|Home | About | Journals | Submit | Contact Us | Français|
Chromatin is the natural context for transcription elongation. However, the elongating RNA polymerase II (RNAPII) is forced to pause by the positioned nucleosomes present in gene bodies. Here, we briefly discuss the current results suggesting that those pauses could serve as a mechanism to coordinate transcription elongation with pre-mRNA processing. Further, histone post-translational modifications have been found to regulate the recruitment of factors involved in pre-mRNA processing. This view highlights the important regulatory role of the chromatin context in the whole process of the mature mRNA synthesis.
Pre-mRNA processing and transcription are two interconnected processes that comprise several events with a very precise regulation. To obtain a final product of messenger ribonucleoproteins (mRNP) ready to be exported to the cytoplasm, each step in the production has to be precisely executed and to be coordinated with the surrounding functions. Thus, defective transcription directly interferes with the processing of the produced pre-mRNA and, conversely, problems in pre-mRNA processing affect transcription (reviewed in1). Adding to the complexity, transcription elongation takes place in the context of chromatin. We discuss here recent results from different groups that indicate that one of the functions of chromatin is in fact to coordinate these processes.
In vitro and in vivo experiments demonstrate that a nucleosome inside of a gene presents an obstacle for transcription elongation and promotes pausing of RNAPII.2-6 However, it is conceivable that pauses in transcription due to the presence of nucleosomes could also have a positive function. Several studies have described the genome-wide distribution of nucleosomes in different organisms,7-10 showing that, inside of genes, nucleosomes are preferentially positioned in certain regions (see below). RNAPII pause sites in transcription elongation often correlate with positioned nucleosomes8,11,12 and, strikingly, the transcription rate varies between genes and even in different regions of the same gene.13,14 It is not yet clear whether elongation rate differences between genes are due to differences in the nucleosome landscape, or whether the characteristic chromatin distribution inside of genes is important for regulating transcription elongation. New results from our laboratory point to an important role of chromatin in controlling the RNAPII elongation rate. Specifically, we have found that decreasing the levels of canonical histones in human cells, which creates a more open chromatin structure, results in faster RNAPII elongation, which strongly suggests that the density of nucleosomes affects the elongation rate.15
Most genes have a similar nucleosome distribution near the transcription start site: the promoter is nucleosome free, while the first (+1) nucleosome (with respect to the promoter element) is strongly positioned. Thus, the +1 nucleosome is located where two events—promoter-proximal pausing and pre-mRNA capping—take place, providing a good opportunity for coordination between transcription elongation, processing machineries and chromatin remodeling. RNAPII pausing has been proposed to act as a transcriptional checkpoint to ensure that all necessary elongation factors have been recruited.16 In an earlier work, Rasmussen and Lis proposed that capping and promoter-proximal pausing are linked processes.17 Later, Reinberg and co-workers demonstrated that DSIF, a factor involved in pausing, enhances capping of pre-mRNA in vitro.18 Maps of Drosophila nucleosomes and paused polymerases are consistent with the +1 nucleosome playing a role in promoter-proximal pausing in a subset of genes.11,19 Moreover, genes with a paused RNAPII have a shift in first nucleosome position, with a direct interaction between the +1 nucleosome and the RNAPII.8 Indeed, we have recently demonstrated by introducing strong nucleosome-positioning sequences that a well-positioned +1 nucleosome near the promoter of the c-Myc gene increases promoter-proximal pausing and pre-mRNA capping efficiency (Fig. 1A).20
Most pre-mRNA splicing events occur co-transcriptionally,21-23 and it is clear that transcription elongation influences the splicing process.24 Two non-mutually exclusive mechanisms, the recruitment coupling and kinetic coupling models, have been postulated to explain how transcription and splicing are coordinated (these have been extensively described in several excellent reviews24-26). The mechanisms proposed depend on the interactions and kinetics of the elongating RNAPII, respectively. Since transcription takes place in chromatin, both the recruitment of factors to the transcription complex and the RNAPII transcription rate are influenced by the chromatin landscape. Numerous results have suggested that two main aspects are involved in chromatin-dependent control of splicing: nucleosome occupancy and histone posttranslational modifications.
In 2009, several laboratories reported biases in nucleosome occupancy in the gene bodies, with a stronger occupancy in exons than in introns, and more specifically, in constitutive exons than in alternative exons, and in exons with weak splice sites than in exons with strong splice sites.27-31 Overall, these results suggest that nucleosome positioning contributes to exon selection. It has been suggested that pausing or delaying of RNAPII, promoted by exonic nucleosome positioning, could help exon inclusion. Several data support this view. First, RNAPII accumulates at spliced exons,11,32-35 and its elongation rate is negatively correlated to the exon density.13,36 Second, changes in progesterone-dependent nucleosome occupancy correlate with alternative splicing and RNAPII accumulation.37 Third, as we have recently shown, perturbing the chromatin structure by depleting canonical histones leads to drastic alterations in splicing.15 In particular, histone depletion promotes skipping of the variable exons in the CD44 gene and leads to a reduction in the levels of RNAPII at that region. Interestingly, this defective phenotype can be rescued by a “slow” RNAPII mutant. Our data therefore suggest that the CD44 splicing defect is a direct consequence of an increased elongation rate caused by an aberrant chromatin organization (Fig. 1B).
