Technical Comments

Comment on “A histone acetylation switch regulates H2A.Z deposition by the SWR-C remodeling enzyme”

See allHide authors and affiliations

Science  22 Jul 2016:
Vol. 353, Issue 6297, pp. 358
DOI: 10.1126/science.aad5921


Watanabe et al. (Reports, 12 April 2013, p. 195) study the yeast SWR1/SWR-C complex responsible for depositing the histone variant H2A.Z by replacing nucleosomal H2A with H2A.Z. They report that reversal of H2A.Z replacement is mediated by SWR1 and related INO80 on an H2A.Z nucleosome carrying H3K56Q. Using multiple assays and reaction conditions, we find no evidence of such reversal of H2A.Z exchange.

Histone variant H2A.Z is deposited in chromatin by the adenosine triphosphate (ATP)–dependent chromatin remodeling complex SWR1/SRCAP/Tip60. Budding yeast SWR1 (also called SWR-C) catalyzes unidirectional, stepwise replacement of two histone H2A-H2B dimers on the canonical nucleosome with H2A.Z-H2B dimers (14). Watanabe et al. reported that histone exchange in the reverse direction—the replacement of nucleosomal H2A.Z-H2B with H2A-H2B dimers—is mediated by SWR1 and the related INO80 complex on nucleosomes harboring the histone H3K56Q substitution, a mimic of H3K56 acetylation (5). We find no evidence of such reverse activity for SWR1 or INO80.

We used three methods to detect histone exchange. In an electrophoretic mobility shift assay (EMSA) (3), replacement of nucleosomal H2A-H2B or H2A.Z-H2B with histone dimers bearing a 3xFlag tag gives a mobility shift of the mononucleosome band on nondenaturing polyacrylamide gel electrophoresis (PAGE), one shift per 3xFlag (Fig. 1). On the (H2A.Z-H2B-H3K56Q-H4)2 (ZQ) nucleosome, SWR1 catalyzed no exchange of H2A.Z-H2B for the H2A-H2B-3xFlag dimer (A-Flag). This was true under the reaction conditions of Watanabe et al. (5) (Buffer I) or Ranjan et al. (3) (Buffer II) (Fig. 1, A and B), but the same dimers could be nonenzymatically exchanged by salt gradient dialysis (Fig. 1C). As previously established on the canonical H2A nucleosome containing wild type histone H3K56 (AK nucleosome), SWR1 catalyzed ATP-dependent exchange of the H2A-H2B dimer for H2A.Z-H2B-3xFlag (Z-Flag) (44% exchange, Buffer I; 58% exchange, Buffer II) (Fig. 1, A and B). These results obtained by F.W. were confirmed by A.R. (Fig. 1D).

Fig. 1 Histone exchange by SWR1.

(A) EMSA shows Cy5-DNA (208 bp) (12) nucleosome products (3) of forward and reverse H2A.Z exchange. Typical 10-μl reaction contains 1 mM ATP, at 30°C for 1 hour in Buffer I [70 mM NaCl, 10 mM Tris-HCl (pH8.0), 5 mM MgCl2, 0.1 mg/ml bovine serum albumin (BSA), and 1 mM dithiothreitol (DTT)]. A 4-μl stop solution (1 mg/ml salmon sperm DNA, 50 mM EDTA, 5 mM ATPγS) was added before loading on 6% native mini-PAGE gel in 0.5X tris-borate EDTA, followed by Cy5 scanning. Forward reaction % replacement = %AZ band/2 + %ZZ band. (B) As in (A), using Buffer II (20 mM HEPES-KOH pH 7.6, 0.3 mM EDTA, 0.28 mM EGTA, 4% glycerol, 0.014% NP40, 0.8 mM DTT, 56 mM KCl, 5.6 mM MgCl2, 80 μg/ml BSA, and protease inhibitors). (C) EMSA nucleosome markers harboring two H2A.Z-H2B (ZZ), one H2A-H2B-3xFlag plus one H2A.Z-H2B (AFZ), and two H2A-H2B-3xFlag dimers (AFAF), produced by salt gradient dialysis after adding A-Flag dimer to the ZQ nucleosome reconstitution mixture. (D) Nucleosomes after forward and reverse H2A.Z exchange by SWR1 and INO80. (E) Anti-Flag Western blot-chemiluminescence of the nondenaturing PAGE gels in (D) (2, 5). (F) As in (A), except for H2A-K121C-Cy3-H2B dimer (A-Cy3) or H2A.Z-K120C-Cy3-H2B (Z-Cy3). *, may be an asymmetrically positioned nucleosome. (G) As in (F) using Buffer II.

Watanabe et al. detected histone exchange indirectly by Western blotting after native PAGE (1, 5). We found no evidence for reversal of H2A.Z exchange when the PAGE gel of Fig. 1D was processed for Western blot-chemiluminescence imaging (Fig. 1E). To eliminate the possibility of tag interference, we used a fluorescently labeled histone dimer free of epitope tags for histone exchange. SWR1 mediated no incorporation of Cy3-labeled H2A-H2B (A-Cy3) on the ZQ nucleosome but substantial ATP-dependent incorporation of Cy3-labeled H2A.Z-H2B (Z-Cy3) on the AK nucleosome in Buffer I and Buffer II (Fig. 1, F and G).

