC (f). A monoexponential
curve is fit to the data and used to obtain an initial rate.
Fig. 3. LCMS analysis of the reaction products of the CYP1A2 oxidation of EOMCC with single ion monitoring in negative mode of (a) m/z ꢁ207 and (b) m/z ꢁ186 and in
positive mode of (c) m/z +236 and (d) m/z +246. The filled peaks are data from reactions including the suicide inhibitor, PBO, while the empty peaks are the response without
inhibitor.

1
90
M.J. Traylor et al. / Archives of Biochemistry and Biophysics 505 (2011) 186–193
concluded that the m/z +236 product is generated from the oxida-
tion of EOMCC at a different site than 7HCC. Thus, EOMCC occupies
at least two binding orientations in the CYP1A2 active site with at
least two sites of oxidation on the substrate.
buffers, yielding maxima at pH 6.7, while r7HCC exhibited clearly
distinct behavior in Tris and phosphate buffers. In Tris–buffered
systems, r7HCC increased with decreasing pH down to the lowest
assayed pH of 7.1. In phosphate buffered systems, r7HCC exhibited
a maximum at pH 7.7.
A strong correlation was observed between -rNADPH and -rEOMCC
as a function of both temperature and buffer concentration, indi-
cating that coupling was not significantly changed by these param-
eters (Fig. 5a). Similarly, the correlation between -rEOMCC and r7HCC
indicates that these parameters also did not alter regioselectivity
Effect of reaction conditions on activity and selectivity
The initial rates of 7HCC generation (r7HCC), EOMCC depletion (-
rEOMCC), and NADPH depletion (-rNADPH) were measured in the reac-
tion of CYP1A2 on EOMCC as a function of buffer type, buffer con-
centration, pH, and temperature (Fig. 4). Each rate increased with
temperature, and appeared to be approaching a maximum at the
highest achievable temperature of 45 °C. Each rate also exhibited
a maximum at a phosphate buffer concentration between 250
and 375 mM. A notable difference in rate was apparent, however,
when varying the buffer type and pH, where -rEOMCC and -rNADPH
both showed qualitatively similar behavior in Tris and phosphate
(Fig. 5c). These correlations, showing constant selectivity, contrast
with the 6-fold change in activity observed with a variation in tem-
perature and buffer concentration at pH 6.7. The selectivity results
are illustrated in Scheme 1 where 7.9% of the reducing equivalents
from NADPH were used for coupled catalysis and 1.7% of the over-
all reducing equivalents or 21.7% of the depleted EOMCC was used
to generate 7HCC via O-dealkylation. Conversely, despite the qual-
itatively similar behavior of -rEOMCC and -rNADPH as a function of
buffer type and pH, much weaker correlations were observed for
both coupling and regioselectivity as a function of these parame-
ters (Fig. 5b and d). Coupling and regioselectivity varied by 14.4%
and 21.6%, respectively. It is important to mention that standard
assays containing known amounts of all three fluorophores indi-
cated that the fluorescence responses are not inherently correlated
(
data not shown). Thus the selectivity trends found in this study
are not a result of spurious correlations.
Discussion
Medium engineering offers a useful tool for the investigation of
P450 function because a tunable range of P450 responses can be
accessed by simply varying the reaction conditions. The practical
utility and potential generality of medium engineering has been
demonstrated by the large rate increases achieved with optimized
P450 reconstitution procedures [6]. The current study seeks to
investigate the connection between environment and P450 selec-
tivity using a novel method to simultaneously measure activity,
coupling, and regioselectivity.
