Core splicing architecture and early spliceosomal recognition determine microexon sensitivity to SRRM3/4

Microexons are essential for the proper operation of neurons and pancreatic endocrine cells, in which their inclusion depends on the splicing factors SRRM3 and SRRM4 (SRRM3/4). However, in pancreatic cells, lower expression of these regulators limits inclusion to only the most sensitive subset among all neuronal microexons. Although various cis-acting elements can contribute to microexon regulation, how they determine this differential dose response and the corresponding high or low sensitivity to SRRM3/4 remains unknown. Here we use massively parallel splicing assays probing 28,535 variants to show that sensitivity to SRRM4 is conserved across vertebrates. Our data support a regulatory model whereby high or low microexon sensitivity is largely determined by the interplay between core splicing architecture and length constraints. This conclusion is further supported by distinct spliceosome activities in the absence of SRRM3/4 and by a mathematical model that assumes that the two types of microexons differ only in their efficiency to recruit early spliceosomal components. Using massively parallel splicing assays and mathematical modeling, Bonnal et al. uncover that conserved splice site strength and exon length encode microexon sensitivity to SRRM3 and SRRM4. This work revises the switch-like splicing regulation model into a dose-responsive continuum.