ENHANCING CTR1-10 ETHYLENE RESPONSE2 is a novel allele involved in CONSTITUTIVE TRIPLE-RESPONSE1-mediated ethylene receptor signaling in Arabidopsis

Background The signal output of ethylene receptor family members is mediated by unknown mechanisms to activate the Raf-like protein CONSTITUTIVE TRIPLE RESPONSE1 (CTR1) in negatively regulating ethylene signaling. The physical interaction between the ethylene receptor histidine kinase (HK) domain and CTR1 N terminus is essential to the CTR1-mediated receptor signal output. To advance our knowledge of the involvement of CTR1-mediated ethylene receptor signaling, we performed a genetic screen for mutations that enhanced the constitutive ethylene response in the weak ctr1-10 allele. Results We isolated a loss-of-function allele of ENHANCING ctr1-10 ETHYLENE RESPONSE2 (ECR2) and found that ecr2-1 ctr1-10 and the strong allele ctr1-1 conferred a similar, typical constitutive ethylene response phenotype. Genetic analyses and transformation studies suggested that ECR2 acts downstream of the ethylene receptors and upstream of the transcription factors ETHYLENE INSENSITIVE3 (EIN3) and EIN3-LIKE1 (EIL1), which direct the expression of ethylene response genes. Signal output by the N terminus of the ethylene receptor ETHYLENE RESPONSE1 (ETR1) can be mediated by a pathway independent of CTR1. Expression of the N terminus of the ethylene-insensitive etr1-1 but not the full-length isoform rescued the ecr2-1 ctr1-10 phenotype, which indicates the involvement of ECR2 in CTR1-mediated but not -independent, ethylene receptor signaling. ECR2 was mapped to the centromere region on chromosome 2. With incomplete sequence and annotation information and rare chromosome recombination events in this region, the cloning of ECR2 is challenging and still in progress. Conclusions ECR2 is a novel allele involved in the ethylene receptor signaling that is mediated by CTR1. CTR1 activation by ethylene receptors may require ECR2 for suppressing the ethylene response.


Background
Ethylene is a gaseous hormone regulating many aspects of plant growth and development. The dicotyledonous model plant Arabidopsis has five ethylene receptors that physically act at the endoplasmic reticulum (ER) with the Raf-like protein CONSTITUTIVE TRIPLE-RESPONSE1 (CTR1) to negatively regulate ethylene signaling [1][2][3][4]. In the absence of ethylene, the receptor signal output is mediated by an unknown mechanism(s) to activate CTR1, and CTR1 is presumably activated to phosphorylate the downstream signaling component ETHYLENE INSENSITIVE2 (EIN2). Phosphorylated EIN2 stays at the ER and cannot induce the ethylene response. With ethylene binding to ethylene receptors, the receptor signal output is prevented and CTR1 is not activated. Unphosphorylated EIN2 undergoes proteolytic cleavage by an unknown mechanism to produce a C-terminal fragment, which enters the nucleus to induce the ethylene response [5,6]. Targets of the EIN2 C terminus remain to be identified. The transcription factors EIN3 and EIN3-LIKE1 (EIL1) direct the expression of ethylene-responsive genes [7], and EIN2 C-terminus-induced ethylene responses are prevented in the ein3-1 loss-of-function mutant [6]. Conceivably, the EIN2 C terminus could mediate the ethylene signaling to EIN3 and EIL1.
The ethylene receptors are structurally similar to histidine kinases (HKs) of prokaryotic two-component modules, and studies have revealed the domain functions of the ethylene receptors. The N terminus has three or four transmembrane domains (TMs) that bind a copper cofactor for ethylene binding and are required for localization at the ER [8][9][10][11]. Following the TMs is the GAF domain for non-covalent receptor interaction to mediate interreceptor signaling [12][13][14]. The C-terminal portion is the HK domain, which is believed to function in the receptor signal output via direct interaction with the CTR1 N terminus [2,4]. CTR1 has serine/threonine kinase activity, and the ethylene response is inversely associated with CTR1 kinase activity [4,15]. These studies suggest that the HK domain mediates ethylene receptor signaling to the CTR1 N terminus, thus activating CTR1 to suppress the ethylene response, although the underlying biochemical mechanisms are elusive.
