Histidine-Kinase Activity of the Ethylene Receptor ETR1 Facilitates the Ethylene Response in Arabidopsis 1

In Arabidopsis thaliana , ethylene is perceived by a receptor family consisting of five members. Subfamily-1 members ETR1 and ERS1 have histidine-kinase activity, unlike the subfamily-2 members ETR2, ERS2, and EIN4 which lack amino-acid residues critical for this enzymatic activity. To resolve the role of histidine-kinase activity in signaling by the receptors, we transformed an etr1-9;ers1-3 double mutant with wild-type and kinase-inactive versions of the receptor ETR1. Both wild-type and kinase-inactive ETR1 rescue the constitutive ethylene-response phenotype of etr1-9;ers1-3 , restoring normal growth to the mutant in air. However, the lines carrying kinase-inactive ETR1 exhibit reduced sensitivity to ethylene based on several growth response assays. Microarray and real-time PCR analyses of gene expression support a role for histidine kinase activity in eliciting the ethylene response. In addition, protein levels of the Raf-like kinase CTR1, which physically associates with the ethylene receptor ETR1, are less responsive to ethylene in lines containing kinase-inactive ETR1. These data indicate that the histidine-kinase activity of ETR1 is not required for but plays a modulating role in the regulation of ethylene responses. Models for how enzymatic and non-enzymatic regulation may facilitate signaling from the ethylene receptors are discussed.


INTRODUCTION
The gaseous hormone ethylene plays roles throughout the plant life cycle (Mattoo and Suttle, 1991;Abeles et al., 1992). Ethylene regulates seed germination, seedling growth, leaf and petal abscission, fruit ripening, organ senescence, as well as stress and pathogen responses. In Arabidopsis, ethylene is perceived by a five-member family of receptors composed of ETR1, ERS1, ETR2, ERS2, and EIN4 (Schaller and Kieber, 2002;Chen et al., 2005;O'Malley et al., 2005;Kendrick and Chang, 2008). The ethylene receptors can be divided into two subfamilies based on phylogenetic analysis and some shared structural features, subfamily 1 being composed of ETR1 and ERS1, subfamily 2 being composed of ETR2, ERS2, and EIN4 (Chang and Stadler, 2001;Schaller and Kieber, 2002;Chen et al., 2005). Genetic analysis indicates that the receptors serve as negative regulators of the ethylene response, that there is functional overlap among the receptors, and that the subfamily-1 receptors generally play the predominant role in ethylene signaling (Hua and Meyerowitz, 1998;Wang et al., 2003;Qu et al., 2007).
The ethylene receptors have a similar overall modular structure, each containing three conserved transmembrane domains near the N-terminus, followed by a GAF domain, and then signal output motifs in the C-terminal half. The transmembrane domains contain the ethylenebinding site (Schaller and Bleecker, 1995;Hall et al., 1999;Rodriguez et al., 1999), and also serve to localize the receptor to the endoplasmic reticulum (ER) and possibly to the Golgi apparatus (Chen et al., 2002;Dong et al., 2008;Grefen et al., 2008). The GAF domain has been implicated in protein-protein interactions among the receptors and may help mediate the formation of higher order receptor clusters (Gao et al., 2008;Grefen et al., 2008). In their Cterminal halves, all five receptors contain histidine kinase-like domains and, excepting ERS1 and ERS2, also contain receiver domains. Histidine kinase and receiver domains are signaling elements originally identified as components in bacterial phosphorelays and are now known to be present in plants, fungi, and slime molds (Schaller et al., 2008;Schaller et al., 2011). In twocomponent systems, histidine kinases autophosphorylate on a conserved histidine residue, often in response to an environmental stimulus (Mizuno, 1997;Stock et al., 2000;Gao and Stock, 2009); this phosphate is then transferred to a conserved aspartic acid residue within a receiver domain. Receiver domains are sometimes found joined to the histidine kinases (as occurs with ETR1, ETR2, and EIN4) and sometimes found in separate proteins referred to as response regulators. The subfamily-1 receptors ETR1 and ERS1 have functional histidine kinase domains based on in vitro analysis, suggesting that they could function like canonical histidine kinases in a two-component signaling pathway (Gamble et al., 1998;Moussatche and Klee, 2004).
