Then, to track distal events, we inserted orange at one of several sites downstream from green (and a euchromatic region (Garcia et al., 2015). fission yeast using a system that quantitatively describes the spreading reaction in live single cells. We show that spreading triggered by noncoding RNA-nucleated elements is stochastic, multimodal, and fluctuates dynamically across time. This lack of stability correlates with high histone turnover. At the mating type locus, this unstable behavior is restrained by an accessory safeguards epigenetic memory against environmental perturbations. Our results suggest that the most prevalent type of spreading, driven by noncoding RNA-nucleators, is epigenetically unstable and requires collaboration with accessory elements to achieve high fidelity. elements that directly recruit H3K9me. (1) which is related to the and repeats at the pericentromere and at the subtelomere (Grewal and Klar, 1997; Hansen et al., 2006). These sequences nucleate H3K9me by at CEP dipeptide 1 least two pathways, which depend on transcription of noncoding RNAs (ncRNAs): the RNAi pathway (Hall et al., 2002; Volpe et al., 2002), and at least one separate pathway dependent on nascent RNA polymerase II transcripts, which requires the budding yeast Nrd1 homology Seb1 (Marina et al., 2013) (collectively ncRNA-nucleation). Separately and unique to the MAT locus, (2) a region downstream of including CEP dipeptide 1 the element, which recruits the H3K9 histone methylase, HP1 proteins and histone deacetylases (HDACs). This is dependent on cells. Using the HSS, we show that ncRNA-dependent elements trigger epigenetically unstable spreading that is stabilized by an accessory RNA-independent gene promoter ((H3K9 methyltransferase. We show that in the absence of heterochromatin, expression of the noise reporter (red) correlates well with that of reporters for both nucleation (green) and spreading (orange) (Figure 1figure supplement 1A,B), especially when all cells in the population are considered without applying a size gate (Figure 1figure supplement 1B, ?~0.83C0.93). This analysis mode is required when cell number is limiting. When a smaller subset is considered where all the cells are of similar size and stage of the cell cycle, the correlation still provides useful noise filtering (Figure 1figure supplement 1A), which becomes evident when the normalization is applied to cells that CEP dipeptide 1 fall in the size gate (Figure 1figure supplement 1C). Thus, cellular noise is mitigated by dividing the signals from the proximal green and distal orange heterochromatic reporters by the signal of the red, euchromatic reporter (green/red; orange/red). Together, these elements constitute our heterochromatin spreading sensor (HSS) (Figure 1A). Open in a separate window Figure 1. Heterochromatin spreading from ncRNA-nucleated elements is stochastic and produces intermediate states.(A)?Overview of heterochromatin spreading sensor. Three transcriptionally encoded fluorescent proteins are inserted in the genome: The clamp site enables isolation of successful nucleation events, the sensor reports on spreading events and the noise filter normalizes for cell-to-cell noise. (B) Overview of the visualized CEP dipeptide 1 by the CEP dipeptide 1 HSS with orange inserted at different distances shown in (B). The red-normalized orange fluorescence distribution of greenOFF cells plotted on a histogram. Inset: 2D-density hexbin plot showing red-normalized green and orange fluorescence within the size gate, with no green or orange filtering. The green’OFF population is HDM2 schematically circled. The fluorescence values are normalized to?=?1 for the derivate of each strain.?(D) TOP: cartoon overview of the FACS experiment for D. and E. green’OFF cells collected from the Error bars indicate standard deviation of two replicate RNA isolations. (E) ChIP for H3K9me2 and H3K4me3 in the same populations as (D). Each ChIP is normalized over input and scaled to?=?1 for a positive control locus (repeat for H3K9me2 and promoter for H3K4me3). Error bars indicate standard deviation of two technical ChIP replicates. Primer pairs for RT-qPCR and ChIP are indicated by solid and dashed line, respectively, in the C. or with (Red) or (High Red) in HSS size-gated cells. LEFT: Plots of green and orange vs. red channel signals of size-gated PAS 135 (or with (Red) or (High Red) in HSS in cells without size gate. Plots and Pearson correlation as above. (C) Effect of red-normalization on distribution of HSS cells. Plots of green and orange vs. red channel signals of PAS 136, which contains the ectopic HSS (Figure 1C). LEFT: effect of using only size gate, without red normalization. RIGHT: effect of red-normalization with and without additional size gate. The distribution of cells is tightened.