Introduction

Soil contamination with cadmium (Cd) is a global environmental problem because of its high toxicity and easy transmission through the food chain. Under excess Cd exposure, plants show limited plant growth, retarded photosynthesis, decreased respiration, leaf chlorosis, inhibited root growth, increased synthesis of reactive oxygen species (ROS), and even death (Smeets et al. 2008; Sharma and Dietz 2009; Wu et al. 2015).

Plants have employed various strategies to resist Cd stress and those include uptake, accumulation, translocation, and detoxification of the metal. In general, there are two major strategies for detoxifying excess Cd absorbed from soil, namely excluder strategy, which avoids Cd entering the plants or extrudes excess Cd (Mills et al. 2005; Moreno et al. 2008; Morel et al. 2009); and tolerance strategy, which chelates Cd in the cytosol and sequesters Cd in vacuoles (Verbruggen et al. 2009; Lin and Aarts 2012). Chelation, such as that with glutathione (GSH) and phytochelatins (PCs), is an important component involved in these mechanisms. GSH functions as a metal chelator, a cellular antioxidant, and a ROS signalling molecule, which is a γ-Glu-Cys-Gly tripeptide (Jozefczak et al. 2012; Seth et al. 2012). In Arabidopsis thaliana, there are two ATP-dependent enzymes, γ-glutamylcysteine synthetase (GSH1; EC.6.3.2.2) and GSH synthetase (GSH2; EC.6.3.2.3), involved in the GSH synthesis pathway (Shanmugam et al. 2012; Noctor et al. 2012). A GSH-deficient mutant, called cad2-1 mutant, is sensitive to Cd stress (Cobbett et al. 1998). Many studies have confirmed that PCs containing (γ-Glu-Cys)n-Gly (n = 2–11) are enzymatically synthesized directly from GSH by the enzyme PC synthase (PCS) (Rauser 1995). In Arabidopsis thaliana, PCSs are encoded by two genes, PCS1 and PCS2, which are strongly induced by Cd stress (May et al. 1998; Noctor et al. 2002; Kühnlenz et al. 2014). Thus, GSH and PCs are known to have key roles in heavy metal detoxification.

It is commonly known that transcription factors (TFs) are critical components in signalling networks owing to their functions as regulators of the defence response by modulating the expression of the genes involved in defence response. WRKY TFs specifically recognize and bind the W-box (contains a TGAC core sequence) (Liu et al. 2014; Li et al. 2016). Many reports have demonstrated that WRKY TFs can be either positive or negative regulators of plant development and defence (Dong et al. 2003; Kalde et al. 2003; Luo et al. 2005; Devaiah et al. 2007; Eulgem and Somssich 2007; Chen et al. 2009, 2012; Wang et al. 2010; Ding et al. 2013; Kim et al. 2013; Liu et al. 2014; Li et al. 2016). WRKY12, which belongs to the subgroup IIc of the WRKY gene family, is involved in secondary cell wall formation by directly inhibiting NST2 expression in pith cells (Wang et al. 2010) and modulates flowering time under short-day conditions (Li et al. 2016). A growing number of studies have found that TFs participate in the response to Cd stress by regulating the expression of downstream target genes such as CaPF1, HsfA4a, BjCdR15/TGA3, and ZAT6 (Tang et al. 2005; Shim et al. 2009; Farinati et al. 2010; Chen et al. 2016). However, the possible role of WRKY12 under Cd stress and the details of the underlying mechanism are yet to be revealed.

The aim of the present study was to investigate the functions and the underlying mechanisms of WRKY12 in plant response to Cd stress.

Results

WRKY12 is repressed by Cd stress

Considering that WRKY13 is involved in the regulation of Cd stress response (Sheng et al. 2018), WRKY12 and WRKY13 belong to the same subgroup in the WRKY gene family, and their amino acid sequences are highly similar (Li et al. 2016), we speculated that WRKY12 may also respond to Cd stress. To verify this hypothesis, we analysed the transcription level of WRKY12 under Cd stress. The results showed that the WRKY12 transcription was inhibited by Cd stress (Fig. 1A), thus providing evidence that WRKY12 gene may be involved in regulating plant Cd tolerance.

