Merits and limitations of recombinant models for the study of human P450-mediated drug metabolism and toxicity: an intralaboratory comparison.

A wide variety of pharmacological and toxicological properties of drugs are determined by cytochrome P450-mediated metabolism. Characterization of these pathways and of the P450 isoenzymes involved constitutes an essential part of drug development. Similarly, because P450s are catalyzing the toxication and detoxication of environmental pollutants, an understanding of these reactions facilitates risk assessment in environmental toxicology. Recently, a variety of recombinant expression systems has been employed to study the role of human P450s in these reactions. These include insect, bacterial, yeast, and mammalian models. As these were developed and characterized by different laboratories, evaluation of their merits and limitations is inherently difficult. To resolve this problem, we have established and characterized the latter three systems and present the key results here. In general, the catalytic properties of P450 isozymes in the various models were rather similar. However, taking technical considerations into account as well as the high level of functional expression of P450s achieved in bacteria make this system ideally suited for drug metabolism research, including the generation of milligram quantities of metabolites for structural determinations. For toxicological studies, however, expression of P450s in mammalian cells was most appropriate. This is exemplified here by studies into the role of human P450s in the activation and inactivation of chemotherapeutic drugs.

Competition between cytochrome P-450 isozymes for NADPH-cytochrome P-450 oxidoreductase affects drug metabolism.

NADPH-cytochrome P-450 oxidoreductase (CPR) is essential for the catalytic activity of cytochrome P-450 (P-450). On a molar basis, the amount of P-450 exceeds that of CPR in human liver. In this study, we investigated whether drug-drug interactions can occur as a result of competition between P-450 isozymes for this ancillary protein. For this purpose, combinations of P-450 isozymes were coexpressed together with P-450 reductase in Escherichia coli. We show that testosterone inhibited the CYP2D6-mediated bufuralol 1'-hydroxylase activity in bacterial membranes containing both CYP2D6 and CYP3A4 but not in membranes containing CYP2D6 alone. Conversely, bufuralol inhibited the CYP3A4-mediated testosterone 6beta-hydroxylase activity in bacterial membranes containing both CYP3A4 and CYP2D6 but not in membranes containing only CYP3A4. In each case, inhibition was seen even at a P-450 to P-450 reductase ratio of 1.9:1, which is more favorable than the ratio of 4 reported for human liver. The physiological significance of this mechanism was demonstrated by the observation that testosterone inhibited several prototypical P-450 enzyme activities, such as bufuralol 1'-hydroxylase, coumarin 7-hydroxylase, and 7-ethoxyresorufin O-dealkylase, in human liver microsomes, but not if tested against a panel of bacterial membranes containing the human P-450 isozymes that mainly catalyze these reactions.

Dimethylsulfide:acceptor oxidoreductase from Rhodobacter sulfidophilus. The purified enzyme contains b-type haem and a pterin molybdenum cofactor.

Dimethylsulfide:receptor oxidoreductase was purified from the purple non-sulfur phototrophic bacterium Rhodobacter sulfidophilus. The native form of the enzyme had a molecular mass of 152 kDa and was composed of three distinct subunits of 94, 38 and 32 kDa. Dimethylsulfide:acceptor oxidoreductase did not oxidise other thioethers which were tested. The enzyme was able to reduce a variety of N-oxides using reduced methylviologen as electron donor but it reduced dimethylsulfoxide at a very low rate. The resting form of dimethylsulfide:acceptor oxidoreductase exhibited a spectrum which was characteristic of a reduced cytochrome with absorbance maxima at 562 nm, 533 nm and 428 nm. Pyridine haemochrome analysis established that the cytochrome contained a b-type haem and a content of 0.65 mol protohaem/mol enzyme was determined. After oxidation of the haem with ferricyanide, the absorbance spectrum of the reduced cytochrome was restored by reduction with dimethylsulfide. Metal analysis revealed that dimethylsulfide:acceptor oxidoreductase contained 0.5 mol Mo and 3.5 mol Fe/mol enzyme. Heat treatment of the enzyme released material with fluorescence excitation and emission spectra which were characteristic of form B of the pterin component of the pterin molybdenum cofactor. From this analysis it is concluded that dimethylsulfide:acceptor oxidoreductase is a molybdenum oxotransferase which may also contain a iron-sulfur cluster. It is suggested that the haem and pterin molybdenum cofactor are associated with the 94-kDa subunit.

Cellular location influences copper-dependent topaquinone formation for phenylethylamine oxidase expressed in Escherichia coli K-12.

Arthrobacter globiformis amine oxidase produced by Escherichia coli cells grown in copper-depleted media was reported to undergo activation due to formation of its topaquinone cofactor in a copper-dependent autocatalytic reaction. Likewise, a mutated E. coli amine oxidase located in the cytoplasm was reported to form topaquinone autocatalytically in an EDTA-sensitive reaction. Here we show unequivocally that formation of an amine oxidase lacking topaquinone is primarily a consequence of the location of the enzyme in the cytoplasm rather than the level of copper in the growth medium. For E. coli, insertion of copper into apoamine oxidase and subsequent topaquinone formation occur after export of the apoenzyme into the periplasm.

Formation in vitro of the 3,4,6-trihydroxyphenylalanine quinone cofactor.

An Escherichia coli K-12 2-phenylethylamine oxidase gene with a mutated leader sequence region produced a largely inactive form of the enzyme in the cytoplasm. This form of the enzyme was activated 30-50-fold on incubation at 30 degrees C in the absence of any added cofactors. After activation the enzyme contained a quinone which was not detected in the non-activated form. This is the first report of the formation in vitro of any quinoenzyme cofactor.