Understanding the stability landscape of LbCas12a by deep analysis of stabilizing mutations and mutation combinations.

Abstract

Cas12a is one of the most widely used Cas nucleases for genome editing and in vitro diagnosis. A number of engineered Cas12a mutants have been identified with improved activity and stability. However, it remains largely unaddressed how these mutations interact. In a previous study, we used a deep learning model to evolve the stability of Lachnospiraceae bacterium Cas12a (LbCas12a) and obtained about 90 mutants with improved stability. In the present study, we performed a deep analysis of these stabilizing mutations and mutation combinations to understand the stability landscape of LbCas12a. It was found that most of the stabilized mutants had shifted fitness, as characterized by higher trans-cleavage activity at high temperatures but lower activity at the "fit" temperature for the parent protein. These stabilizing mutations were found to have sophisticated epistatic effects. Stabilizing mutation S962K improved protein stability in the context of other stabilizing mutations but by itself exhibited minor improvements. Saturation mutagenesis of S962 had differential effects on the stability of wild-type (WT) LbCas12a and C10L/I976L/C1090D variant, despite similar melting temperatures (T_m) for WT (41.9°C) and C10L/I976L/C1090D (41.1°C). Interestingly, 12 out of 19 amino acid substitutions at S962 reduced the T_m in the context of WT LbCas12a, while 18 out of 19 mutations increased T_m in the C10L/I976L/C1090D variant. We also showed that stabilizing mutations could recover the stability and trans-activity of a destabilized LbCas12a variant. Our findings can facilitate the understanding of LbCas12a natural evolution and provide insights to developing novel engineering strategies for Cas nucleases.