WRKY46 functions as a transcriptional repressor of ALMT1, regulating aluminum-induced malate secretion in Arabidopsis



Aluminum (Al) toxicity is the major limiting factor for crop production on acid soils, but the transcriptional regulation of Al tolerance genes is largely unknown. Here, we found that the expression of a WRKY domain-containing transcription factor WRKY46 is inhibited by Al and expressed in root stele, whereas the expression of ALMT1, which encodes a malate efflux transporter, is induced by Al stress and spatially co-localized with WRKY46 in root stele, indicating the possible interaction between WRKY46 and ALMT1 in Arabidopsis. Mutation of WRKY46 by T–DNA insertion leads to better root growth under Al stress, and lower root Al content compared with the wild-type Col–0. The wrky46 mutant shows increased root malate secretion, which is consistent with the higher ALMT1 expression in the mutant. Transient expression analysis using truncated promoter of ALMT1 showed that ALMT1 expression can be inhibited by WRKY46 in tobacco leaves. The yeast one-hybrid assay and ChIP-qPCR analysis revealed that WRKY46 directly binds to ALMT1 promoter through specific W–boxes. Taken together, we demonstrated that WRKY46 is a negative regulator of ALMT1, mutation of WRKY46 leads to increased malate secretion and reduced Al accumulation in root apices, and thus confers higher Al resistance.


Acid soils account for about 30% of the world's total land area and over 50% of the world's potentially arable lands (von Uexküll and Mutert, 1995; Kochian et al., 2004). As the third most abundant element in the Earth's crust, aluminum (Al) is solubilized from aluminosilicate clay minerals to the toxic species Al3+ in soils with pH values of 5 or below (Kochian et al., 2004). As micromolar concentrations of Al3+ can rapidly inhibit root elongation (Kochian, 1995; Poschenrieder et al., 2008), Al stress has been considered as one of the most important constraints for agricultural production in acid soils (Kochian et al., 2004). Therefore, enhancing Al resistance in crops becomes a key to increasing crop productivity on these soils, which would subsequently help solve the problem of food shortage and biofuel production.

The Al-resistant mechanism employed by plants can be classified into external exclusion and internal detoxification, depending on whether Al is detoxified out of or within the plant cell. Although several mechanisms for the external exclusion of Al have been proposed (Kochian et al., 2004; Poschenrieder et al., 2008), the most well-studied strategy is the secretion of organic acid anions, including citrate, oxalate and malate, from the root apices (Ma et al., 2001; Ryan et al., 2001; Kochian et al., 2004). These organic acid anions prevent Al from binding to root cells via chelating Al externally, and thereby detoxify Al outside the cells (Ma et al., 2001; Ryan et al., 2001). Many transporter-encoding genes responsible for the Al-activated secretion of malate and citrate have been identified (Ryan et al., 2011; Delhaize et al., 2012). The malate transporter gene ALMT1 that conferred Al resistance was first cloned in wheat (Sasaki et al., 2004), and its homologous genes have been isolated in Arabidopsis and oilseed rape (Hoekenga et al., 2006; Ligaba et al., 2006); however, the HvALMT1 in Hordeum vulgare (barley) has been proven to be an anion channel to facilitate organic anion transport in stomatal function and expanding cells, but not to be associated with Al resistance (Gruber et al., 2010). Genes involved in the Al-induced or -activated citrate secretion (multidrug and toxic compound extrusion, MATE) have also been identified in barley, Sorghum, Arabidopsis thaliana, Zea mays (maize), Secale cereale (rye), Oryza sativa (rice) and Vigna umbellata (ricebean) (Furukawa et al., 2007; Magalhaes et al., 2007; Liu et al., 2009; Maron et al., 2010; Yokosho et al., 2010, 2011; Yang et al., 2011). But in Triticum spp. (wheat), the TaMATE1B is constitutively expressed and citrate secretion does not require Al activation (Ryan et al., 2009; Tovkach et al., 2013). In Arabidopsis, Al stress induces a high level of root malate exudation and a lower level of root citrate exudation. Using mutants of ALMT1 and MATE demonstrated that ALMT1 is a major contributor to Arabidopsis Al resistance, and MATE makes a significant but smaller contribution (Liu et al., 2009). Moreover, the two transporters function independently to confer Al resistance in Arabidopsis; however, the regulatory mechanisms of expression of these genes are still poorly understood.

