Humans reportedly have 57 intact cytochrome P450 (P450) genes and mice have 93 P450 genes. These proteins catalyze the oxidative metabolism of a large number of lipophilic compounds. A number of the P450 forms have orthologs in different species and are thought to catalyze functionally conserved reactions. Members of cytochrome P450 families 1, 2, and 3 can oxidize a broad range of xenobiotics and are generally considered as xenobiotic-metabolizing enzymes. Orthologous forms of family 1, CYP1A1, CYP1A2, and CYP1B1, are found throughout the animal phyla, from the bony fish to human (http://drnelson.utmem.edu/CytochromeP450.html), their conservation during evolution suggesting that they might have roles other than just xenobiotic metabolism (Stoilov et al., 2001; Schenkman et al., 2003). Consequently, it was of interest to us to compare the abilities of two CYP1B1 orthologs to metabolize substrates of endogenous origin. The compounds investigated were lipid (arachidonate) and retinoids (retinol, retinal, and retinoic acid).

Retinol (vitamin A) is an important compound in the body and is converted to all-trans-retinoic acid (RA), a morphogen necessary for early stages of development in organizing the spatial orientation of the embryo into anterior and posterior segments (Maden, 1999; Swindell et al., 1999). It regulates gene expression by acting as a ligand for nuclear retinoid receptors (Chambon, 1996). Synthesis of RA is mediated by retinol dehydrogenase (ADH-1) plus retinal dehydrogenase-2, and an ideal balance in tissue distribution is maintained by a retinoate-inducible, specific RA-degrading enzyme, CYP26 (Fujii et al., 1997; Swindell et al., 1999). A number of other forms of cytochrome P450 have been reported to be capable of retinoid metabolism (Vanden Bossche et al., 1988; Leo et al., 1989; Roberts et al., 1992). Among human P450 families, CYP1A, CYP1B, and CYP3A subfamily members had major abilities to convert all-trans-retinol (ROL) to all-trans-retinal (RAL); family 1 members (CYP1A1, CYP1A2, and CYP1B1) also showed considerable ability to oxidize RAL to RA (Chen et al., 2000; Zhang et al., 2000).

A large number of biologically active eicosanoid metabolites are produced from arachidonic acid via pathways including epoxidation, allylic oxidation, and ω-hydroxylation, by a number of enzymes, including cytochrome P450 monooxygenases, cyclooxygenases, and lipoxygenases (Capdevila and Falck, 2001). The different hydroxyeicosatetraenoic fatty acids (HETEs) and epoxyeicosatrienoic fatty acids (EETs) interact with a wide range of transduction mediators (Wen et al., 2000; O'Flaherty et al., 2002; Kalyankrishna and Malik, 2003; Shankaranarayanan and Nigam, 2003). The HETEs can be subdivided into ω-terminal (ω- to ω-4-hydroxy), midchain, and bisallylic hydroxylations (Brash et al., 1995). Family 2 P450s metabolize arachidonate primarily to EETs (Daikh et al., 1994; Wu et al., 1996; Keeney et al., 1998; Luo et al., 1998). CYP2E1 differs in that its metabolism of arachidonic acid generates 90% ω- and (ω-1)-HETEs (Rifkind et al., 1995). Family 3 cytochromes P450 have been reported by different laboratories to primarily produce ω- and (ω-1)-HETEs (Rifkind et al., 1995) or midchain HETEs (Bylund et al., 1998). Family 4 cytochromes P450 are considered to be ω- to (ω-2)-HETE producers (Powell et al., 1998; Bylund et al., 2001; Hashizume et al., 2001). Family 1 P450 metabolism of arachidonate has received less attention than families 2 to 4, although the metabolism of this substrate by CYP1A2 has been shown to generate EETs as the major products (Rifkind et al., 1995). Although a considerable number of studies have been carried out on metabolic abilities of human P450s, much less is known about the mouse orthologs. Cyp1b1 was originally purified from mouse embryo fibroblasts and shown to be the major enzyme responsible for oxidative metabolism of the carcinogen, 7,12-dimethylbenz(α)anthracene (DMBA) (Pottenger et al., 1991). A comparison of mouse Cyp1b1 with the human ortholog, CYP1B1, revealed that the latter had only 10% of the specific activity in the metabolism of DMBA (Savas et al., 1997). In the present study, we confirm the greater ability of the mouse enzyme to metabolize DMBA and, in a comparison of the abilities of the two orthologs, show that the efficiency of the mouse ortholog to metabolize arachidonic acid and retinoids is far lower than that of the human enzyme.