Recent studies suggest that the HK domain can be dispensable to ethylene receptor signal output and that CTR1 is not the only component mediating the signaling. Mutations that delete ETHYLENE RESPONSE1 (ETR1) HK and receiver domains have little effect on the receptor signal output [13,16], and expression of the truncated etr1  isoform that lacks the site for interacting with CTR1 rescues the ctr1 loss-of-function mutant phenotype. Conceivably, the ETR1 receptor signal output can be mediated via the N terminus to an alternative pathway independent of CTR1 [3,14]. Of note, ETR1 receptor signaling in kinase-defective ctr1 mutants occurs only when HK-domain-lacking ETR1 isoforms are expressed. A kinase-defective ctr1 isoform may dock at the ETR1 HK domain and actively prevent ETR1 N-terminal signaling to the CTR1-independent alternative pathway [14,17].
The ctr1-10 mutation results from a T-DNA insertion at the 5′-untranslated region (5′-UTR) of CTR1 [18]. The 5′-UTR of a mRNA may be highly structured and contain upstream AUGs (uAUGs) and internal ribosome entry sites that can affect the translation. With increased upstream open reading frames (uORFs) and a highly structured nature, a long 5′-UTR can be translationally inhibitory [19][20][21][22][23]. With a higher CTR1 level in ctr1-10 than in the wild type [18], the T-DNA insertion may substantially increase the 5′-UTR length of the CTR1 transcript to affect translation efficiency, thus reducing the CTR1 level and thus activity as compared with the wild type.
The underlying biochemical mechanisms of CTR1 activation by the ethylene receptors remain to be unraveled. Mutations that enhance the constitutive ethylene response in ctr1-10 could have a role involving CTR1 activation, protein stability or ethylene signaling. Here we report that the loss-of-function mutation of ENHANCING ctr1-10 ETHYLENE RESPONSE2 (ECR2) enhanced the ethylene response in ctr1-10 comparable to the strong allele ctr1-1 and the ethylene-treated wild type. Results from extensive genetic and transformation studies suggested that ECR2 acts downstream of the ethylene receptors and upstream of the positive ethylene response regulators EIN3 and EIL1. We discuss possible roles for ECR2 in the negative regulation of the ethylene response.

ctr1-10 is a weak allele
The ctr1-10 mutation results from a T-DNA insertion at the 5′-UTR, and the mutant shows a weak constitutive ethylene response phenotype [18]. The nature of the mutation that causes the weak phenotype remains to be determined.
Given that the T-DNA inserts at the 5′-UTR, we expected that the T-DNA fragment is transcribed as part of the 5′-UTR of ctr1-10 mRNA ( Figure 1A). Indeed, RT-PCR with primers pairing the 5′-UTR and T-DNA sequences showed the presence of a chimeric cDNA fragment containing the sequence of the ctr1-10 5′-UTR and the T-DNA ( Figure 1B). For the same gene, a longer 5′-UTR can be translationally more inhibitory than a shorter one [19,20,22,23]. With a 4.3-kb T-DNA at the 5′-UTR, we hypothesized that ctr1-10 could produce a lower level of CTR1 than the wild type, for a weak constitutive ethylene response phenotype.
Ethylene inhibits hypocotyl growth and the ethylene response can be quantified by measurement of hypocotyls. Hypocotyls were shorter for ctr1-10 than wild-type (Col-0) seedlings over a wide range of ethylene concentrations ( Figure 1C). When growth for wild-type and ctr1-10 seedlings are normalized to growth with no added ethylene, it becomes apparent that ctr1-10 seedlings are more sensitive to ethylene with a shift from 1 μL L -1 for wild-type to 0.1 μL L -1 for the mutant ( Figure 1D). These data suggest increased ethylene sensitivity with the ctr1-10 mutation.
