A U1-U2 snRNP interaction network during intron definition

Abstract

The assembly of prespliceosomes is responsible for selection of intron sites for splicing. U1 and U2 snRNPs recognize 5' splice sites and branch sites, respectively; although there is information regarding the composition of these complexes, little is known about interaction among the components or between the two snRNPs. Here we describe the protein network of interactions linking U1 and U2 snRNPs with the ATPase Prp5, important for branch site recognition and fidelity during the first steps of the reaction, using fission yeast Schizosaccharomyces pombe. The U1 snRNP core protein U1A binds to a novel SR-like protein, Rsd1, which has homologs implicated in transcription. Rsd1 also contacts S. pombe Prp5 (SpPrp5), mediated by SR-like domains in both proteins. SpPrp5 then contacts U2 snRNP through SF3b, mediated by a conserved DPLD motif in Prp5. We show that mutations in this motif have consequences not only in vitro (defects in prespliceosome formation) but also in vivo, yielding intron retention and exon skipping defects in fission yeast and altered intron recognition in budding yeast Saccharomyces cerevisiae, indicating that the U1-U2 network provides critical, evolutionarily conserved contacts during intron definition.

Fig 1

Identification of SpPrp5-associated proteins. (A) Schematic of S. pombe Prp5 protein, indicating regions previously shown to be sufficient for interaction with U1 and U2 snRNPs (42). (B) Strategy for identification of protein interaction partners of SpPrp5. ProtA, protein A. (C) Silver-stained gel to visualize proteins that copurified with SpPrp5. An untagged extract was used as a negative control. The two large bands in each lane indicated by an asterisk to the right of the gel are the heavy and light chains of IgG, derived from the resin. The positions of molecular weight markers (in thousands) are indicated to the left of the gel.

Fig 2

Rsd1 mediates SpPrp5 interaction with U1 snRNP. (A) In vitro protein-protein interactions indicate that SpPrp5 binds directly to Rsd1, but not to core U1 proteins. GST-tagged proteins were incubated with ³⁵S-labeled in vitro translation products and purified using glutathione-Sepharose. Lanes: In, 1/6 of total input; Pe, copurifying proteins. (B) GST-Rsd1 binds directly to U1A, but not to SpSnu71 (left); GST-Snu71 binds to U1A, but not to Rsd1 (right). (C) Schematic of proposed interaction network between SpPrp5 and U1 snRNP mediated by Rsd1. (D) Rsd1 facilitates SpPrp5 interaction with U1 snRNP. SpPrp5 or Rsd1 was depleted from extract by IgG-Sepharose binding under high-salt conditions (Western blotting), and then the extract was incubated with ATP and GST-SpPrp5 followed by affinity selection. Copurifying snRNAs were analyzed by Northern blotting. The U1-to-U2 snRNA ratio was normalized to the amount of U1 and U2 snRNAs in each TAP-tagged extract. (E and F) SpPrp5 and Rsd1 interact through their RS and RS-like domains. GST pulldowns were performed as described above fir panel A. WT-SpPrp5, SpPrp5-AAAA₃₀₆, and SpPrp5-RSΔ (amino acids [aa] 1 to 172 deleted) were expressed as GST-tagged proteins (bait); Rsd1 full-length and truncated proteins, RS domain (aa 1 to 240), RRM123 (aa 235 to 604), RRM12 (aa 235 to 419), RRM3 (aa 413 to 604), and SRp2 were translated in vitro with ³⁵S labeling.

Fig 3

The DPLD motif and SF3b mediate SpPrp5 association with U2 snRNP. (A) Recombinant GST-SpPrp5 interacts with the SF3b complex partially purified from SF3b145-TAP extract. Signal of SF3b155, detected by Western blotting using anti-SF3b155 antibody, is shown as a representative of the SF3b complex. Pulling down with GST alone was used as a negative control. (B) Interaction between SpPrp5 and SF3b is RNA independent. Recombinant GST-SpPrp5 can pull down the SF3b complex both in the presence (+) and absence (−) of snRNAs, which were detected by Northern blotting. (C) Alanine mutations in the DPLD motif disrupt SpPrp5 interaction with SF3b49, SF3b130, and SF3b145, whereas mutation of the SAT motif in the ATPase domain has no effect on interaction with SF3b proteins. SF3b subunits were either TAP or Flag tagged as indicated and detected by Western blotting. (D) Phylogenetic comparison of the conserved DPLD motif in Prp5. Sequences from Schizosaccharomyces pombe, Homo sapiens, Xenopus laevis, Danio rerio, Drosophila melanogaster, Arabidopsis thaliana, Dictyostelium discoideum, Aspergillus fumigatus, Kluyveromyces lactis, and Saccharomyces cerevisiae are shown. (E) Alanine mutations at D303 and L305 inhibit SpPrp5 interaction with SF3b155 and U2 snRNP. Recombinant GST-tagged proteins containing mutations in the DPLD motif of SpPrp5 were incubated with S. pombe extract, then selected, and assayed for interaction with SF3b proteins (indicated by Western blotting for SF3b155) and U2 snRNP (indicated by Northern blotting for snRNAs). (F) Mutations in the DPLD motif inhibit assembly of prespliceosomes. Recombinant SpPrp5 proteins and ³²P-labeled pre-mRNA substrate were incubated with S. pombe extract depleted of endogenous SpPrp5; their abilities to form prespliceosomes (complex A) were analyzed by 4% native gel.

