Arabidopsis OXIDATIVE STRESS 3 enhances stress tolerance in Schizosaccharomyces pombe by promoting histone subunit replacement that upregulates drug-resistant genes

Abstract

Histone replacement in chromatin-remodeling plays an important role in eukaryotic gene expression. New histone variants replacing their canonical counterparts often lead to a change in transcription, including responses to stresses caused by temperature, drought, salinity, and heavy metals. In this study, we describe a chromatin-remodeling process triggered by eviction of Rad3/Tel1-phosphorylated H2Aα, in which a heterologous plant protein AtOXS3 can subsequently bind fission yeast HA2.Z and Swc2, a component of the SWR1 complex, to facilitate replacement of H2Aα with H2A.Z. The histone replacement increases occupancy of the oxidative stress-responsive transcription factor Pap1 at the promoters of at least three drug-resistant genes, which enhances their transcription and hence primes the cell for higher stress tolerance.

Introduction

Chromatin is comprised of nucleosomes, in which two units of histone H2A, H2B, H3, and H4, are assembled into an octamer with 145–147 base pairs (bp) of DNA wrapped around each nucleosome unit (Luger et al. 1997). It is a dynamic structure that can also control DNA transcription, recombination, repair, and replication (Jarillo and Pineiro 2015; Venkatesh and Workman 2015). Chromatin changes include DNA-methylation, histone modifications, and exchange of histone variants. DNA methylation (5-methylcytosine in various sequence contexts) can prevent gene transcription such as in silencing transposable elements (Asensi-Fabado et al. 2017). Histone modifications include acetylation, phosphorylation, ubiquitination, methylation, and ADP-ribosylation that alter chromatin structure for the recruitment of specific protein complexes (Strahl and Allis 2000; Swygert and Peterson 2014). As for histone exchange, eukaryotes have noncanonical histone variants that are incorporated into chromatin throughout the cell cycle to affect nucleosome stability and histone-protein interactions (Maze et al. 2014; Jarillo and Pineiro 2015). For example, H3.1 is incorporated during DNA replication and is enriched in silenced genomic regions, whereas H3.3 acts as a replacement histone incorporated outside of S-phase in Arabidopsis thaliana and is preferentially enriched in regions of actively transcribed genes (Stroud et al. 2012; Shu et al. 2014).

The H2A family has two universal variants, H2A.X and H2A.Z. H2A.X is characterized by its long carboxyl-terminal tail with a conserved SQE motif, and is a substrate of ATM (Ataxia-telangiectasia-mutated) and ATR (Ataxia telangiectasia and Rad3-related) kinases (Mannironi et al. 1989; Jette and Lees-Miller 2015). H2A.X-Ser139 in mammalian and plant cells (equivalent to Ser129 in budding yeast) can be rapidly phosphorylated in response to DNA double-strand break (DSB), in which the phosphorylated form, γ-H2A.X, helps to recruit DNA repair proteins (Rogakou et al. 1998; Downs et al. 2000). Mammalian H2A.X Tyr142 can also be phosphorylated for maintenance and processing of the DNA damage response (Xiao et al. 2009), while budding yeast H2A.X Ser122 and Ser126 are necessary for response to DNA-damaging agents (Harvey et al. 2005; Moore et al. 2007).

H2A.Z, another histone H2A variant, is the most highly conserved histone variant from yeast to mammals (Jackson et al. 1996) and is essential in Drosophila melanogaster, Tetrahymena thermophila, and Mus musculus (van Daal and Elgin 1992; Liu et al. 1996; Clarkson et al. 1999), but not in budding yeast although its loss leads to slower growth and formamide sensitivity (Jackson and Gorovsky 2000). H2A.Z is involved in multiple biological processes including the regulation of transcription (Santisteban et al. 2000). DNA methylation can also exclude H2A.Z from causing gene silencing, and conversely H2A.Z enrichment can protect genes from DNA methylation (Zilberman et al. 2008; Murphy et al. 2018). Of particular interest is that while H2A.Z at the transcription start site promotes efficient RNA polymerase II recruitment, H2A.Z enrichment across gene bodies correlates with less active transcription (Coleman-Derr and Zilberman 2012; Sura et al. 2017). H2A.Z is required for DSB repair as its acetylation by the histone acetyltransferase KAT5 (Schizosaccharomyces pombe Mst1, human Tip60) (Kim et al. 2009; Xu et al. 2012).

We had previously described Arabidopsis OXIDATIVE STRESS 3 (AtOXS3) as a plant-specific protein that plays a role in stress tolerance (Blanvillain et al. 2009), as loss of AtOXS3 in Arabidopsis led to higher sensitivity to t-BOOH and diamide while expression of AtOXS3 in S. pombe enhanced tolerance to various heavy metals and oxidizing chemicals. AtOXS3 has a putative N-acetyltransferase catalytic domain which was critical for stress tolerance and the protein was co-localized with histone H4 in a speckling subnuclear pattern. This led to speculation that AtOXS3 might be involved in chromatin modification. More recent studies revealed that the rice OXS3 family members OsO3L2 (OXS3-like) and OsO3L3 were co-localized with H2A in the nucleus, and interacted with H2A but not with other histone subunits in a yeast two-hybrid (Y2H) assay (Wang et al. 2016; Xiao et al. 2021). Of particular interest was that overexpression of these rice genes or gene fragments not only enhanced Cd tolerance in rice seedlings, but also lowered Cd accumulation in mature plant shoot, root, and seed, which holds potential for reducing dietary intake of this toxic heavy metal (Wang et al. 2016, 2019).

