The ubiquitin-like protein Hub1/UBL-5 functions in pre-mRNA splicing in Caenorhabditis elegans

Abstract

The ubiquitin-like protein Hub1/UBL-5 associates with proteins non-covalently. Hub1 promotes alternative splicing and splicing of precursor mRNAs with weak introns in yeast and mammalian cells; however, its splicing function has remained elusive in multicellular organisms. Here, we demonstrate the splicing function of Hub1/UBL-5 in the free-living nematode Caenorhabditis elegans. Hub1/UBL-5 binds to the HIND-containing splicing factors Snu66/SART-1 and PRP-38 and associates with other spliceosomal proteins. C. elegans hub1/ubl-5 mutants die at the Larval 3 stage and show splicing defects for selected targets, similar to the mutants in yeast and mammalian cells. UBL-5 complemented growth and splicing defects in Schizosaccharomyces pombe hub1 mutants, confirming its functional conservation. Thus, UBL-5 is important for C. elegans development and plays a conserved pre-mRNA splicing function.

The majority of eukaryotic genes are interrupted by non-coding introns that are precisely excised out to produce mature RNAs by a process known as RNA splicing. Pre-mRNA exons are either constitutively spliced, becoming part of a single mRNA, or are alternatively spliced, producing mRNA variants [1]. Both splicing processes are carried out by two transesterification reactions by a large ribonucleoprotein complex, the spliceosome [2]. In contrast to humans, where more than 90% of genes are alternatively spliced [3], in Caenorhabditis elegans, ~ 35% of genes undergo alternative splicing, but, strikingly, ~ 85% of the genes are trans-spliced. In this process, a 22 nucleotides (nt) splice leader (SL) RNA coded by distinct gene splices to the 5′ ends of transcripts coded by a different set of genes. Thus, a pre-mRNA undergoing trans-splicing contains a 3′ splice site (SS) but lacks a 5’ SS, which is provided by the SL RNA [SL RNAs resemble small nuclear RNAs (snRNAs)] [4–8]. The SL small nuclear ribonucleoprotein (snRNP) donates the 22-nt SL to the trans-splice site (3’ SS on the pre-mRNA) [9].

The C. elegans genome contains more than 1000 operons, and the resulting polycistronic RNA is cut into monocistronic units by 3′-end formation and trans-splicing. SL1 and SL2 trans-splicing occur in different genes. Non-operon genes and the first gene in an operon are trans-spliced to SL1. However, SL2 trans-splicing is mainly used for splicing polycistronic pre-mRNA [10].

The ubiquitin-like protein Hub1 (also referred to as UBL-5 in C. elegans) is a conserved member of the UBL family in eukaryotes reported to function in pre-mRNA splicing [11–17]. In the budding yeast, Saccharomyces cerevisiae, Hub1 binds directly, but non-covalently, to Hub1 interaction domain (HIND) of the splicing factor Snu66, a component of the U4/U6.U5 tri-small nuclear ribonucleoprotein (tri-snRNP), and promotes excision of introns with weak 5’SS [13,15,16]. It also interacts with the DEAD-box helicase Prp5, a regulator of pre-spliceosome assembly, and activates its ATPase activity, thereby enhancing overall splicing efficiency. Interestingly, Hub1-mediated splicing through the non-canonical 5’SS in SRC1 and PRP5 transcripts also promotes their alternative splicing [15,16,18,19]. Consequently, elevated levels of Hub1 affect spliceosomes’ fidelity and causes aberrant splicing by tolerating suboptimal SSs and branch-point sequence (BPS) [18]. Stress-induced upregulation of Hub1 in S. cerevisiae promotes splicing of introns with non-canonical SSs to improve stress tolerance [20].