Materials and Methods

Chemicals. Retinol, retinal, retinoic acid, and arachidonic acid were purchased from Sigma-Aldrich (St. Louis. MO). ¹⁴C-labeled substrates were purchased from PerkinElmer Life and Analytical Sciences (Boston, MA), and ³H-DMBA was obtained from Amersham Biosciences Inc. (Piscataway, NJ). Complete Protease Inhibitor Tablets were obtained from Roche Diagnostics (Indianapolis, IN). The QuikChange Site-Directed Mutagenesis Kit was obtained from Stratagene (La Jolla, CA). Human CYP1A1 and CYP1B1 Supersomes were obtained from BD Gentest (Woburn, MA). Arachidonate metabolite standards 20-HETE, 15-HETE, 11-HETE, 9-HETE, 5-HETE, 8-HETE, and 12-HETE; and 14,15-EET, 11,12-EET, 8,9-EET, and 5,6-EET were obtained from Cayman Chemical (Ann Arbor, MI).

Expression and Purification of CYP1B1 and Cyp1b1. Human CYP1B1 was expressed and purified as described earlier (Jansson et al., 2001). The mouse Cyp1b1 transcript was prepared from a cDNA library using primers with modification at the 5′ end containing the NdeI site (5′-CGA/GGT/CAT/ATG/CTT/TCT/CCA/AAT/GAT/CCA/CAG/CAG/CTG/AGC), cloned into pCR2.1-TOPO plasmid (Invitrogen, Carlsbad, CA). The amino-terminal sequences of CYP1B1 were modified by deletion of amino acid residues 2 to 4 of the wild-type protein to facilitate expression in Escherichia coli. (Shimada et al., 1998; Jansson et al., 2000). The amino acid residues 2 to 8 of Cyp1b1 were replaced by valine for similar reasons. The Cyp1b1 sequence had an additional NdeI site at nucleotide 1137, which was modified (CCA→ CCG, Pro→ Pro) using an in vitro mutagenesis kit (Stratagene), resulting in no additional alteration in the final protein sequence. The sequence of the TOPO-TA construct was verified, and the mCyp1b1 insert was cleaved and liganded between NdeI and XbaI sites of the pCWori⁺ vector, as described for CYP1B1 (Jansson et al., 2000). Transformation of E. coli DH5α and expression were as described for CYP1B1 (Jansson et al., 2000). The enzyme was purified using DEAE-Sephacel, S-Sepharose, and hydroxylapatite column chromatography. The presence of cytochrome P450 was detected and quantified from its reduced minus oxidized CO spectrum, using the extinction coefficient, 91 mM^-1 cm^-1 (Schenkman and Jansson, 1998). Recombinant human CYP1B1, coexpressed with NADPH-cytochrome P450 reductase (Supersomes; BD Biosciences Discovery Labware, Bedford, MA) was used for comparison with the CYP1B1 expressed in E. coli and was found not to differ in metabolite production from our expressed form.

Expression and Purification of hCYP1A2. The CYP1A2 expression system was the generous gift of Dr. R. W. Estabrook and was expressed and purified as described (Fisher et al., 1992).

Incorporation of Reductase and P450. The rabbit NADPH-cytochrome P450 reductase was expressed in E. coli and purified by 2′,5′-ADP Sepharose and hydroxylapatite column chromatography, as previously described in detail (Jansson et al., 2001). In the reconstituted system, the reductase and P450 were incorporated into DLPC at 1:1:160 molar ratios.

