Effects of Alcohol-Induced Increase in CYP2E1 Content in Human Liver Microsomes on the Activity and Cooperativity of CYP3A4

Abstract

We investigate the effect of the alcohol-induced increase in the content of CYP2E1 in human liver microsomes (HLM) on the function of CYP3A4. Membrane incorporation of the purified CYP2E1 into HLM considerably increases the rate of metabolism of 7-benzyloxyquinoline (BQ) and attenuates the homotropic cooperativity observed with this CYP3A4-specific substrate. It also eliminates the activating effect of α-naphthoflavone (ANF) seen in some HLM samples. To probe the physiological relevance of these effects, we compared three pooled preparations of HLM from normal donors (HLM-N) with a pooled preparation from ten heavy alcohol consumers (HLM-A). The composition of the P450 pool in all samples was characterized by the mass-spectrometric determination of 11 cytochrome P450 species. The fractional content of CYP2E1 in HLM-A was from 2.0 to 3.4 times higher than in HLM-N. In contrast, the content of CYP3A4 in HLM-A was the lowest among all samples. Despite that, HLM-A exhibited a much higher metabolism rate and a lower homotropic cooperativity with BQ, similar to CYP2E1-enriched HLM-N. To substantiate the involvement of interactions between CYP2E1 and CYP3A4 in these effects, we probed hetero-association of these proteins in CYP3A4-containing Supersomes™ with a technique employing CYP2E1 labeled with BODIPY-618 maleimide. These experiments evinced the interactions between the two enzymes and revealed an inhibitory effect of ANF on their association. Our results demonstrate that the functional properties of CYP3A4 are fundamentally dependent on the composition of the cytochrome P450 ensemble and suggest a possible impact of chronic alcohol exposure on the pharmacokinetics of drugs metabolized by CYP3A4.

1. Introduction

The core of the drug-metabolizing ensemble in the human liver is the ensemble of multiple P450 species co-localized in the membrane of the endoplasmic reticulum (ER). There is an emerging recognition that multiple P450 enzymes interact with each other with the formation of heteromeric complexes where the functional properties of individual P450 enzymes are largely modified [1–4]. Recent findings point out the association of dissimilar P450s as the primary cause for a weak correlation between the composition of the P450 ensemble and its drug-metabolizing profile demonstrated with studies in a large set of normal human liver microsomes [5–6].

Of particular importance are the changes in P450-P450 crosstalk induced by alcohol consumption. Although the multi-fold increase in the content of CYP2E1 in the liver observed in alcohol consumers is well documented [7–10], the involvement of CYP2E1 in alcohol-drug interactions is commonly considered insignificant due to a minor role of this enzyme in drug metabolism [11]. However, the effects of induction of CYP2E1 may stretch beyond the changes in CYP2E1-dependent drug metabolism and involve the effects of CYP2E1 on the functional properties of other drug-metabolizing enzymes.

To probe the crosstalk between CYP2E1 and other P450 species and elucidate its role in alcohol-drug interactions, we established a model of alcohol-induced increase in CYP2E1 content that implements enrichment of HLM samples with CYP2E1 through membrane incorporation of the purified protein [12–14]. We demonstrated that the adopted CYP2E1 becomes a fully-functional member of the drug-metabolizing ensemble and interacts with other P450 enzymes to form heteromeric complexes [14]. Studying the effect of enrichment of HLM with CYP2E1 on the function of other P450 enzymes, we demonstrated a CYP2E1-induced activation of CYP1A2 and the associated re-routing of the metabolism of 7-ethoxy-4-cyanocoumarin (CEC), the substrate concurrently metabolized by CYP2C19 and CYP1A2, towards the latter enzyme [14].

The present study continues our efforts to reveal the impact of alcohol-induced increase in CYP2E1 content on drug metabolism. Here we explore the effects of CYP2E1 on the functional properties of CYP3A4, the enzyme that metabolizes about 50% of drugs on the market [15]. We used 7-benzyloxyquinoline (BQ) as a probe fluorogenic substrate highly selective for CYP3A enzymes [16]. In addition to studying the effect of CYP2E1 on the parameters of BQ metabolism in a series of microsomal preparations with a thoroughly characterized composition of the P450 pool, we also probed its impact on the effect of α-naphthoflavone (ANF), a prototypical allosteric effector of CYP3A4. To assess the relevance of the results obtained with CYP2E1-enriched microsomes to the changes caused by the effects of chronic alcohol exposure, we also compared the parameters of BQ metabolism in a pooled HLM preparation obtained from 10 alcoholic donors with those exhibited by three pooled HLM preparations obtained from donors with no history of alcohol dependence. Our results demonstrate that the catalytic activity and allosteric properties of CYP3A4 are fundamentally dependent on the composition of the cytochrome P450 ensemble in human liver and imply an impact of chronic alcohol exposure on the pharmacokinetics of drugs metabolized by CYP3A4.

Furthermore, probing possible mechanisms of CYP2E1-dependent activation of CYP3A4 revealed in this study, we investigated the interactions between the two proteins in the microsomal membrane using the technique based on homo-FRET in the oligomers of CYP2E1 labeled with BODYPY-618 maleimide (CYP2E1-BODIPY) [14]. High-affinity interactions between CYP2E1-BODIPY with CYP3A4 and their modulation by α-naphthoflavone revealed in these studies suggest a direct involvement of hetero-association between CYP2E1 and CYP3A4 in the observed effects.

2. Materials and methods

2.1. Chemicals

7-benzyloxyquinoline (BQ) was a product of CypEx (Dundee, UK). 7-hydroxycoumarin was obtained from Acros Organics, a part of Thermo Fisher Scientific (New Jersey, NJ). 7,8-benzoflavone (α-naphthoflavone, ANF) was manufactured by Indofine Chemical Company (Hillsborough, NJ). Amitriptyline and nortriptyline were the products of Cayman Chemicals (Ann Arbor, MI). Ivermectin was obtained from ApexBio (Boston, MA), and acetylacetone was the product of Tokyo Chemical Industry (Tokyo, Japan). Phenacetin was purchased from Sigma-Aldrich (St. Louis, MO). All other reagents were of ACS grade and used without additional purification.

2.2. Protein expression and purification

N-terminally truncated (Δ3-20), and C-terminally His-tagged CYP2E1 [17] was expressed in E. coli TOPP3 cells and purified as described earlier [14].

2.3. Pooled human liver microsomes and their characterization with mass spectrometry

The preparation of Human Liver Microsomes (HLM) obtained from 10 donors (mixed gender) with a history of chronic alcohol exposure (lot FVT) was purchased from BioIVT Corporation (Baltimore, MD). This preparation is referred to from now on as HLM-A. We also studied three different preparations of pooled human liver microsomes from 50 donors (mixed gender) without a reported alcohol exposure history. The preparation referred to as HLM-N1 was obtained by differential centrifugation of pooled human liver S9 fraction (lot number 3212595), the product of BD Gentest, now a part of Corning Life Sciences (Tewksbury, MA) is referred to as HLM-N1. The HLM samples referred to here as HLM-N2N2 and HLM-N3 are IriVitroCYP™ M-class 50-donor mixed gender pooled HLM preparations, lots LFJ and LBA respectively obtained from BioIVT Corporation (Baltimore, MD). The parameters of metabolism of probe substrates and demographic information of the tissue donors for each of the lots may be found in the specification documents provided in the Supplementary Materials. The suppliers of the preparations of HLM used in our studies, BioIvt Corporation and Corning Life Sciences, have declared to adhere to the regulations of the Department of Health and Human Services for the protection of human subjects (45 CFR §46.116 and §46.117) and Good Clinical Practice (GLP), (ICH E6) in obtaining the samples of human tissues used for producing the preparations of human subcellular fractions available from these companies.

The composition of the cytochrome P450 ensemble in these preparations was characterized by mass-spectrometric analysis with a triple quadrupole mass spectrometer using the method of multiple reaction monitoring (MRM) as described previously [14]. Content of NADPH-cytochrome P450 reductase and cytochromes P450 1A2, 2A6, 2B6, 2D6, 2C8, 2C9, 2C18, 2C19, 2E1, 3A4, and 3A5 was determined based on the concentration of protein-specific tryptic peptides in trypsin-digested samples with the use of the stable isotope-labeled internal standards (SIS) [14]. For each protein, one standard peptide with three transitions was used. The peptides, which sequences are provided in Table 1, were arranged into one Selected Reaction Monitoring (SRM) assay. Mass-spectrometry measurements of P450 content were performed using the equipment of “Human Proteome” Core Facilities of the Institute of Biomedical Chemistry (Moscow, Russia).

Table 1.