LCMS analysis verified the presence of multiple coupled prod-
ucts and identified three product peaks. MS/MS analysis revealed
an intact ethoxymethoxy moiety on the product eluting at
1
8.9 min indicating this product is generated via a distinct mecha-
nism from the O-dealkylation product, 7HCC. Furthermore, the
mass and formula of the m/z +236 product are consistent with a
P450 mediated 3,4-epoxidation followed by a ring opening and
decarboxylation similar to the reaction of CYP1A2 with coumarin
[
26,27]. Interestingly, another product eluting over 11 min later
at 30.5 min was identified with a major peak in negative mode
LCMS at m/z ꢁ207 and a minor peak at m/z ꢁ234. This product
could not be captured in LCMS/MS analysis; nonetheless, the mass
is also consistent with an epoxidation and decarboxylation reac-
tion generating the m/z ꢁ234 product that presumably fragments
during ionization to form the major m/z ꢁ207 peak. Although,
the +236 and ꢁ234 products exhibit different elution times and
different ionization behavior, their similar masses suggest that
they may arise from the decarboxylation of a common intermedi-
ate generated from a common substrate-binding mode. Overall,
however, the MS/MS data indicate the presence of two distinct
products that are produced from at least two catalytically relevant
binding modes for EOMCC in the CYP1A2 active site. Thus, the reg-
ioselectivity determined in this system reflects the relative binding
of EOMCC in multiple orientations and the oxidation of multiple
sites on the substrate. The presence of multiple binding orienta-
tions, indicated by the oxidation of multiple sites on the molecule,
is typical of CYP1A2 substrates with caffeine [28] and phenacetin
[29] being well-studied examples.
Fig. 4. Initial rates of (a) 7HCC generation, (b) EOMCC depletion, and (c) NADPH
depletion as a function of (h) pH in 100 mM Tris buffer at 25 °C, (j) pH in 100 mM
phosphate buffer at 25 °C, (s) phosphate buffer concentration at pH 6.7 and 25 °C,
and (d) temperature in 250 mM phosphate buffer at pH 6.7. A table of these rates
can be found in Supplemental materials.

M.J. Traylor et al. / Archives of Biochemistry and Biophysics 505 (2011) 186–193
191
Fig. 5. Correlation of the initial rate of EOMCC depletion, -rEOMCC, with the initial rate of NADPH depletion, -rNADPH, depicting the coupling efficiency (a and b), and of the initial
rate of 7HCC generation, r7HCC, with the initial rate of EOMCC depletion, rEOMCC, depicting the regioselectivity (c and d). The data are from reactions varying the temperature
(d) and phosphate buffer concentration (s), and the phosphate pH (j) and Tris pH (h). A table of these selectivity data can be found in Supplemental materials.
in the present work where the measured rates showed signs of lev-
eling off at 45 °C.
The measured r7HCC was higher in phosphate than in Tris-buf-
fered systems. Similarly, Mäenpää et al. reported higher product-
generation rates in phosphate versus Tris–buffered systems for
the midazolam hydroxylation of human liver microsomes [16].
Other studies, however, have shown similar P450 activity in differ-
ent buffers. For example, the 7-pentoxyresorufin O-dealkylation
activity of rabbit CYP1A2 was comparable in sodium phosphate,
potassium acetate, and HEPES [31]. Interestingly, -rNADPH and -rE-
OMCC showed similar behavior in phosphate and Tris–buffered sys-
tems, demonstrating that buffer type can alter both the rate of
product generation and of substrate depletion independently.
Scheme 1. Selectivity of electron (upper) and substrate (lower) utilization in the
oxidation of EOMCC by CYP1A2 in pH 6.7 phosphate buffer.
Since all three measured rates increased similarly with increasing
ionic strength, as varied by the phosphate buffer concentration,
the inherently larger ionic strength of phosphate compared to Tris
is not an adequate argument to explain the different behavior of
By varying the buffer type, buffer concentration, pH, and tem-
perature in the CYP1A2 oxidation of EOMCC, a 10-fold range of
rates was accessed with the highest activity found at pH 6.7 in
r7HCC compared to -rEOMCC and -rNADPH in these two buffers. P450
reactions are also known to be differentially affected by pH with
different product generation optima for different isoform/substrate
combinations [14,32,33]. The current study extends this result to
show that substrate depletion and product generation can exhibit
disparate pH optima.
A strong correlation was observed between -rEOMCC and -rNADPH
as a function of temperature and phosphate buffer concentration
(Fig. 5a) indicating that coupling is not significantly affected by
these variables. CYP1A2 oxidation of EOMCC was 92.1% uncoupled
for all buffer concentrations and temperatures at pH 6.7. Similarly
high uncoupling percentages have been found for the CYP1A2 oxi-
dation of phenacetin [29,34] and of 7-ethoxycoumarin [35] and for
other human P450 enzymes [36–38]. A much weaker correlation
between -rEOMCC and -rNADPH was found as a function of buffer type
2
50 mM phosphate buffer at 45 °C. Ionic strength is typically var-
ied by varying the buffer concentration, with maximum product-
generation rates occurring between 50 and 300 mM [5,6].
