EIN2 regulates salt stress response and interacts with a MA3 domain-containing protein ECIP1 in Arabidopsis


S.-Y. Chen. Fax: +8610 64873428; e-mail: sychen@genetics.ac.cn; J.-S. Zhang. E-mail: jszhang@genetics.ac.cn


Ethylene signalling regulates plant growth and development. However, its roles in salt stress response are less known. Here we studied functions of EIN2, a central membrane protein of ethylene signalling, and its interacting protein ECIP1 in salt stress responses. Mutation of EIN2 led to extreme salt sensitivity as revealed by phenotypic and physiological changes, and overexpression of C-terminus of EIN2 suppressed salt sensitivity in ein2-5, indicating that EIN2 is required for salt tolerance. Downstream components EIN3 and EIL1 are also essential for salt tolerance because ein3-1eil1-1 double mutant showed extreme salt-sensitive phenotype. A MA3 domain-containing protein ECIP1 was further identified to interact with EIN2 in yeast two-hybrid assay and GST pull-down assay. Loss-of-function of ECIP1 resulted in enhanced ethylene response but altered salt response during seed germination and plant growth. Double mutant analysis revealed that ein2-1 was epistatic to ecip1, and ecip1 mutation partially suppressed ethylene-insensitivity of etr2-1 and ein4-1. These studies strengthen that interactions between ECIP1 and EIN2 or ethylene receptors regulate ethylene response and stress response.


The plant hormone ethylene regulates many aspects of plant growth, development and stress response. The linear signal pathway of ethylene has been established based on analysis of Arabidopsis ethylene-response mutants (Guzman & Ecker 1990; Guo & Ecker 2004). In Arabidopsis, ethylene is perceived by five receptors (ETR1, ETR2, ERS1, ERS2, EIN4) belonging to two subfamilies (Chang et al. 1993; Hua et al. 1995, 1998; Sakai et al. 1998; Bleecker 1999; Kendrick & Chang 2008). Subfamily I has ETR1 and ERS1, whereas subfamily II contains ETR2, ERS2 and EIN4. Ethylene receptors from Arabidopsis, tobacco and rice have His kinase activity and/or Ser/Thr kinase activity (Gamble, Coonfield & Schaller 1998; Xie et al. 2003; Moussatche & Klee 2004; Zhang et al. 2004; Chen et al. 2009; Wuriyanghan et al. 2009). These kinase activities have differential roles in regulation of plant response (Wang et al. 2003; Binder et al. 2004; Qu & Schaller 2004; Chen et al. 2009). The ethylene receptors negatively regulate ethylene responses (Hua & Meyerowitz 1998). RTE1 specifically regulates the function of ETR1 (Resnick et al. 2006; Zhou et al. 2007; Dong et al. 2008; Resnick, Rivarola & Chang 2008). Ethylene receptors are also regulated by protein degradation (Chen et al. 2007; Kevany et al. 2007). The downstream components of ethylene signalling pathway include CTR1, EIN2, EIN3/EIL and ERF1 transcription factors. CTR1 is a kinase and a negative regulator of ethylene signalling (Kieber et al. 1993; Huang et al. 2003). EIN2 positively regulates the levels of EIN3/EIL transcription factors, which results in the activation of transcription of ERF1 and other ethylene responsive genes (Chao et al. 1997; Solano et al. 1998; Alonso et al. 1999; Guo & Ecker 2003; Qiao et al. 2009). F-box proteins EBF1 and EBF2 target EIN3 and EIL1 for degradation via a ubiquitin/proteasome pathway (Guo & Ecker 2003; Potuschak et al. 2003; Gagne et al. 2004; Binder et al. 2007; An et al. 2010). Another MKK9-MPK3/MPK6 cascade was also found to regulate the stabilization of EIN3 through two MAPK phosphorylation sites or to regulate ethylene biosynthesis (Xu et al. 2008; Yoo et al. 2008). All these studies suggest that a complex network regulates the ethylene responses at several levels.

The EIN2 is a central membrane protein with 12 predicated transmembrane helices and has low similarity to the Nramp family of proteins (Alonso et al. 1999). No metal-transporting ability of EIN2 was observed in tested systems (Alonso et al. 1999). The EIN2-like proteins have been identified in other plants such as Medicago, petunia and rice, and have roles both in ethylene signalling and in other developmental processes or defence responses (Jun et al. 2004; Shibuya et al. 2004; Penmetsa et al. 2008). The C-terminal end of EIN2 (CEND) has the ability to constitutively activate the ethylene responses in Arabidopsis and in Medicago (Alonso et al. 1999; Penmetsa et al. 2008). Recent studies have shown that EIN2 is localized at endoplasmic reticulum (ER) membrane and interacts with ethylene receptors in tobacco leaf cells (Bisson et al. 2009; Bisson & Groth 2010). Additionally, two F-box proteins ETP1 and ETP2 have been found to interact with EIN2 and play an important role in ethylene signalling by regulating EIN2 degradation (Qiao et al. 2009).

Ethylene is regarded as a stress hormone and regulates abiotic stress responses. Our previous studies have demonstrated that type II ethylene receptor gene NTHK1 and NTHK2 from tobacco are induced by different stresses (Zhang et al. 1999, 2001a,b). NTHK1 has serine/threonine kinase activity whereas NTHK2 has both serine/threonine and histidine kinase activities in the presence of different ions (Xie et al. 2003; Zhang et al. 2004). Overexpression of the NTHK1 confers salt stress sensitivity in transgenic plants; however, presence of ethylene biosynthesis precursor 1-aminocyclopropane-1-carboxylic acid (ACC) suppresses the salt sensitivity (Cao et al. 2006, 2007). The kinase domain and kinase activity of the NTHK1 are differentially required for the salt stress response and growth response (Zhou et al. 2006; Chen et al. 2009). Arabidopsis ethylene receptor gain of function mutants exhibits similar sensitivity under salt stress (Cao et al. 2007). Ethylene signalling affects plant salt stress response likely through regulation of AtNAC2 and other relevant genes (He et al. 2005; Zhou et al. 2006; Cao et al. 2007; Chen et al. 2009). In addition, we find that mutation of EIN2 in ein2-1 mutant leads to extreme salt sensitivity, suggesting that EIN2 is important for plant salt tolerance (Cao et al. 2007). However, the molecular mechanism of EIN2 function in salt stress response is still unclear.

