Figure 2. ncRNA-dependent and independent nucleation yields qualitatively different spreading reactions in the MAT locus.

(A) Diagram of the reporters within MAT^HSS and ΔREIII^HSS. WT and m for REIII indicate the presence or deletion of the Atf1/Pcr1 binding sites, respectively. (B) 2D-density hexbin plot showing the ‘red’-normalized ‘green’ and ‘orange’ fluorescence for wild-type MAT^HSS cells. Scale bar shows every other bin cutoff as a fraction of the bin with the most cells. Inset: histogram of the ‘red’-normalized ‘orange’ fluorescence distribution of ‘green'^OFF cells. (C) 2D-density hexbin plot and inset as above for ΔREIII^HSS, which contains two 7 bp Atf1/Pcr1-binding site deletions (m) within the REIII element. (D) ChIP for H3K9me2 (red) and H3K9me3 (grey) for amplicons indicated in (A). normalized to dh. WT, wild-type MAT^HSS, m, ΔREIII^HSS. (E) TOP: diagram of the reporters within ΔK^HSS. The cenH nucleator and additional 5’ sequence is deleted and replaced by ‘orange’. ‘green’ is located directly proximal to REIII and serves as the nucleation clamp. ChIP amplicons are indicated as black bars. BOTTOM: 2D- density hexbin plot and inset as above. LEFT: ChIP for H3K9me2 (red) and H3K9me3 (grey) for ‘green’ and ‘orange’ in isolated ΔK^HSS-ON or ΔK^HSS-OFF alleles. In hexbin plots, the Δclr4 derivative of each strain was used to normalize the X- and Y-axes to = 1. Error bars indicate standard deviation of technical replicates.

Figure 2—figure supplement 1. Heterochromatin spreading characteristics of cis-acting elements at the tightly repressed MAT locus.

(A) The MAT^HSS documents tight repression of the wild-type MAT locus. As in Figure 2A and B, with ‘green’ and ‘orange’ switched. (B) Stochastic spreading with intermediate states in pcr1::KAN. pcr1 transcription factor was knocked-out in the PAS217 wild-type MAT^HSS. Plot and inset as in Figure 2B. (C) REII does not contribute to bimodal distribution seen for ΔK^HSS. The REII locus (1 kb) was replaced with the LEU2 gene before clr4+ was introduced by cross. (D) REIII is unable to establish spreading at an ectopic site. 2D density hexbin plots of ura4::REIIIHSS^5kb. Normalized green and orange are near 1.0, indicating a failure to repress both reporters. Inset: 2D density hexbin plots of ura4::REIIIHSS^5kb dcr1::KAN. dcr1 was deleted to release extra heterochromatin factors from RNAi- repressed loci. No additional silencing is detected.

Figure 2—figure supplement 2. REIII is required for heterochromatin formation in ΔK^HSS.

(A) Deletion of both Atf1-/Pcr1-binding sites before introduction of clr4+ in ΔK^HSS blocks gene silencing. In 34/34 strains tested (one representative shown), ΔK^HSSΔs1Δs2 cannot form repressed states. (B) H3K9me2 does not accumulate when both Atf1/Pcr1-binding sites are deleted in ΔK^HSS. H3K9me2 ChIP in ΔK^HSSΔs1Δs2 at ‘green’, ‘orange’ and dh. (ΔK^HSS-OFF accumulates H3K9me2 to similar extent as dh, Figure 2E). Error bars indicate standard deviation of technical replicates. (C) ‘green’ orientation and position does not substantially affect ΔK^HSS behavior. In ΔK^HSS G^flipped‘green’ is flipped in orientation with respect to ΔK^HSS. (D) ‘green’ and ‘orange’ orientations do not substantially affect ΔK^HSS behavior. In ΔK^HSS G^flipped O^flipped‘green’ is located as in C and ‘orange’ is flipped in orientation with respect to ΔK^HSS. ‘green’ in (C) and (D) is 2.1 kb downstream from its location in ΔK^HSS now on the distal side of the mat3m cassette. (E) Increasing distance between REIII and ‘orange’ does not substantially affect ΔK^HSS behavior. The Atf1/Pcr1-binding site proximal to ‘orange’ was deleted (Δs1) and 700 bp of the sib1 ORF inserted to the left of the Δs1 site. 2D-hexbin plots as in Figure 2.