Recently, two zinc-finger proteins, sensitive to proton rhizotoxicity 1 (STOP1) in Arabidopsis and Al3+ resistance transcription factor 1 (ART1) in rice, respectively, were identified by mutant analysis (Iuchi et al., 2007; Yamaji et al., 2009). The two proteins share significant sequence similarity and appear to act as transcription factors to enhance the expression of a range of Al resistance genes (Sawaki et al., 2009; Yamaji et al., 2009). In the absence of STOP1, Al-induced ALMT1 and MATE expression were completely suppressed in Arabidopsis (Liu et al., 2009). Nevertheless, the transcriptional regulation of Al resistance genes is largely unknown. Moreover, both STOP1 and ART1 are not Al-inducible genes (Iuchi et al., 2007; Yamaji et al., 2009), indicating that there may be other Al-responsive factors involved in the regulation of Al resistance genes.

The plant-specific WRKY domain-containing proteins comprise one of the largest transcription factor families in plants (Eulgem et al., 2000; Rushton et al., 2010). There are 74 members in Arabidopsis (Rushton et al., 2010). WRKY proteins contain one or two domains composed of the conserved amino acid sequence WRKYGQK, together with a novel zinc-finger-like motif (Ulker and Somssich, 2004). With the domains, they can activate or repress transcription through directly binding to the W–box, which has a core sequence (T)(T)TGAC(C/T) present in promoters of the target genes (Eulgem et al., 2000). WRKY transcription factors play critical roles in a great number of processes in Arabidopsis or other species (Ulker and Somssich, 2004; Rushton et al., 2010). In particular, they are involved in regulating responses to biotic stress, such as pathogen invasion, or abiotic stress, such as drought and high salinity (Rushton et al., 2010). In this study we found that Arabidopsis WRKY46, which was previously reported to participate in basal resistance against the pathogen Pseudomonas syringae (Hu et al., 2012), is responsive to Al. Furthermore, we demonstrated that it functions in Al sensitivity through regulating the expression of ALMT1.


WRKY46 is responsive to Al

WRKY proteins can be divided into three groups, based on the number of WRKY domains and the features of the zinc-finger-like motif (Eulgem et al., 2000). WRKY46 belongs to group III, the members of which have one WRKY domain that contains a C2HC zinc-finger-like motif instead of the C2H2 pattern found in groups I and II (Eulgem et al., 2000). The public microarray data compiled using the BAR HeatMapper tool suggested the converse expression pattern between WRKY46 and ALMT1 in response to multiple abiotic stresses, such as osmotic, drought, salt and UV–B stress (Figure S1). Moreover, WRKY46 expression was localized to the root apices by using the Arabidopsis eFP Browser (Figure S2), which is similar to ALMT1 localization, as previously described (Kobayashi et al., 2007). As ALMT1, encoding a malate transporter localized in the plasma membrane, plays a critical role in Arabidopsis Al resistance, we firstly performed a time-course expression analysis under Al treatment. As shown in Figure 1(a), WRKY46 transcripts were obviously decreased after 3 h of Al treatment, whereas the expression of ALMT1 was progressively increased, indicating that WRKY46 responds to Al.

Figure 1.

WRKY46 is responsive to aluminium (Al).

(a) Expression of WRKY46 and ALMT1 under Al treatment. RNA was extracted from 4–week-old Col–0 roots in hydroponic culture (pH 4.5) under control or with 50 μm AlCl3 for different time periods. ACT2 was used as internal reference gene. Three independent repeats were performed with similar results. The data are from one experiment. Error bars indicate SEs (n = 3).

(b) Expression of WRKY46 in roots via GUS staining without (i) or with Al treatment (ii). Two-week-old transgenic plants carrying a WRKY46 promoter::GUS construct were used for GUS staining. Scale bars: 50 μm. (c) GUS activity measured in protein extracts from roots with or without Al treatment. Activity units are given in nmol methyl-umbelliferone (μg protein)−1 min−1. Error bars correspond to the means of four independent experiments (*< 0.05).

To further verify the localization of WRKY46 in root, a WRKY46 promoter::GUS reporter construct was generated and used to transform wild-type Arabidopsis. GUS staining revealed that the WRKY46 was mainly expressed in the root stele (Figure 1b), which shows some deviation from Arabidopsis eFP Browser (Figure S2). Moreover, the GUS staining increased in the regions that were closed to the root apices (Figure 1b). Al treatment weakened the GUS staining and activity of WRKY46 (Figure 1b,c), confirming its response to Al.

