ETR1-Specific Mutations Distinguish ETR1 from Other Arabidopsis Ethylene Receptors as Revealed by Genetic Interaction with RTE1^,,

The plant growth regulator ethylene is perceived by a family of homologous receptors that negatively regulate ethylene responses. It is well established that dominant missense mutations within the ethylene-binding domain of any of these receptors result in ethylene insensitivity (Chang et al., 1993; Hua et al., 1995, 1998; Wilkinson et al., 1995; Sakai et al., 1998; Hall et al., 1999; Wang et al., 2006). Furthermore, the deliberate introduction of such missense mutations is known to convert other ethylene receptor genes into dominant mutant forms that confer ethylene insensitivity (Hua et al., 1995, 1998; Hall et al., 1999; Terajima et al., 2001; Shaw et al., 2002; Cui et al., 2004). Here, we present a class of dominant mutations that does not fit this paradigm. That is, certain etr1 ethylene-insensitive mutations in Arabidopsis (Arabidopsis thaliana) fail to confer ethylene insensitivity when introduced at identical positions in the closely related ERS1 ethylene receptor gene, despite the high level of sequence conservation between ERS1 and ETR1. What distinguishes these nontransferable dominant etr1 mutations from the transferable ones is that the former are suppressed by mutations in RTE1, a gene that affects ETR1 signaling (Resnick et al., 2006, 2008).

Ethylene receptors have a membrane-bound N-terminal ethylene-binding domain followed by a cytosolic GAF domain (Hall et al., 2007). At the C terminus is a signaling output domain that has similarity to the His protein kinase module of the two-component signaling system prevalent in prokaryotes. Some ethylene receptors also carry a conserved C-terminal receiver domain, the second component of the two-component system. The biochemical mechanism of ethylene receptor signaling, however, is unknown (Hall et al., 2007). The ethylene receptors fall into two subfamilies: subfamily I contains three N-terminal transmembrane domains and a conserved His kinase domain, whereas subfamily II has four N-terminal transmembrane domains and a degenerate His kinase domain. Despite these differences, the receptors appear to function similarly as negative regulators of ethylene responses; genetic evidence indicates that the receptors signal to repress ethylene responses in the absence of ethylene binding, and when ethylene is bound, receptor signaling is turned off, allowing responses to proceed (Hall et al., 2007). This switch most likely involves a conformational change induced by ethylene binding that is transmitted to the signaling output domain of the receptor (Wang et al., 2006). This model is supported by dominant gain-of-function mutations that lock the signaling domain into the “on” state (thereby repressing ethylene responses), as all known gain-of-function mutations lie within the N-terminal region containing the ethylene-binding domain (with the exception of ers1-10, which lies within the GAF domain; Wang et al., 2006). Some but not all of these mutations block ethylene binding (Hall et al., 1999; Wang et al., 2006).

Arabidopsis has five ethylene receptors, and for each there exist dominant mutant forms conferring ethylene insensitivity (Chang et al., 1993; Hua et al., 1995, 1998; Sakai et al., 1998). Few functional differences between the five receptors have been identified, except that the subfamily I receptors (ETR1 and ERS1) play a more prominent role in ethylene signaling than subfamily II receptors (ETR2, EIN4, and ERS2; Hua and Meyerowitz, 1998; Cancel and Larsen, 2002; Qu et al., 2007). In addition, subfamily I receptors possess His autokinase activity in vitro (Gamble et al., 1998), whereas subfamily II receptors have Ser/Thr kinase activity in vitro (Moussatche and Klee, 2004); ERS1 displays both activities (Moussatche and Klee, 2004). Neither of these phosphorylation activities appears to play a substantial role in ethylene receptor signaling, however (Gamble et al., 2002; Wang et al., 2003; Moussatche and Klee, 2004).