Composition of the cytochrome P450 pool in HLM preparations under study^*

Protein	Peptide used in the assay	HLM-N1		HLM-N2		HLM-N3		HLM-A		A-to-N ratio c
		Specific fraction, %a	Pro-rated content, pmol/mgb	Specific fraction, % a	Pro-rated content, pmol/mgb	Specific fraction, % a	Pro-rated content, pmol/mgb	Specific fraction, % a	Pro-rated content, pmol/mgb
CYP1A2	ASGNLIPQEK	17.0 ± 1.3	33	5.4 ± 0.7	17	5.0 ± 0.6	16	8.0 ± 1.4	22	0.9
CYP2A6	GTGGANIDPTFFLSR	17.8 ±0.7	35	16.9 ± 1.8	53	21.3 ± 2.4	69	15.0 ± 2.9	41	0.8
CYP2B6	GYGVIFANGNR	0.6 ± 0.1	1.2	1.0 ± 0.2	3.2	1.1 ± 0.2	3.7	0.6 ± 0.1	1.6	0.6
CYP2C8	GLGIISSNGK	2.0	4.0	1.1	3.5	0.6 ± 0.4	1.8	2.2 ± 0.3	6.0	1.8
CYP2C9	GIFPLAER	17.0 ± 1.5	33	14.5 ± 2.2	46	15.6 ± 1.6	50	12.0 ± 2.6	33	0.8
CYP2C18	IAENFAYIK	0.5 ± 0.3	1.0	0.33 ± 0.05	1.0	0.24 ± 0.05	0.8	0.9 ± 0.20	2.5	2.5
CYP2C19	ICVGEGLAR	1.8 ± 0.8	3.6	0.33 ± 0.06	1.0	0.28 ± 0.06	0.9	0.6 ± 0.3	1.6	0.7
CYP2D6	LLDLAQEGLK	0.42 ± 0.01	0.8	0.59 ± 0.16	1.9	0.51 ± 0.05	1.7	1.0 ± 0.4	2.6	1.9
CYP2E1	GIIFNNGPTWK	19.4 ±3.1	38	13.9 ± 1.3	44	11.3 ± 1.3	36	38.1 ± 2.2	104	2.6
CYP3A4	LSLGGLLQPEKPVVLK	22.2 ±3.0	43	43.5 ± 5.8	137	43.7 ± 6.6	142	21.0 ± 6.2	57	0.6
CYP3A5	DVEINGVFIPK	1.0 ±1.1	2.0	2.4 ± 1.1	7.5	0.4 ± 0.3	1.3	0.7 ± 0.1	1.9	0.5

^*The values that are given for HLM-N1, HLM-N2, HLM-N3, and HLM-A are the averages of 4, 20, 15, and 11 individual measurements, respectively. In each individual assay, we determined the concentration of all P450 species under analysis, except for CYP2C8, which content in HLM-N1 and HLM-N2 was determined only once. The results of CYP2C8 determination in HLM-N3 and HLM-A represent the averages of 2 and 5 measurements, respectively. The “±” values are the confidence intervals calculated for p=0.05.

^aThe values of the specific fraction of P450 species shows the fractional content of each P450 enzyme in the total of all 11 P450 species analyzed and represent the averages of the values calculated for each individual dataset.

^bApparent molar content of the P450 species per milligram of microsomal protein. Calculated from the respective fractional content and the spectrally-detectable P450 concentration assuming that the total of 11 analyzed P450 proteins represent 80% of the microsomal P450 pool [33].

^c“A-to-N ratio” is the ratio of the fractional content of a given P450 species in HLM-A to an average of the respective contents in HLM-N samples. This ratio characterizes the direction and the amplitude of apparent alcohol-induced changes in the content of P450 enzymes.

2.4. Characterization of the content of protein, phospholipids, NADPH-cytochrome P450 reductase, and cytochromes P450 in HLM

Determinations of protein and phospholipid concentrations in microsomal suspensions were performed with the bicinchoninic acid procedure [18] and through the determination of total phosphorus in a chloroform/methanol extract according to Bartlett [19], respectively. The concentration of NADPH-cytochrome P450 reductase in microsomal membranes was determined based on the rate of NADPH-dependent reduction of cytochrome c at 25 °C and the effective molar concentration of CPR was estimated using the turnover number of 3750 min⁻¹ [14]. The total concentration of cytochromes P450 in HLM was determined with a variant of the “oxidized CO versus reduced CO difference spectrum” method described earlier [14].

2.5. Preparation of CYP2E1-enriched HLM samples

Incorporation of CYP2E1 into HLM was performed by incubating undiluted suspensions of HLM (20-25 mg/ml protein, 10-13 mM phospholipid) in 125 mM K-Phosphate buffer containing 0.25M Sucrose with purified CYP2E1 for 16 - 20 hours at 4°C at continuous stirring. CYP2E1 was added in the amount ranging from 0.25 to 1 molar equivalents to the endogenous cytochrome P450 present in HLM. Following the incubation, the suspension was diluted 4-8 times with 125 mM K-Phosphate buffer, pH 7.4 containing 0.25 M sucrose, and centrifuged at 150,000 g in an Optima TLX ultracentrifuge (Beckman Coulter Inc., Brea, CA, USA) with a TLA100.3 rotor for 90 min at 4 °C. The pellet was resuspended in the same buffer to the protein concentration of 15-20 mg/ml.

2.5. Determination of the degree of incorporation of added CYP2E1 into HLM.

The amount of cytochrome P450 incorporated with the above procedure was routinely calculated from the difference between the heme protein added to the incubation media and the enzyme found in the supernatant. According to this assay results, our procedure resulted in the membrane incorporation of 90 - 98% of the added protein.

To ascertain that the added CYP2E1 protein that misses from the supernatant is indeed incorporated into the membrane of HLM, we used the method for selective determination of the concentration of CYP2E1 with absorbance spectroscopy introduced in our previous study [12]. It is based on our observation that, in contrast to other microsomal P450 proteins, CYP2E1 interacts with dithiothreitol (DTT) with the formation of the hyperporhyrin state, where two thiolate groups serve as axial ligands of the heme iron. This bis-thiolate-ligated state has a peculiar absorbance spectrum with a characteristic split of the Soret band into two peaks positioned around 380 and 460 nm [20]. In our assay procedure, we diluted 10μl of microsomal preparations with 100 μl of the membrane solubilization buffer (MSB, 100 mM K-phosphate buffer pH 7.4, 10% glycerol, 0.5% sodium cholate, 0.4% Igepal CO-630, 1mM EDTA), placed the mixture into a quartz micro cell (absorbance path length 1cm) and performed spectrophotometric titrations with DTT. Typically, 5-6 additions of 1 M stock solution of DTT were made to reach the final concentration of DTT of 50-60 mM. After each addition of DTT, the absorbance spectrum was recorded in the region of 340-700 nm. The resulting series of spectra was subjected to the procedure of principal component analysis (PCA) combined with the approximation of the spectra of principal components with the prototypical spectra of absorbance of the ferric low-spin, ferric high-spin, ferric nitrogen-coordinated (“Type-II complex”), and the bis-thiolate states of CYP2E1. The latter spectral standard was obtained from the titrations of purified CYP2E1 with p-chlorothiophenol. The concentration of the bis-thiolate complex reached at saturation with DTT was used to calculate the concentration of CYP2E1 in the samples.

2.6. Determination of the rates of metabolism of p-nitrophenol and ivermectin

The rate of p-nitrophenol (pNP) hydroxylation was determined with absorbance spectroscopy as previously described [14]. The rate of O-demethylation of ivermectin was quantified with the determination of formed formaldehyde by the Nash method [21] with fluorometric detection. The assays were performed in 0.1 M Na-HEPES buffer, pH 7.4, containing 60 mM KCl. The incubation mixture in the volume 48 μL contained 1 - 2 μl of HLM suspension. Ivermectin was added as a 50 mM solution in acetone. The reaction was started with the addition of 20 mM NADPH to the final concentration of 500μM. The samples were incubated in a shaking water bath for 20 min at 30 °C, and the reaction was terminated by the addition of 12 μL of 20% trichloroacetic acid. The precipitated protein was removed by centrifugation at 9,600g for 10 min, and the supernatant was supplemented with the equal volume of the Nash reagent (solution of 2 M ammonium acetate and 20 mM acetylacetone in 50 mM acetic acid) and incubated for 10 min at 60 °C. The spectra of fluorescence of the samples in 450 – 680 nm region with excitation at 405 nm were recorded in a 3 x 3 mm microcuvette. A spectrum of the sample treated as above and containing all ingredients except for the substrate was used as a reference. The spectrum of fluorescence of the Nash reaction product with 1 μM formaldehyde obtained in calibration experiments was used to quantify the amount of the formed formaldehyde through a linear least-squares approximation using our SpectraLab software [22–23].

