Conformational switching of Arp5 subunit regulates INO80 chromatin remodeling

Abstract

The INO80 chromatin remodeler is a versatile enzyme capable of several functions, including spacing nucleosomes equal distances apart, precise positioning of nucleosomes based on DNA shape/sequence and exchanging histone dimers. Within INO80, the Arp5 subunit plays a central role in INO80 remodeling, evidenced by its interactions with the histone octamer, nucleosomal and extranucleosomal DNA, and its necessity in linking INO80's ATPase activity to nucleosome movement. We find two distinct regions of Arp5 binding near the acidic pocket of nucleosomes. One region has an arginine anchor that binds nucleosomes and is vital for INO80 mobilizing nucleosomes. The other region has a hydrophobic/acid patch of Leu and Asp that binds free histone H2A-H2B dimers. These two regions have different roles in remodeling nucleosomes as seen both in vitro and in vivo and the hydrophobic/acidic patch of Arp5 is likely needed for displacing DNA from the H2A-H2B surface and dimer exchange by INO80.

Graphical Abstract

Figure 1.

The arginine anchor of Arp5 is required for binding the acidic pocket of nucleosomes and mobilizing nucleosomes. (A) The conserved arginines and lysines located in the ‘leg’ and ‘foot’ regions of Arp5 are shown. (B and C) These graphs show (B) nucleosome mobilizing and (C) ATPase activities for INO80 containing wild type (WT) or mutant Arp5 (Arp5^R3) with Arg 482, 488 and 491 mutated to Ala. Remodeling and ATPase assays had 50 nM 70N5 601 nucleosomes, 75 nM INO80 and 80 μM ATP. The ATPase activity is assayed using γ-³²P-ATP and thin layer chromatography. (D and E) The interactions of Arp5 near the acidic pocket of nucleosomes at residue 89 of H2A is assayed by site-specific histone crosslinking. A representative phosphorimage of the 4–20% SDS–PAGE is shown in (D) and the bar graph in (E) shows the changes in photoaffinity labeling between WT and Arp5^R3 complexes. There are a total of three replicates and error bars represent the mean ± SD. (F) The in vivo effect of mutating Arp5’s arginine anchor is assayed using growth spot assays and the AA system to acutely test cell viability under conditions that INO80 is required, see also Supplementary Figure S1B. DMSO: demethylsulfoxide; RAPA: rapamycin (3 μg/ml). In panel (F) (top panel), cells were plated on -URACIL synthetic media (bottom panel) cells were plated on YEPD medium. Rapamycin concentration is 3 μg/ml.

Figure 2.

Early during INO80 remodeling the arginine anchor of Arp5 is required for DNA movement on both sides of nucleosomes, but not for DNA displacement on the entry side. (A) The approach for mapping DNA movement and displacement inside of nucleosomes by histone-DNA crosslinking is depicted. Arrows indicate where histone crosslinks to DNA before and during remodeling or is absent. (B) The rates of DNA movement for 20 bp on the entry side is compared to the 32 bp on the exit side for WT INO80.(C and D) The rates of DNA movement on the (C) entry (20 bp) and (D) exit (32 bp) side of nucleosomes is plotted for WT and mutant Arp5^R3 complexes using a photoreactive probe attached to residue 53 of histone H2B. (E) The rate of DNA displacement on the entry side of nucleosomes is tracked by measuring the net loss of DNA crosslinking for WT and Arp5^R3. Three replicates are performed for each experiment and error bars represent the mean ± SD.

Figure 3.

Site-specific histone crosslinking reveals two distinct modes of Arp5 binding to the acidic patch of nucleosomes important for INO80 remodeling. (A) The hemagglutitin epitope (HA) is fused to the C-terminus of Arp5 subunit to aid in peptide mapping and purified INO80 with and without Arp5 HA tagged are analyzed on a 4–20% SDS–PAGE and stained with Coomassie blue. (B) Photoaffinity labeling of WT and HA-tagged Arp5 is shown at two histone positions, residues 89 of histone H2A (H2A89) and 80 of histone H3 (H380). A representative phosphorimage of a 4–20% SDS–PAGE is shown. (C) The structural organization of the ‘cross’ conformation of the Sc Arp5 subunit generated by AlphaFold2 is shown and the regions crosslinked to H2A89 and H380 (see also Supplementary Figure S4C). (D) Residues 516–565 and 565–620 are crosslinked to residue 89 of histone H2A; whereas (E) residues 385–515 of Arp5 crosslink to residue 80 of histone H3 and the crosslinked regions are highlighted. (F and G) Graphs show the (F) nucleosome mobilizing and (G) ATPase activities of Arp5 and Arp5^Δ516–564 complexes and the error bars show the mean ± SD with 3 replicates (see also Supplementary Figure S5C). (H) The interactions of Arp5 with the histone octamer face of nucleosomes are probed by site-specific histone crosslinking for WT and mutant Arp5^Δ516–564 (see Supplementary Figure S5B). The relative efficiency of crosslinking for Arp5, Ies2 and Ies6 is shown for three replicates and error bars represent the mean ± SD.

