Inhibition of human cytochrome P450 2E1 and 2A6 by aldehydes: Structure and activity relationships

Article abstract of DOI:10.1016/j.cbi.2014.05.014

The purpose of this study was to probe active site structure and dynamics of human cytochrome P450_2E1 and P450_2A6 using a series of related short chain fatty aldehydes. Binding efficiency of the aldehydes was monitored via their ability to inhibit the binding and activation of the probe substrates p-nitrophenol (2E1) and coumarin (2A6). Oxidation of the aldehydes was observed in reactions with individually expressed 2E1, but not 2A6, suggesting alternate binding modes. For saturated aldehydes the optimum chain length for inhibition of 2E1 was 9 carbons (K_I = 7.8 ± 0.3 μM), whereas for 2A6 heptanal was most potent (K_I = 15.8 ± 1.1 μM). A double bond in the 2-position of the aldehyde significantly decreased the observed K_I relative to the corresponding saturated compound in most cases. A clear difference in the effect of the double bond was observed between the two isoforms. With 2E1, the double bond appeared to remove steric constraints on aldehyde binding with K_I values for the 5-12 carbon compounds ranging between 2.6 ± 0.1 μM and 12.8 ± 0.5 μM, whereas steric effects remained the dominant factor in the binding of the unsaturated aldehydes to 2A6 (observed K_I values between 7.0 ± 0.5 μM and >1000 μM). The aldehyde function was essential for effective inhibition, as the corresponding carboxylic acids had very little effect on enzyme activity over the same range of concentrations, and branching at the 3-position of the aldehydes increased the corresponding K_I value in all cases examined. The results suggest that a conjugated π-system may be a key structural determinant in the binding of these compounds to both enzymes, and may also be an important feature for the expansion of the active site volume in 2E1.

Full text of DOI:10.1016/j.cbi.2014.05.014

Accepted Manuscript
Inhibition of human Cytochrome P450 2E1 and 2A6 by aldehydes: Structure
and activity relationships
Suneel K. Kandagatla, Todd Mack, Sean Simpson, Jill Sollenberger, Eric
Helton, Gregory M. Raner
PII:
S0009-2797(14)00172-0
http://dx.doi.org/10.1016/j.cbi.2014.05.014
CBI 7055
DOI:
Reference:
To appear in:
Chemico-Biological Interactions
Received Date:
Revised Date:
Accepted Date:
11 March 2014
5 May 2014
21 May 2014
Please cite this article as: S.K. Kandagatla, T. Mack, S. Simpson, J. Sollenberger, E. Helton, G.M. Raner, Inhibition
of human Cytochrome P450 2E1 and 2A6 by aldehydes: Structure and activity relationships, Chemico-Biological
Interactions (2014), doi: http://dx.doi.org/10.1016/j.cbi.2014.05.014
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Inhibition of human Cytochrome P450 2E1 and 2A6 by aldehydes: Structure
and activity relationships
a
a
a
a
a
Suneel K. Kandagatla , Todd Mack , Sean Simpson , Jill Sollenberger , Eric Helton and
a*
Gregory M. Raner ,
a
The University of North Carolina at Greensboro, Department of Chemistry and Biochemistry,
Greensboro, NC
Address for corresponding author:
412 Sullivan Science Building
Department of Chemistry and Biochemistry
The University of North Carolina at Greensboro
Greensboro, NC 27402-6170
Email:
Phone:
Fax:
gmraner@uncg.edu
(336) 334-4519
(336) 334-5402
*
To whom correspondence should be addressed
1

Abstract:
The purpose of this study was to probe active site structure and dynamics of human
cytochrome P4502E1 and P4502A6 using a series of related short chain fatty aldehydes. Binding
efficiency of the aldehydes was monitored via their ability to inhibit the binding and activation of
the probe substrates p-nitrophenol (2E1) and coumarin (2A6). Oxidation of the aldehydes was
observed in reactions with individually expressed 2E1, but not 2A6, suggesting alternate binding
modes. For saturated aldehydes the optimum chain length for inhibition of 2E1 was 9 carbons
(
K
I
=7.8 ±0.3 μM), whereas for 2A6 heptanal was most potent (K
I
=15.8 ±1.1 μM). A double
relative to the
bond in the 2-position of the aldehyde significantly decreased the observed K
I
corresponding saturated compound in most cases. A clear difference in the effect of the double
bond was observed between the two isoforms. With 2E1, the double bond appeared to remove
steric constraints on aldehyde binding with K
between 2.6 ± 0.1 μM and 12.8± 0.5 μM, whereas steric effects remained the dominant factor in
the binding of the unsaturated aldehydes to 2A6 (observed K
values between 7.0± 0.5 μM and
I
values for the 5-12 carbon compounds ranging
I
>1000 μM). The aldehyde function was essential for effective inhibition, as the corresponding
carboxylic acids had very little effect on enzyme activity over the same range of concentrations,
and branching at the 3-position of the aldehydes increased the corresponding K value in all cases
I
examined. The results suggest that a conjugated π-system may be a key structural determinant in
the binding of these compounds to both enzymes, and may also be an important feature for the
expansion of the active site volume in 2E1.
Abbreviations: NADPH, nicotinamide adenine dinucleotide; HLM, human liver S9 fractions;
2A6, cytochrome P4502A6; 2E1, cytochrome P4502E1;
2

