Arabidopsis thaliana ACS8 plays a crucial role in the early biosynthesis of ethylene elicited by Cu²⁺ ions

Cu²⁺ ions are a key constituent of cytochrome c oxidase and are essential to bacteria and eukaryotes. In plants, Cu²⁺ ions are cofactors for many enzymes, such as plastocyanin, superoxide dismutase and polyphenol oxidase, and participate in photosynthetic electron transport, oxidative stress responses and oxidative phosphorylation. Thus, Cu²⁺ ions play roles in the vegetative growth, development and reproduction of plants (Marschner, 1995; Peñarrubia et al., 2010). A previous study reported that Cu²⁺ associated with the ethylene-binding domain is required for ETR1 (ethylene response 1) to bind the gaseous hormone ethylene. This suggests that Cu²⁺ ions also play a role in hormone perception (Rodríguez et al., 1999). Under copper deficiency, plants show a twisted or malformed phenotype in young leaves (Marschner, 1995). However, excess copper exhibits heavy metal toxicity and causes cell death in plants. Therefore, plants tightly balance the level of Cu²⁺, allowing sufficient metal co-factor to be delivered to target proteins while avoiding toxicity. Plants have evolved specific membrane transporters, such as the copper transporter (COPT) family and P-type ATPase-family copper transporters, to deliver copper intra- and intercellularly (Hirayama et al., 1999; Sancenón et al., 2003). Copper deficiency upregulates expression of the transcription factor SQUAMOSA promoter-binding protein-like 7 (SPL7) in Arabidopsis. SPL7 then targets the cis-element of the GTAC motif present in promoters of some members of the COPT gene family to increase the absorption of Cu²⁺ (Peñarrubia et al., 2010; Yamasaki et al., 2009). However, excess copper suppresses the expression of COPT genes to reduce the concentration of copper in plants (Sancenón et al., 2003; Yuan et al., 2011). In yeast, the CUP2 gene codes for a copper-binding transcription factor protein containing a copper fist. In elevated copper concentrations, CUP2 binds Cu²⁺ and activates transcription of the metallothionein genes CUP1 and CRS5 (Welch et al., 1989).

Cu²⁺ is also a key constituent of many bactericides and fungicides. Copper-based bactericides and fungicides are used extensively in agriculture. Bordeaux mixture, a mixture of copper sulfate (CuSO₄) and slaked lime, has been widely used in vineyards, fruit farms and gardens to protect plants from infestations of downy mildew, powdery mildew and other fungi since 1854. The antimicrobial mechanisms of Cu²⁺ were assumed to be associated with its heavy metal toxicity. Cu²⁺ can suppress microbial growth by denaturing nucleic acids, inhibiting protein activity and changing plasma membrane permeabilization (Borkow and Gabbay, 2004). Interestingly, fewer microbes are reported to be resistant to copper-based bactericides and fungicides, although some microbes are tolerant to copper. For example, Pseudomonas syringae pv. tomato (Pst) isolated from tomato fields is tolerant to copper, dependent on three copper-resistant operon genes (Cha and Cooksey, 1991). However, copper compounds are still effective in controlling plant diseases. Thus, we hypothesize that Cu²⁺ not only inhibits bacterial and fungal growth but also elicits defense responses in plants. In our previous study, we found that Cu²⁺ promoted reactive oxygen species (ROS) accumulation and callose deposition, induced the expression of pathogenesis-related (PR) genes, activated MAP kinase signaling and elicited defense against Pst DC3000 in Arabidopsis (Liu et al., 2015). In addition, the Cu²⁺-mediated defense response is dependent on salicylic acid (SA) and ethylene signaling pathways. Thus, we assumed that Cu²⁺ acts as an elicitor and triggers immune responses in plants.