Genetic screen for ctr1-10 enhancer mutations To isolate components that could be involved in CTR1mediated ethylene receptor signaling, we performed an enhancer screen for ctr1-10. Alleles that enhanced the ctr1-10 mutant phenotype were designated enhancing ctr1-10 ethylene response (ecr). Here we describe the isolation of ecr2-1 ctr1-10.
The silver ion Ag(I) may bind ETR1 and prevent ethylene inhibition of ETR1 receptor signaling [11,24]. Silver treatment largely prevented ethylene-induced inhibition of seedling hypocotyl growth in wild-type (Col-0) and ctr1-10 but not ecr2-1 ctr1-10 seedlings ( Figure 2C). Of note, the seedling hypocotyls were longer for silvertreated than non-silver-treated ecr2-1 ctr1-10 seedlings (Student's t test, P < 0.01), regardless of ethylene treatment; the difference was minor and could be of little biological significance. These results suggest that ETR1 receptor signaling that was mediated in part in ctr1-10 was prevented in ecr2-1 ctr1-10.
Aminoethoxyvinylglycine (AVG) prevents the biosynthesis of the immediate ethylene biosynthesis precursor 1-aminocyclopropane-1-carboxylic (ACC), and the treatment can reduce endogenous ethylene production [25]. AVG (10 μM) treatment did not rescue the reduced growth of ecr2-1 ctr1-10 seedlings, so ecr2-1 did not enhance the production of ethylene to cause reduced growth ( Figure 2D). ETHYLENE OVERPRODUCER1 (ETO1) is a negative regulator of the ACC-synthesing enzyme ACC SYNTHASE5 (ACS5) and the eto1-1 allele produces a higher level of ACC, and therefore ethylene, than does the wild type [26,27]. As a control, eto1-1 seedlings showed a short hypocotyl, and growth was rescued with AVG treatment ( Figure 2D); AVG treatment in this study was sufficient to prevent endogenous ethylene production.
We examined the growth inhibition phenotype at other developmental stages. Grown under light, ethylene treatment inhibited cotyledon expansion and root elongation in wild-type (Col-0) seedlings. Seedlings of ctr1-1, ctr1-10, and  ecr2-1 ctr1-10 showed the growth inhibition phenotype regardless of ethylene treatment ( Figure 2E). At the adult stage, rosettes were larger for the wild type (Col-0) than the three mutants, with ctr1-1 and ecr2-1 ctr1-10 being phenotypically similar and both producing a smaller rosette than ctr1-10 ( Figure 2F).
We attempted to clone ECR2 by map-based cloning and mapped ECR2 to a 598-kb region spanning the centromere at chromosome 2 ( Figure 3C), which agreed with ECR2 and CTR1 (chromosome 5) being in distinct linkage groups. However, we did not obtain more recombinants to advance the mapping to a narrower region. This region contains sequence gaps (indicated as [gapbp]ExpandNs or a string of Ns in Additional file 1) and fragments with repeated sequences, retrotransposon, transposase, and transposable elements. We sequenced 39 annotated genes (excluding those annotated as retrotransposon, transposase, and transposable elements) but did not identify any mutation (Additional file 1). Among the genes, T-DNA insertion mutants for At2g07981and At2g08986 were each genetically crossed with ctr1-10 to test whether any of these was ECR2, and these mutations did not enhance the ctr1-10 phenotype.
The mutation that affects ECR2 remained to be identified and we attempted to isolate the ecr2-1 single mutant. The F2 seedlings from the genetic crossing of ecr2-1 ctr1-10 and the wild type (Col-0) were crossed with ctr1-10, and the individual that produced the ecr2-1 ctr1-10 mutant phenotype in the F2 generation with a 1:15 segregation ratio was considered the ecr2-1 mutant. We isolated ecr2-1, and results from 3 independent crossings with ctr1-10 showed a 1:15 segregation ( Figure 3D). Grown in dark, ecr2-1 seedlings produced a shorter hypocotyl than did the wild type (Col-0) at low ethylene concentrations ( Figure 3E and 3F); without ethylene treatment, the height of ecr2-1 seedlings was 1.355±0.62 mm (99% confidence level) shorter than that of wild-type seedlings. When this data was normalized, we found that wild-type and ecr2-1 seedlings had indistinguishable sensitivity to ethylene ( Figure 3G). Although the relative hypocotyl length was longer (by 4.6±2.9%; 99% confidence level) for ecr2-1 seedlings than wild-type seedlings with 1 μL L -1 ethylene, the difference was small and could be of little biological significance. At the adult stage, the rosette was smaller for ecr2-1 than the wild type ( Figure 3H). We excluded the possibility that the mutant phenotype of ecr2-1 and ecr2-1 ctr1-10 could be due to greater ethylene evolution than the wild type (Col-0), as supported by ethylene evolution ( Figure 3I).