Although exceptions have been found for this trend [5], all mea-
sured rates in the present study increased with increasing phos-
phate buffer concentration to a maximum between 250 and
3
75 mM potassium phosphate. An increase in the measured rates
was also observed with temperature between 25 °C and 45 °C,
the highest temperature setting available on the instrument. The
denaturation temperature of human CYP1A2 in reconstituted
membranes reportedly varied with membrane composition be-
tween 40 °C and 50 °C [30]. This matched the behavior observed

1
92
M.J. Traylor et al. / Archives of Biochemistry and Biophysics 505 (2011) 186–193
and pH (Fig. 5b) indicating that coupling is altered by these vari-
ables. To the best of our knowledge this is the first investigation
of P450 coupling as a function of buffer type, pH, or temperature.
However Schenkman et al. observed an increase in the coupling
of the benzphetamine N-demethylation activity of CYP2B4 en-
riched rabbit liver microsomes with increasing phosphate buffer
concentration [22]. This observation contrasts with our own data
where buffer concentration did not alter coupling, suggesting that
P450 coupling may be differentially regulated by reaction condi-
tions for different P450 systems.
circular dichroism [39]. It is interesting that this change in struc-
ture could accompany a large change in activity without also per-
turbing the active site and altering the substrate binding
interactions relevant to selectivity. Schrag et al., however, observed
constant selectivity in the presence of a structural perturbation of
the active site in the CYP3A4 metabolism of triazolam. Regioselec-
tivity was unaffected by a change in active site topography induced
by magnesium ions [17]. Overall, it is clear that more work is
needed to understand the complex interactions between environ-
mental parameters and P450 function.
Similarly, regioselectivity was also not significantly altered by
temperature or phosphate buffer concentration (Fig. 5c). It was
found that 21.7% of the oxidized EOMCC was used to generate
the O-dealkylation product 7HCC in pH 6.7 phosphate solution.
While no previous study has investigated the effect of temperature
on regioselectivity, similar trends with buffer concentration have
been observed [14,16]. Buffer concentration did not alter the ratio
of several testosterone metabolites from purified rat CYP2A1,
CYP2B1, and CYP2C11 [14] or of midazolam metabolites from hu-
man liver microsomes [16], consistent with our findings that regi-
oselectivity was insensitive to buffer concentration. However, a
change in regioselectivity with pH was reported for the oxidation
of testosterone by rat CYP2A1, CYP2B1, and CYP2C11 [14] and for
cyclosporine metabolism of human CYP3A4 [15]. Likewise, our re-
sults revealed a change in regioselectivity with pH and additionally
with buffer type. In light of previous literature on several P450 iso-
forms and substrates, these data suggest that ionic strength, as var-
ied by buffer concentration, may generally affect the activity of a
P450 reaction but not the regioselectivity, while pH may affect
both. However, further work with additional P450 isoforms and
substrates is needed to confirm this conclusion.
In conclusion, mass spectrometry was used to confirm the exis-
tence of multiple coupled catalytic modes for the CYP1A2 oxida-
tion of EOMCC in addition to the standard uncoupled modes. An
in situ optical method was developed to monitor the activity and
selectivity of human drug metabolizing enzymes and demon-
strated on a cyanocoumarin substrate. This method is anticipated
to be a useful tool for protein engineering studies seeking to under-
stand and control the influence of reaction conditions on P450
function. While the activity of CYP1A2 in this work was found to
vary over 10-fold as a function of buffer type, buffer concentration,
pH, and temperature, the regioselectivity and coupling efficiency
were effectively constant with buffer concentration and tempera-
ture but varied with buffer type and pH. Future work will investi-
gate the generality of these results with other P450 isoforms with
the aim of developing general heuristics relating P450 selectivity
and environmental parameters.
Acknowledgments
We thank Rita Nichiporuk and Ulla Anderson for assistance
with the mass spectrometry experiments and data interpretation,
and Jessica Ryan and Elizabeth Schneider for helpful discussions.
This work was supported by the National Science Foundation and
the National Institutes of Health.
