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 mutant complexes reveal a multilayered ruler. All mutant complexes were assayed like the wild type (WT) INO80 complex ( Figures 5A-E, 6A -D, S1A, S5A,B). WT INO80 was assayed again alongside with matching SGD chromatin. Comparing these replicates ( Figure 5C ) with previous values for WT INO80 ( Figure 2D ) reflected variability in preparing SGD chromatin but at the same time the robustness of the overall effects. All tested INO80 mutants generated steady-state patterns ( Figure S5B ) and differed from WT INO80 in forming aligned arrays in the following ways. First, all mutants, besides the HQ1/2 mutant, which was almost inactive (Figure S5A ), as expected , generated phased regular arrays, but with varying effectiveness and altered distance to one or both barrier types and/or linker lengths compared to WT INO80 ( Figures 5D,E, 6D ). This revealed that also INO80 has a ruler, to which both the Arp8 and the Nhp10 module contribute. Second, the HQ1 showed stronger effects than the HQ2 mutation ( Figure 5D ). Both increased the distances to both barriers. While HQ2 increased linker length at all densities, HQ1 gained clamping activity, i.e., linker length hardly depended on nuclesome density. Both mutations uncoupled distance to DNA ends from linker lengths, in contrast to WT INO80 ( Figure 2D,E) . Only for HQ1, linker 1 length depended on barrier type ( Figure 5B ). We concluded that the Arp8 module, especially via HSAa1 helix-DNA interactions, is threefold involved in spacing, alignment to barrier and responding to nucleosome density. Third, the Nhp10 module subunits contributed to the ruler mainly through the HMG box of Nhp10 as the respective point mutations (HMGII mutant) mimicked the effects upon lack of all Nhp10 module subunits (∆Nhp10 mutant) ( Figure 6C ,D). With these mutations, distances to both barriers were not much affected, but linker length depended less on density, i.e., clamping was gained, similar to the HQ1 mutation. Effects of the combined HMGII-HQ1 and -HQ2 mutations were dominated by the HQ mutations, but with reduced effects on distance to barriers ( Figure 5E ). Even though the Nhp10 HMG box was a prime candidate for sensing extranucleosomal DNA, its contribution was minor compared to the HSA helix contribution. Fourth, the INO80 ∆N mutation affected the distance to Reb1 and even more to DNA ends, but gained clamping less strongly than the HMGII or ∆Nhp10 mutations ( Figure 6D ). The INO80 ∆N mutant lacked the complete Nhp10 module, but also the Ino80 ATPase N-terminus and Taf14 ( Figure S1A ), which may account for the differential effects.