description |
Environmental microbiotas encode a vast diversity of protein homologs. How this diversity impacts protein function can be explored through selection assays aimed to optimize function. While artificially generated gene sequence pools are typically used in selection assays, their usage may be limited by technical or ethical reasons. Here, we investigate an alternative strategy: the use of a soil microbial DNA as a starting point. We demonstrate this approach by optimizing the function of a widely occurring soil bacterial enzyme, 1-aminocyclopropane-1-carboxylate (ACC) deaminase. We identified a specific ACC deaminase domain region (ACCD-DR) that, when PCR-amplified from the soil, produced a variant pool that we could swap into functional ACC deaminase-encoding plasmids. Functional clones of ACC deaminase were selected for in a competition assay based on their capacity to provide nitrogen to E. coli in vitro. The most successful ACCD-DR variants were identified after multiple rounds of selection by sequence analysis. We observed that previously identified essential active site residues were fixed in the original, unselected library and that additional residues went to fixation after selection. We identified a divergent essential residue that hints at alternative substrates and a cluster of neutral residues that did not influence ACCD performance. Using an artificial ACCD-DR variant library generated by DNA oligomer synthesis, we validated the same fixation patterns. Our study demonstrates that soil microbiotas are useful starting pools of protein-encoding gene diversity that can be utilized for protein optimization and functional characterization when synthetic libraries are not appropriate. |