Enzyme Assays.DMBA. For assay of DMBA metabolism, the reconstituted system (150 nM CYP1B1) was incubated with 25 to 200 μM DMBA, 5 μCi of ³H-DMBA, 3 mM MgCl₂, 0.01% cholate, 50 mM phosphate buffer (pH 7.4) at 37°C for 1 min before and, subsequently, for 15 min after the addition of 0.5 mM NADPH. DMBA and its metabolites were extracted twice with 4 ml of solvent mixture (ethyl acetate and acetone in a 2:1 ratio containing 1 mM dithiothreitol). Extracts were pooled, dried under vacuum, dissolved in 30 μl of methanol, and injected onto a Beckman Coulter C18 reverse phase column (Beckman Coulter, Fullerton, CA). HPLC was run using 50% methanol (0–1 min), followed by a gradient of 50 to 100% methanol (1–10 min) and an isocratic elution with 100% methanol (10–45 min). Eluted fractions were collected and counted in an Amersham Biosciences scintillation counter.

Retinoids. The retinoid metabolism assays containing CYP1B1 orthologs (150 nM) were performed in 1-ml volumes at 37°C. The reconstituted system was incubated with 35 μM all-trans-retinol or all-trans-retinal (except when substrate dependence experiments were performed), 10 mM MgCl₂, 1 mM ascorbic acid, 0.01% sodium cholate, 50 mM phosphate buffer (pH 7.4) for 1 min and, subsequently, for 15 min after the addition of 0.5 mM NADPH. The reaction was terminated with the addition of 4 ml of ethyl acetate (0.001% butylated hydroxytoluene). The retinoid metabolites were extracted three times with ethyl acetate. Extracts were pooled, dried under a stream of nitrogen, and dissolved in 30 μl of acetonitrile (ACN); then, 25 μl were injected onto a Beckman Coulter C18 reverse phase column, and HPLC was run using solvent A (ACN/water/acetic acid, 49.75:49.75:0.5) and solvent B (ACN/water/acetic acid, 90:10:0.04), each containing 10 mM ammonium acetate. The elution gradient conditions were 0 to 10 min (20% to 65% solvent B), 10 to 30 min (65% to 100% solvent B) and were subsequently maintained (30–45 min) isocratically at 100% solvent B. The metabolites were identified by the retention time of standards and the spectral features (λ_max). Substrate consumption never exceeded 2%. For determination of kinetic constants in metabolism of substrates, the data were subjected to linear regression analysis using SigmaPlot version 7 (SPSS Inc., Chicago, IL). K_cat values (turnover numbers) for the enzymes are provided in the tables. Catalytic efficiencies (K_cat values divided by the K_m values) of the enzymes are also provided. For comparison we provide the catalytic efficiency ratios (CERs), the ratios of the efficiencies of the orthologs being examined.

Arachidonate. For arachidonic acid metabolism, the reconstituted system containing CYP1A2, CYP1B1, or Cyp1b1 (150 nM), or CYP1A1 Supersomes (20 nM), was incubated in a final volume of 1 ml containing 0.01% sodium cholate, 10 mM MgCl₂, 150 mM KCl, 50 mM Tris-Cl (pH 7.5), 5 mM glucose 6-phosphate, 1 U glucose-6-phosphate dehydrogenase, and 50 μM arachidonic acid (0.1 μCi of ¹⁴C-arachidonate) (except when substrate dependence experiments were performed) for 1 min at 37°C and, subsequently, for 30 min after the addition of 0.5 mM NADPH. The arachidonic acid (in ethanol) was put into assay tubes first, and after dispersal over the bottom of the tube, the ethanol was evaporated away under a stream of dry nitrogen. The reaction was terminated with the addition of 20 μl of 6.7% formic acid and 4 ml of ethyl acetate (0.001% butylated hydroxytoluene). The arachidonic acid metabolites were extracted as described above for retinoids. Substrate consumption never exceeded 2%. The HPLC system was run using mobile phase solvent A (ACN/water/acetic acid, 20:80:0.1) and solvent B (ACN/water/acetic acid, 99.9:0:0.1), elution gradient 38 to 80% B (0–41 min), 80 to 100% solvent B (41–43 min), isocratically at 100% solvent B (43–47 min). The assay system uses separate HETEs and EETs well, but is less effective for the identification of the bisallylic HETEs (Brash et al., 1995). Radiolabeled fractions were collected at 0.2- to 0.5-min intervals for analysis on an Amersham Biosciences RackBeta LSC scintillation counter. For determination of kinetic constants in metabolism of the substrate, the data were subjected to linear regression analysis using SigmaPlot version 7.