2.7. Determination of the rates of metabolism of amitriptyline and midazolam

The rates of N-demethylation of amitriptyline and hydroxylation of midazolam were determined with quantification of the formed products (nortriptyline and 1′-hydroxymidazolam, respectively) by liquid chromatography-mass spectrometry (LC-MS/MS). The incubation conditions were similar to those described above for the ivermectin metabolism assays. The substrates were added as solutions in methanol, which concentration was kept equal to 0.625%. The volume of each probe was 60 μl, and the reaction was started with the addition of 1.5 μl of 20 mM NADPH. After incubation of the probes at 30 °C in a shaking water bath for 2.5 (midazolam) or 15 min (amitriptyline), the reaction was quenched with the addition of 15 μL of 1 M solution of formic acid in acetonitrile containing a known concentration of an internal standard. The probes were centrifuged at 9,300 g for 10 min and subjected to LC-MS/MS analysis.

An LC-20AD series high-performance liquid chromatography system (Shimadzu, Columbia, MD) fitted with an HTC PAL autosampler (LEAP Technologies, Carrboro, NC) was used to perform chromatography on a Kinetex® reverse-phase column (100 × 2.1 mm, Phenomenex, Torrance, CA). Determination of the content of 1′-hydroxymidazolam was performed as described earlier [24] with some modifications in MS tuning parameters [25] and the use of phenacetin as an internal standard. In the case of noramitriptyline, chromatographic separation was performed using a binary method at the total flow rate of 0.3 ml/min. The separation was initiated by 0.3 min flow at 5% mobile phase B, which was then increased to 10% and 100% until 1.0 and 4 min, respectively. The flow of mobile phase B was gradually decreased and equilibrated to initial flow conditions for 1 min. The metabolite quantification was conducted using an API 4000 Q-Trap tandem mass spectrometry system manufactured by Applied Biosystems/MDS Sciex (Foster City, CA) using turbospray ESI operating in positive ion mode. The mass spectrometer parameters were set at curtain gas, 10; collision gas, medium; ion spray voltage, 4500; ion source gas 1, 15; ion source gas 2, 5; temperature, 400, declustering potential, 55; entrance potential, 5; collision energy, 29, collision cell exist potential, 10. The analyte (nortyptyline) and the internal standard (2-methyl-4(3H)-quinazolinone) were detected using MRM mode by monitoring the m/z transition from 264.0 to 91.4 and 161.0 to 120.10, respectively. The determination of the product amounts was achieved with a calibration curve covering the range of nortriptyline concentrations from 0 to 5 μM.

2.8. Fluorimetric assays of BQ metabolism

The rate of debenzylation of 7-benzyloxyquinoline was measured with a real-time continuous fluorometric assay using a Cary Eclipse fluorometer (Agilent Technologies, Santa Clara, CA, USA) or a custom-modified PTI QM-1 fluorometer (Photon Technology International, New Brunswick, NJ [13]. In the experiments with Cary Eclipse, the excitation was performed with a monochromatic light centered at 405 nm with 5 nm bandwidth. In the case of PTI QM-1, the excitation light centered at 405 nm was emitted by a CPS405 collimated laser diode module (Thorlabs Inc, Newton, NJ). The emission wavelength was set at 516 nm with a 20 nm slit. The rate of formation of 7-hydroxyquinoline was estimated by determining the slope of the linear part of the kinetic curve recorded over 2 – 3 min.

All kinetic assays were performed in 0.1 M Na-HEPES buffer, pH 7.4, containing 60 mM KCl. In the case of the use of Cary Eclipse, the total volume of the incubation mixture was equal to 300μl, and a 5 x 5 mm quartz cell was used. In the experiments with PTI QM-1 fluorometer, we used a 3 x 3 mm quartz cell, and the volume of the sample was equal to 60μl. With both instruments, the kinetic assays were carried out at continuous stirring, and the temperature was maintained at 30 °C with a circulating water bath. An aliquot of 15-24 mM stock solution of BQ in acetone was added to attain the desired substrate concentration in the range of 0.5 - 250 μM. The reaction was initiated by adding 20 mM solution of NADPH to the concentration of 200 μM. Fitting of the dependencies of the reaction rate on the substrate concentration to Michaelis-Menten and Hill equations was performed with a combination of Marquardt and Nelder-Mead non-linear regression algorithms as implemented in our SpectraLab software [22–23].

2.9. Monitoring the interactions of CYP2E1-BODIPY with microsomal membranes

Labeling of CYP2E1 with BODIPY-618 maleimide was performed at a 2:1 label to protein ratio, as described earlier [12–14, 26]. The studies of interactions of CYP2E1-BODIPY with microsomal membranes were performed using a Cary Eclipse spectrofluorometer equipped with a Peltier 4-position cell holder. The excitation of donor phosphorescence was performed with monochromatic light centered at 405 with 20 nm bandwidth. Alternatively, the measurements were done using a custom-modified PTI QM-1 fluorometer (Photon Technology International, New Brunswick, NJ) equipped with a thermostated cell holder and a refrigerated circulating bath. In this case, the excitation was performed at 405 nm with a CPS405 collimated laser diode module (Thorlabs Inc, Newton, NJ). The spectra in the 570 – 750 nm wavelength region were recorded repetitively, with the time interval varying from 0.5 to 15 min during the course of monitoring (5 – 16 hours). The experiments were performed at continuous stirring at 4 °C in 100 mM Na-Hepes buffer (pH 7.4) containing 150 mM KCl and 250 mM sucrose.

Calculations of the surface density of cytochromes P450 in microsomal membranes (C_CYP2E1) were based on the molar ratio of membranous phospholipids to incorporated CYP2E1 (R_L/P). For our calculations, we assumed the area of microsomal membrane corresponding to one molecule of phospholipid in a monolayer to be equal to 0.95 nm² [27], similar to the approach used in our earlier reports [12–14]. The use of this value results in the following relationship between C_CYP2E1 and the R_L/P:

$$C_{CYP2E1}=349.6/R_{L/P}\mathrm{pmol}/{\mathrm{cm}}^2$$ (Eq. 1)

Analysis of the series of spectra in fluorescence spectroscopy experiments was done by principal component analysis (PCA) [22] as described [14]. The equation for the equilibrium of binary association (dimerization) used in the fitting of oligomerization isotherms (dependencies of FRET efficiency on the concentration of P450 in membranes) had the following form:

$$\left[X\right]=\frac{1}{2}·K_D+{\left[E\right]}_0-\sqrt{\frac{1}{4}·K_D^2+{\left[E\right]}_0·K_D}$$ (Eq. 2)

where [E]₀, [X], and K_D are the total concentration of the associating compound (enzyme), the concentration of its dimers, and the apparent dissociation constant, respectively. To use this equation in the fitting of the dependencies of the relative increase in fluorescence (R_E) observed at enzyme concentration [E]₀, equation (1) was complemented with the parameter R_max. This parameter corresponds to the value of R_E observed upon a thorough dissociation of completely oligomerized enzyme:

$$R_E=R_{max}·\left(\left[X\right]/{\left[E\right]}_0\right)$$ (Eq. 3)

The parameter R_max is determined by the efficiency of FRET (E_FRET) according to the following relationship:

$$E_{FRET}=R_{max}/\left(1+R_{max}\right)$$ (Eq. 4)

Fitting of the titration isotherms to the above equations was performed with non-linear regression using a combination of Nelder-Mead and Marquardt algorithms as implemented in our SpectraLab software [22].

3. Results.

3.1. Characterization of the composition of the cytochrome P450 pool in HLM samples

The preparations of HLM used in this study were characterized with mass-spectrometry by determining the content of NADPH-cytochrome P450 reductase, cytochrome b₅, and 11 major cytochrome P450 species, namely CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C18, CYP2C19, CYP2E1, CYP3A4, and CYP3A5. The results of our analysis of the composition of the P450 pool in HLM are summarized in Table 1.

The total content of cytochromes P450 and the concentration of NADPH-cytochrome reductase (CPR) and cytochrome b₅ were also quantified with absorbance spectroscopy and determining the rate of NADPH-dependent reduction of cytochrome c. In Table S1 presented in the Supplementary Materials, we compare the results of spectrophotometric and LC-MS/MS assays of the contents of CPR, cytochrome b₅, and cytochrome P450. Significant scatter in the degrees of coverage of the concentrations determined in the spectrophotometric assays by the values derived from the LC-MS/MS analysis suggests a substantial variability in the efficiency of proteolytic digestion in our experiments. To compensate for these variations, which are among the principal causes of errors in the assays with quantitative proteomics [28–29], we used cytochrome b₅ as a natural internal standard. As shown in Fig. S1, the total quantity of cytochromes P450 determined by LC-MS/MS exhibits a strict linear correlation with the quantified amount of cytochrome b₅. This correlation suggests that, despite varying digestion efficiency, our assays provide robust estimates of the fractional contents (percent of each P450 enzyme in the total of 11 analyzed species) of the analyzed P450 proteins.

Furthermore, the concentration of cytochrome b₅ in HLM determined with absorbance spectroscopy provides a robust basis for quantification of cytochromes P450. When scaled proportionally to the LC-MS/MS-detectable cytochrome b₅, the total detected content of the 11 cytochrome P450 species amounts to 140-160% of the cytochrome P450 concentration determined from the CO-differential spectra. The excess of the pro-rated content of the LC-MS/MS-detectable cytochromes P450 over the amount measured with absorbance spectroscopy may be explained by incomplete saturation of the P450 proteins in HLM with the heme cofactor [30–32].