Key Words: Cytochrome P450, Aldehydes, Inhibition, 2E1, 2A6
3

1
. Introduction:
The function of microsomal cytochrome P450 enzymes is generally a protective one, where
foreign chemicals entering the body can be oxidized in a way that makes them more readily excreted.
However, the human microsomal cytochrome P450 isoforms 2E1 and 2A6 have been implicated
in the deleterious effects of a variety of drugs and environmental agents, including
acetaminophen [1], nitrosamines [2,3], and carbon tetrachloride [4], to name a few. There is also
evidence to suggest that the inhibition of these isoforms during times of exposure to these agents
can provide a level of protection against the harmful effects [5-7]. Consequently, identification
of selective inhibitors of these isoforms and understanding the basis for selectivity has been the
focus of a number of recent studies. It has been established through crystallographic analysis
and binding studies that the active sites of these two isoforms are relatively small compared with
other microsomal P450s, with the 2A6 isoform being the more sterically restricted [8-10].
Comparisons between the two enzymes suggest that 2A6 is more structurally rigid, whereas the
2E1 active site may have more flexibility, and can change in response to different substrates
3
3
[
11,12]. Indeed, active site volumes ranging from 190 Å to 470 Å have been measured for 2E1
with different substrates bound [13]. This is consistent with spectroscopic studies indicating
that the active site of this isoform displays a significant level of compressibility under high
pressure, whereas the 2A6 active site is considerably more unyielding [14,15]. Crystallographic
and computational studies suggest that π-stacking interactions in the active site may facilitate the
binding of planar low molecular weight molecules to both isoforms, and in 2E1, these
interactions may be the basis for active site reorganization associated with observed substrate
inhibition [16,17].
The goal of the current study was to explore structure-activity relationships mediating the
binding of a series of related aldehydes to human Cytochrome P4502E1 and P4502A6 in order to
gain additional insight into the structural features contributing to selectivity of substrates and
4

inhibitors toward these two isoforms, along with features that facilitate active site reorganization
and expansion. In particular, factors such as chain length, branching in the 3-position and
unsaturation in the 2-position were evaluated for their influence on K . Aldehydes were selected
I
on the basis of prior studies involving rabbit liver cytochrome P450s identifying aldehydes as a
class of P450 inhibitors, and the availability of a wide range of structurally related compounds
belonging to this class [18]. In addition, aldehydes of the type described in this study can be
generated physiologically via lipid peroxidation, thus insight gained through this study may
guide future studies related to toxicology and cellular oxidative stress.
5

2
. MATERIALS AND METHODS
.1 Chemicals: The aldehyde compounds used in this study were purchased from Acros
2
Organics. Coumarin, 7-hydroxycoumarin (umbelliferone), NADPH, p-nitrophenol and 4-
nitrocatechol were from Sigma Chemical Co. Human liver S9 fractions were purchased from
Moltox Inc, Ashville, NC.
2
.2 Enzymatic assays: P4502E1: All assays for 2E1 (p-nitrophenol) were carried out
according to published protocols (19) with slight modifications. S9 fractions (0.5 mg) were
incubated in 100 mM phosphate buffer (pH 7.4) in the presence of 20-100 μM p-nitrophenol and
o
1
.0 mM NADPH for a total of 30 min at 37 C in a 0.5 mL total reaction volume. Reactions
were terminated by addition of ice-cold 6% perchloric acid and incubation on ice for an
additional 10 min, followed by centrifugation to remove precipitated protein. Cleared
supernatant (50 μL) was then analyzed by RP-HPLC on a 150 x 4.6 C18 column with a mobile
phase consisting of 35% acetonitrile, 64.5% H O, 0.5% acetic acid at a flow rate of 1.0 mL/min.
2
Absorbance detection at 340 nm was used to quantify the 4-nitrocatechol product. A Shimadzu
LC 20A Series HPLC system consisting of an SPD-20A UV/Vis detector, LC 20AT solvent
delivery, and a Sil 20A autosampler was used for analysis of samples.
2
.3 Inhibition of P4502E1: All inhibitors, including the aldehydes and acids, were
prepared as aqueous stocks by diluting 4.0 μg inhibitor into 100 mL DI H O immediately prior
2
to their use. Initially, a dose-response curve was generated for each of the aldehydes with regard
to their relative inhibitory effects on P4502E1. Reactions contained 0.5 mg S9 fractions, 50 μM
p-nitrophenol, 1.0 mM NADPH and aldehyde concentrations ranging from 5- 370 μM in a 0.50
mL volume of 0.1 M phosphate buffer (pH 7.4). Samples were quenched and processed for
HPLC analysis as described earlier. To obtain more detailed kinetic parameters, Michaelis-
Menten kinetic experiments were then performed to probe the mode and potency of inhibition
6