The fission yeast has long served as a model organism for studying oxidative stress caused by H₂O₂ (Veal et al. 2014; Hoffman et al. 2015) and also by diamide that causes a subcategory of oxidative stress known as disulfide stress (Hochgrafe et al. 2007). Since expression of AtOXS3 in S. pombe was able to enhance tolerance to diamide and Cd, it was tantalizing to use fission yeast to shed light on how AtOXS3 might operate in its native host. Here, we report that in S. pombe producing AtOXS3, the transcription factor Pap1 (pombe AP-1-like) and histone H2A.Z showed enhanced binding to the promoters of three drug resistance genes: obr1⁺, caf5⁺, and SPCC663.08c, but this coincided with decreased occupancy of histone H2Aα and γ-H2Aα (phospho Ser129). The ATM and ATR kinases of fission yeast, Tel1 and Rad3, were also necessary for this AtOXS3-mediated response, as were two conserved H2Aα amino acids, Thr125 and Ser129. Reduced stress tolerance and decreased H2A.Z binding to target promoters were also found in a strain lacking Swr1, a Swi2/Snf2-related ATPase within the SWR1 complex and needed for deposition of H2A.Z. As AtOXS3 could only interact with Swc2, another subunit of SWR1 complex, AtOXS3 might work through Swc2 to affect H2A.Z deposition into chromatin. Thus, even though AtOXS3 is not a S. pombe protein, its ability to cause replacement of H2Aα with H2A.Z suggests a possible histone subunit replacement role in Arabidopsis.

Materials and methods

Yeast strains and plasmids

Schizosaccharomyces pombe strain JS23 (h⁺ ura4.294 leu1.32) is the wild type (WT) progenitor of strains in Supplementary Table S1. Strains created by homologous gene disruption were performed according to the Fission Yeast Handbook (https://www.baumannlab.org/documents/Nurselab_fissionyeasthandbook_000.pdf, last accessed 22 September, 2021) and as shown in Supplementary Figure S1. 5’-UTR and 3’-UTR flanking DNAs were amplified from genomic DNA and linked to selection gene kanMX6, hphMX6 or ura4⁺ by overlapping PCR, generating linear DNA for homologous recombination. Positive colonies were selected on YES (yeast extract with supplements) medium with 75 μg/ml G418 (A1720; Sigma-Aldrich) or 75 μg/ml hygromycin B (V900372; Sigma-Aldrich) or EMM (Edinburgh minimal medium) selective media without Uracil; mutants were confirmed by PCR. Expression constructs were derived from pART1 using its alcohol dehydrogenase (adh1) promoter for expression.

Sensitivity assay

Strains were grown overnight on YES or EMM with essential supplements until OD₆₀₀∼2–3, then diluted to OD₆₀₀∼0.15, and grown to OD₆₀₀∼0.3. Serially diluted 10-fold cultures were spotted onto EMM agar plates containing the indicated concentrations of diamide (D3648; Sigma-Aldrich) or CdCl₂ (202908; Sigma-Aldrich) and photos taken after 3–5 days at 30°.

RNA analysis

Cells grown to OD₆₀₀∼0.3 were treated with 2.0 mM diamide, collected by centrifugation at 4000 × g for 5 min, and washed 3 times with PBS (137 mM NaCl, 2.7 mM KCl, and 11.9 mM phosphate buffer, pH 7.4) prior to total RNA extraction (RNeasy Mini Kit 74104; Qiagen). Reverse transcription used PrimeScript RT Reagent Kit with gDNA Eraser (RR047A; TaKaRa); quantitative RT-PCR (qRT-PCR) used GoTaq qPCR Master Mix (A6002; Promega) on LightCycler480 II (Roche); primers listed in Supplementary Table S2. Relative expression level of each gene was normalized to act1⁺ (SPBC32H8.12c) of JS23 transformed with pART1 empty vector (EV) at 0 h.

Yeast two-hybrid (Y2H) assay

Yeast strain Y2H Gold was from Matchmaker™ Gold Yeast Two-Hybrid System (630489; Clontech). The coding sequence of AtOXS3 was amplified and cloned in-frame with the GAL4 DNA-binding domain of pGBKT7 between NdeI/BamHI restriction sites to generate a GAL4 DNA-binding domain fusion construct. Prey proteins used PCR amplified coding sequences of fission yeast proteins Pap1, H2Aα, H2Aβ, H2A.Z, the 14 subunits of the SWR1 complex, and Arabidopsis proteins AtSwc2, AtSwc4 and AtSwc6 fused to the GAL4 activation domain between NdeI/XmaI restriction sites. Yeast transformation was according to the manufacturer’s instructions and cells were incubated on plates at 30° for 3 days prior to photography.

Co-immunoprecipitation (Co-IP)

A construct expressing AtOXS3-FLAG driven by Super_pro, a strong mannopine synthase derived promoter was transformed into Arabidopsis (Col-0) and 2-week-old plants were used for protein extraction in buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 1% Triton X-100, 10% glycerol, 1 mM PMSF, protease inhibitor cocktail). Extracts were centrifuged at 12,000 × g for 10 min, and supernatant incubated (4 hr, 4°) with anti-FLAG antibody (F7425; Sigma-Aldrich) and 25 ul mixture (1:1) of Dynabeads^® Protein A and G (10002D and 10004D; Life Technologies). After washing the beads with extraction buffer 3 times at 4°, the affinity-bound proteins were eluted by boiling for 5 min in protein loading buffer. Western blotting assay used either anti-FLAG or anti-H2A.X.3 antibody at 1:3000 dilution.