In this study, we further investigate the roles of EIN2 in plant responses to salt stress. We find that both the ein2-1 and ein2-5 mutants show salt sensitive response and transgenic plants haboring the C-terminal end (CEND) of EIN2 in ein2-5 mutant (Alonso et al. 1999) exhibit salt tolerant phenotype. EIN3/EIL transcription factor double mutant ein3-1eil1-1 also shows strong sensitivity to salt stress. We further identified a MA3 domain-containing protein (ECIP1) that interacts with EIN2 C-terminal end using yeast two-hybrid method. Mutation of the ECIP1 gene enhanced ethylene response and affected plant performance under salt stress. Genetic interaction between ECIP1 and EIN2, ETR2 or EIN4 was further analysed. These results have significance in understanding the function of both EIN2 and its interacting protein during plant responses to salt stress.


Plant growth and treatments

Arabidopsis seeds (Columbia ecotype, Col-0) were surface-sterilized, plated on ½ Murashige and Skoog (MS) medium, stratified at 4 °C for 3–4 d and then germinated at 23 °C under 16 h photoperiod.

For salt stress, 5-day-old seedlings from Col-0, ein2-1, ein2-5, ein3-1eil1-1 and transgenic line ein2-5:CEND were transferred onto MS containing various concentrations of NaCl. For each NaCl treatment, at least three replicates were performed. After about 10 d, photographs were taken, and relative rosette size and relative electrolyte leakage were measured. The ratio of average rosette size from salt-stressed plate to the average rosette size from MS plates was calculated as relative rosette size. The relative electrolyte leakage was calculated in the same way. The seedlings were also maintained on NaCl for 18 d and then transferred into pots for survival tests.

For triple-response test, the stratified seeds were exposed to light for 4–6 h and incubated with ethylene for 4.5 d at 23 °C in dark. The length of hypocotyls was measured. For ACC effect on cotyledon expansion, the seeds were sowed on 0.8% agar plus 50 µm of ACC without any other nutrients. After stratification, the plants were grown at 23 °C under continuous illumination. After about 5–6 d, the angle of cotyledon expansion was examined and categorized. For analysis of flowering time, 10-day-old seedlings were transferred into pots for normal growth. The ratio and height of bolting were measured.

Analysis of seed germination and other physiological parameters

Seeds (∼100) each from Col-0, ecip1 mutant alleles (ecip1-1and ecip1-2) or ein2-1 were sterilized and sowed on MS or MS containing 100 mm NaCl. Germination (defined by emergence of radicles) was measured with a 12 h interval for 60 h.

Five-day-old seedlings of Col-0, ein2-1, ein2-5 and ein2-5:CEND were transferred onto 1/2MS and 1/2MS containing 100 mm NaCl for 10 d under continuous illumination. Na+ and K+ content in aerial parts were measured according to previous description (Cao et al. 2006). The relative electrolyte leakage was measured according to Cao et al. (2007) and Ouyang et al. (2010).

Analysis of promoter-β-glucuronidase (GUS) staining

The promoters of ECIP1 (1502bp upstream of ORF) were amplified with specific primer pairs (Supporting Information Table S1). The products were used to replace the 35S promoter in pBin121 vector that contains GUS coding region. The construct was introduced into Agrobacterium tumefaciens strain GV3101 by electroporation and transferred into Arabidopsis Col-0 plants by floral dip method (Clough & Bent 1998). The harvested seeds were screened by Kanamycin resistance and for further analysis of GUS staining.

Identification of T-DNA insertion mutants

T-DNA insertion mutants of ecip1 (ecip1-1 is SALK_010461 and ecip1-2 is SALK_011896) were obtained from SALK database. The seed samples were sowed on 1/2MS medium. PCR screening for insertions was carried out using gene specific primer pairs (Supporting Information Table S1). The PCR products carrying plant T-DNA border part were further sequenced to determine the exact site of insertion. The primer for sequencing is LBb1.3. RT-PCR analysis of full length gene expression was used to evaluate the effect of insertion. The Actin2 was amplified as a control.

Subcellular localization analysis

The open reading frame (ORF) of ECIP1 was amplified by RT-PCR with the specific primers (Supporting Information Table S1). The PCR products were digested with BamHI and SalI, and fused to green fluorescence (GFP) in pBIN221. The fusion gene and the GFP control were driven by the 35S promoter. The two constructs were transformed into Arabidopsis protoplasts with polyethylene glycol (PEG) treatment or onion epidermal cells by particle bombardment, and GFP signal was detected by confocal fluorescence microscope.

Yeast two-hybrid interaction assay

Arabidopsis cDNA library was constructed according to the instruction of CytoTrap XR Library construction kit. The cDNAs were ligated into pMyr vector to express the target fusion proteins. The coding region for CD domain of EIN2 C-terminus was cloned into pSos vector resulting in fusion protein of hSOS and CD domain as the bait protein to screen the cDNA library. The primers for bait pSosEIN2CD construction and for interaction test are listed in Supporting Information Table S1. For confirmation, the pSos vectors containing different domains of EIN2 and the pMyr vector containing the C-terminal two MA3 domains of ECIP1, were cotransfected into S. cerevisiae temperature-sensitive mutant strain cdc25H, which contains a point mutation at amino acid position 1328 of the CDC25 gene. The cdc25 mutation in the cdc25H strain prevents growth at 37 °C, but allows normal growth at 25 °C. Growth of the transformants in Gal plate but not in Glu plate at 37 °C in SD(-UL) indicates the positive interactions. The combination of MAFB/MAFB was used as a positive interaction control. The combination of EIN2D or EIN2CD with LaminC served as negative interaction controls. Yeast transformation, growth conditions, and interaction assays were performed referring to the instruction of CytoTrap XR Library construction kit (Stratagene, La Jolla, CA, USA). The coding regions for ethylene receptors without transmembrane domains were also inserted into pSos vector with specific primers (Supporting Information Table S1) and tested for interactions with ECIP1 using procedures mentioned earlier.