These results suggest that the ecr2-1 allele conferred a minor growth inhibition throughout development. As compared with the wild type (Col-0), etiolated ecr2-1 seedlings showed no alteration in ethylene sensitivity.
These data suggest that the ecr2-1 allele prevented the receptor signaling with the full-length etr1-1 and etr1-2 but not etr1-1 N terminus in ctr1-10. Therefore, ECR2 may be involved in CTR1-dependent but not -independent receptor signaling in ctr1-10. Given that the signaling of full-length ETR1 but not ETR1 N terminus is prevented by kinase-defective ctr1 isoforms [14,17], ECR2 could be involved in part in CTR1 kinase activity or protein stability. ctr1-8 is not associated with the ethylene receptors, and the mutant shows a weak constitutive ethylene response. We previously hypothesized that mediation of ethylene receptor signaling in ctr1-8 was independent of CTR1 [4,17]. The possibility that ctr1-8 could mediate in part the receptor signaling, for a weak mutant phenotype, should be considered. Given that ECR2 is involved in CTR1-dependent ethylene receptor signaling, we expected that ecr2-1 would have little effect on ctr1-8 if ctr1-8 did not mediate the ethylene receptor signaling.
The length of etiolated ecr2-1 ctr1-8 seedlings was only 1.4±0.43 mm (99% confidence level) shorter than that of ctr1-8 seedlings ( Figure 4I and 4J), which is in line with the result showing that the ecr2-1 allele conferred minor seedling growth inhibition ( Figure 3E and 3F) but not increased ethylene sensitivity ( Figure 3G). Light-grown seedlings of ctr1-8 and ecr2-1 ctr1-8 were phenotypically similar, with a shorter root and smaller cotyledons than the wild type (Col-0) ( Figure 4K). Consistently, the rosette was larger for the wild type than ctr1-8 and ecr2 ctr1-8 plants ( Figure 4L). Therefore, the ecr2-1 allele may have little effect on the ethylene response phenotype in ctr1-8. We quantified the ethylene response for the two mutants. ERF1 expression was identical in ctr1-8 and ecr2-1 ctr1-8 (Student's t test, P = 0.363) and higher than in the wild type ( Figure 4M). Both ctr1-8 and ecr2-1 ctr1-8 had the same senescence phenotype, and their chlorophyll a content was identical before (Student's t test; P = 0.056) and after (Student's t test; P =0.059) the test ( Figure 4N and 4O).
The ecr2-1 allele had little effect on ctr1-8 ethylene response, which suggests that ethylene receptor signaling in ctr1-8 was predominantly mediated by a pathway independent of CTR1.
Genetic analysis of effect of ein3-1 and eil1-1 on the ethylene response in ecr2-1 ctr1-10 The present data suggested the involvement of ECR2 in ethylene receptor signaling mediated by CTR1. ETHYLENE INSENSITIVE2 (EIN2) acts downstream of CTR1, and ein2 loss-of-function mutation may rescue the ecr2-1 ctr1-10 mutant phenotype. Genetic analyses for the effects of ein2 on ecr2-1 ctr1-10 were not successful because EIN2 is tightly linked with CTR1 and ECR2 remains to be cloned.
EBF1 and EBF2 overexpression rescues ecr2-1 ctr1-10 phenotype EIN3-BINDING F-BOX1 (EBF1) and EBF2 are F-box proteins involved in the ubiquitination of EIN3 and EIL1 to mediate their degradation by the 26S proteosome [32,33]. Excess EBF1 and EBF2 reduces EIN3 and EIL1 levels and suppresses the ethylene response [34]. The argument that ECR2 acts in the ethylene signal transduction pathway upstream of the transcription factors EIN3 and EIL1 can be tested by examining whether overexpression of EBF1 and EBF2 rescues the ecr2-1 ctr1-10 phenotype.
