We have coevolved high activity and hyperstability in subtilisin by sequentially randomizing 12 amino acid positions in calcium-free subtilisin. The optimal amino acid for each randomized site was chosen based on stability and catalytic properties and became the parent clone for the next round of mutagenesis. Together, the 12 selected mutations increased the half-life of calcium-free subtilisin at elevated temperature by 15,000-fold. The catalytic properties of the mutants were examined against a range of substrates. In general, only mutations occurring at or near the substrate-binding surface have measurable effects on catalytic constants. No direct influence of stability on catalytic properties was observed. A high-stability mutant, Sbt140, was a more efficient enzyme in terms of k(cat)/K(m) than a commercial version of subtilisin across a range of substrates but had a lower k(cat) against tight-binding substrates. The reason for this behavior was discerned by examining microscopic rate constants for the hydrolysis of a tight-binding peptide substrate. Burst kinetics were observed for this substrate, indicating that acylation is not rate-limiting. Although acylation occurs at the rate of substrate binding, k(cat) is attenuated by the slow release of the N-terminal product. Natural evolution appears to have optimized catalytic activity against a range of sequences by achieving a balance between substrate binding and the rate of release of the N-terminal product.