The CTD Is Not Essential for the Post-Initiation Control of RNA Polymerase II Activity

Highlights

• The CTD is not essential for promoter-proximal pause-release.

• Even in absence of the CTD, pause-release depends on CDK9 kinase activity.

• The CTD is not vital for productive Pol II transcription elongation.

• The CTD is required for normal RNA processing and transcription termination.

Abstract

Interest in the C-terminal domain (CTD) of the RPB1 subunit of the RNA polymerase II (Pol II) has been revived in recent years, owing to its numerous posttranslational modifications and its “phase-separation” properties. A large number of studies have shown that the status of CTD modifications is associated with the activity of Pol II during the transcription cycle. However, because this domain is essential in living cells, the functional requirement of the full CTD for the control of Pol II activity at endogenous mammalian genes has never been addressed directly in living cells. Using an inducible Pol II-degradation system that we previously established, we investigated here the roles of the CTD in the post-initiation control of Pol II. The selective ablation of the RPB1 CTD, post-initiation, at promoter-proximal pause-sites revealed that this domain, and by extension the CTD heptads and their modifications, is functionally neither absolutely required to maintain pausing in the absence of CDK9 activity nor essential for the release of Pol II into productive elongation.

Introduction

The RNA polymerase II (Pol II) C-terminal domain (CTD), is a long, highly conserved and relatively disordered domain of the largest Pol II subunit RPB1. The CTD in mammals is composed of 52 heptad motifs (with a consensus motif Y₁S₂P₃T₄S₅P₆S₇), whose residues, except for the prolines, are subjected to extensive phosphorylation by a multitude of kinases known to orchestrate the transcription cycle and pre-mRNA processing [[1], [2], [3]]. In particular, S₂- and S₅-phosphorylated residues account for ~ 75% of the total phospho-counts, predominantly found in monophosphorylated repeats [4]. Moreover, the patterns of CTD-repeat modifications are dynamically altered throughout the transcription cycle, with S₅-phosphorylation displaying the highest levels close to the transcription start site (which might also reflect higher levels of Pol II found in promoter-proximal regions) and S₂-phorphorylation being the highest at the 3′ ends of the genes (where Pol II also accumulates to higher levels) [[1], [2], [3]]. Pol II recruited to promoters is largely unphosphorylated. During initiation, the CDK7 subunit of the general transcription factor TFIIH phosphorylates the S₅ (and S₇) residues of the CTD. S₅ phosphorylation has been proposed to disrupt interactions with the Mediator complex to facilitate promoter escape [[5], [6], [7], [8]]. In addition, phosphorylated S₅ residues enhance recruitment and binding of enzymes involved in capping [2,3,6,7]. Following promoter escape, early transcribing complexes encounter an initial checkpoint, promoter-proximal pausing, which may facilitate the integration of multiple cellular signals [6,9]. Promoter-proximal pausing is established within the first 50 bp downstream from the TSS by numerous factors that include DNA/RNA sequences, elongation factors such as DSIF and NELF and the first nucleosome encountered [6,9]. Interestingly, the general initiation factor TFIID has recently been reported to play a role in establishing pausing [10]. Recent evidence suggests that NELF also regulates functions distinct from the initial “intrinsic” pause, namely, a second “regulated” Pol II pausing step whose release is P-TEFb-dependent [11]. The transition to productive elongation is associated with the phosphorylation of NELF, DSIF and CTD S₂ residues by the CDK9 subunit of P-TEFb. Pol II pausing (referring, hereafter, collectively to the first and second pausing steps combined as our system does not allow a distinction between these two steps) appears to be important for proper capping of the nascent transcript, for preventing reinitiation by an additional Pol II and for maintaining the promoter in an open state [6,7]. Following release into productive elongation, as well as later in the transcription cycle, the dynamic changes of CTD modifications appear to play a role in orchestrating the recruitment of factors involved in pre-mRNA processing, such as splicing, and transcription termination [[1], [2], [3]]. In addition, chromatin modifying complexes also bind the modified CTD, suggesting a potential role in transcription elongation by altering the chromatin environment [3,6,12]. Hence, the CTD is often seen as a platform for the recruitment of splicing or other processing factors, although evidence challenging this view has also been reported [13]. Although the requirement of the CTD for normal RNA processing and transcription termination is well established, its function in the control of the post-initiation activity of Pol II remains unclear.

We recently described an inducible degradation system, established in HEK293 cells, that allows depletion of an RPB1 tagged at the very end of the CTD with a minimal auxin-inducible degron (AID) and a fluorescent protein (mKO2) for real-time monitoring (Figure 1(a)) as the sole source of this essential Pol II subunit. The precise characterization of RPB1 depletion following DOX and auxin treatment revealed, unexpectedly, that a fraction of chromatin-bound RPB1 escapes processive degradation by the proteasome, generating actively transcribing and termination-incompetent Pol II complexes lacking the CTD in its entirety (also known as Pol IIB) [15]. An ectopically expressed Pol II with a CTD shortened to five repeats was previously shown to be deficient for initiation in vivo on chromatinized genes [16,17], thereby preventing assessment of post-initiation CTD requirements on endogenous genes. Thus, the special characteristics of our inducible RPB1-truncation system provided an opportunity to assess requirements for the CTD, and by extension the heptad repeats, in the control of Pol II during post-initiation steps such as promoter-proximal pausing, release into productive elongation, and termination. We found that the formation of polymerases that specifically lack the entire CTD preferentially occurred at promoter-proximal pause sites, thereby providing an explanation for the reported termination site read-through observed at genes displaying stably paused Pol II complexes [15]. Here, we further show that these truncated polymerases are nevertheless released into productive elongation in a CDK9-dependent manner despite the lack of all the heptads repeats that are normally phosphorylated by CDK9 during the transition to productive elongation. Hence, these results establish that the CTD is dispensable for pause-release and productive transcription elongation at endogenous genes but, as previously described, play an essential role in RNA processing and transcription termination.