Although serine acetyltransferase (SAT) from Escherichia coli is homologous with a number of bacterial enzymes that catalyze O-acetyl transfer by a sequential (ternary complex) mechanism, it has been suggested, from experiments with the nearly identical enzyme from Salmonella typhimurium, that the reaction could proceed via an acetyl-enzyme intermediate. To resolve the matter, the E. coli gene for SAT was overexpressed and the enzyme purified 13-fold to homogeneity. The results of a steady-state kinetic analysis of the forward reaction are diagnostic for a ternary complex mechanism, and the response of SAT to dead-end inhibitors indicates a random order for the addition of substrates. The linearity of primary double-reciprocal plots, in the presence and absence of dead-end inhibitors, argues that interconversion of ternary complexes is not significantly faster than kcat, whereas substrate inhibition by serine suggests that breakdown of the SAT.CoA binary complex is rate-determining. The results of equilibrium isotope exchange experiments, for both half-reactions, rule out a "ping-pong" mechanism involving an acetyl-enzyme intermediate, and a pre-steady-state kinetic analysis of the turnover of AcCoA supports such a conclusion. Kinetic data for the reverse reaction (acetylation of CoA by O-acetylserine) are also consistent with a steady-state random-order mechanism, wherein both the breakdown of the SAT*serine complex and the interconversion of ternary complexes are partially rate-determining.