Splicing of primary mRNA transcripts occurs co-transcriptionally, but it is unclear to what extent splicing efficiency affects transcription. Beggs and colleagues report that in conditions that block pre-spliceosome complex formation in Saccharomyces cerevisiae, RNA polymerase II (RNAPII) stalls on introns in a Cus2p-dependent manner, which suggests the existence of a splicing-associated transcription elongation checkpoint.

Credit: Jim Zuckerman / Alamy

The formation of pre-spliceosome complexes requires the spliceosomal RNA helicase Prp5p and is facilitated by the U2 small nuclear ribonucleoprotein (snRNP)-associated protein Cus2p, which is then displaced to enable splicing to proceed. Yeast cells with the heat-sensitive prp5-1 mutation fail to displace Cus2p when shifted to the restrictive temperature and are defective in pre-spliceosome formation. To test whether this might affect transcription, the authors performed chromatin immunoprecipitation (ChIP) analysis using RNAPII-specific antibodies along the length of several genes in prp5-1-mutant cells. Shifting mutants to the restrictive temperature resulted in RNAPII accumulation on introns of intron-containing genes but not on intronless genes. RNAPII accumulation at introns was not observed when a different stage of splicing was disrupted, which indicates that it is not a general consequence of defective splicing. Interestingly, the accumulated RNAPII was specifically phosphorylated at Ser5 of its carboxy-terminal domain, which indicates a possible transcription elongation defect in the prp5-1 strain. Next, the authors carried out a genome-wide ChIP followed by sequencing analysis of the prp5-1 strain and found widespread RNAPII enrichment on introns and depletion from downstream exons, which further suggests a defect in transcription elongation.

RNAPII stalling in prp5-1 cells is Cus2p-dependent.

As the requirement for Prp5p to form a pre-spliceosome complex is reduced in the absence of Cus2p, the authors deleted CUS2 from the prp5-1 strain and found for several genes that RNAPII distribution was similar to that in wild-type cells. Reintroducing CUS2 restored RNAPII accumulation on introns at the restrictive temperature, which confirms that RNAPII stalling in prp5-1 cells is Cus2p-dependent.

To confirm a defect in transcription elongation in the prp5-1 strain, the authors measured 4-thio-uracil incorporation into newly synthesized transcripts. They found that the amount of RNA produced in the labelling period from intron-containing genes, but not from intronless genes, was greatly reduced in the prp5-1 mutant strain compared with wild type cells. Furthermore, deletion of CUS2 rescued this defect but not the splicing defects of prp5-1 mutant cells. So, it seems that the prp5-1 mutation reduces the amount of nascent RNA of intron-containing genes in a Cus2p-dependent manner and not simply as a consequence of a splicing defect.

On the basis of their results, the authors propose the existence of a splicing-dependent transcription elongation checkpoint, which is triggered by Cus2p if pre-spliceosome formation is unsuccessful. Interestingly, Cus2p is the putative orthologue of human TAT-specific factor 1 (TATSF1), which is associated with both RNAPII and U2 snRNPs. Whether the dual functions of TATSF1 in transcription and splicing are coupled remains to be investigated.