However, the subfamily-2 receptors ETR2, ERS2, and EIN4 lack the necessary residues for histidine kinase activity and, based on in vitro analysis, are now thought to function as Ser/Thr kinases (Moussatche and Klee, 2004).
Truncation studies using ETR1 demonstrate the importance of the C-terminal half of the protein for signal output, but this importance appears to be largely independent of the enzymatic activity contained in the histidine-kinase domain (Qu and Schaller, 2004;Xie et al., 2006).
Instead the key role for the histidine-kinase domain appears to be as a docking site for the downstream Raf-like kinase CTR1, mutations of which have substantial effects on ethylene signaling (Kieber et al., 1993;Huang et al., 2003). Nevertheless, several studies suggest that the histidine kinase activity of ETR1 may modulate aspects of ethylene signaling. In one study, when a kinase-inactive mutant of ETR1 was introduced into the triple mutant etr1;etr2;ein4, it was found to rescue the mutant phenotype but showed increased sensitivity to ethylene (Qu and Schaller, 2004). A second study suggests that kinase activity regulates the ability of seedlings to

Kinase-Inactive ETR1 Rescues the Constitutive Ethylene-Response Phenotype of etr1-9;ers1-3
We addressed the role of receptor histidine-kinase activity in signaling by examining the ability of kinase-inactive versions of ETR1 to rescue the constitutive ethylene-response phenotype found in the etr1-9;ers1-3 double mutant, following the general strategy illustrated in Fig. 1A. The histidine-kinase domain of ETR1 contains conserved residues essential for activity based on the well-characterized His kinases of bacteria and prior characterization of ETR1 (Gamble et al., 1998;Stock et al., 2000;Gamble et al., 2002;Moussatche and Klee, 2004). These include a His residue (His353) that serves as the autophosphorylation site and a catalytic domain with two groups of conserved Gly residues referred to as the G1 and G2 boxes ( Fig 1B).
Mutations in these conserved residues eliminate autophosphorylation of ETR1 when examined in vitro (Gamble et al., 1998;Gamble et al., 2002;Moussatche and Klee, 2004). Based on this information, we generated three kinase-inactive versions of ETR1 for analysis in plants. The mutant ETR1-G2 contains a mutated G2 box (G545A and G547A), predicted to interfere with catalysis by disrupting ATP binding to the catalytic domain (Gamble et al., 2002). The ETR1-G2 mutant, in addition to interfering with autophosphorylation activity, should be incapable of transphosphorylating other receptors, a consideration because ethylene receptors form higher order clusters (Gao et al., 2008;Grefen et al., 2008). For a second kinase-inactive mutant (ETR1-H/G2), we combined the G2 mutation of ETR1 with a mutation of the His (His353Gln) that serves as phosphor-accepting site (Gamble et al., 1998;Moussatche and Klee, 2004). Inclusion of the His mutation should eliminate autophosphorylation of ETR1, either by other histidine kinases of Arabidopsis or by any residual kinase activity remaining in the G2 box mutant of ETR1. We also generated a third mutant, in which we combined a mutation (Asp659Asn) of the putatively phosphorylated Asp of the receiver domain, with the prior two mutations to create ETR1-H/G2/D. The rationale for the Asp659Asn mutation was that, although the receiver domain would not be phosphorylated by kinase-inactive ETR1, it could potentially serve as a target for other Arabidopsis histidine kinases such as the cytokinin receptors, which like ethylene receptors localize to the ER membrane (Chen et al., 2002;Dong et al., 2008;Caesar et al., 2011;Wulfetange et al., 2011).