Due to considerable variation in the degree of coverage in the individual LC-MS/MS runs, the absolute quantities of the cytochrome P450 species presented in Table 1 were calculated from the respective fractional contents and the amount of spectrally-detectable P450 hemeprotein. In these calculations, we assumed that a total of 11 analyzed P450 proteins represent 80% of the microsomal P450 pool. This estimate was calculated from the data Wang and co-authors, who determined the content of 25 P450 species in a series of 102 individual HLM samples [33]. The values shown in Table 1 for HLM-N1, HLM-N2, HLM-N3, and HLM-A are the mean values averaged over the results of 4, 20, 15, and 11 individual measurements, respectively. As expected, the fractional content of CYP2E1 in HLM-A was from 2.7 to 3.5-fold higher than that detected in the HLM-N samples.

3.2. Rates of metabolism of CYP2E1- and CYP3A4-specific probe substrates by pooled HLM preparations.

Activities exhibited by pooled HLM preparations with probe substrates of CYP2E1 and CYP3A4 enzymes are summarized in Table 2. Here we provide activities of HLM-N2, HLM-N3, and HLM-A with p-nitrophenol, midazolam, ivermectin, and amitriptyline. Table 2 compares the rates of metabolism of the probe substrates normalized on the concentration of CPR with the values obtained upon normalization on the content of cytochrome P450 species primarily involved in the metabolism of the respective substrates. Given a high excess of cytochromes P450 over CPR in HLM, the normalization on the concentration of CPR was considered as the most appropriate approach for comparing the rates of substrate metabolism in the HLM preparations with different compositions of the cytochrome P450 pool. On the other hand, normalization of the rates of metabolism of isoform-specific substrates on the concentration of their primary metabolizing enzyme is expected to reveal the differences in the catalytic activities of the respective individual enzymes. The full sets of kinetic parameters obtained in our studies with p-nitrophenol and amitriptyline may be found in Table S2 (see Supporting Materials).

Table 2.

Rates of metabolism of CYP2E1 and CYP3A4 probe substrates by pooled HLM preparations used in this study^*.

Substrate	Primary metabolizing P450 species	Turnover rate, min−1 (per CPR)				Turnover rate, min−1 (per metabolizing P450 enzyme)
		HLM-N1	HLM-N2	HLM-N3	HLM-A	HLM-N1	HLM-N2	HLM-N3	HLM-A
p-Nitrophenol	CYP2E1	4.8 ± 1.4	5.8 ± 1.2	5.2 ± 3.0	10.3 ±2.8	7.3 ± 2.1	8.9 ±5.1	6.8 ± 1.4	3.4 ± 0.9
Midazolam	CYP3A4	N/A	7.3 ± 2.1	8.2 ± 4.0	11.2 ±5.3	N/A	3.3 ±0.9	3.1 ± 1.5	6.8 ± 3.2
Ivermectin	CYP3A4	6.0 ± 1.2	6.3 ± 2.0	6.5 ± 2.3	8.0 ±1.8	4.7 ± 0.9	2.8 ± 0.9	2.4 ± 0.9	4.9 ± 1.1
Amitriptyline	CYP3A4	N/A	15.7 ± 1.7	10.4 ± 1.9	26.7 ± 2.3	N/A	5.9 ± 0.6	4.6 ± 0.8	16.2 ± 1.4

^*The rates of turnover of ivermectin and midazolam were measured at 100 μM substrate concentration. The rates shown for p-nitrophenol and amitriptyline correspond to the V_max values. The full sets of the kinetic parameters obtained with these substrates may be found in the Table S2 provided in the Supplementary Materials. All values represent the averages of the results of 2 - 6 individual experiments. The “±” values are the confidence intervals calculated for p=0.05.

^aThe rates of turnover normalized on the concentration of CPR determined from the rate of NADPH-dependent reduction of cytochrome c (see Materials and Methods).

^bThe rates of turnover normalized on the concentration of the primary metabolizing cytochrome P450 species calculated from the results of LC-MS/MS assays (see Table 1).

^dN/A – not available.

Of the four substrates used in our assays, midazolam and ivermectin are highly specific probes for CYP3A4 [34–35], while p-nitrophenol is metabolized predominantly by CYP2E1 [36–37]. N-demethylation of amitriptyline is catalyzed primarily by CYP2C19 and CYP3A4 [38–39], of whose the affinity of CYP2C19 to the drug (reported Km values are around 14 μM [38–39]) is much higher than that exhibited by CYP3A4. However, due to a very low abundance of CYP2C19 in HLM, V_max values determined in our experiments covering the range of AMT concentrations up to 500 μM (Fig. S2) are thought to reflect the rate of CYP3A4-dependent metabolism primarily. In conformity with this expectation, the fitting of the AMT titration curves obtained in our experiments with the Michaelis-Menten equation resulted in the K_M estimates of 50-130 μM (Table S2) consistent with the values reported for CYP3A4-dependent N-demethylation [38–39].

As seen from Table 2, the activity of HLM-A preparation with all probe substrates normalized on the concentration of CPR is higher than that exhibited by HLM-N samples. In the case of CYP2E1-specific substrate p-nitrophenol, this effect appears to be entirely due to the high content of the enzyme in HLM from alcohol-exposed donors, as the values obtained upon normalization on the concentration of CYP2E1 did not reveal any activation of the enzyme, but rather suggest some decrease in its specific activity in HLM-A. In contrast, the values of the rate of turnover of CYP3A4-specific substrates by HLM-A are noticeably higher than those exhibited by HLM-N samples with either way of normalization. The rates of turnover obtained with midazolam, testosterone, ivermectin, and amitriptyline normalized on the specific content of CYP3A4 in HLM suggest a multifold increase in the activity of CYP3A4 in HLM-A as compared with the HLM-N samples. In our further experiments, we explored the effect of an increase in the content of CYP2E1 in HLM on the functional properties of CYP3A4 in more detail using 7-benzyloxyquinoline as a CYP3A4-specific probe substrate [16].

3.3. Parameters of BQ metabolism exhibited by four HLM-preparations

To probe further the effect of the composition of the P450 ensemble on the functional properties of CYP3A4, we determined the parameters of the O-debenzylation of BQ by all four microsomal preparations under study. We also probed the effect of ANF, the prototypical allosteric effector of the enzyme. The results of these studies are summarized in Table 3 and illustrated in Figure 1. Similar to the approach discussed above for other probe substrates, the rate of metabolism of BQ in HLM in this study was normalized on the concentration of CPR in the microsomal membrane (see Materials and Methods).

Table 3.

Parameters of BQ metabolism obtained with four HLM preparations and their modulation by α-naphthoflavone^*

HLM Preparationa	S50, μM		N		Vmax, min−1
	No effector	+ANF	No effector	+ANF	No effector	+ANF
HLM-N1	40.2 ± 12.2	10.7 ± 4.8 (0.01)	2.24 ± 0.37	0.83 ± 0.12 (<10−2)	16.4 ± 1.8	24.9 ± 2.8 (<10−2)
HLM-N1+2E1	89.2 ± 20.4 (0.01)	12.4 ± 4.3 (0.01)	1.26 ± 0.12 (0.01)	0.87 ± 0.16 (0.02)	31.9 ± 5.0 (<10−3)	19.9 ± 2.7 (0.01)
HLM-N2	63.3 ± 3.2	19.7 ± 6.7 (<10−3)	1.92 ± 0.02	1.02 ± 0.37 (0.01)	45.7 ± 5.1	49.3 ± 7.9 (0.53)
HLM-N2+2E1	107 ± 55 (0.25)	16.3 ± 6.9 (0.02)	1.13 ± 0.03 (<10−6)	0.86 ± 0.08 (<10−2)	77.3 ± 4.9 (<10−3)	46.8 ± 16.6 (0.08)
HLM-N3	36.7 ± 10.3	7.4 ± 3.8 (<10−2)	2.09 ± 0.44	1.22 ± 0.30 (0.03)	71.4 ± 11.4	32.6 ± 3.5 (<10−2)
HLM-N3+2E1	89.1 ± 34.3 (0.03)	12.0 ± 4.7 (0.01)	1.28 ± 0.09 (0.01)	1.02 ± 0.24 (0.07)	83.7 ± 38.8 (0.6)	46.1 ± 11.1 (0.13)
HLM-A	59.6 ± 12.5	25.0 ± 10.7 (<10−2)	1.02 ± 0.15	0.99 ± 0.20 (0.03)	101 ± 20	113 ± 24 (0.93)
HLM-A+2E1	65.4 ± 25.5 (0.67)	23.0 ± 13.1 (0.04)	1.47 ± 0.30 (0.30)	0.85 ± 0.26 (0.03)	147 ± 30 (0.03)	127 ± 4 (0.3)

^*The values given in the table are the averages of 3 - 7 individual measurements. The “±” values are the confidence intervals calculated for p=0.05. The values in parentheses represent the p-values of Studenťs t-test for the hypothesis of equality of the respective values with the value obtained with untreated microsomes (for the values in “No effector” columns) or at no ANF added (for the values in “+ANF” columns). The values characterized by p-values ≤0.05 are shown in bold to emphasize the effects with high statistical significance.