Chromatin immunoprecipitation

Cells (∼1.25 × 10⁸) were subjected to cross-linking by fresh 30% paraformaldehyde to a final concentration of 1% for 20 min at 30° and terminated with 125 mM glycine for 5 min. Cells pelleted and washed 3 times with PBS were frozen in liquid nitrogen and kept at −80°. Pellets were resuspended in 0.5 ml of nuclear lysis buffer [50 mM HEPES-KOH pH 7.5, 500 mM NaCl, 1 mM EDTA (pH 8.0), 1% Triton X-100, 0.1% sodium deoxycholate, 1% SDS, 1 mM PMSF, protein inhibitor cocktail]. Lysis was performed with a bead beaters (FastPrep-24; MP) in 2.0 ml tubes containing 0.6 ml acid-washing glass beads. Cell lysates were collected by puncturing the bottom of the tubes with a heated 25 gauge syringe needle and separated by centrifuging with a 15 ml collection tube. Chromatin was released and sheared with M220 sonicator (Covaris) to yield an average size of ∼500 bp. Lysis buffer was added up to 1.3 ml, and samples were centrifuged at 12,000 × g for 30 min. One-tenth of the soluble chromatin was saved as input, while the rest was immunoprecipitated overnight at 4° with anti-HA (H3663; Sigma-Aldrich), anti-Histone H2A.Z (39640; Active Motif), or anti-Histone H2A (phospho S129) (ab15083; Abcam). A 25 ul mixture (1:1) of Dynabeads^® Protein A (10002D; Life Technologies) and protein G (10004D; Life Technologies) was added to incubate for 6 hr at 4°. Beads were washed twice in low salt buffer (50 mM HEPES-KOH pH 7.5, 140 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS), twice in high salt buffer (50 mM HEPES-KOH pH 7.5, 500 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS), twice in washing buffer (10 mM Tris pH 8.0, 0.25 M LiCl, 0.5% NP-40, 0.5% sodium deoxycholate, 1 mM EDTA) and twice in TE (10 mM Tris pH 8.0, 1 mM EDTA). DNA was eluted by incubating the beads for 20 min at 65° with 0.3 ml elution buffer (50 mM Tris pH 7.5, 10 mM EDTA, pH 8.0, 1% SDS), and cross-linking was reversed by overnight incubation at 65°. Proteins were digested for 2 hr at 55° with 0.3 mg/ml proteinase K (P6556; Sigma-Aldrich), and 1 µg RNase A (A1101A; TaKaRa) was then added to incubate for 1 hr at 37°. Chromatin was purified by phenol/chloroform/isoamylalcohol extraction and precipitated with ethanol/NaOAC overnight at −20° or −80° for 2 hr. DNA was resuspended in 0.2 ml of TE buffer and recovered DNA fragments amplified by qRT-PCR (GoTaq qPCR Master Mix, A6002; Promega) on LightCycler480 II (Roche). Each IP sample was normalized to its respective input sample, as well as the control region of gpd3⁺. Specific primers for amplifying promoter regions listed in Supplementary Table S2.

Results and discussion

AtOXS3 requires Pap1 to activate obr1⁺, caf5⁺, and SPCC663.08c

S. pombe Pap1 has long been known to as a transcriptional regulator of the stress response (Quinn et al. 2002). In response to diamide, Pap1 along with another nucleocytoplasmic protein, OXS1 (Oxidative stress 1), co-regulates the transcription of at least nine diamide-responsive genes (He et al. 2017). Like OXS1, which cannot enhance diamide tolerance in the absence of Pap1, this was also the case with AtOXS3. As shown in Figure 1A, enhanced tolerance from production of AtOXS3 was significantly abolished when tested in the null mutant pap1Δ.

Figure 1.

AtOXS3 enhanced the expression of obr1⁺, caf5⁺, and SPCC663.08 through Pap1 irrespective of oxidative stress. (A) Tolerance assay of serial dilutions of cells on EMM selective media without or with diamide or CdCl₂; pap1Δ or WT with EV (pART1 empty vector) or pART1 expressing AtOXS3. (B–D) qRT-PCR of pap1⁺ and H₂O₂-upregulated genes (B), diamide-upregulated genes (C), and drug resistance genes (D) in pap1Δ and WT with EV or AtOXS3-expressing plasmid treated or not with 2.0 mM diamide for an hour. (E) qRT-PCR of drug resistance genes from cells grown in EMM selective media treated with 2.0 mM diamide for up to 5 hr. Transcript level of each gene normalized to act1⁺ (SPBC32H8.12c) from WT(EV). Error bars show SEM from three independent experiments; significant differences from unpaired Student’s t-test (*P < 0.05, **P < 0.01, and ***P < 0.001).