In the past decades, many elicitors have been shown to induce the defense response and protect plants from pathogen infection. Among these elicitors, microbe-associated molecular patterns (MAMPs) are the best studied. A number of MAMPs, such as chitin, peptidoglycan, lipopolysaccharide, flagellin and the elongation factor thermo unstable (ER-Tu), have been identified as inducing the immune response in plants (Felix et al., 1993, 1999; Gust et al., 2007; Kunze et al., 2004; Meziane et al., 2005). In addition to MAMPs, there are also many chemical compounds that can induce the plant defense response, including thiamine, riboflavin and rutin (Ahn et al., 2005; Dong and Beer, 2000; Yang et al., 2016). Flagellin 22 (flg22), a 22-amino-acid sequence of the conserved N-terminal region of flagellin, is one of the most well-studied elicitors and can elicit a defense response after being recognized by FLS2, a leucine-rich repeat (LRR) receptor-like kinase (Gómez-Gómez and Boller, 2000). The flg22-induced responses include ion fluxes, oxidative burst, callose deposition, PR gene expression, ethylene production and seedling growth inhibition (Boller and Felix, 2009).

The phytohormone ethylene plays a pivotal role in many aspects of plant growth and development. In addition, it is also a major regulator of plant responses to biotic and abiotic stresses (Pieterse et al., 2012). Pathogen attack and environmental stimuli such as wounding, flooding, heavy metal stress or ozone exposure can induce ethylene production in plants (Keunen et al., 2016; Moeder et al., 2002; Nie et al., 2002; Rao et al., 2002). The ethylene synthesis pathway, which contains three steps from the precursor methionine, has been extensively researched in plants. In the first step, methionine is conversed to S-adenosylmethionine (SAM) by SAM synthase. Then, the conversion of SAM to 1-aminocyclopropane-1-carboxylate (ACC) is catalyzed by ACC synthase (ACS). Finally, ACC is oxidized to ethylene by ACC oxidase (ACO). The conversion of SAM to ACC by ACS is the rate-limiting step of the ethylene biosynthesis pathway. ACS genes play a pivotal role in ethylene biosynthesis and their regulation has been intensively studied. In Arabidopsis, the ACS gene family includes 12 members, and only 8 of them are involved in the biosynthesis of ethylene (Yamagami et al., 2003). Tsuchisaka and Theologis (2004) reported that exogenous indole-3-acetic acid enhances the expression of AtACS2, 4, 5, 6, 7, 8 and 11 in the root, and that wounding of hypocotyl tissue induced the expression of AtACS2, 4, 6, 7, 8 and 11. In Arabidopsis, AtACS2 and AtACS6 were strongly upregulated upon challenge with Pst DC3000 or Botrytis cinerea (Guan et al., 2015; Han et al., 2010).

In a previous report, we showed that Cu²⁺-elicited defense responses in Arabidopsis were dependent on SA and ethylene signaling pathways (Liu et al., 2015). However, whether and how Cu²⁺ activates the production of ethylene remains unclear. In this study, we report that Cu²⁺ rapidly induces the production of ethylene. In addition, Cu²⁺-induced ethylene production mainly relies on the AtACS8 gene. We demonstrate that the copper-response cis-element (CuRE) plays an important role in Cu²⁺-activated AtACS8 transcription. Together, our results suggest that AtACS8 is responsible for the production of ethylene and Cu²⁺-elicited defense responses.

In a previous study, we showed that Cu²⁺ could elicit the defense response in Arabidopsis, and that the Cu²⁺-activated defense required the SA and ethylene signaling pathways (Liu et al., 2015). Moreover, Arteca and Arteca (2007) reported that copper stress increased ethylene production in Arabidopsis. However, the molecular mechanism remains unknown. To further test the relationship between Cu²⁺ and ethylene production, we examined the production of ethylene in Arabidopsis plants treated with CuSO₄ or MgSO₄. We found that CuSO₄-treated plants produced more ethylene than MgSO₄-treated plants 24 h post treatment (hpt) (Fig. 1A). We also examined the transcription of the ethylene responsive factor 1 (ERF1) gene, which is a marker of the ethylene signaling pathway (Alonso and Stepanova, 2004), to confirm the activation of ethylene signaling by Cu²⁺. Compared with the MgSO₄-treated plants, the transcript level of ERF1 was significantly greater in CuSO₄-treated Arabidopsis (Fig. 1B). These results demonstrate that Cu²⁺ promotes ethylene production and activates ethylene signaling in Arabidopsis.