using inhibitor concentrations of either 30 μM or 60 μM. These concentrations were chosen
based on the results of the screening experiments in which it was determined that this was the
lowest concentration to produce measurable inhibition for all of the compounds being examined
except for dodecyl aldehyde. The substrate concentrations were varied from 20.0-100 μM in a
total reaction volume of 0.50 mL for 30 min, and reactions were processed as described above.
A minimum of 4 independent trials were carried out for each experiment and the standard
deviations are reported.
2.4 Enzymatic assays: P4502A6: The assay that was used to measure inhibition of 2A6 is
one that utilizes the conversion of coumarin into 7-hydroxycoumarin. The procedure is modified
from that of Waxman and Chang [20]. Liver S9 fractions (0.50 mg) were incubated in a reaction
mixture containing 3.0 μM coumarin, 100 mM potassium phosphate buffer (pH 7.4) and 1.0 mM
o
NADPH in a total volume of 0.50 mL. Following a 30 min incubation at 37 C the protein was
precipitated with 100 μL of 6% perchloric acid and placed on ice for 10 minutes. Samples were
centrifuged at 13500 rpm for 10 minutes and 40 μL of the cleared supernatant was injected onto
a RP-C18 HPLC column with a mobile phase consisting of 59% DI water, 40% methanol, and
1
% acetic acid, at a flow rate of 1.0 mL/min.
.5 Inhibition of P4502A6: Inhibitors were prepared as aqueous stock solutions by
diluting 4.0 μg inhibitor into 100 mL DI H O immediately prior to their use. Dose-response
2
2
curve was generated for each of the aldehydes with regard to their relative inhibitory effects on
P4502A6. Reactions contained 0.5 mg S9 protein, 3.0 μM coumarin, 1.0 mM NADPH and
aldehyde concentrations ranging from 5- 370 μM in a 0.50 mL volume of 0.1 M phosphate
buffer (pH 7.4). Samples were quenched and processed for HPLC analysis as described in the
previous sections. Michaelis-Menten kinetic experiments were then performed using inhibitor
concentrations of either 30 μM or 60 μM. Again, these inhibitor concentrations were selected
7

based on the results of the screening experiments, as with the 2E1 reactions. The substrate
concentrations were varied from 0.5-6.0 μM in a total reaction volume of 0.50 mL for 30 min,
and reactions were processed as described above. A minimum of 4 independent trials were
carried out for each experiment and the standard deviations are reported.
2
.6 Reversibility Studies: To determine whether the inhibition of each P450 isoform by
the individual aldehydes was reversible or irreversible, microsomal samples were pre-incubated
with (1) Buffer alone, (2) Buffer + Aldehyde and (3) Buffer + Aldehyde + NADPH, as described
in previously by Raner et al. [18].
2
.7 Aldehyde Oxidation by human P450 enzymes: Oxidation of trans-2-octenal was
carried out in a total reaction volume of 200 μL, and contained expressed human P450 (2E1 or
A6 supersomes, Gentest, 50 nM final), 100 mM potassium phosphate buffer (pH 7.4), 100 μM
2
trans-2-octenal, and 1 mM NADPH. The reaction was quenched with 6% perchloric acid as
described previously. A C18 RP HPLC column (4.6 x 250) from Phenominex was used with a
mobile phase consisting of 55% acetonitrile and 45% water, both containing 0.1% TFA at a flow
rate of 1.4 mL/min and detection was at 230 nm. A sample of trans-2-octenoic acid was
acquired from Sigma-Aldrich and used to generate a standard curve for quantification of product.
o
The reactions with human 2E1 supersomes were linear for 45 min at 37 C, so all additional
reactions were carried out under these conditions. The 2A6 supersomes failed to produce
NADPH-dependent acid product. The oxidation of trans-2-decenal, trans-2-nonenal, trans-2-
octenal and trans-2-heptenal by the expressed cytochrome P4502E1 in supersomes was also
carried out with each of the aldehydes present at the following concentrations: trans-2-decenal
(
110 μM), trans-2-nonenal (120 μM), trans-2-octenal (100 μM) and trans-2-heptenal (110 μM),
along with 1.0 mM NADPH in a 100 mM phosphate buffer at pH 7.4. Reactions were carried
8

0
out for 45 min at 37 C and quenched as described above. Control reactions without NADPH
were carried out in parallel, and the reaction mixtures were analyzed using HPLC under the same
conditions as described for trans-2-octenoic acid alone. The effects of ethanol (0.001%, 0.01%,
0.10% and 1.0%) on the activity in 2E1 supersomes was also monitored under identical
conditions.
2.8 Analysis of the kinetic data. Michaelis-Menton plots were prepared for each of the
various activity data sets in the presence and absence of inhibitor. All data was fit using
Slidewrite version 4.1 (Advanced Graphics Software Inc) and the non-linear regression analysis
function using the Michaelis-Menton equation. Based on Michaelis-Menton plots for each
aldehyde, the inhibition could be described using the competitive model, so values for K
I
were
app
m
calculated using this model, where Vmax remains constant, αis defined as K
/K
m
and K
I
=
[
I]/(α -1).
.9 Graphical images of P450 active sites. The active site images of the 2E1 and
2
P4502A6 enzymes were generated using crystal structure coordinates deposited in the protein
databank. The file PDB: 1Z10 was used for the image of P4502A6 with coumarin bound; PDB:
3T3Z was used for P4502E1 with pilocarpene bound; PDB: 3E6I was used to generate the
P4502E1 image with indazole bound; and the PDB: 3LC4 coordinates were used for P4502E1
image with imidazolyl-dodecanoic acid bound. RasWin 2.7.5.2 Molecular Graphics software
was used to generate each of these images based on the respective pdb files.
9

