Volume 6, Issue 5, Pages 1089-1098 (November 2000) The Apoptosis-Promoting Factor TIA-1 Is a Regulator of Alternative Pre-mRNA Splicing  Patrik Förch, Oscar Puig, Nancy Kedersha, Concepción Martínez, Sander Granneman, Bertrand Séraphin, Paul Anderson, Juan Valcárcel  Molecular Cell  Volume 6, Issue 5, Pages 1089-1098 (November 2000) DOI: 10.1016/S1097-2765(00)00107-6

Figure 1 Sequence of the 5′ ss Regions of Transcripts Used in This Study Intronic uridines are in bold, and the exon–intron boundary is represented by a slash. Molecular Cell 2000 6, 1089-1098DOI: (10.1016/S1097-2765(00)00107-6)

Figure 2 Proximity of the Downstream U-Rich Sequence Is Important for msl-2 5′ ss Recognition by U1 snRNP (A) Effect of increasing the distance between the 5′ ss and the U-rich element in spliceosome assembly. Radioactively labeled msl-2 and msl-2 SPA pre-mRNAs were incubated with nuclear extracts (NE) under splicing conditions in the absence or presence of ATP, and the complexes were fractionated by electrophoresis on agarose/polyacrylamide composite gels, which were dried and exposed to film. Complex H indicates hnRNP protein–RNA complexes. Complex A corresponds to the prespliceosome, formed by assembly of U2 snRNP on the 3′ ss branch point sequence. Complexes B/C correspond to two conformations of the fully assembled spliceosome. (B) Effect of increasing the distance between the 5′ ss and the U-rich element in U1 snRNA cross-linking. Radioactively labeled msl-2 and msl-2 SPA 5′ half RNAs were incubated in the absence (−) or presence (+) of either nuclear extracts (NE) or nuclear extracts in which the 5′ end of U1 snRNA was removed by oligodeoxynucleotide-mediated RNase H cleavage (H), in the presence (+) or absence (−) of psoralen (Psor), and irradiated with 365 nm UV light. The RNA was then purified and fractionated by electrophoresis on a denaturing polyacrylamide gel. The position of the U1 snRNA/msl-2 cross-link is indicated. H* represents a reaction in which U1 snRNA was cleaved by RNase H–mediated degradation of the region immediately 3′ to the 5′ end of U1 snRNA that establishes base pairing interactions with the 5′ ss. Disappearance of the U1/msl-2 signal in lane 4 demonstrates that the 5′ end of U1 snRNA is required for the formation of the cross-linked species. Disappearance of the signal in lane 5 demonstrates that the cross-linking species involved U1 snRNA. Molecular Cell 2000 6, 1089-1098DOI: (10.1016/S1097-2765(00)00107-6)

Figure 3 TIA-1 Promotes U1 snRNP Binding to Specific 5′ ss (A) msl-2/U1 snRNA cross-linking depends on poly-U binding proteins. Psoralen-mediated UV cross-linking assays were performed as in Figure 2B using either nuclear extracts (NE) or nuclear extracts depleted of poly-U-binding proteins by chromatography on poly-U Sepharose (NEΔU). The 2 M KCl eluate from the column was also added to the reaction in lane 4. (B) U1 snRNA cross-linking to a 5′ half pre-mRNA containing a consensus 5′ ss does not depend on poly-U-binding proteins. Psoralen-mediated UV cross-linking assays were performed as in (A) using a 5′ half RNA containing a consensus 5′ ss, in the absence or in the presence of standard nuclear extracts (NE) or poly-U Sepharose-depleted nuclear extracts (NEΔU). The identity of the U1 snRNA cross-link was confirmed using RNase H–treated nuclear extracts (see [E]). (C) Recombinant TIA-1 restores U1 snRNA cross-linking to msl-2 5′ half RNA in poly-U-depleted nuclear extracts. Psoralen-mediated UV cross-linking assays were performed and analyzed as in Figure (A), in the absence (−) or in the presence (+) of 7 ng/μl of recombinant purified TIA-1. (D) Recombinant TIA-1 promotes the association of purified U1 snRNP to msl-2 5′ half RNA. Psoralen-mediated UV cross-linking assays were performed and analyzed as in (A), except that nuclear extracts were replaced by partially purified U1 snRNP. (E) TIA-1 does not promote the association of purified U1 snRNP to an RNA containing a consensus 5′ ss. Psoralen-mediated UV cross-linking assays were performed as in (D). TIA-1 was added to a final concentration of 20 ng/μl. H indicates inactivation of the 5′ end of U1 snRNA by oligodeoxynucleotide-mediated RNase H degradation. (F) TIA-1 enhances U1 snRNA cross-linking to msl-2 in nuclear extracts, but only marginally if intronic position +5 is mutated to G (msl-2+G5). Assays were performed as in (C) using 5′ labeled synthetic RNAs containing 10 nucleotides of exonic sequences and 36 nucleotides of intronic sequences. Molecular Cell 2000 6, 1089-1098DOI: (10.1016/S1097-2765(00)00107-6)

