Kinetic properties of prothrombinase were investigated as a function of composition and structure of the membrane component. The kinetic properties were quite diverse, giving linear or nonlinear Eadie-Hofstee plots and substrate concentrations at half-maximum velocity ([S]0.5) that varied from 5 to more than 200 nM. This reaction might be described as a 'catalytic system' in order to distinguish it from standard models that have been developed to describe the kinetics of soluble enzymes. The latter do not anticipate a key feature of prothrombinase and probably other membrane-bound enzymes, which is the presence of reaction steps that do not contain an enzyme (E) term. At least four kinetic mechanisms can arise from a logical series of steps that may occur during the prothrombinase reaction. All of these mechanisms appeared to contribute to reaction properties under some conditions. In some cases, one mechanism dominated at low substrate concentration and another at high substrate concentration. This change in the course of a titration was referred to as 'mechanism switching'. Only membranes of low phosphatidylserine (PS) content displayed Michaelis-Menten behavior. Transfer of substrate from the membrane surface to the enzyme was not important so that the enzyme was involved in capture of substrate directly from solution. As PS content increased, transfer of substrate from the membrane surface to the enzyme occurred. In these cases, multiple mechanisms contributed to the reaction so that K(M) and apparent K(M), properties that describe an enzyme active site, were not appropriate, even when Eadie-Hofstee plots were linear. At high PS content, the enzyme captured every substrate molecule that became bound to the same vesicle. Reaction velocity was governed entirely by protein-membrane binding rather than by enzyme properties. Eadie-Hofstee plots were often nonlinear and/or V(max) was less than k(cat)[E1]. A small impact from collision-limited kinetics was also detected. Small unilamellar vesicles (SUV, 30 nm diameter) gave higher [S]0.5 values than large unilamellar vesicles (LUV, 100 nm diameter) of the same phospholipid composition. There appeared to be two bases for this behavior. First, LUV may provide a better relationship between the phospholipid surface and the enzyme, giving a better substrate binding site. Second, for membranes containing high PS, the number of substrate binding sites per vesicle contributed to the enhanced function of LUV. These studies showed that mechanism-switching was important to prothrombinase reaction in vitro and suggest that various mechanisms, generated by the nature of the membrane, may be an important regulator for prothrombinase behavior in vivo.