You are here

David Tollervey

Co-workers:

Stefan Bresson, Clémentine Delan-Forino, Tatiana Dudnakova, Hywel Dunn-Davies, Aziz El Hage, Aleksandra Helwak, Rebecca Holmes, Laura Milligan, Elisabeth Petfalski, Camille Sayou, Vadim Shchepachev, Tomasz Turowski, Marie-Luise Winz
Tollervey Lab Homepage

Nuclear RNA Processing and Surveillance

David gives a brief overview of his research.

Our aim is to understand the nuclear pathways that process newly transcribed RNAs and assemble the RNA-protein complexes, the mechanisms that regulate these pathways and the surveillance activities that monitor their fidelity. A prominent feature of our recent work has been the application of in vivo UV crosslinking techniques to precisely identify sites of RNA-protein interaction (CRAC) and RNA-RNA basepairing.

Applying the CRAC approach to the RNA polymerases allowed high-resolution, strandspecific mapping of transcriptionally engaged RNA polymerase I, II and III. In each case, this revealed markedly uneven crosslinking efficiency along genes. We believe this reflects strong effects of chromatin structure on polymerase processivity at a local level. This was particularly striking for RNA polymerase I within the 5’ region of the rDNA, where strong, highly regular pausing was evident. Apparent pause sites were associated with binding of RNA surveillance factors, reflecting our finding that quality control mechanism are deeply embedded within eukaryotic gene expression systems. In the case of RNAPII, we further mapped the distribution of multiple modified forms of the polymerase (mCRAC) and applied a Hidden Markoff Model (HMM) to help interpret the resulting, highly complex datasets (Fig. 1). We also used CRAC to identify direct targets for the major pre-mRNA binding protein Npl3. We found that Npl3 binds diverse sites on large numbers of transcripts, and that the loss of Npl3 results in transcriptional read-through on many genes. One effect of this transcription read-through is that the expression of numerous flanking genes is strongly down-regulated. This underlines the importance of faithful termination for the correct regulation of gene expression. The effects of the loss of Npl3 are seen on both mRNAs and non-protein coding RNAs. These have distinct but overlapping termination mechanisms, with both classes requiring Npl3 for correct RNA packaging. Analyses of RNA polymerase III showed that many tRNA genes generate long, 3´-extended forms due to read-through of the canonical poly(U) terminators. The steady-state levels of 3´-extended pre-tRNA transcripts are low, due to targeting by the nuclear surveillance machinery. In addition, several previously unidentified RNA polymerase III transcripts were mapped. Together, these findings increased our understanding of the interdependency of RNA synthesis and surveillance pathways.

Selected publications:

Turowski, T.W., Lebaron, S., Zhang, E., Peil, L., Dudnakova, T., Petfalski, E., Granneman, S., Rappsilber, J. and Tollervey, D. (2014) Rio1 mediates ATPdependent final maturation of 40S ribosomal subunits. Nucleic Acids Res., 42, 12189-12199. PMID: 25294836
El Hage, A., Webb, S., Kerr, A. and Tollervey, D. (2014) Genome-wide distribution of RNA-DNA hybrids identifies RNase H targets in tRNA genes, retrotransposons and mitochondria. PLoS Gen., doi: pgen.1004716. PMID: 25357144.
Holmes, R.K., Tuck, A.C., Zhu, C., Dunn-Davies, H.R., Kudla, G., Clauder-Munster, S., Granneman, S., Steinmetz, L.M., Guthrie, C. and Tollervey, D. (2015)
Loss of the yeast SR protein Npl3 alters gene expression due to transcription readthrough. PLoS Gen., 11, e1005735


1. Total RNA polymerase II density across all protein-coding genes in the sense and antisense orientation. On almost all genes the polymerase density is higher close to the transcription start site (TSS). In addition, many genes show clear antisense transcripts in the 3’ region.
2. Gene segmentation by HMM analyses. The model uses machine learning to define 8 major “states”, based on the sequence data for the positions of modified forms of RNA polymerase II, and segments the genome based on the probability that each position is in one of these states. A. Representative gene. B. Metagene analysis of 1,500 highly expressed genes. There is a progression from initiation states (I1 and I2) to early elongation (EE) and late elongation (E1-E3) states. Low modification (L) and noise (N) states frequently flank genes.