Structural Insights into CRISPR-Cas12a Target Recognition and Cleavage Mechanisms

Cas12a nucleases offer a single-crRNA-guided, T-rich PAM-dependent alternative to Cas9 with intrinsic pre-crRNA processing activity and staggered double-strand break generation, but the structural basis for how target-DNA mismatches are tolerated or rejected at different positions along the guide-target duplex remains incompletely resolved for the Lachnospiraceae bacterium orthologue (LbCas12a) most widely used in genome editing applications. We determined cryo-EM structures of LbCas12a at four discrete stages of target engagement -- PAM-bound pre-R-loop, partial R-loop (10 bp), complete R-loop with un-activated RuvC domain, and the catalytically engaged cleavage-competent state -- at resolutions between 2.9 and 3.4 angstrom, complemented by single-molecule FRET measurements of R-loop propagation kinetics and cleavage assays across a panel of 32 systematically positioned single-mismatch target variants. The structures show that R-loop propagation proceeds through a previously unresolved kinked intermediate at position 10-12 of the protospacer, coinciding with a conformational checkpoint at which the WED and PI domains must complete a rigid-body rotation before propagation can continue; mismatches placed within this checkpoint region reduced cleavage rate 40- to 90-fold relative to matched targets, far more than equivalently positioned mismatches either proximal or distal to the checkpoint. Single-molecule FRET confirmed that checkpoint-region mismatches stall R-loop propagation at the partial state for a median of 340-fold longer than matched targets, rarely proceeding to complete R-loop formation within the experimental observation window. Mutation of two checkpoint-proximal WED domain residues (R306A/K318A) identified from the structures reduced off-target cleavage at three of four tested genomic off-target sites by 60-85% while retaining greater than 90% on-target activity, without a corresponding rational design being derivable from primary sequence or prior lower-resolution structural data alone. These results identify a discrete structural checkpoint governing Cas12a target discrimination and demonstrate that this checkpoint is a tractable engineering target for improving genome-editing specificity.