GST pull-down analysis

In vitro transcribed and translated 35S-labelled CD domain of EIN2 C-terminus and C-terminal domains of ETR2 and EIN4 ethylene receptors without transmembrane domain were generated referring to the instruction of TNT-Coupled Wheat Germ Extract systems (Promega, Madison, WI, USA). The primers for PCR amplification of the gene fragments are listed in Supporting Information Table S1. The PCR products were digested with EcoRI and SalI, and cloned into pTNT vector. For ECIP1, the coding region for ECIP1 C-terminal two MA3 domains was amplified and cloned into pGEX6p-1 vector and then introduced into BL21 strain for GST-fusion protein expression. The GST pull-down analysis was performed as described (Cancel & Larsen 2002).

Statistical analysis

The data were analysed by one-way analysis of variance (anova) or t-test with the SPSS program (version 11.5, SPSS Inc., Chicago, IL, USA).


Analysis of ein2 mutants in response to salt stress

Our previous analysis finds that ethylene receptor signalling regulates salt stress responses, and EIN2 of ethylene signalling may also be involved in this process (Cao et al. 2006, 2007; Zhou et al. 2006; Chen et al. 2009). Roles of EIN2 in salt response were studied through mutant analysis. The ein2-1 harbours a stop codon after the sequence encoding transmembrane domains, and ein2-5 has a frame-shift near 5′-terminus of EIN2 coding for transmembrane region (Alonso et al. 1999). In MS plate, mutants and Col-0 plants showed normal growth (Fig. 1a). However, in plates containing NaCl, both ein2-1 and ein2-5 mutants exhibited a severe salt-sensitive phenotype as can be seen from epinastic backward growth of leaf blade and petiole, and small rosette size (Fig. 1a, b). These results indicate that mutation of EIN2 results in salt-sensitive response in plants.

Figure 1.

Phenotypic and physiological changes of ein2 mutants in response to salt stress. (a) Phenotypes of Col-0, ein2-1, ein2-5 and ein2-5:CEND line under salt stress. The ein2-5:CEND line overexpressing the C-terminal end of EIN2 (CEND) in ein2-5 (Alonso et al. 1999) has a salt-tolerant phenotype. (b) Relative rosette size in different plants. Values represent means ± standard deviation (SD; n = 3). (c) Relative electrolyte leakage in various plants. Values represent means ± SD (n = 3). (d) Na+ and K+ contents in Col-0, ein2-5 and ein2-5:CEND. Each data point represents average from three samples and bars indicate SD. The asterisks indicate significant difference between Col-0 and ein2-5 or ein2-1:CEND (P < 0.05 for * and P < 0.01 for **).

Relative electrolyte leakage is a parameter for evaluation of salt-induced damage. Both the ein2-1 and ein2-5 mutants showed substantially higher levels of relative electrolyte leakage compared with Col-0, in response to increasing concentrations of NaCl, suggesting that the two mutants are more sensitive to salt stress than Col-0 (Fig. 1c). Na+ content was significantly affected in ein2-5 compared with Col-0 under salt stress condition (Fig. 1d, left panel). K+ content was also significantly reduced in stressed and non-stressed ein2-5 (Fig. 1d, middle panel). Na+/K+ ratio appeared not affected in Col-0 and ein2-5 under salt stress (Fig. 1d, right panel). These results indicate that salt stress disrupted Na+ and K+ accumulation in ein2 mutant.

Overexpression of CEND in ein2-5 constitutively activates ethylene response (Alonso et al. 1999). We then studied whether salt-sensitive response in ein2-5 can be altered through CEND introduction. Transgenic plants (ein2-5:CEND) harbouring the CEND in ein2-5 (Alonso et al. 1999) were analysed. After salt stress, ein2-5:CEND plants did not show the salt sensitive phenotype as observed in ein2-5 or ein2-1 (Fig. 1a, b). Relative electrolyte leakage in salt-stressed ein2-5:CEND plants were also similar to those in salt-stressed Col-0 (Fig. 1c). Na+ and K+ contents in ein2-5:CEND plants were partially or fully recovered to the level of Col-0 (Fig. 1d). Additionally, the Na+/K+ ratio was reduced in salt-stressed ein2-5:CEND plants compared with that in Col-0, implying a potential for better performance under salt stress. These results indicate that overexpression of CEND suppresses salt sensitivity of ein2 mutant, suggesting that EIN2 is beneficial for salt tolerance.

Double mutant ein3-1 eil1-1 is sensitive to salt stress

EIN3 transcription factor family acts downstream of EIN2 in ethylene signalling transduction pathway (Chao et al. 1997; Solano et al. 1998; Guo & Ecker 2003). Our previous study found that mutation of EIN3 in ein3-1 mutant led to no or very weak phenotypic change but high electrolyte leakage after salt stress compared with salt-stressed Col-0, suggesting that EIN3 is only partially involved in salt stress response (Cao et al. 2007). Considering the possible functional redundancy of the EIN3 family members, other EIN3-like genes may also be involved in the salt stress response. We tested the performance of double mutant ein3-1eil1-1 (Alonso et al. 2003) under NaCl treatment. Similar to ein2-1, the ein3-1eil1-1 showed a severe salt-sensitive phenotype compared with Col-0 (Fig. 2a, b). Relative electrolyte leakage was also higher in ein3-1eil-1 than that in wild type Col-0 when plants were treated with NaCl (Fig. 2c). These results indicate that both EIN3 and EIL1 contribute to salt tolerance.