The Arabidopsis REVERSION-TO-ETHYLENE SENSITIVITY1 (RTE1) gene is a positive regulator of ETR1 that was identified in a genetic screen for suppressors of the dominant ethylene-insensitive mutation etr1-2 (Resnick et al., 2006). RTE1 encodes a novel integral membrane protein of unknown function in plants, animals, and some protists. The RTE1 protein colocalizes with ETR1 at both the endoplasmic reticulum and the Golgi apparatus in Arabidopsis (Dong et al., 2008). Loss-of-function mutations, including the strong allele rte1-2 and the null allele rte1-3, confer ethylene hypersensitivity identical to that in the etr1 null mutant, whereas overexpression of RTE1 confers ethylene insensitivity (Resnick et al., 2006). Overexpression of GREEN-RIPE, a tomato (Solanum lycopersicum) RTE1 homolog, similarly results in ethylene-insensitive phenotypes (Barry and Giovannoni, 2006). The insensitivity conferred by RTE1 overexpression is largely dependent on ETR1 and not on the other ethylene receptor genes (Resnick et al., 2006; Zhou et al., 2007; J.S. Resnick and C. Chang, unpublished data), suggesting that RTE1 is specific to ETR1. rte1 is capable of suppressing additional dominant etr1 alleles besides etr1-2, but interestingly, this ability was found to be allele specific (Resnick et al., 2008). Out of 13 etr1 dominant mutant alleles tested, seven were suppressed by rte1-2 (Resnick et al., 2008). All 13 mutations result in amino acid substitutions in the N-terminal domain of the ETR1 receptor, preventing ETR1 signaling from being turned “off.” There was no correlation between ethylene-binding ability and suppression by rte1-2 (Resnick et al., 2008). RTE1 is proposed to promote the signaling “on” state of ETR1 by acting on the ETR1 N-terminal domain (Zhou et al., 2007; Resnick et al., 2008).

In the work presented here, we wanted to determine more definitively whether RTE1 action is specific to ETR1 or whether RTE1 could similarly affect other ethylene receptors in Arabidopsis. Our approach was to test rte1-2 for the ability to suppress dominant mutations in ethylene receptor genes other than ETR1. Previously, we had tested single gain-of-function alleles for each of the other Arabidopsis ethylene receptor genes (ers1-10, ein4-1, etr2-1, and ers2-2) and found that they were not suppressed by rte1-2 (Resnick et al., 2006). However, we could not rule out the possibility that this result was a function of the particular mutations tested, given our finding that only certain etr1 alleles are suppressed by rte1. Therefore, we set out to reexamine this question using mutations known to be RTE1 dependent when carried by etr1.

We focused on ERS1, the only other Arabidopsis ethylene receptor in the same subfamily as ETR1 (subfamily I). Among Arabidopsis ethylene receptors, ERS1 has the most closely related ethylene-binding domain to that of ETR1 (75% identity, 83% similarity). We tested five amino acid substitutions that are known to confer dominant ethylene insensitivity when present in the ETR1 receptor (Table I); four (Y32A, E38A, F58A, and A102T) are dependent on RTE1 for ethylene insensitivity, and one (C65Y) is RTE1 independent. Each of the substitutions lies within a highly conserved region in one of the three predicted transmembrane domains of the ethylene-binding domain (Fig. 1).

Table I.

RTE1-dependent ethylene-insensitive etr1 mutations do not confer insensitivity when carried by an ERS1 ethylene receptor transgene in Arabidopsis

Amino Acid Substitution in the Ters1 Transgene	Ethylene Insensitivity Conferred by the Ters1 Mutant Transgene		Equivalent etr1 Mutation Suppressed by rte1-2?c
	in the Wild-Type Background?a	in rte1-2?b
C65Y	Yes (88%)	Yes (82%)	No (93%)
Y32A	No (37%)	n/a	Yes (42%)
E38A	No (38%)	n/a	Yes (45%)
F58A	No (39%)	n/a	Yes (46%)
A102T	No (35%)	n/a	Yes (36%)

^aValues are average percentage hypocotyl lengths of two independent representative transgenic lines (driven by the native ETR1 promoter) on 10 μm 1-aminocyclopropane-1-carboxylic acid with respect to that on 0 μm 1-aminocyclopropane-1-carboxylic acid in the wild-type background. Hypocotyl length was measured using 12 to 15 homozygous seedlings per transgenic line.