To gain a better understanding of the role of Cu²⁺ in the plant defense response, we generated the transcriptome profiles from seedlings exposed to CuSO₄ using RNA sequencing. There were six samples in total, with each treatment having two biological replicates. More than 0.14 billion reads were generated, with approximately 24 million reads from each sample (Table S3). In CuSO₄-treated Arabidopsis plants, 2206 and 1009 genes showed greater than twofold up- or down-regulation (P<0.05) at 2 hpt, whereas 2423 and 590 genes were up- or down-regulated at 24 hpt, respectively, compared with control plants (Fig. 2A; Table S4). Most of the Cu²⁺-regulated genes fell into very broad GO categories such as metabolic process, cellular process, localization and response to stimulus (Fig. S1). In addition to a large group of genes with unknown functions (516 genes), a considerable number of the genes upregulated at 2 hpt can be classified as being involved in signal perception (113 genes encoding receptor-like kinases and resistant proteins), signal transduction (121 genes), transcriptional regulation (79 genes) and ethylene signal (42 genes encoding ethylene biosynthesis proteins and ethylene responsive factors) (Table S5).

Among the rapidly (2 hpt) upregulated genes, 370 genes were also identified as being upregulated by flagellin (Zipfel et al., 2004) and EF-Tu (Zipfel et al., 2006) (Fig. 2B; Table S4), such as WRKY29 (At4g23550), WRKY33 (At2g38470), WRKY53 (At4g23810), ERF1 (At4g34410), ERF5 (At5g47230) and ERF6 (At4g17490). To confirm Cu²⁺-mediated upregulated expression in the RNA sequencing experiment, we performed quantitative real-time polymerase chain reaction (qRT-PCR) assays. The qRT-PCR results showed that the expression patterns of all the selected genes in the plants treated with CuSO₄ were similar to those in the RNA sequencing experiment (Fig. 2C; Fig. S2). GO-based functional profiling showed that 79.30% of the upregulated genes can be classified into metabolic process, cellular process and response to stimulus (Fig. S1E).

Among the genes that are rapidly induced more than twofold at the transcriptional level at 2 hpt, we detected the following genes: MTO3 (At3g17390), SAM1 (At1g02500), SAM2 (At4g01850), ACS2 (At1g01480), ACS6 (At4g11280), ACS7 (At4g26200), ACS8 (At4g37770), ACS11 (At4g08040), ACO1 (At2g19590), ACO4 (At1g05010) and ACO5 (At1g77330) (Fig. 3A). These genes have previously been shown to be involved in the biosynthesis of ethylene in Arabidopsis (Mao et al., 2015; Wang et al., 2002; Yang and Hoffman, 1984). We performed qRT-PCR using gene-specific primers to confirm the expression patterns of these genes, and found that the expression of all the selected genes in Arabidopsis treated with CuSO₄ was similar to that in the RNA sequencing experiment (Fig. 3B). These results suggest that Cu²⁺ induces the expression of genes in the ethylene biosynthesis pathway in Arabidopsis.

The genes involved in the ethylene biosynthesis pathway and ethylene signaling pathway were upregulated at 2 hpt (Figs 2C and 3), suggesting that Cu²⁺ activates ethylene within 2 h. To confirm this hypothesis, we examined the production rate of ethylene in Arabidopsis treated with CuSO₄ or MgSO₄ at different time points. We found that, at 0.25 hpt CuSO₄- and MgSO₄-treated plants showed similar ethylene production rates. However, CuSO₄-treated plants showed a significantly faster rate of ethylene production than MgSO₄-treated plants at 0.5 hpt. In addition, the rate of ethylene production peaked at 2 hpt (Fig. 4B). Compared with MgSO₄-treated plants, the level of ethylene in CuSO₄-treated plants was 109% higher at 0.5 hpt. After 1- and 2-h treatments of wild-type seedlings with CuSO₄, the level of ethylene was 168% and 131% higher than that in MgSO₄-treated seedlings, respectively (Fig. 4A). These results suggest that Cu²⁺ rapidly induces ethylene production in Arabidopsis.

To confirm that Cu²⁺ rapidly promotes ethylene production, we examined the transcription of ERF1 in Arabidopsis at 2 and 24 h after CuSO₄ treatment. At 2 hpt, ERF1 gene expression was significantly induced by Cu²⁺ (Fig. 2C), suggesting that Cu²⁺ activates ethylene signaling within 2 h.

