Histone Deacetylase HDA9 and WRKY53 Transcription Factor Are Mutual Antagonists in Regulation of Plant Stress Response

Journal Pre-proof Histone deacetylase HDA9 and transcription factor WRKY53 are mutual antagonists in regulation of plant stress response Yu Zheng, Jingyu Ge, Chun Bao, Wenwen Chang, Jingjing Liu, Jingjie Shao, Xiaoyun Liu, Lufang Su, Lei Pan, Dao-Xiu Zhou PII: DOI: Reference:

S1674-2052(19)30408-3 https://doi.org/10.1016/j.molp.2019.12.011 MOLP 871

To appear in: MOLECULAR PLANT Accepted Date: 23 December 2019

Please cite this article as: Zheng Y., Ge J., Bao C., Chang W., Liu J., Shao J., Liu X., Su L., Pan L., and Zhou D.-X. (2020). Histone deacetylase HDA9 and transcription factor WRKY53 are mutual antagonists in regulation of plant stress response. Mol. Plant. doi: https://doi.org/10.1016/ j.molp.2019.12.011. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. All studies published in MOLECULAR PLANT are embargoed until 3PM ET of the day they are published as corrected proofs on-line. Studies cannot be publicized as accepted manuscripts or uncorrected proofs. © 2019 The Author

1

Histone deacetylase HDA9 and transcription factor WRKY53 are mutual

2

antagonists in regulation of plant stress response

3 4

Yu Zheng1*#, Jingyu Ge1#, Chun Bao1, Wenwen Chang1, Jingjing Liu1, Jingjie Shao1,

5

Xiaoyun Liu1, Lufang Su1, Lei Pan1 and Dao-Xiu Zhou2*

6

1

7

Hubei Province Engineering Research Center of Legume Plants, Jianghan University,

8

Wuhan 430056, China

9

2

10

Institute

for

Interdisciplinary

Research

and

Institute of Plant Sciences Paris-Saclay, CNRS, INRAE, Université Paris-Saclay,

91405 Orsay, France

11 12

# Co-first authors

13

* Correspondence author:

14

Yu Zheng ([email protected])

15

Dao-Xiu Zhou ([email protected])

16 17

Running title: Opposite roles of HDA9 and WRKY53 in stress response

18 19

Short Summary

20

HDA9 represses stress-tolerance by regulating DNA-binding, lysine acetylation and

21

activity of WRKY53 that functions as a higher hierarchy transcription activator of

22

stress-responsive genes and as a repressor of HDA9 histone deacetylase activity.

23 24 25 26 27 28 29 30 1

31

ABSTRACT

32

Epigenetic regulation of gene expression is important for plant adaptation to

33

environmental changes. Previous results showed that Arabidopsis RPD3-like histone

34

deacetylase HDA9 has a function to repress plant response to stress. The underling

35

mechanism by which this epigenetic regulator targets to specific chromatin loci to

36

control gene expression networks involved in plant response to stress remains

37

unknown. Here, we show that HDA9 represses stress-tolerance by interacting with

38

and regulating DNA-binding and transcriptional activity of WRKY53 that functions

39

as a higher hierarchy positive regulator of stress resistance. We found that WRKY53

40

is post-translationally modified by lysine acetylation at multiple sites, some of which

41

are removed by HDA9 resulting in inhibition of WRKY53 transcription activity.

42

Conversely, WRKY53 displays a negative function on the HDA9 histone deacetylase

43

activity. The results indicate that HDA9 and WRK53 are reciprocal negative

44

regulators of their activities and reveal functional interplay between a chromatin

45

regulator and a transcription factor to regulate stress tolerance in plants.

46 47

2

48

INTRODUCTION

49

Salt-triggered abiotic stress is one of the most common environmental stresses

50

endured by plants, which exerts evolutionary pressure on plants that, in turn, have

51

developed sophisticated responses to cope and to survive (Zhu, 2002). Thus, the

52

knowledge of mechanisms underlying response to salt stress has profound

53

implications in many biotechnological applications, including increasing plant

54

productivity and environment protection. Regulation of response to salt stress is a

55

complex process controlled by developmental and environmental signaling pathways

56

(Bohnert et al., 2006; Ingram and Bartels, 1996; Jeandroz and Lamotte, 2017;

57

Shinozaki et al., 2003; Xiong et al., 2002). Recent results have shown that histone

58

acetylation/deacetylation is an essential process in the epigenetic regulation of gene

59

expression involved in plant response to environmental changes and stress

60

(Kawashima and Berger, 2014; Liu et al., 2015; Shen et al., 2015). The dynamic

61

equilibrium between histone acetyltransferases (HATs) and histone deacetylases

62

(HDACs) controls histone acetylation levels of nucleosomes affecting chromatin

63

structure and gene activity (Mikkelsen et al., 2007). HATs promote gene

64

transcription , while HDACs are generally associated with transcriptional repression

65

and gene silencing (Verdin and Ott, 2015).

66

HDACs are highly conserved enzymes in eukaryotes. In Arabidopsis, there are

67

18 HDACs identified, of which 12 belong to the reduced potassium dependency3

68

(RPD3/HDA1) subfamily, 4 to the histone deacetylase2 (HD2) subfamily, and 2 to the

69

silent information regulator2 (SIR2) or sirtuin subfamily (Alinsug et al., 2009; Pandey

70

et al., 2002). Many Arabidopsis HDACs have been shown to be involved in plant

71

response to stress. For instance, several members of the RPD3/HDA1 subfamily

72

including HDA6, HDA9, and HDA19 are involved in plant stress response by

73

regulating gene expression through histone deacetylation (Chen et al., 2010; Chen and

74

Wu, 2010; Hu et al., 2019; Kim et al., 2017; Luo et al., 2017; Shen et al.; Zheng et al.,

75

2016). Although the mechanism by which HDAC target to specific loci is generally

76

unclear, HDA19 was shown to form transcriptional repressor complexes with other

77

proteins to regulate plant response to abiotic stress through histone deacetylation at 3

78

specific target genes (Song et al., 2005; Wang et al., 2013).

79

HDA9 has been shown to regulate histone acetylation and expression of genes

80

involved in different biological processes including flowering time, seed germination,

81

stress resistance, and senescence. For example, HDA9 deacetylates H3K9 and H3K27

82

in the promoter of AGL19, a flowering-promoting gene, and represses its expression

83

and flowering time (Kang et al., 2015; Kim et al., 2013; Kim et al., 2016). In addition,

84

HDA9-mediated histone deacetylation suppresses the expression of key negative

85

regulators of senescence genes, which in turn induces the de-repression of their

86

downstream genes to promote leaf senescence (Chen et al., 2016; Kim et al., 2016).

87

Furthermore, HDA9-mediated histone deacetylation represses genes involved in

88

photo-autotrophy and photosynthesis and thus inhibits seed germination (Zanten et al.,

89

2014). It was shown that HDA9 loss-of-function resulted in a more resistant

90

phenotype to stress and HDA9 regulates histone acetylation levels of a large number

91

of stress-responsive genes to negatively regulate plant sensitivity to salt and drought

92

stresses (Zheng et al., 2016). These results indicate that HDA9 is a potent epigenetic

93

regulator of histone acetylation and genes expression involved in both developmental

94

and stress-responsive pathways. Unlike other HDACs that play positive roles in plant

95

response to stress, HDA9 appears as a negative regulator of plant tolerance to stress.

96

However, the underling mechanism by which HDA9 represses plant stress tolerance is

97

unclear. Recent results of proteomic screening revealed that HDA9 interacts with

98

transcription factors such as the MYB protein POWERDRESS and WRKY53 (Chen

99

et al., 2016). Interaction between HDA9 and POWERDRESS is shown to promote

100

senescence in Arabidopsis (Chen et al., 2016). WRKY53 was previously reported to

101

play a central role in stress response and during early events of leaf senescence (Chen

102

et al., 2019; Miao et al., 2007; Miao and Zentgraf, 2007; Murray et al., 2007; Sarwat,

103

2017; Zentgraf et al., 2010). The functional significance of interaction between HDA9

104

and WRKY53 is unclear.

105

In this work, we identified domains within the WRKY53 and HDA9 proteins

106

involved in their interaction and studied their functional relationship and their

107

reciprocal regulation in controlling gene expression and plant tolerance to salt stress. 4

108

We show that wrky53 loss-of-function mutants and HDA9 over-expression (OE)

109

plants displayed higher sensitivity to salt stress. Conversely, hda9 loss-of-function

110

mutants and WRKY53-OE plants were more resistant to salt stress. Importantly, we

111

show that WRKY53 was posttranslational modified by lysine acetylation and that

112

HDA9 repressed the transcription activity of WRKY53 by mediating its lysine

113

deacetylation. Conversely, our data suggest that the deacetylase activity of HDA9

114

could be inhibited by WRKY53 interaction. The results uncover a novel mechanism

115

involved in plant response to stress and provide mechanistic insight into how

116

interaction between a histone deacetylase and a transcription factor contributes to a

117

regulatory network of plant tolerance to stress.

118 119

RESULTS

120 121

HDA9 deacetylase domain interacts with WRKY53 DNA binding domain

122

Previous results showed interaction between HDA9 and WRKY53 (AT4G23810)

123

in Y2H (Chen et al., 2016). We confirmed the interaction in Y2H and by BiFC in

124

Arabidopsis leaf protoplasts (Figure 1A, B). To further study the in vivo interaction

125

between HDA9 and WRKY53, we performed co-immunoprecipitation (Co-IP) assays

126

with transgenic Arabidopsis plants expressing GFP-tagged WRKY53 (WRKY53-OE,

127

Supplemental Figure 1) using anti-GFP and anti-HDA9 antibodies (Supplemental

128

Figure 2). The HDA9 protein could be precipitated by anti-GFP, and the

129

WRKY53-GFP fusion could be precipitated by anti-HDA9 (Figure 1C). To identify

130

interacting domains, we truncated both HDA9 and WRKY53 into three segments and

131

tested the pairwise combinatorial interaction between the segments in Y2H assays.