We also examined other aspects of the ethylene response affected by the transgenes. ERF1 expression was greater in ecr2-1 ctr1-10 and ctr1-1 than the wild type; the expression of each transgene largely reduced ERF1 levels ( Figure 6J). Of note, ERF1 expression was greater in 35S:EBF1 ctr1-1 than 35S:EBF1 ecr2-1 ctr1-10 lines, likely because of its lower EBF1 expression ( Figure 6H).
Consistent with the suppression of ecr2-1 ctr1-10 by ein3-1 and eil1-1, the respective elevation in EBF1 and EBF2 levels in ecr2-1 ctr1-10 rescued the inhibition in mutant growth and reduced ERF1 levels, which supports that ECR2 acts upstream of the transcriptions factors EIN3 and EIL1 in the ethylene signal transduction pathway.

Discussion
Arabidopsis ethylene signaling is negatively regulated by ethylene receptors and the Raf-like protein CTR1 [1][2][3]35]. The biochemical nature of receptor signaling is unknown, as are the underlying mechanisms by which CTR1 is activated [3,14,36]. A genetic screen for enhancers facilitates the isolation of genes or mutations that play a role in the same biological process as the known mutation. Given that the ctr1-10 mutation does not disrupt the CTR1 ORF and that ctr1-10 is a weak allele, an enhancer screen for ctr1-10 could lead to isolation of components involved in CTR1-mediated ethylene receptor signaling.
In this study, we reported the isolation of ecr2-1 as an enhancer mutation for ctr1-10. With a relative weak constitutive ethylene response phenotype for ctr1-10 and a strong phenotype for ecr2-1 ctr1-10, the mutations ecr2-1 and ctr1-10 synergistically facilitated ethylene signaling. The assumption that CTR1 level in ctr1-10 could be reduced needs to be verified biochemically. Nevertheless, extensive studies suggest that for a common transcript, a long 5′-UTR can be translationally inhibitory as compared with a short one [22,37,38]. Conceivably, the T-DNA insertion that increases the 5′-UTR length may add structures of higher order and uORFs to inhibit the translation of CTR1, thus reducing CTR1 level, so that ctr1-10 shows a weak constitutive ethylene response phenotype. We considered 3 scenarios explaining the synergistic effect of ecr2-1 and ctr1-10: 1) ECR2 could be required in part for CTR1 activity or the protein stability, 2) ECR2 and CTR1 act in different pathways, or 3) ECR2 could be a signaling molecule acting with CTR1 to suppress the ethylene signaling. In the first scenario, CTR1 level, and possibly activity, in ctr1-10 is presumably reduced to a level requiring ECR2 to suppress the ethylene signaling.
The nature of ctr1-10 and ctr1-8 mutations is distinct: CTR1 level could be reduced in ctr1-10 and the ctr1-8 protein is not associated with the ethylene receptors [4,18]. The docking of kinase-defective ctr1 isoforms at the HK domain of ETR1 prevents receptor signaling to an alternative CTR1-independent pathway. In ctr1-8, the ethylene receptors are free from association with ctr1-8; thus, the receptor signaling is mediated to the alternative pathway and the mutant shows a weak phenotype [14,17]. Conceivably, in ecr2-1 ctr1-10, the activity of CTR1 docking at the ethylene receptors could be greatly reduced, and ethylene receptor signaling that is mediated by CTR1 as well as the alternative pathway was prevented. In contrast, the ctr1-8 protein does not dock at the ethylene receptors and the receptor signaling can be mediated by the alternative, CTR1-independent pathway; thus, ecr2-1 had little effect on ctr1-8 phenotype. In line with these results, our data showing that ecr2-1 prevented etr1-1 and etr1-2 but not etr1-1 1-349 receptor signaling in ctr1-10 also support that ECR2 is involved in the ethylene receptor signaling that is dependent on CTR1.
