The coronavirus main protease (M(pro)) represents an attractive drug target for antiviral therapy of coronavirus (CoV) infections, including severe acute respiratory syndrome (SARS). The SARS-CoV M(pro) and related CoV proteases have several distinct features, such as an uncharged Cys-His catalytic dyad embedded in a chymotrypsin-like protease fold, that clearly separate these enzymes from archetypical cysteine proteases. To further characterize the catalytic system of CoV main proteases and to obtain information about improved inhibitors, we performed comprehensive simulations of the proton-transfer reactions in the SARS-CoV M(pro) active site that lead to the Cys(-)/His(+) zwitterionic state required for efficient proteolytic activity. Our simulations, comprising the free enzyme as well as substrate-enzyme and inhibitor-enzyme complexes, lead us to predict that zwitterion formation is fostered by substrate binding but not inhibitor binding. This indicates that M(pro) employs a substrate-induced catalytic mechanism that further enhances its substrate specificity. Our computational data are in line with available experimental results, such as X-ray geometries, measured pKa values, mutagenesis experiments, and the measured differences between the kinetic parameters of substrates and inhibitors. The data also provide an atomistic picture of the formerly postulated electrostatic trigger involved in SARS-CoV M(pro) activity. Finally, they provide information on how a specific microenvironment may finely tune the activity of M(pro) toward specific viral protein substrates, which is known to be required for efficient viral replication. Our simulations also indicate that the low inhibition potencies of known covalently interacting inhibitors may, at least in part, be attributed to insufficient fostering of the proton-transfer reaction. These findings suggest ways to achieve improved inhibitors.