Inhibition of the mitochondrial permeability transition pore by cyclosporin A or trifluoperazine is transient on the time scale of cell injury studies (hours). However, these agents act synergistically and produce long-lasting inhibition when used in combination. The cause of this synergism has been investigated from the perspective of the known action of trifluoperazine as an inhibitor of mitochondrial phospholipase A2. Free fatty acids, which are phospholipase reaction products, facilitate pore opening in a concentration-dependent manner (I50 approximately 2 nmol/mg of mitochondrial protein). Endogenous and exogenous fatty acids are similarly effective. Fatty acids of differing structure are also similarly effective, but long-chain alcohols and alkanes are ineffective. Free fatty acids accumulate in cyclosporin A-treated mitochondria when Ca2+ plus tert-butyl hydroperoxide or Ca2+ plus N-ethylmaleimide is present, but do not accumulate when Ca2+ plus inorganic phosphate is present. In the presence of cyclosporin A, bovine serum albumin markedly delays pore opening induced by tert-butyl hydroperoxide or N-ethylmaleimide, but has little effect on pore opening induced by inorganic phosphate, which is subject to long-lasting inhibition by cyclosporin A without trifluoperazine. Free fatty acid accumulation is thus a factor which limits pore inhibition by cyclosporin A. However, trifluoperazine has no effect on free fatty acid accumulation in intact, cyclosporin-inhibited mitochondria and thus does not act by inhibiting phospholipases. Comparing the actions of free fatty acids, trifluoperazine, long-chain acyl cations, and other effectors on the pore suggests that a more negative membrane surface potential favors pore opening and a more positive potential favors a closed pore. Expected surface potential effects of trifluoperazine can explain the synergism between this compound and cyclosporin A as pore inhibitors. Surface potential may influence the pore through the voltage-sensing element which responds to transmembrane potential. The present data also suggest that long-lived, solute-selective forms of the pore exist when it is opened in the presence of inhibitors. The implications of these findings for pore regulation and for the use of cyclosporin A to identify pore opening as a component of cell injury mechanisms are discussed.