ABSTRACT
AIMS: A homology modeling methodology was developed and used to obtain the 3D structures for four putative catalases of Rhizopus oryzae in order to assess their functionality. MAIN METHODS: Homology models were built using different modeling strategies using non-protein compounds as steric constraints, a symmetry constraint to force identical chains and an additional loop modeling algorithm. Percent structural overlap values (SO) were calculated for each model-template pair to qualify the homology models. KEY FINDINGS: Comparing the different modeling strategies by the SO values revealed that the quality of the models, i.e. the similarity to the template was greatly increased in the presence of the prosthetic groups, modeling multiple protein chains together, enforcing symmetrical chains and applying additional loop modeling. For the best homology models achieved this way, the SO values express similar evolutionary relationships between the proteins modeled and the templates that were previously established by phylogenetic analysis. In three out of the four catalases of R. oryzae the highest quality models, the active center, i.e. the heme molecule and the surrounding amino acids showed a spatial arrangement identical to that observed experimentally in other catalases. The remaining protein is missing an 11 residue long fragment and has mutated residues within the active center. SIGNIFICANCE: Better homology models can be obtained with templates chosen by phylogenetic relationship, although building an accurate model needs structural constraints too. Calculating the structural overlap between the models and the templates may also help to find the appropriate templates.