To test whether AtOXS3 affects the expression of Pap1 upregulated genes, we examined three antioxidant genes (srx1⁺, ctt1⁺, and trr1) (Calvo et al. 2012) and four diamide-induced genes (hsp90⁺, ssa2⁺, SPBC36.02c, and wis2⁺) (He et al. 2017) that required an oxidized Pap1. In addition, we also included three drug resistance genes (obr1⁺, caf5⁺, and SPCC663.08c) that are activated independent of Pap1 oxidation (Calvo et al. 2012; Chen et al. 2018). Of the antioxidant genes, statistically significant differences in the absence of diamide were observed with a 1.7-fold increased expression of srx1⁺ and 0.5-fold decreased expression of ctt1⁺ in WT(AtOXS3) (WT expressing AtOXS3) compared with the control WT(EV) (WT with empty vector) (Figure 1B). When treated with diamide, only srx1⁺ showed less than one fold AtOXS3-induced expression. With respect to the diamide-induced genes (hsp90⁺, ssa2⁺, SPBC36.02c and wis2⁺), AtOXS3 expression in the absence of diamide enhanced expression of SPBC36.02c by 1.2-fold, and that of wis2⁺ by 0.5-fold (Figure 1C). When treated with diamide, AtOXS3-enhanced expression could still be observed with SPBC36.02c and wis2⁺.

In contrast, the expression of the drug resistance genes obr1⁺, caf5⁺ and SPCC663.08c increased 5–17-fold in WT(AtOXS3) compared with WT(EV) (Figure 1D), and this AtOXS3-induced expression could be observed for at least 5 hr (Figure 1E). When treated with diamide, the AtOXS3-induced expression was even more pronounced, from 8- to 30-fold above WT(EV) (Figure 1D). Most noticeable is that the AtOXS3 effect required Pap1. These drug resistance genes encode proteins with various functions: obr1⁺ produces a ubiquitinated histone-like protein in heterochromatin assembly and silencing (Naresh et al. 2003); caf5⁺ encodes a plasma membrane spermine family transmembrane transporter described as an efflux pump for enhanced tolerance to caffeine (Benko et al. 2004; Calvo et al. 2009); and SPCC663.08c produces a short chain dehydrogenase annotated as having oxidoreductase activity (Wang et al. 2017).

To test whether those genes were involved in AtOXS3-enhanced stress tolerance, AtOXS3 was produced in the triple null mutant obr1Δcaf5ΔSPCC663.08cΔ. Whereas expression of AtOXS3 in WT enhanced tolerance to diamide or Cd, this effect was mostly eliminated in the triple mutant (Figure 2). With single mutants, loss of caf5⁺ or SPCC663.08c reduced AtOXS3-mediated diamide tolerance, while loss of obr1⁺ abolished it entirely. With respect to Cd treatment, loss of caf5⁺ reduced AtOXS3-mediated tolerance, while loss of obr1⁺ or SPCC663.08c had less of an effect.

Figure 2.

Plate assay of obr1⁺, caf5⁺, and SPCC663.08 tested for AtOXS3-mediate stress tolerance. 10-fold serially diluted cells containing EV or pART1 expressing AtOXS3 spotted on EMM selective media without or with diamide or CdCl₂.

Since obr1⁺, caf5⁺, and SPCC663.08c require Pap1 for AtOXS3-enhanced expression, it begs the question of how AtOXS3 can exert this transcription activation effect. Unlike the Pap1 activation of others genes, Pap1 does not need to be in an oxidized form to activate these three genes (Calvo et al. 2012), as the mere retention of Pap1 in the nucleus has been known to activate them, such as from the overproduction of an NES (nuclear export sequence) that competes for Pap1 export (Chen et al. 2018). However, AtOXS3 lacks an NES to serve as a competitive substrate to retain Pap1 in the nucleus. AtOXS3 also did not interact with Pap1 in an Y2H assay (Figure 3B) to suggest the possibility of chaperoning Pap1 to the promoters of these drug resistance genes (henceforth abbreviated as DRGs).

Figure 3.

AtOXS3 interactions in Arabidopsis and in S. cerevisiae, and H2A.Z tested for genes upregulated by AtXOS3. (A) Co-immunoprecipitation (Co-IP) of AtOXS3 and Arabidopsis H2A.X.3 from WT (Col-0); AtOXS3: FLAG transgenic plants detected by α-FLAG or α-H2A.X.3 antibodies. (B) Y2H assays for interaction between AtOXS3 and S. pombe H2A.Z. SD-2DO, SD medium lacking Leu, Trp; SD-4DO, SD medium without Leu, Trp, His, Ade; AD, activation domain; BD, DNA binding domain; representative photos from 3 independent experiments. (C) Growth of pht1Δ and WT with EV or AtOXS3-expressing plasmid. (D) Tolerance assay as in Figure 1(A) of pht1Δ vs WT with EV or AtOXS3-expressing plasmid. (E) qRT-PCR of indicated genes on pht1Δ or WT with EV or AtOXS3-expressing plasmid. Transcript level of each gene normalized to act1⁺ from WT(EV). SEM (error bars) from three independent experiments; significant differences from unpaired Student’s t-test (*P < 0.05 and **P < 0.01).

AtOXS3 enhanced tolerance requires H2A.Z

In Arabidopsis, we previously reported that AtOXS3 co-localized with histone H4 in a speckling sub-nuclear pattern (Blanvillain et al. 2009). More recent Y2H assays found interaction between AtOXS3 and some Arabidopsis H2A variants (H2A.X.3, H2A.X.5, H2A.W.7, H2A.W.12, H2A.Z.8, H2A.Z.9, and H2A.Z.11); and a BiFC (Bimolecular Fluorescent Complimentary) experiment showed AtOXS3 interaction with Arabidopsis H2A.X.3 and H2A.X.5 in plant cells (Xiao et al. 2021). We also conducted Co-IP (co-immunoprecipitation) on protein extracts from 2-week-old Arabidopsis plants expressing AtOXS3-FLAG from a mannopine synthase derived promoter (Super_pro). After incubation with anti-FLAG antibody, western blotting by histone H2A.X antibody detected the H2A.X.3 band (Figure 3A). All these interaction data suggest that histone H2A.X might play a role in AtOXS3-mediated stress tolerance.