132

The analysis revealed that the central deacetylase domain of HDA9 interacted with

133

the DNA-binding domain of WRKY53 (Figure 1D). The HDA9 and WRKY53

134

interaction appeared to be specific, as no interaction was detected between WRKY53

135

and other HDACs (i.e. HDA6, HDA9, HDA19 and SRT1) or between HDA9 and

136

other WRKY proteins (i.e. WRKY22, WRKY38, WRKY53 and WRKY62) in Y2H

137

(Supplemental Figure 3). 5

138 139

WRKY53 and HDA9 have opposite functions in regulating plant response to salt

140

stress

141

We have shown that HDA9 plays a negative role in regulating plant sensitivity to

142

salt and drought stresses (Zheng et al., 2016). To study functional relationship

143

between HDA9 and WRKY53, we produced HDA9-overexpression (HDA9-OE)

144

(oe9-2 and oe9-3) and WRKY53-OE (oe53-wt-1 and oe53-wt-2) lines and

145

characterized two T-DNA insertion lines of WRKY53 (wrky53-1 and wrky53-2)

146

(Supplemental Figure 1). These lines together with hda9-1 and hda9-2 mutants (Kim

147

et al., 2013) and the wild type (col-0) were tested for germination rates on normal MS

148

media and MS supplemented with NaCl (100 mM), PEG (-0.75 Mpa), or mannitol

149

(150 mM). The germination rates were scored 78 h later. Seeds from three different

150

harvests were tested. Under normal conditions, the germination rates of wild type,

151

mutants, and the transgenic lines were more than 90% (Figure 2A). When treated with

152

NaCl, PEG, or mannitol, the germination rates of wrky53 and HDA9-OE plants were

153

lower (7.36%-12.27%) than wild type (14.11%-18.89%, p<0.05). Conversely, the

154

germination rates of hda9-1 and WRKY53-OE plants were much higher

155

(35.30%-47.48%) than wild type (p<0.05, Figure 2A). The data indicated that

156

mutation or overexpression of WRKY53 and HDA9 had opposite effects on seed

157

germination sensitivity to the tested stresses.

158

To study effects of over-expression and mutation of HDA9 and WRKY53 on

159

stress resistance, we tested seedling growth of the wild type, mutants, and

160

overexpression plants in soil in the presence of 100 mM NaCl for 30 days. Compared

161

with wild type, hda9 (hda9-1, hda9-2) and WRKY53-OE (oe53-wt-1, oe53-wt-2)

162

plants were more resistant, while the wrky53 (wrky53-1, wrky53-2) and HDA9-OE

163

(oe9-2, oe9-3) plants were more sensitive to the salt stress (Figure 2B, Supplementary

164

Figure 4A). Under 100 mM NaCl, the ABA contents of wrky53 and HDA9-OE plants

165

were also significant lower (p<0.05), whereas that of hda9 was much higher (p<0.05)

166

than the wild type. WRKY53-OE lines had similar ABA content as wild type (Figure

167

2C). We also tested fresh water loss rates of leaves detached from plants of the 6

168

different genotypes during 6.5 h. At all of the tested time points the fresh water loss

169

rate of wrky53 was higher, whereas that of hda9 and WRKY53-OE plants were lower

170

than the wild type. The fresh water loss rate of HDA9-OE lines was lower than the

171

wild type in the first two hours, but subsequently increased and was higher than wild

172

type (Figure 2D). Thus, the fresh water loss rate order was wrky53 > HDA9-OE >

173

wild type > WRKY53-OE > hda9.

174

To further test WRKY53 function in stress response, we transformed

175

35S:WRKY53-GFP into different genetic backgrounds (Supplemental Figure 1). In

176

wrky53-1 background, 35S:WRKY53-GFP likely complemented the mutation as the

177

plant displayed a similar phenotype as wild type under NaCl stress (Supplemental

178

Figure 4B). In hda9-1 and HDA9-OE (oe9-2) backgrounds, 35S:WRKY53-GFP

179

transgenic lines showed similar phenotypes as hda9-1 or HDA9-OE (oe9-2)

180

respectively under both control and stress conditions (Figure 2B, Supplemental Figure

181

4B). The results indicated that HDA9 and WRKY53 had opposite functions in plant

182

response to abiotic stress, with HDA9 being a negative regulator and WRKY53 a

183

positive regulator.

184 185

HDA9 represses WRKY53 expression and its downstream targets

186

Our data are in line with previous results showing that WRKY53 plays a critical

187

role in regulating response to diverse stresses including drought, ozone, hydrogen

188

peroxide, salicylic acid as well as pathogen infection (Miao et al., 2007; Miao and

189

Zentgraf, 2007; Murray et al., 2007; Sarwat, 2017; Zentgraf et al., 2010). WRKY53

190

expression is also induced by drought stress (Sun and Yu, 2015). The physical

191

association between WRKY53 and HDA9 and their apparently opposite functions in

192

stress response suggested that the two proteins may antagonistically regulate

193

stress-responsive gene expression. Therefore, we tested whether HDA9 regulated

194

WRKY53 expression by analyzing WRKY53 transcript levels in wild type, hda9-1, and

195

HDA9-OE (oe9-2) plants. Under normal conditions, WRKY53 transcript level was

196

significantly higher (p<0.05, >2 fold) in hda9-1 and lower (p<0.05, >2 fold) in

197

HDA9-OE than in wild type. Under salt stress, the WRKY53 expression was not 7

198

clearly changed in wild type plants, but induced in hda9-1 (p<0.05, >2 fold) mutant

199

plants (Figure 3). In HDA9-OE plants, the NaCl treatment only increased the

200

WRKY53 expression to the wild type level. The data suggested that HDA9 repressed

201

WRKY53 expression under normal conditions and inhibited induction of the gene

202

under salt stress. Interestingly, the salt stress further increased WRKY53 expression in

203

the WRKY53-OE plants. A possible explanation would be that WRK53-OE overcame

204

the inhibitory effect of HDA9 on salt-stress induction of the gene (see below).

205

WRKY53 expression levels in the different genotypes correlated with plant resistance

206

phenotypes to 100 mM NaCl (Figure 2), indicating that WRKY53 expression level

207

was an important factor to confer plant resistance to the stress, which was negatively

208

regulated by HDA9.

209

In addition to its auto-activation, WRKY53 was shown to also regulate other

210

WRKY family members (Miao et al., 2004). Under normal condition, WRKY13,

211

WRKY15, WRKY18, WRKY22, WRKY29 and WRKY62 are activated by WRKY53

212

whereas WRKY6 and WRKY42 are negatively regulated by WRKY53 (Miao et al.,

213

2004). Therefore, we examined expression of these WRKY genes in different

214

genotypes of HDA9 and WRKY53 in comparison with wild type. We found that under

215

normal conditions, the expression of WRKY13, WRKY15, WRKY18, WRKY22,

216

WRKY29, and WRKY62 was all higher (p<0.05, >2 fold) in hda9-1 and WRKY53-OE

217

(oe53-wt1) and lower (p<0.05, >2 fold) in wrky53-1 and HDA9-OE (oe9-2) than in

218

wild type, whereas that of WRKY6 and WRKY42 was clearly reduced (p<0.05, >2 fold)

219

in hda9-1 and WRKY53-OE, but increased (p<0.05, >2 fold) in wrky53-1 and

220

HDA9-OE compared with wild type (Figure 3). The data confirmed the positive or

221

negative regulation of the genes by WRKY53 and indicated that HDA9 negatively

222

regulated the activity of WRKY53 to control the downstream target genes under

223

normal conditions. Under salt (100 mM NaCl) stress, WRKY13, WRKY22, WRKY29,

224

and WRKY62 displayed a similar expression profile as at normal conditions in the

225

different genotypes. No further significant induction of the downstream genes by salt

226

stress suggested that their expression controlled by increased WRKY53 activity in

227

hda9 and WRKY53-OE plants might be already saturated. However, the expression of 8

228

WRKY18 was more induced in wrky53 mutation or HDA9-OE than in WRKY53-OE,

229

hda9 mutation or wild type backgrounds, suggesting that the induction of the gene by

230

salt stress was independent of WRKY53 or HDA9. WRKY6 expression was

231

unchanged under stress compared with under normal condition. The expression of

232

WRKY15 and WRKY42 displayed no difference in the different genotypes under stress.

233

Thus, 4 of the 6 downstream genes activated by WRKY53 under normal conditions

234

remained to be controlled by WRKY53 and HDA9 under salt stress.

235

In order to study whether WRKY13, WRKY22, WRKY29, and WRKY62 were

236

direct targets of WRKY53 or HDA9, we performed ChIP analysis of the

237

wrky53-complementation lines grown under normal conditions by using anti-GFP and

238

anti-HDA9 antibodies. The precipitated chromatin fragments were analyzed by qPCR

239

using a primer set corresponding to the W-box-containing region, and another primer

240

set to the coding region of the genes (Supplemental Figure 5). The analysis revealed

241

that WRKY53 bound to the W-box-containing promoters. However, HDA9 was not

242

detected to be associated with the promoters. The results suggested that HDA9 might

243

regulate WRKY53 function off its chromatin targets.

244 245

Repression of WRKY53 expression by HDA9 is independent of histone

246

deacetylation

247

Next, we examined whether HDA9-mediated repression of WRKY53

248

transcription was achieved through histone deacetylation of the locus. By

249

chromatin immunoprecipitation (ChIP) assays, we compared the levels of histone

250

modifications (H3K9ac, H3K27ac, H3K4me2, H3K4me3, H3K27me2, and

251

H3K27me3) in WRKY53 promoter and coding regions in wild type, hda9-1, and

252

HDA9-OE (oe9-2) plants grown under normal or salt-stressed conditions.

253

Surprisingly, the analysis revealed no significant change (p>0.05, <2 fold) of

254

H3K9ac, H3K27ac or H3K27me3 at the locus in the different genotypes under

255

both normal and stressed conditions, but significantly higher H3K4me2/me3

256

(p<0.05, >2 fold) in the promoter region induced by salt stress in hda9-1

257

(Supplemental Figure 6). H3K4me2 and H3K4me3 are two marks associated with 9

258

actively transcribed genes in both plants and animals (Lachner et al., 2004; Li et

259

al., 2008). The increases correlated well with the high induction of WRKY53 by

260

salt stress in hda9 (Figure 3). The data suggested that HDA9-mediated repression

261

of WRKY53 was independent of histone deacetylation of the locus and supported

262

the above observation that HDA9 was not associated with the promoters.