It is relevant to emphasize that 7HCC is a minor product in the
overall oxidation of EOMCC by CYP1A2. This low regioselectivity
toward O-dealkylation suggests that studies employing 7-alkoxy-
3
-cyanocoumarin substrates as convenient fluorescent probes for
CYP1A2 activity are measuring only a fraction of the reaction prod-
ucts. A further elaboration of the product profile of commonly used
coumarin substrates would be a valuable contribution to the liter-
ature, especially in light of our current finding that 7HCC accounts
for only 21.7% of the substrate depletion.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.abb.2010.10.002.
Because regioselectivity and coupling are governed by enzyme–
substrate binding interactions and neither temperature nor buffer
concentration altered regioselectivity or coupling, it follows that
these parameters do not alter the enzyme–substrate binding
interactions. Variations in P450 function with buffer concentration
have been previously attributed to the resulting variation in ionic
strength and not to specific effects of phosphate ions
References
[1] P.R. Ortiz de Montellano, Cytochrome P450: Structure, Mechanism, and
Biochemistry, Plenum Press, New York, 2005.
[
2] M.K. Julsing, S. Cornelissen, B. Buhler, A. Schmid, Curr. Opin. Chem. Biol. 12
2008) 177–186.
[3] L.C. Wienkers, T.G. Heath, Nat. Rev. Drug Discov. 4 (2005) 825–833.
(
[
4] E.M.J. Gillam, T. Baba, B.R. Kim, S. Ohmori, F.P. Guengerich, Arch. Biochem.
Biophys. 305 (1993) 123–131.
[
5,14,16,22,39]. Although ion-specific effects are known in P450
systems [7,40,41], the correlation of phosphate buffer effects with
ionic strength effects is supported by the similar activity trends ob-
served in the CYP1A2 de-ethylation of 7-ethoxyresorufin and 7-
ethoxycoumarin as a function of phosphate buffer and sodium
chloride concentration [39]. Variations in P450 function with ionic
strength have been attributed to an alteration of the P450 – reduc-
tase interaction [5,22] or the heteromeric oligomerization state
[
5] H. Yamazaki, E.M.J. Gillam, M.S. Dong, W.W. Johnson, F.P. Guengerich, T.
Shimada, Arch. Biochem. Biophys. 342 (1997) 329–337.
6] H. Yamazaki, T. Shimada, Methods Mol. Biol. 320 (2006) 61–71.
7] H. Yamazaki, Y.F. Ueng, T. Shimada, F.P. Guengerich, Biochemistry 34 (1995)
[
[
8380–8389.
[8] A. Sigel, H. Sigel, R.K.O. Sigel, The Ubiquitous Roles of Cytochrome P450
Proteins, John Wiley & Sons, Ltd., Chichester, U.K., 2007.
[
10] R.E. White, M.B. Mccarthy, K.D. Egeberg, S.G. Sligar, Arch. Biochem. Biophys.
228 (1984) 493–502.
11] A. Seifert, S. Tatzel, R.D. Schmid, J. Pleiss, Proteins: Struct., Funct., Bioinf. 64
9] E.M.J. Gillam, Arch. Biochem. Biophys. 464 (2007) 176–186.
[
[
[
[
31,42], both of which have been shown to be mediated by ionic
interactions [31,42–44]. In addition, the reduction of cytochrome
c by P450 reductase, studied as a function of pH and ionic strength,
yielded activity maxima at pH 7.5 [45] and an ionic strength of
(
2006) 147–155.
12] M.R. Wester, E.F. Johnson, C. Marques-Soares, P.M. Dansette, D. Mansuy, C.D.
Stout, Biochemistry 42 (2003) 6370–6379.
5
00 mM [46], similar to the behavior observed in the current work.
[13] G.T. Miwa, A.Y.H. Lu, Bioessays 7 (1987) 215–219.
[
[
[
[
14] B. Gemzik, M.R. Halvorson, A. Parkinson, J. Steroid, Biochem. Mol. Biol. 35
1990) 429–440.
15] M. Hermann, E.T. Kase, E. Molden, H. Christensen, Curr. Drug Metab. 7 (2006)
265–271.
It is conceivable that selectivity is regulated by the enzyme–sub-
strate binding interaction, which controls substrate positioning,
while activity is regulated by the P450–reductase interaction,
which controls the flow of electrons to the P450.
Ionic strength, as varied by sodium chloride concentration, has
been shown to alter the structure of rabbit CYP1A2 as measured by
(
16] J. Maenpaa, S.D. Hall, B.J. Ring, S.C. Strom, S.A. Wrighton, Pharmacogenetics 8
(
1998) 137–155.
