Transcriptional heterogeneity shapes stress-adaptive responses in yeast

In response to stress, cells activate signaling pathways that coordinate broad changes in gene expression to enhance cell survival. Remarkably, complex variations in gene expression occur even in isogenic populations and in response to similar signaling inputs. However, the molecular mechanisms underlying this variability and their influence on adaptive cell fate decisions are not fully understood. Here, we use scRNA-seq to longitudinally assess transcriptional dynamics during osmoadaptation in yeast. Our findings reveal highly heterogeneous expression of the osmoresponsive program, which organizes into combinatorial patterns that generate distinct cellular programs. The induction of these programs is favored by global transcriptome repression upon stress. Cells displaying basal expression of the osmoresponsive program are hyper-responsive and resistant to stress. Through a transcription-focused analysis of more than 300 RNA-barcoded deletion mutants, we identify genetic factors that shape the heterogeneity of the osmostress-induced transcriptome, define regulators of stress-related subpopulations and find a link between transcriptional heterogeneity and increased cell fitness. Our findings provide a regulatory map of the complex transcriptional phenotypes underlying osmoadaptation in yeast and highlight the importance of transcriptional heterogeneity in generating distinct adaptive strategies. Overall design: S. cerevisiae (BY4741) were genetically barcoded either with resistance markers (WT and hog1 dataset) or through a modification of the yeast knock out collection deletion (deletion based scRNA-seq screening). For both experiments, all cells were grown individually and pooled before subjecting them or not to osmotic stress (0.4M NaCl, 15 mins). Cells were harvested by centrifugation and immediatelly fixed with ice cold methanol 80% and stored at -20ºC following 10X genomics meethanol fixation instructions. Frozen cells were equilibrated at 4ºC for 30 minutes and rehydrated following manufacturer's instructions. scRNA-seq profiling was performed using 10X genomics, each condition or time point was loaded into a lane.