263 264

HDA9 regulates lysine acetylation of the WRKY53 protein

265

Besides histones, many proteins including transcription factors also show lysine

266

acetylation that controls their function (Shen et al., 2015). To study whether HDA9

267

regulated the transcriptional function of WRKY53 through lysine deacetylation, we

268

immunoprecipitated the WRKY53-GFP protein from the transgenic plants in wild

269

type, hda9-1, and HDA9-OE backgrounds using anti-GFP antibody and analyzed

270

WRKY53-GFP lysine acetylation levels by immunoblotting using an anti-acetylated

271

lysine (anti-acLys) antibody. Higher and lower levels of WRKY53-GFP lysine

272

acetylation were observed respectively in hda9-1 and HDA9-OE than in the wild type

273

background (Figure 4A). To confirm the results, we incubated the immunoprecipitated

274

WRKY53-GFP and GFP alone produced in tobacco leaves with E. coil-produced

275

HDA9-GST or GST alone (Supplemental Figure 2), and analyzed lysine acetylation

276

levels by immunoblotting using anti-acLys. Incubation with HDA9-GST but not GST

277

reduced

278

Immunoprecipitated GFP did not show any lysine acetylation signal (Figure 4B). To

279

further study lysine acetylation sites regulated by HDA9, we analyzed the

280

immunoprecipitated WRKY53-GFP proteins incubated with or without HDA9-GST

281

by mass spectrometry (MS). The analysis detected 7 acetylated lysine residues (K12,

282

K26, K27, K58, K169, K175 and K268) in the absence of HDA9-GST and 2 lysine

283

residues (K169 and K175) in the presence of HDA9-GST (Figure 4C, Supplemental

284

Figure 8; Supplemental dataset 2). The results indicated that WRKY53 was post

285

translationally modified by lysine acetylation, most of which could be removed by

286

HDA9. The HDA9-targeted acetylation sites were located outside the DNA-binding

287

domain of WRKY53. Comparison of WRKY protein sequences revealed that most of

the

lysine

acetylation

level

10

of

WRKY53-GFP

(Figure

4B).

288

HDA9-targeted Lys residues were not conserved in other WRKY proteins

289

(Supplemental Figure 7), suggesting HDA9 specifically regulates WRKY53 lysine

290

acetylation.

291 292

HDA9 inhibits WRKY53 DNA-binding activity to its own promoter

293

It was shown that WRKY53 could bind to its own promoter and activate its own

294

expression (Miao et al., 2004). To test whether HDA9 regulated WRKY53

295

DNA-binding activity, we produced WRKY53-GST (Supplementary Figure 2) and

296

used the protein for gel mobility assays with the WRKY53 promoter region

297

containing the W-box (-470 to -380) or a mutant version labeled by biotin as probes. A

298

clear binding of the protein to the wild type but not mutant promoter region was

299

observed (Figure 5A). Incubation with HDA9-GST, but not SRT1-GST (Liu et al.,

300

2017), reduced the binding signal (Figure 5A). To study whether HDA9 affected

301

WRKY53 binding to its promoter, we performed ChIP analysis of 35S:WRKY53-GFP

302

plants in WT, hda9-1, and HDA9-OE backgrounds using anti-GFP. The precipitated

303

chromatin fragments were analyzed by qPCR using a primer set corresponding to the

304

W-box-containing region, and another primer set to a more upstream region (Figure

305

5B). The analysis revealed that WRKY53 bound in vivo to the W-box-containing

306

promoter sequence. The binding was significantly higher in the hda9-1 mutant

307

background confirming that HDA9 inhibited WRKY53 binding to its target promoter.

308

However, HDA9-OE had no clear effect on the binding, suggesting that endogenous

309

HDA9 might be not limiting to regulate WRKY53 binding activity. In addition, there

310

was no clear difference observed in plants grown on 1/2MS supplemented with 100

311

nM Trichostatin A (TSA, an inhibitor of HDAC) or without TSA (Figure 5B),

312

suggesting that the negative effect of HDA9 on WRKY53 DNA-binding activity was

313

independent of lysine deacetylation. This was consistent with the observation that the

314

target lysine deacetylation sites of HDA9 were located outside the DNA-binding

315

domain of WRKY53 (Figure 4C).

316 317

HDA9 inhibits WRKY53 trans-activation function 11

318

To study whether HDA9 regulated WRKY53 transcription activity, we

319

co-transformed leaf protoplasts isolated from wild type, hda9-1, and HDA9-OE

320

(oe9-2) plants with pWRKY53Pro:GUS (reporter) and p35S:WRKY53-GFP (effecter)

321

or p35S:GFP as effecter control in the presence or absence of TSA. GUS activities

322

were determined to analyze the effect of the hda9-1 mutation and HDA9-OE on

323

WRKY53 auto-activation. In wild type protoplasts, we detected a significantly higher

324

GUS activity when transfected with p35S:WRKY53-GFP compared with the

325

p35S:GFP control (Figure 5C). Treatment with TSA further increased the GUS

326

activity. In hda9-1 protoplasts transformed with p35S:WRKY53-GFP, the GUS

327

activity was much higher than that in wild type. However, TSA treatment did not

328

reduce the activity. By contrast, in HDA9-OE protoplasts transformed with

329

p35S:WRKY53-GFP, the GUS activity was at the same level as the control,

330

suggesting that endogenous HDA9 might not be limiting to inhibit WRKY53.

331

Treatment with TSA increased the GUS activity. These results suggested that HDA9

332

represses the transcriptional activity of WRKY53, which is sensitive to TSA.

333

To further study the role of lysine acetylation in WRKY53 transcriptional

334

function, we made the Lys to Arg substitution mutations (to conserve the positive

335

charge) of K169 and K175 and constructed p35S:WRKY53M-GFP effecter vector.

336

The reporter (pWRKY53Pro:GUS) and the wild type (p35S:WRKY53-GFP) or the

337

mutant (p35S:WRKY53M-GFP) effecter were co-transfected into Arabidopsis leaf

338

protoplasts. GUS activities were significantly reduced when transfected with

339

p35S:WRKY53M-GFP compared with p35S:WRKY53-GFP (Figure 5D). Treatment

340

with TSA highly increased the GUS activity with p35S:WRKY53-GFP and slightly

341

with p35S:WRKY53M-GFP. The results indicated that lysine acetylation was

342

important for WRKY53 trans-activation function.

343 344

WRKY53 inhibits HDA9 histone deacetylase activity

345

To study whether the mutation or over-expression of HDA9 affected histone

346

acetylation, we compared histone acetylation levels of HDA9 mutant and

347

over-expression lines with that of wild type plants by immunoblotting using 12

348

anti-H3K9ac, anti-H3K27ac, anti-H3ac, and anti-H3 antibodies. The analysis revealed

349

clear increases of H3K9ac and H3K27ac levels in hda9 and slight decreases of the

350

levels in HDA9-OE compared to wild type plants while no clear change of overall H3

351

acetylation level was observed (Figure 6A). This was consistent with previous results

352

showing that HDA9 is involved in deacetylation of H3K9ac and H3K27ac (Kim et al.,

353

2016). Interestingly, we found some increase of H3K9ac and H3K27ac levels in

354

WRKY53-OE and decreases of the levels in wrky53-1 plants. The observation raised

355

a possibility that WRKY53 might also regulate the activity of HDA9.

356

To test this hypothesis, we immuno-purified HDA9 protein from wild type

357

(Col-0), wrky53 mutant (wrky53-1) and OE-WRKY53 (oe53-wt-1) plants grown

358

under normal conditions using anti-HDA9 antibody and tested HDAC activity using

359

the Epigenase HDAC activity/inhibition direct assay kit. The tests revealed a higher

360

HDAC activity of HDA9 isolated from wrky53 and a low activity of the protein from

361

WRKY53-OE compared with that from wild type plants (Figure 6B). The HDAC

362

activities were largely reduced in the presence of TSA. We also tested the HDAC

363

activity of the HDA9 isolated from wild type plants by adding E. coli-produced

364

WRKY53-GST or GST alone in the assays. The tests indicated that WRKY53-GST,

365

but not GST alone, reduced the in vitro HDAC activity of HDA9. To test whether

366

WRKY53 interaction affected HDA9 HDAC activity, we separated WRKY53 into

367

N-terminal (Nter), DNA binding (DB), and C-terminal (Cter) parts (Figure 1D) and

368

produced WRKY53-Nter-GST, WRKY53-DB-GST and WRKY53-Cter-GST proteins

369

in E.coli (Supplemental Figure 2). The proteins were included in HDA9 HDAC

370

activity assays. The presence of WRKY53-Nter-GST, WRKY53-Cter-GST, or both

371

had no effect on HDAC activity of HDA9 isolated from wild type plants. By contrast,

372

WRKY53-DB-GST significantly reduced HDA9 activity (Figure 6C), suggesting that

373

the interaction between the WRKY53 DB domain and HDA9 inhibited the HDAC

374

activity. Computational analysis of the putative interaction docking model revealed

375

that 4 residues (Gly213, Thr214, His215 and Thr216) located in DNA binding domain

376

of WRKY53 were in contact with 4 residues (Cys32, Met33, Thr34 and His25) in the

377

catalytic domain of HDA9. In addition, Pro277 of WRKY53 was found to be close to 13

378

Val305 and Ala306 of HDA9 (Figure 6D). The contacting residues in WRKY53 and

379

HDA9 are not conserved in other WRKY or HDAC proteins, supporting the above

380

data that the interaction between the two proteins was specific. The interaction might

381

mask HDA9 HDAC activity.

382 383 384

DISCUSSION

385

Antagonistic actions between HDA9 and WRKY53 to regulate plant response to

386

stress

387

Recent results have shown that HDA9 acts in a complex with a SANT

388

domain-containing protein POWERDRESS (PWR) and WRKY53 to promote

389

age-related and dark-induced leaf senescence (Chen et al., 2016; Kim et al., 2016).

390

However, the precise molecular function of WRKY53 in the complex is not yet

391

known. The present work showing mutual inhibition between HDA9 and WRKY53 in

392

regulating stress response suggests that the two proteins may antagonistically regulate

393

gene expression in multiple pathways.

394

We have previously reported that HDA9 has a negative function in salt stress

395

responses. HDA9 represses a large number of stress-responsive genes by regulating

396

histone acetylation in their promoter (Zheng et al., 2016). In this study, we provide

397

evidence that HDA9 interacts with and inhibits DNA binding and transcriptional

398

activity of WRKY53 to repress WRKY53 auto-activation and activation of its

399

downstream WRKY genes (which are also stress regulators) in both normal and

400

stressed conditions, indicating that HDA9 constitutively inhibits the gene regulatory

401

network controlled by WRKY53. The higher stress tolerance of hda9 mutants and

402

impaired stress tolerance of HDA9-OE plants (Figure 2) are at least in part related

403

respectively to de-repressed and inhibited transcription activity of WRKY53. Thus,

404

the negative function of HDA9 in stress tolerance may be achieved by repressing

405

DNA-binding and transcriptional activity of WRKY53 in addition to mediating

406

histone deacetylation of other stress-responsive genes. Similar, previous results

407

showed that HDA19 interacts with WRKY38 and WRKY62 and has antagonistic 14

408

function of the transcription factors in basal defense (Kim et al., 2008).

