Mutations in the Caenorhabditis elegans U2AF large subunit UAF-1 alter the choice of a 3' splice site in vivo

Abstract

The removal of introns from eukaryotic RNA transcripts requires the activities of five multi-component ribonucleoprotein complexes and numerous associated proteins. The lack of mutations affecting splicing factors essential for animal survival has limited the study of the in vivo regulation of splicing. From a screen for suppressors of the Caenorhabditis elegans unc-93(e1500) rubberband Unc phenotype, we identified mutations in genes that encode the C. elegans orthologs of two splicing factors, the U2AF large subunit (UAF-1) and SF1/BBP (SFA-1). The uaf-1(n4588) mutation resulted in temperature-sensitive lethality and caused the unc-93 RNA transcript to be spliced using a cryptic 3' splice site generated by the unc-93(e1500) missense mutation. The sfa-1(n4562) mutation did not cause the utilization of this cryptic 3' splice site. We isolated four uaf-1(n4588) intragenic suppressors that restored the viability of uaf-1 mutants at 25 degrees C. These suppressors differentially affected the recognition of the cryptic 3' splice site and implicated a small region of UAF-1 between the U2AF small subunit-interaction domain and the first RNA recognition motif in affecting the choice of 3' splice site. We constructed a reporter for unc-93 splicing and using site-directed mutagenesis found that the position of the cryptic splice site affects its recognition. We also identified nucleotides of the endogenous 3' splice site important for recognition by wild-type UAF-1. Our genetic and molecular analyses suggested that the phenotypic suppression of the unc-93(e1500) Unc phenotype by uaf-1(n4588) and sfa-1(n4562) was likely caused by altered splicing of an unknown gene. Our observations provide in vivo evidence that UAF-1 can act in regulating 3' splice-site choice and establish a system that can be used to investigate the in vivo regulation of RNA splicing in C. elegans.

Figure 1. uaf-1 gene and proteins.

(A) Genomic structure of the uaf-1 isoforms uaf-1a and uaf-1b (adapted from Wormbase WS189) . The locations of the n4588 missense mutation and the n5222 deletion allele are indicated. Black boxes: coding exons. Open box: 3′ UTR. Positions of start (ATG) and stop codons (TAA) are indicated. SL1 and SL2, splice leaders associated with the uaf-1a transcript . (B) Predicted UAF-1 protein domains encoded by uaf-1a and uaf-1b cDNAs, the position of the T180I change caused by the n4588 mutation and the domains affected by the n5222Δ deletion are shown. RNAi fragment: the portion of the uaf-1a cDNA used within a dsRNA-expressing plasmid for RNAi. RS: Arginine-Serine rich domain. W: U2AF small subunit-interacting domain. RRM: RNA recognition motif. UHM: U2AF homology motif.

Figure 2. uaf-1(n4588) dramatically alters unc-93(e1500) but not unc-93(+) exon 9 splicing.

(A) Genomic structure of the unc-93 gene. Exons of unc-93 are indicated by black boxes and introns by thin lines. The part of exon 9 that was removed by alternative splicing in uaf-1(n4588) unc-93(e1500) animals is marked as a red box. Exonic primers (arrows) flanking each intron were used to amplify mRNA regions of each exon-exon junction. The red primer pair was used to identify the alternative splice products of unc-93(e1500) exon 9. (B) RT–PCR experiments to examine the splicing of unc-93 mRNA. En-En+1 (arrows) indicates the junction of two adjacent exons. Genotypes of each sample are indicated at the top. Small arrowheads indicate primer dimers, which form variably in RT–PCR experiments and are template-independent. (C) A weak alternatively spliced unc-93 exon 9 was detected in unc-93(e1500) animals (lane 3, lower arrow). In uaf-1(n4588) unc-93(e1500) animals (lane 4, lower arrow), this alternative spliced product was dramatically enhanced. The diagrams on the left illustrate the splicing events responsible for the generation of each band. The upper band reflects the splicing seen for wild-type unc-93, while the lower band reflects the use of a cryptic 3′ splice site generated by the unc-93(e1500) missense mutation. Red arrows indicate the positions of PCR primers. Genotypes of each sample are listed at the top. (D) Reducing UAF-1 levels with RNAi did not cause increased splicing at the cryptic 3′ splice site found in unc-93(e1500) exon 9. (E) Partial genomic sequences of unc-93 intron 8 (lowercase letters) and exon 9 (uppercase letters) in the wild type (above) and in unc-93(e1500) mutants (below). The G-to-A nucleotide change of the e1500 mutation is indicated with an arrow. The AG sequence (red) forms a cryptic 3′ splice site that is recognized by the splicing machinery in uaf-1(n4588) animals.

