De Novo enzyme design of a minimal and soluble carbonic anhydrase from Mesorhizobium loti assisted by molecular...

Carbonic anhydrases (CAs) are critical biocatalysts in the carbon capture and utilization due to their remarkable efficiency in converting carbon dioxide into bicarbonate. Among all, the CA from Mesorhizobium loti (MlCA) exhibits the highest catalytic activity. However, the poor expression and low solubility in Escherichia coli significantly restricts its application. To overcome the challenges, we employed a protein minimization strategy to improve both enzymatic expression and solubility. Enhanced expressions were observed in genetic constructs with short N-terminal tags which MlCA expression is strongly affected by mRNA secondary structure near the start codon. For solubility issue, a de novo protein design workflow guided by molecular dynamics simulations was developed. The process consists of four stages: (1) size-constrained de novo design considered using AlphaFold and RFdiffusion, (2) sequence recovery replying on solubleMPNN and ESMfold, (3) In silico screening by FoldX, SASA and molecular dynamics evaluation, and (4) experimental validation. This enabled the generation of compact, stable, and catalytically active MlCA variants with 28 % reduction in protein size among 500 candidates. Finally, the most promising design, dM22, of minimal CA showed improved solubility from 16.1 % to 61.2 % in B7G, a GroELS integrated BL21(DE3) strain.