Perturbation of transcription elongation influences the fidelity of internal exon inclusion in Saccharomyces cerevisiae

Abstract

Unknown mechanisms exist to ensure that exons are not skipped during biogenesis of mRNA. Studies have connected transcription elongation with regulated alternative exon inclusion. To determine whether the relative rates of transcription elongation and spliceosome assembly might play a general role in enforcing constitutive exon inclusion, we measured exon skipping for a natural two-intron gene in which the internal exon is constitutively included in the mRNA. Mutations in this gene that subtly reduce recognition of the intron 1 branchpoint cause exon skipping, indicating that rapid recognition of the first intron is important for enforcing exon inclusion. To test the role of transcription elongation, we treated cells to increase or decrease the rate of transcription elongation. Consistent with the "first come, first served" model, we found that exon skipping in vivo is inhibited when transcription is slowed by RNAP II mutants or when cells are treated with inhibitors of elongation. Expression of the elongation factor TFIIS stimulates exon skipping, and this effect is eliminated when lac repressor is targeted to DNA encoding the second intron. A mutation in U2 snRNA promotes exon skipping, presumably because a delay in recognition of the first intron allows elongating RNA polymerase to transcribe the downstream intron. This indicates that the relative rates of elongation and splicing are tuned so that the fidelity of exon inclusion is enhanced. These findings support a general role for kinetic coordination of transcription elongation and splicing during the transcription-dependent control of splicing.

FIGURE 1.

Cloning and in vivo expression of dyn2-Cup1 constructs. (A) PCR amplification and cloning of DYN2 sequences into CUP1 expression vectors. The 5′ and 3′ PCR primers (L and R, respectively) introduce BamHI (b) and KpnI (k) sites for cloning. Lengths of the exons (filled boxes represent coding regions) and introns (lines) are indicated in the “pre-mRNA”. Variability in the DYN2 third exon length (see Materials and Methods) enabled us to use different constructs as Cup1 protein expression reporters for exon inclusion and skipping splicing. In schemes i–iii, full-length or truncated proteins are predicted, depending on the number of coding bases (indicated above) in the exon-included and exon-skipped mRNAs. For example, the pGSC1 plasmids allow Cup1p expression only from the exon included mRNA (scheme i), whereas exon skipped mRNA encodes a truncated polypeptide due to translation termination early in the CUP1 coding region. In contrast, the pGSC3 plasmids (scheme ii) permits Cup1p expression only from mRNAs lacking the internal exon, whereas exon-included mRNAs encode truncated polypeptide. Scheme iii was used for constructs derived from the low/moderate stringency screen described in the text and in Figure 2 ▶. (B) RT analysis of precursor and spliced mRNAs from wild type and mutant dyn2-Cup1 expression constructs (scheme ii translational reading frame). Sequences of the mutated splice sites are indicated (changes from the genomic sequence are shown in lowercase). WT indicates wild-type sequence. Expected products are as follows: P, pre-mRNA containing both introns; I1+I2−, intron 1 retained and intron 2 removed; I1−I2+, intron 1 removed and intron 2 retained; E2+, exon 2 included; E2−, both introns removed as a single intron and exon 2 skipped; and m, pUC13 Sau3AI DNA size markers. Reactions were normalized to cDNA derived from reverse transcription of SCR1, an RNAP III transcript.

FIGURE 2.

Cis-acting exon skipping mutations. (A) Schematic of positions altered by mutation. Each bar represents a base position changed within a unique mutant sequence, and the height of the bar indicates the frequency with which a change at that position was identified in the library (many of the sequences carried multiple mutations). Dark bars indicate mutations within the conserved splicing signals. Shown in detail is the branchpoint region in intron 1 (numbering relative to the conserved branchpoint sequence). Changes are indicated beneath the wild type sequence. X indicates XhoI site used for subcloning. (B) Primer extension analysis of exon skipping mutants. Single base mutants derived from the high CuR selection (left gel, lanes 2–10) and low/moderate Cu^R selection (right gel, lanes 12,13) are shown. Products were normalized to SCR1 cDNA and are labeled as described in Figure 1B ▶.

FIGURE 3.

The U2 C41G mutation promotes exon skipping. The U2 snRNA/branchpoint interaction sequence with the bulged adenosine branch residue is shown. Wild-type U2 snRNA is present in lanes 1, 2, and 4 and is absent in lanes 3 and 5. Figure labels are as described in Figure 1B ▶.

FIGURE 4.