While Hub1’s splicing function has been well studied in yeast and cultured human cells, a similar role has not been reported in multicellular animals, including the nematode Caenorhabditis elegans. In this organism, UBL-5 has been reported to function in mitochondrial unfolded protein response (UPRmt) by associating with the transcription factor DVE-1 for inducing the expression of the mitochondrial chaperones HSP-60 and HSP-6. However, pre-mRNA splicing defects could not be observed in ubl-5 mutant worms [21,22]. Notably, both UBL-5 and homologues of the HIND-containing splicing factors Snu66 and Prp38 are present in C. elegans. Thus, UBL-5 might play a role in RNA splicing. It has been independently suggested that UBL-5’s role in UPRmt could be an outcome of its pre-mRNA splicing function [23].

In this study, we have investigated the potential role of C. elegans UBL-5 in pre-mRNA splicing. The ubl-5 mutant animals did not grow beyond the Larval 3 stage, indicating a crucial role of UBL-5 in the animal’s development. By making a splicing-sensitive microarray of a subset of genes and hybridizing total RNA isolated from L3-stage mutant worms, we found an accumulation of intron-containing transcripts, outron-containing transcripts and intercistronic regions in ubl-5 mutant worms. The data presented here suggest that UBL-5 plays a role in cis- and trans-RNA splicing.

Methods summary

Yeast strains described in this study are listed in Tables S3 and S4.

Protocols used for protein purification, protein–protein interaction, Schizosaccharomyces pombe complementation, splicing assays, C. elegans strain maintenance and rescue experiments are described in the Data S1 section.

Results

C. elegans UBL-5 binds to HIND-containing spliceosomal proteins

Hub1 binds to HIND-containing pre-mRNA splicing factor Snu66 to promote alternative splicing in yeast and human cell lines [14,16]. We observed two putative HIND elements in C. elegans proteins; at the N terminus of Snu66/SART-1 and the C terminus of PRP-38 (Fig. S1). To verify whether C. elegans UBL-5 binds to the putative elements in these proteins, we carried out yeast two-hybrid assays. Yeast cells transformed with C. elegans UBL-5 and SART-1 (HIND) constructs were selected, and transformants were grown on plates lacking histidine. The growth on these plates is indicative of the interaction (Fig. 1A). The importance of the salt bridge between Asp22 on UBL-5 and Arg62 on SART-1 for the interaction was confirmed by altering the residues, which abolished the interaction (Fig. 1A). Thus, we infer that the binding of C. elegans UBL-5 to SART-1 is also mediated through a salt bridge, showing a conserved mode of interaction between these proteins in C. elegans. However, yeast cells transformed with C. elegans UBL-5 and PRP-38 (HIND) constructs, lacked growth on plates lacking histidine (data not shown). The absence of two-hybrid interaction between UBL-5 and PRP-38 may be attributed to one or more of the following reasons; PRP-38 (HIND) harbours a weaker UBL-5-binding motif when compared to SART-1 (HIND), possible transient interaction between PRP-38 and UBL-5, technical limitations of the assay.

Fig. 1. Caenorhabditis elegans UBL-5–HIND interaction.

(A) Yeast two-hybrid interaction assay (Y2H). Conserved mode of interaction between UBL-5 and SART-1 through salt bridge forming residues in C. elegans. Plasmids expressing C. elegans UBL-5 and UBL-5(D22A) mutant fused to Gal4-binding domain (BD; with uracil marker) and SART-1 HIND and SART-1 HIND(R62A) mutant fused to activation domain (AD; with leucine marker) were expressed in yeast cells. Transformants were five-fold diluted and spotted on control (-ura -leu) and selective (-ura -leu -his) plates. The expression of the HIS3 reporter gene, which allows growth on -his selection medium, indicates interaction between the two fusion proteins. The UBL-5(D22A) mutant did not bind SART-1 HIND, and SART-1 HIND(R62A) mutant did not bind UBL-5. (B) C. elegans UBL-5 interacts with SART-1 HIND. GST pulldown assays with recombinant GST–SART-1 HIND fusion protein and 6xHis–UBL-5. GST–SART-1 HIND bound UBL-5. GST was used as a negative control. Input represents about one-tenth of the total proteins used in the pulldown. (C) C. elegans UBL-5 interacts with PRP-38 HIND. GST pulldown assays from lysates of bacteria (Escherichia coli) expressing GST–PRP-38 HIND fusion protein and 6xHis–UBL-5. GST–PRP-38 HIND bound UBL-5. GST was used as a control. Input represents about one-tenth of the total proteins used in the pulldown. GST–PRP-38 HIND pulldown samples were immunoblotted with a C. elegans UBL-5 antibody raised in rabbits. (D) UBL-5 co-immunoprecipitates with the splicing factors in worms. 3xFlag–UBL-5-expressing C. elegans cells were immunoprecipitated using an anti-FLAG antibody, and bound proteins were analysed by mass spectrometry (MS). The table shows a list of splicing factors co-purified with UBL-5 (MS analysis of anti-FLAG IP material from untagged worms was used as negative control). Human homologues are shown for illustration. Hs, Homo sapiens; Ce, Caenorhabditis elegans.