^bValue is the average percentage hypocotyl length of two independent representative transgenic lines in the rte1-2 background with respect to that in the wild-type background on 10 μm 1-aminocyclopropane-1-carboxylic acid. n/a, Not applicable for lines showing no insensitivity in the wild-type background. Hypocotyl length was measured using 12 to 15 homozygous seedlings per transgenic line.

^cValues are percentage hypocotyl lengths of a representative transgenic line in the rte1-2 background with respect to that in the wild-type background on 20 μm 1-aminocyclopropane-1-carboxylic acid. Data are from Resnick et al. (2008).

Figure 1.

Sequence alignment of the ethylene-binding region of the five Arabidopsis ethylene receptors showing the amino acid substitutions examined in this study. Amino acids that were substituted are shown in boldface within the sequences, with the corresponding substitutions indicated above the alignment. Asterisks denote conserved residues, colons denote conserved substitutions, and periods denote semiconserved substitutions aligned using ClustalW (Chenna et al., 2003). The approximate positions of the three predicted transmembrane domains (I, II, and III) are underlined.

Mutations encoding these amino acid substitutions were introduced to the corresponding conserved positions in the ERS1 coding sequence using in vitro site-directed mutagenesis (Supplemental Materials and Methods S1). In order to control for possible differences in expression between ETR1 and ERS1, two separate ers1 transgene (Ters1) constructs were created for each mutation, one driven by the native ERS1 promoter and the other driven by the native ETR1 promoter. Each construct was stably transformed into wild-type and rte1-2 mutant plants, and six transformed lines were analyzed for each construct in each genetic background. Ethylene insensitivity was assessed using the classic triple-response assay in dark-grown seedlings (Guzmán and Ecker, 1990). Interestingly, none of the Ters1 lines carrying Y32A, E38A, F58A and A102T in the wild-type background (nor in the rte1-2 background) exhibited ethylene insensitivity (Fig. 2; Table I). All of these mutations require RTE1 in order to confer dominant ethylene insensitivity when carried by etr1, unlike the C65Y mutation, which does not require RTE1. The absence of ethylene insensitivity was unexpected, since other dominant mutations are known to confer ethylene insensitivity when engineered into various ethylene receptor genes (Hua et al., 1995, 1998). Indeed, a different result was obtained for the C65Y mutation (which is not suppressed by rte1-2 when carried by the etr1 gene). All 12 lines carrying Ters1 (C65Y) showed strong ethylene insensitivity in the wild-type background (Fig. 2), and there was no detectable difference between transgenes expressed under the control of the ETR1 or ERS1 promoter (data not shown). Similar insensitivity was observed when Ters1 (C65Y) was expressed in the rte1-2 mutant background, indicating that rte1-2 cannot suppress Ters1 (C65Y; Fig. 2; Table I). These results suggest a correlation between RTE1 dependence and the inability to confer ethylene insensitivity when carried by the ers1 transgene, since all four etr1 mutations that require RTE1 in order to confer ethylene insensitivity did not cause insensitivity when carried by ers1.

Figure 2.

The seedling triple-response assay shows the presence or absence of ethylene insensitivity conferred by ethylene receptor transgenes carrying various amino acid substitutions. Wild-type Arabidopsis was transformed with the transgenes Tetr1 and Ters1 (expressed under the control of the native ETR1 promoter) and Tein4 (expressed under the control of the cauliflower mosaic virus 35S promoter) carrying the specified substitutions (Y32A, E38A, F58A, A102T, and C65Y). Shown are representative dark-grown 4-d-old seedlings germinated in the presence of 10 μm 1-aminocyclopropane-1-carboxylic acid, an ethylene precursor. The untransformed wild type (WT) displays the triple-response phenotype (inhibition of hypocotyl and root elongation and exaggeration of the apical hook) in response to 10 μm 1-aminocyclopropane-1-carboxylic acid. [See online article for color version of this figure.]