The heavy metal cadmium was recently shown to induce the production of ethylene in Arabidopsis, mainly via increased transcription of AtACS2 and AtACS6 (Schellingen et al., 2014). The transcription levels of AtACS2 and AtACS6 were also elevated by CuSO₄ treatment (Fig. 3). To test the role of AtACS2 and AtACS6 in Cu²⁺-induced ethylene production, we studied the effects of Cu²⁺ on acs2-1/acs6-1 double knockout mutant plants. Compared with MgSO₄-treated plants, CuSO₄-treated wild-type and acs2-1/acs6-1 mutant plants exhibited significantly increased ethylene emission at 2 hpt. In addition, no significant difference was observed in the ethylene levels of wild-type and acs2-1/acs6-1 mutant plants. However, exposure to CuSO₄ for 24 h led to significantly lower ethylene levels in the mutants than in the wild-type plants (Fig. 5A). Consistent with the increased ethylene emission, the transcriptions of ERF1 and ERF5 could be induced with CuSO₄ treatment in acs2-1/acs6-1 mutant plants at 2 hpt (Fig. 5B). The induction expression level was attenuated in acs2-1/acs6-1 mutants compared with the wild-type plants (Fig. 5B). This result shows that the AtACS2 and AtACS6 genes are required for later (24 h) ethylene production, but not early (2 h) ethylene production.

Our RNA sequencing and qRT-PCR results suggest that Cu²⁺ rapidly induces expression of the AtACS8 gene (Fig. 3). To verify the role of the AtACS8 gene in Cu²⁺-induced ethylene production, ethylene emission was measured in wild-type and acs8 knockout mutant plants after MgSO₄ or CuSO₄ treatments. At 2 hpt, no significant differences in ethylene level were observed between acs8 knockout mutant plants treated with MgSO₄ and CuSO₄ (Fig. 5C). By contrast, 24-h exposure to CuSO₄ resulted in a significantly higher level of ethylene than MgSO₄ exposure in acs8 mutant plants (Fig. 5C). We also examined the transcription of ERF1 and ERF5 in acs8 mutant plants. As shown in Fig. 5D, CuSO₄ treatment completely lost the induction of ERF1 and ERF5 at 2 hpt. This result suggests that the rapid ethylene production induced by Cu²⁺ relies on the AtACS8 gene. To confirm this result, we examined ethylene production in acs1-1 acs2-1 acs6-1 acs4-1 acs5-2 acs9-1 acs7-1 acs11-1 octuple mutant plants after MgSO₄ or CuSO₄ treatments. The octuple mutant plants could produce more ethylene and expressed a higher level of the AtACS8 gene than wild-type plants. However, CuSO₄ treatment resulted in similar ethylene levels in the wild-type and octuple mutant plants (Fig. S3). The results suggest that AtACS8 is required for Cu²⁺-activated early ethylene signaling.

To determine whether the AtACS8 gene is required for the Cu²⁺-mediated defense response, we assessed the effects of Cu²⁺ on protection against Pst DC3000 in wild-type, acs2-1/acs6-1 double knockout mutant and acs8 knockout mutant plants. Compared with MgSO₄ treatment, CuSO₄ treatment resulted in a significant (P<0.01) decrease in the Pst DC3000 population in wild-type and acs2-1/acs6-1 double knockout mutant plants. Meanwhile, in acs8 mutant plants, there were no significant (P<0.05) differences in Pst DC3000 infection between acs8 knockout mutant plants treated with MgSO₄ and CuSO₄ (Fig. 5E). Taken together, our results strongly suggest that the AtACS8 gene plays an essential role in Cu²⁺-mediated early ethylene production and defense responses.

