SPTing across condensates: SEC‐mediated translocation of SPT complex from pausing condensates to elongation condensates

Plain language summary

Several independent studies in the last few years have suggested that phase separation and biomolecular condensation play a critical role in regulating different transcription steps from initiation to pausing and elongation. However, how components of the transcription machinery translocate among different types of condensates during transcription remains poorly understood. Guo et al have now identified a potential mechanism underlying translocation of the DSIF complex from pausing to elongation condensates during promoter pause release, as reported in this issue of EMBO reports.

The super elongation complex (SEC) induces phase transition of SPT5 from pausing clusters into elongation droplets, and disease‐associated SEC mutations impair phase properties of the elongation clusters and transcription.

Eukaryotic transcription by RNA Polymerase II (Pol II) is regulated at multiple steps to control mRNA output during growth, development, and stress. The binding of negative elongation factor (NELF) and DRB sensitivity‐inducing factor (DSIF) to transcribing Pol II results in its promoter‐proximal pausing about 40–60 bp downstream of the transcription start site (TSS). The release of paused RNA Pol II involves a series of phosphorylation events on RNA Pol II, NELF, and DSIF by positive transcription elongation factor PTEFb. One of the active PTEFb‐containing complexes is the super elongation complex (SEC) containing eleven‐nineteen Lys‐rich leukemia (ELL) proteins and several frequent mixed lineage leukemia (MLL) translocation partners. As a result of phosphorylation by PTEFb, NELF dissociates from Pol II, and Pol II goes into productive elongation mode accompanied by DSIF. Thus, PTEFb‐mediated phosphorylation switches DSIF, from being a pausing factor to become a positive elongation factor to facilitate the progression of RNA Pol II through the gene body (Cramer, 2019).

In recent years, evidence suggests that different transcriptional steps occur in phase‐separated biomolecular condensates regulated by low‐complexity sequences and post‐translational modifications (PTMs; Rippe & Papantonis, 2022; Sabari, 2020). RNA Pol II can itself undergo phase separation owing to its low‐complexity C‐terminal domain (CTD; Boehning et al, 2018). Phosphorylation of the CTD of RNA Pol II results in its switch from initiation to elongation condensates (Guo et al, 2019). Despite a number of recent reports, the functional relevance of condensation for transcriptional regulation still remains unresolved. Several questions need to be addressed in order to understand the molecular dynamics of the transcription process within and across condensates. How do transcriptional condensates exchange components to proceed to the subsequent transcriptional steps? How do the regulatory proteins partition and regulate these condensates in space and time?

As a first step to address these questions, Guo et al (2023) characterized the propensity of the DSIF complex to undergo condensation in the context of transcriptional pausing and the subsequent step of transcriptional elongation. The DSIF complex consists of two proteins SPT4 and SPT5. The authors demonstrate the formation of stable condensates containing NELF, SPT5, and Ser5 phosphorylated Pol II (S5P‐Pol II) at paused promoters upon serum starvation of human cells. They refer to them as pausing condensates. These findings align with the previously described stress‐induced NELF condensates that attenuate global transcription (Rawat et al, 2021). SPT5 contains two IDRs in the N and C‐terminal sequences. Deleting either of the IDRs abrogates SPT5 function in both transcription pausing and productive elongation. However, N‐terminal IDR deletion leads to aberrant aggregation, whereas C‐terminal IDR deletion leads to the complete loss of condensation. Even though the underlying molecular mechanisms of the pausing condensates remain uncharacterized, this study paves a way to investigate the nature of intra and intermolecular interactions among IDRs of different constituent proteins and PTMs. Such interactions may result in stable multicomponent and multilayer pausing condensates with different FRAP dynamics for the core and the shell (Fig 1).

Figure 1. DSIF translocates from pausing to elongation condensates.

Super elongation complex (SEC)‐directed PTEFb‐mediated phosphorylation of the DSIF complex results in translocation from gel‐like pausing condensates to elongation condensates with fast FRAP dynamics. NELF complex constitutes the more mobile core of pausing condensates, whereas DSIF forms the less mobile, more gel‐like shell. The CHOPS syndrome mutations in AFF4 result in elongation condensates with slower FRAP dynamics and hyperactive transcriptional output in human cells.