3
. RESULTS
3
.1 Inhibition of P4502E1 by alkanals and alkenals:
Dose-response inhibition studies using p-nitrophenol oxidation as a measure for 2E1
activity were initially carried out using saturated straight-chain aldehydes ranging from 8-12
carbon atoms and 2-unsaturated aldehydes with 8-, 9- and 12-carbon atoms (Figure 1) at fixed
inhibitor concentrations of 30 μM and 230 μM. Inhibition was observed for each aldehyde as a
dose-dependent reduction in activity, relative to a control in which no aldehyde was present. The
selection of these concentrations was based empirically on their ability to highlight the
differences in potency among this group of aldehydes. Maximum inhibition was observed for
the nonyl and decyl aldehydes, and unsaturation significantly increased the level of inhibition,
suggesting the extended π-system of the unsaturated aldehydes leads to more favorable binding
in the active site.
The reversibility of inhibition was examined for each of the aldehydes represented in
Figure 1 by pre-incubating the aldehydes with S9 fractions in the presence and absence of
o
NADPH for 15 min at 37 C. Dilution of the pre-incubation mixture resulted in complete
restoration of the p-nitrophenol oxidation activity of the samples , thus all of these saturated and
unsaturated aldehydes (data not shown) appeared to be reversible inhibitors of 2E1 under the
conditions used. As a positive control, the experiment was carried out using 11-undecylenic
o
aldehyde at a concentration of 25 μM. Following a 15 min pre-incubation at 37 C in the
presence of 1 mM NADPH, a loss of ~50% of the activity of 2E1 was observed (data not
shown). This compound possesses a terminal olefin group, which is known to cause irreversible
damage to the heme [21-23].
Detailed inhibition studies were subsequently carried out in which the inhibition
constants (K ) were calculated for the aldehydes, both saturated and unsaturated, with carbon
I
1
0

chain lengths, between 5-12, using the standard Michaelis-Menten approach (Table 1). For all
the aldehydes tested, Vmax was unaltered, once again indicating competitive-reversible type
inhibition, thus the competitive model of inhibition was applied for K
I
determination.
Consistent with the preliminary screening data, the K values indicate that the 9- and 10-carbon
I
aldehydes, possessing a 2-double bond (nonenal and decenal) were the most potent inhibitors of
the human 2E1 isoform. It is also worth noting that a strong dependence on chain length was
observed for the saturated compounds, whereas less dependence on chain length was observed
with unsaturated aldehydes. Furthermore, the unsaturated compounds were all much more
potent than the respective saturated compounds, with K
I
values as much as 30-fold lower in the
case of the 11- and 12-carbon aldehydes.
Table 1
P4502A6
P4502E1
Aldehyde K (μM) K (μM)
K (μM) K (μM)
I
I
I
I
Saturated Unsaturated
Saturated Unsaturated
1
26 ± 16
1270 ± 150
155 ± 13
13 ± 1
6.7 ± 0.2
5.6 ± 0.1
4.1 ± 0.1
2.9 ± 0.2
3.3 ± 0.1
3.4 ± 0.1
2.6 ± 0.1
Pentyl
Hexyl
Heptyl
Octyl
5
7 ± 7
33 ± 3
26 ± 2
1
6 ±1
5 ± 2
5 ± 6
9 ± 1
32 ± 1
2
5
7 ± 1
13 ± 1
8.0 ±1
34 ± 3
Nonyl
Decyl
1
06 ± 29
59 ± 6
17 ± 1
6
4
8 ± 7
8 ± 9
103 ± 22
128 ± 35
117 ± 15
85 ± 8
Undecy
Dodecyl
1
1

1
2

3
.2 Effects of branching on inhibition of 2E1:
Additional aldehydes (citral and citronellal) with branching at the 3-position were
evaluated for inhibitory potency against 2E1. Citral (3,7-dimethyl 2,6-octadienal) as shown in
Figure 2, is a naturally occurring mixture of 2,3 -unsaturated aldehyde isomers with branching
methyl groups at carbon 3 and 7. Citronellal (3,7-dimethyl 6-octenal ), shown in figure 3, also
has a 3-methyl branch but does not have a double bond in the 2-position. Both aldehydes
showed reversible competitive inhibition profiles when evaluated using the Michaelis-Menten
model. K
the linear 10-carbon saturated and unsaturated aldehydes. By comparing citral to citronellal it is
clear that the double bond at the 2-position dramatically impacts inhibition; K values of 18.8
I
values were measured for both of these compounds and were compared with those of
I
μM vs 57.8 μΜ, respectively. As for the effects of branching, both decanal and citronellal are
saturated at the 2-position and possess 10 carbon atoms, like decanal. The K for Decanal,
I
however, is about 8-fold lower than the branched citronellal (7.0 μM vs 57.8 μM). Likewise, a
comparison between decenal and citral showed a 5-fold increase in K
the inclusion of a branching methyl at the 3-position of the aldehyde.
I
(3.3 μM to 15.7 μM) with
3
.3 Effects of aldehyde function:
The importance of the aldehyde group for inhibition was probed by monitoring 2E1
inhibition using the corresponding 8- to 12-carbon saturated carboxylic acids. Over a range of
1
5
.0-100 μM, only lauric acid inhibited 2E1 significantly (data not shown). At a concentration of
0 μM, lauric acid reduced the activity of 2E1 by 50%, which is still modest inhibition relative
to the aldehydes. Due to the lack of observed inhibition at high concentrations of carboxylic
acid, K values were not determined.
I
1
3