The S. pombe genome harbors three H2A genes, hta1, hta2, and pht1, encoding histone H2Aα, H2Aβ, and H2A.Z, respectively. Similar to plant and animal H2A.Xs, H2Aα and H2Aβ contain an SQE motif in their C-termini that is a consensus phosphorylation substrate for phosphatidylinositol-3-OH kinase-related kinases (PIKKs), and the two proteins differ only in three amino acids in their C-termini (Supplementary Figure S5). Surprisingly, although AtOXS3 could interact with Arabidopsis H2A.X.3 and H2A.X.5, the Y2H assay showed interaction only with S. pombe H2A.Z (Figure 3B). Since histone variant H2A.Z could be involved in transcription (Zlatanova and Thakar 2008), we tested whether H2A.Z could affect expression of the DRGs. A null mutant of H2A.Z, pht1Δ, generated by homologous recombination, showed slower growth than its WT progenitor irrespective of AtOXS3 overexpression (Figure 3C), and pht1Δ(AtOXS3) did not show enhanced tolerance to diamide or Cd (Figure 3D). Furthermore, the pht1Δ strain showed reduced AtOXS3-induced transcript accumulation of the DRGs, but not reduced expression of ctt1⁺, trr1⁺, or pap1⁺ (Figure 3E). This suggested that H2A.Z plays a positive role in expression of the DRGs.

AtOXS3 enhanced tolerance requires Tel1 and Rad3

According to the DSB repair model for mammalian cells, the H2A.X SQE motif can be directly phosphorylated by ATM (Ataxia-telangiectasia-mutated) and ATR (Ataxia telangiectasia and Rad3-related) kinases (Ciccia and Elledge 2010). We tested whether Tel1 and Rad3, the fission yeast homologs of ATM and ATR, respectively, were needed for AtOXS3-mediated stress tolerance. Single and double mutants of Tel1 and Rad3 (tel1Δ, rad3Δ, and tel1Δrad3Δ) were generated with a deletion of the PIK-related kinase domain responsible for H2A.X phosphorylation (Supplementary Figure S2). As shown in Figure 4A, the presence of AtOXS3 in each genotype provided higher tolerance to Cd or diamide compared to the no-AtOXS3 control, but not comparable to WT(AtOXS3), especially with the double mutant tel1Δrad3Δ. Correspondingly, expression of the DRGs, but not trr1⁺ and ctt1⁺ declined significantly in tel1Δrad3Δ(AtOXS3) (Figure 4B), indicating that in addition to H2A.Z (Figure 3, D and E), Tel1 and Rad3 are also necessary for AtOXS3-mediated DRG expression and enhanced stress tolerance.

Figure 4.

Tel1 and Rad3 tested for AtOXS3-H2A.X mediated stress tolerance. (A) Plate tolerance assay as in Figure 1A;tel1Δ, rad3Δ, and tel1Δrad3Δ mutants containing EV or AtOXS3-expressing plasmid. (B) qRT-PCR of indicated genes in tel1Δrad3Δ with EV or AtOXS3-expressing plasmid. Transcript level of each gene normalized to act1⁺ from WT(EV). Error bars show SEM from three independent experiments; significant differences from unpaired Student’s t-test (*P < 0.05 and **P < 0.01).

Key downstream components of the ATM/ATR-mediated DSB repair pathway were also tested, namely rad32, mrc1, cds1, chk1, and crb2. Rad32 is required for meiotic recombination and repair of radiation-induced DNA damage (Wilson et al. 1999); Mrc1 is required for the sensor kinase Rad3-Rad26 to activate the effector kinase Cds1 (Zhao et al. 2003); Cds1 is a protein kinase phosphorylated and activated by DNA-replication arrest caused by DNA damage (Lindsay et al. 1998); Chk1 is an essential kinase required for the G2/M DNA damage checkpoint (Walworth and Bernards 1996); and Crb2 is a mediator of the DNA damage checkpoint (Nakamura et al. 2004). For each of these genes, we generated and tested the null mutants, but AtOXS3 expression in none of them showed greater sensitivity to diamide or Cd (Supplementary Figure S3). This suggested that the ATM/ATR-mediated DSB repair pathway is not involved in AtOXS3-mediated stress tolerance.

Thr125 and Ser129 alanine substitutions affect AtOXS3-mediated stress tolerance

In mammalian cells, failure to phosphorylate H2A.X-Ser139 causes sensitivity to agents that generate DNA DSB (Bassing et al. 2002; Celeste et al. 2002). We tested whether the corresponding substitution in fission yeast H2Aα-Ser129 (corresponding to Ser139 in mammalian cells) would play a role in AtOXS3-mediated stress tolerance, but the H2AαS129A alanine substitution mutant failed to show an effect (Supplementary Figure S4). In budding yeast, at least two other amino acid residues, Ser122 and Thr126, in the H2A C-terminus have been implicated in stress tolerance (Harvey et al. 2005; Moore et al. 2007). The S. pombe H2Aα C-terminus harbors four potential phosphosites, Thr121, Ser122, Thr125, and Ser129, while H2Aβ has Ser122 and Ser128 (Supplementary Figure S5). We replaced those Ser/Thr residuals with alanine or aspartate to create the phospho-deficient or phospho-mimetic mutants, respectively. The combination mutant producing hta1-T121A, S122A, T125A, S129A, and hta2-S122A, S128A, henceforth abbreviated as hta1^A4/hta2^A2, showed no significant tolerance change to Cd or diamide. However, when AtOXS3 is produced in this hta1^A4/hta2^A2 strain, decreased tolerance to diamide or Cd was observed. The same effect was detected in hta1^A4, but not in hta2^A2 (Figure 5A). This seemed surprising considering that H2Aα and H2Aβ differ only in three amino acids, and with H2Aα containing an additional threonine (Thr125) in its C-terminus (Supplementary Figure S5). Nonetheless, this showed that putative phosphorylation of the C-terminal residues of H2Aα, but not of H2Aβ, affects AtOXS3 function.