409 410

HDA9 -mediated lysine deacetylation inhibits WRKY53 transcriptional activity

411

Previous work showed that the WRKY DNA-binding domain of the RRS1

412

protein can be lysine-acetylated by the bacterial pathogen effector PopP2, resulting in

413

inhibition of the DNA-binding activity (Sarris et al., 2015). The present work revealed

414

several lysine acetylation sites in WRKY53 including two sites within the WRKY

415

DNA-binding domain and that HDA9 could remove acetylation from most of the sites

416

except those within the DNA-binding domain (Figure 4) that actually interacted with

417

the catalytic domain of HDA9 (Figure 1). In RRS1 WRKY domain, two lysine

418

residues are acetylated (Sarris et al., 2015). One of the sites was also identified in the

419

WRKY53 DNA-binding domain (Figure 4C). The observation that TSA treatment did

420

not affect HDA9 inhibition of WRKY53 binding to its target in vivo (Figure 5B)

421

suggests that HDA9 inhibition of WRKY53 DNA-binding activity may be

422

independent of lysine deacetylation but may be simply due to the physical interaction.

423

By contrast, HDA9-mediated lysine deacetylation was involved in the inhibition of

424

WRKY53 transcription activity, as the inhibitory effect of HDA9 could be alleviated

425

by addition of TSA in the transient expression assays in HDA9-OE or wild type

426

protoplasts (Figure 5C). The results showing that Lys to Arg substitution mutation of

427

WRKY53 reduced the transactivation function (Figure 5D) confirms that WRKY53

428

transcriptional activity is modulated by lysine acetylation. The effects of HDA9 on

429

WRKY53 lysine acetylation indicate that in addition to histones HDA9 also

430

deacetylates non-histone proteins to regulate gene expression. Regulation of

431

transcription factor activity by HDAC-mediated lysine deacetylation is emerging in

432

plants (Hartl et al., 2017). For instance, HDA6 can interact with and deacetylate BIN2

433

to repress its kinase activity in Arabidopsis (Hao et al., 2016). Arabidopsis sirtuin-like

434

histone deacetylase SRT1 deacetylates and enhances the stability of AtMBP-1, a

435

transcriptional repressor of stress-responsive and glycolytic genes (Liu et al., 2017).

436

Similarly,

437

dehydrogenase (GAPDH) and inhibits its nuclear localization and function as a

the

rice

SRT1

could

deacetylate 15

glyceraldehyde-3-phosphate

438

transcriptional activator of glycolytic genes (Zhang et al., 2017).

439

WRKY53 has been reported to play a central role in several stress pathways

440

including senescence regulation (Hinderhofer and Zentgraf, 2001; Miao and Zentgraf,

441

2007; Zentgraf et al., 2010), drought stress response (Sun and Yu, 2015), and basal

442

resistance against Pseudomonas syringae (Murray et al., 2007). The present data

443

confirmed that WRKY53 functions as a higher hierarchical transcriptional regulator

444

of downstream stress-responsive WRKY genes. In addition, our data suggest that

445

WRKY53 may also enhance stress tolerance by inhibiting HDA9 histone deacetylase

446

activity. Although the precise mechanism of the inhibition remains unclear at this

447

stage, our data suggest that interaction with WRKY53 may mask the HDAC activity

448

of HDA9 or prevent its targeting to genomic loci for histone deacetylation. Thus,

449

HDA9 and WRKY53 form an antagonistic couple reciprocally repress their activities.

450

In conclusion, our results indicate that HDA9 represses stress-responsive genes

451

expression by mediating lysine deacetylation of both histones and WRKY53, which,

452

in acetylated form, functions as a higher hierarchy activator of downstream WRKY

453

genes. Conversely, WRKY53 can inhibit HDA9 histone deacetylase activity (Figure

454

7). The work uncovers functional interplay between a chromatin regulator (i.e. HDA9)

455

and a transcription factor (i.e. WRKY53) to regulate stress resistance/tolerance.

456 457

METHODS AND MATERIALS

458 459

Plant materials and production of transgenic plants

460

The Arabidopsis ecotype Columbia-0 (Col-0) was used as wild type in all

461

experiments. The hda9 T-DNA mutants (hda9-1, SALK, N507123; hda9-2, GABI-kat,

462

305G0320) were previously reported (Kim et al., 2013). The T-DNA insertion mutants

463

wrky53-1 (SALK_034157) and wrky53-2 (SALKSEQ_110288.2) were obtained from

464

the Arabidopsis Biological Resource Center (ABRC). After surface-sterilized in 30%

465

bleach, Arabidopsis seeds were kept at 4 °C for 48 h before sowing. The seeds except

466

those used in germination tests were grown in vitro on half-strength MS (Murashige

467

Skoog) with 0.5% sucrose media (pH 5.7, 1.2% agar) in a growth chamber (20 °C) 16

468

under white light (120 µmol m-2 s-1) in 16 h light/day photoperiods for 10 days, and

469

then transplanted into soil. The temperature, light intensity and photoperiods for

470

seedlings in soil were the same as in chamber.

471

For producing WRKY53 and HDA9 overexpression plants, the full length

472

cDNAs of WRKY53 and HDA9 were amplified and cloned into a modified p1300

473

vector (p35S:GFP ), separately to obtain p35S:HDA9-GFP and p35S:WRKY53-GFP

474

vectors. These vectors were transformed by Agrobacterium-mediated infection into

475

wild type plants to obtain HDA9-OE and WRKY53-OE lines. p35S:WRKY53-GFP

476

vector was also used to transform hda9-1 mutant and wrky53-1 mutant to obtain

477

WRKY53-OE-hda9, and WRKY53-complementation lines. Two lines from each

478

transgene

479

co-overexpressing

480

p35S:WRKY53-GFP vector into HDA9-OE (oe9-2).

were

chosen HDA9

for and

phenotype WRKY53

observation were

and

obtained

analysis. by

Plants

transforming

481 482

Tests of salt stress, germination rate, ABA content and fresh water loss rate of

483

detached leaves

484

Salt stress was imitated by adding PEG (-0.75 Mpa), NaCl (100 mM) and

485

mannitol (150 mM) in MS medium (Verslues et al., 2006). For measuring germination,

486

approximately 100 seeds were sowed. The germination (fully emerged radical) rates

487

were scored after 78 hours. Three seed batches (harvested in December 2016, March

488

2017, and May 2017) were used to avoid harvest and storage effects in germination

489

tests. Two alleles of each mutant and two lines of each transgenic plant were tested

490

with three biological repeats. For wild type samples, six biological repeats were tested.

491

The multiple comparisons using LSD (Fisher's least significant difference) method

492

were conducted to compare significant differences among samples.

493

Seedlings grown in soil were treated with 100 mM NaCl (by pouring 100 mM

494

NaCl into trays containing the seedling pots) for 30 days. Seedlings treated with water

495

for same period were used as control. Photographs of plants were taken from different

496

biological repetitions. ABA contents of the 2 whole seedlings were determined by

497

ELISA (Kmaels Shanghai Biotechnology Co., Ltd) by following the instruction. 17

498

Three biological repetitions were performed. Water loss rates of detached leaves were

499

assayed as described by (Zhang et al., 2004) with minor modifications. The sixth

500

rosette leaves from six plants per genotype were excised at the same developmental

501

stage and exposed to light (110 µmol m−2 s-1) at 22 °C. The leaves for each repeats

502

were weighed at various time points, and the loss of fresh weight (%) was used to

503

represent water loss rate. Two alleles of each mutant and two lines of each transgene

504

were tested with three biological repeats. For wild type samples, six biological repeats

505

were tested. The multiple comparisons using LSD (Fisher's least significant difference)

506

method were conducted to compare significant differences among samples.

507 508

Yeast two-hybrid (Y2H) assays

509

Full length or fragmented cDNA of WRKY53, WRKY22, WRKY38, WRKY62,

510

HDA9, HDA6, HDA19 and SRT1 were amplified by PCR and cloned into the prey

511

vector pGADT7 and bait vector pGBKT7

512

binding motif of WRKY53 (421-663 bp) and the deacetylase motif of HDA9 (64-945

513

bp) were cloned into pGADT7 and pGBKT7, respectively. The N- (1-63 bp of HDA9,

514

1-420bp of WRKY53), and C- (946-1278 bp of HDA9, 664-972 bp of WRKY53)

515

terminal fragments were also ligated into both the prey and bait vectors.

respectively (Clontech, USA). The DNA

516

Y2H assays were performed using MatchmakerTM Gold two-hybrid system

517

following manufacturer’s protocols. In this study, SD/-Leu/-Trp dropout medium

518

(Double dropout medium, DDO) was used to select for the bait and prey plasmids.

519

Cells harboring Matchmaker bait and prey plasmids were able to grow because the

520

plasmids encode tryptophan and leucine biosynthesis genes, respectively, that were

521

otherwise absent from the cell. SD/-Ade/-His/-Leu/-Trp dropout medium (Quadruple

522

dropout medium, QDO) was used to confirm protein interactions to activate HIS3 and

523

ADE2.

524 525

Co-immunoprecipitation analysis

526

The anti-HDA9 antibody was produced using synthesized peptides by (ABclonal

527

Biotech Co., Ltd, Wuhan, China). Co-IP was performed in WRKY53 overexpression 18

528

plants (oe53-wt-1). Total proteins extracted from 1.5 g plant materials were incubated

529

with 50 ml Protein A magnetic beads (Thermo Fisher Scientific 88802) and anti-GFP

530

(Abcam ab13970) for WRKY53-GFP purification or anti-HDA9 for HDA9

531

purification. Total and immunopreciptated proteins were detected by immunoblotting

532

with anti-GFP and anti-HDA9 antibodies using the ECL Super signal system.

533 534

Bimolecular fluorescence complementation experiments (BiFC)

535

For BiFC, HDA9 and WRKY53 were tagged respectively with the N-terminal

536

part (YFPN) or the C-terminal part (YFPC) of YFP, as previously described (Schütze

537

et al., 2009). The two vectors were co-transformed into Arabidopsis

538

protoplasts as previously described (Yoo et al., 2007). DAPI staining was used to

539

visualize the nuclei. The fluorescence was observed with a laser scanning confocal

540

microscope (Leica DM IRE2). The empty vectors pYFPC and pWRKY53-YFPN

541

were co-transformed into Arabidopsis protoplasts as the negative control.

mesophyll

542 543

RT-qPCR analysis of gene expression

544

Total RNA was extracted from seedlings grown in soil with or without NaCl

545

treatment using Trizol isolation reagents, and first-strand cDNA was synthesized from

546

1 µg of total RNA using the First Strand cDNA Synthesis Kit (Transgene Biotech)

547

following the protocol. Three biological replicates for each sample were used for

548

RNA extraction. For quantitative real-time PCR (RT-qPCR) assays, transcript levels

549

were normalized against the average of the reference genes: the tubulin2 and actin2

550

genes (AT5G62690 and AT3G18780). The Ct values and the real-time PCR

551

efficiencies were obtained using LinRegPCR (12.X) (Ruijter et al., 2009), and then

552

the normalized relative quantities and standard errors for each sample were obtained

553

by pBaseplus (Hellemans et al., 2007). The relative fold differences for genes

554

expression in different samples were calculated on the base of the normalized relative

555

quantities obtained above, with the normalized relative quantity of wild type (control)

556

set as 1. Each biological replicate had three technical repetitions, and reaction

557

mixtures without cDNA templates were used as negative controls to evaluate the 19

558

specificity of each real-time PCR. The multiple comparisons using LSD (Fisher's least

559

significant difference) method were conducted to compare significant differences

560

among samples. The primers for all genes and annealing temperatures used in

561

qRT-PCR are given in Supplemental Table S2.