To understand UBL-5-HIND interaction further, we performed GST pulldown assays. The recombinant fusion protein of SART-1 (HIND) with GST was able to pulldown UBL-5, which confirmed a direct interaction between them (Fig. 1B). To verify whether C. elegans UBL-5 binds to the putative HIND elements of PRP-38, we carried out the GST pulldown assay using bacterial lysates containing soluble proteins. We were unsuccessful in purifying full-length recombinant GST-PRP-38 (HIND)-containing protein and thus used crude bacterial lysate. GST-PRP-38 (HIND)-containing protein lysate was able to pulldown UBL-5 (Fig. 1C). The interaction was verified further by immunoblotting with a UBL-5-specific antibody (Fig. 1C). Thus, C. elegans UBL-5 interacts with the HIND-containing splicing factors SART-1 and PRP-38, suggesting a potential role of UBL-5 in pre-mRNA splicing.

To test if UBL-5 interacted with the spliceosome or splicing factors in C. elegans, worms expressing 3xFlag–UBL-5 constructs were immunoprecipitated using anti-FLAG beads. The UBL-5 co-immunoprecipitated complex was eluted and analysed by mass spectrometry. We report the identification of different splicing factors enriched in UBL-5-purified complexes (Fig. 1D and Table S1). Certain splicing factors co-purified with UBL-5 suggest its association with spliceosomal factors in vivo in C. elegans. We also detected additional proteins related to other biological processes co-purifying with UBL-5 (Table S2). However, the splicing factors co-purifying with UBL-5 represent the spliceosome only partially, either because of weak/transient associations or technical challenges associated with spliceosome purification from the worms. Nonetheless, both in vitro and in vivo evidence indicate that C. elegans UBL-5 associates with components of the spliceosome.

UBL-5 is essential for the development of C. elegans

UBL-5 is essential for the viability of C. elegans. As ubl-5 mutant worms did not survive after the Larval stage 3 (L3), the mutants were generated by using UV as a mutagen, giving rise to a chromosomal deletion in the ubl-5 that was balanced using a GFP-marked translocation. Hence, heterozygous worms show pharyngeal GFP signals, while homozygous mutant worms showed early larval arrest [24]. This phenotype suggested that UBL-5 activity is crucial at the L3 stage of worm development. We next carried out an expression profile of the ubl-5 gene in a larval stage-specific manner in WT animals by quantitative RT-PCR. The expression of ubl-5 varied in a stage-specific manner; the transcript levels were higher at the L3 stage of development (Fig. 2A). These data indicate that the expression/activity of ubl-5 at the L3 stage may be required for the normal development of worms. Furthermore, to confirm that the lethality of the mutant worms is due to the absence of ubl-5, we expressed a ubl-5 genomic DNA clone in ubl-5 mutant worms. The lethality was rescued (Fig. 2B). These results suggested that UBL-5 is essential for C. elegans development.

Fig. 2. Ubl-5 expression is critical in Caenorhabditis elegans.

