NADH:Cytochrome b5 Reductase and Cytochrome b5 Can Act as Sole Electron Donors to Human Cytochrome P450 1A1-Mediated Oxidation and DNA Adduct Formation by Benzo[a]pyrene

Benzo[a]pyrene (BaP) is a human carcinogen that covalently binds to DNA after activation by cytochrome P450 (P450). Here, we investigated whether NADH:cytochrome b5 reductase (CBR) in the presence of cytochrome b5 can act as sole electron donor to human P450 1A1 during BaP oxidation and replace the canonical NADPH:cytochrome P450 reductase (POR) system. We also studied the efficiencies of the coenzymes of these reductases, NADPH as a coenzyme of POR, and NADH as a coenzyme of CBR, to mediate BaP oxidation. Two systems containing human P450 1A1 were utilized: human recombinant P450 1A1 expressed with POR, CBR, epoxide hydrolase, and cytochrome b5 in Supersomes and human recombinant P450 1A1 reconstituted with POR and/or with CBR and cytochrome b5 in liposomes. BaP-9,10-dihydrodiol, BaP-7,8-dihydrodiol, BaP-1,6-dione, BaP-3,6-dione, BaP-9-ol, BaP-3-ol, a metabolite of unknown structure, and two BaP-DNA adducts were generated by the P450 1A1-Supersomes system, both in the presence of NADPH and in the presence of NADH. The major BaP-DNA adduct detected by 32P-postlabeling was characterized as 10-(deoxyguanosin-N2-yl)-7,8,9-trihydroxy-7,8,9,10-tetrahydro-BaP (assigned adduct 1), while the minor adduct is probably a guanine adduct derived from 9-hydroxy-BaP-4,5-epoxide (assigned adduct 2). BaP-3-ol as the major metabolite, BaP-9-ol, BaP-1,6-dione, BaP-3,6-dione, an unknown metabolite, and adduct 2 were observed in the system using P450 1A1 reconstituted with POR plus NADPH. When P450 1A1 was reconstituted with CBR and cytochrome b5 plus NADH, BaP-3-ol was the predominant metabolite too, and an adduct 2 was also generated. Our results demonstrate that the NADH/cytochrome b5/CBR system can act as the sole electron donor both for the first and second reduction of P450 1A1 during the oxidation of BaP in vitro. They suggest that NADH-dependent CBR can replace NADPH-dependent POR in the P450 1A1-catalyzed metabolism of BaP.

) is a polycyclic aromatic hydrocarbon (PAH) that has been classified as human carcinogen (Group 1) by the International Agency for Research on Cancer. 1 It is a pro-carcinogen requiring metabolic activation catalyzed by cytochrome P450 (P450) enzymes prior to reaction with DNA. 2 Among the P450s, P450 1A1 is the most important enzyme, in combination with microsomal epoxide hydrolase (mEH), involved in the metabolic activation of BaP to species forming DNA adducts. 2,3 First, P450 1A1 oxidizes BaP to an epoxide (i.e. BaP-7,8-epoxide) that is then converted to a dihydrodiol by mEH (i.e. BaP-7, 8-dihydrodiol). Further bioactivation by P450 1A1 leads to the ultimate reactive species, BaP-7,8-dihydrodiol-9,10-epoxide (BPDE) that can react with DNA, forming adducts preferentially at guanine residues ( Figure 1). The 10-(deoxyguanosin-N 2 -yl)-7,8,9-trihydroxy-7,8,9,10-tetrahydrobenzo[a]pyrene (dG-N 2 -BPDE) adduct is the major product of the reaction of BPDE with DNA in vitro and in vivo. [4][5][6][7][8][9][10] However, BaP is also oxidized to other metabolites such as other dihydrodiols, BaP-diones and further hydroxylated metabolites. 2,6,[11][12][13][14][15][16] Although most of these metabolites are detoxification products, BaP-9-ol is the precursor of 9-hydroxy-BaP-4,5-epoxide that can form another adduct with deoxyguanosine in DNA ( Figure 1). 7,8,10,15,[17][18][19] The P450 enzymes, including P450 1A1, are components of a mixed-function oxidase (MFO) system located in the membrane of endoplasmic reticulum (microsomes). This enzymatic system also contains other enzymes, the multidomain flavoprotein NADPH:cytochrome P450 oxidoreductase (POR) and cytochrome b 5 accompanied by its NADH:cytochrome b 5 reductase (CBR). Mammalian microsomal P450s function by catalyzing the insertion of one atom of molecular oxygen into a variety of xenobiotics, including BaP, while reducing the other atom to water, a reaction that requires two electrons. 