562 563

Chromatin immunoprecipitation (ChIP) assay

564

Chromatin fragments were isolated from seedlings of wild type and mutant

565

or transgenic plants and immunoprecipitated with antibodies of H3K9ac

566

(Millipore, 07-352), H3K27ac (Millipore 05-1334), H3K27me2 (Millipore

567

07-452), H3K27me3 (Millipore 07-449), H3K4me2 (Millipore 07-030) and

568

H3K4me3 (Millipore 07-473). The amounts of inmmunoprecipitated chromatin

569

fragments relative to input chromatin were determined by qPCR analysis using 2

570

primer pairs specific to WRKY53, and calculated according to the 2-

ΔΔCT

method.

571

To analyze WRYK53-GFP binding to the WRKY53 promoter in vivo, we

572

used anti-GFP to immunoprecipitate the chromatin fragments isolated from

573

transgenic plants expressing WRKY53-GFP and analyzed relative enrichment

574

(IP/input) of two promoter regions of WRKY53 by qPCR.

575

To analyze whether WRKY13, WRKY22, WRKY29, and WRKY62 are direct

576

targets of WRKY53 or HDA9, we performed ChIP-PCR analysis of

577

WRKY53-complementation lines grown under normal conditions using anti-GFP

578

and anti-HDA9.

579

Each sample had three biological replicates and the multiple comparisons

580

using Fisher's least significant difference (LSD) method were conducted to

581

compare significant differences among samples. The primer pairs used for

582

ChIP-PCR are listed in Supplemental Table S2.

583 584

Immunoblotting analysis of WRKY53 and histone lysine acetylation

585

To determine lysine acetylation of the WRKY53 protein, 5 week-old seedlings

586

overexpressing WRKY53-GFP in wild type (oe53-wt-1), hda9 mutant (oe53-hda9-1)

587

and OE-HDA9 (oe53-oe9-2) backgrounds were used to isolated total proteins. The 20

588

same plant materials were also used to prepare mesophyll protoplasts for total protein

589

extraction. In addition, p35S:WRKY53-GFP and p35S:GFP were introduced into 5

590

week-old tobacco (Nicotiana benthamiana) leaf cells by injection (5 weeks) and total

591

proteins were extracted. WRKY53-GFP or GFP proteins from the above protein

592

extracts were purified by using anti-GFP agarose after a series of washing steps.

593

Eluted proteins were detected by immunoblotting with anti-GFP (Abcam ab13970)

594

and anti-acetyl lysine (Abcam ab80178) antibodies using the ECL Super signal

595

system.

596

For histone acetylation analysis, histone proteins were extracted from 1.5 g plant

597

tissues of wild type and F2 transgenic lines as previously described (Benhamed et al.,

598

2006). The antibodies of H3K9ac (Millipore, 07–352), H3K27ac (Millipore 05-1334),

599

H3ac (Active Motif, 39139), H3 (Abcam, ab1791) were used. All immunoblots were

600

developed using the ECL Super signal system.

601 602

Transient expression assays

603

For transient expression assays in Arabidopsis protoplasts, the 1 kb promoter of

604

WRKY53 was amplified and cloned upstream of the GUS coding sequence in a

605

modified p1301 binary vector to obtain pWRKY53Pro:GUS reporter vector. The

606

p35S:WRKY53-GFP was used as the effecter vector and p35S:GFP as effecter control

607

vector. Lysine 169 and 175 of WRKY3 cDNA were mutated to Arginine using the

608

One Tube Fast Mutagenesis Kit (Aidlab Biotechnologies Co., Ltd) and the mutated

609

WRKY53 cDNA was cloned into p35S:GFP vector to obtain the effecter vector

610

p35S:WRKY53M-GFP. Seedlings (5 weeks) of wild type (Col-0), hda9 mutant

611

(hda9-1), wrky53 mutant (wrky53-1) and HDA9 overexpression lines (oe9-2) grown

612

on 1/2 MS or 1/2MS supplemented with TSA (100 nM) were used to produce

613

Arabidopsis mesophyll protoplasts. Co-transformation of Arabidopsis mesophyll

614

protoplasts was performed as previously described (Yoo et al., 2007). GUS activity

615

assays were carried out as described (Schledzewski and Mendel (1994).

616 617

Electrophoretic Mobility Shift Assays (EMSA) 21

618

The full-length coding sequence of HDA9 and WRKY53 were cloned into a

619

modified pTOPO-D1 vector and transformed into Escherichia coli strain BL21 (DE3).

620

The recombinant proteins were affinity-purified using the GST label protein

621

purification kit (Beyotime Biotechnology).

622

For EMSA, the WRKY53 promoter region containing the WRKY-box (-470 to

623

-380) were used as probes. The same region without WRKY-box were used as

624

mutated probes by substituting the 5'-TTGACC-3' motifs by 5'-TTGAAA-3'

625

(Supplemental Table S2). All the probes were synthesized and biotin labeled by

626

Beyotime

627

Chemiluminescent EMSA Kit (Thermo Scientific, Waltham, MA, USA) according to

628

the manufacturer’s instructions. The WRKY53-GST, HDA9-GST fusion proteins

629

were mixed with biotin-labelled wild type or mutant probes and incubated at 24°C for

630

30 min. The unlabeled probes were used for competition. The free and bound probes

631

were separated by acrylamide gel electrophoresis.

Biotechnology.

EMSAs

were

performed

using

the

LightShift

632 633

LC-MS/MS and data processing

634

Total proteins were extracted from tobacco leaves that express WRKY53-GFP,

635

and immunoprecipitated with anti-GFP antibody-conjugated magnetic beads. Next,

636

0.5 g of the eluate was incubated with E.coli-produced HDA9-GST or GST protein in

637

HDAC assay buffer (1.4 mM NaH2PO4, 18.6 mM Na2HPO4, 0.25 mM EDTA, 10 mM

638

NaCl, 10% (v/v) glycerol and 10 mM β-mercaptoethanol). The deacetylation assay

639

were performed according to the described as (Zhang et al., 2017; Lu et al., 2018).

640

Then the two mixtures were resolved by SDS/PAGE. The WRKY53-GFP bands were

641

excised from SDS/PAGE gel, fully trypsinized, and analyzed on Thermo Fisher LTQ

642

Obitrap ETD mass spectrometry. Briefly, samples were loaded onto an HPLC

643

chromatography system named Thermo Fisher Easy-nLC 1000 equipped with a C18

644

column (1.8 mm

645

B contained 100% acetonitrile. The elution gradient was from 4% to 18% in 182 min,

646

18% to 90% in 13 min solvent B at a flow rate of 300 nL/min. Mass spectrometry

647

analysis were carried out at the AIMS Scientific Co. Ltd. (Shanghai, China) in the

0.15×1,00 mm). Solvent A contained 0.1% formic acid and solvent

22

648

positive-ion mode with an automated data-dependent MS/MS analysis with full scans

649

(350-1600 m/z) acquired using FTMS at a mass resolution of 30,000 and the ten most

650

intense precursor ions were selected for MS/MS. The MS/MS was acquired using

651

higher-energy collision dissociation at 35% collision energy at a mass resolution of

652

15,000. Raw MS files were analyzed by Proteome Discoverer 1.4 with TAIR data

653

(uniprot-Arabidopsis thaliana.fasta). The related parameters used for raw MS files

654

analyzed were as follow: EnFLAGme Name was Trypsin (full); Max.Missed

655

Cleavage Sites were 2; Static Modification was Carbamidomethyl (C); Dynamic

656

Modification were Oxidation (M) , Deamidated (N,Q) ,Acetyl(K); Precursor Mass

657

Tolerance was 20ppm; Fragment Mass Tolerance was 0.05D; Validation based on

658

q-Value. The MS files are uploaded to the iProX databases under accession

659

“PXD016822”.

660 661

In vitro HDAC Assays of HDA9 protein isolated from plant extracts

662

HDA9 protein was immuno-purified using anti-HDA9 and Protein A bead from

663

total protein extracts of wild type (Col-0), wrky53 mutant (wrky53-1) and WRKY53

664

over-expression (oe53-wt-1) plants grown under normal conditions. After incubation

665

at 4°C overnight, HDA9 was eluted from the Protein –A beads after a series of washes.

666

HDA9 protein (about 0.2 µg) was used to test in vitro HDAC activity. Epigenase

667

HDAC Activity/Inhibition Direct Assay Kit (Epigentek) was used to measure the

668

HDAC activity according to the manufacturer’s instructions. Read the fluorescence on

669

a fluorescence microplate reader within 2 to 10 min at 530EX/590EM. The HDAC

670

inhibitor Trichostatin A (TSA 100 nM) was used. The activity of the HDAC enzyme is

671

proportional to the OD intensity was calculated with formula: HDAC Activity (RFU/min/mg) =

Sample RFU– Blank RFU × 1000 Protein Amount (µg) ∗ x min ∗∗

672 673

Protein structural modelling

674

Homology modeling of three-dimensional structure prediction of WRKY53 and

675

HDA9 based on their sequence similarity to five proteins of known structure 23

676

downloaded from PDB database (http://www.rcsb.org/), Protein structure files

677

1wj2.pdb, 22yd.pdb, 2lex.pdb, 5w3x.pdb and 6ir8.pdb were used as WRKY53's

678

template, and protein structure files 4bkx.pdb, 5icn.pdb, 4a69.pdb, 4lxz.pdb and 4ly1

679

were used as HDA9's template. Homology modeling was performed with

680

MODELLER (https://salilab.org/modeller/). This model with the lowest energy and

681

highest

682

(http://zdock.umassmed.edu/) was used for primary docking and the top1 model

683

produced by ZDOCK was optimized by PyROSETTA (http://www.pyrosetta.org/)

684

according to the software's protein-protein interaction protocol. Structural figures

685

were prepared using PyMOL (https://pymol.org/2/).

score

was

selected

for

further

docking

analysis.