Figure 5.

Phosphorylation of S129 and/or T125 tested for AtOXS3 mediated stress tolerance. (A) Plate tolerance assay as in Figure 1A of hta1^A4/hta2^A2, hta1^A4, hta2^A2, and hta1^5,9A containing EV or AtOXS3-expressing plasmid. Panels (i–v) show relevant C-terminus aa sequence, dash in H2Aβ shows absent aa residues, potential phosphosites shown in blue lettering, S/T to A substitution in red lettering; SEQ motif highlighted in orange. (B) qRT-PCR of indicated genes in hta1^5,9A strain with EV or AtOXS3-expressing plasmid. Transcript level of each gene normalized to act1⁺ from WT(EV). SEM (error bars) from three independent experiments; significant differences from unpaired Student’s t-test (*P < 0.05 and **P < 0.01). (C) Alignment of histone H2A.X C-terminal tails from 16 species. SEQ motif highlighted in orange and the potential phosphosites shown in blue lettering.

To test the specific amino acid residues for AtOXS3 function, we generated three triple mutants hta1^1,2,9A(T121A, S122A, and S129A), hta1^1,5,9A(T121A, T125A, and S129A), and hta1^2,5,9A(S122A, T125A, and S129A). Two of them, hta1^1,5,9A and hta1^2,5,9A, showed greater sensitivity to diamide or Cd treatment, similar to that shown by H2Aα^A4 (Supplementary Figure S6A). Since Thr125 and Ser129 are common in hta1^1,5,9A and hta1^2,5,9A, they seemed likely to be involved in AtOXS3 function. To confirm, the double mutant of hta1^5,9A(T125A and S129A) was constructed with alanine substitutions of Thr125 and Ser129. As shown in Figure 5A, hta1^5,9A lost stress tolerance to the same extent as triple mutants hta1^1,5,9A and hta1^2,5,9A, while single alanine substitution mutants in Thr121, Ser122, Thr125, or Ser129 were unaffected (Supplementary Figure S6B). qRT-PCR showed that although AtOXS3 could still enhance DRG expression in strain hta1^5,9A, significantly reduced expression compared to WT(AtOXS3) was found with the DRGs, but not with srx1⁺ or wis2⁺ (Figure 5B). This indicated that the phosphorylated Thr125 and Ser129 were needed for AtOXS3-mediated stress tolerance. Coincidentally, Thr125 in H2Aα is also a highly conserved residue of H2A.X among eukaryotes (Figure 5C).

As for the aspartic acid substitution mutants to mimic phosphorylation, hta1^D4(hta1-T121D, S122D, T125D, and S129D) or hta1^D2(hta2-S122D and S128D) as well the double mutant hta1^D4/hta2^D2 behaved the same as the WT for stress tolerance, regardless of whether AtOXS3 was expressed (Supplementary Figure S6C). This indicated that although phosphorylation of these serine/threonine residues of H2Aα is necessary, it is not sufficient to substitute for AtOXS3-mediated stress tolerance.

AtOXS3 alters the binding of H2A.Z and H2Aα to the DRG promoters

Since Pap1 activation of DRG expression in the absence of oxidative stress depends on AtOXS3, we tested whether Pap1 binds the promoters and whether the binding depends on AtOXS3. A chromatin immunoprecipitation (ChIP) assay was conducted on a HA-pap1⁺ strain using anti-HA antibody followed by qPCR. Figure 6A shows that the control gene gpd3⁺ did not show a change in Pap1 binding under all conditions, but for the DRGs, Pap1 was recruited to their promoters at a 40% to 90% higher rate when AtOXS3 was produced. Moreover, the combination of AtOXS3 and diamide led to the highest Pap1 occupancy for obr1⁺ and caf5⁺. This indicates that while AtOXS3 caused the recruitment of Pap1 to the DRG promoters, this recruitment could be further enhanced by diamide treatment, at least for two of the three promoters.

Figure 6.

AtOXS3 effect on occupancy of histones H2A.Z and H2Aα on the DRG promoters. ChIP-qPCR of (A) Pap1 by anti-HA antibody in HA-pap1⁺ strain; (B) H2Aα by anti-HA antibody in HA-H2Aα strain; (C) H2AS129p by anti-H2A (phospho Ser129) in WT strain; (D) H2A.Z by anti-H2A.Z antibody in (D) WT strain and (E) hta1^5,9A strain. Each strain contained EV or AtOXS3-expressing plasmid treated with or without 2.0 mM diamide. Fold enrichment normalized against gpd3⁺ control region in strains harboring EV without diamide. SEM from three independent experiments; significant differences from unpaired Student’s t-test (*P < 0.05 and **P < 0.01).