3
.4 Inhibition of P4502A6
by alkanals and alkenals:
As with the 2E1
isoform, all of the aldehydes
tested, both saturated and
unsaturated,
coumarin hydroxylation via
reversible competitive
mode. Values for K were
inhibited
a
I
determined for each of the
saturated and unsaturated
aldehydes used with 2E1,
and these values are
Figure 2
presented in Table 1. A trend similar to that observed with 2E1 was also observed for 2A6.
However, the optimal chain length appeared to be slightly shorter for 2A6, with a carbon chain
length of 7-8 being most potent. Comparison of the saturated vs unsaturated aldehydes again
suggested that the double bond in the 2-position improved the interaction with the active site, as
indicated by a 2- to 3-fold decrease in K for the unsaturated aldehydes over nearly the entire
I
range. It is worth pointing out that with 2A6 inhibition, dodecanal and undecanal appeared to be
outliers in the general trends observed, as both of these saturated aldehydes inhibited more
potently than the smaller decanal, and more potently than the corresponding unsaturated
counterparts. Another clear distinction between the 2E1 and 2A6 data can be seen in the
behavior of the unsaturated aldehydes. Where the 2E1 isoform appeared to be almost insensitive
to chain length, the inhibitors of the 2A6 isoform had an optimal chain length of 7 or 8, and
1
4

above or below this value the potency of inhibition decreased in a continuous manner, with the 2
notable exceptions previously pointed out.
3
.5 other structural determinants for P4502A6 inhibition by aldehydes:
As with P4502E1, saturated carboxylic acids were ineffective at inhibiting the 2A6
isoform at concentrations as high as 100 μM, thus K
I
values were not determined. Interestingly,
branching in the aldehyde also resulted in a dramatic loss in potency as demonstrated by the use
of citral and citronellal as inhibitors of 2A6. Only citral inhibited 2A6, and here the K
I
was still
>100 μM (data not shown).
3
.6 Oxidation of aldehydes by human S9 fractions and expressed cytochrome P4502E1 and
P4502A6: To determine that the aldehyde function of each inhibitor was inside the active site,
reactions involving 4 of the unsaturated aldehydes, trans-2-heptenal, trans-2-octenal, trans-2-
nonenal and trans-2-decenal were evaluated for the formation of the corresponding carboxylic
acid. The substrate trans-2-octenal was examined initially to establish linearity over time, and
the effects of ascorbate and glutathione on the reaction. This aldehyde was selected on the basis
of its ability to inhibit both 2A6 and 2E1 isoforms effectively. The reaction with 2E1
o
supersomes was linear for 45 min at 37 C, and beyond 45 min a marked decrease in activity was
observed. In addition, inclusion of either 1.0 mM ascorbate or 1.0 mM glutathione had no effect
on the rate of acid formation. Furthermore, for the 2E1-dependent reaction, increasing
concentrations of ethanol (a P4502E1 inhibitor), between 0.001 and 1.0% caused a dose-
dependent decrease in acid formation, with >80% inhibition at 1% and 40% inhibition at 0.1%
ethanol. In contrast, 2A6 supersomes did not catalyze oxidation of trans-2-octenal. An
additional experiment was carried out in which the four aldehydes; trans-2-heptenal, trans-2-
octenal, trans-2-nonenal and trans-2-decenal were incubated with either expressed 2E1 or
1
5

expressed 2A6, and the NADPH-dependent oxidation of each aldehyde was monitored.
Supersomes containing expressed 2E1 were effective in the oxidation of all 4 aldehydes, while
2A6 supersomes again yielded no acid products.
4. DISCUSSION
Prior studies involving rabbit cytochrome P4502B4 indicated that mammalian cytochrome
P450s were susceptible to aldehyde-mediated irreversible mechanism-based inhibition [18]
Mechanistically, the inactivation was the result of heme destruction and was correlated to the
aldehyde deformylation reaction mechanism [24,25]. No prior studies have been reported
involving interactions between aldehydes and the human isoforms, including 2E1 and 2A6, both
of which are of considerable toxicological importance. In the current study, no evidence for
irreversible inhibition was observed with any of the aldehydes or either cytochrome P450, rather,
reversible competitive inhibition was observed. However, oxidation of the aldehyde function by
the expressed 2E1 enzyme was observed indicating an “aldehyde in” binding orientation in the
active site for these compounds. With 2A6, a different binding orientation must be envisioned
based on the inability to oxidize the aldehyde function. It should be noted that in the rabbit liver
study, aldehyde concentrations used to calculate inactivation rates were in the mM range,
whereas in the current study, the highest aldehyde concentrations used were mostly below 100
μM. It is conceivable that inactivation of the human P450s could occur at higher aldehyde
concentrations.
All of the saturated aldehydes inhibited both P450 isoforms, with varying potency, as
determined by measured K
I
values. Several interesting trends were observed in this data. For
both 2E1 and 2A6, there appeared to be an optimal carbon chain length as it related to inhibition
1
6

