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

The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.02.28.969618 doi: bioRxiv preprint generate linkers of about 20 and 12 bp, respectively, the 18 bp average linker speaks for ISW1a contributing globally more than Chd1. Indeed, lack of Isw1 in vivo globally shortened linkers, while lack of Chd1 affected global spacing only mildly (Kubik et al., 2019; Ocampo et al., 2016) . Locally, high transcription rate correlates with shorter spacing Ocampo et al., 2016) , which points to increased Chd1 contribution, probably due to increased Chd1 recruitment by elongating RNA polymerase (Simic et al., 2003) . Remodeler-specific rulers explain how ISW1a, ISW2 and INO80 affect +1 nucleosome positioning in vivo (Kubik et al., 2019; Parnell et al., 2015; Whitehouse et al., 2007; Yen et al., 2012) and in vitro , especially in combination with RSC. RSC and SWI/SNF are the only yeast remodelers that disassemble nucleosomes (Clapier and Cairns, 2009; Clapier et al., 2017) , particularly at promoter NDRs Brahma and Henikoff, 2019; Ganguli et al., 2014; Hartley and Madhani, 2009; Kubik et al., 2019; Kubik et al., 2018; Parnell et al., 2008; Rawal et al., 2018; van Bakel et al., 2013; . By definition, a promoter NDR has low nucleosome density. Therefore, remodeler rulers will set distances to NDR-bound barriers as measured here at low or medium nucleosome density. In vivo distances between Reb1 and +1 nucleosomes are 60-80 bp ( Figure S3B , (Rhee and Pugh, 2011) ), which are within remodeler-specific distances to Reb1 at medium or low density (81-86 bp for INO80, 70-74 bp for ISW2, 58-60 bp for ISW1a). ISW2 and INO80 contribute more to +1 nucleosome positioning in vivo than ISW1a (Kubik et al., 2019) as their long rulers are more suited for setting long distances across NDRs. Conversely, the short Chd1-ruler hardly contributes to +1 positioning in vivo (Kubik et al., 2019; Ocampo et al., 2016; van Bakel et al., 2013) . These different ruler characteristics explain why ISW1a and Chd1 are mainly involved in spacing nucleosomes into densely packed arrays and why ISW2 and INO80 mainly use their ruler for +1 alignment at NDRs in vivo. This resolves the conundrum why yeast has two remodelers, INO80 and ISW2, that seemingly generate "too wide" spacing compared to average in vivo spacing. We do not preclude that other mechanisms, like recruitment via histone modifications or transcription factors, also affect where each remodeler is active.