Molecular cloning and functional analysis of cytochrome P450 1A2 from Japanese monkey liver: comparison with marmoset cytochrome P450 1A2

Qualitative de novo analysis of full length cDNA and quantitative analysis of gene expression for common marmoset (Callithrix jacchus) transcriptomes using parallel long-read technology and short-read sequencing

The common marmoset (Callithrix jacchus) is a non-human primate that could prove useful as human pharmacokinetic and biomedical research models. The cytochromes P450 (P450s) are a superfamily of enzymes that have critical roles in drug metabolism and disposition via monooxygenation of a broad range of xenobiotics; however, information on some marmoset P450s is currently limited. Therefore, identification and quantitative analysis of tissue-specific mRNA transcripts, including those of P450s and flavin-containing monooxygenases (FMO, another monooxygenase family), need to be carried out in detail before the marmoset can be used as an animal model in drug development. De novo assembly and expression analysis of marmoset transcripts were conducted with pooled liver, intestine, kidney, and brain samples from three male and three female marmosets. After unique sequences were automatically aligned by assembling software, the mean contig length was 718 bp (with a standard deviation of 457 bp) among a total of 47,883 transcripts. Approximately 30% of the total transcripts were matched to known marmoset sequences. Gene expression in 18 marmoset P450- and 4 FMO-like genes displayed some tissue-specific patterns. Of these, the three most highly expressed in marmoset liver were P450 2D-, 2E-, and 3A-like genes. In extrahepatic tissues, including brain, gene expressions of these monooxygenases were lower than those in liver, although P450 3A4 (previously P450 3A21) in intestine and P450 4A11- and FMO1-like genes in kidney were relatively highly expressed. By means of massive parallel long-read sequencing and short-read technology applied to marmoset liver, intestine, kidney, and brain, the combined next-generation sequencing analyses reported here were able to identify novel marmoset drug-metabolizing P450 transcripts that have until now been little reported. These results provide a foundation for mechanistic studies and pave the way for the use of marmosets as model animals for drug development in the future.

Characterization of marmoset CYP2B6: cDNA cloning, protein expression and enzymatic functions

The common marmoset is a promising species for evaluating the safety of drug candidates. To further understand the capacity for drug metabolism in marmosets, a cDNA encoding a CYP2B enzyme was cloned from the total RNA fraction of marmoset liver by 3'- and 5'-RACE methods. Nucleotide and deduced amino acid sequences showed 90.8 and 86.2% identity, respectively, with human CYP2B6. The marmoset CYP2B6 (marCYP2B6) protein was expressed in insect cells, and its enzymatic properties were compared with those of human (humCYP2B6) and cynomolgus monkey (cynCYP2B6) orthologs in liver and insect cell microsomes. Enzymatic functions were examined for the oxidation of 7-ethoxy-4-(trifluoromethyl)coumarin (7-ETC), bupropion (BUP) and efavirenz (EFV). The kinetic profiles for the oxidation of the three substrates by liver microsomal fractions were similar between humans and cynomolgus monkeys (biphasic for 7-ETC and monophasic for BUP and EFV), but that of marmosets was unique (monophasic for 7-ETC and biphasic for BUP and EFV). Recombinant enzymes, humCYP2B6 and cynCYP2B6, also yielded similar kinetic profiles for the oxidation of the three substrates, whereas marCYP2B6 showed activity only for 7-ETC hydroxylation. In silico docking simulations suggested that two amino acid residues, Val-114 and Leu-367, affect the activity of marCYP2B6. In fact, a marCYP2B6 mutant with substitutions V114I and L367V exhibited BUP hydroxylase activity that was 4-fold higher than that of humCYP2B6, while its EFV 8-hydroxylase activity was only 10% that of the human enzyme. These results indicate that the amino acids at positions 114 and 367 affect the enzymatic capacity of marmoset CYP2B6.

Regio- and stereoselective oxidation of propranolol enantiomers by human CYP2D6, cynomolgus monkey CYP2D17 and marmoset CYP2D19

Toxic and pharmacokinetic profiles of drug candidates are evaluated in vivo often using monkeys as experimental animals, and the data obtained are extrapolated to humans. Well understanding physiological properties, including drug-metabolizing enzymes, of monkeys should increase the accuracy of the extrapolation. The present study was performed to compare regio- and stereoselectivity in the oxidation of propranolol (PL), a chiral substrate, by cytochrome P450 2D (CYP2D) enzymes among humans, cynomolgus monkeys and marmosets. Complimentary DNAs encoding human CYP2D6, cynomolgus monkey CYP2D17 and marmoset CYP2D19 were cloned, and their proteins expressed in a yeast cell expression system. The regio- and stereoselective oxidation of PL enantiomers by yeast cell microsomal fractions were compared. In terms of efficiency of expression in the system, the holo-proteins ranked CYP2D6=CYP2D17>>CYP2D19. This may be caused by the bulky side chain of the amino acid residue at position 119 (leucine for CYP2D19 vs. valine for CYP2D6 and CYP2D17), which can disturb the incorporation of the heme moiety into the active-site cavity. PL enantiomers were oxidized by all of the enzymes mainly into 4-hydroxyproranolol (4-OH-PL), followed by 5-OH-PL and N-desisopropylpropranolol (NDP). In the kinetic analysis, apparent K(m) values were commonly in the μM range and substrate enantioselectivity of R-PL<S-PL was observed in both K(m) and V(max) values for the formation of the three metabolites from PL enantiomers. The activity to produce NDP tended to be higher for the monkey enzymes, particularly CYP2D17, than for the human enzyme. These results indicate that in the oxidation of PL enantiomers by CYP2D enzymes, stereoselectivity is similar but regioselectivity is different between humans and monkeys.