Figure 2.

Phenotype and gene expression in Col-0, ein2-1 and ein3-1eil1-1 under salt stress. (a) Phenotypic comparison in salt-stressed plants. (b) Relative rosette size in salt-stressed plants. (c) Relative electrolyte leakage in plants after salt stress. For (b) and (c), each data point is average of three replicates and bars indicate standard deviations.

MA3 domain-containing protein ECIP1 interacts with the CEND

Although EIN2 regulated salt stress responses (Figs 1 & 2), the relevant mechanism was not known. We used a CytoTrap yeast two-hybrid system for screening proteins that interact with EIN2. The C-terminal domain of EIN2 without transmembrane regions was adopted as bait (Fig. 3a). However, because full-length C-terminal domain (CEND or ABCD) had false positive response, we chose CD domain of EIN2-CEND as a bait protein (Fig. 3a, c). A series of prey proteins, e.g. SAUR, etc., were identified (Supporting Information Table S2), and among these, a novel MA3 domain-containing protein, designated as ECIP1 (EIN2 C-terminus Interacting Protein 1, At4g24800), was investigated further.

Figure 3.

The MA3 domain-containing protein ECIP1 interacts with C-terminus of EIN2. (a) Schematic representation of EIN2 structure. The N-terminal black boxes indicate putative transmembrane regions. The C-terminal end (CEND) was divided into A, B, C and D subdomains. The numbers indicate amino acid positions. (b) Predicted structure of EIN2-interacting protein ECIP1 and phylogenetic analysis. The phylogenetic analysis was performed with the Clustal W program using DNASTAR software (Madison, WI, USA). Os08g0120500, RCOM_1507670, POPTRDRAFT_570193, LOC100243296 and HsPDCD4 are from rice, Ricinus communis, Populus trichocarpa, Vitis vinifera and Homo sapiens, respectively. ECIP1, At5g63190 and At3g48390 are from Arabidopsis. (c) The D and CD domains of EIN2 CEND interact with the ECIP1 in yeast two-hybrid assay. Growth of transformants in Gal plate but not in Glu plate at 37 °C indicates positive interactions. (d) EIN2 interacts with ECIP1 in GST pull-down assay. Equal amount of GST-ECIP1 or GST was incubated with 35S-labelled EIN2CD for GST pull-down assay. The 35S-labelled EIN2CD was also loaded as a control (input). Arrow indicates position of EIN2CD. (e) Subfamily II ethylene receptors ETR2 and EIN4 interact with ECIP1 in yeast two-hybrid assay. Five ethylene receptors were tested for interactions with ECIP1. For (c) and (e), combination of MAFB/MAFB was used as positive interaction control, and combinations of EIN2D or EIN2CD with LaminC were served as negative interaction controls.

ECIP1 has 702 amino acids with four tandem MA3 domains (Fig. 3b) based on SMART analysis (Letunic, Doerks & Bork 2009). MA3 domain is a protein-protein interaction module (Shibahara et al. 1995; Yang et al. 2004). In Arabidopsis, two other homologous proteins (At3g48390 and At5g63190) were identified and ECIP1 showed identity of 66.4 and 70.2% to them, respectively (Fig. 3b). Homologous proteins are present in other plants including rice (Os08g0120500), Ricinus communis (RCOM_1507670), Populus trichocarpa (POPTRDRAFT_570193) and Vitis vinifera (LOC100243296). Identities between ECIP1 and the above four proteins were 67.2, 75.4, 75.4 and 76.3%, respectively. Cluster analysis revealed that ECIP1 was more closely related to proteins from Ricinus, Populus and Vitis (Fig. 3b). Interestingly, ECIP1 also shared 29% identity to HsPCDC4 (Homo sapiens), which contain two MA3 domains. HsPCDC4 functions as a tumour suppressor and affects multiple signal transduction pathways leading to carcinogenesis (Lankat-Buttgereit & Goke 2009). These results indicate that ECIP1-like MA3 domain-containing proteins are widely existed in plants and/or in animals, suggesting conserved roles in signal transduction.

A series of truncated mutants of EIN2 C-terminal region were generated to identify minimum domain required for association with ECIP1 by yeast two-hybrid system. The C-terminal two MA3 domains (amino acids 434–702) of ECIP1 were used in the assay. Only cells harbouring pSosEIN2D (amino acids 1099–1294) or pSosEIN2CD (amino acids 889–1294) plus pMyrECIP1 grew well under inductive (Gal) condition but not under non-inductive (Glu) condition on SD-UL at 37 °C (Fig. 3c). Cells harbouring the other plasmid combinations did not show this growth pattern. These results indicate that only EIN2-D (amino acids 1099–1294) and EIN2-CD (amino acids 889–1294) interact with ECIP1. These results also reveal that D domain containing the 196 amino acids in the CEND is enough for its interaction with ECIP1.

To further examine the interaction between EIN2 and ECIP1, an in vitro GST pull-down analysis was performed. Equal amount of GST and GST-ECIP1 fusion proteins, which were produced and purified from Bl-21 strain, was incubated with in vitro-translated 35S-Met-labelled EIN2-CD (amino acid 889–1294). As shown in Fig. 3d, a clear signal was observed in GST-ECIP1 interacted proteins. In contrast, no signal was detected in GST interacted protein lane. These results indicate that an interaction is present between ECIP1 and EIN2CD.

