Furthermore, maternal presence of ERI-1 was sufficient to ensure uniform expression in homozygous mutant progeny, consistent with previous observations of maternal rescue of some defects (Zhuang and Hunter, 2011). tissue specification. In contrast, in the case of tissues made MIV-247 from a single blastomere (e.g., intestine from the E blastomere), any variation between cells must arise after tissue specification. Thus, tissues such as the intestine provide an opportunity to examine cell-to-cell variation within a tissue after fate specification. Cell-to-cell variation in the activity of genes associated with repetitive DNA has been observed in many animals, often between cells of the same tissue. Repetitive DNA can variably effect the expression of nearby genes in different cells in a process called position effect variegation (PEV) in (Elgin and Reuter, 2013). An early example showed that the location of the gene near repetitive DNA results in a variegated expression such that some cells of the eye express the gene but others do not (Muller, 1930). We now know that such repeat-associated gene silencing can occur through RNA-directed mechanisms associated with chromatin modifications and/or DNA methylation (Volpe and Martienssen, 2011; Elgin and Reuter, 2013). However, the origins of the variation between cells and the developmental mechanisms, if any, that control such variation are unclear. Furthermore, despite repetitive sequences constituting an estimated 45% (Lander et al., 2001) to 69% (de Koning et al., 2011) of the human genome, we do not understand how these large parts of animal MIV-247 genomes are regulated during development. Studies in using repetitive transgenes have provided some insight into expression from repetitive DNA. Genetic screens have identified many conserved factors that promote expression from repetitive DNA through mechanisms that are unclear (Hsieh et al., 1999; Fischer et al., 2013). Insights from the analysis of a few protein factors, MIV-247 however, suggest that expression from repetitive DNA requires the inhibition of RNAi triggered by some form Rabbit Polyclonal to TRAPPC6A of double-stranded RNA (dsRNA). First, loss of the adenosine deaminases acting on RNA (ADAR) enzymes, which deaminate adenosines in dsRNA, results in the silencing of expression from repetitive DNA (Knight and Bass, 2002) and the recruitment of RNAi on many targets (Wu et al., 2011). Second, loss of the exonuclease ERI-1 (enhancer of RNAi-1), which can trim 3 overhangs in dsRNA, causes silencing of expression from repetitive DNA (Kennedy et al., 2004). Third, preventing the spread of forms of dsRNA between cells increases the number of cells that show expression from repetitive DNA (Jose et al., 2009). Fourth, silencing observed upon loss of ERI-1 (Kim et al., 2005) MIV-247 or upon loss of ADAR enzymes (Knight and Bass, 2002) can both be relieved by loss of genes required for RNAi. A curious feature of silencing in many genetic backgrounds that lack is that it varies from cell to cell (e.g., see Fig. S3 in Kim et al.  and Fig. 1 in Jose et al. ). However, the precise source of dsRNA and the source of cell-to-cell variability are unknown. Here, we analyze expression from repetitive DNA in the intestine at single-cell resolution to uncover a source of cell-to-cell variation and to reveal a developmental mechanism that reduces such variation. Results Rearrangements in repetitive DNA generate double-stranded RNA and hairpin RNA To examine repetitive DNA expression in individual cells without the disruption of cellular function or development in repetitive transgene that expresses GFP MIV-247 in all somatic cells, with particularly high levels in intestinal cells. This transgene was generated by transforming worms with a circular plasmid that expresses (Fig. S1 A) and integrating the resultant multicopy array into the genome (first used in Winston et al., 2007). Estimations from Illumina sequencing reads suggested that this transgene had 213 26 adjacent copies of the plasmid (Figs. 1 A and S1 B). Consistent with early experiments (Stinchcomb et al., 1985), we detected abundant inversions and deletions (Fig. 1 B and Fig. S1, CCE) and a few translocations (Fig. S1, D and E) among the copies of the plasmid. The rearrangements were flanked by short sequences with homology (Fig. 1 C), consistent with their generation by recombinases that cause inversions and.