Fig. 1
figure 1

Expression patterns of WRKY12. A The expression of WRKY12 was induced by Cd stress. Two-week-old WT seedlings were treated with CdCl2 (50 µM) for 0 and 6 h for qRT-PCR analysis. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as the internal control. Data present means ± SE (n = 3). Bars with different lowercase letters are significantly different at P < 0.05 (Tukey’s test). B The expression pattern of WRKY12 was analysed by qRT-PCR in different tissues of wild-type (WT) plants. GAPDH was used as internal control. The samples were isolated from the root, rosette leaves, stem, cauline leaves, inflorescences, and seeds of the WT plants

Full size image

We also examined the expression pattern of WRKY12 in different tissues and detected the expression of WRKY12 in all tissues examined, but the highest level was found in the stem (Fig. 1B).

Loss-of-function mutations of the WRKY12 gene lead to enhanced Cd tolerance

To further verify the potential regulatory role of WRKY12 in response to Cd stress, two T-DNA insertion mutants were obtained from the SALK Arabidopsis T-DNA mutant collection, named wrky12-1 (CS 435919) and wrky12-2 (CS 374453) (Alonso et al. 2003). The two insertion sites of the wrky12-1 and wrky12-2 were located upstream of the gene coding region (Fig. 2A). qRT-PCR results detected no transcripts of WRKY12 in the two mutants (Fig. 2B). There were no significant differences in growth between the WT and mutants in the control 1/2 MS medium. However, under Cd stress, the mutants exhibited enhanced Cd tolerance compared with the WT (Fig. 2C). Compared with WT plants, the root length and FW of the mutants were significantly increased in a Cd concentration-dependent manner (P < 0.05; Fig. 2D, E). Together, these results indicated that the WRKY12 loss-of-function results in increased Cd tolerance.

Fig. 2
figure 2

The wrky12 mutants are tolerant to Cd stress. A Genetic map of the WRKY12 gene. Black boxes and black lines indicate exons and introns, respectively, in WRKY12. The positions of T-DNA insertion are indicated by triangles. B The expression of WRKY12 in the wrky12 mutants and wild-type (WT) plants by qRT-PCR using glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as internal control. Data are presented as the means ± SE (n = 3). Bars with different lowercase letters are significantly different at P < 0.05 (Tukey’s test). C The WRKY12 loss-of-function mutation produced phenotypes resistant to Cd stress. Seedlings were grown on half-strength Murashige and Skoog (1/2 MS) medium for 3 days then treated with 0, 50, or 7 5 µM CdCl2 for 2 weeks. Bar = 1 cm. D and E Root length (D) and fresh weight (E) of plants described in C. Data present the means ± SE (n = 3). Bars with different lowercase letters are significantly different at P < 0.05 (Tukey’s test)

Full size image

To test whether wkry12 mutants are also involved in the regulation of other heavy metal stresses, the wkry12 mutants were grown on 1/2 MS medium with or without Pb(NO3)2, Na3AsO4, and ZnSO4. When Na3AsO4 or ZnSO4 was added, no significant differences between WT and wrky12 mutant plants were observed. However, the wrky12 mutants showed enhanced tolerance to Pb(NO3)2 (Supplementary Fig. S1).

Overexpression of the WRKY12 gene results in enhanced Cd sensitivity

To further examine whether WRKY12 regulates Cd tolerance, we generated a 35S:WRKY12 construct and transformed it into the WT plants. At least five transgenic lines were obtained, of which two lines OE5 and OE11 (Fig. 3A) were chosen for further study. There were no differences in growth and development among WT, wrky12 mutant, and WRKY12-OE plants under normal growth conditions (Supplementary Fig. S2).

Fig. 3
figure 3

Overexpression of WRKY12 leads to sensitivity to Cd stress. A qRT-PCR test of WRKY12 expression in wild-type (WT) and WRKY12-overexpressing (OE) lines. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as internal control. Data present the means ± SE (n = 3). Bars with different lowercase letters are significantly different at P < 0.05 (Tukey’s test). B Growth of WT and WRKY12-OE lines (OE5 and OE9) under Cd stress. Seedlings were grown on half-strength Murashige and Skoog (1/2 MS) medium with 0, 50, or 75 µM CdCl2 for 2 weeks. Bar = 1 cm. C and D Root length (C) and fresh weight (D) of plants described in B. Data present the means ± SE (n = 3). Bars with different lowercase letters are significantly different at P < 0.05 (Tukey’s test)

Full size image

In 1/2 MS media, no significant differences were observed between the WT and the two overexpressing plants; however, under Cd stress, the WRKY12-OE plants displayed enhanced Cd sensitivity compared with the WT plants (Fig. 3B). This was further confirmed by quantitative analysis of the root length (P < 0.05; Fig. 3C) and the fresh weight (P < 0.05; Fig. 3D). These results suggest that WRKY12 negatively regulates Cd tolerance.

