Rapid bursts and long-term stasis: phylogeny, evolutionary dynamics, and expression of B3 transcription factors in Arabidopsis, Cannabis, and Humulus

The B3 superfamily is one of the largest plant transcription factor gene families, yet its evolutionary dynamics in dioecious species like Cannabis sativa remain largely unexplored. Here, we present a comprehensive phylogenetic, molecular evolution, and transcriptomic analysis of B3 domain-containing genes in Arabidopsis thaliana, C. sativa, and Humulus lupulus. Using a gene mining pipeline, we identified 110, 101, and 88 B3 genes across the five established subfamilies (ARF, ABI3, HSI, RAV, and REM), respectively. We identified two distinct evolutionary trajectories with the B3 superfamily. On the one hand the REM subfamily was by far the largest and characterized by massive lineage-specific tandem duplications subjected to relaxed selective constraint, suggesting distinct or partially redundant roles in reproductive development. On the other hand, specific non-tandem clades remained highly conserved under strong purifying selection. Detailed analysis of the candidate sex-determination gene CsREM16 showed low dN/dS ratios, clear one-to-one orthology across studied species, and unique persistent sex-biased expression across all six analyzed developmental stages in C. sativa. Comparative analysis of its ortholog HlREM16 confirmed a conserved female-biased expression in H. lupulus. Together, these results demonstrate a dual-mode evolutionary dynamic in the B3 superfamily and particularly in the REM genes and establish CsREM16 as a strong candidate for sex determination in C. sativa, with a potentially conserved role in H. lupulus.