1. The in vitro cooperativity exhibited by cytochrome P450 (CYP) 3A4 is influenced by the nature of the recombinant system in which the phenomenon is studied. Diclofenac, piroxicam and R-warfarin were used as model substrates, and quinidine was the effector. 2. The 5-, 5'- and 10-hydroxylation of diclofenac, piroxicam and R-warfarin, respectively, were enhanced five- to sevenfold by quinidine in human liver microsomal incubations. Whereas these cooperative drug interactions were apparent in incubations with CYP3A4 expressed in human lymphoblast cells, similar phenomena were not observed with the enzyme expressed in insect cells. 3. Insect cell microsomes were treated with a detergent and CYP3A4 was solubilized into a buffer medium. In incubations with CYP3A4 'freed' from its host membrane, the 5-hydroxylation of diclofenac increased with increasing quinidine concentrations, reaching a maximal eightfold elevation relative to controls. The metabolism of piroxicam and warfarin was similarly enhanced by quinidine. 4. Kinetically, enhancement by quinidine of the 5-hydroxylation of diclofenac in incubations with solubilized CYP3A4 was characterized by increases in the rate of metabolism with little change in the substrate-binding affinity. Conversely, the 3-hydroxylation of quinidine was not affected by diclofenac. 5. The data suggest that certain properties of CYP3A4 are masked by expression of the protein in insect cells and reinforce the concept that the enzyme possesses multiple binding domains. The absence of cooperative drug interactions with quinidine when CYP3A4 was expressed in insect cells might be due to an absence of enzyme conformation changes on quinidine binding, or the inability of quinidine to gain access to a putative effector-binding domain. 6. Caution should be exercised when comparing models for CYP3A4 cooperativity derived from different recombinant preparations of the enzyme.