Figure 4.

Leu 567, Leu 568 and Asp 571 (LLD) of Arp5 are required for INO80’s in vivo activity. (A) Growth spot assays are shown with WT, Arp5 (ΔArp5) and Arp8 (ΔArp8) deleted strains including those with plasmids expressing WT or mutant Arp5 / Apr8 as indicated. Arp8^ΔN has amino acids 1–197 removed from the N-terminus of Arp8. (B) The conserved residues between amino acid 514 and 571 of Sc Arp5 are highlighted for identical or similar amino acids. Sc (CAA95933.1), Schizosaccharomyces pombe (CAB44762.1), Thermochaetoides thermophila (XP_006693704.1), Drosophila Melanogaster (NP_650684.1), mouse (NP_780628.3) and human (NP_079131.3). (C) Spot assays are shown for scanning the effect of mutating the conserved residues in (B) to alanine with endogenous Arp5 being tagged for AA (Arp5^AA) and mutant Arp5 complementing for the loss of Arp5. DMSO: demethylsulfoxide; RAPA: rapamycin (3 μg/ml), and others as in Figure 1. In panel C (top) cells were plated on -URACIL synthetic media (bottom) cells were plated on YEPD medium.

Figure 5.

Leu 567, Leu 568 and Asp 571 (LLD) is required for binding H2A–H2B dimers, Arp5 binding to the acidic pocket of nucleosomes and INO80 mobilizing nucleosomes. (A) The nucleosome remodeling activity of WT and mutant Arp5^LLD, Arp5^DBD and Arp5^R3 complexes are measured using EMSA and the extent of mobilized nucleosome plotted versus time (see also Supplementary Figure S7B). (B) Relative histone crosslinking efficiency is plotted for Arp5 near the acid pocket (H2A89) and another lateral position on the histone octamer (H380) for WT and mutant Arp5^LLD, Arp5^DBD and Arp5^LLD+DBD (see also Supplementary Figure S8C). (C and D) Pulldown assays are done with (C) nucleosomes and (D) H2A/B dimer using biotinylated LANA peptide and WT and mutant Arp5 Arg anchor and LLD peptides. DNA in the nucleosomes is P³² radiolabeled and is also tracked in the nucleosome pulldown assays as shown in the lower panel of (C). (E) Nucleosomes are preincubated with 6, 18 and 50 μM of LANA, WT or mutant LLD peptides for 30 min before adding 75 nM INO80 and incubating for 30 min. Remodeling reactions are stopped 10 min after addition of 80 μM ATP and analyzed on a 5% native polyacrylamide gel.

Figure 6.

Mutating Leu 567 and 568 and Glu 571 to Ala is not expected to perturb the structure of the grappler domain or the interactions of the arginine anchor. (A–C) Models showing the composition and placement of residues in the ‘arm’, ‘foot’ and ‘leg’ regions of the grappler domain along with the average distances between key residues. (D) Shown are the structural models generated by AlphaFold3 of Arp5’s grappler domain for WT and mutated LLD, that illustrate no predicted change in the grappler domain due to mutation of Leu 567 and 568 and Glu 571 to Ala.

Figure 7.

The LLD patch of Arp5 is most critical for DNA movement and displacement on the entry side of nucleosomes. (A–D) DNA movement is tracked on the (A and B) exit and (C) entry sides of nucleosomes using a photoreactive probe attached to residue 53 of histone H2B and (D) DNA displacement on the entry side with INO80 containing WT Arp5 and mutant Arp5^DBD, Arp5^LLD and Arp5^R3 as indicated. The movement of (A and B) 32 bp on the exit side and (C) the loss of the initial crosslinking position on the entry side is plotted versus time with graph (B) showing the reactions over a more extended time. In (D), the net loss of DNA crosslinking on the entry side is plotted. Three replicates are performed for each condition and error bars represent the mean ± SD.