20 The oxygen is activated in the active center of P450s by two electrons. It is generally accepted that POR with NADPH serves as donor of electrons for both reductions of P450 in the MFO reaction cycle. 20 However, the second electron may also be provided by CBR with cytochrome b 5 and NADH, but cytochrome b 5 seems to have also additional roles in the monooxygenase system. [20][21][22][23][24][25][26][27][28] Although POR is considered an essential constituent of the electron transport chain towards P450, 20 its exact role in the P450-mediated reaction cycle is still not clearly established. Recently we used two mouse models, in which the expression of POR has been permanently (the Hepatic P450 Reductase Null (HRN) line) or conditionally (the Reductase Conditional Null (RCN) line) deleted in hepatocytes leading to a lack of almost all hepatic POR activity. Despite this lack of POR the levels of the P450-mediated dG-N 2 -BPDE adducts in the livers of mice of both lines exposed to BaP were higher than in BaP-treated wild-type mice. 7,8,19 These findings suggested BaP activation in other liver cells other than hepatocytes (e.g. Kupffer or endothelial cells), 29 or bioactivation of BaP by non-P450 enzymes (e.g. prostaglandin H synthase and lipoxygenases), 30,31 or combinations of these mechanisms were operative. However, these phenomena might also indicate that another reductase such as microsomal CBR can contribute to the P450-mediated BaP oxidation in these animal models.
The latter possibility is supported by experiments with rat P450s indicating the involvement of cytochrome b 5 in the NADH-dependent hydroxylation of BaP in a reconstituted P450containing system. 32,33 Studies using cytochrome b 5 -knockout mouse lines, namely, HBN mice (Hepatic cytochrome b 5 Null) with a conditional hepatic deletion of cytochrome b 5 , and HBRN mice (Hepatic cytochrome b 5 /P450 Reductase Null), in which POR and cytochrome b 5 are deleted specifically in the liver, also indicate the involvement of CBR in the P450mediated metabolism of some other P450 substrates. 27,34,35 We have recently demonstrated that the NADH/cytochrome b 5 /CBR system is indeed able to function as the sole electron donor for both reduction steps of rat P450 1A1 during 6 oxidation of BaP in vitro. 15 Although this function of the NADH/cytochrome b 5

Preparation of Human Recombinant P450 1A1 and Rat Recombinant POR.
Human recombinant P450 1A1 (EC 1.14.14.1) and rat POR (EC 1.6.2.4) were prepared by heterologous expression in E. coli and purified to apparent homogeneity (i.e. as single bands on sodium dodecyl sulfate-polyacrylamide gel electrophoresis) as described recently. 36 The specific content of human recombinant CYP1A1 was 11.5 nmol/mg protein. Cytochrome b 5 reductase (CBR) (E.C. 1.6.2.2) was isolated from rat liver microsomes by a procedure described by Perkins and Duncan. 38 The specific activity of rat CBR measured as NADH-ferricyanide reductase was 49.2 µmol ferricyanide/min/mg protein. Cytochrome b 5 was isolated from rabbit liver microsomes as described. 39 Both proteins purified to apparent homogeneity 38,39 were utilized in the reconstitution experiments. and purified rat CBR (200 pmol in a ratio of 1:1), without or with cytochrome b 5 (in a ratio of P450 1A1 with reductase to cytochrome b 5 of 1:5) were added to the prepared dispersion and incubated at 20°C for 10 min. As shown in previous studies, 19,25,37,42,43 the enzymatic activity 9 of human P450 1A1 reconstituted with POR and cytochrome b 5 from several animal models was the same as that of the enzyme reconstituted with the human orthologs of these enzymes. substrates has been found previously. 7,8,19,25,28,44 The reaction was initiated by adding the NADPH or NADH.