ECIP1 interacts with subfamily II ethylene receptor ETR2 and EIN4

EIN2 has been found to interact with all ethylene receptors (Bisson et al. 2009; Bisson & Groth 2010). As EIN2 interacts with ECIP1 (Fig. 3c, d), we tested if ECIP1 has any interaction with ethylene receptors using yeast two-hybrid assay. Both subfamily I receptor (ETR1 and ERS1) and subfamily II receptor (ETR2, EIN4, ERS2) genes were constructed in pSos vector. Transfected cells harbouring the pSosETR2 plus pMyrECIP1 or pSosEIN4 plus pMyrECIP1 grew well under inductive condition (37 °C, Gal) but not under non-inductive condition (37 °C, Glu) (Fig. 3e). On the contrary, cells containing other plasmid combinations did not grow under inductive condition. The pSosMAFB plus pMyrMAFB was used as a positive interaction control, and pSosEIN2D plus pMyrLaminC or pSosEIN2CD plus pMyrLaminC were used as negative interaction controls (Fig. 3e, right panel). GST pull-down assay was also performed by incubating GST-ECIP1 with in vitro-translated 35S-Met-labelled C-terminal domains of ETR2 or EIN4, and ECIP1 showed weak interactions with the C-terminal domains of ETR2 and EIN4 (Supporting Information Fig. S1). These results indicate that ECIP1 interacts with ETR2 and EIN4 but not other ethylene receptors.

ECIP1 gene expression and protein localization

Organ-specific expression pattern of ECIP1 was investigated by promoter-GUS analysis and 1.5 kb promoter region of the ECIP1 gene was used to drive the GUS gene in pBIN121 vector. At seedling stage, relatively strong GUS activity was observed in the root tips, root vascular tissues and the junction region of hypocotyl and root (Fig. 4a, left panel). At flowering stage, the GUS activity was mainly detected at the cauline leaf tips, the connection of petiole and stem, and the young flower buds (Fig. 4a, middle and right panels). Relatively weak expression was found in sepals and pistil (Fig. 4a, right panels).

Figure 4.

ECIP1 expression and subcellular localization of ECIP1 protein. (a) ECIP1 expression in seedlings and flower organs revealed by promoter-GUS analysis. (b) Subcellular localization of ECIP1 revealed by GFP fusion in Arabidopsis protoplasts. Bar = 5 µm. (c) ECIP1 subcellular localization in onion epidermal cells. Two different sections of the same cells were scanned and merged. Plasmolysis was also conducted with 250 mm NaCl to observe change in ECIP1 localization.

Subcellular localization of ECIP1 was examined. ECIP1 was fused to GFP to form the fusion gene ECIP1-GFP driven by the 35S promoter in pBIN121. The fusion gene plasmid and GFP control plasmid was transformed into Arabidopsis mesophyll protoplasts and onion epidermal cells, and the green fluorescence was observed under a confocal microscope. In protoplasts, ECIP1-GFP fusion protein was mainly localized in cytoplasm, similar to the GFP control that was largely present in the cytoplasm (Fig. 4b). Examination of different scanning sections (Fig. 4c) and plasmolysis with 250 mm NaCl treatment (Fig. 4c) in onion epidermal cells also suggest ECIP1 localization in cytoplasm. These results indicate that ECIP1 is mainly localized in cytoplasm of plant cells.

Germination of ecip1 mutant seeds is sensitive to salt stress

To elucidate ECIP1 functions in plants, two T-DNA insertion mutants were identified and designated as ecip1-1 (SALK_010461) and ecip1-2 (SALK_011896). The T-DNA insertion in ecip1-1 was located in the third exon at 2412 bp after ATG and another allele ecip1-2 had an insertion in the third exon at 1940 bp after ATG (Fig. 5a). No full-length transcript of ECIP1 was detected in the two mutants by RT-PCR (Fig. 5a), suggesting that ecip1-1 and ecip1-2 were loss-of-function mutants.

Figure 5.

The ecip1 seed germination is sensitive to salt stress. (a) Identification of ecip1 T-DNA insertion mutants. Left penal: insertion of T-DNA in ECIP1 genomic sequence. Solid black boxes represent exons and lines between boxes represent introns. Positions of T-DNA insertion in ecip1-1 and ecip1-2 are indicated by triangles. Right panel: full-length ECIP1 transcript levels in Col-0, ecip1-1 and ecip1-2 detected by RT-PCR. Actin2 was amplified as a control. (b) The ecip1 seed germination under salt stress. Various seeds were plated on Murashige and Skoog (MS) or MS containing 100 mm NaCl. Germination of around 100 seeds was scored for each plant line. Each data point is average of three replicates and bars indicate standard deviation.

Previous study has shown that mutation of EIN2 affected seed germination under salt stress (Wang et al. 2007). Presently, under MS condition, germination of ecip1-1, ecip1-2 and ein2-1 seeds was slightly delayed compared with Col-0 (Fig. 5b, left panel). However, under 100 mm NaCl, seed germination was significantly reduced for all the three mutants ecip1-1, ecip1-2 and ein2-1 compared with Col-0 at the time points examined (Fig. 5b, right panel). These results indicate that germination of ecip1 seed is sensitive to salt stress, similar to the case in ein2-1 mutant. It should be noted that ecip1-2 had more severe phenotype compared with ecip1-1 in seed germination although both mutants had no ECIP1 transcript detectable. This phenomenon should be further investigated.

Mutations of ECIP1 gene alters ethylene response

The ecip1 etiolated seedlings were also treated with ethylene to see if ethylene response was changed. Both ecip1-1 and ecip1-2 mutants exhibited shorter hypocotyls especially at higher concentrations of ethylene compared with Col-0, suggesting an enhanced response to ethylene (Fig. 6a, b). To study the genetic relationship between ecip1 and ein2-1, ecip1-1ein2-1 double mutant was generated. The ecip1-1ein2-1 double mutant had a hypocotyl length comparable with those of ein2-1 and untreated Col-0 at all the ethylene concentrations (Fig. 6a, b), indicating an insensitivity to ethylene treatment. These results suggest that EIN2 function is required for the enhanced ethylene response in ecip1-1 mutant.

Figure 6.

