The hypothesis that His159 in yeast enolase moves on a polypeptide loop to protonate the phosphoryl of 2-phosphoglycerate to initiate its conversion to phosphoenolpyruvate was tested by preparing H159N, H159A, and H159F enolases. These have 0.07%-0.25% of the native activity under standard assay conditions and the pH dependence of maximum velocities of H159A and H159N mutants is markedly altered. Activation by Mg2+ is biphasic, with the smaller Mg2+ activation constant closer to that of the "catalytic" Mg2+ binding site of native enolase and the larger in the mM range in which native enolase is inhibited. A third Mg2+ may bind to the phosphoryl, functionally replacing proton donation by His159. N207A enolase lacks an intersubunit interaction that stabilizes the closed loop(s) conformation when 2-phosphoglycerate binds. It has 21% of the native activity, also exhibits biphasic Mg2+ activation, and its reaction with the aldehyde analogue of the substrate is more strongly inhibited than is its normal enzymatic reaction. Polypeptide loop(s) closure may keep a proton from His159 interacting with the substrate phosphoryl oxygen long enough to stabilize a carbanion intermediate.