All ETR1 constructs were derived from a genomic fragment that contains both promoter and coding regions of ETR1 (Chang et al., 1993). Wild-type and kinase-inactive versions of ETR1 were transformed into the etr1-9;ers1-3 double mutant, which lacks the histidine-kinase containing receptors of subfamily-1 and exhibits a strong constitutive ethylene-response phenotype (Qu et al., 2007). Since the homozygous double mutant is sterile, constructs were initially transformed into etr1-9/etr1-9;ers1-3/ERS1 plants, with plants homozygous for etr1-9, ers1-3, and the transgene identified in subsequent generations. Several independent lines were isolated for each construct and characterized for their ability to rescue dark-grown and lightgrown phenotypes of etr1-9;ers1-3.
Dark-grown seedlings of etr1-9;ers1-3 exhibit a constitutive ethylene-response phenotype when grown in the absence of ethylene (in air) (Qu et al., 2007). As shown in Figure   2A, this mutant phenotype is characterized by inhibition of root and hypocotyl elongation, an exaggerated apical hook, and a thickening of the hypocotyl. These features contrast sharply with the etiolated phenotype observed in wild-type seedlings as well as in the single etr1-9 and ers1-3 mutants ( Fig. 2A). As expected, transgenic expression of wild-type ETR1 (tETR1-wt) rescues the constitutive ethylene-response phenotype of dark-grown etr1-9;ers1-3 seedlings ( Fig. 2A) (Qu et al., 2007). The kinase-inactive versions of ETR1 (tETR1-G2 and tETR1-H/G2) also rescue the constitutive ethylene-response phenotype of etr1-9;ers1-3, the transgenic seedlings proving phenotypically indistinguishable from seedlings rescued by expression of wild-type ETR1 ( Fig. 2A).
The etr1-9;ers1-3 mutant also exhibits a pronounced constitutive ethylene-response phenotype when grown in the light. Compared to wild type, the mutant plants are dwarfed, late flowering, sterile, exhibit premature leaf senescence, and have altered floral morphology and development ( Fig. 2B-D) (Qu et al., 2007). Expression of the wild-type transgene tETR1-wt in the etr1-9;ers1-3 background rescues all these phenotypes, indicating that they originate from a lack of the ethylene receptors ( Fig. 2B-D) (Qu et al., 2007). The kinase-inactive versions of ETR1 (tETR1-G2 and tETR1-H/G2) also rescue these light-grown phenotypes of etr1-9;ers1-3 ( Fig. 2B-D), the transgenic plants proving indistinguishable from plants rescued by expression of the wild-type ETR1. Overall, these data indicate that the kinase-inactive versions of ETR1 can functionally replace a wild-type version of ETR1 in terms of their ability to rescue the constitutive ethylene-response phenotypes observed in the etr1-9;ers1-3 double mutant.

Plants with Kinase-Inactive ETR1 Exhibit Reduced Ethylene Responsiveness
The etr1-9;ers1-3 lines containing kinase-inactive ETR1 were indistinguishable from those containing wild-type ETR1 or from wild-type plants themselves when examined in the absence of ethylene (in air). This raises the question as to whether there might be a difference among the lines in terms of their response to ethylene. Growth of dark-grown seedlings in the presence of 1 µL/L ethylene suggested that this might be the case (Fig. 3A). All the transgenic lines exhibited the triple response to the ethylene characterized by the reduction in hypocotyl and root length, formation of an apical hook, and thickening of the hypocotyl, but this response was less pronounced in the kinase-inactive lines (Fig. 3A). In particular, the hypocotyls of the kinaseinactive transgenic lines were longer than those containing wild-type ETR1. Immunoblot analysis, performed with membrane proteins from dark-grown seedlings, confirmed expression of ETR1 in the transgenic lines and revealed a range of ETR1 protein levels (Fig. 3B).
Significantly, levels of the kinase-inactive ETR1 fell within the expression range exhibited in the wild-type ETR1 lines, indicating that phenotypic differences were not owed to differences in the protein levels.
We confirmed the apparent difference in ethylene responsiveness by performing a quantitative ethylene dose-response analysis of growth in the transgenic lines (Fig. 3C).