(A) ubl-5 expression in the worms. ubl-5 mRNA was monitored in the L1-L2, L2-L3, L3-L4 and L4-adult stages of the worm using real-time PCR. ubl-5 levels are normalized against act-1. L3-L4 worms showed the highest expression of all other developmental stages. (B) Lethality in adult ubl-5 knockout worms. A genomic ubl-5 construct rescued the lethality in ubl-5 mutants. The total number of adult worms surviving from a single parent worm is counted for WT, ubl-5/+, Δubl-5 and ubl-5 rescue lines.

C. elegans ubl-5 mutants show defects in cis- and trans-splicing

To understand the role of UBL-5 in RNA splicing, we designed a splicing-sensitive microarray for a subset of C. elegans genes. The array was rich in neuronal genes that undergo alternative splicing or SL1/SL2-mediated trans-splicing. It contained probes specific for introns, exon–exon junction and mRNA-specific splice variants (as a measure of alternative splicing) for a subset of genes to monitor cis-splicing events (Fig. 3A,B). In addition, various oligonucleotide probes were also included for measuring trans-splicing: outronic probes detected pre-mRNA and trans-spliced junction probes (SL1) detected mature SL1-mRNA. For SL2 trans-splicing, we used 60-bp-long probes, 30 bp upstream and 30 bp downstream of the trans-splice site of the candidate gene, for detecting the pre-mRNAs. The trans-spliced junction probes (SL2) detected mature SL2-mRNA (Fig. 3C).

Fig. 3. UBL-5 is required for cis- and trans-splicing in Caenorhabditis elegans.

(A) Analysis of total RNA from WT and ubl-5 mutant worms using a splicing-sensitive microarray to monitor cis-splicing events. The microarray heat map represents the log2-fold-change values of the mutant samples compared to WT samples for exon–exon ligated (EL) and intron-containing transcripts. Yellow represents accumulation, black denotes no change and blue shows a reduction in signals. From 439 targets analysed for cis-splicing using intron-specific probes, 49 targets showed ≥ 2-fold change (P ≤ 0.05). ≥ 2-fold changes were chosen as a conservative estimate (stringent criteria). Statistical significance was P ≤ 0.05. q-value is not used due to a lower sample number. (B) Analysis of alternative splicing events using the splicing-sensitive microarray. The mutant samples are compared to WT samples for mRNA-specific splice variants (SV). From 570 targets analysed for alternative splicing using splice variant probes, 41 targets showed ≥ 2-fold change (P ≤ 0.05). (C) Analysis of trans-splicing events using splicing-sensitive microarray. The mutant samples are compared to WT samples for outron and SL1 trans-spliced transcripts (J1). From 1267 targets analysed using trans-splicing outron probes, 94 targets showed ≥ 2-fold change (P ≤ 0.05). The mutant samples are compared to WT samples for intercistronic regions containing transcripts (IC) and SL2 trans-spliced transcripts (J2). And, from 156 targets analysed using SL2 trans-splicing intercistronic probes, 9 targets showed ≥ 2-fold change (P ≤ 0.05). (D) C. elegans UBL-5 is required for cis-splicing. Semi-quantitative RT-PCR reveals the accumulation of intron-containing b0350.2b transcripts. The block diagrams (not drawn to the scale) represent exons and introns. Primers are depicted with arrows on exons. RT-PCR of act1 (actin) pre-mRNA is used as a control. (E) Semi-quantitative RT-PCR for outron-containing transcripts. Primers are indicated with arrows. Forward primers were specific to the SL1 sequence (black box) or the outron (red line). Reverse primers were specific to the exons (unfilled box). RT-PCR of act-1 (actin) pre-mRNA is used as a control. (F) Semi-quantitative RT-PCR showing accumulation of intron-containing tos-1 transcripts. The block diagrams (not drawn to the scale) represent exons and introns. 2′ exon is cryptic and provides the 3′ splice site to the preceding intron. Primers are depicted with arrows. RT-PCR of act-1 (actin) pre-mRNA is used as a control. Lanes unrelated to the probe were cropped to assemble the figure. The unedited image is shown in Fig. S2. PCR product sizes are deduced from [38]. (G) Semi-quantitative RT-PCR showing potential role of ubl-5 in trans-splicing. Forward primers were specific to SL1 and SL2 (black boxes) and reverse primers were specific to the exons (unfilled box). RT-PCR of gpd-2 pre-mRNA is used as a control. Relative signal intensities in WT and mutant worms are analysed using IMAGEJ software (National Institutes of Health, Bethesda, MA, USA and the Laboratory for Optical and Computational Instrumentation University of Wisconsin, Madison, WI, USA).