^aThe CYP2E1-enriched samples represented in this table are the samples where the content of added CYP2E1 was equal to the total P450 content in untreated HLM (250 – 400 pmol CYP2E1 per mg of protein).

Figure 1. BQ metabolism in intact and CYP2E1-enriched HLM samples.

The figure shows the dependencies of the rate of BQ turnover on substrate concentration in intact (black) and CYP2E1-enriched (red) HLM preparations. The dependencies obtained in the absence of added effector are shown in open circles, while the data obtained in the presence of 25 μM ANF are designated by closed circles. The CYP2E1-enriched samples exemplified in this figure are the samples where the content of added CYP2E1 was equal to the total content of spectrally-detectable P450 in HLM (“1:1 incorporation”, 250 - 400 pmol CYP2E1 pmol per mg of protein). The datasets shown on this figure were obtained by averaging the results of 3 – 5 individual experiments. The solid lines show the results of the fitting of the datasets with the Hill equation.

In all three samples of HLM-N, the metabolism of BQ revealed a prominent homotropic cooperativity characterized by the Hill coefficient of 1.9 – 2.2 (Figure 1, curves shown in black open circles), in a good agreement with the previous reports [16, 40]. In HLM-N samples, the maximal reaction rate increases concomitant with increasing fractional content of CYP3A4 in microsomes (Table 3). Similar to that observed with other CYP3A4-specific substrates, despite the low content of CYP3A4 in HLM-A, this preparation is characterized by the highest maximal rate of BQ metabolism over all four HLM samples studied. V_max value exhibited by HLM-A is over six times higher than the value obtained with HLM-N1, where the fractional content of CYP3A4 (26%) is comparable to that observed in HLM-A (21%). The values of V_max obtained with HLM-N2 and HLM-N3, where the fractional content of CYP3A4 is around 40%, are respectively 2.2 and 1.4 times lower than that characteristic to HLM-A. This activation of BQ metabolism in HLM-A is associated with the attenuation of homotropic cooperativity, which is revealed in a considerable decrease of the Hill coefficient (Table 2).

The contrasting difference between HLM-A and HLM-N preparation in the rate of BQ metabolism is illustrated in Figure 2, where the V_max values are plotted against the molar ratio of CYP3A4 to CPR determined in these preparations with LC-MS/MS. As might be expected, the V_max value observed with HLM-N preparations and calculated per molar content of CPR displays an evident proportionality to the content of CYP3A4A (Figure 2a). However, the rate observed with HLM-A, where the content of CYP3A4 is the lowest out of all four preparations, deviates from this proportionality being the highest of all four values under comparison. The difference between HLM-A and HLM-N becomes even more contrasting when the turnover rates are calculated per molar content of CYP3A4. This way of normalization attenuates the difference between the three HLM-N samples. Simultaneously, the turnover number calculated for HLM-A becomes almost an order of magnitude higher than the values calculated for HLM-N samples (Figure 2b).

Figure 2. The relationship between the CYP3A4 content in HLM preparations and the rate of BQ metabolism at no added effector.

The maximal velocity of BQ debenzylation is plotted against the molar ratio of CYP3A4 to CPR in microsomal membranes calculated from the results of LC-MS/MS analysis (Table 1). The panel A shows the V_max values normalized on the concentration of CPR. The panel B shows the same results normalized per CYP3A4 content.

3.4. The effect of ANF

Consistent with the previous observations [26, 40–41], the addition of 25 μM ANF eliminated homotropic cooperativity observed with BQ and considerably decreased the S₅₀ values in all four HLM samples studied (Table 3). However, the effect of ANF on the maximal rate of reaction differs dramatically between the HLM samples. Whereas in the case of HLM-N1, the addition of ANF increased the maximal rate of reaction almost two times, it resulted in over two-fold inhibition of BQ metabolism in HLM-N3, while having no statistically significant effect on V_max in HLM-N2 and HLM-A (Table 3). The effect of ANF on substrate saturation curves in four HLM samples is illustrated in Figure 1 (curves in black closed circles).

As shown in Figure 3, where the V_max values observed in the presence of ANF are plotted against the content of CYP3A4 in HLM preparations, the addition of this allosteric effector eliminates the proportionality between the rate of BQ turnover and the content of CYP3A4 in HLM-N samples. This ANF-dependent leveling of the turnover rates is achieved through the activating effect of ANF in HLM-N1, the sample with the lowest CYP3A4 content, and a substantial inhibition of BQ metabolism in CYP3A4-rich HLM-N3. On the other side, the addition of ANF does not affect the contrasting difference between HLM-A and HLM-N samples in the rate of BQ turnover (Figure 3).

Figure 3. The relationship between the CYP3A4 content in HLM preparations and the rate of BQ metabolism in the presence of 25 μM ANF.

The maximal velocity of BQ debenzylation is plotted against the molar ratio of CYP3A4 to CPR in microsomal membranes calculated from the results of LC-MS/MS analysis (Table 1). The panel A shows the V_max values normalized on the concentration of CPR. The panel B shows the same results normalized per CYP3A4 content.

3.5. Incorporation of purified CYP2E1 into HLM and the characterization of CYP2E1 enriched HLM preparations.

To probe if the loss of homotropic cooperativity with BQ and a multifold increase in the rate of its metabolism in HLM-A is caused by an alcohol-induced increase in the fraction of CYP2E1 in the P450 pool, we studied the effect of incorporation of exogenous CYP2E1 into HLM. This approach has been successfully used and thoroughly characterized in our previous publications [12, 14]. In these earlier studies, we confirmed the completeness of incorporation of the added protein into the microsomal membranes with selective determination of incorporated CYP2E1 with absorbance spectroscopy [12].

In these assays, we used the method based on our observation that CYP2E1 interacts with dithiothreitol (DTT) to form a bis-thiolate complex of the heme iron. We found that the addition of DTT to purified CYP2E1 in solution resulted in a disappearance of the Soret absorbance bands of the low- and high-spin states of the heme-protein (at 392 and 418 nm, respectively) and a concomitant appearance of the new bands around 370 and 460 nm. These changes are indicative of the formation of the “hyperporphyrine” state of the heme protein, where the 6-th ligand of the heme iron is provided by the thiolate group of an external thiol-containing compound [20]. The formation of bis-thiolate complexes of cytochromes P450 with such organic thiols as chlorothiophenol, where the thiol group is highly ionized at neutral pH, is well documented [20]. However, thiolate compounds with alkaline pK_a, such as DTT, do not form this type of complexes with most P450 enzymes. CYP2E1 appears to be the only known eukaryotic P450 that exhibits this feature.

Our initial experiments with the use of this unique feature of CYP2E1 for monitoring its incorporation into the microsomal membranes were performed with insect cell microsomes (Supersomes) [12]. For the present study, we elected to reproduce them with HLM to probe the completeness of the CYP2E1 incorporation. Due to the limited amounts of HLM preparations HLM-N1, HLM-N2, and HLM-N3 in hands, these control experiments were performed with the pooled HLM preparation from 50 donors lot ODN obtained from BioIVT corporation.

A series of spectra of purified CYP2E1 recorded at increasing DTT concentrations is shown in Fig. 4A. DTT-induced spectral changes illustrated in this figure were completely reversible - the protein returns into its initial state after the removal of DTT by dialysis or size-exclusion chromatography (data not shown). Application of PCA to this spectral series combined with the approximation of the spectra of the first two principal components with a set of prototypical spectra of absorbance of CYP2E1 in each of the four states of ligation of the heme iron (see Materials and Methods) allowed us to quantify the changes in the fraction of P450 in the bis-thiolate state (Fig. 4B). According to our results, the maximal fractional content of the bis-thiolate state reached at the saturation with DTT is equal to approximately 62%. (Fig. 4B).

Figure 4. Spectral transitions in purified CYP2E1 caused by its interactions with DTT.

Panel A shows a series of absorbance spectra of 2.6 μM CYP2E1 in microsome-solubilization buffer (MSB) recorded at DTT concentration increasing from 0 to 62.5 mM (individual spectra correspond to 0, 3.3, 6.6, 10, 16, 32 and 62.5 mM DTT). Inset shows the same series as differential spectra relative to the spectrum taken at no DTT added. Panel B shows the dependence of the fractional content of the bis-thiolate state of the heme protein on DTT concentration. The solid line shows the result of fitting of the dataset to the Hill equation (A_max = 62.4%, S₅₀ = 6.7 mM, N = 2.4). The use of the Hill equation for fitting is arbitrary and has no interpretational value. The inset shows the spectrum of the first principal component of the observed changes (dotted line) along with its approximation with a combination of prototypical standards of the individual states of CYP2E1 (solid line).