Results

In a seminal observation, an unusual pattern of DMBA metabolism in a mouse fibroblast cell was linked to a new form of cytochrome P450, Cyp1b1 (Pottenger et al., 1991). The mouse enzyme was suggested to be more active than the human ortholog, CYP1B1 (Savas et al., 1997). In the present study, the kinetics of DMBA metabolism were examined using our E. coli-expressed CYP1B1 orthologs (Table 1). It was observed that the mouse ortholog has a considerably higher specific activity than the human. Cyp1b1 had a k_cat almost 3-fold higher than that of CYP1B1. In addition, it had an apparent K_m [K_m(app)] for DMBA about 60% of that of the human ortholog, resulting in a CER of the human to mouse ortholog of 0.23. A comparison of the primary structure of the orthologs revealed a sequence identity of 81%. The proportion of identical residues in the substrate recognition sites (SRSs) of the proteins was found to be 100% for SRS-2 and SRS-4; 90% for SRS-1, SRS-5, and SRS-6; and only 80% for SRS-3 (Table 2). In contrast, in comparison with CYP1B1, the sequence identity for CYP1A1 and CYP1A2 did not exceed 50% and 40% in SRS 1–4, respectively, whereas greater identities were seen in SRS-5.

CYP1B1 is capable of oxidative metabolism of ROL to RAL, and oxidation of RAL to RA (Chen et al., 2000; Zhang et al., 2000). Figure 1A shows an HPLC tracing of metabolites generated from 35 μM ROL in a reconstituted system containing CYP1B1. The upper tracing shows the pattern of metabolites in the absence of NADPH and the lower tracing shows the pattern in the presence of NADPH, a necessary cofactor for the monooxygenase activity. RAL was the principal metabolite produced. Other small metabolite peaks were also observed, but these have, as yet, not been characterized. Cyp1b1 was similarly incubated with ROL under the same conditions (Fig. 1B), and a lower rate of RAL production was obtained. The metabolism of RAL by CYP1B1 was considerably faster than its metabolism of ROL (Fig. 2A), and, as in the oxidation of ROL, a number of minor metabolites were observed in addition to RA. Comparable levels of Cyp1b1 again produced a much lower amount of products (Fig. 2B). The lower activity of the mouse enzyme toward metabolism of all-trans retinol and all-trans-retinal might reflect an intrinsically lower activity of the mouse enzyme toward this substrate. Alternatively, the substrate dependence of the mouse enzyme could have been altered over evolution. The kinetic constants for the enzyme were therefore determined (Table 1). The k_cat for the oxidation of ROL by Cyp1b1 was almost twice that of CYP1B1, but the CYP1B1 to Cyp1b1 CER was 13.5, reflecting the more than 20-fold higher K_m(app) of Cyp1b1 for ROL. The rates of oxidation of RAL by both orthologs were considerably greater than their rates of oxidation of ROL. Furthermore, the K_m(app) of Cyp1b1 for substrate was much higher, 18-fold, resulting in a 27-fold greater CER. Neither human nor mouse CYP1B1 ortholog was able to generate metabolites from RA, when incubated under conditions as above with retinol or retinal (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Reverse-phase HPLC profile of all-trans-retinol metabolism by CYP1B1 (A) and Cyp1b1 (B) monitored at 380 nm. The reconstituted system consisting of a 0.15 μM concentration, each, of cytochrome P450 and NADPH-cytochrome P450 reductase, 24 μM DLPC in 50 mM sodium phosphate buffer (pH 7.4), 10 mM MgCl₂, 0.01% sodium cholate, 1 mM ascorbic acid, and 35 μM ROL was incubated for 15 min at 37°C in the absence (upper tracing) or presence (lower tracing) of 0.5 mM NADPH.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Reverse-phase HPLC profile of all-trans-retinal metabolism by CYP1B1 (A) and Cyp1b1 (B) monitored at 350 nm. The reconstituted system consisting of a 0.15 μM concentration, each, of cytochrome P450 and NADPH-cytochrome P450 reductase, 24 μM DLPC in 50 mM sodium phosphate buffer (pH 7.4), 10 mM MgCl₂, 0.01% sodium cholate, 1 mM ascorbic acid, and 35 μM RAL was incubated for 15 min at 37°C in the absence (upper tracing) or presence (lower tracing) of 0.5 mM NADPH.