Ethylene responses in ecip1 and ecip1-1ein2-1 double mutant. (a) Phenotype of etiolated seedlings in response to ethylene. The bar indicates 1 cm. (b) Comparison of hypocotyl length in etiolated seedlings in response to ethylene. Each column represents average from 50 to 90 seedlings and bars indicate standard deviation (SD). The asterisks indicate significant difference between Col-0 and mutants at P < 0.01 for ‘**’ and at P < 0.05 for ‘*’. (c) Types of cotyledon expansion in response to ACC. Type I represents well-expanded cotyledons. Type II cotyledons have angles between 0 and 180 °. Type III cotyledons are completely closed. (d) Cotyledon expansions under normal growth condition or in response to 50 µm ACC. About 60 seeds for each line were used to determine the types of cotyledon expansion. Each column represents average from three replicates and bars indicate SD.

Another ethylene regulated-phenotype was further examined. When Col-0 seeds are germinated and grown in the light in agar medium without any other nutrients, the two cotyledons are fully expanded (type I) (Fig. 6c). When ethylene biosynthesis precursor ACC was applied, the two cotyledons were partially (type II) or completely (type III) closed (Fig. 6c). The ratio of the three type seedlings was evaluated. Under normal growth condition or in response to ACC treatment, both ecip1-1 and ecip1-2 showed enhanced ethylene response as can be seen from increased ratio of type II or type III seedlings (Fig. 6d). However, in the presence or absence of ACC, both ein2-1 and ecip1-1ein2-1 remained as type I seedlings (Fig. 6d). All these results indicate that ecip1-1 and ecip1-2 had enhanced ethylene response, and ein2-1 and ecip1-1ein2-1 double mutant are insensitive to ACC treatment.

The ecip1 mutant plants are tolerant to salt stress

We further tested if plant growth was affected by salt stress. Both ecip1-1 and ecip1-2 mutant seedlings showed a better growth compared with Col-0 on 1/2MS containing 150 mm NaCl (Fig. 7a). In contrast, the mutant ein2-1 and double mutant ecip1-1ein2-1 exhibited salt-sensitive phenotype at 120 mm and 150 mm NaCl. After 18-d salt treatment, the seedlings were transferred into pots and recovered for 16-d under normal growth condition (Fig. 7b). The survival rate of the two ecip1 mutant seedlings was higher than that of Col-0, whereas both ein2-1 and ecip1-1ein2-1 had low survival rates (Fig. 7b, c). These results indicate that growth of the ecip1 mutant plants is tolerant to salt stress and ein2-1 mutation in ecip1-1 mutant leads to salt sensitivity.

Figure 7.

Performance of Col-0 and various mutants under salt stress. (a) Phenotype of mutant plants grown on salt plates. (b) Recovery of the salt-stressed plants. The 150 mm NaCl-treated seedlings were transferred to soil for recovery for 16 d under normal condition (lower panel). Plants from ½ Murashige and Skoog (1/2MS) plates were transferred as controls (upper panel). (c) Survival rate of the plants after salt stress. Results are averages of three replicates and bars indicate standard deviation. Asterisks indicate significant difference between Col-0 and mutants at P < 0.05.

Disruption of ECIP1 reduces cotyledon size

Alteration of ethylene signalling affects rosette size in Arabidopsis (Guzman & Ecker 1990; Alonso et al. 1999; Cao et al. 2007; Chen et al. 2009). The ein2-1 mutant exhibited the largest cotyledons compared with Col-0 and other tested mutants, whereas ctr1-1 had the smallest cotyledons (Fig. 8a, b). Both ecip1-1 and ecip1-2 showed smaller cotyledons than Col-0. However, cotyledons of ecip1-1ein2-1 were slightly larger than that of Col-0 but significantly smaller than that of ein2-1. These results indicate that disruption of ECIP1 reduces cotyledon size, and an additive effect may be present between ecip1 and ein2-1 in the control of cotyledon development.

Figure 8.

The ecip1 mutants have small cotyledons. (a) Comparison of cotyledon size among various plants. (b) Cotyledon expansion width from various seedlings. The results are averages from 10 to 20 seedlings with standard deviation (SD). Different letters above each column indicate significant difference between the compared seedlings (P < 0.05, Duncan test). (c) Flowering time variation among the compared plants. (d) Bolting rate of plants grown in (c). The results are average from three replicates with SD. Asterisks indicate significant difference between Col-0 and mutants at P < 0.05.

Ethylene signalling also affects flowering in Arabidopsis and rice (Guzman & Ecker 1990; Ogawara et al. 2003; Achard et al. 2007; Cao et al. 2007; Wuriyanghan et al. 2009). Early flowering was observed in both ecip1 mutants compared with Col-0 (Fig. 8c, d). The result indicates that disruption of ECIP1 promotes early flowering. The double mutant ecip1ein2-1 had a bolting rate similar to Col-0, but much lower than ecip1 mutants and higher than ein2-1 (Fig. 8c, d). The data suggest that ein2-1 substantially suppressed the early flowering phenotype of ecip1.

Genetic interactions of ecip1 with etr2-1 and ein4-1

As ECIP1 interacts with ETR2 and EIN4 in yeast two-hybrid assay (Fig. 3), we further studied the genetic interaction of ECIP1 with the two subfamily II ethylene receptor genes ETR2 and EIN4. Double mutants ecip1-1etr2-1 and ecip1-1ein4-1 were generated and the ethylene response was examined (Fig. 9a, b). In the presence or absence of ethylene, ecip1-1ein4-1 mutant hypocotyls were significantly shorter than those of ein4-1, suggesting that ecip1-1 mutation partially suppresses ethylene-insensitivity in ein4-1. However, ecip1-1etr2-1 mutant was only slightly but significantly shorter than etr2-1 single mutant in the absence of ethylene, implying that ecip1-1 mutation slightly reduces hypocotyl length of etiolated etr2-1 seedlings.