Incubations to Study the Metabolism of BaP by Human
In separate experiments 100 nM pure human recombinant P450 1A1 reconstituted individually with other components of the MFO-system was used instead of P450 1A1 in Supersomes™. Negative control reactions lacked either P450 1A1, reductases or BaP. After incubation (37°C, 20 min), 5 µl of 1 mM phenacetin in methanol was added as an internal standard. BaP metabolism by the P450 1A1 systems has been shown to be linear up to 30 min of incubation. 15,16,19 BaP metabolites were extracted twice with ethyl acetate (2 × 1 ml), the solvent was evaporated to dryness, the residues were dissolved in 25 µl methanol and subsequently BaP metabolites were separated by HPLC as reported. 19,45 BaP metabolite peaks were analyzed by HPLC by comparison with metabolite standards whose structures were determined previously by NMR and/or mass spectrometry. 19 Determination of BaP-DNA Adduct Formation by 32 P-postlabeling. Incubation mixtures used to assess DNA adduct formation by BaP activated with all enzymatic systems containing human P450 1A1 consisted of 50 mM potassium phosphate buffer (pH 7.4), 1 mM NADPH or NADH, 100 nM human recombinant P450 1A1 plus other enzymes as indicated in the figures (or as described above), 0.1 mM BaP (dissolved in 7.5 µl DMSO), and 0.5 mg of calf thymus DNA in a final volume of 0.75 ml as described previously. 19 The reaction was initiated by adding 0.1 mM BaP and incubations were carried out at 37°C for 60 min. BaP-DNA adduct formation has been shown to be linear up to 90 min. 7,19 Control incubations were carried out without P450 1A1, or without reductases, or without cytochrome b 5 , or without NADPH (or NADH), or without DNA, or without BaP. After the incubation, DNA was isolated from the residual water phase by standard phenol/chloroform extraction. DNA adduct formation was analyzed using the nuclease P1 version of the 32 P-postlabeling technique. 7,19 Resolution of the adducts by thin-layer chromatography (TLC) using polyethyleniminecellulose plates (Macherey and Nagel, Düren, Germany) was carried out as described. 7,19,46 DNA adduct levels (RAL, relative adduct labelling) were calculated as described. 47 Statistical Analyses. Statistical analyses were carried out on the means ± standard deviations of three parallel experiments with Student's t-test (UNISTAT Statistics Software v6, Unistat Ltd., Highgate, London N6 5UQ, UK) and p < 0.05 was considered significant.
Statistical association between amounts of BaP metabolites formed by oxidation of BaP by the P450 reconstituted systems containing POR or CBR and levels of BaP-DNA adduct 2 formed by the same systems were determined by the Spearman correlation coefficients using version 6.12 Statistical Analysis System software. Spearman correlation coefficients were based on a sample size of 6. All Ps are two-tailed and considered significant at the 0.05 level.

Oxidation of BaP by Human P450 1A1 Expressed in Supersomes™ and Pure Human
Recombinant P450 1A1 Reconstituted with POR or CBR. In order to evaluate the role of POR and CBR in the reduction of human P450 1A1 during BaP oxidation, two enzymatic systems containing this human P450 were utilized. The first system used Supersomes™ containing human recombinant P450 1A1 expressed with POR, CBR, mEH and/or cytochrome b 5 . In the second system, pure human recombinant P450 1A1 was reconstituted in liposomes with either pure POR or CBR. The latter enzymatic system was utilized with or without cytochrome b 5 to examine the function of both reductases as electron donors to P450 1A1 during BaP metabolism with NADPH (cofactor of POR) or NADH (cofactor of CBR).
