Fig. 2

Pol II and GTF mutants confer polar shifts in TSS usage across all promoter classes in S. cerevisiae. a Heat maps show relative TSS distribution changes in a fast (rpb1 E1103G) or a slow (rpb1 H1085Y) Pol II mutant relative to WT. Four hundred one nucleotide promoter windows were anchored on measured median TSS position in the WT strain, with TSS distributions for WT or mutant strains normalized to 100% for each promoter (heat map row). Differences in distribution between WT and mutant TSS usage were determined by subtracting the normalized WT distribution from normalized mutant distributions. Promoters are separated into those classified as Taf1-enriched (Taf1 Enriched), Taf1-depleted (Taf1 Depleted), or neither (−), and rank ordered on the y-axis based on total reads in WT (from high to low). Gain in relative mutant TSS usage is positive (orange) while loss in relative mutant usage (cyan) is negative. b Polar shifts in TSS usage are apparent for examined rpb1 mutants (except rpb1 F1084I) across promoter classes. All box plots are Tukey plots unless otherwise noted (see the “Methods” section). Promoters examined are n = 3494 (> 200 reads total expression in WT). Pol II mutants are rank ordered by relative strength of in vitro elongation defect (slow to fast) and colored from blue (slow) to green (fast) in similar fashion to allow visual comparison of same mutants between promoter classes. All median TSS shift values for mutants are statistically distinguished from zero at p < 0.0001 (Wilcoxon signed rank test), except F1084I Taf1 Enriched (p = 0.0021) or F1084I Taf1 Depleted (not significant). c Polar shifts in TSS usage are apparent for examined GTF mutants and an rpb1 tfg2 double mutant shows exacerbated TSS shifts relative to the single mutants (compare c to b). Promoters examined are as in b. All median TSS shift values for mutants are statistically distinguished from zero at p < 0.0001 (Wilcoxon signed rank test). d Average TSS shifts in Pol II rpb1 mutants correlate with their measured in vitro elongation rates. Error bars on TSS shifts and elongation rates are bounds of the 95% confidence intervals of the means. Elongation rates are from [82, 84]. Mutants slower than WT in vitro exhibit downstream shifts in TSS distributions while mutants faster than WT in vitro exhibit upstream shifts in TSS distributions. Linear regression line is shown along with the 95% confidence interval of the fit (dashed lines). R squared = 0.8969. Note the log₂ scale on the x-axis. Break in x-axis is to allow Pol II slow mutants to be better visualized. Promoters examined are as in b and c