Seedlings were grown in the dark in ethylene at concentrations ranging from 0 to 1,000 µL/L ethylene and the hypocotyl lengths measured after 4 d growth. The ethylene biosynthesis inhibitor aminoethylvinyl-Gly (AVG) was included in the growth media to inhibit endogenous ethylene production by the seedlings. In the absence of ethylene, both wild-type and transgenic lines exhibited a similar hypocotyl length, consistent with our earlier observation that both wildtype and kinase-inactive versions of ETR1 rescue the constitutive ethylene-response phenotype of etr1-9;ers1-3 ( Fig. 2A). The two transgenic lines of tETR1-wt exhibited ethylene responsiveness, indicating that the addition of the wild-type ETR1 restored ethylene responsiveness to the etr1-9;ers1-3 mutant line (Fig. 3C). Ethylene responsiveness of tETR1-wt was slightly less than that observed in native wild-type WS seedlings, which could be due to additional sequence not contained on the 7.3 kb genomic fragment used for transformation, the chromosome location of the transgenes, or the fact that the genetic background still lacks ERS1 and thus is not identical to wild-type WS. In contrast, transgenic lines containing kinase-inactive tETR1-G2, tETR1-H/G2, or tETR1-H/G2/D exhibited a substantially different ethylene dose response from those containing tETR1-wt (Fig. 3C). The kinase-inactive lines demonstrated a partial ethylene-insensitive phenotype observable at all concentrations from 0.1-1000 µL/L ethylene, with this response being most pronounced at 1-10 µL/L. Several additional points can be made from analysis of the dose response curves. First, the dose response curves are quite similar for the two independent lines examined for each transgene, in spite of differences in ETR1 protein levels, indicating that the effects of the mutations on the ethylene response phenotype outweigh potential effects of ETR1 expression. Second, the ethylene insensitivity exhibited by the tETR1-H/G2 and tETR1-H/G2/D lines was slightly greater than that of the tETR1-G2 lines (e.g. at 1 µL/L ethylene, the tETR1-wt lines averaged 46.2%, the tETR1-G2 lines 68.6%, the tETR1-H/G2 lines 78.7%, and the tETR1-H/G2/D lines 80.6% hypocotyl length compared to their untreated controls), suggesting that His353 may serve as a phosphorylation site in tETR1-G2, albeit at a reduced level compared to tETR1-wt. Third, the ethylene insensitivity exhibited by the tETR1-H/G2/D lines was similar to that observed in the tETR1-H/G2 lines (e.g. at 1 µL/L ethylene, the tETR1-H/G2/D lines averaged 80.6% and the tETR1 lines averaged 78.7% hypocotyl length compared to their untreated controls), suggesting that the maximal contribution of kinase activity to ETR1 signaling has been reached.
We also examined the ethylene responsiveness of seedlings grown in the light (Fig. 4).
For this purpose, seedlings were grown under continuous light for 7 d in the absence or presence of 10 µL/L ethylene. We did not observe any difference in shoot growth between the tETR1-wt and the kinase-inactive lines, whether grown in the absence or presence of ethylene (Fig. 4A). However, the root-growth response of the kinase-inactive lines to ethylene was reduced compared to that of the tETR1-wt lines, the kinase-inactive lines thus exhibiting reduced ethylene sensitivity for root growth (Fig. 4B). Overall, these phenotypic data support a role for histidine-kinase activity in the establishment of ethylene responses since kinase-inactive versions of ETR1 exhibit reduced ethylene sensitivity compared to wild-type ETR1.

Gene Expression Analysis of Seedlings with Kinase-Inactive ETR1
To gain information at the molecular level on how the ethylene response differed between kinase-active and kinase-inactive ethylene-receptor lines, we performed a microarray analysis. For this purpose, we used the tETR1-wt #2 and tETR1-H/G2 #2 lines, choosing these lines for comparison because they exhibit similar levels of ETR1 protein to each other as well as to the native level of ETR1 found in wild-type seedlings (Fig. 3B). RNA was prepared from seedlings grown in the dark in the presence or absence of 1 µL/L ethylene, we having observed maximal differences in the hypocotyl growth response under these growth conditions ( Fig. 3A, C). Samples were prepared in triplicate for analysis.