Figure 9.

Double mutant analysis of ecip1 with ethylene receptor mutants etr2-1 and ein4-1. (a) Comparison of dark-grown hypocotyls in air and ethylene. (b) Hypocotyl lengths of dark-grown double and single mutants in response to ethylene. Each column represents average from 50 to 70 seedlings and bars indicate standard deviation. For ecip1-1, ‘**’ indicate significant difference compared with Col-0 at P < 0.01. For ecip1-1etr2-1, ‘*’ indicates significant difference compared with etr2-1 at P < 0.05. For ecip1-1ein4-1, ‘**’ indicate significant difference compared with ein4-1 at P < 0.01.


We previously demonstrate that ethylene receptor signalling confers salt sensitivity and ethylene precursor ACC suppresses this sensitivity (Cao et al. 2007), consistent with a negative regulation between ethylene and its receptors (Hua & Meyerowitz 1998). The results indicate that ethylene signalling is beneficial for salt stress tolerance. Now we studied the roles of EIN2 (Alonso et al. 1999), a central component of ethylene signalling, in salt stress responses and found that EIN2 is essential for salt tolerance. A MA3 domain-containing protein ECIP1 was identified to interact with EIN2 and regulate salt response and ethylene response.

Two loss-of-function mutant alleles of EIN2 gene were extremely sensitive to salt stress as revealed by phenotypic change, rosette size and relative electrolyte leakage levels (Fig. 1a–c), and introduction of the CEND of the EIN2 in the ein2-5 background (Alonso et al. 1999) resulted in recovery of the salt-tolerant phenotype and relevant physiological parameters (Fig. 1a–d), supporting that the EIN2 is absolutely required for plant salt tolerance. These results are consistent with previous findings that ein2-1 mutant has reductions in both root growth and seed germination under salt stress (Cao et al. 2007; Wang et al. 2007). The roles of EIN2 in salt stress response suggest that whole ethylene signalling pathway may also be involved in this process. In fact, overexpression of the subfamily II ethylene receptor NTHK1 from tobacco causes salt sensitivity in both transgenic tobacco and Arabidopsis plants, and ethylene receptor gain-of-function mutants such as etr1-1 and ein4-1 exhibit the same salt sensitivity as the ein2-1 mutants (Cao et al. 2006, 2007). The kinase domain and the serine/threonine kinase activity of the NTHK1 ethylene receptor from tobacco are required for the salt responses through transgenic analysis (Zhou et al. 2006; Chen et al. 2009).

The downstream component EIN3 was also analysed previously in relation to salt stress; however, single ein3-1 mutant showed no significant phenotypic change but had some physiological changes under 100 mm NaCl (Cao et al. 2007). On 200 mm NaCl, the ein3-1 mutant exhibited reduced salt tolerance (Achard et al. 2006). Additionally, ebf1-1ebf1-2 double mutant with stabilized EIN3 protein showed increased salt tolerance, and EIN3 may promote salt tolerance by enhancing DELLA function (Guo & Ecker 2003; Achard et al. 2006). In the present study, the ein3-1eil1-1 double mutant exhibited severe salt sensitivity as evaluated by phenotypic change, rosette size and relative electrolyte leakage (Fig. 2a, b), very similar to the case in ein2 mutants with NaCl treatment (Figs 1 & 2). Therefore, EIN3 and EIL1 may function redundantly downstream of EIN2 to regulate salt tolerance. All these analyses indicate that ethylene receptors-EIN2-EIN3/EIL1 pathway regulates salt stress responses.

The salt sensitivity of ein2 mutant may result from disrupted Na+ and K+ accumulation (Fig. 1d). Altered Na+/K+ ratio has been observed in transgenic plants overexpressing tobacco ethylene receptor NTHK1 when salinity was applied (Cao et al. 2006). Recent studies demonstrate that ethylene is involved in low potassium deprivation (Shin & Schachtman 2004; Jung, Shin & Schachtman 2009). The induction of a low K+-inducible potassium transporter gene HAK5 was significantly reduced in ein2-1 compared with the wild type under K+ deficient conditions (Jung et al. 2009). It is speculated that EIN2 and ethylene signalling may affect ion homeostasis to alter salt stress response.

As CEND is enough for recovery of salt tolerance in ein2 mutant (Fig. 1), a portion of the CEND was used as bait to screen EIN2-interacting proteins and the ECIP1 with four tandem MA3 domains was identified after yeast two-hybrid assay and GST pull-down assay (Fig. 3). The D domain of the EIN2 CEND has 196 amino acids and is enough for interaction with the C-terminal two MA3 domains of ECIP1 (Fig. 3c, d). Qiao et al. (2009) reported that the region containing the last 250 amino acids of EIN2 is necessary and sufficient for the interaction of EIN2 with F-box proteins ETP1 and ETP2. Therefore, the EIN2 C-terminal region with ∼200 to 250 amino acids may be required for interaction with ETP1/2 and ECIP1. How the specific region of EIN2 discriminates between different proteins remains to be further investigated.