As illustrated in Fig. 5A, the DTT-induced changes in the spectra of absorbance of CYP2E1-enriched HLM are similar to those observed with the purified protein. Here again, the amount of formed bis-thiolate state may be quantified using our PCA-based approach (Fig. 5B). Using this technique and assuming that the maximal fraction of CYP2E1 present in the bis-thiolate state at DTT saturation in HLM is similar to that observed with the purified enzyme (62%), we may calculate the apparent concentration of CYP2E1 in the membrane. In Table 4, we characterize the degree of incorporation of added CYP2E1 in three samples with the different amounts of added enzyme. As seen from this table, when the amount of added CYP2E1 corresponds to 25% of the total P450 present in HLM, the degree of incorporation is equal to 100%. Although the degree of incorporation decreases upon increasing the amount of added CYP2E1, it remains as high as 76% at 1:1 molar ratio of the added protein to the microsomal P450 content (Table 4).

Figure 5. Spectral transitions caused by addition of DTT to CYP2E1-enriched HLM..

Panel A shows a series of absorbance spectra of a suspension of HLM enriched with CYP2E1 by incubation with the purified protein added in the amount of 375 pmol per mg of protein. The spectra were recorded at DTT concentration increasing from 0 to 62.5 mM (individual spectra correspond to 0, 3.3, 10, 26, 51 and 62.5 mM DTT). HLM suspension was prepared in MSB buffer and had the protein concentration of 1.8 mg/ml. Inset shows the same series as differential spectra relative to the spectrum taken at no DTT added. Panel B shows the dependence of the concentration of the bis-thiolate state of the heme protein on DTT concentration. The solid line shows the result of fitting of the dataset to the Hill equation (A_max = 0.39μM, S₅₀ = 10.4 mM, N = 1.0). The use of the Hill equation for fitting is arbitrary and has no interpretational value. The inset shows the spectrum of the first principal component of the observed changes (dotted line) along with its approximation with a combination of prototypical standards of the individual states of CYP2E1 (solid line).

Table 4.

Incorporation of purified CYP2E1 into HLM and its effect on the rate p-nitrophenol hydroxylation.

Molar ratio of added CYP2E1 to P450 content in HLM	added CYP2E1, pmol/mg	CYP2E1 detected, pmol/mga	% of incorporation	Rate of p-nitrophenol hydroxylation, min−1 b
0	0	53 ± 48	N/A	2.5 ± 0.4
0.25	99	152 ± 19	100 ± 19	10.2 ± 0.9
0.5	196	215 ± 93	83 ± 48	12.9 ± 1.4
0.96	377	338 ± 61	76 ± 16	8.5 ± 1.4

^*The values given in the table are the averages of 2 individual measurements. The “±” values are the confidence intervals calculated for p=0.05.

^aContent of CYP2E1 determined from titrations of HLM preparations with dithiothreitol.

^bRate of hydroxylation of p-nitrophenol determined at 250 μM concentration of the substrate.

In our previous publication [14], where we used the same HLM-N1 and HLM-N2 preparations as employed here (designated in [14] as HLM1 and HLM2, respectively), we demonstrated that in both these samples, the activity of HLM with p-nitrophenol and chlorzoxazone, another CYP2E1-specific substrate, increases proportionally to the concentration of CYP2E1 up to its content of 200-250 pmol/mg of protein. After this point, the activity of CYP2E1 stabilizes (p-nitrophenol with HLM-N2) or decreases (HLM-N1 with p-nitrophenol and HLM-N2 with chlorzoxazone, see Fig. 6 in [14]). This eventual decrease may be explained by an increased probability of forming non-functional P450 complexes with CPR at very high P450 concentrations in the membrane[14]. Results of the measurements of p-nitrophenol hydroxylation activity of CYP2E1-enriched HLM samples shown in Table 4 are well consistent with this earlier report. These results demonstrate that added CYP2E1 efficiently incorporates into HLM and becomes a fully functional member of the P450 ensemble.

Figure 6. Effect of incorporation of CYP2E1 into HLM-N on the parameters of BQ metabolism at no added effector.

Dependencies of S₅₀ (panel A), V_max (panel B), and N (panel C) on the fractional content of CYP2E1 in a total of 11 analyzed P450 species. The datasets obtained with HLM-N1, HLM-N2, HLM-N3, and HLM-A are shown in black, blue, red, and green, respectively. The error bars show the confidence intervals calculated for p=0.05. Solid lines show the linear approximations of the data.

3.5. Effect of incorporation of CYP2E1 into HLM on BQ metabolism

The dependencies of the reaction rate on substrate concentration with and without the addition of ANF obtained with all four studied HLM samples enriched with CYP2E1 are exemplified in Figure 1 (curves are shown in red). The parameters of BQ metabolism in these preparations may be found in Table 3. As seen from these results, incorporation of CYP2E1 virtually eliminates the heterotropic cooperativity in all samples studied. Furthermore, in two of the three HLM-N preparations, HLM-N1 and HLM-N2, this incorporation results in a significant increase in the maximal rate of BQ turnover. Therefore, in some sense, the effect of incorporating CYP2E1 into HLM on the homotropic cooperativity of CYP3A4 is analogous to the effect of the addition of ANF.

Notably, the addition of ANF to CYP2E1-enriched HLM preparations largely eliminates their differences from the untreated HLM samples in BQ metabolism parameters. With all three CYP2E1-enriched HLM-N samples, the addition of ANF inhibits BQ turnover at substrate saturation, while decreasing S₅₀ values and eliminating heterotropic cooperativity. In the case of CYP2E1-enriched HLM-A, the effects of ANF on V_max and S₅₀ are marginal, while the effect on the Hill coefficient is retained.

The effect of increasing amounts of incorporated CYP2E1 on the parameters of BQ metabolism in HLM-N and HLM-A preparations in the absence and the presence of ANF is illustrated in Figure 6 and Figure 7, respectively. The dependencies of S₅₀, V_max, and the Hill coefficient on CYP2E1 content obtained in the absence of ANF reveal a striking contrast between the “normal” HLM samples and the sample obtained from the alcohol-exposed donors. Enrichment of HLM-N in CYP2E1 resulted in a stepwise decrease in the Hill coefficient, while increasing the values of S₅₀ and V_max. In contrast, incorporation of CYP2E1 into HLM-A, while also increasing V_max, had no statistically significant effect on the values of S₅₀ and the Hill coefficient (Fig 6).

Figure 7. Effect of incorporation of CYP2E1 into HLM-N on the parameters of BQ metabolism in the presence of ANF.

Dependencies of S₅₀ (panel A), V_max (panel B), and N (panel C) on the fractional content of CYP2E1 in a total of 11 analyzed P450 species. The datasets obtained with HLM-N1, HLM-N2, HLM-N3, and HLM-A are shown in black, blue, red, and green, respectively. The error bars show the confidence intervals calculated for p=0.05.

On the other side, the dependencies of the parameters of BQ metabolism on CYP2E1 content obtained in the presence of ANF (Figure 7) show that the addition of this allosteric effector eliminates most of the effects of CYP2E1. However, the increased activity of CYP3A4 in HLM-A preparation remains to be seen (Figure 7). This observation suggests that the grounds of heterotropic cooperativity in CYP3A4 and the origins of its activation by CYP2E1 may involve some common mechanistic elements.

3.6. Interactions of CYP2E1 with CYP3A4 in the membranes of Supersomes and their modulation by ANF

In our recent study, we introduced a homo-FRET-based technique to monitor the incorporation of CYP2E1 into the microsomal membrane and probe the effect of microsomal P450 proteins on the dissociation of CYP2E1 homo-oligomers [14]. This method employs the use of CYP2E1 labeled with BODIPY-618 maleimide in the molar ratio 1:2. Its principle is based on a deep quenching of the label’s fluorescence in the CYP2E1-BODIPY homo-oligomers due to homo-FRET between the labels attached to different loci of the interacting protein molecules [14]. More specifically, the intensity of fluorescence of one of the two BODIPY residues attached to the CYP2E1 molecule is considerably (over ten times) lower than that from the other one, presumably due to extensive FRET to the heme located in its proximity [14]. Association of CYP2E1 into homooligomers results in homo-FRET from the more-fluorescent BODIPY residues to the less-fluorescent ones in the neighboring subunits. As a result, the total fluorescence of BODIPY is substantially (over 90%) decreased in the homo-oligomers of the labeled protein. Upon incorporation of these oligomers into the membrane, their subsequent partial dissociation and exchange of subunits with the oligomers of the endogenous P450 species results in the increase of fluorescence, which is used to judge as to the level of CYP2E1 homo-oligomerization after establishing the steady state [14]. Our interpretation of the observed increase in fluorescence as a result of the changes in the efficiency of homo-FRET we validated through monitoring the depolarization of BODIPY fluorescence that accompanies the process of incorporation and establishing the steady state [14].