To identify the mechanisms underlying Cu²⁺-induced AtACS8 expression, we used the GFP reporter gene to analyze the AtACS8 promoter. The GFP open reading frame (ORF) was inserted downstream of the AtACS2, AtACS6 or AtACS8 promoter. The constructs were transiently expressed in Nicotiana benthamiana (NB) and analyzed for their responsiveness to CuSO₄ treatment. GFP under the AtACS6 promoter showed a strong fluorescence signal in leaves treated with MgSO₄ and CuSO₄. However, in contrast to MgSO₄ treatment, CuSO₄ treatment could activate GFP expression driven by the AtACS2 and AtACS8 promoters. In addition, the fluorescence in AtACS8-driven GFP was stronger than AtACS2-driven GFP (Fig. S4). The promoter of AtACS8 was chosen for further analysis. Site-directed deletions of the AtACS8 promoter at -1155, -901 and -661 were generated (Fig. 6A). Compared with MgSO₄ treatment, CuSO₄ treatment activated GFP driven by the -1655 and -1155 promoters. However, when driven by the -901 and -661 promoters, the GFP gene was not induced by CuSO₄ treatment (Fig. 6A and B). The above result suggests that the DNA fragment from -902 to -1155 is required for Cu²⁺-induced expression of the AtACS8 gene. We performed in silico scanning for Cu²⁺-induced elements using Jaspar (Mathelier et al., 2015) with a profile score threshold of 80%. The analysis revealed one CuRE located at -967 to -957 in the promoter region (Fig. S5), indicating the candidate cis-element controlling Cu²⁺-induced AtACS8 expression. Then, we used site-directed promoter deletions at -967 and -956 to drive the GFP reporter gene and analyzed these constructs for their responsiveness to Cu²⁺ (Fig. 6C). Cu²⁺ induced GFP expression when GFP was driven by the -967 promoter but not when GFP was driven by the -956 promoter, which only lacks the CuRE (Fig. 6C and D). Using in silico scanning, we found that the promoters of those ethylene biosynthesis genes induced by CuSO₄ treatment contain 3–9 (7.10±1.79) CuREs, which is significantly (P<0.05) more than that of 100 random genes (Fig. S6). We also constructed CuRE-35S mini and the 35S mini to drive the GFP reporter gene. In addition, Cu²⁺ induced a GFP signal in leaves transiently expressing CuRE-35S mini: GFP, but not in leaves expressing 35S mini: GFP (Fig. 6C,D). The function of CuRE in the Cu²⁺-induced expression of AtACS8 was further confirmed in transgenic Arabidopsis plants. The constructs containing CuREs showed strong GFP signal after CuSO₄ treatment, whereas constructs lacking CuREs showed weak GFP signal (Fig. 6E,F). The results suggest that CuREs are required for Cu²⁺ to regulate the expression of gene AtACS8.

Previously, we have demonstrated that Cu²⁺ elicited defense responses and protected plants against pathogens (Liu et al., 2015). CuSO₄ treatment promotes ROS accumulation and callose deposition, activates MAP kinase signaling and upregulates the expression of PR genes (Liu et al., 2015). All these response reactions elicited by Cu²⁺ ions imply that they act in a similar manner to other MAMP molecules, such as flg22 and elf18 (Zipfel, 2008). About 9.5% of the genes upregulated by Cu²⁺, flg22 and elf18, were associated with a response to stimulus (Fig. S1). In this study, we have demonstrated that Cu²⁺ also elicited early ethylene production and activated ethylene signaling in Arabidopsis. Ethylene emission is one of the early events in the Arabidopsis response to MAMPs such as flg22 and elf18 (Zipfel et al., 2006). CuSO₄ treatment rapidly induced the production of ethylene 30 min after treatment (Fig. 4). This is the second study to identify similarities between Cu²⁺ and MAMP molecules. Furthermore, 2 h after CuSO₄ treatment, WRKY22, WRKY29, WRKY33, WRKY53 and WRKY55 were upregulated (Fig. 2 and Table S4); the expression of these genes is also induced by MAMPs such as flg22, elf18, Atpep1 and chitin (Wan et al., 2008; Yamaguchi et al., 2010; Zipfel et al., 2004, 2006). These five WRKY genes are involved in plant responses to both bacterial and fungal pathogens (Asai et al., 2002; Deslandes et al., 2002; Zheng et al., 2006). This is especially true for WRKY22 and WRKY29, which are considered early defense markers of flg22 and function downstream of MPK3 and MPK6 (Asai et al., 2002). We conclude that Cu²⁺ acts as an elicitor, inducing a MAMP-triggered immunity (MTI)-like response in Arabidopsis. This can explain, at least partially, why Cu²⁺ has broad bactericide and fungicide effects.