Figure 4 TIA-1 Cross-Linking to U-Rich Sequences Downstream of msl-2 5′ ss (A) Pattern of polypeptides cross-linked to msl-2 5′ half pre-mRNA. Reactions set up as in Figure 2A were irradiated with 254 nm UV light, followed by digestion with RNase A, the polypeptides fractionated on an SDS-polyacrylamide gel, and the cross-linked species visualized by autoradiography. #4 and #5 indicate fractions from the 2 M KCl eluate of the poly-U Sepharose column. The sizes of molecular weight markers are indicated to the left. (B) Immunoprecipitation of TIA-1 from the cross-linked species. Cross-linking reactions carried out as in (A) were immunoprecipitated using either ML-29 anti-TIA-1 monoclonal antibody (α TIA-1) or a control monoclonal antibody against a Myc epitope (α ctrl), followed by fractionation of the samples on a denaturing SDS-polyacrylamide gel. The cross-linked proteins were present in the supernatant of the immunoprecipitation reaction with the control antibody, ruling out that the absence of signal was due to protein degradation. Relevant molecular weight markers are indicated. (C) Low levels of TIA-1 cross-linking/immunoprecipitation to AdML 5′ half pre-mRNA. Reactions were set up and analyzed as in (B). (D) Cross-linking of TIA-1 to site specifically labeled msl-2 pre-mRNAs. The location of radioactive phosphates in each of the two mutants is indicated by asterisks. The position of the labeled nucleotide with regard to the exon/intron junction is also indicated. Cross-linking/immunoprecipitation was as in (B). (E) TIA-1 cross-linking requires two stretches of uridine-rich sequences. Assays were performed as in (B) using the indicated RNAs, whose sequences are shown in Figure 1. Molecular Cell 2000 6, 1089-1098DOI: (10.1016/S1097-2765(00)00107-6)

Figure 5 TIA-1 Is Required for Splicing of msl-2 Pre-mRNA Reactions were set up as in Figure 3A using an RNA containing the complete intron of msl-2 5′ UTR and flanking sequences (Gebauer et al. 1998). After incubation at 30°C for 45 min, RNAs were purified and fractionated on a 13% polyacrylamide denaturing gel, which was exposed to film. Because relatively small amounts of spliced products accumulated in this experiment, splicing intermediates (intron-exon 2 lariat intermediate and free exon 1) are shown (schematically represented on the left). The concentrations of recombinant purified TIA-1 (50 ng/μl) and U2AF65 (5 ng/μl) used were optimized in preliminary experiments to provide maximal levels of complementation. Higher concentrations of either or both of the two proteins resulted in equal or lower levels of activity. Molecular Cell 2000 6, 1089-1098DOI: (10.1016/S1097-2765(00)00107-6)

Figure 6 Regulation of Fas Alternative Pre-mRNA Splicing by TIA-1 (A) Nucleotide sequence comparison of the 5′ ss regions of msl-2, Fas exon 5, and Fas mutant (Fas-M) pre-mRNAs. The exon/intron boundaries are indicated by a slash. Intronic uridines are in bold. (B) Alternative splicing patterns of human Fas exon 6 and encoded protein isoforms. The filled circle represents the position of a putative TIA-1-binding site. (C) TIA-1 affects the accumulation of alternatively spliced transcripts from a transiently transfected Fas minigene. Cytoplasmic RNA was isolated from the indicated cell lines transfected with an expression vector containing human Fas genomic sequences from exons 5 to 7. In lane 5, the reporter construct was cotransfected with an expression construct containing TIA-1 cDNA. RNAs were analyzed by quantitative RT-PCR using oligonucleotide primers corresponding to vector sequences and the products analyzed on agarose gels. The sizes of molecular weight markers are indicated on the left, and the sizes expected for the alternatively spliced transcripts are schematically indicated to the right. The identity of the products was confirmed by sequencing. The band indicated by an asterisk is a PCR artifact generated only in the presence of both 5-6-7 and 5-7 amplification products, which has proven impossible to clone and sequence and whose pattern of digestion with restriction endonucleases is not compatible with the presence of Fas exon 6 (data not shown). The intensities of the different bands were quantified using NIH Image software. The numbers inside the bars indicate the values of the ratios between 5-6-7 and 5-7 amplification products. Lane 1: 100 bp molecular weight marker (GIBCO–BRL). (D) TIA-1 cross-links to Fas intron 5 5′ ss region. Cross-linking/immunoprecipitation was performed as in Figure 4B using a radioactively labeled RNA corresponding to exon 5 and 42 nucleotides of intron 5 of human Fas. Consensus 5′ ss is as in Figure 3E. (E) Accumulation of alternatively spliced products upon mutation of the U-rich stretch downstream of Fas intron 5 5′ ss. Transfections were performed and analyzed as in (C) The sequence of the mutant Fas-M 5′ ss region is indicated in (A). Molecular Cell 2000 6, 1089-1098DOI: (10.1016/S1097-2765(00)00107-6)