Comparative study of the oxidation of propranolol enantiomers in hepatic and small intestinal microsomes from cynomolgus and marmoset monkeys

Oxidative metabolism of propranolol (PL) enantiomers (R-PL and S-PL) to 4-hydroxypropranolol (4-OH-PL), 5-OH-PL and N-deisopropylpropranolol (NDP) was examined in hepatic microsomes from cynomolgus and marmoset monkeys and in small intestinal microsomes from monkeys and humans. In hepatic microsomes, levels of oxidation activities were similar between the two monkey species, and substrate enantioselectivity (R-PL<S-PL) was observed in the formation of 5-OH-PL and/or NDP. Kinetic experiments revealed that the formation of all metabolites was biphasic in cynomolgus monkeys, whereas only the formation of NDP was biphasic in marmosets. Inhibition experiments employing human CYP antibodies and chemical inhibitors suggested that mainly CYP2D enzymes and partially CYP1A and 2C enzymes are involved in the oxidation of PL in both monkey liver microsomes. In small intestinal microsomes, activity levels were much higher in cynomolgus monkeys than in marmosets and humans and reversed substrate enantioselectivity (R-PL>S-PL) was seen in the formation of NDP in cynomolgus monkeys and humans and in the formation of 5-OH-PL in marmosets. The formation of the three metabolites in cynomolgus monkeys and the formation of NDP in marmosets were biphasic, while the formation of 4-OH-PL in humans was monophasic. From the inhibition experiments using CYP antibodies, CYP2C9 and 2C19 were thought to be involved as N-deisopropylases and CYP2D6 and 3A4 as 4-hydroxylases in human small intestine. Furthermore, CYP1A, 2C and 3A enzymes could be involved in cynomolgus monkeys and CYP2C and 3A enzymes in marmosets. These results indicate that the oxidative profile of PL in hepatic and small intestinal microsomes differ considerably among cynomolgus monkeys, marmosets and humans.

Purification, molecular cloning, heterologous expression and characterization of pig CYP1A2

Porcine cytochrome P450 (CYP) 1A2 was purified to electrophoretic homogeneity from the hepatic microsomes of beta-naphthoflavone-treated male pigs. In a reconstituted system, this enzyme showed a good catalytic activity towards caffeine, acetanilide, and methoxyresorufin, all known markers of mammalian CYP1A2. Using 3'- and 5'-rapid amplification of coding DNA (cDNA) ends (RACE), we amplified from the liver RNA of control pigs a full-length 1827 bp cDNA containing an open reading frame of 1548 bp which encoded a putative CYP1A2 protein of 516 amino acids and an estimated Mr of 58 380 Da. Reverse transcriptase-polymerase chain reaction (RT-PCR) experiments showed that the messenger RNA (mRNA) of CYP1A2 was expressed in liver, heart and nasal mucosa but not in lung, small intestine, kidney and brain. Using the pCW vector containing a N-terminal modified cDNA, pig CYP1A2 was expressed in Escherichia coli. 3-[(3-Chloroamidopropyl)dimethylmmonio]-1-propane-sulfonate (CHAPS)-solubilized E. coli preparations expressing CYP1A2 produced a functionally isoform which, in a reconstituted system, was catalytically active toward ethoxyresorufin and methoxyresorufin showing K(m)'s similar to those obtained with CYP1A2 purified from pig liver or human recombinant CYP1A2. Taken together, these results demonstrate that domestic pigs have a functionally active CYP1A2 gene well expressed in the liver with biochemical properties quite similar to those corresponding to the human enzyme.

