To investigate the roles of catalytic residues and the nature of the nucleophilic intermediate in citrate synthase, calculations have been carried out on models of the first stage of the reaction with ab initio (MP2/6-31+G(d)//6-31+G(d)) and semiempirical (AM1 and PM3) methods. The first stage of the reaction involves deprotonation of acetyl-CoA, with Asp-375 identified as the likely base. The resulting intermediate is believed to be stabilized by a hydrogen bond from a neutral histidine side chain (His-274), which has been suggested to be a "short, strong" or "low-barrier" hydrogen bond. Such bonds have been suggested to have exceptionally high energies and to stabilize many enzyme reaction intermediates. Transition state and stable hydrogen bonded complex geometries have been fully optimized for models of proton transfer between Asp-375 and acetyl-CoA and between the enolate of acetyl-CoA and His-274. The results support the proposal that Asp-375 is the base in the reaction and show that stabilization of the thioester enolate at the active site is required if it is to exist as a reaction intermediate. The effective basicities of methylimidazolate (representing His-274) and the thioester enolate in the hydrogen bonded complex are calculated to be clearly different. The enolate-methylimidazole complex (with the proton localized on methylimidazole) is found to be significantly lower in energy than the enol complex (in which the hydrogen bonded proton is transferred from methylimidazole), which is not stable. Unless the pK(a) for deprotonation of neutral His-274 is lowered by conditions in the active site, it appears that the hydrogen bond with the enolate will not be of the low-barrier type. The highly favorable energy of the interaction between the thioester enolate and methylimidazole, as well as calculations on a larger model including all three components, indicate that a normal hydrogen bond with His-274 can make an important contribution to stabilization of the enolate intermediate.