Lac repressor is a DNA-binding protein which inhibits the expression of a series of genes involved in lactose metabolism. Lac repressor can bind at a random DNA site via nonspecific interactions; then, it rapidly translocates through the double chain of DNA until it finds the specific binding site. Therefore, the site transform between these two modes is essential for the specific recognition between Lac repressor and DNA. Here, the recognition mechanism between Lac repressor and DNA was illustrated with molecular dynamics simulations and correlation network analyses. We have found that the correlation network of the specific system (2KEI) is more centralized and denser than that of the nonspecific system (1OSL). The significant difference in the networks between the nonspecific and specific systems is apparently due to the different binding modes. Then, different interaction modes were found where electrostatic and hydrogen bonding interactions in the nonspecific system are stronger than those in the specific system. Hydrophobic interactions were found only in specific complexes and mostly focused on the hinge helices. Furthermore, the hinge helix will induce the bending of DNA for the specific system. At the same time, a common specific sequence of DNA was revealed for three specific systems. Then, two design systems (positive and control) were used to evaluate the specific recognition between DNA and Lac repressor. These combined methods can be used to reveal the recognition mechanism between other transcription factors and DNA.

No abstract

Covariant residues identified by computational algorithms have provided new insights into enzyme evolutionary routes. However, the reliability and accuracy of routine statistical coupling analysis (SCA) are unable to satisfy the needs of protein engineering because SCA depends only on sequence information. Here, we set up a new SCA algorithm, SCA.SIM, by integrating structure information and MD simulation data. The more reliable covariant residues with high‐quality scores are obtained from sequence alignment weighted by residual movement for eight related subfamilies, belonging to α/β hydrolase family, with Candida antarctica lipase B (CALB). The 38 predicted covariant residues are tested for function by high‐throughput quantitative evaluation in combination with activity and thermostability assays of a mutant library and deep sequencing. Based on the landscapes of both activity and thermostability, most mutants play key roles in catalysis, and some mutants gain 2.4‐ to 6‐fold increase in half‐life at 50°C and 9‐ to 12‐fold improvement in catalytic efficiency. The activity of double mutants for A225F/T103A is higher than those of A225F and T103A which means that SCA.SIM method might be useful for identifying the allosteric coupling. The SCA.SIM algorithm can be used for protein coevolution and enzyme engineering research.