From the microarray analysis, we identified a group of 40 genes whose expression is induced 3-fold or more by ethylene in the tETR1-wt sample (Supplemental Table 1). So as to work with a robust set of ethylene-induced genes, independent of effects due to transgenes and experimental variation, we compared the genes identified in this dataset to those previously identified in wild-type seedlings grown under similar conditions for ethylene treatment (Alonso et al., 2003). Almost half (19) of the ethylene-induced genes from our microarray met the criterion of also being induced in this independent experiment (Table 1). We used box-plot analysis to visualize how expression of this group of ethylene-induced genes compares between tETR1-wt and tETR1-H/G2 (Fig. 5A). In both cases, we observe an induction of this gene set by ethylene. However, we observe differences in the expression levels of the genes that relate to the kinase activity. Most pronounced is a decrease in the basal level expression (minus exogenous ethylene) for the genes in the tETR1-H/G2 line compared to the tETR1-wt line. There is also a decrease under ethylene induction conditions (plus exogenous ethylene) for the genes in the tETR1-H/G2 background compared to tETR1-wt, although this difference is not as pronounced as that observed under basal conditions.
We performed qRT-PCR on a subset of the induced genes, examining expression in two independent lines each for tETR1-wt and tETR1-H/G2 (Fig. 5B). Some differences in ethyleneregulated gene expression are predicted between lines due to the differing protein levels for ETR1 ( Fig. 3B). tETR1-wt-line #1 and tETR1-H/G2-line #1 exhibit the higher protein levels, while tETR1-wt-line #2 and tETR1-H/G2-line #2 exhibit the lower protein levels, being similar to the native ETR1 protein level (Fig. 3B). Because ETR1 serves as a negative regulator of ethylene responses, the lines with higher ETR1 protein levels are predicted to more strongly suppress ethylene-responsive gene expression. In general, this prediction is born out by the qRT-PCR analysis. For example, the basal expression level for ethylene-regulated genes is generally lower in tETR1-wt-line #1 than in tETR1-wt-line #2 (four of the five genes exhibiting lower expression and one exhibiting similar expression). In addition, the expression level after ethylene induction is lower in tETR1-wt-line #1 compared to tETR1-wt-line #2 for four of the five genes examined.
Several points can be made based on comparison of the tETR1-wt to the kinase-inactive tETR1-H/G2 lines. First, the qRT-PCR analysis ( Fig. 5B)  exhibit reduced ethylene induction in the kinase-inactive lines. We thus observe reduced expression of ethylene-induced genes in the tETR1-H/G2 lines, particularly under the basal expression conditions, providing a potential molecular basis for the reduced phenotypic response to ethylene observed in these kinase-inactive lines.
Using a similar strategy employed for analyzing ethylene-induced genes, we also examined genes whose expression is repressed by ethylene. In this case, of the 151 genes whose expression was repressed 3-fold or more by ethylene in the tETR1-wt sample (Supplemental Thus, overall, analysis of the ethylene-repressed genes is consistent with our analysis of the ethylene-induced genes, with various genes in the kinase-inactive tETR1-H/G2 lines demonstrating a decreased ability to induce the ethylene response at the molecular level.