ZDOCK

686 687

FUNDING

688

This study was supported by Discipline Innovation Team Foundation of Jianghan

689

University (03100074), Natural Science Foundation of Hubei Province (Grant

690

No.2016CFB630), National Natural Science Foundation of China (NSFC31701100,

691

NSFC31600981 and NSFC31600801), and French ANR-19-CE12-0027-01.

692 693

REFERENCE

694

Alinsug, M.V., Yu, C.-W., and Wu, K. (2009). Phylogenetic analysis, subcellular

695

localization, and expression patterns of RPD3/HDA1 family histone

696

deacetylases in plants. BMC plant biology 9:37.

697

Benhamed, M., Bertrand, C., Servet, C., and Zhou, D.-X. (2006). Arabidopsis GCN5,

698

HD1, and TAF1/HAF2 interact to regulate histone acetylation required for

699

light-responsive gene expression. The Plant Cell 18:2893-2903.

700

Bohnert, H.J., Gong, Q., Li, P., and Ma, S. (2006). Unraveling abiotic stress tolerance

701

mechanisms

702

9:180-188.

getting genomics going. Current opinion in plant biology

703

Chen, L.-T., Luo, M., Wang, Y.-Y., and Wu, K. (2010). Involvement of Arabidopsi s

704

histone deacetylase HDA6 in ABA and salt stress response. Journal of

705

experimental botany 61:3345-3353. 24

706

Chen, L.-T., and Wu, K. (2010). Role of histone deacetylases HDA6 and HDA19 in

707

ABA and abiotic stress response. Plant signaling & behavior 5:1318-1320.

708

Chen, Q., Xu, X., Xu, D., Zhang, H., Zhang, C., and Li, G. (2019). WRKY18 and

709

WRKY53 coordinate with HISTONE ACETYLTRANSFERASE1 to regulate

710

rapid responses to sugar. Plant Physiology:pp.00511.02019.

711

Chen, X., Lu, L., Mayer, K.S., Scalf, M., Qian, S., Lomax, A., Smith, L.M., and

712

Zhong, X. (2016). POWERDRESS interacts with HISTONE

713

DEACETYLASE 9 to promote aging in Arabidopsis. Elife 5:e17214.

714

Hao, Y., Wang, H., Qiao, S., Leng, L., and Wang, X. (2016). Histone deacetylase

715

HDA6 enhances brassinosteroid signaling by inhibiting the BIN2 kinase.

716

Proceedings of the National Academy of Sciences 113:10418-10423.

717

Hartl, M., Füßl, M., Boersema, P.J., Jost, J.-O., Kramer, K., Bakirbas, A., Sindlinger,

718

J., Plöchinger, M., Leister, D., Uhrig, G., et al. (2017). Lysine acetylome

719

profiling uncovers novel histone deacetylase substrate proteins in Arabidopsis.

720

Molecular Systems Biology 13:949.

721

Hellemans, J., Mortier, G., De Paepe, A., Speleman, F., and Vandesompele, J. (2007).

722

qBase relative quantification framework and software for management and

723

automated analysis of real-time quantitative PCR data. Genome Biol 8:R19.

724 725 726

Hinderhofer, K., and Zentgraf, U. (2001). Identification of a transcription factor specifically expressed at the onset of leaf senescence. Planta 213:469-473. Hu, Y., Lu, Y., Zhao, Y., and Zhou, D.-X. (2019). Histone Acetylation Dynamics

727

Integrates Metabolic Activity to Regulate Plant Response to Stress. Frontiers

728

in Plant Science10:1236. doi: 10.3389/fpls.2019.01236.

729 730 731 732 733

Ingram, J., and Bartels, D. (1996). The molecular basis of dehydration tolerance in plants. Annual review of plant biology 47:377-403. Jeandroz, S., and Lamotte, O. (2017). Plant responses to biotic and abiotic stresses: lessons from cell signaling. Frontiers in plant science 8:1772. Kang, M.J., Jin, H.S., Noh, Y.S., and Noh, B. (2015). Repression of flowering under a

734

noninductive photoperiod by the HDA9

735

New Phytologist 206:281-294. 25

AGL19

FT module in Arabidopsis.

736 737 738

Kawashima, T., and Berger, F. (2014). Epigenetic reprogramming in plant sexual reproduction. Nature Reviews Genetics 15:613-624. Kim, J.-M., To, T.K., Matsui, A., Tanoi, K., Kobayashi, N.I., Matsuda, F., Habu, Y.,

739

Ogawa, D., Sakamoto, T., and Matsunaga, S. (2017). Acetate-mediated novel

740

survival strategy against drought in plants. Nature plants 3:17097.

741

Kim, W., Latrasse, D., Servet, C., and Zhou, D.-X. (2013). Arabidopsis histone

742

deacetylase HDA9 regulates flowering time through repression of AGL19.

743

Biochemical and Biophysical Research Communications 432:394-398.

744

Kim, Y.J., Wang, R., Gao, L., Li, D., Xu, C., Mang, H., Jeon, J., Chen, X., Zhong, X.,

745

and Kwak, J.M. (2016). POWERDRESS and HDA9 interact and promote

746

histone H3 deacetylation at specific genomic sites in Arabidopsis. Proceedings

747

of the National Academy of Sciences 113:14858-14863.

748

Lachner, M., Sengupta, R., Schotta, G., and Jenuwein, T. (2004).

Trilogies of

749

histone lysine methylation as epigenetic landmarks of the eukaryotic genome.

750

In: Cold Spring Harbor symposia on quantitative biology: Cold Spring Harbor

751

Laboratory Press. 209-218.

752

Li, X., Wang, X., He, K., Ma, Y., Su, N., He, H., Stolc, V., Tongprasit, W., Jin, W., and

753

Jiang, J. (2008). High-resolution mapping of epigenetic modifications of the

754

rice genome uncovers interplay between DNA methylation, histone

755

methylation, and gene expression. The Plant Cell 20:259-276.

756

Liu, X., Luo, M., Yang, S., and Wu, K. (2015).

Role of epigenetic modifications in

757

plant responses to environmental stresses. In: Nuclear Functions in Plant

758

Transcription, Signaling and Development: Springer. 81-92.

759

Liu, X., Wei, W., Zhu, W., Su, L., Xiong, Z., Zhou, M., Zheng, Y., and Zhou, D.-X.

760

(2017). Histone Deacetylase AtSRT1 Links Metabolic Flux and Stress

761

Response in Arabidopsis. Molecular plant 10:1510-1522.

762

Lu, Y., Xu, Q., Liu, Y., Yu, Y., Cheng, Z.Y., Zhao, Y., and Zhou D.X. (2018) Dynamics

763

and functional interplay of histone lysine butyrylation, crotonylation, and acetylation in

764

rice under starvation and submergence. Genome Biol. 2018 Sep 25;19(1):144. doi:

765

10.1186/s13059-018-1533-y. 26

766 767 768 769

Luo, M., Cheng, K., Xu, Y., Yang, S., and Wu, K. (2017). Plant Responses to Abiotic Stress Regulated by Histone Deacetylases. Frontiers in plant science 8:2147. Miao, Y., Laun, T., Zimmermann, P., and Zentgraf, U. (2004). Targets of the

770

WRKY53 transcription factor and its role during leaf senescence in

771

Arabidopsis. Plant molecular biology 55:853-867.

772

Miao, Y., Laun, T.M., Smykowski, A., and Zentgraf, U. (2007). Arabidopsis MEKK1

773

can take a short cut: it can directly interact with senescence-related WRKY53

774

transcription factor on the protein level and can bind to its promoter. Plant

775

molecular biology 65:63-76.

776

Miao, Y., and Zentgraf, U. (2007). The antagonist function of Arabidopsis WRKY53

777

and ESR/ESP in leaf senescence is modulated by the jasmonic and salicylic

778

acid equilibrium. The Plant Cell 19:819-830.

779

Mikkelsen, T.S., Ku, M., Jaffe, D.B., Issac, B., Lieberman, E., Giannoukos, G.,

780

Alvarez, P., Brockman, W., Kim, T.-K., and Koche, R.P. (2007). Genome-wide

781

maps of chromatin state in pluripotent and lineage-committed cells. Nature

782

448:553-560.

783

Murray, S.L., Ingle, R.A., Petersen, L.N., and Denby, K.J. (2007). Basal resistance

784

against Pseudomonas syringae in Arabidopsis involves WRKY53 and a

785

protein with homology to a nematode resistance protein. Molecular

786

Plant-Microbe Interactions 20:1431-1438.

787

Pandey, R., MuÈller, A., Napoli, C.A., Selinger, D.A., Pikaard, C.S., Richards, E.J.,

788

Bender, J., Mount, D.W., and Jorgensen, R.A. (2002). Analysis of histone

789

acetyltransferase and histone deacetylase families of Arabidopsis thaliana

790

suggests functional diversification of chromatin modification among

791

multicellular eukaryotes. Nucleic acids research 30:5036-5055.

792

Ruijter, J.M., Ramakers, C., Hoogaars, W.M., Karlen, Y., Bakker, O., van den Hoff,

793

M.J., and Moorman, A.F. (2009). Amplification efficiency: linking baseline

794

and bias in the analysis of quantitative PCR data. Nucleic acids research

795

37:e45. 27

796

Sarris, P.F., Duxbury, Z., Huh, S.U., Ma, Y., Segonzac, C., Sklenar, J., Derbyshire, P.,

797

Cevik, V., Rallapalli, G., and Saucet, S.B. (2015). A plant immune receptor

798

detects pathogen effectors that target WRKY transcription factors. Cell

799

161:1089-1100.

800

Sarwat, M. (2017).

Leaf Senescence in Plants: Nutrient Remobilization and Gene

801

Regulation. In: Stress Signaling in Plants: Genomics and Proteomics

802

Perspective, Volume 2: Springer. 301-316.

803

Schütze, K., Harter, K., and Chaban, C. (2009). Bimolecular fluorescence

804

complementation (BiFC) to study protein-protein interactions in living plant

805

cells. Plant Signal Transduction: Methods and Protocols:189-202.

806

Schledzewski, K., and Mendel, R.R. (1994). Quantitative transient gene expression:

807

comparison of the promoters for maize polyubiquitin1, rice actin1,

808

maize-derivedEmu andCaMV 35S in cells of barley, maize and tobacco.

809

Transgenic Research 3:249-255.

810

Shen, Y., Lei, T., Cui, X., Liu, X., Zhou, S., Zheng, Y., Guérard, F., Issakidis-Bourguet,

811

E., and Zhou, D.-X. Arabidopsis histone deacetylase HDA15 directly represses

812

plant response to elevated ambient temperature. The Plant

813

Journal100(5):991-1006. doi: 10.1111/tpj.14492.