Alteration of WRKY12 gene expression affects Cd accumulation

To test whether the expression change of WRKY12 affects Cd content, we measured Cd content in WT, wrky12 mutant, and WRKY12-OE plants under Cd stress. A higher Cd content was observed in roots and shoots of the wrky12 mutant plants than in the WT (P < 0.05; Fig. 4). In contrast, Cd content was reduced in the WRKY12-OE lines under Cd stress compared with that in the WT (P < 0.05; Fig. 4). These results suggest that the sequestration mechanism may be involved in WRKY12-mediated Cd tolerance.

Fig. 4
figure 4

Cd contents in shoots and roots of wild-type (WT), wrky12 mutants (A), and WRKY12-overexpressing (OE) lines (B) treated with CdCl2. These plants were grown on half-strength Murashige and Skoog medium containing 50 µM CdCl2 for 2 weeks, and roots and shoots of these samples were collected. Data present the means ± SE (n = 3). Bars with different lowercase letters are significantly different at P < 0.05 (Tukey’s test)

Full size image

WRKY12 gene negatively regulates Cd tolerance via the GSH-dependent PC synthesis pathway

The GSH-dependent PC synthesis pathway is an important pathway for heavy metal detoxification in plants. Accordingly, we tested whether WRKY12 regulates Cd tolerance through the GSH-dependent PC synthesis pathway by treating plants with BSO, a GSH synthesis inhibitor. The WT and wrky12 mutants exhibited a similar growth when grown in 1/2 MS media with or without BSO alone (Fig. 5A, B). When both BSO and Cd were applied to the medium, the tolerant phenotypes of the wrky12 mutants disappeared (Fig. 5A, B). These results suggest that the accumulation of Cd caused by the knockout of WRKY12 is GSH-dependent.

Fig. 5
figure 5

WRKY12 mediates Cd tolerance through the glutathione-dependent pathway. A Analysis of buthionine sulfoximine (BSO) effect on the growth of wild-type (WT) and wrky12 mutant plants. Seedlings were grown on half-strength Murashige and Skoog (1/2 MS) medium with no CdCl2, 50 µM CdCl2, or 0.1 mM BSO for 2 weeks. B Root length of plants described in A. Data are presented as means ± SE. Bars with different lowercase letters are significantly different at P < 0.05 (Tukey’s test). C and D Measurements of total glutathione (GSH plus 2 glutathione disulphide [GSSG]) (C) and phytochelatin (PC) (D) contents in the WT, wrky12 mutants, and the WRKY12-overexpressing (OE) lines. Seedlings were grown on 1/2 MS medium for 2 weeks, and after treating them with 0 or 50 µM CdCl2 for 24 h, their GSH/GSSG (C) and PC (D) contents were quantified. Data present the means ± SE (n = 3). Bars with different lowercase letters are significantly different at P < 0.05 (Tukey’s test)

Full size image

Next, we further analysed the levels of GSH and PCs in WT, wrky12, and WRKY12-OE plants subjected to Cd treatment. No significant difference was detected in total glutathione (GSH plus 2 glutathione disulphide [GSSG]) between the wild type, the wrky12 mutants, and the WRKY12-OE lines without Cd treatment (Fig. 5C). Under Cd stress, GSH concentrations decreased significantly in the wild type, wrky12 mutants, and WRKY12-OE plants; however, compared with WT plants, its content was higher in wrky12 mutants and lower in WRKY12-OE lines (Fig. 5C). In addition, the PC content was increased significantly in the wrky12 mutants and decreased in the WRKY12-OE lines compared with the WT (Fig. 5D). These results suggest that WRKY12 negatively regulates Cd tolerance via the GSH-dependent PC synthesis pathway.

WRKY12 protein represses the expression of the PCs synthesis-related genes

On the basis of the above results, we hypothesized that WRKY12 regulates the expressions of PC synthesis-related genes. Therefore, the genes GSH1, GSH2, PCS1, and PCS2, which are involved in this pathway, were analysed (Zhu et al. 1999a, b; Brunetti et al. 2011; Kühnlenz et al. 2014). Compared with the expressions in the WT, the transcription levels of these genes induced by CdCl2 were significantly increased in the wrky12 mutants but decreased in the WRKY12-OE seedlings (Fig. 6A). These results imply that WKY12 represses the expressions of the PCs synthesis-related genes. In addition, we also detected the expression pattern of PDR8, a gene encoding a Cd extrusion pump conferring Cd tolerance. Our data showed no significant differences in the transcription of PDR8 between the WT, wrky12 mutants, and WRKY12-OE seedlings (Supplementary Fig. S3).