The DSIF complex is one of the few components that transverses with RNA Pol II from pausing to elongation state as a result of PTEFb‐mediated phosphorylation. First, to test whether SEC components (EFF4 or ELN) play a role in SPT5 switching from pausing to elongation condensates, the authors performed in vitro reconstitution assays with NELFA‐IDR and SPT5‐CTR with or without AFF4‐IDR. The NELFA‐IDR droplets enrich SPT5‐CTR in the absence of AFF4‐IDR, whereas AFF4‐IDR readily forms droplets partitioning SPT5‐CTR but not NELFA‐IDR. These results indicate an apparent affinity of SPT5 toward AFF4 condensates. Second, the authors used ENL KO cells and pharmacological inhibition of the SEC complex to better understand the interplay between pausing and elongation condensates. Inhibition of transcription elongation results in increased partitioning of SPT5 into pausing condensates characterized by increased co‐localization with NELF complex members and S5P‐Pol II. These findings complement a previous report of phosphorylation by PTEFb antagonizing NELF condensates (Rawat et al, 2021). Finally, to directly investigate the role of SPT5 phosphorylation in its translocation to elongation condensates, the authors generated a phosphor‐mimic mutant of SPT5‐CTR. Phospho‐mimic SPT5‐CTR showed better partitioning into AFF4 droplets. On the contrary, Phospho‐mimic SPT5‐CTR showed complete exclusion from NELFA‐IDR droplets. These results align with published biochemical evidence on pause release and transcription elongation (Cramer, 2019). Thus, the results reported by Guo et al (2023) provide a first biochemical glimpse into the dynamics of transcription regulators within condensates when transitioning from one stage to another. Interestingly, the authors find that disease‐associated mutations alter the properties of transcription condensates and their functional outcome. AFF4 mutants resembling the CHOPS syndrome's germline mutation stabilize AFF4 protein levels and enhance its phase separation properties, adding to recent studies linking phase separation properties and disease states (Alberti & Hyman, 2021).

In conclusion, the study sheds light on the importance of switching components among biomolecular condensates for their function and regulation based on a mix of in vitro and cellular experiments focusing on the DSIF complex. To our knowledge, SEC‐mediated phosphorylation remains the only mechanism reported so far that governs the switching of components (Pol II and SPT5) between different transcriptional condensates (initiation or pausing to elongation). The study opens up several aspects for further exploration: How do different transcriptional compartments maintain their integrity and switch components precisely to regulate transcriptional output? Are there more mechanisms at play in addition to SEC‐mediated phosphorylation? Recently, two models were proposed for Pol II switching among transcriptional condensates (Portz & Shorter, 2020). The first model describes Pol II switching from one condensate to another in their spatially and compositionally maintained states. The second model suggests a change in the concentrations of condensate components, such as the mediator complex or splicing machinery in Pol II containing condensates. Both models remain probable for DSIF, in line with findings reported here (Guo et al, 2023). Additionally, it is possible that insulator complexes function in preventing the mixing of components from different transcriptional condensates. One such candidate for an insulator complex would be the NELF complex. The NELF complex remains inert to all transcriptional components except Pol II (Lyons et al, 2023). It would be interesting to identify whether transient NELF condensates can act as barriers between initiation and elongation condensates under no‐stress conditions (Rawat et al, 2021). Findings on the role of SEC in transcriptional condensates raise several questions: do SEC condensates differ from previously reported elongation condensates with splicing components (Guo et al, 2019)? Do these SEC condensates fuse partially or not with pausing condensates to phosphorylate NELF and DSIF and achieve productive elongation? In the second case, the fate of the NELF core in pausing condensates remains elusive upon pause release. Finally, understanding the regulation of these biomolecular condensates, the functional consequences of aberrant phase transitions, and cell fate outcomes will help design better therapeutics for targeting human diseases involving transcriptional deregulation.

EMBO reports (2023) 24: e56810