814

Shen, Y., Wei, W., and Zhou, D.-X. (2015). Histone acetylation enzymes coordinate

815

metabolism and gene expression. Trends in plant science 20:614-621.

816

Shinozaki, K., Yamaguchi-Shinozaki, K., and Seki, M. (2003). Regulatory network of

817

gene expression in the drought and cold stress responses. Current opinion in

818

plant biology 6:410-417.

819

Song, C.-P., Agarwal, M., Ohta, M., Guo, Y., Halfter, U., Wang, P., and Zhu, J.-K.

820

(2005). Role of an Arabidopsis AP2/EREBP-type transcriptional repressor in

821

abscisic acid and drought stress responses. The Plant Cell 17:2384-2396.

822

Sun, Y., and Yu, D. (2015). Activated expression of AtWRKY53 negatively regulates

823

drought tolerance by mediating stomatal movement. Plant cell reports

824

34:1295-1306.

825

Verdin, E., and Ott, M. (2015). 50 years of protein acetylation: from gene regulation 28

826

to epigenetics, metabolism and beyond. Nature reviews Molecular cell biology

827

16:258-264.

828

Verslues, P.E., Agarwal, M., Katiyar Agarwal, S., Zhu, J., and Zhu, J.K. (2006).

829

Methods and concepts in quantifying resistance to drought, salt and freezing,

830

abiotic stresses that affect plant water status. The Plant Journal 45:523-539.

831

Wang, Z., Cao, H., Sun, Y., Li, X., Chen, F., Carles, A., Li, Y., Ding, M., Zhang, C.,

832

and Deng, X. (2013). Arabidopsis paired amphipathic helix proteins SNL1 and

833

SNL2 redundantly regulate primary seed dormancy via abscisic acid

834

ethylene antagonism mediated by histone deacetylation. The Plant Cell

835

25:149-166.

836 837 838

Xiong, L., Schumaker, K.S., and Zhu, J.-K. (2002). Cell signaling during cold, drought, and salt stress. The plant cell 14:S165-S183. Yoo, S.-D., Cho, Y.-H., and Sheen, J. (2007). Arabidopsis mesophyll protoplasts: a

839

versatile cell system for transient gene expression analysis. Nature protocols

840

2:1565-1572.

841

Zanten, M., Zöll, C., Wang, Z., Philipp, C., Carles, A., Li, Y., Kornet, N.G., Liu, Y.,

842

and Soppe, W.J. (2014). HISTONE DEACETYLASE 9 represses seedling

843

traits in Arabidopsis thaliana dry seeds. The Plant Journal 80:475-488.

844

Zentgraf, U., Laun, T., and Miao, Y. (2010). The complex regulation of WRKY53

845

during leaf senescence of Arabidopsis thaliana. European journal of cell

846

biology 89:133-137.

847

Zhang, H., Zhao, Y., and Zhou, D.-X. (2017). Rice NAD+-dependent histone

848

deacetylase OsSRT1 represses glycolysis and regulates the moonlighting

849

function of GAPDH as a transcriptional activator of glycolytic genes. Nucleic

850

acids research 45:12241-12255.

851

Zhang, W., Qin, C., Zhao, J., and Wang, X. (2004). Phospholipase Dα1-derived

852

phosphatidic acid interacts with ABI1 phosphatase 2C and regulates abscisic

853

acid signaling. Proceedings of the National Academy of Sciences of the United

854

States of America 101:9508-9513.

855

Zheng, Y., Ding, Y., Sun, X., Xie, S., Wang, D., Liu, X., Su, L., Wei, W., Pan, L., and 29

856

Zhou, D.-X. (2016). Histone deacetylase HDA9 negatively regulates salt and

857

drought stress responsiveness in Arabidopsis. Journal of experimental botany

858

67:1703-1713.

859 860

Zhu, J.-K. (2002). Salt and drought stress signal transduction in plants. Annual review of plant biology 53:247-273.

861 862 863

AUTHOR CONTRIBUTIONS

864

Conceptualization, Y.Z and D.X.Z; Investigation, Y.Z, J.Y.G, C.B, J.J.L, J.J.S, X.Y.L,

865

L.F.S and L.P; Formal Analysis, Y.Z and C.B; Visualization, Y.Z; Writing-Original

866

Draft, Y.Z and D,X.Z; Writing-Review & Editing, Y.Z and D,X.Z; Funding

867

Acquisition, Y.Z, X.Y.L, L,P and D.X.Z

868 869 870 871

FIGURE LEGENDS

872

Figure 1. HDA9 deacetylase domain interacts with the DNA-binding domain of

873

WRKY53

874

A. Yeast two hybrid analysis of interaction between HDA9 and WRKY53. Yeast Gold

875

cells were co-transformed with a combination of the indicated plasmids. Yeast cells

876

were plated on selective media DDO/X/A and QDO/X/A and then incubated for 5

877

days. DDO/X/A means double dropout medium: SD/-Leu/-Trp supplement with

878

X-α-Gal and Aureobasidin A. QDO/X/A means quadruple dropout medium:

879

SD/-Ade/-His/-Leu/-Trp/ supplement with X-α-Gal and Aureobasidin A.

880

B. BiFC analysis of HDA9 and WRKY53 interaction. pHDA9-YFPC and

881

pWRKY53-YFPN were co-transformed into Arabidopsis protoplasts. The empty

882

pYFPC and pWRKY53-YFPN were co-transformed as the negative control. DAPI

883

staining was used to visualize nuclei. Images were obtained using a confocal

884

microscope (Leica DM IRE2).

885

C. Co-IP assays of HDA9 and WRKY53 interaction in vivo. Total proteins extracted 30

886

from WRKY53-GFP transgenic plants were incubated with or without (input) Protein

887

A beads and anti-GFP or anti-HDA9. The total proteins and the elution from the

888

agarose resins were analyzed by immunoblotting with anti-GFP and anti-HDA9

889

antibodies using the ECL Super signal system.

890

D. The coding regions of HDA9 and WRKY53 were fragmented into three parts (top

891

panel), used with GAL4-AD and BD domains and tested in pairwise combination to

892

test interaction in yeast 2-hybrid as in (A).

893 894

Figure 2. WRKY53 and HDA9 have opposite functions to regulate response to

895

salt stress.

896

A. Seed germination rates of the different genotypes under NaCl (100mM), PEG

897

(-0.75Mpa) and mannitol (150mM) treatments. Seeds (100 seeds per genotype per

898

treatment) were grown on stress media. Germination (fully emerged radicle) was

899

scored after 78 h. Seeds grown on 1/2MS were used as control. Two mutant alleles of

900

hda9 (hda9-1 and hda9-2) and wrky53 (wrky53-1 and wrky53-2) and two lines of

901

HDA9-OE (oe9-2 and oe9-3) and WRKY53-OE (oe53-wt1 and oe53-wt2) were tested

902

with 3 biological replicates for each line. Bars represent means +/- standard errors

903

with six (3 x 2 lines) biological replicates and multiple comparisons (Fisher's LSD)

904

were conducted for significant differences among samples.

905

B. Plant phenotypes of the different genotypes under NaCl (100 mM) treatment.

906

Seeds were sowed on 1/2MS medium for 4 days to obtain uniform germination, and

907

then seedlings were transferred to soil for 30 days with and without NaCl treatment

908

before taking photographs. The bar = 0.5 cm.

909

C. ABA contents of the different genotypes under 100 mM NaCl treatment. ABA

910

contents were determined in 2 whole seedlings per genotype by ELISA method. Bars

911

represent means +/- standard errors with six (3 x 2 lines) biological replicates and

912

multiple comparisons (Fisher's LSD) were conducted for significant differences

913

among samples.

914

D. Water loss rates of the different genotypes. The sixth rosette leaves were excised at

915

the same developmental stage (early senescence phase, approximately 25% yellowing) 31

916

and exposed to light (110 µmol m−2 s-1) at 22°C. Six plants per genotypes were used

917

for the analysis. Leaves were weighed at various time points, and the loss of fresh

918

weight (%) was used to represent water loss rates. The same lines as those used in A,

919

and C were tested. Bars represent means +/- standard errors with six (3 x2 lines)

920

biological replicates and multiple comparisons (Fisher's LSD) were conducted for

921

significant differences among samples.

922 923

Figure 3. HDA9 represses WRKY53 expression and its downstream targets

924

RT-qPCR analysis of WRKY53 and its downstream WRKY genes expression in WT,

925

hda9-1, wrky53-1, HDA9-OE (oe9-2), and WRKY53-OE (oe53-wt1) plants under

926

normal (control) and 100 mM NaCl conditions for 30 days. Bars represent means +/-

927

standard errors from three biological replicates and multiple comparisons (Fisher's

928

LSD) were conducted for significant differences among samples. Different letters

929

represent significant differences between samples.

930 931

Figure 4. WRKY53 displays lysine acetylation that is reduced by HDA9

932

A. Lysine acetylation levels of WRKY53-GFP produced in wild type, hda9-1 and

933

HDA9-OE seedlings (upper panel) or leaf protoplasts (lower panel). WRKY53-GFP

934

was

935

immunoblotting with antibody of acetylated lysine (anti-AcLys). Anti-GFP was used

936

as loading controls. The bands were quantified with the Image J software. Bars are

937

means +/- SD from three measures of each of the two repeats.

938

B. HDA9-GST could reduce lysine acetylation level of WRKY53-GFP. Tobacco

939

leaves expressing WRKY53-GFP or GFP alone were immunoprecipitated by

940

anti-GFP and incubated with or without HDA9-GST or GST alone. The incubations

941

were then analyzed by immunoblotting with anti-acLys. Anti-GFP was used as

942

loading control. The bands were quantified with the Image J software. Bars are means

943

+/- SD from three measures of each of the two repeats.

944

C. Analysis of lysine acetylation site of WRKY53 by mass spectrometry.

945

WRKY53-GFP was transiently expressed in tobacco leaves and purified with

immunoprecipitated

from

the

different

32

genotypes

and

analyzed

by

946

anti-GFP

antibody-conjugated

beads

magnetic

beads

and

incubated

with

947

E.coli-produced HDA9-GST or GST alone. Then the mixtures were resolved by

948

SDS/PAGE and the WRKY53-GFP bands were excised for MS analysis. The

949

positions of detected lysine acetylation sites (red lines) are indicated. Sequence

950

alignment shows comparison of the two lysine acetylation sites (in red) in WRKY53

951

and RRS1 WRKY DNA-binding domains. The MS graphs are shown in Figure S8.

952 953

Figure 5.

HDA9 inhibits DNA-binding and transcriptional activity of WRKY53.