Fig. 6
figure 6

The expressions of genes in glutathione (GSH)-dependent phytochelatin (PC) synthesis pathway were suppressed by WRKY12. A The transcriptions of GSH/PC synthetic genes GSH1, GSH2, PCS1, and PCS2 in the wild type (WT), wrky12 mutant, and WRKY12-overexpressing (OE) lines by quantitative analysis. Two-week-old seedlings were treated with 0 or 50 µM CdCl2 for 6 h. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as internal control. All the data shown here are presented as the means ± SE (n = 3). Bars with different lowercase letters are significantly different at P < 0.05 (Tukey’s test). B Schematics of all constructs used for transient expression assays in Nicotiana benthamiana leaves. The GSH1, GSH2, PCS1, or PCS2 promoter was fused to the β-glucuronidase (GUS) reporter gene. 35S:WRKY12 acted as an effector. The 35S promoter was fused to the green fluorescent protein (GFP) gene as an internal control. C GUS staining showed expressions of ProGSH1, ProGSH2, ProPCS1, or ProPCS2 after co-expression with WRKY12. ProGSH1:GUS, ProGSH2:GUS, ProPCS1:GUS, or ProPCS2:GUS was co-transformed with 35S:WRKY12 into N. benthamiana, and the 35S-empty vector was used as effector plasmid control

Full size image

To further test whether WRKY12 suppresses the transcriptions of GSH1, GSH2, PCS1, and PCS2, we observed the GUS activity by transient expression analysis in Nicotiana benthamiana. As shown in Fig. 6B, C, WRKY12 inhibited the expressions of ProGSH1:GUS, ProGSH2:GUS, ProPCS1:GUS, and ProPCS2:GUS. These results suggest that WRKY12 has the capacity of inhibiting reporter activity driven by the promoters of GSH1, GSH2, PCS1, and PCS2.

The WRKY12 gene directly binds to the promoter of GSH1

It is known that WRKY proteins can bind to W-box motifs present in the promoters of target genes (Ulker and Somssich 2004; Jiang et al. 2014). Using promoter sequence analysis, we detected many W-boxes in the promoter regions of GSH1, GSH2, PCS1, and PCS2 (Fig. 7A) and speculated that WRKY12 may directly bind to these boxes and affect their expressions. To test this speculation, we performed ChIP-qPCR assay using seedlings of a transgenic line WRKY12-GFP. As shown in Fig. 7B, WRKY12 could directly bind to the W-box of the promoter in GSH1 but not in GSH2, PCS1, and PCS2. These results indicate that WRKY12 directly regulates the transcription of GSH1 by binding to its promoter, and indirectly represses the expression of GSH2, PCS1, and PCS2.

Fig. 7
figure 7

WRKY12 directly binds to the promoter of GSH1 in vivo. A Schematic diagram of the GSH1, GSH2, PCS1, and PCS2 promoters showing W-box presence in different regions. Bars indicate W-box (TGAC); lines beneath the bars represent the sequences for chromatin immunoprecipitation (ChIP) assays. B The direct binding of WRKY12 to the W-box of the GSH1, GSH2, PCS1, and PCS2 promoters using ChIP-real-time PCR assay. Input DNAs were used as internal control. Data present the means ± SE (n = 3). Bars with different lowercase letters are significantly different at P < 0.05 (Tukey’s test)

Full size image

Discussion

The WRKY TF family has been reported to be involved in a variety of heavy metal stresses. For example, WRKY46 responds to Al stress by negatively regulating ALMT1 expression, and loss of function of WRKY46 leads to increased malate secretion and reduced Al accumulation (Ding et al. 2013). The activities of Zea mays ZmWRKY4 and Thellungiella salsuginea EsWRKY33 were induced by CdCl2 and AgNO3, respectively (Mucha et al. 2015; Hong et al. 2017). WRKY53 in Thlaspi caerulescens has also been reported to play a potential role in Cd stress response (Wei et al. 2008). ThWRKY7 can specifically bind to the promoter of ThVHAc1 and improve Cd stress tolerance by regulating ROS homeostasis in Tamarix hispida (Yang et al. 2016). We also reported that WRKY13 positively regulates Cd tolerance in Arabidopsis (Sheng et al. 2018). In the present study, we found that WRKY12 plays a negative role in regulating Cd stress tolerance and its expression is inhibited by Cd stress. Moreover, the wrky12 mutants displayed enhanced Cd tolerance; conversely, the WRKY12-overexpressing lines showed a phenotype sensitive to Cd stress. These data suggested that WRKY12 was able to mediate the Cd stress response in Arabidopsis. Additionally, no differences were observed between the mutants and the WT under As and Zn stresses, but the tolerance phenotype was observed under Pb stress. One possible explanation is that plant response to Pb stress may also require phytochelatins (Clemens 2006; Verbruggen et al. 2009; Chen et al. 2015).

