dig1Δ, but not dig2Δ, cells display increased noise in yeast mating pathway outputs. a. Schematic of the yeast mating MAPK pathway. Note: for simplicity the Ste12-target gene is illustrated as having one Ste12-binding site. b. Mean output for pFUS1-YFP, pAGA1-YFP, and pPMP1-GFP in wild-type (blue), dig1Δ (red), dig2Δ (green), and dig1Δdig2Δ (black) mutants in absence of α-Factor. Error bars indicate the standard deviation of three replicate experiments. The Y-axis is broken between 10,000 and 20,000 AU. c. Bar graphs illustrating the coefficient of variation (CV) for each strain as in b. The Y-axis is broken between 0.7 and 0.8. T-test was used to calculate P = 0.0003 for difference between pAGA1-YFP and pAGA1-YFP dig1Δ and P = 0.0014 for difference between pFUS1-YFP and pFUS1-YFP dig1Δ. d. Probability density functions (PDFs) of wild-type (blue) dig1Δ (red) and dig2Δ (green) for each reporter: pFUS1-YFP (left), pAGA1-YFP (middle) and pPMP1-GFP (right). Solid lines represent the average PDF for three replicates while the envelope indicates the standard deviation. (b-d. Data shown is for gate 5, see Methods.) e. CV vs. gate radius for pFUS1-YFP strains (left), pAGA1-YFP strains (middle) and pPMP1-GFP strains (right). b-e. See Methods for gate sizes and numbers of cells analyzed. Error bars represent the standard deviation of three replicate experiments.

Sorted dig1Δ cells can regenerate the entire pAGA1-YFP output distribution. dig1Δ cells expressing mean levels of pAGA1-YFP were sorted and re-grown over time. At t = 0, 60, 120, 180, 240 and 300 min, cells were removed and the fluorescence distribution was determined.

Extrinsic and intrinsic noise in output of mating pathway increase in dig1Δ cell populations. A. Schematic of two-colour experiment. pAGA1-GFP is in the endogenous AGA1 locus, while pAGA1-mCherry-AGA1 3’UTR is inserted into the LYS1 locus. B. Density plots of wild-type (left), dig1Δ (middle) and dig2Δ (right). c. Quantification of intrinsic, extrinsic and total noise of wild-type (blue), dig1Δ (red), dig2Δ (green) populations. Each value is the mean of three replicates and error bars indicate the standard deviation. T-test was used to calculate P = 0.035 for increase in intrinsic noise and P = 0.009 for increase in extrinsic noise in dig1Δ mutant. Total noise was calculated as . d,e. Plots of intrinsic (d) and extrinsic (e) noise vs. gate radius to control for cell size and shape. Error bars indicate the standard deviation of three replicates. f. Quantification of intrinsic, extrinsic and total noise in Ste12-independent reporter strain in which two fluorophores are driven by pPMP1 (inset).

Increased noise in mating pathway outputs in dig1Δ cells is not due to feedback at the STE12 promoter. a. Model for the increased noise in the mating pathway in dig1Δ cells. b. pSTE12 was replaced with the Ste12-independent promoter pEAF3 (above). PDFs of pAGA1-YFP strains containing pEAF3-STE12 (below). c. pSTE12 was replaced with the Ste12-independent promoter pTAF4 (above). PDFs of pste12::pTAF4 strains (below). b,c. wild-type in blue, dig1Δ in red and dig2Δ in green. Solid line is the mean of three replicate experiments and the envelope reflects standard deviation of three replicates.

Ste12-GFP forms nuclear foci in dig1Δ cells. a. Fluorescence microscopy images of Ste12-GFP/Nup188-mCherry (left) and Reb1-GFP/Nup188-mCherry (right) from top to bottom: wild-type, dig1Δ, and dig2Δ. Ste12-GFP forms nuclear foci in dig1Δ (see white arrow heads). b. Quantification of foci seen in (a) n = 100 (wt), n = 116 (dig1Δ), n = 95 (dig2Δ). Distributions of foci in all mutants were compared to wild-type by the Chi-square test and only the distribution of foci in dig1Δ was statistically significant (P < 0.001). c. Ste12-GFP localization upon addition of 5 μm α-Factor for 1 hr. d. Normalized ChIP signal of Dig1-GFP at AGA1, FUS1 and LYS1 promoters in absence and presence of 5 μM pheromone. Scale bar for all images is 5 μm.

ChIP-chip of Ste12-target locus reveals long-range interactions with other Ste12-target genes. a. Experimental set-up. LacI-mCherry was immunoprecipitated from cells and the FUS1 locus (green), marked with an array of Lac operators (light blue), was efficiently pulled down (see peaks centred on Lac operators in graph below) in wild-type (blue), dig1Δ (red) and dig1Δste12Δ (orange) cells. The graph below illustrates the ChIP-chip signal along chromosome 3, where the array of Lac operators is inserted. See Methods for experimental details. b. Difference maps were calculated and genes were aligned by increasing median value of the region spanning −500bp to 0bp, with respect to the translation start site. The top 5% of differences (269 genes, Supplementary Information, Table S1) were analyzed for enrichment of target genes of 203 transcription factors21. The dashed line indicates the Bonferroni-corrected P value of 0.05. c. Bonferroni-corrected P values for enrichment of Ste12- and Tec1-target genes in the top 5% of genes with the greatest differences in dig1Δ vs. WT and dig1Δste12Δ vs. WT (Supplementary Information, Table S1). d. Ste12- and Tec1 target genes (in presence and absence of pheromone21) found in the list of 5% of genes with the largest differences in dig1Δ vs. WT datasets. P values calculated by hypergeometric testing.

dig1Δ cells display defects in growth and cell-cell fusion. a. Log phase wild-type (blue squares), dig1Δ (red circles), dig2Δ (green triangles) and dig1Δste12Δ (orange diamonds) cells were grown for 9 hrs and OD₆₀₀ was measured every hour. b. Error in growth. OD₆₀₀ of mutant (dig1Δ in red, dig2Δ in green and dig1Δste12Δ in orange) was compared to that of wild type by calculating: 1 – [OD₆₀₀(mutant)]/OD₆₀₀(wild-type)]. c. Schematic for FACS-based cell fusion assay. See Methods for details. d. Fraction of GFP+/mCherry+ cells over time for wild-type (blue), dig1Δ (red), dig2Δ (green) and fus3Δ (pink). Samples were analyzed at 0, 60, 90, 120 and 150 min. Error bars indicate the standard deviation of three replicates. e. Cell-cell fusion error for mutants dig1Δ in red, dig2Δ in green and fus3Δ in pink was calculated in the following manner: 1 – [(fraction mutant fused)/(fraction wild-type fused)].

Time course of induction of pAGA1-YFP and pFUS1-YFP after treatment with pheromone. a. Heat map of induction of pAGA1-YFP (left) and pFUS1-YFP (right) in wild-type and dig1Δ cells. Samples were analyzed at 0, 15, 30, 60, 90, 120, 150, 180 and 240 min. b. left: Probability density functions for 150 min time point of pAGA1-YFP (dark blue), pAGA1-YFP dig1Δ (dark red), pFUS1-YFP (light blue) and pFUS1-YFP dig1Δ (pink). The inset is a blow-up of the area marked by the grey-dashed line. right: The percentage of cells in the low fluorescence population versus time. c. Mean fluorescence of the transcriptionally induced populations in part a versus time. Transcriptionally induced and un-induced populations (see a) were separated and the means of the high expressing populations were calculated for each time point. The colours are as in part b. a-c. Data shown are from a single representative experiment, but results have been replicated in three separate experiments.