954

A. DNA-binding of WRKY53 is impaired by HDA9 in vitro. EMSA of

955

WRKY53-GST binding to its promoter region. Upper panel, blue box represents the

956

probe and the red line indicates the W-box motif (TTGACC) which were substituted

957

by (TTGAAA) in mutated probe. Both wild type and mutant versions of the promoter

958

fragment were biotin-labeled and used for incubation with or without WRKY53-GST

959

(middle). Lower panel: EMSA of WRKY53-GST binding to the wild type promoter

960

region in the presence or absence of HDA9-GST or SRT1-GST.

961

B. ChIP-PCR analysis of WRKY53-GFP association to its promoter regions in

962

wrky53-1, hda9-1 and HDA9-OE plants treated with or without TSA. Two regions

963

(P1 and P2) shown in A of the locus were tested. Bars represent means +/- standard

964

errors from three biological replicates and multiple comparisons (Fisher's LSD) were

965

conducted for significant differences among samples. Different letters represent

966

significant differences between samples.

967

C. Transient expression assays of pWRKY53Pro:GUS reporter in protoplasts isolated

968

from WT, hda9-1, and HDA9-OE (oe9-2) plants with or without TSA treatment. The

969

protoplasts were co-transformed with the reporter (pWRKY53Pro:GUS) and the

970

effector (p35S:WRKY53-GFP) or the control (35S:GFP) vectors, and GUS enzyme

971

activity was tested after 30 min. Bars represent means +/- standard errors from three

972

biological replicates and multiple comparisons (Fisher's LSD) were conducted to

973

compare significant differences among samples.

974

D. Transient expression assays of protoplasts co-transformed with the reporter

975

(pWRKY53Pro:GUS) and the wild type effector (p35S:WRKY53-GFP) or the mutant 33

976

effector (p35S:WRKY53M-GFP) vectors. Protoplasts were isolated from Arabidopsis

977

plants treated with or without TSA. GUS enzyme activity was tested after 30 min.

978

Bars represent means +/- standard errors from three biological replicates and multiple

979

comparisons (Fisher's LSD) were conducted to compare significant differences among

980

samples.

981 982

Figure 6. WRKY53 displays an inhibitory effect on HDA9 deacetylase activity

983

A. WRKY53 affects histone acetylation in plants. Histone acetylation levels were

984

detected by immunoblots with antibodies of the indicated histone acetylation marks in

985

WT, hda9, wrky53, HDA9-OE (oe9-2), and WRKY53-OE (oe53-wt1) seedlings. Two

986

repeats of the immunoblots are shown. Anti-H3 antibody was used control loading.

987

B. In vitro deacetylase activity assays, HDA9 protein were isolated by

988

immunoprecipitation with anti-HDA9 from wild type, wrky53-1 and WRKY53-OE

989

plants and the HDAC activity was tested in the presence of E. coli-produced

990

WRKY53-GST or GST alone. TSA were added to the assays as controls. The

991

deacetylase activity was expressed by relative fluorescent intensity using a

992

fluorescence microplate reader. Bars represent means +/- standard errors from three

993

biological replicates and multiple comparisons (Fisher's LSD) were conducted to

994

compare significant differences among samples.

995

C. In vitro HDA9 HDAC activity assays. HDA9 protein was isolated by

996

immunoprecipitation with anti-HDA9 from wild type plants. HDAC activity was

997

tested

998

WRKY53-DB-GST. The deacetylase activity was expressed by relative fluorescent

999

intensity using a fluorescence microplate reader. Bars represent means +/- standard

1000

errors from three biological replicates and multiple comparisons (Fisher's LSD) were

1001

conducted to compare significant differences among samples.

1002

D. Computational model of protein interaction between HDA9 (green) and WRKY53

1003

(red). Residues Gly213, Thr214, His215 and Thr216 within WRKY53 DNA binding

1004

domain showed direct contacts with Cys32, Met33, Thr34, and His35 within HDA9

1005

HDAC domain. Pro277 of WRKY53 was close to Val305 and Ala306 of HDA9.

in

the

presence

of

WRKY53-Nter-GST,

34

WRKY53-Cter-GST

or

1006 1007

Figure 7. Proposed model of HDA9 and WRKY53 functional interaction

1008

HDA9 interacts with and reduces lysine acetylation of WRKY53, thereby inhibiting

1009

WRKY53 transcriptional activity which functions as a higher hierarchy activator of

1010

downstream

1011

WRKY53 can reduce HDA9 activity to deacetylate nucleosomal histones. Ac, lysine

1012

acetylation.

W-box-containing

stress-responsive

WRKY

genes.

Conversely,

1013 1014

Supplemental Files

1015 1016

Supplemental Figure1. Characterization of wrky53 mutants and WRKY53 and HDA9

1017

transgenic plants.

1018

A. wrky53 T-DNA insertion lines and genotyping by PCR. The insertion positions of

1019

the 2 mutants (wrky53-1, wrky53 -2) are indicated. The primers sued for SALK

1020

T-DNA

1021

(http://signal.salk.edu/tdnaprimers.2.html). LP and RP are respectively left and right

1022

boarder primers. LB is the left T-DNA boarder primer. Results of genotyping of

1023

wrky53 T-DNA mutants including DNA-level homozygous analysis and RNA-level

1024

expression analysis are shown.

1025

B. Production of 35S:HDA9-GFP lines. HDA9 cDNA was fused to GFP under the

1026

control of 35S promoter. The construct was used to transform wild type plants.

1027

Transcript and protein levels of the fusion protein in independent lines were tested by

1028

RT-qPCR and immunoblotting with anti-GFP. Bars represent means +/- standard

1029

errors from three biological replicates and multiple comparisons (Fisher's LSD) were

1030

conducted for significant differences among samples. Different letters represent

1031

significant differences (p<0.05 and >2 fold) between samples.

genotyping

were

designed

with

T-DNA

Primer

Design

1032

C. Production of 35S: WRKY53-GFP in wild type and the different mutant and

1033

transgenic backgrounds. The construct was used to transform wrky53-1 mutant plant

1034

to obtain the complementation lines (c1 to c3), wild type plants to obtain the

1035

WRKY53-OE lines (oe53-wt1, oe53-wt2), hda9-1 mutants to obtain the oe53-hda9-1 35

1036

and oe53-hda9-3 lines, and HDA9-OE plants to obtain oe53-oe9 double

1037

overexpression lines. Transcript and proteins levels of WRKY53-GFP in the different

1038

background were tested by RT-PCR and immunoblotting using anti-GFP antibody.

1039

Anti-actin was used as control. Bars represent means +/- standard errors from three

1040

biological replicates and multiple comparisons (Fisher's LSD) were conducted for

1041

significant differences among samples. Different letters represent significant

1042

differences (p<0.05 and >1.5 fold) between samples.

1043 1044

Supplemental Figure 2. Production of HDA9-GST, WRKY-GST, and GST

1045

recombinant proteins and anti-HDA9 antibody.

1046

A. Production and purification of the HDA9-GST protein.

1047

B. Tests of rabbit antiserum raised against HDA9-GST by immunoblotting of nuclear

1048

proteins extracted from wild type and hda9-1 and hda9-2 mutant plants grown at

1049

normal or NaCl-treated conditions as indicated. HDA9-GST was loaded as a positive

1050

control. Anti-actin was used for loading control of plant proteins.

1051

C. E. coli- production and purification of WRKY53-GST and GST proteins.

1052

D. E. coli- production and purification of WRKY53-Nter-GST, WRKY53-DB-GST

1053

and WRKY53-Cter-GST proteins.

1054 1055

Supplemental Figure3.

Yeast two hybrid analysis of interaction between HDAC

1056

(HDA9, HDA6, HDA19 and SRT1) and WRKY (WRKY53, WRKY22, WRKY38,

1057

WRKY62) proteins. Yeast gold cells were co-transformed with a combination of the

1058

indicated plasmids. Yeast cells were plated on selective media DDO/X/A and

1059

QDO/X/A and then incubated for 5 days. DDO/X/A is the double dropout medium:

1060

SD/-Leu/-Trp supplemented with X-α-Gal and Aureobasidin A. QDO/X/A is the

1061

quadruple dropout medium: SD/-Ade/-His/-Leu/-Trp/ supplemented with X-α-Gal and

1062

Aureobasidin A.

1063 1064

Supplemental Figure 4. Phenotypes of the mutant and transgenic plants.

1065

A Repetition of growth phenotype of the different lines grown in normal and stress 36

1066

(100 mM NaCl) conditions shown in Fig 2B.

1067

B. Growth phenotypes of WRKY53 overexpression wkry53-1, hda9-1, and HDA9-OE

1068

backgrounds in normal and salt (100 mM NaCl) conditions.

1069 1070

Supplemental Figure 5. ChIP-PCR analysis of WRKY53-GFP and HDA9 association

1071

to promoters of WRKY53 downstream genes. Chromatin fragments were isolated

1072

from wrky53-complementation plants and immunoprecipitated with anti-GFP or

1073

anti-HDA9 and ananyzed by qPCR with primer sets corresponding to two regions

1074

(PW and P) of the genes (the red line indicates W-box). Y-axis: levels of precipitated

1075

chromatin relative to input. Bars represent means +/- standard errors from three

1076

biological replicates and multiple comparisons (Fisher's LSD) were conducted for

1077

significant differences among samples. Different letters represent significant

1078

differences between samples

1079 1080

Supplemental Figure 6. Repression of WRKY53 expression by HDA9 is independent

1081

of histone deacetylation

1082

H3K9ac, H3K27ac, H3K27me3, H3K27me2, H3K4me2 and H3K4me3 modifications

1083

of the WRKY53 locus were detected by ChIP-PCR in WT, hda9-1, and HDA9-OE

1084

(oe9-2) plants under normal and stress conditions (100 mM NaCl). Two regions (P1

1085

and P3, as indicated in the top panel) of the genes were tested by ChIP-PCR. Y-axis:

1086

levels of precipitated chromatin relative to input. Bars represent means +/- standard

1087

errors from three biological replicates. Multiple comparisons (Fisher's LSD) were

1088

conducted to test significant differences among samples.

1089 1090

Supplemental Figure 7. Amino acid sequence comparison of WRKY proteins.

1091

All the WRKYs protein sequences were downloaded from TAIR, the comparison

1092

were carried out by ClustalX. The lysines in red were acetylated in WRKY53, the

1093

green frame indicated the conserved motif among WRKYs.

1094 1095

Supplemental Figure 8. Representative MS graphs of acetylated lysine sites in 37

1096

WRKY53 in the presence or absence of HDA9-GST.

1097 1098

Supplemental Figure 9. The original western blot image of Figures 4, 5 and 6.

1099 1100

Supplemental Table S1. WRKY53 peptides identified by mass spectrometry.

1101

Supplemental Table S2. The primers used in this study.

1102

38