Analysis of CTR1 Protein Levels
The Raf-like kinase CTR1 serves as a negative regulator of ethylene responses and physically associates with the ethylene receptors to suppress ethylene signal transduction (Clark et al., 1998;Cancel and Larsen, 2002;Gao et al., 2003). The subfamily-1 mutant etr1-9;ers1-3 employed as the genetic background for our studies results in a substantial loss of the membraneassociated CTR1 (maximal CTR1 protein levels at the membrane is approximately 35% of that found in wild-type) (Qu et al., 2007), accounting in part for the mutant's constitutive ethyleneresponse phenotype. Based on the role of CTR1 in suppressing ethylene responses, one potential explanation for the reduced ethylene sensitivity observed in the kinase-inactive ETR1 lines would be if kinase-inactive ETR1 localized greater levels of CTR1 to the membrane than wildtype ETR1. We therefore immunologically characterized the levels of membrane-associated CTR1 in the different lines (Fig. 3B). The null mutant ctr1-2 served as a negative control for expression of CTR1 (Kieber et al., 1993;Gao et al., 2003). All the transgenic lines exhibited CTR1 protein levels similar to or greater than those found in wild-type, consistent with the ability of these lines to suppress the constitutive-ethylene-response phenotype observed in the etr1-9;ers1-3 background. Significantly, the lines containing kinase-inactive versions of ETR1 did not exhibit any more CTR1 protein than tETR1-wt lines.
Previous work has demonstrated that CTR1 protein levels are not constant in wild-type Arabidopsis membranes but increase in response to ethylene (Gao et al., 2003), potentially owed to short-term transcriptional induction of CTR1 (Shakeel and Schaller, unpublished data) (Winter et al., 2007). We therefore followed up on our initial analysis of CTR1 protein by quantitatively characterizing the levels of membrane-associated CTR1 in etiolated seedlings of tETR1-wt and tETR1-H/G2 grown in the absence or presence of 1 µL/L ethylene (Fig. 7), the same ethylenetreatment conditions used for gene expression analysis. In both the ETR1-wt and ETR1-H/G2 lines, CTR1 is present and as in wild type exhibits an ethylene-induced increase in protein levels.
However, similar to what we observed in our analysis of ethylene-induced gene expression (Table 1, Fig. 5), we observed a slightly decreased level of CTR1 protein in the ETR1-H/G2 lines compared to ETR1-wt in the absence of ethylene. Perhaps most significantly, the degree to which CTR1 levels were induced by exogenous ethylene was substantially less in the tETR1-H/G2 lines than in wild-type or the tETR1-wt lines. Immunological examination of ETR1 in the lines reveals a small ethylene-dependent increase in protein levels but this does not correlate with the differences in CTR1 induction (Fig. 7). Thus, differences in CTR1 protein levels in the ETR1-H/G2 lines compared to ETR1-wt lines cannot directly account for the difference in phenotypes, but rather seem to reflect the differences in ethylene sensitivity. A key conclusion from our study, consistent with the earlier study by Wang et al. (2003), is that histidine-kinase activity is not absolutely required for signaling by the receptors, either for repressing ethylene responses in air or for establishing ethylene responses upon ethylene binding. transgenic lines reported here. Thus the rescue of the double-mutant phenotype is consistent with both wild-type and kinase-inactive versions of ETR1 reforming a functional signaling complex with CTR1, which is then able to repress the constitutive ethylene-response phenotype of the etr1-9;ers1-3 mutant in air. Significantly, kinase-inactive ETR1 can also induce an ethylene response based on phenotypic and gene expression analyses. This indicates that ETR1 has an alternative mechanism, not involving its histidine-kinase activity, by which to regulate the kinase activity of CTR1. The physical interaction of CTR1 with the receptors (Clark et al., 1998;Cancel and Larsen, 2002;Gao et al., 2003) suggests that its regulation could be accomplished due to conformational changes in the receptors being transmitted to the associated CTR1.
Although the histidine-kinase is not absolutely required for signaling by ETR1, our study does demonstrate that this enzymatic activity modulates signaling. Kinase-inactive ETR1 was less effective than wild-type ETR1 at inducing an ethylene response, based on phenotypic analysis of dark-and light-grown seedlings as well as the molecular analysis of ethyleneregulated gene expression. A concern with any genetic study involving site-directed mutations is whether the phenotype is owed to the known effect of the site-directed mutations or due to an unintended side effect. To minimize the possibility of side effects, we made relatively conserved changes in the amino acids in which we incorporated established modifications based on work with bacterial and fungal two-component signaling elements (Iuchi, 1993;Pan et al., 1993;Yang and Inouye, 1993;Posas et al., 1996). In addition, we made use of three mutant versions of ETR1, the effects of these mutations being consistent with a shared mechanistic role in a His-Asp phosphorelay. Specifically, we found that the reduced ethylene sensitivity observed in ETR1-G2 is further accentuated in the ETR1-H/G2 mutant suggesting that His353 and the G2 box, even though separate in the primary sequence, play a role in the same signaling mechanism.