While ubl-5 knockout mutants are lethal in C. elegans, the worms survived until the L3 stage. Therefore, the mutant worms were maintained in a heterozygous state with the help of a balancer chromosome. We collected L3-stage WT and ubl-5 mutant worms, isolated total RNA, reverse transcribed it to double-stranded cDNA, generated labelled cRNA by in vitro transcription of the cDNA, and fragmented the cRNA and hybridized on the microarray. The microarray heat maps in Fig. 3A–C represent fold-change values obtained by comparing mutant samples with WT. Splicing patterns of multiple genes were altered in the ubl-5 mutants, as evidenced by the enhanced accumulation of splice variants, intron-, outron- and intercistronic regions containing transcripts compared to WT worms. We also examined the splicing of b0350.2b (selected candidate) and tos-1 (tos-1 provides a sensitive readout for studying alternative splicing in C. elegans) genes by using RT-PCR experiments [25]. In the case of the b0350.2b transcript, an accumulation of intron-containing products was observed in ubl-5 mutant worms (Fig. 3D). Increased retention of pre-mRNA, skipped 2′ cryptic exon (isoform 1) and skipped exon 3 (isoform 2) of the tos-1 transcript were also observed in the ubl-5 mutant worms (Fig. 3F). The splicing of targets specific for SL1 trans-splicing (rps-22, rpl-22, rps-3 and egal-1) and SL2 trans-splicing (rla-1) was monitored using RT-PCR (trans-splicing of rps-3 and rla-1 was reported previously [26]). We also attempted to detect non-trans-spliced transcripts of rps-22 and rpl-22; interestingly, an accumulation of non-trans-spliced rps-22 pre-mRNA was seen in the ubl5 mutants (Fig. 3E). In general, detecting non-trans-spliced pre-mRNA is quite challenging, presumably due to rapid processing or instability. Furthermore, the amount of trans-spliced rps-3 and egal-1 transcripts was comparatively lower in the ubl5 mutant than in WT animals (Fig. 3G). Nevertheless, the splicing defect in the ubl5 mutant was neither complete nor seen for every transcript. These data suggest some specificity in the UBL-5 targets in C. elegans and its regulatory activity in pre-mRNA splicing. The pre-dominant type of trans-splicing defect is unclear due to the limited number of targets analysed. Moreover, from the targets analysed, we did not observe any specific bias towards a particular kind of splicing defects in ubl-5 mutant worms. Altogether, these data showed pre-mRNA splicing defects in ubl-5 mutant animals, suggesting that UBL-5 is required for selected cis-, trans- and alternative splicing in C. elegans.

C. elegans UBL-5 complemented the splicing and growth defects of S. pombe hub1-1 mutants

We next studied the functional conservation of C. elegans UBL-5 by complementing Hub1 functions in the fission yeast S. pombe (Hub1 is essential for viability and pre-mRNA splicing in S. pombe). For this, we expressed C. elegans UBL-5 in hub1 temperature-sensitive and deletion mutants of S. pombe. The expression of C. elegans UBL-5 rescued the lethality in S. pombe hub1-1 mutants at 37 °C, similar to the complementation by S. pombe Hub1 (Fig. 4A). Interestingly, C. elegans UBL-5 also complemented the lethality of S. pombe hub1Δ mutant at all temperatures (Fig. 4B). These results indicate that C. elegans UBL-5 and S. pombe Hub1 are functionally conserved. We also monitored the complementation of the splicing defects in S. pombe hub1 mutants by the C. elegans homologue. RT-PCR-based splicing assays showed restoration of splicing defects in S. pombe hub1 mutants by the C. elegans UBL-5 protein (Fig. 4C). This ability of C. elegans UBL-5 to functionally complement the growth and splicing defects of S. pombe hub1 mutants implies that the protein’s splicing function might be conserved across the eukaryotic kingdom.

