Nuclear coupling: RNA processing reaches back to transcription

Abstract

790 nature structural biology • volume 9 number 11 • november 2002 Synthesis of messenger RNA (mRNA) in a eukaryotic cell involves numerous successive steps: first transcription initiation and elongation; then RNA processing reactions, such as capping, splicing and polyadenylation; and finally termination of transcription. Remarkably, in the last five years or so, all these reactions have been found to be coupled to one another, such that protein(s) involved in one step also participate directly in another (reviewed in refs 1–5). A key factor in this integration is RNA polymerase II (RNAP II) itself, and specifically the repetitive C-terminal domain (CTD) of its largest subunit. These initially surprising observations now seem to make good sense: what better way to ensure accurate and efficient processing of nascent transcripts in the complex and dynamic nuclear milieu than to involve the transcribing enzyme itself in recruiting and coordinating processing factors? But now several papers suggest that these interactions can also work ‘backwards’, so that splicing and capping factors influence transcription. Although to some degree these findings may have been anticipated years ago, they are mechanistically fascinating and point to additional complexity in the elaborate process of eukaryotic gene expression. The first direct evidence that components of the splicing machinery could influence mRNA transcription came from recent work by Fong and Zhou6, who were studying the role of the transcription elongation factor P-TEFb (a heterodimer of the protein kinase Cdk9 and cyclin T1) in TAT-stimulated HIV transcription. They identified a biochemical fraction in nuclear extracts capable of enhancing transcription that is associated with cyclin T1. This activity, which seemed to enhance transcriptional elongation, but not initiation, was found to contain the previously identified TAT cofactor TATSF1. Intriguingly, TAT-SF1 is similar to the Saccharomyces cerevisiae protein Cus2p, a splicing factor associated with the U2 small nuclear ribonucleoprotein (snRNP)7. Consistent with this, the TATSF1 complex was found to contain U snRNPs, notably U1 as well as U2, and a snRNP-containing complex was shown to stimulate transcription elongation. More generally, transcription from TATindependent templates was also stimulated by TAT-SF1–U snRNP, with the important proviso that an intron with functional splicing signals must be present in the transcribed RNA. The observations are consistent with a model in which the elongation factor P-TEFb recruits the TAT-SF1–snRNP complex to elongating RNAP II holoenzyme, probably involving the RNAP II CTD. Although the mechanism is unclear, this complex is more proficient in transcription elongation in vitro, pointing to a stimulatory role of splicing snRNPs on transcription. These results may also help to explain the converse stimulation of splicing by RNAP II8,9 and indicate that co-transcriptional splicing may increase the efficiency of both processes. Another link between splicing and transcription now comes from an entirely different direction. Motivated by the recent observation10,11 that P-TEFb contains an RNA component, 7SK (an abundant snRNA of previously unknown function), Kwek et al.12 examined the RNAP II transcription machinery for other small RNAs and report, on page 800 of this issue of Nature Structural Biology, that highly purified preparations of one component, TFIIH, contain roughly stoichiometric amounts of U1 snRNA. No snRNP proteins were uncovered, and U1 appeared to be associated specifically with the cyclin H subunit of TFIIH. By comparing U1depleted and reconstituted TFIIH preparations, the authors provide evidence that U1 snRNA-containing TFIIH is more active both in an ‘abortive initiation’ assay that measures synthesis of the first transcribed phosphodiester bond, and in directing subsequent rounds of transcription (that is, reinitiation) from a TFIIHcontaining promoter ‘scaffold’. These findings raise several questions. First, unlike 7SK, U1 snRNA already has a well-defined function — in splicing as a component of U1 snRNP. Is there enough U1 snRNA in the nucleus to go around? TFIIH is much less abundant than U1 snRNP, so only a relatively small amount (<10%) of the total U1 snRNA would be needed in TFIIH complexes, and this seems feasible. And what could U1 snRNA be doing in TFIIH? One could imagine a purely structural function, helping to hold the TFIIH complex together. More intriguingly, it might play a regulatory role, perhaps modulating TFIIH kinase activity via the cyclin H interaction, analogous to the apparent regulatory role of 7SK in P-TEFb. While additional work will be required to address these questions, a clue may come from the puzzling observation that the addition of small RNAs containing the 5′ splice site consensus sequence (that is, complementary to the 5′ end of U1 snRNA) to nuclear extracts specifically inhibits RNAP II transcription at a very early step13. Could it be that, as in splicing, the 5′ end of U1 snRNA must be free and somehow functions to modulate TFIIH activity? Perhaps consistent with this, a promoter-proximal 5′ splice site was shown to be required for U1 snRNAdependent stimulation of reinitiation12. Another issue is how the results of Kwek et al.12 relate to those of Fong and Zhou6 described above. Both studies showed that components of the splicing machinery can stimulate transcription, but this is where the similarity ends. The results of Fong and Zhou6 suggest that functional snRNPs are recruited to elongating Nuclear coupling: RNA processing reaches back to transcription

DOI: 10.1038/nsb1102-790

Statistics

02040'04'06'08'10'12'14'16
Citations per Year

129 Citations

Semantic Scholar estimates that this publication has 129 citations based on the available data.

See our FAQ for additional information.

Cite this paper

@article{Manley2002NuclearCR, title={Nuclear coupling: RNA processing reaches back to transcription}, author={James L Manley}, journal={Nature Structural Biology}, year={2002}, volume={9}, pages={790-791} }