Furthermore, the Asp mutation had no additive effect, consistent with its role being downstream and dependent on the histidine-kinase activity of ETR1.
When expressed in the etr1-7;etr2-3;ein4-4 background, kinase-inactive ETR1 was found to be slightly less effective than wild-type ETR1 in rescuing the constitutive-ethylene response phenotype of the background (Qu and Schaller, 2004 Two possibilities, not mutually exclusive, could account for phosphorylation playing a modulating role in ethylene signal transduction (Mason and Schaller, 2005). First, ETR1 could transmit a signal through a multi-step phosphorelay involving downstream phosphotransfer proteins (AHPs) and response regulators (ARRs); this alternative and secondary pathway could augment output from the primary CTR1-dependent pathway. Support for such an alternative ethylene-signaling pathway comes from evidence that the ethylene receptors interact with AHP proteins (Urao et al., 2000;Scharein et al., 2008), that the response regulator ARR2 may modulate ethylene signaling (Hass et al., 2004;Mason et al., 2005), and that mutants of CTR1 still demonstrate a residual ethylene response (Larsen and Chang, 2001;Hall and Bleecker, 2003).
A second possibility is that phosphorylation of ETR1 affects signaling through wellestablished components of the pathway, a possibility consistent with the broad effect on transcription of ethylene-regulated genes in the kinase-inactive mutant. Phosphorylation is a commonly used mechanism to elicit conformational changes in proteins as well as to modulate interactions between proteins, and therefore could, for instance, modulate the conformational information passed between ETR1 and the physically associated CTR1. Consistent with such a possibility is the finding that both the histidine-kinase and receiver domains of the ethylene receptors associate with CTR1 (Clark et al., 1998;Cancel and Larsen, 2002;Gao et al., 2003;Zhong et al., 2008). Based on our data the kinase-inactive ETR1 is more effective at suppressing the ethylene response than wild-type ETR1 and thus, according to this model, autophosphorylation of ETR1 would serve as a part of the means by which CTR1 is inactivated (i.e. in the kinase-inactive ETR1 lines, CTR1 is more active and thus it is more difficult to induce an ethylene response). Interestingly, we observed lower levels of CTR1 associated with the kinase-inactive ETR1 lines than with the wild-type ETR1 lines, even though these lines were more effective at suppressing the ethylene response. The reduced levels of CTR1 in the kinaseinactive lines may reflect their reduced ethylene responsiveness, because one of the ethylene responses is to induce increased levels of membrane-associated CTR1 (Gao et al., 2003).
