Multiplex CRISPR/Cas12a editing of ROC5, GS3, GW2, and LARGE2 reveals effects on yield-related traits in rice cultivar Kasalath.

Abstract

Sustainable improvement of rice yield requires coordinated modification of multiple agronomic traits, yet the agronomic outcomes of multiplex genome editing remain difficult to predict. Here, we applied CRISPR/Cas12a-mediated multiplex editing in the indica rice cultivar Kasalath to target six yield-related genes, Gn1a, TAD1, ROC5, GS3, GW2, and LARGE2, using a single Agrobacterium-delivered construct. Sanger sequencing of 28 T₀ plants detected edits in Gn1a (3.57%), ROC5 (7.14%), GS3 (35.71%), and GW2 (10.71%), whereas TAD1 and LARGE2 showed no detectable T₀ edits. Single, double, and triple edits were recovered, and homozygous lines were established for four single mutants, two double mutants, and two triple mutants. Molecular analysis showed that the recovered GW2 alleles were intronic and did not alter the GW2 coding sequence. Phenotypic evaluation across 13 agronomic traits showed that allele type and locus combination, rather than simple additive effects, determined agronomic outcome. The gs3-13d frameshift allele caused strong pleiotropic effects, including increased tillering and panicle number but reduced grain length and yield. The roc5-7d gw2-8d gs3-3i triple mutant maintained wild-type-level yield performance with altered leaf architecture, whereas gw2-8d gs3-3i large2-10d showed increased grain width, grain thickness, and 1000-grain weight but severe reductions in plant height, tillering, seed setting, yield per panicle, and yield per plant. These results demonstrate the feasibility of CRISPR/Cas12a multiplex editing in Kasalath and highlight the importance of allele design, locus combination, and source-sink balance in multiplex genome editing targeting yield-related traits.