Metabolism of arachidonic acid can generate a wide variety of compounds with transcriptional regulatory activity, e.g., HETEs and EETs. Figure 3 is a sketch showing the many isomeric sites of metabolism of arachidonic acid by different forms of cytochrome P450 and the products (EETs and HETEs) formed. The HETEs may be classified by regions of oxy-metabolites formed: the midchain metabolites (hydroxyl group at position 5, 8, 9, 11, 12, or 15), the bisallylic HETEs (with a hydroxyl group at position 7, 10, or 13), and the ω-terminal metabolites (hydroxy group at position 16, 17, 18, 19, or 20). Figure 4 shows the metabolites generated from arachidonic acid by CYP1B1 (Fig. 4, top), and no apparent metabolite production by Cyp1b1 (Fig. 4, bottom) at 50 μM arachidonate in the reconstituted system. The ω-terminal HETEs eluted first in the reverse-phase HPLC. The peak at 15 min was identified as 20-HETE, based upon retention time and comparison with authentic standard. However, as demonstrated by Rifkind et al. (1995), using the same assay conditions, this peak may also contain 19-HETE. The peak that elutes at 17.5 min has been shown to contain 16-HETE, 17-HETE, and 18-HETE under similar HPLC conditions (Falck et al., 1990). Using authentic standards, midchain HETEs were found to follow with peaks at 20 min (15-HETE), 21 min (11-HETE), 22 min (mixture of 12-HETE and 8-HETE), and 23.5 min (5-HETE). Authentic EET standards demonstrated that EET peaks appear after elution of the midchain HETEs in the order of 14,15-EET (26 min), 11,12-EET (27.3 min), 8,9-EET (27.9 min), and 5,6-EET (28.5 min). Unmetabolized arachidonic acid (not shown) eluted at 43 min.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Arachidonic acid metabolites generated by different cytochromes P450.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Reverse-phase HPLC profile of arachidonic acid metabolism by CYP1B1 (top) and Cyp1b1 (bottom). The reconstituted system consisting of a 0.15 μM concentration, each, of cytochrome P450 and NADPH-cytochrome P450 reductase, 24 μM DLPC in 50 mM Tris-Cl (pH 7.4), 10 mM MgCl₂, 150 mM KCl, 0.01% sodium cholate, 50 μM arachidonic acid (0.1 μCi of ¹⁴C-arachidonic acid), and regenerating NADPH system (5 mM glucose 6-phosphate and 1 unit/ml glucose-6-phosphate dehydrogenase) was incubated for 30 min at 37°C in the absence (open circles) or presence (closed circles) of 0.5 mM NADPH.