Fig. 4. Caenorhabditis elegans UBL-5 complements growth and splicing defects in Schizosaccharomyces pombe hub1 mutant.

(A) Rescue of temperature sensitivity in S. pombe hub1-1 cells by C. elegans UBL-5. A construct expressing S. pombe hub1 was used as the positive control. The proteins were expressed from the weak nmt81 promoter. Five-fold serial diluted cells were spotted on the indicated media, followed by incubation at 30 and 37 °C for 4 days. (B) C. elegans UBL-5 complements S. pombe hub1Δ lethality. Lethality in S. pombe hub1Δ cells was rescued by expressing S. pombe Hub1 and C. elegans UBL-5 at 30 and 37 °C. Five-fold serial diluted cells were spotted on control or FOA-containing plates followed by incubation at 30 and 37 °C for 4 days. Plasmids were expressed from the weak nmt81 promoter. (C) Rescue of splicing defects in S. pombe hub1-1 cells by C. elegans UBL-5. S. pombe hub1-1 mutant showed accumulation of intron-containing transcripts for mug161 and gnd1. UBL-5 complemented the splicing defect to an extent similar to S. pombe Hub1. Block diagrams represent exons and introns. Primers are depicted as arrows. The PCR bands with genomic DNA template show the size expected from pre-mRNAs.

Discussion

Conserved splicing function of C. elegans UBL-5

Through multiple lines of evidence, we have shown that UBL-5 performs a conserved function of RNA splicing in C. elegans. Similar to yeast and human cultured cells, the non-covalent interactions of Hub1 with HIND-containing splicing factors are also seen with C. elegans proteins. In the budding yeast S. cerevisiae, Hub1 is not essential for viability, possibly because of fewer Hub1-dependent introns [16]. In contrast, the protein becomes essential in S. pombe and human cells because of a larger number of Hub1-dependent introns [12,16,17]. A similar proposition for UBL-5′s involvement in a larger number of pre-mRNA splicing events can also be suggested for the nematode C. elegans, where UBL-5 is also essential for development. Nevertheless, the protein is not critical for all splicing reactions in C. elegans. These observations are consistent with the Hub1/UBL5 function in selected pre-mRNAs splicing in yeast and mammalian cells [14,16,17]. Introns containing non-canonical 5’SS in yeast require Hub1 for splicing [16]. Similar features in C. elegans and mammalian targets would be interesting to identify. Interestingly, C. elegans UBL-5 complemented the lack of Hub1 in S. pombe and rescued the growth and splicing defects of S. pombe hub1 mutants. These observations suggest a conserved splicing function of UBL-5 across eukaryotes.

Splicing defects were missed in ubl-5 knockdown worms in a previous report [21], possibly because those assays missed analysing the right targets. Alternatively, low levels of UBL-5 protein in the knockdown worms used in the study might have been enough for most of its splicing function. Transcripts containing introns, outron and intercistronic regions have generally been challenging to capture, possibly due to their rapid clearance from the worms. We could partially circumvent these problems by using ubl-5 knockout worms at the L3 stage and testing against a larger set of genes.

UBL-5 associates with spliceosomal components

The Hub1–Snu66 interaction is well-studied in S. cerevisiae and human cells [14,16]. Besides the interaction with Snu66, Hub1 makes additional contact with Prp5, a DEAD-box helicase, to promote alternative splicing in S. cerevisiae [15,18]. Caenorhabditis elegans UBL-5 binds both HIND-containing spliceosomal proteins SART-1 and PRP-38. The mode of this interaction through salt bridges also appears to be conserved.