An alternative explanation for the distinct effects of ecr2-1 on ctr1-10 and ctr1-8 could arise from a role of ECR2 in CTR1 stability. In this scenario, very small amount of the ctr1-8 protein that could associate with the ethylene receptors could have a higher turnover rate in the absence of ECR2; meanwhile, ctr1-8 reserved in the soluble fraction could be continuously recruited to the receptors and mediate the receptor signaling. Contrarily, in ctr1-10, CTR1 level is highly reduced without ECR2 and the receptor signaling is prevented. However, this scenario is not consistent with the argument that the ethylene receptor signaling can be mediated to an alternative pathway that is independent of CTR1. In other words, if ECR2 were involved in CTR1 stability, the ethylene receptor signaling that was mediated independent of CTR1 still occurred and the ecr2-1 allele would have little effect on the degree of ctr1-10 ethylene responses.

Conclusions
CTR1 is presumably activated by the ethylene receptors, and the mechanism is unknown [1]. Little is known about the presence of any other components that are involved in the ethylene receptor signaling to CTR1. The present data suggest that ECR2 is a component involved in part in CTR1-mediated ethylene receptor signaling, which indicates a regulatory mechanism for the receptor signaling. We favor the argument that CTR1 activity but not level was affected by ecr2-1 in ctr1-10. With limited sequence information for the 598-kb region that contains ECR2, the cloning of ECR2 is currently hampered. Complementation of ecr2-1 by a large DNA fragment that contains ECR2 is ongoing. However, with repeated sequences, transposons, retrotransposons, and sequence gaps in this region, cloning of a large DNA fragment is still highly challenging. Nevertheless, the cloning of ECR2 will advance our knowledge of ethylene signaling that involves CTR1 activation.

Plant materials and growth conditions
ctr1-10 and the PCR primers for genotyping were previously described [18]. For mutagenesis, following ethyl methanesulfonate treatment, ctr1-10 seeds (M 0 ) were washed with continuous flowing water and grown on soil as 380 pools (20-25 M 1 plants in a pool). Seedlings at the M 2 generation showing the constitutive ethylene response phenotype were candidates carrying the enhancer mutation and were characterized for other aspects of the ethylene response. We screened 162 pools, and ecr2-1 ctr1-10 was isolated. For the growth of etiolated seedlings, seeds were stratified at 4°C for 72 hr and then moved to 22°C for germination (80 hr) in the dark. Seedling hypocotyl length was measured by use of Video tesT (Moscow) [29]. For measuring growth of seedlings and rosettes under light, stratified seeds were germinated and grown under 16-hr light/8-hr dark at 25°C; seedlings were phenotyped 7 days after germination and rosettes 4 weeks after germination. The isolation of ecr2-1 and the genetic crossing of ecr2-1 with etr1-1, etr1-2, ctr1-8, ein3-1, and eil1-1 are described in the Results section and the segregation was verified by chisquare test (α=0.01). For ethylene treatment, the ethylene concentration was 20 μL L -1 , or otherwise as indicated.

Leaf senescence test
Detached leaves from plants were incubated in a Petri dish with wet filter paper in the dark for 5 days. Leaf senescence was quantified by measuring chlorophyll a content [39].

Ethylene measurement
Ethylene evolution in light-grown seedlings was measured by use of the Ethylene Detector (ETD-300 by Sensor Sense) with the "stop-and-flow" measurement. In brief, 25-30 seedlings were grown in a vented vial for 5 days after germination. The vial was closed for 3.5 hr to let ethylene accumulate, and the ethylene amount was measured. We measured 3 independent biological samples for each genotype, and ethylene evolution is represented as mean (nL L -1 seedling -1 hr -1 )±SD.

Statistical analysis
Data are expressed as mean±SD or mean±SE for gene expression. Student's t test (α = 0.01) was used for comparing 2 groups and Scheffe test (α = 0.01) for multiple groups. Chi-square test (α = 0.01) was used for testing segregation ratio.

Additional file
Additional file 1: BAC clones and sequences encompassing ECR2.