Structural mining and engineering of metagenome-derived Cas12a orthologs expands the CRISPR genome editing and...

CRISPR-Cas12a is a compact, RNA-guided nuclease widely deployed in genome editing and molecular diagnostics, yet its broader utility is limited by suboptimal cis-cleavage efficiency and incompletely defined trans-cleavage behavior. To overcome these constraints, we developed an artificial intelligence-guided structural discovery pipeline powered by AlphaFold2, which identified 1,261 previously uncharacterized Cas12a orthologs. From this set, 21 structurally conserved but sequence-divergent candidates were selected for biochemical characterization. Using structure-informed engineering, we generated PcuCas12a MAX, a high-fidelity variant that achieves genome-editing efficiencies in human cells comparable to the benchmark AsCas12a Ultra while retaining robust activity in murine and porcine systems. In addition, four orthologs (LcoCas12a, FcaCas12a, EsoCas12a, and Mac2Cas12a), when paired with specifically engineered CRISPR RNAs, exhibited distinct single-stranded DNA trans-cleavage signatures. These properties enabled construction of a multiplex CRISPR sensor capable of simultaneously detecting multiple nucleic acid targets. Together, these findings expand the Cas12a endonuclease repertoire and enhance its utility in genome engineering and next-generation diagnostics.