The setup of our experiments employs registration of the process of dissociation of homooligomers after their incorporation into the microsomal membrane through monitoring the consequent increase in the intensity of fluorescence (Figure 8). The dependence of the amplitude of the observed increase on the amount of added microsomes (phospholipid-to–CYP2E1-BODIPY ratio, R_L/p) allows determining the apparent dissociation constant (K_D) of the homo-oligomers, or the value of R_L/P at which the degree of homo-oligomerization is equal to 50% (R_L/P,50%). Effect of the presence of a given cytochrome P450 species in the microsomal membrane on R_L/P,50% is interpreted as reflecting its ability to form mixed oligomers with CYP2E1 (thus promoting the dissociation of the homo-oligomers) [14].

Figure 8. Process of dissociation of homo-oligomers of CYP2E1-BODIPY after their incorporation into microsomal membrane monitored by the changes in BODIPY fluorescence.

The spectra shown in the main panel were recorded in the process of incubation of 0.081 μM CYP2E1-BODIPY with SS(3A4) added to the phospholipid concentration of 158 μM (R_L/P = 1950). The spectra were recorded at 0.5, 2, 8, 16, 34, 65, 126, 253, 382, 510, 766, and 1009 min after the addition of Supersomes to CYP2E1-BODIPY. The inset shows the changes in the relative intensity of fluorescence (Re) during incubations performed at R_L/P ratios of 50 (blue), 150 (red), and 1950 (black). The solid lines represent the approximations of the datasets with a three-exponential equation.

To probe the relevance of the effects of CYP2E1 on the functional properties of CYP3A4 to direct physical interactions of these proteins, we explored the ability of CYP2E1 to form mixed oligomers with CYP3A4 with the use of the homo-FRET based approach. In these experiments, we compared the Supersomes containing recombinant CYP3A4 overexpressed together the recombinant CPR and cytochrome b₅ (SS(3A4)) with the preparation of Supersomes containing recombinant CPR and cytochrome b₅, but lacking any P450 protein (SS(CPR)). Dependencies of the amplitude of an increase in the intensity of fluorescence on the resulting surface density of CYP2E1-BODIPY in the membrane (or R_L/p ratio) are shown in Figure 9. The parameters deduced by fitting these dependencies to Eq. 2 are summarized in Table 3.

Figure 9. The dependencies of the changes in fluorescence caused by incorporation of CYP2E1-BODIPY into Supersomes on the concentration (surface density) of CYP2E1-BODIPY in the membrane.

The datasets obtained with SS(CPR), SS(3A4) at no effector added, and SS(3A4) in the presence of 25 μM ANF are shown in blue, black, and red, respectively. Solid lines show the approximations of the data sets with a combination of Eq. 2 and Eq. 3. Panel A shows the changes in the relative increase in the intensity of fluorescence (R_E in Eq. 3). Dashed lines shown in this panel indicate the maximal level of increase (R_max) estimated from the fitting. Panel B shows the same datasets scaled to the percent of maximal fluorescence.

A comparison of the dependencies obtained with SS(3A4) and SS(CPR) reveals a dramatic effect of microsomal CYP3A4 on homo-oligomerization of the incorporated CYP2E1-BODIPY. If in the membranes of SS(CPR), 50% dissociation of the homo-oligomers requires the molar ratio of membrane phospholipids to CYP2E1 to be as high as 14,400:1, the presence of CYP3A4 decreases this ratio to 176:1 (Table 5). A similar effect of the presence of another P450 species in the microsomal membrane on CYP2E1 homooligomerization has been reported in our earlier studies of CYP2E1 interactions with CYP1A2 [14]. In the same study, we also demonstrate that the presence of CYP2C19 in the membrane elicits no effect on CYP2E1 homooligomerization. Together with the present results on CYP2E1-CYP3A4 interactions, these observations demonstrate the high potential of our approach in probing the specificity of interactions of CYP2E1 interactions with other individual P450 enzymes. Another important finding in the present study is a two-fold decrease in the apparent maximal amplitude of the increase in fluorescence intensity in the presence of CYP3A4. This observation may reveal a conformational effect of the interactions of CYP3A4 with CYP2E1, resulting in a decreased intensity of BODIPY fluorescence in CYP3A4-CYP2E1 complexes.

Table 5.

Effect of membrane incorporation of CYP2E1-BODIPY on the degree of its homooligomerization explored with homo-FRET^*

Membranous preparation	[CYP2E1]50%a, pmol/cm2	(Lipid/CYP2E1)50% b	FRET Efficiency, %
SS(CPR)	0.024 ± 0.005	14,400 ± 2,850	84 ± 10
SS(3A4)	2.0 ± 0.15	176 ± 14	74 ± 9
SS(3A4)+ANF	0.58 ± 0.06	604 ± 62	84 ± 20

^*The values given in the table were obtained by fitting the titration curves with the binary association equation (Eq. 2) in combination with equation defining the relationship between the relative increase in the intensity of fluorescence and the steady-state concentration of the homo-oligomers (Eq. 3). The “±” values show the confidence interval calculated for p = 0.05.

^aSurface density of P450 in the membranes at which the amplitude of the titration curves reaches 50% of the maximal level. This value is equal to the value of K_D determined from the fitting of the titration curves with Eq. 2 combined with Eq. 3. The surface density was calculated from the R_L/P ratio, as described in Materials and Methods using Eq. 1.

^bLipid-to-P450 ratio at which the amplitude of the titration curves reaches 50% of the maximal level, which was calculated from the respective [CYP2E1]_50% values.

^cFRET efficiency was calculated from the values of R_max (Eq. 3) according to the relationship given by Eq. 4.

We also probed the effect of ANF on hetero-association of CYP2E1 with CYP3A4 in a series of experiments with the incorporation of CYP2E1-BODIPY into SS(3A4) in the presence of 25 μM ANF. As seen from Table 5, the addition of ANF increases the value of R_L/P,50% to 604:1, and eliminates the effect of CYP3A4 on the maximal intensity of fluorescence. These observations indicate that the addition of ANF attenuates the ability of CYP3A4 to form mixed oligomers with CYP2E1 and eliminates its apparent conformational effect revealed in the decreased intensity of fluorescence of CYP2E1-BODIPY.

4. Discussion

4.1. Changes in the composition of the cytochrome P450 ensemble under the influence of chronic alcohol exposure

Analysis of the content of 11 drug-metabolizing cytochrome P450 species in pooled HLM preparation obtained from alcoholic donors revealed a 2.0 – 3.4 fold increase in the fractional content of CYP2E1 as compared to the microsomal samples unaffected by alcohol exposure (Table 1). This observation is in good agreement with the previous reports [7–10].

Importantly, our results reveal no alcohol-induced increase in the contents of CYP3A4 and CYP2A6, notwithstanding the existing reports on the induction of expression of these enzymes by ethanol in model systems [42–44]. This discrepancy may reveal a limited relevancy of the results obtained in model animals and cell culture to alcohol effects in humans. Furthermore, in the studies of the effect of alcohol on CYP3A4 expression in human hepatocytes [42] and HepG2 cells [43], a noticeable elevation in the CYP3A4 content was observed only upon prolonged exposure of the cells to the concentrations of ethanol of 50 mM (0.26 %) and above, which is considerably higher than the expected physiologically-relevant concentrations of alcohol in human liver [45], at least for a long time exposure.

An intriguing and hard-to-explain peculiarity of the HLM-A preparation is a low degree of coverage of the spectrally determined quantities of both cytochrome P450 and cytochrome b₅ by their amounts recovered in LC-MS/MS assays (Table S1). Since this low degree of recovery is equally characteristic to both cytochrome P450 and cytochrome b₅, it is likely to stem from a low efficiency of proteolytic digestion observed with the HLM-A sample. It may reflect some characteristic features of the lipid environment in HLM-A that cause decreased accessibility of membrane-incorporated proteins for trypsinolysis.

4.2. Multifold activation of CYP3A4 by alcohol exposure is stipulated by the effect of CYP2E1 on the landscape of P450-P450 interactions

The most striking result of this study is a demonstration that the increase in the content of CYP2E1 observed in alcohol-exposed donors is associated with a considerable amplification of the rate of metabolism of BQ, a highly-specific substrate of CYP3A enzymes not metabolized by CYP2E1. Although the fractional content of CYP3A enzymes in HLM-A was the lowest among all four studied preparations, these microsomes exhibited a much higher rate of metabolism and lower homotropic cooperativity with BQ as compared to HLM-N samples (Figure 1). Notably, in contrast to two of the three HLM-N preparations, HLM-A does not show any ANF-induced increase in V_max of BQ debenzylation. Incorporation of additional CYP2E1 into HLM-A results in a further increase in BQ metabolism rate, making it over an order of magnitude faster than that exhibited by the HLM-N1 sample, which is characterized by a similar CYP3A content.

