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

Ruler-regulated sliding: the unifying principle for nucleosome positioning by remodelers. As nucleosome positions are defined by the DNA sequence bound by the histone octamer, all mechanisms that generate consistent nucleosome positions across many genome copies, must select certain DNA sequences in competition with other sequences. As shown here and in the accompanying paper (Oberbeckmann & Krietenstein et al.) , remodelers may mediate this selection in two ways. On the one hand, a remodeler may directly choose a sequence, e.g., INO80 turns DNA shape features into +1 nucleosome positions at promoters (accompanying paper Oberbeckmann & Krietenstein et al.) On the other hand, a remodeler ruler may place a nucleosome at a ruler-determined distance to a barrier, e.g., ISW2 aligns nucleosomes to Reb1 and generates a regular array by aligning a second nucleosome to the first and so on. In the former case, the resulting nucleosomal sequence is directly selected for its sequence features, while in the latter case, it is indirectly selected without regards for its sequence features but merely for its position relative to the barrier, as we show here by using Reb1 sites in S. pombe and E. coli genomes. Our ruler model unifies these positioning mechanisms. The generalized barrier also encompasses DNA sequence elements, with which a remodeler ruler interacts such that sliding direction bias is regulated ( Figure 7C ). This explains observations for hybrid Chd1 remodelers where the Chd1 DBD was replaced with heterologous sequence-specific DBDs (Donovan et al., 2019; McKnight et al., 2011; McKnight et al., 2016) . Such hybrid Chd1 remodelers slide nucleosomes faster towards the cognate site of the heterologous DBD, if it was in reach of this site, until the nucleosome became positioned on the site. In our model, the heterologous DBD is a remodeler ruler. As a DNA sequence element as barrier is no hindrance for nucleosome sliding, the remodeler may slide the nucleosome onto this site. This prevents ruler binding to the site, abolishes the increase in sliding rate linked to ruler binding and makes a nucleosome on the cognate site a poorer nucleosome sliding substrate than at neighboring positions ( Figure 7C , right), which corresponds to the kinetic release model as noted . Our model now adds that sliding from neighboring positions will always (within ruler reach) convene at the cognate site and stabilize this position, even if there is no local sliding rate minimum, as long as the ruler regulates sliding direction bias according to the three key elements outlined above ( Figure 7C, left) In vivo there are many ways that may regulate nucleosome positioning by remodelers, e.g., by recruitment, by architectural factors, by nucleosome density fluctuations or by histone variants and modifications, possibly in the context of elongating polymerases. Nonetheless, we expect that the regulation of nucleosome sliding direction bias via built-in sensing and processing of information in the nucleosome environment, i.e., a remodeler ruler, will be at the heart of each nucleosome positioning mechanism.