Although Cu²⁺ is similar to flg22 and elicits a defense response in Arabidopsis, there are also many differences between the responses elicited by Cu²⁺ and flg22. Flg22 treatment rapidly induced the expression of FLS2, MEKK1, MKK4 and MPK3, which are upstream of WRKY22 and WRKY29 (Asai et al., 2002; Zipfel et al., 2004). These genes were demonstrated to be part of the MAP kinase signaling cascade in the immune response induced by flg22 (Asai et al., 2002). However, Cu²⁺ did not upregulate the expression of these receptors or MAP kinase genes (Table S4), suggesting that Arabidopsis perceives Cu²⁺ with different MAP kinase signaling pathways. The following genes are some of the MAP kinase genes that are rapidly induced at the transcriptional level by Cu²⁺ but not flg22: MAPKKK5 (At5g66850), MEKK3 (At4g08470) and MAPKKK19 (At5g67080). These genes are members of the MEKK family, and few studies have reported their functions in response to abiotic and biotic stresses. The functions of these three MEKK family members in the Cu²⁺-mediated defense response require further research.

The phytohormone ethylene plays roles in plant growth, development, and the response to biotic and abiotic stresses. Various stress stimuli, such as pathogen infection, exposure to a heavy metal, wounding and ozone, have been reported to increase ethylene production in plants (Boller, 1991; O'Donnell et al., 1996; Keunen et al., 2016; Rao et al., 2002). Exposure to 5, 10, 25 or 100 µM cadmium increased the level of ethylene in Arabidopsis after 24 and 72 h (Schellingen et al., 2014). Copper-induced increases in ethylene production were also found in Arabidopsis (Arteca and Arteca, 2007) and wheat (Groppa et al., 2003). However, in another study, exposing Arabidopsis to 5 µM cadmium for 16 days resulted in decreased ethylene production (Carrió-Seguí et al., 2015). In addition, no significant changes in ethylene level were detected for Arabidopsis seedlings grown in the presence of 25 or 50 µM copper for 9 days (Lequeux et al., 2010). These results suggest that heavy metal-mediated ethylene production is dependent on the concentration of metal, plant age and exposure time.

The amino acid methionine is the biological precursor of ethylene, and the conversion of methionine to SAM by SAM synthase is the first step in ethylene biosynthesis (Yang and Hoffman, 1984). The sam1/sam2 double mutants showed significantly lower ethylene levels than wild-type plants (Mao et al., 2015). In addition, SAM1 overexpression lines produced more SAM and ethylene than Col-0 (Mao et al., 2015). These data suggest that upregulation of SAM synthase genes increases the level of ethylene. However, very few studies have reported the biotic or abiotic stress-mediated regulation of the transcription of SAM synthase genes (Broekaert et al., 2006). Interestingly, CuSO₄ treatment induced the expression of SAM1 (At1G02500), SAM2 (At4G01850) and MTO3 (At3G17390) by more than twofold at 2 hpt in our study (Fig. 2; Table S4).

ACOs catalyze the last step in ethylene biosynthesis, and members of the ACO gene families are induced by different biotic and abiotic stress stimuli. For example, wounding, flooding, heavy metal stress and ozone exposure were shown to enhance ACO gene expression in potato (Nie et al., 2002), tomato (Moeder et al., 2002; Nakatsuka et al., 1998) and Arabidopsis (Herbette et al., 2006; Schellingen et al., 2014). However, in a microarray experiment, none of the ACO genes were induced by copper treatment, although enhanced production of ethylene was found (Weber et al., 2006). In our study, AtACO1, 2, 4 and 5 were induced by Cu²⁺ (Fig. 3), suggesting that these ACO genes are also regulated in Cu²⁺-induced ethylene production.