Mutation of WRKY46 leads to increased Al resistance

To investigate the role of WRKY46 in Al response, we identified its T–DNA insertion mutant wrky46–1 (SAIL_1230_H01) that carries the T–DNA insertion in the last exon of the WRKY46 gene (Figure S3). Reverse transcription–polymerase chain reaction (RT–PCR) analysis confirmed the expression of WRKY46 in the mutant and complementation lines (Figure S3). The wrky46–1 mutant grew as well as wild type Col–0 in soil, in low-pH hydroponic culture and on agar plates. Figure 2(a) shows wrky46–1 and Col–0 grown on solid medium; however, when different concentrations of Al3+ were introduced into the medium, the mutant clearly showed greater Al resistance, with significantly longer roots than Col–0 (Figure 2b), suggesting that mutation of WRKY46 increases Al resistance. To verify that these phenotypes resulted from the loss of WRKY46 function, we performed experiments with wrky46–1 mutants transformed with the WRKY46 gene driven by its native promoter. The complemented lines had similar growth with Col–0 under Al treatment (Figure 2c). In order to confirm that the phenotype resulted from the mutation of WRKY46, we tested another T–DNA inserted line, wrky46–2 (SALK_134310C), and found that it exhibited a very similar trend to wrky46–1 when exposed to Al stress (Figure S4). These results demonstrated that WRKY46 is involved in Arabidopsis Al resistance.

Figure 2.

Mutation of WRKY46 enhances aluminium (Al) resistance.

(a) Wild-type Col–0 and wrky46–1 mutant plants grown on solid medium with or without Al treatment (pH 4.5). Scale bar: 1.0 cm.

(b) Comparison of relative root growth between wrky46–1 and Col–0 on solid medium with different Al concentrations. Three independent experiments were performed with similar results, each with about 30 seedlings. Error bars indicate SEs (n = 3, *< 0.05).

(c) Complementation tests. The relative root growth of Col–0, wrky46–1 and two independent complementation lines were measured under 200 μm AlCl3. Three independent experiments were performed with similar results, each with about 25 seedlings. Error bars indicate SEs (n = 3, *< 0.05).

As Al resistance is generally related to the levels of Al accumulated in the roots, Al content in Col–0 and wrky46–1 roots was analyzed. Four-week-old Col–0 and wrky46–1 plants grown in a hydroponic environment (pH 4.5) were exposed to 50 μm AlCl3 for a period of time. There was only a minor difference of the total Al content in the roots between Col–0 and wrky46–1 after 24 h of Al treatment, but wrky46–1 accumulated significantly less Al in the roots than Col–0 after 72 h of treatment (Figure 8), indicating the possible operation of an Al exclusion mechanism in the mutant.

WRKY46 represses ALMT1 expression and malate secretion

As the Al-induced secretion of malate contributes to Al exclusion in Arabidopsis (Liu et al., 2009), we performed a time-course expression analysis of ALMT1, a gene encoding the malate transporter, in Col–0 and wrky46–1 plants under Al treatment. The qPCR revealed a higher transcriptional level of ALMT1 in wrky46–1 than in Col–0 (Figure 3a), implying that WRKY46 represses ALMT1 expression. Similar results were obtained in wrky46–2 (Figure S4). Besides, ALMT1 expression was further detected in the complementation lines, with expression levels similar to that in wild-type Col–0 (Figure S5), thereby confirming the regulatory role of WRKY46. We also tested whether the loss of WRKY46 would affect the expression of other Al resistance genes, including previously reported STOP1 (Iuchi et al., 2007), MATE (Liu et al., 2009), ALS1 (Larsen et al., 2007), ALS3 (Larsen et al., 2005) and STAR1 (Huang et al., 2010); however, none of them had significantly altered transcripts between Col–0 and wrky46–1 mutant under normal and Al treatment (Figure 3b), indicating that WRKY46 affects Al resistance probably mainly through regulating ALMT1 expression.

Figure 3.

WRKY46 regulates ALMT1 expression. Expression of ALMT1 (a) and other aluminium (Al) resistance genes (b) in Col–0 and wrky46–1 after different time periods of Al treatment. RNA was extracted from 4–week-old Col–0 roots in hydroponic culture (pH 4.5) under control or with 50 μm AlCl3 for the indicated time periods. ACT2 was used as the internal reference gene. Three independent repeats were performed with similar results. The data shown are from one experiment. Error bars indicate SEs (n = 3).

As the expression of ALMT1 is altered in the mutant, we further determined the malate secretion of Col–0 and wrky46–1 plants under normal and Al treatment. The two genotypes were hydroponically cultured for 4 weeks (pH 4.5), and subsequently transferred to 0.5 mm CaCl2 solution (pH 4.5) for 12 h pre-treatment, followed by the same concentration of CaCl2 with or without 50 μm AlCl3. As shown in Figure 4, wrky46–1 roots released more malate than Col–0, either in normal conditions or under Al treatment, which is consistent with the relatively higher ALMT1 expression level and Al resistance in the mutant.

Figure 4.

Mutation of WRKY46 increases malate secretion. Malate secretion was detected in Col–0 and wrky46–1 roots. Four-week-old plants were exposed to 0.5 mm CaCl2 solution (pH 4.5) with or without 50 μm AlCl3; FW, fresh weight.