Fig 4

Prp5-DPLD motif mutants yield intron retention and exon skipping in S. pombe. (A) In vivo, mutations at D303 and L305 in the DPLD motif of SpPrp5 affect the growth of fission yeast. (Top) Strategy; (middle) tetrad dissection and growth on G418 (only Kan⁺ cells grow; wild-type Kan⁻ cells leave a faint background of dead cells; for the AAAA mutant, two of the spots are blank, because this mutant is inviable); (bottom) temperature growth assay. (B) RT-PCR analysis reveals that D303A and L305A mutants of Prp5 inhibit pre-mRNA splicing of intron-containing genes, yielding intron retention. Expression levels of intronless genes were not affected. P/P+M, precursor/(precursor + mature), is an estimate of the fraction of unspliced RNA for various transcripts. (C) D303A and L305A mutants of Prp5 trigger skipping of exon 2 of the cdc2 gene. RT-PCRs were tested by various sets of primers across five exons of cdc2, and the PCR products were confirmed by sequencing. (D) D303A and L305A (but not P304A) mutants of Prp5 inhibit pre-mRNA splicing of multiple-intron-containing genes, yielding intron retention, as shown here by increased levels of pre-mRNA and decreased levels of mRNA (most notably for erf1 and SF3b155, but decreased effects for other genes like cgs2 and nda3). D303A and L305A mutants of Prp5 result in skipping of exon 2 of the cdc2 gene (panel C), but no detectable exon skipping for four other multiple-intron-containing genes. RT-PCRs were performed using primers listed in Table 2.

Fig 5

Prp5-DPLD motif mutants modulate substrate selectivity of suboptimal branch regions in S. cerevisiae. (A) Schematic of ACT1-CUP1 pre-mRNA, indicating intron mutations at 5′SS, BS, and 3′SS used in panel B. (B) Analysis of prp5-DPLD mutant alleles that alter splicing of branch region mutants. Graphs of maximum copper concentration tolerated (top) and growth on selected copper plates (bottom) are shown. Previously described prp5-N399D and -TAG448 alleles (25) were tested for comparison. prp5-DPLD mutants improved the copper tolerance of branch region mutants U257C and A258C that decrease pairing with U2 snRNA but do not alter splicing and growth on copper of 5′SS, 3′SS, or branch site C or G mutants. The presence of additional base pairs between the branch region and U2 snRNA (25) abrogates the effects of both DPLD mutants and ATPase mutants on the U257C branch region mutation. max, maximum. (C) Additional potential base pairs between U2 snRNA and the intron branch region partially suppress the U257C defect; prp5 alleles provide no additional improvement. (Top) Schematic of base pairing interactions between U2 snRNA and the intron branch region, indicating BS-U257C ACT1-CUP1 reporter mutation, which is improved by prp5 alleles, and (bottom) BS-U257C plus five additional base pairs to U2 snRNA, which is not improved by prp5 alleles, shown in panel B.

Fig 6

Protein interaction network for U1-U2 snRNP communication during intron specification and prespliceosome assembly. (A) During intron definition, the U1 snRNP core protein U1A binds to an SR-like protein, Rsd1. Rsd1 also contacts SpPrp5, mediated by SR-like domains in both proteins. SpPrp5 then contacts U2 snRNP through SF3b, mediated by a conserved DPLD motif in Prp5. The ATPase domain of Prp5 is not required for these protein-protein interactions but instead is required for the remodeling of U2 snRNP for its stable binding with the branch site. (B) Prp5 contributes to communication between the 5′SS and branch region, helping to define the intron (left). Loss of Prp5-SF3b interaction results in failure of U2 snRNP to engage the intron, leading to intron retention (right). (C) In multi-intron genes, loss of Prp5-SF3b interaction at a weak PPT can result in a U1 connection to U2 snRNP at a downstream branch region, resulting in exon skipping (right). The asterisk indicates a prp5 mutant that is impaired or slow in interactions with U2 snRNP.