Cloning of a cDNA encoding a novel marmoset CYP2C enzyme, expression in yeast cells and characterization of its enzymatic functions

We cloned a cDNA encoding a novel CYP2C enzyme, called P450 M-2C, from a marmoset liver. The deduced amino acid sequence showed high identities to those of human CYP2C8 (87%), CYP2C9 (78%) and CYP2C19 (77%). The P450 M-2C enzyme expressed in yeast cells catalyzed p-methylhydroxylation of only tolbutamide among four substrates tested, paclitaxel as a CYP2C8 substrate, diclofenac and tolbutamide as CYP2C9 substrates and S-mephenytoin as a CYP2C19 substrate. p-Methylhydroxylation of tolbutamide by marmoset liver microsomes showed monophasic kinetics, and the apparent K(m) value (1.2 mM) for the substrate was similar to that of the recombinant P450 M-2C (1.8 mM). Although all of the recombinant human CYP2C8, CYP2C9 and CYP2C19 expressed in yeast cells catalyzed tolbutamide p-methylhydroxylation, the kinetic profile of CYP2C8 was most similar to that of P450 M-2C. Tolbutamide oxidation by the marmoset liver microsomes and the recombinant P450 M-2C was inhibited most effectively by quercetin, a CYP2C8 inhibitor, followed by omeprazole, a CYP2C19 inhibitor, whereas sulfaphenazole, a CYP2C9 inhibitor, was less potent under the conditions used. These results indicate that P450 M-2C is the major tolbutamide p-methylhydroxylase in the marmoset liver.

cDNA cloning and characterization of feline CYP1A1 and CYP1A2

Deficiency of drug glucuronidation in the cat is one of the major reasons why this animal is highly sensitive to the side effects of drugs. The characterization of cytochrome P450 isoforms belonging to the CYP1A subfamily, which exhibit important drug oxidation activities such as activation of pro-carcinogens, was investigated. Two cDNAs, designated CYP1A-a and CYP1A-b, corresponding to the CYP1A subfamily were obtained from feline liver. CYP1A-a and CYP1A-b cDNAs comprise coding regions of 1554 bp and 1539 bp, and encode predicted amino acid sequences of 517 and 512 residues, respectively. These amino acid sequences contain a heme-binding cysteine and a conserved threonine. The cDNA identities, as well as the predicted amino acid sequences containing six substrate recognition sites, suggest that CYP1A-a and CYP1A-b correspond to CYP1A1 and CYP1A2, respectively. This was confirmed by the kinetic parameters of the arylhydrocarbon hydroxylase and 7-ethoxyresorufin O-deethylase activities of expressed CYPs in yeast AH22 cells and by the tissue distribution of each mRNA. However, theophylline 3-demethylation is believed to be catalyzed by CYP1A1 in cats, based on the high V(max) and low K(m) seen, in contrast to other animals. Because feline CYP1A2 had a higher K(m) for phenacetin O-deethylase activity with acetaminophen, which cannot be conjugated with glucuronic acid due to UDP-glucuronosyltransferase deficiency, it is supposed that the side effects of phenacetin as a result of toxic intermediates are severe and prolonged in cats.

Oxidative metabolism of 5-methoxy-N,N-diisopropyltryptamine (Foxy) by human liver microsomes and recombinant cytochrome P450 enzymes

In vitro quantitative studies of the oxidative metabolism of (5-methoxy-N,N-diisopropyltryptamine, 5-MeO-DIPT, Foxy) were performed using human liver microsomal fractions and recombinant CYP enzymes and synthetic 5-MeO-DIPT metabolites. 5-MeO-DIPT was mainly oxidized to O-demethylated (5-OH-DIPT) and N-deisopropylated (5-MeO-IPT) metabolites in pooled human liver microsomes. In kinetic studies, 5-MeO-DIPT O-demethylation showed monophasic kinetics, whereas its N-deisopropylation showed triphasic kinetics. Among six recombinant CYP enzymes (CYP1A2, CYP2C8, CYP2C9, CYP2C19, CYP2D6 and CYP3A4) expressed in yeast or insect cells, only CYP2D6 exhibited 5-MeO-DIPT O-demethylase activity, while CYP1A2, CYP2C8, CYP2C9, CYP2C19 and CYP3A4 showed 5-MeO-DIPT N-deisopropylase activities. The apparent Km value of CYP2D6 was close to that for 5-MeO-DIPT O-demethylation, and the Km values of other CYP enzymes were similar to those of the low-Km (CYP2C19), intermediate-Km (CYP1A2, CYP2C8 and CYP3A4) and high-Km phases (CYP2C9), respectively, for N-deisopropylation in human liver microsomes. In inhibition studies, quinidine (1 microM), an inhibitor of CYP2D6, almost completely inhibited human liver microsomal 5-MeO-DIPT O-demethylation at a substrate concentration of 10 microM. Furafylline, a CYP1A2 inhibitor, quercetin, a CYP2C8 inhibitor, sulfaphenazole, a CYP2C9 inhibitor and ketoconazole, a CYP3A4 inihibitor (5 microM each) suppressed about 60%, 45%, 15% and 40%, respectively, of 5-MeO-DIPT N-deisopropylation at 50 microM substrate. In contrast, omeprazole (10 microM), a CYP2C19 inhibitor, suppressed only 10% of N-deisopropylation by human liver microsomes, whereas at the same concentration the inhibitor suppressed the reaction by recombinant CYP2C19 almost completely. These results indicate that CYP2D6 is the major 5-MeO-DIPT O-demethylase, and CYP1A2, CYP2C8 and CYP3A4 are the major 5-MeO-DIPT N-deisopropylase enzymes in the human liver.