Mutations in RNAP II subunit RPB2 rescue exon inclusion. (A) Splicing of dyn2 U[−2]C. (B) Splicing of dyn2 A[2]G. All reactions were normalized to the E2⁺ signal before loading on the gel, so that levels of skipping can be evaluated by observing changes in the amount of the E2⁻ signal. (C) Quantitation of dyn2 cDNA signals and calculation of alternative splicing ratios (±SEM; n = 3 for U[−2]C and n = 4 for A[2]G). Each dyn2 cDNA was measured and is expressed as the percentage of the total dyn2 cDNA signal in each lane. Measures of the statistical confidences of E2⁺, E2⁻, and E2⁻/E2⁺ values relative to the wild-type RPB2 strain were calculated. Student’s t test was applied to compare E2⁺, E2⁻, or E2⁻/E2⁺ quantities in mutant with the wild-type strain. P values <0.05 indicate that the differences between the compared values are unlikely to be due to sampling error.

FIGURE 5.

Rescue of exon skipping by treatment with inhibitors of nucleotide biosynthesis. (A) Primer extension reactions of RNAs isolated from untreated (lanes 1,2) cells and from cultures treated with 6AU (lanes 3,4) or MPA (lanes 5,6). Reactions were normalized to the signal for SCR1. (B) Statistical analysis of dyn2 cDNA signals and ratios of exon skipping-to-inclusion. The 1-h or 2-h E2⁺, E2⁻, or E2⁻/E2⁺; quantities were compared with the 1-h or 2-h MOCK treatment, respectively. The 2-h MOCK was compared with the 1-h measurement. P value indicates the likelihood that the differences between the compared values are due to sampling error.

FIGURE 6.

A transcriptional block to elongation abrogates TFIIS-dependent exon skipping. (A, upper panel) Presence of TFIIS enhances exon skipping in dyn2 U[−2]C. Strains JAY698 (ppr2Δ) and JAY699 (PPR2) were transformed with pU[−2]C and grown at 30°C. (A, lower panel) Analysis of signals, as described in previous figure legends (n = 3). E2⁺, E2⁻, or E2⁻/E2⁺ values were compared between TFIIS⁺ and TFIIS⁻ strains; P values indicate the likelihood that the differences in these values are due to sampling error. (B) Schematic of the pGSlacIAt template with the lacI operator sequence (lac_o). Yeast strain CMKY21 (ppr2Δ) carrying plasmids pGSlacIAt and a methionine-repressible lacI expression construct were transformed with a high-copy (H) or low-copy (L) PPR2 expression construct or appropriate high-copy or low-copy empty vector (−). (C) Primer extension (upper panel) and graph of exon skipping ratios (lower panel; E2⁻/E2⁺ ±SEM, n = 3).

FIGURE 7.

Differences in levels of exon skipped and included mRNAs are not caused by differences in their relative steady-state stabilities or decay rates. (A) Strains HFY114 (UPF1) and HFY870 (upf1Δ) (Table 1 ▶) carry dyn2-CUP1 plasmids pU[−1]A and pU[−1]A*, which are identical except that they possess translational opposite reading frames (RFs). E2⁺ and E2⁻ levels (percentage of all dyn2 RNAs) and alternative splicing ratios (E2⁻/E2⁺) were calculated (±SEM, n = 3). Probability values (P) comparing E2⁺ or E2⁻, or E2⁻/E2⁺ from both reading frames were determined using Student’s t test. P values >0.05 signify the compared values are not significantly different. (B) 1,10-Phenanthroline treatment of strain KH51B4 expressing pU[−2]C. Because the drug blocks nascent transcription by all DNA-dependent RNA polymerases, absolute decay rates cannot be measured (equivalent amounts of SCR1 cDNA were loaded in each lane except lane 12). Rather, relative decay rates of the E2⁺ and E2⁻ mRNAs were measured by taking the E2⁻/E2⁺ ±SEM ratio (n = 4). P values were determined by comparing E2⁻/E2⁺ values in each lane to lane 1. Time points for lanes 1–12, respectively, are (in minutes): 0, 3, 6, 10, 15, 20, 30, 40, 60, 90, 120, 180.

FIGURE 8.

A model showing how relative rates of elongation and intron recognition can lead to exon skipping or inclusion. Factors that affect the relative rates of spliceosome assembly and transcript elongation will influence the amount of time available for establishment of splice site pairing as in (A), which will lead to exon inclusion. Slow recognition of the first intron or rapid transit of polymerase through the second intron provides a greater opportunity for exon skipping, as in (B).