In addition to SAM synthase genes and ACO genes, we also found that ACS genes were upregulated by CuSO₄ treatment (Fig. 2; Table S4). The ACS-catalyzed conversion of SAM to ACC is the rate-limiting step in the ethylene biosynthesis pathway. The transcriptional regulation of ACS genes to promote ethylene production has been thoroughly studied. Using beta-glucuronidase and GFP as reporters, Tsuchisaka and Theologis (2004) examined the spatial and temporal expression patterns of ACS genes in Arabidopsis and showed that different abiotic stress stimuli result in different patterns of expression among the various ACS gene family members. Peng et al. (2005) found that hypoxia stress specifically induced the expression of AtACS2, AtACS6, AtACS7 and AtACS9. In a recent study, Schellingen et al. (2014) noticed that cadmium-induced ethylene production relied on AtACS2 and AtACS6. Meanwhile, a transgenic Arabidopsis line with enhanced glutathione content showed higher levels of ACC and ethylene than wild-type plants. In addition, glutathione induced the transcription of AtACS2 and AtACS6, which is dependent on WRKY33 (Datta et al., 2015). Guan et al., (2015) reported that acs2-1/acs6-1 double mutant plants produce less ethylene and were more susceptible to pathogens. In our study, RNA sequencing and qRT-PCR revealed that Cu²⁺ induced the expression of AtACS2, AtACS6, AtACS7, AtACS8 and AtACS11 but suppressed the expression of AtACS4 (Fig. 3). CuSO₄ treatment enhanced the fluorescence signal in leaves expressing GFP driven by AtACS2, AtACS6 or AtACS8 promoters (Fig. S4). However, the acs2-1/acs6-1 double mutant showed similar levels of ethylene to wild-type plants 2 h after MgSO₄ or CuSO₄ treatment (Fig. 5A). Consistent with this, Cu²⁺ ions were still able to protect acs2-1/acs6-1 mutants from Pst DC3000 (Fig. 5C). At the same time, the levels of ethylene and the immune response in acs8 mutant plants were significantly lower than those of wild-type plants after CuSO₄ treatment (Fig. 5B and C). Therefore, it is possible that AtACS8 is the main gene regulating Cu²⁺-induced early ethylene production and defense responses. Interestingly, the expression of AtACS8 is controlled by the circadian clock, which is tightly correlated to ethylene production (Thain et al., 2004; Vandenbussche et al., 2003). It was further concluded that the enhanced transcription level of AtACS8 is crucial for the production of ethylene in Arabidopsis.

The Cu²⁺-induced expression of AtACS8 is dependent on the copper-response cis-element, termed CuRE (Fig. 6). This cis-element was first found in the promoter of the CUP1 gene, which codes for a metallothionein protein that protects yeast against heavy metal toxicity. Cu²⁺ activates the expression of CUP1 through the CUP2 transcription factor protein (Welch et al., 1989). The CUP2 protein is a copper sensor that contains a copper fist, which has the function of binding Cu²⁺ and DNA (Buchman et al., 1989). The copper sensor in plants is still unknown, and no similar CUP2 protein had been found in plants. It is possible that Cu²⁺ binds to a novel transcription factor containing a copper fist domain to regulate expression of the AtACS8 gene. Completion of a screen to identify the CuRE-binding protein will answer this question, which needs further research.

The bacterial strains and plasmids used in this study are described in Table S1. The Escherichia coli strains were cultured on Luria-Bertani (LB) medium at 37°C. The Pseudomonas syringae pv. tomato strain DC3000 was cultured on King's B medium containing 50 μg ml⁻¹ rifampicin at 28°C. The Agrobacterium tumefaciens strain GV3101 was cultured on LB medium containing 50 μg ml⁻¹ rifampicin at 28°C.

For promoter deletion analysis, 1655-bp, 1155-bp, 901-bp, 967-bp, 956-bp and 661-bp promoter fragments located immediately upstream of the AtACS8 start codon were cloned upstream of the GFP gene in the pCXGFP-P binary vector (Chen et al., 2009). The CuRE-35S mini promoter and 35S mini promoter were also inserted into the pCXGFP-P plasmid. These fragments were amplified using primers as described in the supplementary material (Table S2). All plasmids were validated by sequencing.

Seeds of Arabidopsis (Arabidopsis thaliana ecotype Col-0), acs2-1/acs6-1 (CS16581) and acs8 (SALK_006628) were obtained from The Arabidopsis Information Resource (TAIR). Arabidopsis plants were grown in nutrient substrates in a growth chamber with 12-h days (at 23°C and 60–75% relative humidity) and 12-h nights (at 21°C and 60–75% relative humidity). According to previous reports, 100 μM CuSO₄ can trigger the defense response but is not toxic to Pst DC3000 (Liu et al., 2015). Four- to five-week-old plants were sprayed with 100 μM (or other concentrations) CuSO₄ or MgSO₄ supplemented with 0.05% (v/v) Tween 20 for 4 h and then pathogen inoculation (Liu et al., 2015), unless stated otherwise. Two-week-old plants were used for ethylene measurement after being sprayed with 100 μM CuSO₄ or MgSO₄ supplemented with 0.05% (v/v) Tween 20 for different time periods.