The GSH-dependent PC synthesis pathway is an important mechanism for chelating and detoxifying Cd in higher plants (Cobbett and Goldsbrough 2002; Lin and Aarts 2012). Under Cd stress, GSH and PC could chelate Cd ions to form Cd-GSH and Cd-PCs complexes, which are then sequestered to the vacuoles (Cobbett and Goldsbrough 2002). Therefore, PCs play a vital role in Cd accumulation and detoxification though PC-conjugated vacuolar sequestration (Clemens 2006). In the present study, when Cd and BSO co-existed, the growth of the mutants was the same as that of the WT plants. When treated with Cd stress, compared with WT plants, the concentrations of GSH and total PCs were decreased in WRKY12-OE plants and increased in the mutants. It is well known that the genes, including GSH1, GSH2, PCS1, and PCS2, involved in the synthesis of GSH and PCs are closely related with this phenomenon. Overexpression of GSH1, GSH2, PCS1, and PCS2 enhanced Cd tolerance by increasing the contents of GSH and PCs (Zhu et al. 1999a, b; Cazalé and Clemens 2001; Gasic and Korban 2007; Brunetti et al. 2011; Kühnlenz et al. 2014). Co-overexpressing GSH1 and PCS1 increased the tolerance and accumulation of Cd in Arabidopsis thaliana (Guo et al. 2008). In our study, we found that the transcription levels of these genes were induced in the wrky12 mutants, while they were the opposite in the WRKY12-OE plants. Therefore, WRKY12-mediated Cd tolerance is dependent on the synthesis of GSH and PCs, and WRKY12 is a negative regulator.

The WRKY TFs respond to abiotic and biotic stresses as either positive or negative regulators. They can activate or repress the transcriptions of stress-related and co-regulated genes by directly recognizing and binding to the W-box, which contains a core sequence TGAC present in the promoters of those genes (Eulgem and Somssich 2007). Promoter sequence analysis revealed that W-box is enriched in the promoters of PCs synthesis genes. ChIP-qPCR assays demonstrated that WRKY12 directly bound to the W-box in the promoter of GSH1, but not in the promoter of GSH2, PCS1, and PCS2. Interestingly, the GUS staining assays showed that WRKY12 repressed the activities of the promoters of GSH1, GSH2, PCS1, and PCS2. These results suggest that WRKY12 directly regulates the expression of GSH1 but indirectly suppresses the expression of GSH2, PCS1, and PCS2.

Moreover, we further studied the relationship between WRKY12 and PDR8 using qRT-PCR. The results showed that PDR8 was not involved in WRKY12-mediated Cd tolerance. Previously, we found that WRKY13 positively regulates Cd stress by activating the expression of PDR8 (Sheng et al. 2018). Moreover, we also demonstrated that a GSH-dependent PC synthesis pathway is not involved in the mechanism of WRKY13-mediated Cd tolerance (Sheng et al. 2018). Thus, the opposite role of WRKY12 and WRKY13 in the regulation of plant Cd tolerance is though different pathways, although they both belong to the same subgroup within the WRKY gene family.

We propose a possible working mode for WRKY12 in regulating Cd accumulation and tolerance (Fig. 8): WRKY12 directly targets GSH1 and indirectly represses the expressions of PC synthesis-related genes to negatively regulate Cd tolerance.

Fig. 8
figure 8

A working model of the role of WRKY12 in plant responses to Cd stress. Cd stress inhibits the expression of WRKY12, releasing its binding to the promoter of GSH1 and co-ordinately facilitating phytochelatin (PC) synthesis-related gene expression. This results in increased contents of glutathione (GSH) and PCs, which leads to increased Cd tolerance

Full size image

Materials and methods

Plant materials and growth condition

All seeds of Arabidopsis thaliana (L.) Heynh. (including wild-type [WT] ecotype Col-0, wrky12 mutants, and transgenic plants) were surface-sterilized and planted on half-strength Murashige–Skoog (1/2 MS) media (Murashige and Skoog 1962) containing 1% sucrose and 1% agar [Sangon Biotech] (pH 5.8). The seeds were vernalized in the dark at 4 °C for 3 days and then grown in a controlled culture room under long day (16/8 h light/dark) at 22 °C (light intensity of 100 µmol m−2 s−1).