The INO80 complex mediated no exchange of nucleosomal H2A.Z-H2B for H2A-H2B-3xFlag on a centrally positioned ZQ nucleosome (Fig. 2A), whereas INO80 displayed its known histone octamer sliding activity from one end of a long-linker nucleosome to the mobility-retarded central position (Fig. 2B). This absence of H2A.Z reverse exchange demonstrated by F.W. (Fig. 2A) was confirmed by A.R. using EMSA (Fig. 1D) and Western blot-chemiluminescence (Fig. 1E). Furthermore, contrary to the report of Watanabe et al. (5), we detected no reversal of H2A.Z exchange using SWR1 complex purified from a Swc2Δ strain, which is deficient for the major DNA- and H2A.Z-binding component of SWR1 (3, 6) (Fig. 2C).

Fig. 2 Analysis of INO80 and SWR1 activities.

(A) Nucleosome products of H2A.Z exchange by INO80, as in Fig. 1A. (B) Histone octamer sliding (13) of end-positioned ZQ nucleosome (60-bp linker) in Buffer I at 30°C for 1 hour. (C) SWR1 (Swc2Δ) mutant exhibits neither reverse nor forward H2A.Z exchange, as in Fig. 1A. (D) Nucleosomes after reverse H2A.Z exchange by SWR1, as in Fig. 1A, except that ZQ nucleosomes were reconstituted with no subsequent λ DNA treatment, and with (*) and without (**) sucrose gradient purification. (E) H2A.Z (ZZ) nucleosomes (Fig. 1A), mixed with increasing amounts of nucleosomes bearing A-Flag dimers produced by salt dialysis (Fig. 1C). % band intensities (Cy-5) of AFAF nucleosome over AFAF + ZZ nucleosome indicated. (F) Forward H2A.Z exchange by SWR1 on AK and AQ nucleosomes, as in Fig. 1A. (G) EMSA and SYBR Green I DNA stain of purified nucleosomes in sucrose gradient fractions.

The source of these discrepancies is unclear, but it is not due to the use of different affinity tags for enzyme purification, as INO80 from the Peterson laboratory also showed no reversal of H2A.Z exchange in our hands. It is unlikely to be due to different nucleosome preparations, as two reconstitution procedures gave the same results (Figs. 1, A and B, and 2D). Under the same buffer conditions and enzyme-substrate concentrations, we find no evidence for their claim of 30% nucleosomal H2A.Z exchange for H2A, within the detection limits of quantitative EMSA (~0.03% H2A-H2B-3xFlag) (Fig. 2E) or Western blot-chemiluminescence.

Watanabe et al. cite reduced promoter occupancy in vivo by H2A.Z in the H3K56Q yeast strain as evidence consistent with SWR1-mediated reverse H2A.Z exchange on the ZQ nucleosome. Although we observed no reverse exchange, we detected ~two-fold lower H2A.Z exchange in the forward direction on the AQ nucleosome (Fig. 2F), consistent with their in vitro observations (5). Because steady-state H2A.Z occupancy is a function of deposition and eviction pathways, the lower forward H2A.Z exchange could suffice to explain reduced H2A.Z occupancy for the H3K56Q strain.

Our findings conflict with Peterson and colleagues’ claims of reversal of H2A.Z exchange by SWR1 and INO80. Irreproducibility of their findings despite use of three detection modalities compels consideration of other mechanisms for H2A.Z eviction.

Experimental procedures were as follows: Preparation of nucleosomes, SWR1-Flag or INO80-Flag complexes, and H2A/H2A.Z-H2B dimers was as published (1, 3, 7). DNA purity of nucleosomes were reconstituted by our laboratory protocol in Fig. 2G. Histone exchange assays were as published (2, 5, 8, 9).

Recombinant yeast histones were purified individually following the Tsukiyama Laboratory ( or copurified as a dimer (H2A-H2B) (7). Nucleosomes were reconstituted by mixing yeast histones and unlabeled or Cy5-labeled 208 base pairs (bp) Widom 601 DNA (31N30) or Cy5-N60 DNA, followed by salt gradient dialysis according to Luger et al. (10). Cy5-labeled reconstituted nucleosomes were subsequently treated with λ DNA (1 μg λ DNA per 4 pmol 601 DNA, incubated for ~2 hours at 4°C) to bind residual unincorporated histones and purified by 5 to 20% sucrose gradient sedimentation (100 μl to 150 μl sample load per 4.8 ml sucrose gradient, Beckman SW55Ti rotor, 30,000 rpm, 17.1 hours, 5°C) following Ranjan et al. (3). Purity of nucleosome fractions was confirmed by EMSA and SYBR Green I staining (Fig. 2G). Alternatively, the λ DNA treatment was omitted, and nucleosomes were used directly or purified on a sucrose gradient as above (Fig. 2D). Yeast histones H2A, H2A.Z, H2A.Z-K121C, H2B (for H2A.Z-K121C-H2B), and H2B-3xFlag proteins were expressed individually in Escherichia coli, except for coexpression of H2A-K120C-H2B. Cell pastes for H2A/Z and H2B (or H2B-3xFlag) were resuspended and mixed before sonication and histone dimer purification using an established protocol (7). H2A-K120C-H2B and H2A.Z-K121C-H2B were labeled with Cy3 following a protocol modified from Joo and Ha (11).


Acknowledgments: We thank S. Watanabe and C. Peterson for an INO80 sample and our colleagues for helpful comments. This work was supported by the Intramural Research Program of the Center for Cancer Research, National Cancer Institute, and by the Janelia Research Campus, Howard Hughes Medical Institute.
View Abstract

Navigate This Article