Alternatively, the reduced CTR1 levels may reflect part of a feedback mechanism to modulate output from the receptors, we having observed that CTR1 levels do not always directly correlate with the total receptor levels, actually increasing in the etr1-7 and etr1-9 null mutants (Gao et al., 2003;Qu et al., 2007). The simplest mechanistic model, consistent with our data indicating that autophosphorylation serves a role in establishing an ethylene response, is that ethylene stimulates the histidine-kinase activity and autophosphorylation of ETR1. The kinase-inactive ETR1would thus be less effective at initiating an ethylene response, resulting in the reduced sensitivity to ethylene in our growth response assays. Similarly, we also observed a reduced ability by the kinase-inactive ETR1 to regulate gene expression in response to ethylene, particularly when considered on a per protein basis with wild-type ETR1. In some cases, the effects on gene expression were strong enough that the kinase-inactive ETR1 lines differed from the wild-type ETR1 lines, regardless of the ETR1 protein level. This effect was most apparent when examining the basal expression levels of ethylene-induced genes, which were consistently lower in the kinase-inactive ETR1 lines. Only endogenous ethylene is present under basal expression conditions, some ethylene being produced by seedlings even when the ethylene biosynthesis inhibitor AVG is incorporated into the growth media (Sanders et al., 1991). This low level of ethylene is apparently sufficient to induce expression of the genes in the wild-type ETR1 lines to a greater extent than in the kinase-inactive ETR1 lines. A more substantial role for autophosphorylation at low ethylene levels is consistent with our growth response curves (

Plant Growth Conditions and Ethylene Response Assays
Treatment and analysis of the triple response of dark-grown Arabidopsis seedlings to ethylene (Chen and Bleecker, 1995;Roman and Ecker, 1995)

Immunoblot Analysis
Immunoblot analysis was performed using microsomal fractions isolated from darkgrown Arabidopsis seedlings essentially as described (Gao et al., 2008). Briefly, plant material was homogenized in a buffer containing 30 mM Tris (pH 8.3 at 4°C), 150 mM NaCl, 1 mM EDTA, and 20% (v/v) glycerol with protease inhibitors (Sigma Aldrich, P9599) and then centrifuged at 8,000 g for 15 min (Chen et al., 2002;Zhao et al., 2002). The supernatant was then centrifuged at 100,000 g for 30 min, and the resulting membrane pellet resuspended in 10 mM Tris (pH 7.6 at 22°C), 150 mM NaCl, 0.1mM EDTA, and 10% (v/v) glycerol with protease inhibitors. Protein concentration was determined by use of the BCA reagent (Pierce) according to the manufacturer after first adding 0.1 ml 0.5% (w/v) sodium dodecyl sulfate to solubilize membrane proteins. Bovine serum albumin was used as a standard for protein assays. ETR1 was identified by use of a polyclonal anti-ETR1 antibody generated against amino acids 401-738 of ETR1 (Chen et al., 2002). CTR1 was identified by use of a polyclonal anti-CTR1 antibody (Gao et al., 2003). An anti-BiP antibody was used as a loading control (Stressgen Biotech, #SPA-818E). Relative expression levels for ETR1 and CTR1 were determined using the program ImageJ version 1.38x (http://rsbweb.nih.gov/ij/), with quantification of the scanned exposed film from the immunoblots being made by comparison to a dilution series (Zhao et al., 2002). imported into GeneSpring GX 11.5 (Agilent Technologies, Palo Alto, CA) for data analysis. The data were normalized by baseline transformation, which is equivalent to per chip to the 50 th percentile and per gene to the median. Analysis of these data indicate that only 1.64% of the probe sets vary in expression by two-fold or greater when comparing tETR1-wt to tETR1-H/G2, and that ethylene treatment of tETR1-wt results in a two-fold or greater change in expression for 3.00 % of the 22,810 probe sets present on the array, consistent with the majority of the genes not being differentially expressed under the experimental conditions. The raw data was filtered by expression, with the requirement that at least one sample out of six for a line exceed a lower cut-off percentile of 20%. The data were statistically analyzed using an unpaired T-test, and data accepted with p-values less than or equal to 0.05. For interpretation of data, the Genespring foldchange mode was used and putative ethylene-regulated genes identified based on their exhibiting a three-fold change in the ethylene-treated sample compared to the control sample. The complete dataset is deposited in Array-Express (http://www.ebi.ac.uk/arrayexpress) with accession number E-MEXP-3574 and is also available as Supplemental Table S3.

Gene Expression Analysis
Real-time PCR was performed as described (Argyros et al., 2008), using primer sets specific to genes for pEARL1-like (At4g12470)     Note early developmental arrest of etr1-9;ers1-3 mutant, which is rescued by introduction of the kinase-deficient versions of ETR1. Scale-bars = 2 mm.     and etr1-9 null mutants serve as negative controls for the expression of CTR1 and ETR1,