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The copyright holder for this preprint (which was not peer-reviewed) is the .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 org/10. /2020 input, the overall regulation of sliding direction bias by the ruler will share three key elements that constitute the ruler mechanism. First, the ruler has a certain reach (regions A + B in Figure 7A ), within which it interacts with the barrier. Second, if the position, from where the remodeler slides the nucleosome, is within region B, the interaction between ruler and barrier biases overall sliding direction towards the barrier (red curve is above green curve), e.g., due to binding energy gained upon orienting the remodeler towards vs. away from the barrier. Third, if the nucleosome is in region A, the ruler-barrier interaction disfavors sliding towards relative to sliding away from the barrier (green curve is above red curve), e.g., because the ruler gets sterically in the way. Our study determined the length of region A for different remodeler and barrier types and conditions. Region B and exact curve shapes will have to be determined in future studies. If these three key elements are met, resulting fluxes lead to steady-state nucleosome placement at a defined position relative to the barrier (stippled vertical arrows throughout Figure 7 ). This position is a selfstabilizing dynamic equilibrium point (intersection of red and green curves) without sliding direction bias here, but with biases towards this point from neighboring positions. This model applies to how a remodeler with ruler stably positions a nucleosome next to a GRF as well as to another nucleosome and therefore explains both spacing and phasing. It also explains density-independent clamping. As long as a remodeler is processive enough to fortuitously bring nucleosomes into region B of a barrier also at low density, the ruler mechanism will keep the nucleosome at the dynamic equilibrium point. Nonetheless, the model can also accommodate sensing of nucleosome density and barrier type, e.g., if the ruler offers a hierarchy of interaction points (Brahma et al., 2018) . INO80 mutants showed not concerted but uncoupled effects on distances to Reb1, DNA ends and nucleosomes, even if the same module, like the Nhp10 module, was differentially mutated. Chd1 generated shorter linker lengths (12-16 bp) than distances to DNA ends or Reb1 (35-40 bp). For Chd1, Reb1 may be a "hard" barrier while nucleosomes are "soft" barriers as they are partially "invaded" by the ruler. Indeed, Chd1 partially unwraps nucleosomal DNA (Farnung et al., 2017) . The way how different remodeler rulers interact with different barriers requires clarification, and we outline our model (Figure 7) in terms of extension-less point particles, but actual footprints have to be taken into account. The model is fully compatible with the ruler, i.e., the DNA binding domain (DBD) of Chd1 or Drosophila ACF (Yang et al., 2006) , introducing bias via sensing extranucleosomal DNA length. Indeed, differently long extranucleosomal DNA in mono-or oligonucleosome sliding assays amounts to different distances to barriers like DNA ends or other nucleosomes. Our model is fully consistent with previous data and models but offers an alternative interpretation and is more widely applicable, e.g., to stable nucleosome positioning at only one barrier and not only in-between two barriers. We introduced our model in terms of overall sliding direction bias. More specifically, the model may refer to differential regulati