Development and Evaluation of a Model to Predict Substrate Specificity of a b-glycosidase
Julio H.K. Rozenfeld and Sandro R. Marana
Instituto de Química – Depto. de Bioquímica - USP
A model was created to predict the specificity of a digestive b-glycosidase from Spodoptera frugiperda (AF 052729; Sfbgli50) towards fucosides, glycosides and galactosides. This model assumes that non-covalent interactions between two residues of the Sfbgly50 active site (E451 and Q39) and glycone hydroxyls 4 and 6 (4OH and 6OH) are independent and that their energies may be combined to predict the total energy available to stabilize the ES‡ complex. Based on that, energies of interactions with 4OH and 6OH of different residues introduced in the position 451 and 39 may be added to predict the glycone specificity of Sfbgly50 double-mutants.
In order to test this model, residues E451 and Q39 were replaced through site-directed mutagenesis, resulting in 3 double-mutants: A451E39, S451N39 and S451E39. The correct incorporation of these mutations in pET46 expression vector coding for the mutant enzymes was checked using DNA sequencing. Following that, competent E. coli (BL21DE3) cells were individually transformed with those vectors and induced to express the mutant proteins, which were then partially purified through affinity chromatography using a Ni-sepharose column.
Each of the three mutants was assayed with 3 different p-nitrophenyl beta-glycosides substrates (NPbfucoside, NPbglucoside and NPbgalactoside) and their kinetic parameters (Vmax and Km) were determined.
The lack of activity towards the glucoside substrate indicated that the model was fit for predicting preferences when comparing glucosides and galactosides. However, comparisons of Vmax/Km ratios showed an unexpected enhancement of activity towards galactosides when compared to fucosides in all 3 mutants. This phenomenon indicates not only that the model is not fit for this specific prediction, but also that the assumption that the interactions are independent might be wrong.
More structural data and site-directed mutagenesis combined with enzyme kinetic experiments are needed to approach this problem as well as to refine the model.
Supported by Fapesp and CNPq
|