Nucleosomes not only affect splicing due to their elongation barrier properties. For instance, some exon-specific histone post-translational modifications are involved in the recruitment of splicing factors. A paradigm for a histone modification that affects splicing is the trimethylation of H3 lysine 36 (H3K36me3), which is significantly enriched on exons.27,29,31,38 Silencing or mutating the H3K36 methyltransferase SETD2 has been associated with differential exon inclusion and intron retention.39,40 Moreover, H3K36me3 binding proteins, such as MRG1539 and Psip1/Ledgf,41 are able to recruit splicing factors to chromatin that regulate alternative splicing. Two recent studies have reported that the protein BS69/ZMYND11 is able to specifically bind K36me3 that occurs in the histone variant H3.3.42,43 BS69/ZMYND11 also interacts with a number of U5 snRNP components of the spliceosome.42 Strikingly, depletion of BS69/ZMYND11 strongly decreases intron retention, suggesting that BS69/ZMYND11 promotes retention of specific introns.42 We have reported that depletion of canonical histones increases the levels of the histone variant H3.3 at gene bodies and, intriguingly, promotes also intron retention.15 Therefore, higher levels of H3.3 at the body of transcribed genes under histone depletion could facilitate the recruitment of BS69/ZMYND11, which eventually would enhance intron retention. Other histone post-translational modifications, such as acetylation or methylation of H3K9, have also been found to be related to RNA splicing.44-46
In summary, many lines of evidence indicate that both alternative and constitutive splicing are affected by chromatin structure, which works either by modulating the RNAPII elongation rate or by promoting the recruitment of splicing factors (Fig. 1B). These two aspects may be two sides of the same coin, whose influence on splicing depends on the specific context of the exon.
Nucleosome occupancy around transcription termination sites is also well organized. In fact, polyadenylation (polyA) sites are strongly depleted of nucleosomes, whereas downstream regions are nucleosome enriched.27 It is well known that RNAPII pauses around the 3′ end of most mammalian and yeast genes,47-49 and that a forced pausing increases termination in vivo in an XRN2-dependent manner.50 Interestingly, a correlation between RNAPII accumulation and nucleosome organization around the polyA site has been reported.48,51 Moreover, a chromatin remodeling mutant of the ISWI sub-family in yeast shows termination defects.52 Based on these results, RNAPII pausing created by the positioned nucleosomes downstream of the polyA site has been proposed to help the termination process, although how the pausing affects 3′ pre-mRNA processing is still unclear.
Proudfoot and collaborators have recently reported a possible mechanism by which chromatin increases pausing at the termination region.53 They show that R-loop (RNA-DNA hybrids)–induced antisense transcription occurs after the polyA site promoting recruitment of DICER, AGO1, AGO2 and the G9a histone H3 lysine 9 methyltransferase. Following this, the level of the H3K9me2 repressive mark increases, and the heterochromatin protein 1γ (HP1γ) is recruited, to induce a heterochromatin-like structure that enhances RNAPII pausing and promotes termination. A role of facultative heterochromatin in controlling alternative polyA sites has been also reported in Arabidopsis thaliana (reviewed in54). In some cases, antisense transcription and RNA-interference machinery are also involved in alternative 3′-processing in plants.
Finally, consistent with the important role that chromatin plays in termination, our analyses of cells depleted of canonical histones have revealed defects in 3′-processing and polyA in some genes (S. Jimeno-González and J. C. Reyes, unpublished results).
Both the nucleosome organization and the combination of histone post-translational modifications at gene bodies strongly affect how pre-mRNA is processed. This mostly occurs through two interrelated mechanisms: by modulating the RNAPII elongation rate and by determining the recruitment of additional factors that directly affect RNA processing. A common aspect that arises from our data and the work from other groups is that the nucleosome-imposed slow progression of RNAPII provides windows of opportunities to recruit RNA processing factors. Hence, an “overly” permissive chromatin causes accumulation of pre-mRNA intermediaries (e.g., uncapped or unspliced pre-mRNA) that will eventually be degraded by the RNA quality control machinery. The cell can then use these chromatin-based mechanisms to change pre-mRNA processing efficiency in order to modulate gene expression, through the involvement of chromatin remodeling machineries and epigenetic factors far from their traditionally-studied working sites to promoters and enhancers. Finally, it is known that senescence and aging are characterized by changes in histone availability.55,56 Therefore, it is possible that RNA processing effects associated to senescence or aging are caused by chromatin alterations.
No potential conflicts of interest were disclosed.
Work in the Silvia Jimeno-González and José C. Reyes laboratory was supported by the Spanish Ministry of Economy and Competitiveness (MINECO) [Grants number BFU-2011-23442 and BFU2014-53543-P] and Andalusian Government [P06-CVI-4844].