A new player in Pol II pausing

Abstract

EMBO J 32 13, 1829–1841 doi:; DOI: 10.1038/emboj.2013.111; published online May 24 2013

Transcriptional control plays a primary role in gene expression and can be exerted at multiple levels of the transcription process, such as RNA polymerase recruitment and promoter-proximal pausing. A recent report published in The EMBO Journal (Li and Gilmour, 2013) provides new insights into the regulation of a large class of paused genes through the identification of a novel transcription factor (TF; Motif 1 binding protein, M1BP) that binds to promoters of paused genes and controls their expression.

RNA polymerase II (Pol II) transcription is regulated primarily at two steps early in the production of mRNA (Nechaev and Adelman, 2011). The first is the recruitment of Pol II to promoters where, with the aid of the general transcription factors (GTFs), Pol II assembles into a pre-initiation complex (PIC) that then rapidly initiates transcription (Core et al, 2012). The second is promoter-proximal Pol II pausing and the subsequent escape into productive elongation (Figure 1). Interestingly, these steps of recruitment and pausing are tightly linked both spatially and mechanistically, and DNA sequences of the core promoter that provide strong PIC interactions also participate in Pol II proximal pausing (Kwak et al, 2013). A plethora of sequence-specific TFs and Pol II-interacting factors modulate, either negatively or positively, the efficiency of both the recruitment and the pausing steps. While the general role of Pol II recruitment in gene regulation has been studied intensively for decades, the genome-wide role of Pol II pausing in transcription regulation has only recently gained a broader appreciation (Adelman and Lis, 2012).

Figure 1.

Genes are regulated at either of two stages in transcription: recruitment of Pol II or promoter-proximal pausing. (A) For genes regulated via Pol II recruitment, levels of transcriptionally engaged Pol II are relatively uniform across the gene (bottom panel, grey line). Once Pol II is recruited, it initiates transcription and transcribes through the promoter-proximal region and any nucleosomes in the gene body at a relatively constant rate. These genes can be regulated by sequence-specific TFs that enhance or repress Pol II recruitment and initiation, and increase (green line) or decrease (red line) levels of transcriptionally engaged Pol II, respectively. (B, C) For genes regulated at the pausing stage, levels of transcriptionally engaged Pol II are dramatically higher at the 5′ end of the gene (grey line). Pol II is recruited, initiates transcription and transcribes 30–60 nucleotides of RNA. Only after release from the pause does Pol II continue to transcribe across the gene. (B) One class of paused genes is enriched for GAF binding, and is further controlled by a set of sequence-specific TFs that regulate expression by increasing the rate of pause escape. Thus, the level of transcriptionally engaged Pol II on the gene body increases upon activation, resulting in a large range of expression levels depending on the regulation status, pre-activation (grey line) or post-activation (green line). (C) Li and Gilmour (2013) have discovered Motif 1 and M1BP at a subset of GAF-less paused genes. Pausing is less efficient for M1BP than for GAF-bound genes. In contrast to GAF-bound genes, the paused Pol II is directly upstream of the first nucleosome, and pausing may be nucleosome-dependent. M1BP genes are constitutively expressed, tend not to have other TF motifs and have a narrower range of expression levels (bottom panel).

The first Pol II pausing factors, identified by Handa and colleagues, were shown to interact directly with Pol II shortly after the start of transcription. These factors, DSIF (Spt5, Spt4) and NELF, stabilize and maintain Pol II at the pause region. Both proteins were shown to inhibit elongation in vitro and to localize to pause regions. In addition, depletion of these factors reduces pausing in vivo (Yamaguchi et al, 2013). In contrast, the kinase P-TEFb (Cyclin T1, CDK9) phosphorylates Spt5 and the C-terminal domain of the Pol II subunit Rpb1, and triggers Pol II escape from pausing (Peterlin and Price, 2006). Together, these regulators have featured prominently in models of promoter-proximal pausing, but more recent studies show that they are not the only players.

Sequence-specific TFs have also been implicated in pausing. GAF and its binding motif, the GAGA element, are enriched on paused genes in Drosophila, but only about 20% of paused genes are GAF-bound (Hendrix et al, 2008; Lee et al, 2008). The large number of paused genes not associated with a sequence-specific TF motivated Li and Gilmour, 2013 to search for sequence motifs enriched on GAF-less paused genes. They found that the previously identified Motif 1 (Ohler et al, 2002) was enriched around 30 bases upstream of the transcription start site (TSS) on paused genes. In addition, they isolated a ubiquitously expressed 55 kDa protein that binds to Motif 1, which they termed Motif 1 binding protein (M1BP).

M1BP bound nearly all Motif 1-containing promoters and its binding appeared critical for the control of transcription at many genes. RNAi-depletion of M1BP altered expression of 1152 genes, and dramatically reduced promoter Pol II levels at all three of the M1BP-bound genes tested. In general, the level of expression of M1BP-bound genes was higher and more uniform than GAF-bound genes, and moreover, individual M1BP-bound genes showed a more consistent expression at different developmental stages than either GAF-bound or genes containing a TATA box, a core promoter element. Consistent with their uniform expression, M1BP-bound genes were enriched for genes involved in basic cellular processes, and were not associated with any other specific TF motifs (Figure 1C). In contrast, GAF-bound genes were enriched for developmental genes and TATA box-containing genes were enriched for genes that respond to external stimuli and have tissue-specific functions. Consistent with their distinct functions, GAF-bound and TATA box-containing genes are enriched for unique sets of TF motifs (Figure 1B).

M1BP-bound and GAF-bound genes may have distinct pausing mechanisms. Li and Gilmour (2013) found that GAF-bound genes had a higher level of pausing than M1BP-bound genes, suggesting less-efficient pausing on M1BP-bound genes. A recent base-pair resolution mapping of engaged Pol II revealed two distinct classes of pausing. The more ‘focused&proximal’ pause sites are upstream of the first (+1) nucleosome, whereas a large subset of ‘dispersed&distal’ pause sites was located within the +1 nucleosome-protected region (Kwak et al, 2013), suggesting nucleosomes may mediate pausing at some sites by blocking further elongation. Interestingly, GAF-bound genes were enriched for ‘focused&proximal’ pausing. In agreement with these results, Li and Gilmour (2013) now found that pausing sites on many GAF-bound genes were upstream of the +1 nucleosome, suggesting pausing on these GAF-bound genes is largely nucleosome-independent (Figure 1B). In contrast, most paused polymerase on M1BP-bound genes was within +1 nucleosome-protected region, suggesting pausing at these genes may be nucleosome-dependent (Figure 1C). Given the lower level of pausing at M1BP-bound genes, these results may also be explained by a recent study showing that there is a competition between the paused polymerase and the +1 nucleosome for binding to the promoter-proximal region (Gilchrist et al, 2010). Thus, it will be mechanistically important to determine whether both Pol II and the nucleosome co-occupy these promoters to give further credence to a nucleosome-dependent mechanism for pausing.

The distinct set of functions and different expression patterns between the GAF and M1BP groups of genes indicate that unique transcriptional programmes are dictated by these two factors. Interestingly, differences in Pol II and nucleosome distribution on these gene sets suggest different pausing mechanisms exist for each. GAF and M1BP now account for ∼50% of paused genes. Perhaps other pause-specifying TFs exist, and they too may direct unique regulatory programmes for their target genes.