Ruler elements in chromatin remodelers set nucleosome array spacing and phasing

Elisa Oberbeckmann; Vanessa Niebauer; Shinya Watanabe; Lucas Farnung; Manuela Moldt; Andrea Schmid; Patrick Cramer; Craig L. Peterson; Sebastian Eustermann; Karl-Peter Hopfner; Philipp Korber

INO80, ISW2, ISW1a and Chd1, but not Fun30 align regular arrays at the barrier Reb1. We tested all yeast remodelers with known spacing activity, INO80, ISW2, ISW1a and Chd1 Lusser et al., 2005; Stockdale et al., 2006; Torigoe et al., 2013; Tsukiyama et al., 1999; Udugama et al., 2011) as well as the Fun30 remodeler, for which it was unclear if it has spacing activity (Awad et al., 2010) . INO80, ISW2, ISW1a and Chd1, each in combination with Reb1, generated phased regular arrays at promoters with Reb1 sites (red shaded top of heat maps in Figure 1C ), while Fun30 did not ( Figure S1B ). This clarifies that Fun30 does not have regular array generation and alignment activity. Previously, Chd1 purified from budding yeast did not show much effect in genome-wide reconstitutions . This was maybe due to full-length Chd1 tending to aggregate in vitro, which is why truncated Chd1 constructs were often used Patel et al., 2011) . Here, we leveraged our finding that recombinant full-length Chd1 is stabilized in complex with recombinant FACT complex (Farnung et al., 2017) and achieved in vitro array generation and alignment also by Chd1. The heat map patterns ( Figure 1C ) and even more the corresponding composite plots for the Reb1bound genes only ( Figure 1D ) suggested that the distance of arrays to the barrier Reb1 as well as the linker lengths varied with nucleosome density in a remodeler-specific way. For all remodelers with spacing activity, array extent increased with growing density, consistent with greater nucleosome availability and processive spacing activity. Array extent at high density was larger than in our previous reconstitutions , i.e., we achieved higher densities here. Adding more remodeler after half of the incubation time did not change the array distances of resulting patterns confirming non-limiting remodeling activity and steady state conditions ( Figure S1C ).