Since the Y2H assay did not detect interaction between Pap1 and AtOXS3 (Figure 3B), this raised the question on how AtOXS3 was able to recruit Pap1 to the DRG promoters. Considering that both histone H2Aα and H2A.Z are required for AtOXS3-enhanced tolerance, ChIP assays on H2Aα and H2A.Z were performed to test the possibility of histone replacement. For H2Aα, the ChIP assay was conducted with anti-HA antibody on a HA-H2Aα strain. As shown in Figure 6B, diamide treatment alone caused 2–3-fold decrease of H2Aα binding to the DRG promoters, and AtOXS3 also caused 2–4-fold decrease of H2Aα occupancy. This suggested that H2Aα was evicted from these promoters upon diamide treatment or by AtOXS3, but there was not an additive effect from the combination of diamide plus AtOXS3. Strikingly, AtOXS3 or diamide led to almost the same extent of occupancy decrease, showing that AtOXS3 could trigger the eviction of H2Aα that mimics the effect of diamide exposure.

Since H2Aα Thr125 and Ser129, as well as Rad3 and Tel1 are all necessary for AtOXS3-mediated stress tolerance (Figures 4A and 5A), we tested whether H2Aα phosphorylation is needed for DRG promoter occupancy in a ChIP-qPCR assay using anti-Histone H2A-S129p (phosphoSer129) antibody. As shown in Figure 6C, phospho-H2Aα binding to the DRG promoters decreased from the inclusion of AtOXS3, but not from diamide treatment, consistent with AtOXS3 promoting the eviction of phosphorylated H2Aα. Since the data in Figure 6, B and C used the same control gene, gpd3⁺, a rough comparison between the two figures would suggest the following: (i) in the absence of stress (EV), H2Aα occupancy was higher than phospho-H2Aα occupancy, which meant that H2Aα was not phosphorylated in the absence of stress; (ii) under diamide stress, phospho-H2Aα occupancy was about the same as that for H2Aα occupancy, consistent with stress-induced phosphorylation of H2Aα; (iii) regardless of whether H2Aα or phospho-H2Aα was measured, AtOXS3 promoted low occupancy of this histone subunit. Taken together, the correlation of AtOXS3 or diamide to low phospho-H2Aα occupancy would suggest the phosphorylation of H2Aα might be the cause for eviction from these promoters.

Eviction of H2Aα could suggest a possible replacement by another histone H2A variant. Since AtOXS3 interacted with H2A.Z in the Y2H assay (Figure 3B) and that H2A.Z deposition near the 5’ ends of genes is associated with transcriptional activation (Probst et al. 2020), we tested the binding of H2A.Z to the DRG promoters. Using anti-H2A.Z antibody, the ChIP-qPCR data showed that production of AtOXS3 enhanced H2A.Z binding to the DRG promoters by 3–14-fold (Figure 6D). An additive effect by diamide to AtOXS3 was also observed in the obr1⁺ promoter. Surprisingly, H2A.Z occupancy did not increase in response to diamide without AtOXS3 (EV vs EV+diamide). This would suggest that H2A.Z occupancy does not naturally occur under diamide stress, but was unnaturally induced by the heterologous AtOXS3. This would also imply that diamide stress normally replaces H2Aα with a histone variant other than H2A.Z.

We also tested H2A.Z binding to the DRG promoters in H2Aα (hta1^5,9A) cells with the T125A and S129A substitutions. Inclusion of AtOXS3 with or without diamide failed to cause an increase in H2A.Z binding as in Figure 6D, but instead caused a 30%–70% decrease of H2A.Z occupancy at the caf5⁺ and SPBC663.08c promoters (Figure 6E). Therefore, this suggested the possibility that H2Aα phosphorylation at Thr125 and Ser129 was required for H2Aα eviction before replacement by H2A.Z.

The deposition of H2A.Z is mediated by chromatin remodeling enzyme SWR1, a Swi2/Snf2 ATPase family complex, which in budding yeast contains up to 14 proteins including the catalytic subunit Swr1, an ATPase that works as a scaffold and contributes to the enzymatic activity of the SWR1 complex (Krogan et al. 2003; Mizuguchi et al. 2004). To explore a possible involvement of Swr1, a null mutant in S. pombe swr1 was generated by gene disruption. Contrary to WT control, enhanced tolerance to Cd or diamide was largely abolished when overproducing AtOXS3 in the swr1Δ strain (Figure 7A). Y2H assays were also used to determine the interaction between AtOXS3 and the 14 subunits of SWR1 complex, respectively. As shown in Figure 7B, the interaction between AtOXS3 and Swc2 was detected. The N-terminal 29 aa of Swc2 was not needed, but removal of its C-terminal 37 aa abolished AtOXS3 interaction (Figure 7B). This interaction with Swc2 suggested that AtOXS3 could be promoting the deposition of H2A.Z by binding to the SWR1 complex through Swc2, and not just by binding H2A.Z.

Figure 7.

SWR1 dependence on AtOXS3-mediated stress tolerance. (A) Tolerance assay as in Figure 1A of swr1Δ strain with EV or AtOXS3-expressing plasmid. (B) Y2H assays as in Figure 3B on interaction between AtOXS3 and the 14 subunits of the SWR1 complex, N-terminal and C-terminal deletions of Swc2, AtSwc2, AtSwc4 and AtSwc6; H2A.Z used as positive control. (C) ChIP-qPCR on binding of H2A.Z to the DRG promoters in swr1Δ strain with EV or AtOXS3-expressing plasmid treated with or without 2.0 mM diamide; anti-H2A.Z antibody used. Fold enrichment normalized against control region gpd3⁺ of strains harboring EV without diamide. SEM from three independent experiments; significant differences from unpaired Student’s t-test (*P < 0.05 and **P < 0.01).