The T-DNA insertion lines for WRKY12 (wrky12-1:CS435919; wrky12-2:CS 374453) were obtained from the Arabidopsis Biological Resource Center (ABRC) at Ohio State University, USA. The T-DNA insertion sites were confirmed by PCR amplification using the primers showed in Supplementary Table S1.

Generation of transgenic plants

To generate WRKY12-overexpressing (OE) and WRKY12-green fluorescent protein (GFP) plants, full-length WRKY12 cDNA was amplified with specific primers (primers are listed in Supplementary Table S1), cloned into the pCAMBIA1301 or pXB94 (pART27 with expanded restriction sites, 35S promoter, and GFP reporter) at the XbaI and HindIII restriction sites (named 35S::WRKY12 and 35S::WRKY12::GFP), and transformed into Arabidopsis using the floral dip method (Clough and Bent 1998). The obtained homozygous lines were selected for further study.

Transcript analysis by qRT-PCR

Plant total RNA was extracted from the seedlings, grown as described above using TRIzol reagent (Invitrogen), and the first-stand cDNA was synthesized by a RevertAid™ First Strand cDNA Synthesis Kit (Fermentas). Quantitative real-time PCR was conducted using SYBR Green qPCR SuperMix-UDG (Invitrogen) according to the instructions provided for the Bio-Rad iCycler iQ system. The PCR amplifications for each sample were quantified at least in triplicate and normalized using glyceraldehyde 3-phosphate dehydrogenase (GAPDH) gene as an internal control. The primers used are listed in Supplementary Table S1.

Chemical treatments

For Cd tolerance assay, following a 3-day growth, Arabidopsis seedlings were transplanted to 1/2 MS medium in the absence or presence of indicated concentrations of CdCl2. The root lengths and FW were measured after 10 days of cultivation. All the tests were conducted in triplicate, and approximately 30 plants were used for each measurement.

To test the function of GSH under WRKY12-mediated Cd tolerance, buthionine sulfoximine (BSO, Sigma Aldrich), a GSH synthesis-inhibitor, was used to treat the wrky12 mutants cultivated on media with or without CdCl2.

Analysis of Cd content

Cd content assay was performed as described by Kim et al. (2006). Briefly, the roots and aboveground parts of the plants (WT, wrky12 mutants, and WRKY12-OE) grown on 1/2 MS medium containing CdCl2 for 2 weeks were harvested apart, digested, and the digested samples were analysed using an atomic absorption spectrometer (Solaar M6; Thermo Fisher).

Measurement of GSH and PC contents

Two-week-old WT, wrky12 mutant, and WRKY12-OE plants were treated or not with CdCl2 for 24 h and then sampled for determination of GSH and PC contents. The specific method was carried out as described previously (Chen et al. 2015).

Cloning and transient expression assay

The full-length promoters of GSH1, GSH2, PCS1, and PCS2 were PCR-amplified from Arabidopsis genomic DNA using specific primers (Supplementary Table S1). These promoters were cloned into the vector pXB93 (pART27 with expanded restriction sites and β-glucuronidase [GUS] reporter), named ProGSH1::GUS, ProGSH2::GUS, ProPCS1::GUS, and ProPCS2::GUS. These four constructs were co-transformed into epidermal cells of Nicotiana benthamiana with the 35S:WRKY12 construct. Samples were then collected and labelled, and GUS staining was performed as described previously (Xu et al. 2006).

ChIP-qPCR assay

Chromatin immunoprecipitation coupled with real-time PCR (ChIP-qPCR) assays were performed as previously described (Kaufmann et al. 2010) with an anti-GFP antibody (Abmart). In brief, 3 g of 10-day-old 35S::WRKY12:GFP plants as the experimental group and 35S::GFP plants as the control group were harvested and fixed. The fixed plant tissues were sonicated with ultrasonic cell disruption and divided into three parts: one part was used for input DNA, and the two other parts were incubated with anti-GFP antibody. The relative concentrations of DNA fragments were analysed by qRT-PCR using primers listed in Supplementary Table S1.