Three independent repeats were performed, with similar results. The data shown are from one experiment. Error bars indicate SEs (n = 5).

WRKY46 binds to ALMT1 promoter in vitro and in vivo

The WRKY proteins can recognize the W–boxes containing the TGAC core sequence present in the promoters of target genes. As there are 10 W–boxes with TGACT/C sequences in the ALMT1 promoter (1.7 kb), and six of them are enriched within a 500–bp region (Figure 5a), we speculated that WRKY46 may directly bind to these W–boxes. To verify this hypothesis, we firstly performed the transient transformation of Nicotiana benthamiana (tobacco) leaves by agro-infiltration, which has been used to analyze the plant promoters and transcription factors (Yang et al., 2000). We used the dual luciferases vector as a reporter system, following Hellens et al. (2005). The intact and 5′-deleted ALMT1 promoters were constructed into the reporter vector as P1 and P2 (Figure 5b), which were infiltrated alone or co-expressed with the CaMV35-driven WRKY46 (W46), respectively. It was clear that WRKY46 had the capacity of inhibiting reporter activity driven by the full-length ALMT1 promoter, but not the 5′-deleted one (Figure 5b), indicating that the deleted promoter region containing the certain enriched W–boxes is critical for the interaction between WRKY46 and the ALMT1 promoter.

Figure 5.

Transient assay for the interaction between WRKY46 and the ALMT1 promoter in Nicotiana benthamiana (tobacco).

(a) Characterization of the ALMT1 promoter. The full length of the promoter, from the translational start site (ATG), is shown. The elements containing TGACC/T were boxed as the potential W–boxes. The sequences in black containing the W–boxes were later used in the yeast one-hybrid experiment.

(b) Transient expression assay in N. benthamiana. Full-length and 5′–deleted ALMT1 promoters were constructed into the report vector, and WRKY46 was cloned into the effect vector, respectively; LUC, Firefly luciferase activity; REN, Renilla luciferase activity (used as the control). Data show ratios of LUC to REN. Three independent repeats were performed with similar results. The data shown are from one experiment. Error bars indicate SEs (n = 4).

To test whether WRKY46 specifically binds to the regions of the W–boxes in the ALMT1 promoter, three regions of about 50 bp in length surrounding the W–boxes, as shown in Figure 5(a), and the same regions with the mutant W–boxes were used as the baits for binding assays in the yeast one-hybrid system (Figure 6a) (Vidal and Legrain, 1999). The interactions between WRKY46 and these promoter fragments were tested by growth on media lacking Trp, Leu and His. Increasing concentrations of the His synthase inhibitor 3–aminotriazole (3AT) were added to the media to suppress background activation and assess the strength of the interaction (Figure 6b). It was evident that yeast co-transformed with WRKY46 and the natural promoter regions, but not with WRKY46 and the corresponding mutant fragments, grew well in the selective media (Figure 6b), indicating that WRKY46 was able to bind these regions in the ALMT1 promoter through the W–boxes.

Figure 6.

WRKY46 binds to ALMT1 promoter in yeast.

(a) The promoter fragments were indicated. Elements in black were potential W–boxes. Bases in red were indicated as the mutant sites.

(b) Yeast cells were co-transformed with a bait vector, containing a promoter fragment in (a) fused to a HIS2 reporter gene, and a prey vector, containing WRKY46 fused to a GAL4 activation domain. Cells were grown in liquid media to an OD600 of 0.1 (10−1) and diluted in a 10× dilution series (from 10−2 to 10−3). From each dilution 5 μL was spotted on media selecting for both plasmids (SD, –Trp, –Leu) and selecting for interaction (SD, –Trp, –Leu, –His), supplemented with 25 or 50 mm 3AT to suppress background growth and test the strength of the interaction.

We further performed chromatin immunoprecipitation (ChIP) analysis using seedlings of a transgenic line expressing GFP-Tagged WRKY46 to determine whether WRKY46 directly binds to the ALMT1 promoter in vivo. After the ChIP with anti-GFP antibody, the enrichment of specific ALMT1 promoter fragments in the immune precipitant was determined by qPCR using two primer pairs (AF1 and AF2; Figure 7a). The normal mouse lgG and a specific STOP1 promoter fragment were used as negative controls. We noticed that the AF1 region had a significantly stronger enrichment of WRKY46 than the negative control, which was not detected at all, whereas the AF2 region had very little enrichment of WRKY46 (Figure 7b). These results suggest that WRKY46 directly binds to the ALMT1 promoter, with the binding site mainly at the W–boxes enriched across a 500–bp region upstream of the transcription start site.

Figure 7.