We next tested EIN4, a subfamily II ethylene receptor. EIN4, like ETR1, possesses a C-terminal receiver domain, which ERS1 lacks. We introduced a mutation coding for A125T (equivalent to ETR1 A102T) into the EIN4 cDNA sequence driven by the cauliflower mosaic virus 35S promoter (Supplemental Materials and Methods S1). Consistent with the results for ERS1, wild-type plants transformed with the Tein4 (A125T) transgene failed to display ethylene insensitivity (Fig. 2).

These findings suggest that etr1 missense mutations requiring RTE1 to confer ethylene insensitivity do not confer the same insensitivity when introduced into the corresponding conserved positions in other Arabidopsis ethylene receptor genes. If this hypothesis is correct, then existing ethylene-insensitive mutations in the four other ethylene receptor genes should not be suppressed by rte1 when carried by the etr1 gene. We tested this using two amino acid substitutions, I62F and P36L (numbered here based on the ETR1 sequence). The corresponding I62F substitution is encoded by an existing mutant allele in four different ethylene receptor genes: etr1-4 (Chang et al., 1993), ers1-1 (Hua et al., 1995), ein4-1 (Hua et al., 1998), and ers2-2 (Hua et al., 1998; Fig. 1). We have shown previously that rte1-2 does not suppress ein4-1 and ers2-2 (Resnick et al., 2006), and here we found that rte1-2 does not suppress ers1-1 (Table II; Supplemental Materials and Methods S1). Therefore, we predicted that etr1-4 would not be suppressed by rte1. To test this, we created an etr1 transgene encoding the corresponding I62F substitution (driven by the native ETR1 promoter), followed by transformation into wild-type and rte1-3 (null mutant) plants (Supplemental Materials and Methods S1). As predicted, the Tetr1 (I62F) transgene conferred ethylene insensitivity that was not suppressed by rte1-3 (Table II). The second mutation that we analyzed, encoding the P36L substitution, is identical to that encoded by etr2-1 (Sakai et al., 1998). We used an existing etr1 transgene encoding P36L driven by the native ETR1 promoter (Wang et al., 2006) and transformed it into both the wild type and the rte1-3 mutant to test for ethylene insensitivity and suppression, respectively. Similar to the results for Tetr1 (I62F), the Tetr1 (P36L) transgene conferred ethylene insensitivity that was not suppressed by rte1-3 (Table II). These results are consistent with the prediction that existing mutations in the other ethylene receptor genes would not be RTE1 dependent when carried by etr1.

Table II.

Ethylene-insensitive mutations in each Arabidopsis ethylene receptor gene are RTE1 independent when carried by ETR1

Amino Acid Substitution	Equivalent Ethylene-Insensitive Mutant Allele or Transgenea	Suppressed by rte1?
I → F	ers1-1 (I62F)	No (84%)b
	ein4-1 (I84F)	No (96%)bc
	ers2-2 (I94F)	No (101%)bc
	Tetr1 (I62F)	No (94%)d
P → L	etr2-1 (P66L)	No (98%)bc
	Tetr1 (P36L)	No (106%)d

^aConserved positions of corresponding amino acid substitutions are given in parentheses for each allele. Each allele has been previously shown to confer ethylene insensitivity as measured by hypocotyl length (Hua et al., 1995; Hall et al., 1999; Resnick et al., 2006; Wang et al., 2006).

^bValues are percentage hypocotyl lengths for the double mutant (carrying both the receptor mutant allele and rte1-2) with respect to that of the receptor mutant allele alone on 20 μm 1-aminocyclopropane-1-carboxylic acid.

^cData are from Resnick et al. (2006).

^dValues are average percentage hypocotyl lengths of two independent representative transgenic lines in the rte1-3 background with respect to that in the wild-type background on 20 μm 1-aminocyclopropane-1-carboxylic acid.