It has been known for some time that a number of different forms of cytochrome P450 are able to metabolize arachidonate. In the present study, we examined the intrinsic activity of the human CYP1 enzymes and their isomeric specificities of metabolite formation. The apparent kinetic constants of arachidonate metabolism of these forms, CYP1B1, CYP1A1, and CYP1A2, are compared in Table 3. CYP1A1 had the highest specific activity, more than 50% higher than that of CYP1A2 and more than 24-fold higher than that of CYP1B1. Whereas CYP1A1 and CYP1B1 have similar K_m(app) values for arachidonate, CYP1A2 required higher substrate levels for maximum activity. The catalytic efficiency of CYP1A1 was more than 20-fold higher than that of CYP1B1, reflecting their different k_cat values, and 4-fold greater than CYP1A2, reflecting their different K_m values. The explanation for lack of observable metabolism of arachidonate by Cyp1b1 (Fig. 4, bottom) under standard assay conditions (Rifkind et al., 1995) was forthcoming when the substrate dependence of this enzyme was examined (Table 3). Cyp1b1 proved capable of producing metabolites when a higher range of arachidonate was used; the K_m for arachidonate metabolism by Cyp1b1 was 0.5 mM, a value 17-fold higher than the K_m for this substrate with CYP1B1. Furthermore, the k_cat of Cyp1b1 was about 25% and its catalytic efficiency was about 1.5% of the human ortholog.

Clear differences exist between the isomeric specificities of these family 1 enzymes toward arachidonic acid: CYP1A2 showed a high specificity for epoxide production (EETs) approximating 52% of its metabolites. This is about twice the proportion formed by the much less active CYP1B1. CYP1A1, the most active catalyst of arachidonate oxidation (Table 3), showed specificity for the ω-terminal region of the substrate, generating almost 75% of its metabolites as ω-terminal region HETEs. It generated the lowest proportion of epoxymetabolites among the family 1 monooxygenases, about 5%. CYP1B1 preferentially produced midchain HETEs, almost 55% of its total metabolites (Table 3). This proportion, however, is equivalent to a specific rate of about 10% of the rate of midchain HETE formation by CYP1A1, due to the much greater k_cat of the latter. In contrast to CYP1B1, the major proportion of metabolites by the mouse ortholog, Cyp1b1, was EETs (50%), with a fairly high proportion of midchain HETEs (37%).

Discussion

Orthologous forms of CYP1B1 are found in a number of vertebrates, from bony fish to mouse to human, suggesting, perhaps, an ability to use compounds involved in intermediary metabolism of endobiotics. In earlier studies, a form of mouse cytochrome P450, Cyp1b1, was shown to metabolize the carcinogen, DMBA (Pottenger et al., 1991). The metabolism of this xenobiotic could also be accomplished by its human ortholog, CYP1B1, but at a slower rate (Savas et al., 1997). In the present study, the kinetics of DMBA metabolism by both orthologs were examined, and a human to mouse CER of 0.23 was found. The turnover number for the mouse enzyme was almost 3-fold greater than that of the human enzyme, when examined using the same reconstituted system. Since human CYP1B1 has been reported to metabolize a number of endogenous compounds (endobiotics), we compared the metabolism of several lipophilic endobiotic substrates, arachidonic acid and retinoids, by the two orthologs.

Cyp26 has been shown to be especially effective in maintaining the homeostasis of RA, functioning as a retinoate-inducible, specific RA-degrading enzyme (Fujii et al., 1997). Prior studies revealed that RA was converted to polar metabolites by skin microsomes, and that the rate was increased by cytochrome P450 inducers and decreased by cytochrome P450 inhibitors (Vanden Bossche et al., 1988). A number of purified forms of human cytochrome P450 were subsequently shown to convert RA to polar and 4-hydroxy metabolites (Leo et al., 1989; Roberts et al., 1992; Marill et al., 2000; McSorley and Daly, 2000; Smith et al., 2003). Of interest, a number of human P450 family 1 and family 3 members were also shown to be capable of functioning in RA formation, converting ROL to RAL, and RAL to RA (Chen et al., 2000; Zhang et al., 2000). The most active forms for the first step were the family 1 members, with all exhibiting high activities. Differences were seen, however, in the conversion of the RAL to RA. In the present study, we were able to confirm the ability of CYP1B1 to oxidize ROL to RAL (Table 1; Fig. 1). In comparison, the ability of Cyp1b1 to synthesize RA was lower. It had a higher k_cat for conversion of ROL to RAL than did CYP1B1 (Table 1), but a much lower capability (Fig. 1). This effect was due to a considerably higher K_m(app) for the substrate by the mouse enzyme. The human to mouse CER was almost 14 (Table 1). For the conversion of RAL to RA, CYP1B1 and Cyp1b1 had k_cat values 5-fold and 2-fold higher, respectively, than for ROL to RAL. However, the human to mouse ortholog CER for oxidation of RAL was more than 25 because the K_m(app) of Cyp1b1 for RAL was 18-fold higher.

