The role of phase separation in RNA polymerase II transition from initiation to elongation

High resolution mapping of elongating RNA polymerase II (RNAPII) in yeast using Native Elongating Transcript sequencing (NET-seq) revealed frequent pausing events in the 5’end of genes (Churchman and Weissman, 2011). More recently, we found that these early elongation pausing events are reduced dramatically when the RNAPII C-terminal domain (CTD) is mutated to replace its tyrosines with phenylalanines (Collin et al., 2019). Because the CTD is well known to recruit different factors to the elongating RNAPII, we then performed a proteomic analysis of this RNAPII-Y1F mutant but found no tyrosine-dependent RNAPII interactor likely to provide an explanation to the loss of pausing phenotype. Mounting evidence suggests that the CTD can form protein condensates in cells via liquid-liquid phase separation (Boehning et al., 2018; Guo et al., 2019; Lu et al., 2018). Protein condensates formed by transcription factors, co-activators and RNAPII (via the CTD) are major players in gene expression regulation, notably by providing local environments favorable for the assembly of macromolecular assemblies such as super-enhancers (Boija et al., 2018; Cho et al., 2018; Sabari et al., 2018). Interestingly, CTD phosphorylation, which is a very dynamic process during transcription elongation (Jeronimo et al., 2013), was shown to affect the ability of the CTD to form condensates (Boehning et al., 2018) and to drive RNAPII from one type of condensate to another (Guo et al., 2019; Lu et al., 2018). Based on these data, and because tyrosine residues are often determinant for the ability of a protein to form condensates (Wang et al., 2018), we hypothesise that 1) the transition from initiation to elongation involves a redistribution of RNAPII from initiation-type condensates into elongation-type condensates; 2) this transition is a rate limiting step resulting in pausing and 3) the phosphorylation of the CTD provides the driving force for this transition. This hypothesis will be addressed using a combination of in vitro and in vivo phase separation assays, combined with genomic technologies such as NET-seq and others.