Swc2 is a widely conserved H2A.Z-binding module essential for ATP-dependent histone exchange (Wu et al. 2005). Phylogenetic analysis of Swc2s from S. pombe, Saccharomyces cerevisiae, Homo sapiens, A. thaliana, and Zea mays shows two main clades between the two yeasts from human and higher plants (Supplementary Figure S7). Y2H interaction was observed between AtOXS3 and AtSwc2, but not with AtSwc4 or AtSwc6 (Figure 7B). Hence, this conserved interaction adds strength to the hypothesis that in Arabidopsis, AtOXS3 might also promote the deposition of H2A.Z through AtSwc2.

The ChIP assay with anti-H2A.Z antibody showed that inclusion of AtOXS3 failed to enhance H2A.Z occupancy at the DRG promoters in a swr1Δ strain, but instead decreased promoter occupancy of caf5⁺ and SPBC663.08c (Figure 7C); hence H2A.Z binding to DRG promoters required Swr1. Similar to the dependence on H2A.X phosphorylation for recruitment of chromatin remodeling enzymes to DSBs (Lee et al. 2010; van Attikum and Gasser 2009), H2Aα phosphorylation appears to be required for the AtOXS3-induced H2A.Z deposition at the DRG promoters by the SWR1 complex, and that AtOXS3 acts through the SWR1 complex subunit Swc2. Although AtOXS3 is a plant-specific protein, its ability to enhance stress tolerance in S. pombe had enabled us to carry out this study using this more advanced genetic model system. Based on this study, we propose the following working model (Figure 8). When triggered by abiotic stress (diamide in this study), H2Aα becomes phosphorylated at its C-terminal Thr125 and/or Ser129 by Tel1/Rad3 kinases in stress-associated gene promoters (Figures 4A and 5A), which initiates the eviction of H2Aα (Figure 6, B and C). While we do not know which histone variant replaces H2Aα during diamide stress, AtOXS3 causes the replacement of H2Aα by H2A.Z, which also depends on deposition by the SWR1 complex (Lee et al. 2010). Since AtOXS3 can bind H2A.Z and Swc2, it is possible that AtOXS3 behaves like a H2A.Z chaperon through binding to Swc2 to configure the SWR1 complex for H2A.Z deposition. The incorporation of H2A.Z into specific promoter-binding nucleosomes has been known to prime genes for expression (Guillemette et al. 2005), and this may be the mechanism for recruiting Pap1 to the DRG promoters (Figure 1, B and C) that leads to enhanced diamide and Cd tolerance (Figure 1, A and E). However, the role of AtOXS3 has to include the eviction of H2Aα, since this occurs even in the absence of diamide stress, therefore, it is possible AtOXS3 mimics a stress response signal, like diamide, to initiate the phosphorylation of H2Aα.

Figure 8.

Model of AtOXS3-mediated stress tolerance. AtOXS3 enhances the replacement of H2Aα with H2A.Z, by interacting with the Swc2 subunit of SWR1 complex, which subsequently promotes the recruitment of Pap1 to promoters of three drug-resistant genes to activate their expression and primes the cells for stress tolerance. P: phosphorylation. Insert shows interaction among AtOXS3, Swc2, and H2A.Z.

Aside from oxidative stress and Cd tolerance, AtOXS3 has also been reported to positively regulate freezing tolerance and cold acclimation (Catala et al. 2014). Another Arabidopsis family member, AtO3L3 (At3g03170), appears to be a transcription factor in early abiotic stress-response especially to oxidative stress in mitochondria (Ben Daniel et al. 2016); and overproduction of a third Arabidopsis member, AtO3L1 (SIP1, SOS2 interaction protein 1), was shown to enhance salt tolerance (Wang et al. 2018). More recently, by generating quadruple null mutant of the gene family in Arabidopsis (in AtOXS3, AtOXS3b, AtO3L4, and AtO3L6), we found that phosphorylation of these gene products play a role in the activation of abscisic acid responsive genes (Xiao et al., 2021). Outside of Arabidopsis, Brassica napus BnKCP1 (BnO3L1) acts as a cold-induced transcription factor that regulates gene expression by interacting with histone deacetylase HDAC19 (Gao et al. 2003). Hence this family of proteins could be involved in many stress response pathways. Whether this is how AtOXS3 and the other family members operate in their native hosts has yet to be investigated, but this study points to a probable mechanism on how a heterologous protein reorganizes the chromatin landscape to define specific genes for transcription activation.

Data availability

Data supporting conclusions of this article are in this published article and its supplemental files, and materials reported in this study are freely available upon request. Strains and primers used in this study are listed in Supplementary Tables S1 and S2, respectively. Supplementary material is available at figshare: https://doi.org/10.25386/genetics.15176115.

Author contributions

D.L., C.W., and D.W.O. designed the project. D.L. and C.W. performed the experiments. D.L., D.W.O., and C. W. wrote the manuscript.

Funding

This work was supported by the National Key Research & Development Project (Grant No. 2016YFD0101904) from the Ministry of Science and Technology of China, and by the Key Research Program of Frontier Sciences, Chinese Academy of Sciences (Grant No. QYZDY-SSW-SMC010).

Conflicts of interest

The authors declare that there is no conflict of interest.