Enzyme engineering based on rational variation of amino acid sequence drives the quest for higher activity or selectivity within the seemingly infinite landscape of possible variants. Directed laboratory evolution and iterative saturation mutagenesis (ISM) at specific target sites of an enzyme are efficient approaches but very laborious, as the process of generating mutant libraries by natural evolution requires iterative rounds of gene mutations and benchmarking. Given the experimental hurdles of high-throughput screening, developing more efficient selection methods remains the primary challenge toward smaller, active variant-enriched libraries. Virtual screening based on transition state (TS) calculations can significantly reduce experimental costs, but the associated computational cost is considerable. Aiming to reduce the workload for simulation-based ISM, Jianping Lin, Dawei Zhang, and colleagues propose the use of transition state analogues (TSAs) to predict the most promising targets for site-directed mutagenesis in the laboratory.
TSAs correspond to simplified and computationally less expensive proxies for their complex TS structures that mimic the change in geometry and charge during the reaction. The combination of computational TSA modeling and simulation-based ISM (TSA-CS-ISM) consisted of a three-part strategy: computational modeling of TSA states based on the catalytic mechanism; computational ISM design, in which the calculation evaluates a large library of mutations; and experimental validation. This approach was exemplified on chitinase A1 (BcChiA1) from Bacillus circulans, which catalyzes the hydrolytic degradation of chitin, an abundant renewable biomass raw material. Although the combination of positive single-site mutants is a generally applicable strategy, the experimental data evidenced that the optimal combination of two positions was not necessarily a direct consequence of the best single position variants. This highlighted a core problem of mutation-based enzyme optimization: in the search for beneficial mutations, one tends to be trapped in a local optimal solution space of catalytic activity. By taking advantage of a computational combination library (23,340 mutations in three iterations), the authors achieved breaking out of the local optimal solution space. Guided by energy evaluation and experimental data, the library was reduced from 107 to 102, significantly reducing the workload for experimental validation. In addition, the TSA-CS-ISM strategy offered insight into the correlation between mutation and activity, allowing for an increased proportion of active mutants. Notably, among the 144 samples tested, the enrichment ratio of variants with increased activity reached 83%, including a mutant with 29.3-fold increased activity, the highest reported to date.
This is a preview of subscription content, access via your institution