The Effects of Modulated Global Levels of Sumoylation on Gene Expression

Abstract

Hundreds of proteins are modified by SUMO (small ubiquitin-like modifier) peptides, the majority of which are nuclear and involved in gene expression. Through SUMO chromatin immunoprecipitation sequencing (ChIP-seq), sumoylated proteins can be readily detected at numerous distinct chromatin sites. This includes promoters of protein-coding genes thus linking sumoylation with transcription regulation. Cells can coordinately modulate levels of sumoylation globally. For example, there is a surge in SUMO conjugation in response to heat shock, an effect that we attribute here to the degradation of the major SUMO protease, Ulp1, in budding yeast. Whereas the effects of sumoylation have been examined for many individual target proteins, not much is known about how coordinated global changes to sumoylation levels impact transcription. To address this, we investigated whether changing cellular sumoylation levels in yeast impacts (1) cell growth in normal and stress conditions, and (2) global transcription and gene expression patterns. Primarily, this was accomplished using strains that express mutant forms of Ubc9, the sole E2 conjugating enzyme, or Ulp1, which harbour constitutively reduced or elevated sumoylation levels, respectively. We find that cells with dramatically reduced levels of sumoylation grow near-normally in optimal, non-stress conditions. While they tolerate multiple stress conditions very well, they show strong sensitivity to heat shock specifically. Both reduced and elevated sumoylation levels lead to widespread changes to transcription, but intriguingly, both result in a gene expression pattern that resembles that of stressed cells. Therefore, heat-shock genes have high levels of expression, even in non-stress conditions. This indicates that both sumoylation and desumoylation are important for suppressing stress response gene expression in non-stress conditions, and paradoxically, our results correlate constitutive expression of stress response genes with temperature sensitivity. Finally, our data implicate activation of the key stress response factors, Msn2 and Hog1, in driving the inappropriate induction of heat-shock genes in Ubc9 and Ulp1 mutants even in the absence of stress.