Free energy perturbation mutations of two core residues in B. licheniformis beta-lactamase

SICA, MP; SANTOS, J; RISSO, VA; ERMÁCORA, MR

Buenos Aires, Argentina

Workshop; Giambiagi Winter School; 2006

Abstract: One of the most important problems of the structural biology is the way the protein sequence codes its 3D structure. In this context, our laboratory developed some mutants of B. licheniformis Exo?samll b?lactamse (ESBL) with SeràCys mutations in unexposed to solvent positions. Although this mutation only implies an S for O replacement, the mutants (S126C?, S265C? and S126C/S265C?ESBL) presented interesting experimental features. Spectroscopic and hydrodynamic studies revealed that wild?type ESBL and its mutants achieved a rather similar native conformation, while kinetic studies showed that the intermediate structures explored during the folding pathway was different in every case. Thermal and urea?induced denaturations revealed that, with respect to the wild?type, S126C?ESBL was more stable, while S265C? and S126C/S265C?ESBL were less and similarly stable, respectively. In order to understand these results, we carried out molecular dynamics (MD) simulations of the SeràCys mutations using the GROMOS96 (ff43a1) force field and the GROMACS 3.2 package. Basically, after minimization and solvation in SPCE?water, we simulated ESBL during 18 ns at 300 K in a NPT ensemble. Also, we implemented free energy perturbation (FEP) mutations using the thermodynamic integration (TI) method to finally simulate the native mutated conformations. Inputs for the native state mutations were taken at 2 and 10 ns from the wild?type ESBL simulation. The unfolded state was approximated to one 21?residues peptide for each SeràCys mutation, where the mutating residue was in the 11th position. The mutations consisted in eleven 200?ps steps of fixed l (Dl = 0.1) connected by 100 ps linearly increasing l. The final native conformations were further run during 2 ns. The mutations DG were obtained from the fixed?l steps of the native and unfolded states mutations (DGSàC,N and DGSàC,U, respectively) as the area under the vs l curve and DDGWT?mut (DDG between wild type and mutant proteins) were calculated as DDGWT?mut = DGSàC,N - DGSàC,U. The calculated DDGWT?mut values were in good agreement with those obtained experimentally (DDGWT-S126C = ?1.9 ± 0.4, DDGWT-S126C = 0.9 ± 0.4 for the 10 ns FEP mutation). The Ca RMSD, accessible surface area as well as the visual inspection, suggested that the mutants achieved a globally similar conformation with respect to wild?type ESBL. The S265C simulation revealed that an H?bond between Ser265 and Arg244 was disrupted. As for S126C mutant, cysteine side chain was more efficiently accommodated in its environment, which is correlated to a slight movement of favoring contacts with Val77 side chain and Lys73 Ca. Besides, S126C mutation triggers a conformational change that reaches 3/10 helix (residues 221-226) and implies the formation of a new H?bond for Arg222. The analysis of the energetic interactions of each mutating residue and its 9 Å surrounding environment are also in accordance to the observed stability. Coulombic potential is less favorable along the S265C mutation, while van der Waals contributions are more favorable along the S126C mutations. In conclusion, this work has contributed to the insight of the effect of two subtle mutations on the stability of a complex system as the ESBL.