The finding that the activation of BQ metabolism and attenuation of its cooperativity may be reproduced in HLM-N samples via incorporation of the exogenous CYP2E1 into their membranes suggests that these effects, at least in part, represent direct results of the alcohol-induced increase in CYP2E1 content. It is extremely unlikely that the changes in the parameters of BQ metabolism upon CYP2E1 incorporation into HLM may result from the activity of CYP2E1 itself or be due to participation of any other P450 enzyme, as BQ metabolism is known to be almost entirely selective to CYP3A4. CYP2E1 itself lacks any potential to metabolize this substrate [12, 16]. According to the data of Stresser and co-authors [16], the only other enzymes present in adult HLM and capable of BQ metabolism are CYP1A2 and CYP3A5. However, the turnover numbers exhibited by these enzymes with BQ are respectively 8.5 and 4.3 times lower than that characteristic to CYP3A4 [16]. Thus, due to the relatively low abundance of CYP1A2 and CYP3A5 in HLM (Table 1), their sizable participation in BQ metabolism is unlikely. Therefore, we can conclude that the effects of CYP2E1 incorporation demonstrate that the increased rate and decreased homotropic cooperativity in BQ metabolism seen in HLM-A are attributable, at least in part, to direct effects of the increased CYP2E1 content on the functional properties of CYP3A4.

The striking effects of one P450 protein on the activity of another P450 enzyme observed in this study and evidenced in earlier reports from our group and by others [12–13, 46–48] may be best explained with a concept of “positional heterogeneity” in P450 oligomers [1–2, 12, 14, 49]. This concept is based on considering the microsomal cytochrome P450 ensemble as existing in equilibrium between the monomeric enzymes and their oligomers, and the heterooligomers of multiple P450 species, in particular. The composition of these heterooligomers is determined by the composition of the P450 pool, mutual affinities of different P450s, and the presence of their substrates. According to our hypothesis, the P450 subunits forming these oligomers are not identical in their conformation and orientation, but are characterized by different abilities to be reduced, bind substrates, and interact with redox partners [12, 14]. In effect, oligomerization results in abstracting a large part of the P450 pool from the catalytic activity. A discussion of the possible structural basis for positional heterogeneity in P450 oligomers may be found in a comprehensive review by Reed and Backes [49].

From the standpoint of the above concept, the activation of one P450 enzyme by its interaction with another one might reveal a difference between the two proteins in their propensities for occupying the positions of two different types. The results of our experiments with the incorporation of CYP2E1-BODIPY into CYP3A4-containing Supersomes (Figure 9, Table 3) provide strong evidence of high-affinity interactions between CYP3A4 and CYP2E1, resulting in the formation of their heteromeric complexes. This conclusion is supported by our earlier studies with an LRET-based technique [12]. Therefore, it might be hypothesized that, when CYP3A4 interacts with CYP2E1, the former preferentially occupies the “active” positions in the oligomers, whereas CYP2E1 fills the “restrained” ones. The resulting redistribution of CYP3A4 between the “active” and “restrained” positions increases the active fraction of this enzyme in HLM and results in increased catalytic turnover of CYP3A4-specific substrates.

4.3. Variability of heterotropic cooperativity of CYP3A4 in HLM and its modulation by P450-P450 interactions

Four preparations of HLM examined in this study revealed notable differences in their response to the addition of ANF, a prototypical allosteric effector of CYP3A enzymes. Although the attenuation of the homotropic cooperativity with BQ and an increase in the affinity to this substrate was revealed with all four samples, the effect of ANF on the maximal rate of BQ turnover was contrastingly different. Whereas in the case of HLM-N1, the addition of ANF substantially increased the maximal rate of reaction, it resulted in over two-fold inhibition in HLM-N3 while having no statistically significant effect on V_max in HLM-N2 and HLM-A (Table 3).

Notably, an increase in the fractional content of CYP2E1, either due to chronic alcohol exposure or through the incorporation of the exogenous protein, resulted in a substantial decrease in the degree of heterotropic cooperativity seen with BQ (Table 2). Therefore, the effect of enrichment of the microsomal membrane with CYP2E1 on BQ metabolism reveals some parallelism with the effect of ANF.

The mechanisms of heterotropic cooperativity in CYP3A4 became a subject of extensive studies and a vigorous discussion over the last two decades. The initial interpretation of all known examples of CYP3A4 cooperativity, either homo- or heterotropic, is that the large substrate-binding pocket of the enzyme sometimes requires simultaneous binding of several substrate molecules to assure a productive orientation of at least one of them (see [50–52] for review). The presence of at least two molecules of some substrates in the binding pocket of CYP3A4 is well established [53–55]. Although this model provides a reasonable explanation for most cases of homotropic cooperativity, it fails to explain complex instances of heterotropic activation in CYP3A4 [26].

More recent studies compellingly demonstrated the presence of a separate allosteric effector binding site and revealed its role in the instances of heterotropic activation of CYP3A4 by various effectors, including ANF [26, 56–61]. This allosteric site is formed at the vicinity of the F’ and G’ helices and located at the interface between the two CYP3A4 molecules in the crystallographic dimer of the CYP3A4 complex with peripherally-bound progesterone (PDB 1W0F [62]). As we discussed earlier [26], the high-affinity interactions at this site require the protein to be oligomeric – ligands are unlikely to bind with high affinity at this site in the monomeric enzyme due to the small size and low degree of “buriedness” of the pocket formed by F – F′ and G – G′ loops. Therefore, if this peripheral site is involved in the mechanism of heterotropic activation of CYP3A4, the enzymes’ allosteric properties are expected to be determined by the degree of its oligomerization. A close interconnection between the allosteric properties of CYP3A4 and the degree of its oligomerization we demonstrated in our earlier study of the activating effect of ANF in model microsomes with variable surface density of CYP3A4 in the membrane [26].

The allosteric site’s location at the interface between two subunits of the P450 oligomer suggests a critical dependence of allosteric properties of the enzyme in HLM on the formation of mixed oligomers between multiple P450 species. Although in some pairs of CYP3A4 with interacting partners, this allosteric site may be retained, its affinity to the ligands and the effect of their binding on the functional properties of CYP3A4 must be considerably modified by heterooligomerization. Our observation that the addition of ANF attenuates the ability of CYP2E1-BODIPY to form heterooligomers with CYP3A4 may be interpreted as an indication of decreased ability of CYP2E1-CYP3A4 complexes to bind this allosteric effector, as compared to CYP3A4 homo-oligomers. This difference may explain the attenuation of the effect of ANF on CYP3A4 in CYP2E1-rich microsomes demonstrated in this study.

Certainly, unveiling the sources of the effect of the composition of the P450 ensemble on CYP3A4 allostery requires further rigorous investigation. At the current state of knowledge, we are far away from understanding the mechanisms of P450 hetero-association and their relationship to the allosteric properties of the P450 ensemble. Nevertheless, the model based on the hypothesis of positional heterogeneity and the involvement of the allosteric binding site located at the interface between two subunits of P450 oligomer provides a reasonable explanation for a remarkable connection between CYP3A4 allostery and the composition of the P450 pool revealed in our studies.

5. Concluding remarks

Our study with human liver microsomes obtained from alcoholic donors demonstrated that chronic alcohol exposure results in a multifold activation of CYP3A4, the enzyme responsible for the metabolism of a wide variety of drugs currently on the market. This activation and the associated attenuation of homo- and heterotropic cooperativity of CYP3A4 can be reproduced in model HLM preparations, where the content of CYP2E1 was elevated by incorporation of the purified enzyme. These results reveal a profound effect of the alcohol-induced increase in CYP2E1 content on the functional properties of CYP3A4 and suggest that CYP2E1-CYP3A4 interactions may play an important role in the effects of alcohol exposure on drug metabolism. CYP2E1-dependent activation of CYP3A4 demonstrated in this study is consistent with the general enhancement of drug metabolism observed in alcoholics [63–65]. In particular, it provides a credible explanation for the increased rate of metabolism of such drug substrates of CYP3A4 as diazepam and doxycyclin observed under the influence of chronic alcohol exposure in humans [66–68]. Furthermore, our findings compellingly demonstrate that the profile of human drug metabolism cannot be unambiguously determined from a simple superposition of the properties of the P450 species constituting the drug-metabolizing ensemble. This profile instead reflects a complex, “non-linear” combination of the functionalities of the individual enzymes, which is largely affected by their intermolecular interactions and the formation of heteromeric complexes.

Abbreviations:

BQ

7-benzyloxyquinoline

ANF

α-naphthoflavone (7,8-benzoflavone)

MRM

multiple reaction monitoring

SRM

single reaction monitoring

R_L/P

the ratio of the molar content of phospholipids to that of cytochromes P450 in microsomal membranes

CPR

NADPH-cytochrome P450 reductase

HLM

human liver microsomes

SS2E1, SS1A2, SS2B6, SS2C8, SS2C9, SS2C19, SS2D6 and SS3A4

insect cell microsomes (Supersomes™) containing recombinant human CYP2E1, CYP1A2, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, and CYP3A4, respectively