Recently it has been found that the kinetic mechanism for Escherichia coli 5-enolpyruvoylshikimate-3-phosphate synthase (EPSPS) in the forward direction is random with synergistic binding of substrates and inhibitors (K. J. Gruys, M. C. Walker, and J. A. Sikorski, 1992, Biochemistry 31, 5534). This work, however, did not address the reverse reaction with 5-enolpyruvoylshikimate-3-phosphate (EPSP) and phosphate (Pi) as substrates where a similar question of random versus ordered addition of substrates remained. Previous transient-state kinetic results led to a proposal for an equilibrium-ordered mechanism, where binding of EPSP occurs first followed by Pi (K. S. Anderson, and K. A. Johnson, 1990, Chem. Rev. 90, 1131). Steady-state kinetic results of the reverse reaction presented here suggest that, like the forward reaction, addition of substrates occurs randomly. Initial velocity studies with EPSP and Pi show a normal intersecting pattern in the reciprocal plots, consistent with a random or steady-state-ordered mechanism, but not with equilibrium-ordered addition of substrates. Inhibition of the EPSPS reverse reaction by 5-amino-S3P or the S3P-glyphosate hybrid molecule gave the expected competitive patterns versus EPSP, but mixed noncompetitive patterns versus Pi. These results also disfavor an equilibrium-ordered model, but again are consistent with a random or steady-state-ordered mechanism. A more quantitative mechanistic analysis of the inhibition data to determine the true rather than apparent Ki values provides evidence for a random over a steady-state-ordered addition of substrates. These results in combination with previous findings lead to the conclusion that the mechanism is random addition of EPSP and Pi since it is the only possible model for substrate addition that is consistent with the cumulative data from all kinetic (transient- as well as steady-state) and direct binding studies.