potency. For 2E1, the nonyl and decyl aldehyde had the lowest measured K , whereas heptyl and
I
octyl aldehydes were most effective against 2A6, and neither isoform was inhibited by
corresponding carboxylic acids. Several recent comparative studies have been carried out
concerning structural and activity relationships between the 2E1 and the 2A6 human P450
isoforms that support these experimental observations [11,12,14]. Both enzymes have relatively
small active sites, and therefore binding of larger molecules may be disfavored via steric
interactions, consistent with the observed pattern of inhibition. The 2A6 active site has a
slightly smaller volume than 2E1 based on crystal structures with the small molecule pilocarpine
3
3
bound (280 Å vs 337 Å ) [13], which is consistent with the smaller optimum chain length for
inhibition of 2A6 (7-8 carbons for 2A6 vs 9-10 for 2E1). In addition, DeVore et al. [12], using
both experimental and computational approaches showed that the 2E1 active site was slightly
less polar than 2A6, however both contained a cluster of phenylalanine groups that appeared to
define the active site boundaries. The reduced interaction of both isoforms with carboxylic acids
relative to aldehydes is most easily justified by the less polar character of the aldehyde group.
The crystal structure of 2A6 reveals an Asn side chain (Asn 297) that may help to stabilize the
carbonyl group of the aldehydes via hydrogen bonding. Indeed, the substrates pilocarpene and
coumarin both have carbonyl oxygens that form H-bond interactions with Asn 297 (Figure 3
shows the structure of the 2A6 active site with coumarin bound).
1
7

Figure 3
The 2-unsaturated aldehydes were generally more potent inhibitors of both 2E1 and 2A6,
with some K values nearly 10-fold lower than the corresponding saturated compounds.
I
Computational studies by Xu et al. [26] highlight the importance of π−π interactions in the
binding of 4-(methylnitrosamino)- 1-(3-pyridyl)-1-butanone (NNK) in the active site of 2A6,
thus similar interactions with the planar π-system of the 2-unsaturated aldehydes could be
envisioned as a basis for more efficient binding. Furthermore, based on the crystal structure of
the human 2A6 isoform with coumarin bound, the nature of this π-π interaction involves the
perpendicular arrangement of the π-orbitals on the aromatic ring systems of coumarin with those
on Phe107 ring. [27],
It is reasonable to suggest that the π-system of the 2-unsaturated
aldehyde interacts favorably with Phe-107 of 2A6 in an analogous fashion, which serves to
1
8

orient the substrate and restrict its mobility. Thus, increasing size would create steric
interference in the rigid 2A6 active site, if these binding interactions were to be maintained [12].
This binding mode can also be used to rationalize the lack of aldehyde oxidation, as the carbonyl
group of the aldehyde would be hydrogen bonded to Asn-297, which is removed from the
activated oxygen in the catalytic site. The increase in potency toward 2A6 for the undecanal and
dodecanal, relative to the corresponding unsaturated compounds was noted, however, it is not
clear what the molecular basis for this observation is. It is tempting to rationalize this based on
the greater flexibility of the large saturated aldehydes in the rigid active site. This does not
account for the increased potency of these larger aldehydes relative to the smaller decanal, which
should have less steric restrictions on binding. A more likely explanation is that given the
hydrophobic nature of these compounds, disruption of the P4502A6/P450-reductase complex may
result in the more potent observed inhibitory action.
Interestingly, for the unsaturated aldehydes, the dependence on chain length was almost
completely abolished in 2E1, with the aldehydes containing 7-12 carbon atoms all inhibiting with
K
I
in the 3-5 μM range. This is in stark contrast to the behavior of the un-saturated aldehydes
with 2A6. Although the π-interactions appear to be strongly driving the binding of these
aldehydes to both enzymes, the 2E1 active site appears to have the ability to accommodate larger
molecules in response to the presence of the π-system of the unsaturated aldehydes. First, 2E1
lacks the H-bond donor to the carbonyl that is present in 2A6, as Asn-297 is replaced in 2E1 with
Asp-295, which is excluded from the active site. According to DeVore et al. [12] the substrate
pilocarpine appears to adapt to the 2A6 active site upon binding, whereas the 2E1 active site may
change topology in response to substrate binding. In support of this flexibility model, the 2E1
3
active site volume can expand to 420-470 Å in response to the binding of the large substrate
imidazolyl-dodecanoic acid [10].
1
9

