Intron self-complementarity enforces exon inclusion in a yeast pre-mRNA

Abstract

Skipping of internal exons during removal of introns from pre-mRNA must be avoided for proper expression of most eukaryotic genes. Despite significant understanding of the mechanics of intron removal, mechanisms that ensure inclusion of internal exons in multi-intron pre-mRNAs remain mysterious. Using a natural two-intron yeast gene, we have identified distinct RNA-RNA complementarities within each intron that prevent exon skipping and ensure inclusion of internal exons. We show that these complementarities are positioned to act as intron identity elements, bringing together only the appropriate 5' splice sites and branchpoints. Destroying either intron self-complementarity allows exon skipping to occur, and restoring the complementarity using compensatory mutations rescues exon inclusion, indicating that the elements act through formation of RNA secondary structure. Introducing new pairing potential between regions near the 5' splice site of intron 1 and the branchpoint of intron 2 dramatically enhances exon skipping. Similar elements identified in single intron yeast genes contribute to splicing efficiency. Our results illustrate how intron secondary structure serves to coordinate splice site pairing and enforce exon inclusion. We suggest that similar elements in vertebrate genes could assist in the splicing of very large introns and in the evolution of alternative splicing.

Figure 1

Splicing phenotype of YL8A mutants suggests an intron definition mechanism. (A) Intron–exon structures of the known multiply interrupted genes in S. cerevisiae. (B) Structure of L8A-CUP1 expression constructs used to test cis-acting factors in YL8A splicing. The sequences shown are at the junction of the second exon and intron. Wild-type (WT) and mutant (WK5′; X-5′) 5′ splice sites were tested (mutant residues are underlined). (C) The 5′ splice site mutations do not induce exon skipping in vivo. Splicing was analyzed by reverse transcription of total cell RNA using a 5′-labeled (³²P) primer complementary to CUP1 sequences. Lane m, DNA size markers. E1E3 is a marker for exon skipping expressed from a construct in which exons 1 and 3 are directly fused. Expected products are diagrammed at the right. scr1 is a small cytoplasmic RNA used as an internal control for total RNA amount. A strong RT stop (○) correlates with the 5′ end of snR39, a small nucleolar RNA encoded in intron 2 (24) and could be generated by processing events related to snR39 biosynthesis.

Figure 2

Compatibility of YL8A splice sites and influence of internal exon sequences on exon inclusion. (A) Structure of mutant substrates. Single intron substrates with chimeric introns of 305 and 692 nucleotides (deleted regions are depicted as gaps) and 1,020 nucleotides (the inverted region spans exon 2) are shown. The two-intron miniE2 substrate lacks 72 of 94 nucleotides from the internal exon. (B) Splicing phenotypes of the mutant constructs. Splicing was analyzed by reverse transcription of total cell RNA using a 5′-labeled (³²P) primer complementary to CUP1 sequences. Lane m, DNA size markers. E1E3 is a marker for exon skipping expressed from a construct in which exons 1 and 3 are directly fused. Expected products are diagrammed at the right. The different unspliced pre-mRNAs are indicated by an asterisk (∗). Spliced products were measured relative to scr1; the amount of exon included mRNA in wild type was taken as 100%.

Figure 3

Intron self-complementarities ensure exon-inclusion. (A) Sequences of intron complementarities and mutant derivatives. Intron complementarities between regions downstream of the 5′ splice site and upstream of the branchpoint in introns 1 and 2 (designated A/B and C/D, respectively), and their predicted secondary structures. Additional secondary structure is predicted for the regions internal to each intron (data not shown). Mutations (X, Y, and ¥) that replace intron complementarities are shown next to the affected element. Complementarity exists between X/Y and between X/¥. For each intron, 5′ and 3′ splice site guanosine and branchpoint adenosine residues are indicated. Intron sequences are numbered 1–1 through 1–458 for intron 1, and 2–1 through 2–468 for intron 2. (B) Predicted effect of the mutations on substrate secondary structure. Presence or absence of pairing is shown schematically for each construct. (C) Splicing phenotypes of the mutant constructs. Splicing was analyzed by reverse transcription of total cell RNA using a 5′-labeled (³²P) primer complementary to CUP1 sequences. Lane m, DNA size markers. E1E3 is a marker for exon skipping expressed from a construct in which exons 1 and 3 are directly fused. Expected products are diagrammed at the right. The different unspliced pre-mRNAs are indicated by an asterisk (∗). Spliced products were measured relative to scr1; the amount of exon included mRNA in wild type was taken as 100%.

Figure 4

Intron complementarities and splicing. (A) Intron complementarities (inverted arrows) coordinate splice site pairing in yeast YL8A. Open arrows represent one complementarity and filled arrows represent a distinct complementarity. Pairing between complementarities defines introns and enforces exon inclusion. (B) New complementarity within a splicing substrate induces a new pattern of splicing. In YL8A a new engineered complementarity (arrows) causes exon skipping as well as allowing residual exon inclusion (Fig. 3). In higher organisms, the introduction of new complementarity by the transposition of a mobile repeat element (arrows) is proposed to induce an alternative splicing pathway and increase the coding potential of the genome (see text).