WRKY46 binds to ALMT1 promoter in vivo.

(a) Diagram of ALMT1 promoter showing the TGACC/T core sequence present in different regions. AF1 and AF2 indicate genomic DNA fragments around the ALMT1 promoter for the ChIP assay.

(b) ChIP analysis showing the direct binding of WRKY46 to the ALMT1 promoter in vivo. Pro46-WRKY46-sGFP seedlings (at 7 days old) were used for ChIP analysis. The lgG was used as a negative control. A fragment of STOP1 promoter was also used as a negative control. Three independent repeats were performed with similar results. Error bars indicate SEs (n = 3).


The secretion of organic acid from roots is an essential strategy for external detoxification of Al for Al-resistant plants (Ma et al., 2001; Ryan et al., 2001; Kochian et al., 2004). There are two distinct patterns of secretion that have been well reviewed (Ma, 2000a; Ma et al., 2001). Pattern I is monitored in plant species like wheat (Delhaize et al., 1993) and Fagopyrum esculentum (buckwheat; Zheng et al., 2005), where the secretion of organic acid anion occurs immediately after exposure to Al; in pattern II, plant species such as Cassia tora (Ma et al., 1997), triticale (Ma et al., 2000b), sorghum (Magalhaes et al., 2007), maize (Maron et al., 2008) and ricebean (Yang et al., 2011) exhibit a significant lag phase of several hours (even 3–6 days in sorghum), prior to the substantial efflux of organic acid anions, indicating that de novo protein biosynthesis is required. The first gene TaALMT1 responsible for Al-activated malate secretion was identified nearly 10 years ago in wheat by Sasaki et al. (2004). Although TaALMT1 is constitutively expressed in the root apices, which results in pattern–I organic acid anion secretion (Sasaki et al., 2004), many other reported ALMT genes in Arabidopsis (Hoekenga et al., 2006), Brassica napus (oilseed rape; Ligaba et al., 2006) and rye (Collins et al., 2008) are Al inducible. Besides, most of the reported MATE genes responsible for citrate secretion are significantly induced by Al (Magalhaes et al., 2007; Liu et al., 2009; Ryan et al., 2009; Maron et al., 2010; Yokosho et al., 2010, 2011; Yang et al., 2011), except HvAACT1 and TaMATE1B, which are constitutively expressed in barley and wheat, respectively (Furukawa et al., 2007; Ryan et al., 2009; Tovkach et al., 2013). These results suggest that transcriptional regulation is critical for the expression of most of these genes, and is closely related to Al resistance. Although two genes, STOP1 and ART1, encoding homologous transcription factors that regulate a number of Al resistance genes, including ALMT1 and MATE, were recently identified in Arabidopsis and rice, respectively (Iuchi et al., 2007; Sawaki et al., 2009; Yamaji et al., 2009), these two genes are not induced by Al, indicating that there may be other Al-responsive factors involved in the regulation of Al resistance genes. In this study, we demonstrated that a WRKY transcription factor WRKY46 functions as a transcriptional repressor of ALMT1 in modulating Arabidopsis Al tolerance, supported by the following evidence: (i) WRKY46 was inhibited, whereas ALMT1 was induced by Al, so they are negatively co-related; (ii) the wrky46 mutant showed more Al resistance, higher ALMT1 expression and malate secretion in roots; (iii) WRKY46 and ALMT1 expression overlapped spatially; (iv) WRKY46 bound to the ALMT1 promoter for direct trans-repression.

The transcription factor WRKY46 could be induced by multiple abiotic stresses (Figure S1) and biotic stress, as previously reported (Hu et al., 2012), but was inhibited by Al treatment (Figure 1), suggesting a specific role in the Al response that differs from other stress responses. We found that WRKY46 expression was significantly inhibited by Al from 3 h (Figure 1a), whereas the expression of ALMT1 was obviously induced after 6 h (Figure 1a), indicating a visible precedence relationship between the two genes of expression alteration. Further phenotype analysis using T–DNA insertion mutants showed that the mutation of WRKY46 increased Al resistance (Figure 2). The qPCR analysis revealed a higher expression level of ALMT1 in wrky46–1 than in Col–0 (Figure 3a), and the corresponding malate secretion of wrky46–1 was more than that of Col–0 (Figure 4). This is consistent with the lower Al content in the roots of the wrky46–1 mutant (Figure 8) when the plants were subjected to Al treatment for 72 h, and subsequently leads the mutant to be more Al resistant. There was a minor difference in Al content of the whole root system in the first 24 h of Al treatment, and this might be because the exclusion of Al by malate secretion is only functional in the root apices, so that in short-term exposure to Al the root apices were much less affected by malate secretion, compared with the whole root system, thereby reducing the overall efficiency in expelling Al. But when the exposure time was extended the root sections affected by malate secretion increased too, thus at 72 h of exposure a significant difference could be detected (Figure 8). Therefore, it is clear that WRKY46 may negatively regulate the temporal expression of ALMT1, the secretion of malate and Al resistance.