Figure 3. sfa-1 gene and protein.

(A) Predicted sfa-1 gene structure (adapted from Wormbase WS189) . The locations of the n4562 nonsense mutation and the n5223 deletion allele are indicated. SL2: splice leader associated with the sfa-1 transcript . (B) Partial sequence alignment of the conserved zinc finger domains of SF1/BBP orthologs from C. elegans, D. melanogaster and human. The amino acid numbers, the accession numbers and the Cys458Opal mutation are indicated. (C) Total RNAs were prepared from animals with the indicated genotypes, and RT–PCR experiments were performed to detect the splicing of unc-93 exon 9. sfa-1(n4562) did not increase the alternative splicing of unc-93(e1500) exon 9. Genotypes are listed at the top. (D) Total RNAs were prepared from wild-type or unc-93(e1500) animals treated with control or sfa-1 RNAi, and RT–PCR experiments were performed to detect the splicing of unc-93 exon 9. Reducing SFA-1 by RNAi did not increase the alternative splicing of unc-93(e1500) exon 9. RNAi treatments are listed at the top.

Figure 4. Taqman Real-time PCR quantification of unc-93 alternative splicing.

(A) Location of PCR primers (red arrows) and Taqman probes (dark short lines) for the quantification of unc-93 exon 9 splicing. (B) Proportion of alternatively spliced exon 9 expressed as a percentage of total spliced (wild-type and alternative splice products) unc-93 exon 9 in animals of different genotypes and animals treated with RNAi targeting the indicated genes. For every genotype except the RNAi-treated samples, each data set represents the average value of duplicate measurements of two biological replicates. For the RNAi-treated samples, each data set represents the average value of duplicate measurements of one biological sample. Error bars: standard errors of two biological replicates.

Figure 5. uaf-1 mutations define a UAF-1 region that affects the recognition of different 3′ splice sites.

Comparison of the sequences of the U2AF large subunit between the RS domain and the first RRM domain from C. elegans, D. melanogaster and human. The mutations caused by the uaf-1(n4588), uaf-1(n5120), uaf-1(n5123), uaf-1(n5125) and uaf-1(n5127) mutations are indicated with arrows. RS: Arginine-Serine rich domain. W: U2AF small subunit-interacting domain. RRM: RNA recognition motif. UHM: U2AF homology motif. Black box: region of UAF-1a that might regulate the recognition of 3′ splice sites.

Figure 6. Nucleotide substitutions at the intron 8 endogenous 3′ splice site and the exon 9 cryptic 3′ splice site affect splicing at these different sites in wild-type and uaf-1(n4588) animals.

(A) Exon structures of transgenes used for 3′ splice site nucleotide substitution analysis. Location of PCR primers (red arrows) and Taqman probes (dark short lines) are indicated. (B–D) List of 3′ splice site nucleotide substitutions and the proportion of alternatively spliced exon 9 expressed as a percentage of total spliced (wild-type and alternatively spliced products) unc-93 exon 9 in wild-type and uaf-1(n4588) animals carrying the corresponding transgenes. Nucleotide bases altered are in red, and bases that remain the same as in the original 3′ splice site are in black. The designated positions (−7 to −1) of each base are indicated in transgene No. 1 of (B). For each transgene, two stable lines were established for both wild-type and uaf-1(n4588) animals, except in cases labeled with *, for which only one stable transgenic line was established. Each dataset represents the average value of duplicate measurements of each biological sample. Error bars: standard deviations. (E) Summary graph of the nucleotide substitution analysis shown in (B–D), indicating the % usage of each 3′ splice site in wild-type and uaf-1(n4588) animals carrying the corresponding transgenes. −, 0 to 1%; +, 1 to 10%; ++, 10 to 30%; +++, 30 to 70%; ++++, 70 to 90%; +++++, 90 to 99%; ++++++, 99 to 100%.