Another endogenous substrate, arachidonic acid, the metabolites of which are also known to be important in signaling, is metabolized to biologically active molecules by three pathways, cyclooxygenase, lipoxygenase, and cytochromes P450 (Capdevila and Falck, 2001). The P450 pathway (Fig. 3) generates two biologically active classes of eicosanoids, the epoxy (EETs) and hydroxy (HETEs) derivatives. These metabolites have important in vivo activities, such as modulation of electrolytes flux, fluid transport, and smooth muscle tone (vascular and airway) (McGiff, 1991; Ma et al., 1993), growth of prostate cells (O'Flaherty et al., 2002), activation of peroxisomal proliferator-activated receptor-γ (Shankaranarayanan and Nigam, 2003) and the p38-mitogen-activated protein kinase (Kalyankrishna and Malik, 2003), and inhibition of corneal Na⁺, K⁺-ATPase (Schwartzman et al., 1987). Our studies show, for the first time, that human CYP1A1 readily metabolizes arachidonate and produces ω-terminal HETEs as 75% of its metabolites. CYP1A2 had a catalytic efficiency less than 25% of CYP1A1 toward arachidonic acid and, in agreement with earlier reports (Rifkind et al., 1995; Bylund et al., 1998), produced more than half of its metabolites as EETs. CYP1B1 had a metabolite pattern different from that of the other CYP1 members, producing mainly midchain HETEs as 54% of its metabolites. Whereas it is often difficult to compare activity of an enzyme assayed in a microsomal preparation, like Supersomes, with a reconstituted pure enzyme, we have compared CYP1B1 in Supersomes with the purified, reconstituted human enzyme in an arachidonate assay, and the two had essentially the same metabolite pattern; but the Supersome preparation had only 42% of the total specific activity (data not shown). In a condition of equivalent levels of CYP1B1 and CYP1A1 in a tissue, the expected proportion of midchain HETE metabolites contributed by CYP1B1, which has a k_cat of 4% of CYP1A1, would only be about 12%. For the first time, we show Cyp1b1 to metabolize arachidonic acid. At the arachidonic acid concentration (50 μM) usually used for demonstration of metabolism by cytochromes P450, Cyp1b1 did not appear to metabolize this endogenous compound (Fig. 4, bottom). This is because it has a K_m for arachidonic acid of 0.5 mM, resulting in a human to mouse ortholog CER of 64 (Table 3). Of interest, although the mouse enzyme similarly produces a high proportion of midchain HETEs, it differs from the human ortholog by primarily producing EETs (>50%). The extremely high concentrations of arachidonate necessary to demonstrate metabolism by the mouse ortholog is interesting, since it is highly homologous to the human enzyme.

Although the amino acid sequence identity between the human and mouse CYP1B1 orthologs is 81%, higher values (90–100%) were observed for the SRSs (Table 2), using the alignments for these regions (Gotoh, 1992) as modified for the CYP1 members and based upon CYP2C5 alignment (Lewis et al., 2003). The considerable degree of similarity might explain the similarity in substrate recognition, but not the very large differences in kinetic constants and efficiencies observed with the endobiotics tested. CYP1A1 has a closer sequence identity with CYP1B1 in the SRS regions than has CYP1A2. However, human CYP1A1 and CYP1A2 both have a lower proportion of sequence identity to human CYP1B1 in all SRSs than has Cyp1b1. In SRS-5, however, the human family 1A forms reach a sequence identity of 80 to 90% with CYP1B1. This possibly explains the considerable overlap in polycyclic aromatic hydrocarbon substrate recognition by all of the family 1 forms.