Figure 8.

Aluminium (Al) accumulation analysis in Col–0 and wrky46–1 roots.

The total Al accumulation in the roots of Col–0 and wrky46–1 was determined via ICP-OES analysis. Mean ± SE values were determined from five samples for each. CaCl2 and nutrient solutions (pH 4.5) were used for less than 24 and 72 h of Al treatment, respectively. Three independent experiments were performed with similar results. The data shown are from one experiment. Error bars indicate SEs (n = 3).

So, is WRKY46 spatially co-expressed with ALMT1? Using transgenic GUS staining we observed that WRKY46 was preferentially expressed in the stele of the root, and the expression increased in the regions that are closer to the root apex (Figure 1b). Kobayashi et al. (2007) also found that the expression of ALMT1 was significantly induced in the elongation zone and the stele of mature root. This indicates that the two proteins have an overlapping localization, mainly in the central cylinder of the root. Because of the specific expression of ALMT1, malate was released mainly from Arabidopsis root apices. Such a restricted pattern of secretion can achieve the maximum protection for the root apex, and the metabolic cost to plants will be greatly reduced (Kinraide et al., 2005; Kobayashi et al., 2007). Then why are these two genes intensively expressed in the root stele? We think that the significantly elevated ALMT1 expression in the stele of mature roots may help the plant to translocate malate from mature roots to the root tips through the phloem, and subsequently enhance malate secretion. Diminishing the expression of WRKY46 under Al stress may further enhance the above scheme, and facilitate malate secretion.

The in vitro and in vivo binding assays demonstrated that WRKY46 directly binds to the ALMT1 promoter through the specific W–boxes (Figures 6 and 7), suggesting that WRKY46 is a direct regulator of ALMT1. Moreover, we found that the promoter region more than 700 bp upstream of the translational start site might be important for ALMT1 expression, as the deleted promoter exhibited significantly lower activity than the full-length one in the transient assay (Figure 5). Further truncated promoter analysis using transgenic methods will be helpful in isolating the exact promoter region that confers the high expression of ALMT1, especially during exposure to Al.

Recently, ALMT1 expression was reported to be induced through shoot infection by Pseudomonas syringae pv tomato (Pst DC3000) and pathogen-derived microbe-associated molecular patterns (MAMPs) (Rudrappa et al., 2008; Lakshmanan et al., 2012). The elevated ALMT1 expression resulted in increased malate secretion, and the subsequent recruitment of beneficial soil bacteria (Rudrappa et al., 2008). As WRKY46 was also responsive to Pst DC3000 (Hu et al., 2012), the induction of ALMT1 by foliar infection might not be through regulating WRKY46 expression in the root. Although we don't know whether WRKY46 is involved in the feed-forward inhibition of ALMT1 expression when the latter is over-induced by aerial infection, at least there must be another unknown mechanism involved in the high induction of ALMT1 expression upon foliar infection.

The WRKY transcription factors play roles of importance in a great number of processes in many plant species (Rushton et al., 2010), such as seed development and germination, senescence, biotic and abiotic stresses, but none have been reported in Al response. W–box, the binding site of WRKY proteins, is enriched in the ALMT1 promoter region, enabling WRKY46 to interact with it and subsequently regulate gene expression. Promoter sequence analysis revealed that W–boxes are also enriched in the promoters of many Al resistance genes (Figure S6). Some of them have more than six W–boxes in the 2–kb promoter region, implying possibly strong enrichment of WRKY proteins in these regions. Thus, it's interesting to identify more Al-responsive WRKYs, and help widen the transcriptional regulation of Al resistance genes, as well as the certain signaling pathways.

In conclusion, we found that the transcription factor WRKY46 is a negative regulator of ALMT1 expression, but it may not only be involved in the Al response, as ALMT1 expression was inhibited by multiple stresses (Figure S1). As malate is an important metabolic intermediate involved in the tricarboxylic acid cycle (TCA cycle), which is responsible for the oxidation of respiratory substrates to drive ATP synthesis (Sweetlove et al., 2010), plants should increase intermediates such as malate in their cells to acquire more energy, and cope with environmental stresses by inhibiting its release. Besides, malate is also required for cellular osmoregulation, such as is typically found in the guard cells (Shimazaki et al., 2007). Thus, the plant must have a tight control of the secretion of these small molecular organic acids, which play multiple roles in plant physiology.

