Dynamic sumoylation of promoter-bound general transcription factors facilitates transcription by RNA polymerase II

Abstract

Transcription-related proteins are frequently identified as targets of sumoylation, including multiple subunits of the RNA polymerase II (RNAPII) general transcription factors (GTFs). However, it is not known how sumoylation affects GTFs or whether they are sumoylated when they assemble at promoters to facilitate RNAPII recruitment and transcription initiation. To explore how sumoylation can regulate transcription genome-wide, we performed SUMO ChIP-seq in yeast and found, in agreement with others, that most chromatin-associated sumoylated proteins are detected at genes encoding tRNAs and ribosomal proteins (RPGs). However, we also detected 147 robust SUMO peaks at promoters of non-ribosomal protein-coding genes (non-RPGs), indicating that sumoylation also regulates this gene class. Importantly, SUMO peaks at non-RPGs align specifically with binding sites of GTFs, but not other promoter-associated proteins, indicating that it is GTFs specifically that are sumoylated there. Predominantly, non-RPGs with SUMO peaks are among the most highly transcribed, have high levels of TFIIF, and show reduced RNAPII levels when cellular sumoylation is impaired, linking sumoylation with elevated transcription. However, detection of promoter-associated SUMO by ChIP might be limited to sites with high levels of substrate GTFs, and promoter-associated sumoylation at non-RPGs may actually be far more widespread than we detected. Among GTFs, we found that TFIIF is a major target of sumoylation, specifically at lysines 60/61 of its Tfg1 subunit, and elevating Tfg1 sumoylation resulted in decreased interaction of TFIIF with RNAPII. Interestingly, both reducing promoter-associated sumoylation, in a sumoylation-deficient Tfg1-K60/61R mutant strain, and elevating promoter-associated SUMO levels, by constitutively tethering SUMO to Tfg1, resulted in reduced RNAPII occupancy at non-RPGs. This implies that dynamic GTF sumoylation at non-RPG promoters, not simply the presence or absence of SUMO, is important for maintaining elevated transcription. Together, our findings reveal a novel mechanism of regulating the basal transcription machinery through sumoylation of promoter-bound GTFs.

Waves of sumoylation support transcription dynamics during adipocyte differentiation

“…First, ML-792 caused the disappearance of about 35% of RXR peaks and second, the appearance of 5616 specific peaks. These data are consistent with the previously reported role of sumoylation in altering the capacity and the specificity of transcription factors to bind the DNA ( 9 , 11 , 58 ). Third, ML-792 triggered an increase of RXR occurrence at PPAR/RXR/SUMO MA target genes, which correlated with transcriptional inhibition.…”

SUMOylation controls the rapid transcriptional reprogramming induced by anthracyclines in Acute Myeloid Leukemias

“…Their SUMOylation is likely to occur on chromatin as both SUMO conjugating (E1, E2 and E3s) and deconjugating enzymes can bind to the chromatin (17, 22–24). Although SUMOylation of chromatin-bound proteins has often been associated with gene silencing or gene expression limitation (24–28), it can also participate in the activation of certain genes such as ribosomal genes (19, 20). Overall, the impact of SUMOylation on transcription appears to be dependent on both genes and signaling contexts, as well as on the nature of the conjugated proteins and of the chromatin environment (16).…”

“…Numerous transcription factors and co-regulators, as well as histones and the basal transcription machinery are SUMOylated (16). Moreover, genome-wide studies have revealed that SUMOylated proteins are highly enriched at gene regulatory regions, including promoters and enhancers (17)(18)(19)(20)(21). Their SUMOylation is likely to occur on chromatin as both SUMO conjugating (E1, E2 and E3s) and deconjugating enzymes can bind to the chromatin (17,(22)(23)(24).…”

The SUMO stress response in transcriptional regulation: Causal relationships or secondary bystander effects?

“…For instance, whereas as discussed above heat stress-induced sumoylation at heat shock genes in human cells appears to reduce transcription of heat shock genes via RNAPII pausing, the opposite has been reported in Arabidopsis cells, where heat stress-induced sumoylation at heat shock genes enhances their transcription. [4] Finally, adding even more complexity to transcriptional regulation by SUMO, a recent study that investigated the effect of sumoylation on transcription in budding yeast identified the TFIIF subunit Tfg1 as a major SUMO target; [5] surprisingly, both preventing the sumoylation of Tfg1 (by mutating the lysine sites that are sumoylated into nonsumoylatable arginine residues) as well as increasing Tfg1 sumoylation (by directly tethering SUMO to Tfg1) resulted in reduced RNAPII occupancy at a selection of highly active promoters. This shows that it is not simply the presence of SUMO, but rather its dynamic turnover at promoters that determines transcriptional output.…”

“…Finally, adding even more complexity to transcriptional regulation by SUMO, a recent study that investigated the effect of sumoylation on transcription in budding yeast identified the TFIIF subunit Tfg1 as a major SUMO target; [ 5 ] surprisingly, both preventing the sumoylation of Tfg1 (by mutating the lysine sites that are sumoylated into nonsumoylatable arginine residues) as well as increasing Tfg1 sumoylation (by directly tethering SUMO to Tfg1) resulted in reduced RNAPII occupancy at a selection of highly active promoters. This shows that it is not simply the presence of SUMO, but rather its dynamic turnover at promoters that determines transcriptional output.…”