Modeling the role of charged residues in thermophilic proteins by rotamer and dynamic cross correlation analysis

Discerning the determinants of protein thermostability is very important both from the theoretical and applied perspective. Different lines of evidence seem to indicate that a dynamical network of salt bridges/charged residues plays a fundamental role in the thermostability of enzymes. In this work, we applied measures of dynamic variance, like the Gini coefficients, Kullback-Leibler (KL) divergence and dynamic cross correlation (DCC) coefficients to compare the behavior of 3 pairs of homologous proteins from the thermophilic bacterium Thermus thermophilus and mesophilic Escherichia coli. Molecular dynamic (MD) simulations of these proteins were performed at 303 K and 363 K. In the characterization of their side chain rotamer distributions, the corresponding Gini coefficients and KL-divergence both revealed significant correlations with temperature. Similarly, a DCC analysis revealed a higher trend to de-correlate the movement of charged residues at higher temperatures in the thermophilic proteins, when compared with their mesophilic homologues. These results highlight the importance of dynamic electrostatic network interactions for the thermostability of enzymes.

Dynamic cross correlation analysis of Thermus thermophilus alkaline phosphatase and determinants of thermostability

BACKGROUND

Understanding the determinants of protein thermostability is very important both from the theoretical and applied perspective. One emerging view in thermostable enzymes seems to indicate that a salt bridge/charged residue network plays a fundamental role in their thermostability.

METHODS

The structure of alkaline phosphatase (AP) from Thermus thermophilus HB8 was solved by X-ray crystallography at 2.1 Å resolution. The obtained structure was further analyzed by molecular dynamics studies at different temperatures (303 K, 333 K and 363 K) and compared to homologous proteins from the cold-adapted organisms Shewanella sp. and Vibrio strain G15-21. To analyze differences in measures of dynamic variation, several data reduction techniques like principal component analysis (PCA), residue interaction network (RIN) analysis and rotamer analysis were used. Using hierarchical clustering, the obtained results were combined to determine residues showing high degree dynamical variations due to temperature jumps. Furthermore, dynamic cross correlation (DCC) analysis was carried out to characterize networks of charged residues.

RESULTS

Top clustered residues showed a higher propensity for thermostabilizing mutations, indicating evolutionary pressure acting on thermophilic organisms. The description of rotamer distributions by Gini coefficients and Kullback-Leibler (KL) divergence both revealed significant correlations with temperature. DCC analysis revealed a significant trend to de-correlation of the movement of charged residues at higher temperatures.

SIGNIFICANCE

The de-correlation of charged residues detected in Thermus thermophilus AP, highlights the importance of dynamic electrostatic network interactions for the thermostability of this enzyme.