At present, there is no general standard automated method for engineering metalloenzymes, industrially-relevant systems able to catalyze environmentally friendly reactions. One of the most studied natural metalloenzymes is the second isoform of human carbonic anhydrase (HCA II), capable of reversibly hydrating CO2 to bicarbonate. The objective of this thesis is to develop computational methods capable of identifying (i) the positions of peptidic chains which can be altered without compromising the structural stability of the protein and (ii) the specific mutations at those positions that maximize advantageous chemical properties, e.g., desired catalytic activities. We intend to show that protein scaffolds of medium size, like the ß1 domain of the staphylococcal protein G (Gß1), can act as a generic template to generate systems with desirable catalytic activities. Gß1 is a 56-residue peptidic chain of known thermal stability; previous research conducted by Bozkurt et al. characterized a highly stable metal variant and a moderately stable and hydrolytically active mutant. With these results as a starting platform, the objective pursued in this thesis is to optimize a metal-containing mutant of Gß1 both for stability and catalytical activity, so as to mimic HCA II activity. An innovative computational scanning mutagenesis protocol was implemented, where each residue of wild-type Gß1 was mutated to several classes of amino acids and the resulting energy changes were computed. Despite some limitations, a series of positions whose substitution was either well tolerated or very destabilizing was identified and our predictions agreed largely with recent experimental literature. A twelve-position polymutant candidate originally designed by Bozkurt, promising both in terms of functionality and stability, was examined and found to be stable using molecular dynamics techniques, including "classical" molecular dynamics and hybrid quantum mechanics/molecular mechanics (QM/MM) at the density functional theory level of theory. The mutant features a Zn2+ coordinated by three histidines at positions 5, 16 and 18 and a water molecule or a hydroxide ion, reminiscent of the catalytic site in HCA II. To study the capability of the Gß1 mutant to bind CO2, a series of well-tempered metadynamics simulations was run and a potential binding site was identified. In parallel, the propensity of the zinc-bound water molecule to deprotonate to a hydroxide ion was investigated by means of a QM/MM thermodynamic integration procedure, as this step is rate-limiting in HCA II catalysis. Results show a concerted shuttle movement of protons among several water molecules ending up on the negative side chain of Asp33 as the final acceptor. A novel approach based on a multiobjective genetic algorithm (GA) optimization was utilized to tune the domain stability together with catalytically relevant features. Mimicking natural evolution, our GA approach cyclically generates populations of mutants, selects the best ones and passes the selected features to the following population, until a convergence criterion is met and quasi-optimal solutions are found. The fittest mutants were identified via scoring functions for CO2 binding affinity and deprotonation propensity of the Zn2+-bound water. The optimized mutants for the chemical features thus obtained lay the steps for the next generation of stable Gß1 mutants with catalytic activity.