Experimental procedures

Plant materials and growth conditions

Arabidopsis thaliana (Columbia ecotype, Col–0) was used for all of the control experiments. The T–DNA insertion mutants wrky46–1 (SAIL_1230_H01; Col–0 background) and wrky46–2 (SALK_134310C; Col–0) were obtained from the Arabidopsis Biological Resource Center (ABRC, http://abrc.osu.edu). The mutant was identified by amplifying the flanking regions and sequencing them. Plants were grown in an environmental controlled growth chamber programmed for a 16–h light/8–h dark cycle, with a daytime temperature of 24°C and a night temperature of 21°C. For hydroponic culture, the nutrient solution was as described previously (Zhu et al., 2012). For root growth assay, solid medium with 0.8% (wt/vol) agar was as described by Hoekenga et al. (2006), with the pH modified to 4.5. Surface-sterilized and stratified seeds were sown in the medium in single files parallel with the outer edge of the magenta box. Seedlings were grown vertically, and photographed after 10 days of growth.

For complementation, a 1650–bp length of native promoter coupled with WRKY46 cDNA sequence was cloned into a binary vector (pCAMBIA1300). WRKY46 sequence without terminator was fused to the GFP gene in the vector. For the transgenic GUS line, a 2400–bp length of WRKY46 promoter was cloned into the binary vector (pCAMBIA1301). The recombinant plasmid was transformed into Arabidopsis through the Agrobacterium GV3101, and homozygous lines were selected. The primers used are listed in Table S1.

Gene expression analysis and sequence analysis

For real-time qPCR, total RNAs from seedlings or roots of different genotypes were extracted using RNAprep pure Plant Kit (TianGen Biotech Co. Ltd., Qiagen, http://www.qiagen.com), according to the manufacturer's protocols. Total RNAs treated with DNase I (TianGen Biotech Co. Ltd., Qiagen) were converted into cDNAs using PrimeScript RT® reagent Kit (TaKaRa, now Clontech, http://www.clontech.com/takara). Real-time qPCR analysis was carried out using the SYBR® Premix Ex Taq™ II (TaKaRa) on a Roche LightCycler480 real-time qPCR system, following the manufacturer's instructions (Roche, http://www.roche.com). Transcript levels of each mRNA were determined and normalized with the level of ACT2 mRNAs using the ΔCt method (Czechowski et al., 2005; Schmittgen and Livak, 2008). Gene-specific primers are listed in Table S1.

The Expression Browser, Arabidopsis eFP Browser and Heat-Mapper tools provided by BAR (The BioArray Resource for Arabidopsis Functional Genomics; http://bbc.botany.utoronto.ca) were employed to display the heat map of the gene expression patterns.

GUS staining

GUS staining of 2–week-old transgenic lines, with or without Al treatment for 6 h, was performed by immersing seedlings in a staining solution [100 mm sodium-phosphate buffer, pH 7, 2 mm K4Fe(CN)6, 2 mm K3Fe(CN)6, 0.2% Triton X–100, 10 mm EDTA, 2 mm X–Gluc] in a 10–ml tube for 12 h at 37°C in the dark, followed by two washes with 70% ethanol to remove the chlorophyll. Samples were photographed using a stereoscope (Nikon, http://www.nikon.com) equipped with a charge-coupled device (CCD) camera.

For GUS activity quantification, roots from 2–week-old plants with or without Al treatment for 6 h were lysed for assays in an extraction buffer (50 mm NaH2PO4, pH 7.0, 10 mm EDTA, 0.1% Triton X–100, 0.1% sodium lauryl sarcosine, 10 mm β–mercaptoethanol) by freezing with liquid nitrogen and grinding with a mortar and pestle. The GUS activity assay was performed by using the substrate 4–methyl umbelliferyl glucuronide (MUG; Sigma-Aldrich) and the standards of MU (Sigma-Aldrich), as previously described (Jefferson et al., 1987).

Determination of root Al content

Four-week-old Col–0 and wrky46–1 plants grown in a hydroponic environment (pH 4.5) were transferred to the same fresh nutrient solution or 0.5 mm CaCl2 solution. Plants in nutrient solution were treated with 50 μm AlCl3 for 72 h. Plants in CaCl2 solution were treated with 50 μm AlCl3 for 6 h or 24 h. Root samples were further harvested, weighed and digested with HNO3 : HClO4 (4 : 1, V/V). Al concentrations in the extracts were determined by inductively coupled plasma-atomic emission spectrometry (IRIS/AP optical emission spectrometry).