Figure 4 shows a comparison of the active site structure of P4502E1 with (A) pilocarpene
(
B) imidazolyl-dodecanoic acid and (C) indazole bound (substrate molecules are omitted from
the structure). Although each of these substrates is structurally distinct from the aldehydes used
in the current study, what is apparent is that there are residues in the active site that can rearrange
in response to substrate and/or inhibitor binding. Four Phe side chains (Phe 116, Phe 207, Phe
298 and Phe 478) that form the upper surface of the active site are shown in this figure. The
most significant differences on substrate binding occur in the position of Phe 298 and Phe 478
with respect to the heme. For example, in panel A, Phe 298 and Phe 478 have moved up relative
to the heme, creating a larger volume directly above the heme when the bulky pilocarpene
molecule binds. With indazole bound (panel C) Phe 478 is angled down toward the heme,
reducing the overall volume of this cavity. Panel B shows the additional movement of Phe 298
in response to binding imidazolyl-decanoic acid. This motion allows the acyl chain of the
substrate to access a binding channel running parallel to the I-helix just above Phe 298, which
presumably accounts for the larger active site volume in the presence of this substrate. It is
conceivable that π-interactions between unsaturated aldehydes and the 2E1 active site may have
an analogous effect on Phe 298, essentially holding this door open so that longer chain aldehydes
can gain access to the more remote areas of the active site. Alternatively, computational studies
by Li et al. [17] suggest that the phenyl ring of Phe 478 can rotate to become almost
perpendicular to the heme as a substrate binds to an “effector” site, which lies along the substrate
access channel, directly above the γ-meso position on the heme. This motion repositions both the
active substrate and a key catalytic residue T303, resulting in negative cooperativity. Li et al.
[
17] indicate the orientation of Phe 478 is significantly altered via π-π stacking interactions
when an aromatic ligand occupies the effector site. As the saturated aldehydes would not be
expected to bind to this effector site, it is also reasonable to suggest that the π-system of the 2-
2
0

unsaturated aldehyde allows it to gain access to the effector site, thus increasing the available
active site volume for unsaturated aldehydes. The lack of a hydrogen bonding group analogous
to Asn-297 in 2A6 appears to allow the carbonyl group of the substrate more freedom to explore
the active site, resulting in its oxidation by 2E1, but not 2A6.
A
B
C
Figure 4
5
. Conclusions: In summary, the current work demonstrates that a series of saturated and
unsaturated aliphatic aldehydes are effective competitive inhibitors of both human P450 2E1 and
A6. The relatively small size of the 2E1 and 2A6 active sites result in a preference for binding
2
the 7-10 carbon saturated aldehydes. In addition, 2-unsaturated aldehydes display 5 to 30-fold
higher affinity for both human enzymes than their corresponding unsaturated counterparts, which
2
1

is likely due to the potential for π-stacking interactions in Phenylalanine-rich active site of these
enzymes. This study also provides direct experimental support for the flexibility model of
substrate binding to 2E1, and suggests that an extended π-system may be a key determinant in
active site expansion.
Acknowledgements: Funding for this research was provided by The National Science
Foundation (#0414301), Research Corporation (CC4924) and National Center for
Complementary and Alternative Medicine (R15 AT007860-01A1) to G.M.R.
2
2

References:
1
2
. Guengerich, F.P., Kim, D.H., Iwasaki, M. (1991) Chem. Res. Toxicol. 4: 168-179.
. Chowdhury, G., Calcutt, M.W., Guengerich, F.P. (2010) J. Biol. Chem. 285:8031-8044.
. Liu, Y., Glatt, H. (2008) Mutat. Res. 643:64-69.
3
4
. Castillo, T., Koop, D.R., Kamimura, S., Triadafilopoulos, G. and Tsukamoto, H. (1992)
Hepatology 16: 992-996.
5
6
. Tung YT, Wu JH, Huang CC, Peng HC, Chen YL, Yang SC, Chang ST. (2009) Food
Chem Toxicol. 47:1385-1392.
. Karamanakos PN, Trafalis DT, Geromichalos GD, Pappas P, Harkitis P, Konstandi M,
Marselos M. (2009) Arch Toxicol. 83:571-580.
7
8
. von Weymarn, L.B., Chun, J.A., Hollenberg, P.F. (2006) Carcinogenesis. 27:782-90.
. Yano, J.K., Denton, T.T., Cerny, M.A., Zhang, X., Johnson, E.F., Cashman, J.R. (2006)
J. Med. Chem. 49:6987-7001.
9
1
. Porubsky, P.R., Meneely, K.M., Scott, E.E. (2008) J. Biol. Chem. 283:33698-707.
0. Wang, M.H., Wade, D., Chen, L., White, S., Yang, C.S. (1995) Arch. Biochem. Biophys.
3
17:299-304.
1
1. Anzenbacherova, E., Hudecek, J., Murgida, D., Hildebrandt, P., Marchal, S., Lange, R.,
Anzenbacher, P. (2005) Biochem. Biophys. Res. Commun. 338:477-482.
1
1
2. Porubsky, P.R., Battaile, K.P., Scott, E.E. (2010) J. Biol. Chem. 285:22282-22290.
3. DeVore, N.M., Meneely, K.M., Bar,t A.G., Stephens, E.S., Battaile, K.P., Scott, E.E.
(
2012) FEBS J. 279:1621-1631.
1
1
4. Skopalík, J., Anzenbacher, P., Otyepka, M. (2008) J. Phys. Chem. B. 112:8165-8173.
5. Hendrychova, T., Berka, K., Navratilova, V., Anzenbacher, P., Otyepka, M. (2012) Curr.
Drug Metab. 13:177-189.
1
6. Ping, J., Wang, Y.J., Wang, J.F., Li, X., Li, Y.X., Hao, P. (2012) Curr. Drug Metab.
1
3:1024-1031.
1
1
1
7. Li, J., Wei, D.Q., Wang, J.F., Li, Y.X. (2011) J. Chem. Inf. Model. 51:3217-3225.
8. Raner, G.M., Chiang, E.W., Vaz, A.D., Coon, M.J. (1997) Biochemistry. 36:4895-4902.
9. Yang, S.-P., Medling, T., Raner, G.M. (2003) Comp. Biochem. Physiol. C. 136, 297-
3
08.
2
3