The BaP metabolites formed by human P450 1A1 in these enzyme systems were analyzed by HPLC ( Figure 2A Figure 2B. In addition, a metabolite of unknown structure (Mx) was detected. Essentially no BaP metabolites were found when both NADPH and NADH were omitted from the incubation mixtures containing the P450 1A1-Supersomes™ ( Figure 2C).
NADH was less effective than NADPH to act as electron donor to human P450 1A1 in Supersomes™ ( Figure 2C). Addition of cytochrome b 5 to the incubation mixtures led to an increase in P450 1A1-mediated BaP oxidation both in the presence of NADPH and in the presence of NADH ( Figure 2C).
In the second enzymatic system, where pure human recombinant P450 1A1 was reconstituted with POR with or without cytochrome b 5 in liposomes, this P450 enzyme was also able to oxidize BaP in the presence of NADPH (Figure 3). Only five BaP metabolites were formed, which were BaP-3-ol (M7) as the major metabolite, BaP-9-ol (M6), BaP-1,6dione (M4), BaP-3,6-dione (M5) and metabolite Mx ( Figure S1). Because mEH was absent from the system, no dihydrodiols were formed. Addition of cytochrome b 5 to this P450 1A1 system increased formation of all BaP metabolites, in particular BaP-3-ol (Figure 3 and Figure S1). In contrast, NADH did not lead to significant BaP oxidation by P450 1A1 reconstituted with POR ( Figure 3 and Figure S1), which confirms that NADH functions as a coenzyme of POR at very slow rate that is negligible relative to NADPH, as we showed recently with cytochrome c as a substrate for POR. 15 No BaP metabolites were found when both NADPH and NADH were omitted from the incubation mixtures containing the human recombinant P450 1A1 reconstituted with POR or CBR (Figure 3). In the enzymatic system where human P450 1A1 was reconstituted with CBR and cytochrome b 5 in liposomes, BaP was predominantly oxidized to BaP-3-ol (M7) and to a lower extent to BaP-9-ol (M6), BaP-1,6-dione (M4), BaP-3,6-dione (M5) and a metabolite Mx ( Figure S1). Cytochrome b 5 as a substrate of CBR was necessary for BaP oxidation in the system of human P450 1A1 reconstituted with CBR with NADH as cofactor. Without this protein, essentially no BaP oxidation was detectable by P450 1A1 reconstituted with CBR ( Figure 3 and Figure S1). Addition of POR to the reconstituted system of human P450 1A1 with CBR and cytochrome b 5 did not change BaP metabolite levels ( Figure 3 and Figure S1). BaP was activated with human P450 1A1 in Supersomes™, in the presence of either NADPH or NADH. No such BaP adducts were found when both NADPH and NADH were omitted from the incubation mixtures containing the P450 1A1-Supersomes™ (Figure 4). Adduct 1 was the BaP-DNA adduct predominantly formed in this system (Figure 4). Comparison with previous 32 P-postlabeling analyses 10,19 showed that adduct 1 is the dG-N 2 -BPDE adduct. 10,19 The other adduct, which was formed in this enzymatic system as only a minor product, if detectable at all (e.g. there was no formation of this adduct in Supersomes™ in the presence of NADPH), has similar chromatographic properties by TLC to a guanine adduct derived from reaction with 9-hydroxy-BaP-4,5-epoxide (see adduct spot 2 in insert of Figure 1 and  Figures 5B and 6). Because of the absence of mEH, no adduct 1 (i.e. dG-N 2 -BPDE) was formed. Addition of cytochrome b 5 to this P450 1A1 system decreased the levels of adduct 2, but this decrease was not statistically significant ( Figure 6). NADH was essentially ineffective to mediate the activation of BaP by P450 1A1 reconstituted with POR ( Figures 5B and 6), which again indicates that NADH functions as a coenzyme of POR at very slow rate that is negligible relative to NADPH.