In conclusion, not all dominant ethylene-insensitive alleles of etr1 are transferable to other ethylene receptor genes. The data suggest that dominant etr1 mutations that require RTE1 to confer ethylene insensitivity are essentially silent when transferred to the identical conserved positions in other ethylene receptor genes. It remains to be seen whether this extends to all of the subfamily II receptors and whether other known RTE1-dependent etr1 alleles behave similarly. Nevertheless, our results are unexpected given the strong sequence conservation among the ethylene receptors, particularly within the ethylene-binding domain. Moreover, several dominant missense mutations that confer ethylene insensitivity have been previously transferred to or identified in the corresponding positions in other ethylene receptor isoforms even between different species. For example, this is how the Arabidopsis ERS1 and ERS2 genes were shown to encode functional ethylene receptors in the absence of endogenous mutant alleles for these genes: the etr1-4 (I62F) mutation was introduced to an ERS1 transgene (Hua et al., 1995), and the etr1-4 (I62F) and etr2-1 (P66L) mutations were each introduced to an ERS2 transgene (Hua et al., 1998). Similarly, the ethylene-insensitive mutation of tomato Never-ripe encoding P36L (a substitution identical to that of etr2-1) was introduced into the tobacco (Nicotiana tabacum) NtERS1 ethylene receptor gene to create ethylene-insensitive tobacco plants (Terajima et al., 2001), and a mutation encoding I62F was introduced into a Brassica oleracea ERS1 homolog to delay flower senescence in petunia (Petunia hybrida; Shaw et al., 2002). In another example, a mutation encoding H69A, which confers ethylene insensitivity in Arabidopsis etr1 (Hall et al., 1999; Wang et al., 2006), was engineered into melon (Cucumis melo) CmETR1 to produce ethylene insensitivity in Nemesia strumosa (Cui et al., 2004). Thus, it had been generally accepted that dominant mutations identified in one ethylene receptor gene could be introduced into another ethylene receptor gene of the same or different species in order to create a dominant mutant transgene that would confer ethylene insensitivity in wild-type plants.

Our findings also demonstrate that RTE1 is specific for ETR1. It is unknown why the RTE1 mechanism that promotes signaling in ETR1 lacks this role for ERS1 (and probably for the three other ethylene receptors as well). The specificity of RTE1 does not appear to be based on differences between ETR1 and ERS1 expression patterns, as the same results were obtained using either the ERS1 or ETR1 promoter to drive the expression of the mutant transgenes. The RTE1 protein may act to maintain or stabilize active conformations of certain ETR1 dominant mutant forms (Resnick et al., 2008). RTE1-dependent etr1 mutations presumably lead to an altered conformation of the ETR1 ethylene-binding domain that confers ethylene insensitivity through constitutive signaling by the ETR1 signaling domain. Conceivably, the same substitutions in ERS1 (and EIN4) do not result in the same altered conformation and/or these other receptors lack the action of RTE1 to bring about ethylene insensitivity. This reveals ETR1 to be distinct from the other Arabidopsis ethylene receptors, possibly in terms of its steric structure. These distinctions apply not only to the dominant mutant forms but to the wild-type receptors as well, since wild-type ETR1 is also dependent on RTE1. By creating chimeras between ETR1 and ERS1, it might be possible to determine which regions of ETR1 are responsible for RTE1 dependence.

The ETR1-RTE1 genetic interaction may have coevolved to distinguish ETR1 from the other ethylene receptors. The inability of the ers1 and ein4 mutant transgenes to confer insensitivity suggests that ERS1 and EIN4 signaling is not only independent of RTE1 but independent of an RTE1-like protein. If a protein similar to RTE1 were acting on ERS1 and EIN4, then we should have seen ethylene insensitivity for all the dominant mutations carried by Ters1 and Tein4. The Arabidopsis genome carries a second copy of RTE1, called RTH, which does not appear to play a role in ethylene signaling (M. Rivarola and C. Chang, unpublished data). In plants other than Arabidopsis, it might be possible to determine which ethylene receptors, if any, are dependent on RTE1 orthologs by introducing RTE1-dependent mutations into the receptors and testing them for the ability to confer ethylene insensitivity, as in this paper. These findings have implications for practical applications that involve the engineering of dominant mutations into heterologous ethylene receptor genes for the purpose of generating ethylene insensitivity in plants.

Supplemental Data

The following materials are available in the online version of this article.

• Supplemental Materials and Methods S1. Description of plant strains and growth conditions; transgene constructs and plant transformation; double mutant construction; measurement of hypocotyl length.