2
2
0. Waxman, D.J., Chang, T.K. (2006) Methods Mol. Biol. 320:91-96.
1. Ortiz de Montellano, P.R., Beilan, H.S., Kunze, K.L., Mico, B.A. (1981) J. Biol Chem.
2
56:4395-4399.
2
2
2. Ortiz de Montellano, P.R., Beilan, H.S., Kunze, K.L. (1981) J. Biol. Chem. 256:6708-
713.
6
3. Ortiz de Montellano, P.R, Beilan, H.S., Kunze, K.L. (1981) Proc. Natl. Acad. Sci. U S A.
8:1490-1494.
7
2
2
4. Kuo, C.L., Raner, G.M., Vaz, A.D., Coon, M.J. (1999) Biochemistry. 38:10511-10518.
5. Vaz, A.D., Pernecky, S.J., Raner, G.M., Coon, M.J. (1996) Proc. Natl. Acad. Sci. U S A.
9
3:4644-4648.
2
2
6. Xu, Y., Shen, Z., Shen, J., Liu, G., L,i W., Tang, Y. (2011) J. Mo.l Graph. Model. 30:1-
9
.
7. Yano, J.K., Hsu, M.H., Griffin, K.J., Stout, C.D., Johnson, E.F. (2005)
Nat.Struct.Mol.Biol.12:822-823.
2
4

Figure Captions:
Figure 1. Effects of saturated and 2-unsaturated aldehydes on P4502E1 activity. (A) Effect
of chain length on P450 inhibition by satuated aldehydes at 30 and 230 μM aldehyde
concentration. (B) Comparison of the effects of saturated vs 2- unsaturated aldehydes on
P4502E1 inhibition at 30 and 230 μM aldehyde concentration. (C) Structures of aldehydes
used for these comparative studies.
Figure 2. Structure and measured inhibition constants for branched saturated and
unsaturated aldehydes, citral, citronellal, and the corresponding linear, saturated and
unsaturated 10 carbon aldehydes, decanal and trans-2-decenal.
Table 1. Experimentally determined inhibition constants (K
cytochrome P4502A6 by straight chain saturated and unsaturated aldehydes ranging in size
from 5-12 carbons. K values were calculated using a competitive model for inhibition and a
I
) for the inhibition of Human
I
minimum of 3 trials were carried out for each aldehyde.
Figure 3. Key residues in the active site structure of human cytochrome P4502A6 with the
substrate coumarin bound. Hydrogen bond interactions between the carbonyl oxygen on the
substrate and the amide group from Asn-297 and π-π interactions between Phe-107 and the
aromatic ring of coumarin are shown ( generated using RasWin 2.7.5.2 and coordinates from
PDB: 1Z10)
2
5

Figure 4. Active site structure of P4502E1 showing the position of key Phe residues in the
presence of the substrates (A) pilocarpene (PDB: 3T3Z), (B) indazole (PDB: 3E6I) and (C)
imidazolyl-dodecanoic acid (PDB: 3LC4). Images were prepared using RasWin 2.7.5.2
Molecular Graphics software.
2
6

P4502A6
P4502E1
Aldehyde K (μM) K (μM)
K (μM) K (μM)
I
I
I
I
Saturated Unsaturated
Saturated Unsaturated
126 ± 16
1270 ± 150
155 ± 13
13 ± 1
6.7 ± 0.2
5.6 ± 0.1
4.1 ± 0.1
2.9 ± 0.2
3.3 ± 0.1
3.4 ± 0.1
2.6 ± 0.1
Pentyl
Hexyl
Heptyl
Octyl
57 ± 7
33 ± 3
26 ± 2
1
6 ±1
5 ± 2
5 ± 6
9 ± 1
32 ± 1
2
5
7 ± 1
13 ± 1
8.0 ±1
34 ± 3
Nonyl
Decyl
106 ± 29
59 ± 6
17 ± 1
6
8 ± 7
8 ± 9
103 ± 22
128 ± 35
117 ± 15
85 ± 8
Undecy
Dodecyl
4

Highlights:
•
•
•
•
•
Human P450 2E1 and 2A6 active sites were probed
using a series of small aldehydes
Both enzymes preferred α,β -unsaturated aldehydes to
corresponding saturated forms
Branching in the aldehyde reduced binding to both
isoforms
The 2E1 active site appeared to expand in response to
α,β -unsaturation
The aldehydes were oxidized by 2E1, but not 2A6
2
7