Both splicing and disorder are more common in multicellular eukaryotes than in lower organisms. Dunker wondered whether this trend might be more than coincidence. Structures are available for only five pairs of alternatively spliced isoforms but, in three pairs, the regions present in one splice form but absent in another are found within disordered regions.
To expand the dataset, the authors compared various databases of disorder and of splicing. Indeed, alternatively spliced regions were strongly biased toward disorder. Their flexibility probably improves the odds that an addition or deletion will not impede folding and thus lead to aggregates.
Splicing in disordered regions might also increase functional diversity. Unlike structured domains, which use far-flung residues to build a single functional unit, disordered regions often use a compact and linear series of residues to create a particular functional unit. “You get more bang for your buck,” says Dunker. “Just splice out ten consecutive residues, and a whole function is gone.”
Disordered domains can evolve quickly, given their structural freedom, and often bind several partners. As disordered regions are commonly signaling and regulatory domains, more disorder and more splicing might have contributed to the emergence of cellular specialization. “With different splicing in different cells,” says Dunker, “the signaling network becomes radically altered.” Splicing out a piece of a disordered region in BRCA1, for example, eliminates p53 binding.
Testing this evolutionary hypothesis, however, will take time. To start, the group would like to work out the major signaling network differences between a pair of similar but distinct cell types, perhaps from a simple multicellular organism like the sponge, and then determine whether the changes relate to alternative splicing within disordered regions.