Detection of root organic acid exudation

Four-week-old Col–0 and wrky46–1 plants cultured in nutrient solutions were transferred to 0.5 mm CaCl2 solution (pH 4.5) for 12 h pre-treatment, followed by the same concentration of CaCl2 with or without 50 μm AlCl3 for 24 h. Every six plants in 1.25 L CaCl2 solution made a repeat. Root exudate samples were passed through cationic and anionic chromatography columns (Econo-Pac columns; Bio-Rad, http://www.bio-rad.com) filled with 5 g of Amerlite IR-120B resin (H+ form; Muromachi Chemical, http://www.muro-chem.co.jp/en/index) or 1.5 g Dowex 1 × 8 resin (100–200 mesh, formate form), separately. The organic acid anions absorbed in the Dowex 1 × 8 resin were desorbed with 15 ml of 1 m HCl, and the eluate was concentrated to dryness using a rotary evaporator at 40°C. The residue was re-dissolved in 1 ml of Milli–Q water. Malate concentrations were then measured according to the enzymatic method described in Delhaize et al. (1993).

Transient expression in Nicotiana benthamiana leaves

For construction of effector vector, the full length of WRKY46 mRNA was amplified and cloned into the binary vector (pCambia1300) under the control of the 35S promoter. The intact (1600 bp) and 5′–deleted ALMT1 promoters were amplified and constructed into the reporter vector pGreenII0800-LUC (Hellens et al., 2005). The recombinant plasmids were transferred into the Agrobacterium GV3101 lines. It should be noted that the reporter vector must be co-transferred with a pSoup plasmid to ensure amplification in Agrobacterium cells (Hellens et al., 2005). The GV3101 lines were co-infiltrated into the N. benthamiana leaves as described previously (Yang et al., 2000). The Firefly and Renilla luciferase activities were measured using a Dual Luciferase assay kit (Promega, http://www.promega.com). The primers used are listed in Table S1.

Yeast one-hybrid experiment

WRKY46 coding regions were amplified and cloned into the pGADT7-rec2 prey vector (Clontech, http://www.clontech.com), creating a translational fusion between the GAL4 activation domain and the transcription factor. For construction of the pHIS2 vector, forward and reverse oligonucleotides (Figure 6a) were annealed, digested by EcoRI/SacI, and subcloned into EcoRI/SacI-linearized pHIS2 vector. Competent yeast cells were prepared according to the Clontech Yeast Protocols Handbook using the Y187 yeast strain. For yeast transformation, 50 μL of competent yeast cells were incubated with 100 ng of pHIS2 bait vector and 100 ng of pGADT7-Rec2 prey vector, 50 μg salmon sperm carrier DNA (Invitrogen, now Life Technologies, http://www.lifetechnologies.com) and 0.5 ml PEG/LiAc solution. Cells were transformed according to the manufacturer's instructions. Transformations were plated onto SD media (–Leu, –Trp) to select for co-transformed cells and incubated at 28°C for 4 days. Transformed yeast cells were subsequently grown in SD (–Leu, –Trp) liquid media to an OD600 of 0.1 and diluted in a 10× dilution series. From each dilution, 5 μl was spotted on SD (–Leu, –Trp) and on SD (–His, –Leu, –Trp) media plates supplemented with 25 or 50 mm 3AT (Sigma-Aldrich). The plates were then incubated for 3 days at 28°C.

ChIP assay

The EpiQuik Plant ChIP kit (Epigentek, http://www.epigentek.com) was used to perform the ChIP assays. In brief, 0.8–1.0 g of 8–day-old Pro46-WRKY46-sGFP seedlings was fixed with 20 mL of 1.0% formaldehyde by vacuuming for 10 min. The chromatin DNA was extracted and sheared to 200–1000–bp fragments by sonicating. Sheared DNA (100 μl) was immunoprecipitated with 3–5 μg anti-GFP (Abcam, http://www.abcam.com) for 90 min at 50–100 rpm at 25°C. In addition, 1 μl of normal mouse IgG (Epigentek) was used as a negative control. DNA fragments that specifically associated with WRKY46 were released, purified, and used as templates for real-time qPCR using specific primers (Table S1).

Accession numbers

Sequence data from this article can be found in the Arabidopsis Genome Initiative or GenBank/EMBL databases under the following accession numbers: WRKY46 (At2g46400), ALMT1 (At1g08 430), MATE (At1g51340), STOP1 (At1g34370), ALS1 (At5g39040), ALS3 (At2g37330), STAR1 (At1g67940).


We thank Prof. Roger P. Hellens (HortResearch, Mt Albert Research Centre, Auckland, New Zealand) for kindly providing pGreenII 0800-LUC and pSoup plasmids. This research was supported by the Program for Innovative Research Team in Universities (IRT1185), the Specialized Research Fund for the Doctoral Program of Higher Education (SRFDP) and the Fundamental Research Funds for the Central Universities.