UBL-5 appears to associate with other splicing factors, notably the arginine/serine-rich family proteins, including RSP-1, RSP-2, RSP-3 and RSP-6. It also associates with other splicing factors such as HRPF-1 (hnRNP F homologue) and HRPA-2 (hnRNP A1 homologue). The relevance and mode of these interactions would be interesting to study. We could not detect SART-1 and PRP-38 proteins in the Co-IP experiments. Possible reasons for missing these proteins include their weaker affinities with UBL-5, lower expression levels and technical challenges associated with biochemical purifications from C. elegans.

Trans-acting splicing factors such as SR proteins bind with the regulatory sequences on pre-mRNA for splicing control [27,28]. In contrast, heterogeneous ribonuclear proteins (hnRNPs) bind to regulatory sequences on exons and inhibit splicing by preventing SR proteins from binding to exons [29,30]. UBL-5 interestingly appears to associate with both types of trans-acting splicing factors. Further studies are needed to understand the functional relevance of these interactions.

Human Hub1 has been reported to interact with coilin (a core component of Cajal bodies) and colocalize with Cajal bodies, a subnuclear domain where assembly or modification of spliceosomal components occurs [31]. Therefore, Hub1 might have evolved to perform a broader function in RNA splicing by binding to other spliceosomal proteins in a HIND-independent manner, for example, Prp5 and the SR protein kinases Cdc2/Cdc28-like kinases [32]. Hub1 also associates with proteins through hydrophobic interactions and might function beyond RNA splicing [33].

Potential role of C. elegans UBL-5 in trans-splicing

In C. elegans, the 5′ ends of many pre-mRNAs undergo trans-splicing with either SL1 or SL2 splice leader RNAs, and ~ 85% of the genes undergo trans-splicing [7]. Data from our splicing-sensitive microarray suggest a potential role of UBL-5 in trans-splicing. SR proteins’ function in trans-splicing has been established by in vitro splicing assays. They can recruit U2 snRNP to the branch point of natural trans-splicing substrates [34,35]. SL snRNPs provide the 5′ splice site, and the pre-mRNA provides the trans-splice site (3′ splice site). SL snRNP, U2 snRNP and U4/U6.U5 tri-snRNP assemble into a trans-spliceosome to perform trans-splicing. Extending the observation from S. cerevisiae, where Hub1 interacts with the U1/U2 snRNP factor Prp5 and the tri-snRNP factors Snu66 and Prp38, we hypothesize that UBL-5 might bridge the snRNPs during trans-splicing. Moreover, SR proteins also promote the entry of the U4/U6.U5 snRNP into the cis-spliceosome [36] and are also critical for the formation of catalytically active trans-spliceosome for trans-splicing [37]. UBL-5 complexes with SR proteins (RSP-1, RSP-2, RSP-3 and RSP-6); these proteins might facilitate trans-spliceosomes assembly for trans-splicing selected transcripts in C. elegans.

In conclusion, in most eukaryotes, Hub1-binding sites, HINDs, are located in the homologues of the RNA splicing factor Snu66/SART1. However, the plant Snu66 homologue lacks HIND, but this absence is likely compensated by the HIND in the splicing factor Prp38 [16]. Interestingly, in organisms such as Plasmodium and C. elegans, HINDs are present in the homologues of both Snu66 and Prp38. The implications of more than one splicing factor associating with Hub1 are not yet clear, but it is tempting to suggest that the dual occurrence of HINDs might be linked to the higher prevalence of cis- and trans-splicing in these organisms [9]. Supporting this, C. elegans UBL-5 appears to play a role in cis- and trans-splicing.

Accession codes

Splicing-sensitive microarray data are deposited with GEO (accession number GSE157943). Mass-spectrometry data are deposited with the PRIDE data-base (accession number PXD033914).

Abbreviations

AS

alternative splicing

BPS

branch-point sequence

CBs

Cajal bodies

Co-IP

co-immunoprecipitation

HIND

Hub1 interaction domain

hnRNP

heterogeneous ribonuclear protein

SL

splice leader

snRNAs

small nuclear RNAs

SR

ser/arg proteins

SS

splice-site

UPRmt

mitochondrial unfolded protein response.