In animals, MA3 domains have important roles in directing the interaction between proteins (Waters et al. 2007; Lankat-Buttgereit & Goke 2009). PDCD4 is tumour-suppressor protein that contains two C-terminal MA3 domains and is known to inhibit translation by binding to two eukaryotic translation initiation factors eIF4A and eIF4G (Shibahara et al. 1995; Suzuki et al. 2008). In the present study, the ECIP1 with four MA3 domains was isolated and the C-terminal two MA3 domains of ECIP1 were shown to interact with the D domain of EIN2, a central membrane component of ethylene signalling in Arabidopsis plants (Fig. 3). Interaction between full-length of ECIP1 and the D domain of EIN2 was not observed in yeast two-hybrid assay (data not shown), possibly because of the conformational blocking or other unknown mechanisms. Whether the two MA3 domains at N-terminal end of ECIP1 suppress the interaction or whether each MA3 domain has any specificity for interaction remains to be further investigated. In PDCD4, the two MA3 domains can act synergistically to interact with eIF4A whereas a single MA3 is sufficient to inhibit translation (Suzuki et al. 2008; Loh et al. 2009). Through mutant analysis, we find that mutation of the ECIP1 leads to increased plant survival rate under salt stress and enhanced ethylene responses (Figs 6 & 7). These roles appear to be negatively correlated with that of EIN2, whose mutation results in salt sensitivity and ethylene insensitivity (Figs 1, 2 & 6). Therefore, the ECIP1 and EIN2 seem to have a negative regulatory relationship in the above two responses. However, it should be mentioned that the ecip1 mutant seeds had a low germination rate under salt stress, comparable with the low germination for salt-stressed ein2-1 (Fig. 5), indicating a positive correlation of the two genes during seed germination. It is thus likely that the ECIP1 and EIN2 have negative or positive regulatory mechanisms depending on different processes or responses.

The interaction of the ECIP1 and EIN2 was further disclosed using ecip1-1ein2-1 double mutant. This double mutant is ethylene insensitive regarding hypocotyl length (Fig. 7a, b) and cotyledon expansion behaviour (Fig. 7c, d), and is also salt sensitive in terms of plant survival (Fig. 6). These features all resemble those of ein2-1 mutant, suggesting that ecip1-1 phenotype is solely dependent on ethylene signalling and ein2-1 is epistatic to ecip1-1 genetically in regulation of the above responses.

It is interesting to note that an unexpected interaction between ECIP1 and subfamily II ethylene receptors ETR2 and EIN4 was observed (Fig. 3e, Supporting Information Fig. S1). Double mutant analysis also indicates genetic interaction of ECIP1 with EIN4 or ETR2 (Fig. 9a, b). Considering that ECIP1 also interacts with EIN2, it is possible that the ECIP1 may mediate the signal transduction between subfamily II ethylene receptors and EIN2 to regulate salt stress responses and ethylene responses. It should be mentioned that another subfamily II member ERS2 and subfamily I members ETR1 and ERS1 do not interact with ECIP1 (Fig. 3e), suggesting that the diverged kinase domain in ETR2 and EIN4 plus the receiver domain may determine the interaction specificity. Other methods should be used to further confirm the interactions among these proteins.

CEND was also found to interact with EER5, a protein with a PAM domain, to affect ethylene signalling (Christians et al. 2008). Recent studies have demonstrated that EIN2 is localized at the ER membrane and interacts with all the five Arabidopsis ethylene receptors in tobacco leaf cells (Bisson et al. 2009; Bisson & Groth 2010). The kinase activity of the ETR1 modulates the interaction between receptor and EIN2 (Bisson & Groth 2010). Ethylene receptors can form protein complexes through formation of homodimers or heterodimers (Schaller et al. 1995; Gao et al. 2008; Grefen et al. 2008). ETR1 and ERS1 can interact directly with CTR1 and co-localized in the ER membrane (Clark et al. 1998; Gao et al. 2003). In tomato, all the three CTR1-like proteins LeCTR1, LeCTR3 and LeCTR4 mainly interact with subfamily I ethylene receptors LeETR1, LeETR2 and NR, but not interact with subfamily II members LeETR4, LeETR5 and LETR6. For LeCTR2, it interacts with LeETR1 and LeETR2 but not NR or other subfamily II receptors (Lin et al. 2008; Zhong, Lin & Grierson 2008; Lin & Grierson 2010). A newly identified protein SlTPR1 binds the tomato NR and LeETR1 (Lin et al. 2008) and the homologous protein AtTPR1 in Arabidopsis also interacts with ERS1 (Lin, Ho & Grierson 2009). It is possible that a number of proteins including ethylene receptors, EIN2, CTR1, ECIP1, EER5 and TPR1, etc., can form signalling complexes and perform signal transductions for ethylene signal. The complex may have some major components and additional components recruited in different responses or at different developmental stages or in different cells, tissues or organs of the plants to finish the signalling processes with specific intensity and duration. Considering that ethylene receptors and EIN2 mainly localized in ER membrane (Bisson et al. 2009) whereas the ECIP1 was mainly located in cytoplasm, how and where the ECIP1 would transfer signal and whether ethylene would affect ECIP1 translocation require further investigation. Additionally, whether the present ECIP1 and/or the above proposed complex interact also with EIN3/EIL1 represents a new speculation and further studies should be conducted to test these possibilities.

Ethylene has long been regarded as a stress hormone, however, its roles in abiotic stress is unclear. Our previous analysis finds that application of ethylene precursor ACC inhibits salt-sensitive response caused by overexpression of ethylene receptor NTHK1, supporting that ethylene plays a positive role in salt stress tolerance (Cao et al. 2007). Presently, we further find that ethylene-signalling components EIN2 and EIN3/EIL1 are positive regulators of salt stress tolerance. The newly identified EIN2-interacting protein ECIP1, acting as a negative regulator for both ethylene response and salt tolerance during plant growth, may function between subfamily II receptors and EIN2.

Collectively, we demonstrate that EIN2 has important roles in adaptation to salt stress. ECIP1, as a novel EIN2-interacting partner with four tandem MA3 domains, also participates in ethylene response and salt stress responses. Further study should be performed to elucidate the function of ECIP1 in interactions with ethylene signalling components and in stress responses.


We thank Dr. Hong-Wei Guo (Beijing University) and Dr. J.R. Ecker (The Salk Institute for Biological Studies) for kindly providing ein2-5:CEND and ein3-1eil1-1 double mutant seeds, and ABRC for ecip1 mutant seeds. This work was supported by National Natural Science Foundation of China (90717005, 30925006), CAS Project (KSCXZ-YW-N-010) and National Transgenic Research Projects (2009ZX08009-054B, 2008ZX08004-002-3-3).