17] M.L. Schrag, L.C. Wienkers, Drug Metab. Dispos. 28 (2000) 1198–1201.
[18] P.J. Loida, S.G. Sligar, Biochemistry 32 (1993) 11530–11538.

M.J. Traylor et al. / Archives of Biochemistry and Biophysics 505 (2011) 186–193
193
[
[
19] T.M. Makris, K. von Koenig, I. Schlichting, S.G. Sligar, Biochemistry 46 (2007)
4129–14140.
20] I.G. Denisov, T.M. Makris, S.G. Sligar, I. Schlichting, Chem. Rev. 105 (2005)
253–2277.
[32] M. Delaforge, P. Ladam, G. Bouille, J.G. Benarous, M. Jaouen, J.P. Girault, Chem.-
Biol. Interact. 85 (1992) 215–227.
[33] J.M. Hutzler, F.J. Powers, M.A. Wynalda, L.C. Wienkers, Arch. Biochem. Biophys.
417 (2003) 165–175.
1
2
[
[
21] F.P. Guengerich, Drug Metab. Rev. 36 (2004) 159–197.
22] J.B. Schenkman, A.I. Voznesensky, I. Jansson, Arch. Biochem. Biophys. 314
[34] Q.B. Huang, G.D. Szklarz, Drug. Metab. Dispos. 38 (2010) 1039–1045.
[35] H. Mayuzumi, C. Sambongi, K. Hiroya, T. Shimizu, T. Tateishi, M. Hatano,
Biochemistry 32 (1993) 5622–5628.
(
1994) 234–241.
[
23] D.M. Stresser, S.D. Turner, A.P. Blanchard, V.P. Miller, C.L. Crespi, Drug Metab.
Dispos. 30 (2002) 845–852.
24] M. Murray, Curr. Drug Metab. 1 (2000) 67–84.
25] F.P. Guengerich, Chem. Res. Toxicol. 14 (2001) 611–650.
26] S.L. Born, D. Caudill, K.L. Fliter, M.P. Purdon, Drug Metab. Dispos. 30 (2002)
[36] X.J. Fang, Y. Kobayashi, J.R. Halpert, FEBS Lett. 416 (1997) 77–80.
[37] C.W. Locuson, P.M. Gannett, T.S. Tracy, Arch. Biochem. Biophys. 449 (2006)
115–129.
[38] A. Perret, D. Pompon, Biochemistry 37 (1998) 11412–11424.
[39] C.H. Yun, M. Song, T. Ahn, H. Kim, J. Biol. Chem. 271 (1996) 31312–31316.
[40] J.S. Kim, T. Ahn, S.K. Yim, C.H. Yun, Biochemistry 41 (2002) 9438–9447.
[41] J.S. Kim, C.H. Yun, Toxicol. Lett. 156 (2005) 341–350.
[42] R.W. Kelley, D.M. Cheng, W.L. Backes, Biochemistry 45 (2006) 15807–15816.
[43] D.R. Davydov, A.A. Kariakin, N.A. Petushkova, J.A. Peterson, Biochemistry 39
(2000) 6489–6497.
[
[
[
4
83–487.
[
[
27] S.L. Born, P.A. Rodriguez, C.L. Eddy, L.D. LehmanMcKeeman, Drug Metab.
Dispos. 25 (1997) 1318–1323.
28] K.A. Regal, K.L. Kunze, R.M. Peter, S.D. Nelson, Drug. Metab. Dispos. 33 (2005)
1
837–1844.
[
[
[
29] C.H. Yun, G.P. Miller, F.P. Guengerich, Biochemistry 39 (2000) 11319–11329.
30] T. Ahn, C.H. Yun, D.B. Oh, Biochemistry 44 (2005) 9188–9196.
31] R.W. Kelley, J.R. Reed, W.L. Backes, Biochemistry 44 (2005) 2632–2641.
[44] W.L. Backes, R.W. Kelley, Pharmacol. Ther. 98 (2003) 221–233.
[45] D.S. Sem, C.B. Kasper, Biochemistry 32 (1993) 11539–11547.
[46] D.S. Sem, C.B. Kasper, Biochemistry 34 (1995) 12768–12774.