Human P450 1A1 reconstituted with CBR and cytochrome b 5 also generated adduct 2.
The presence of cytochrome b 5 as a substrate of CBR and NADH as its cofactor was essential for this adduct formation; NADPH, the cofactor of POR, was ineffective ( Figures 5B and   6).No BaP-DNA adduct 2 was found when both NADPH and NADH were omitted from the incubation mixtures containing human recombinant P450 1A1 reconstituted with POR or 14 CBR ( Figure 6). Levels of adduct 2 formed by human P450 1A1 in the presence of the NADH/cytochrome b 5 /CBR system were lower than those formed in the system of human P450 1A1 containing POR and NADPH (Figures 5 and 6). These results corresponded to lower levels of BaP-9-ol (a precursor of 9-hydroxy-BaP 4,5-epoxide generating this adduct) and other BaP metabolites formed in this system (see Figure S1). Significant correlations were found between levels of adduct 2 and BaP-9-ol (r = 0.943, P < 0.001, Spearman's correlations) or BaP-3-ol (r = 0.943, P < 0.001), whereas no such correlations were found between levels of this adduct and BaP-1,6-dione, BaP-3,6-dione or a metabolite Mx (r < 0.5).

DISCUSSION
The metabolism of BaP has been extensively studied over the past decades 2 and various studies have examined the role of P450 enzymes, particularly P450 1A1 of several species, to metabolize this carcinogen. 2,6,7,8,15,16,19,28,44 However, the mechanism of the reaction cycle of BaP oxidation catalyzed by P450 1A1, particularly the roles of POR and CBR as electron donors to P450 1A1, has not yet been fully resolved. 7,8,15,19 Enigmatic results have been found in two mouse models where deletion of POR in hepatocytes did not lead to an expected decrease in BaP-DNA adduct formation in liver in vivo, but instead to higher BaP-DNA binding. 7,8,19 We also showed that in livers of HRN mice P450 1A1, cytochrome b 5 and mEH can effectively activate BaP to DNA binding species, even in the presence of very low amounts of POR. 19 Because this feature has biological significance, studying the role of the enzymes reducing P450 1A1 is important to better understand the mechanism(s) involved in BaP metabolism. Recently we demonstrated that the NADH/cytochrome b 5 /CBR system is able to function as the sole donor of electrons for both reduction steps of rat P450 1A1 during BaP oxidation in vitro. 15 This finding indicates that CBR as an NADH-dependent reductase might substitute POR in the P450 1A1-mediated BaP metabolism and might help to explain our enigmatic results in the POR-knockout mouse models. 7,8,19 However, the question whether this novel function of the NADH/cytochrome b 5 /CBR system as electron donor to rat P450 1A1 in BaP metabolism represents a general feature for the P450 1A1 reaction cycle in other species including humans remained to be answered. Therefore, the primary aim of this study was to determine whether the NADH/cytochrome b 5 /CBR system can be the exclusive donor Using the first system, human P450 1A1-Supersomes™, we proved that NADH acts as electron donor for both reductions of the P450 1A1 human orthologue in BaP oxidation independent of NADPH and POR if CBR and cytochrome b 5 is present. This conclusion is supported by the fact that NADH functions as a poor coenzyme of POR when cytochrome c is used as its substrate. 15 Our results therefore confirm the novel feature on the mechanism of the catalytic cycle of human P450 1A1 during BaP oxidation previously described for rat P450 1A1. We demonstrated that the reaction cycle of BaP oxidation catalyzed by human P450 1A1 can proceed by ways that differ from the generally accepted mechanism, where the first reduction of P450 is considered to be catalyzed by POR without cytochrome b 5 . 20,48-52 Considering the redox potentials of cytochrome b 5 (+20 mV) 53,54 and ferric substrate-bound P450 (-237 mV) 27,53 it is thermodynamically impossible for cytochrome b 5 to provide the first electron in the P450 catalytic cycle 55 . Given that the redox potential of oxyferrous P450 is also approximately +20 mV, it is feasible that cytochrome b 5 can supply the second electron into the catalytic cycle. 