From this paper:
http://cshperspectives.cshlp.org/content/6/6/a016071.full.pdf+html
Origin of Spliceosomal Introns and Alternative Splicing
Manuel Irimia and Scott William Roy
ORIGIN AND ESTABLISHMENT OF THE SPLICEOSOMAL SYSTEM DURING EUKARYOGENESIS
In this section, we discuss the major steps leading to the origin and establishment of the spliceosomal system in eukaryotes. Despite the persistence of disagreement on some important aspects, there is a general consensus about how this process may have unfolded (Fig. 2). According to this general model, spliceosomal introns evolved from invading group II introns, perhaps derived from the early mitochondrion (thought to be descended from an engulfed member of the a-proteobacteria, whose modern members contain group II introns). For some reason(s), these introns then proliferated to an unprecedented level in the host genome. Over time, the self-splicing activities of these many intron copies degenerated, which was associated with the increase of trans-encoded RNAs and proteins that promoted efficient intron splicing, setting the basis of the protospliceosomal machinery, and further releasing selective pressure on cisintronic splicing elements. As this protospliceosomal machinery recruited more proteins and became more efficient, introns became increasingly reliant on the emerging spliceosome for proper splicing.
The funny thing is, I reasoned just above that this is pretty much what must have happened simply by understanding the basics of catalyzed RNA hydrolysis and ligation. That you start with self-splicing RNA introns and then gradually build the more complex protein-dependent versions on top.
Turns out that this somewhat ad-hoc rationalization of mine (which would seem to predict that more complex protein-dependent intron splicing evolved from self-splicing RNA's) actually has strong empirical support from phylogenetics(from the same paper):
Transfer of Group II Introns to the Host Genome
Structural and functional evidence suggests that spliceosomal and group II self-splicing introns are evolutionarily related. Both types of introns are spliced through a similar two-step catalytic reaction that relies on an endogenous adenosine (the BP), and releases the excised intron as a lariat structure. The two intron types have similar boundary sequences (GT-AY in group II introns and usually GT-AG in spliceosomal introns, although some U12 introns are AT-AC) (Lambowitz and Zimmerly 2011), and there are striking structural similarities between key regions of group II intron domains and spliceosomal snRNAs (Lambowitz and Zimmerly 2011). These include at least (i) domain DV and U6 snRNA, with divalent metal-ion binding sites involved in catalysis and similar base-pairing interactions (Jarrell et al. 1988; Peebles et al. 1995; Yu et al. 1995; Abramovitz et al. 1996; Konforti et al. 1998; Yean et al. 2000; Shukla and Padgett 2002), further supported by crystal structure (Toor et al. 2008; Keating et al. 2010); (ii) ID3 subdomain and d –d0 motifs and U5 snRNA stem loop, involved in the recognition of 50 and 30 exons (Hetzer et al. 1997); and (iii) DVI and the U2-intron pairing that include the BP adenosine (Schmelzer and Schweyen 1986; Parker et al. 1987; Li et al. 2011). In addition, it
has also been shown that extracts of snRNAs can catalyze both splicing reactions without proteins in vitro (Valadkhan et al. 2007, 2009), in a similar manner to complete group II introns. Whereas it is theoretically possible that group II introns evolved from spliceosomal introns (or that both share a distinct common ancestor [Vesteg et al. 2012]), it seems much more likely that group II introns gave rise to spliceosomal introns (Cech 1986). The presence of spliceosomal introns in all extant eukaryotic supergroups indicates that this transformation occurred before LECA.
Understanding the basics implies(predicts it), phylogenetics supports it. Sorry, game over.