A bacterial growth assay was performed in Arabidopsis plants at 0 and 3 days after inoculation. The concentration of Pst DC3000 suspension in buffer containing 10 mM MgCl₂ was adjusted to 10⁶ CFU ml⁻¹ and infiltrated into 4- to 5-week-old Arabidopsis plants using a needleless syringe. Leaves were harvested, surface sterilized in 70% ethanol solution for 1 min and then rinsed in sterile distilled water for 1 min. Leaf disks were excised from the leaves of three independent plants with a 0.5 cm² cork borer and were ground in 1.5-ml microfuge tubes with 100 μl sterile distilled water. The samples were serially diluted (1:10) and plated on King's B medium containing 50 μg ml⁻¹ rifampicin at 28°C for colony counting. The experiment was repeated three times. The means were compared using a t-test. The standard error and t-test results were recorded.

Arabidopsis plants were treated with CuSO₄ or MgSO₄ and whole plants were collected to extract total RNA with TRIzol Reagent (Invitrogen, USA). Using a HiFi Script cDNA Synthesis Kit (CWBIO, China), 1 μg of total RNA was used to synthesize first-strand cDNA as described in the manufacturer's instructions. qRT-PCR was performed with a QuantStudio 6 Flex Real-Time PCR System (Life Technologies, USA) using UltraSYBR Mixture (CWBIO, China) as described in the manufacturer's instructions. The primers used in qRT-PCR are described in the supplementary material (Table S2). The expression levels of genes were analyzed using the 2^−⊿⊿Ct analysis method or normalized with AtActin2 (At3g18780).

Two biological replicates of RNA samples were collected from two-week-old wild-type Arabidopsis plants treated with 100 μM CuSO₄ for 0, 2 or 24 h. As previously reported (Ju et al., 2017), six libraries were constructed and sequencing performed with a BGISEQ-500 by the Beijing Genomic Institution (www.genomics.org.cn, BGI, Shenzhen, China). Clean tags were mapped to the reference genome and genes that are available in the Arabidopsis TAIR10.2 reference genome. For gene expression analysis, the matched reads were calculated and then normalized to RPKM (reads per kilobase of transcript per million mapped reads) using RESM software. The significance of differential gene expression was confirmed with the BGI bioinformatics service using the combination of the absolute value of log2-Ratio≥1 and P≤0.05 in this research.

Ethylene biosynthesis rates were measured by gas chromatography as described previously (Spollen et al., 2000). Briefly, 2-week-old seedlings were treated with CuSO₄ or MgSO₄, and leaves from 20 plants were detached and placed in a 20-ml penicillin bottle. At various times after CuSO₄ or MgSO₄ treatment, a 1-ml gas sample was injected into a cold trap. Ethylene was quantified using a gas chromatograph equipped with a photoionization detector.

Agrobacterium GV3101 containing the pCXGFP-P binary vector with the GFP gene driven by different promoter fragments were resuspended with Agro-infiltration buffer (10 mM MgCl₂, 10 mM MES, pH 5.6, 200 mM acetosyringone), and the OD₆₀₀ adjusted to 0.5. Bacterial suspensions were infiltrated into 4- to 6-week-old N. benthamiana leaves with a needleless syringe. Two or three days later, whole plants were sprayed with 100 μM CuSO₄ or MgSO₄ supplemented with 0.05% (v/v) Tween 20 for 12 h. Plants were then observed using an M165 FC fluorescence stereomicroscope (Leica, Germany).

Stable transgenic Arabidopsis plants expressing the GFP gene driven by different promoter fragments were generated by floral dipping of Col-0 plants. Multiple positive T₁ lines were selected for production of the T₂ generation for experiments.The transgenic plants were sprayed with 100 μM CuSO₄ or MgSO₄ supplemented with 0.05% (v/v) Tween 20 for 12 h and then observed using a LSM880NLO confocal microscope (Zeiss, Germany).

We are grateful to TAIR for providing seeds of Arabidopsis mutants.