27,53,55 However, based on the redox potential of CBR determined under the anaerobic conditions (-265 mV) 56 it could provide the first electron in the P450 catalytic cycle. 27,57 Moreover, considering the effect of Le Chatelier's principle given, the reduced P450 will rapidly bind dioxygen under aerobic conditions of the experiments. Indeed, in the present study we demonstrate that the first reduction of P450 1A1 which previously had been considered to be mediated exclusively by the NADPH/POR system can be substituted by the NADH/cytochrome b 5 /CBR system. This reaction was found not only in Supersomes™ containing human recombinant P450 1A1, but, even more importantly, in the system of pure human P450 1A1 reconstituted with CBR and cytochrome b 5 . Such NADH/cytochrome b 5 /CBR-mediated activity of the human P450 1A1 systems was proven by the formation of up to seven BaP metabolites [BaP-9,10-dihydrodiol (M1), BaP-7,8-dihydrodiol (M3), BaP-1,6dione (M4), BaP-3,6-dione (M5), BaP-9-ol (M6), BaP-3-ol (M7) and a metabolite of unknown structure (Mx)]. In addition, these systems generated two BaP-DNA adducts, the dG-N 2 -BPDE adduct (adduct 1) and/or a guanine adduct derived from the reaction with 9hydroxy-BaP-4,5-epoxide (adduct 2). The BaP metabolite profiles formed by P450 1A1 with the NADH/cytochrome b 5 /CBR system were the same as in the system where NADPH and POR were used as electron donors. BaP-4,5-dihydrodiol (M2), which was previously found to be formed by rat P450 1A1 15,19 was not generated by the human P450 1A1 orthologue ( Figure   2). This finding is in line with previous findings showing that this BaP metabolite was not formed in human bronchoalveolar H358 cells expressing P450 1A1 after BaP exposure. 13 Interestingly, essentially no differences in the levels of dG-N 2 -BPDE adducts were seen when using either NADPH or NADH as cofactors in the P450 1A1-Supersomes™ system (see Figure 4). However, it should be noted that the oxidation reaction of BaP to its metabolites was catalyzed less efficiently by NADH than by NADPH in this system (see Figure 2). One In summary, this study demonstrates for the first time that NADH, CBR and cytochrome b 5 can act as sole electron donors for both the first and second reduction of human P450 1A1 during the oxidative metabolism of BaP and formation of BaP-DNA adducts in vitro. These findings confirm our results of a recent study where rat P450 1A1 was utilized to study BaP metabolism. 15 However, although the role of the NADH/cytochrome b 5 /CBR system in both reductions of rat 15 and human P450 1A1 present work is proven in our in vitro studies, further investigations are needed in the future. In this context, the mechanism of both the first and second reduction of P450 1A1 remains to be examined in detail. In addition, as shown by Guengerich and coworkers for the P450 3A4-mediated 6-hydroxylation of testosterone 57 , the question whether NADH/cytochrome b 5 /CBR system might also reduce other P450 enzymes needs to be addressed. This might be the case in rat hepatic microsomes where various P450s were induced by their specific inducers; in these microsomes, BaP was oxidized not only in the presence of NADPH, but also in the presence of NADH. 15 Another crucial question that remains to be addressed relates to the impact of the NADH/cytochrome b 5 /CBR system on BaP metabolism in vivo. Recent in vivo experiments using cytochrome b 5 -knockout mouse lines (i.e. HBN and HBRN) provide evidence that in the absence of POR, cytochrome b 5 /CBR are capable of supplying electrons for P450 catalytic function (i.e. in metabolism of the P450 3A substrate midazolam). 27 It is anticipated that these mouse lines will help to elucidate the different mechanisms of P450-catalyzed BaP biotransformation in vivo.

Notes
The authors declare no competing financial interest.

Description of the Supporting Information material
This material is available free of charge via the Internet at http://pubs.acs.org.