Canonical histidine kinase activity of the transmitter domain of the ETR1 ethylene receptor from Arabidopsis is not required for signal transmission

Abstract

Ethylene signaling in plants is mediated by a family of receptors related to bacterial two-component histidine kinases. Of the five members of the Arabidopsis ethylene receptor family, members of subfamily I (ETR1 and ERS1) contain completely conserved histidine kinase domains, whereas members of subfamily II (ETR2, EIN4, and ERS2) lack conserved residues thought to be necessary for kinase activity. To examine the role of the conserved histidine kinase domain in receptor signaling, ers1;etr1 loss-of-function double mutants were generated. The double mutants exhibited a severe constitutive ethylene response phenotype consistent with the negative regulator model for receptor function. The adult ers1-2;etr1-6 and ers1-2;etr1-7 phenotypes included miniature rosette size, delayed flowering, and both male and female sterility, whereas etiolated-seedling responses were less affected. Chimeric transgene constructs in which the ETR1 promoter was used to drive expression of cDNAs for each of the five receptor isoforms were transferred into the ers1-2;etr1-7 double-mutant plants. Subfamily I constructs restored normal growth, whereas subfamily II constructs failed to rescue the double mutant, providing evidence for a unique role for subfamily I in receptor signaling. However, transformation of either the ers1-2;etr1-6 or ers1-2;etr1-7 mutant with a kinase-inactivated ETR1 genomic clone also resulted in complete restoration of normal growth and ethylene responsiveness in the double-mutant background, leading to the conclusion that canonical histidine kinase activity by receptors is not required for ethylene receptor signaling.

Ethylene serves as an important signaling molecule in plants, both in regulating developmental processes and mediating responses to environmental signals (1). Genetic analysis of ethylene signaling in Arabidopsis revealed the presence of a small family of ethylene receptors that are related to the bacterial two-component histidine kinase superfamily of signaling molecules (2).

Although the five members of the ethylene receptor family from Arabidopsis share a high degree of sequence similarity, each has distinguishing characteristics. All members contain an N-terminal, membrane-associated sensor domain that shows high-affinity ethylene binding when expressed in yeast (3, 4). Additional studies with ETR1 indicated that ethylene binding is mediated through a copper cofactor (5). In all receptor isoforms, the ethylene sensor domain is followed by a domain showing varying degrees of sequence similarity to the histidine kinase catalytic domains characteristic of bacterial two-component regulators. The bacterial systems transduce signal via autophosphorylation of a histidine residue in the kinase transmitter domain, followed by transfer of phosphate to an aspartate residue in the receiver domain of a response regulator protein (6). The residues thought to be essential for histidine kinase activity are conserved in ETR1 and ERS1 (7, 8), but are not completely conserved in ETR2, EIN4, and ERS2 (9, 10). Based on these distinguishing features and overall sequence similarity, the members of the ethylene receptor family from Arabidopsis can be divided into two subfamilies: subfamily I, which includes ETR1 and ERS1, and subfamily II, which includes ETR2, EIN4, and ERS2.

The roles of ethylene receptor isoforms in signal transduction must be considered in the context of the prevailing model for the signal transduction pathway based on genetic studies in Arabidopsis (2, 11). According to this model, the five receptor isoforms interact with CTR1, a RAF-related kinase, to negatively regulate response pathways. This model accounts for the observations that (i) loss-of-function mutations in either CTR1 or combinations of receptor isoform genes leads to a constitutive response phenotype (12, 13) and (ii) dominant gain-of-function mutations in any of the five receptor isoforms confer global ethylene insensitivity in the plant, presumably by locking the receptor into an active signaling state. According to this model, ethylene acts as an inverse agonist by negatively regulating receptor-CTR1 signaling. The target of receptor-CTR1 signaling is thought to be the NRAMP-related EIN2 protein (14), null mutations in which cause global ethylene insensitivity, which is thought to act downstream of the ethylene receptors and CTR1 based on epistasis analysis (12).

It is of interest to consider the specific roles that subfamily I and subfamily II receptors play in receptor signaling. Recent studies demonstrated that all five receptor isoforms can bind ethylene with similar high affinity (F. I. Rodriguez and A.B.B., unpublished results). Lack of an ethylene-related phenotype in single loss-of-function receptor mutants indicates functional redundancy between receptor isoforms (13). In addition, the observation that triple-mutant lines containing loss-of-function mutations in various combinations of three receptor isoforms showed constitutive response phenotypes indicates that all members of the family contribute to ethylene-related signaling.

If transmission of the ethylene signal occurs through canonical two-component phosphotransfer, ETR1 and ERS1 could perform a specialized function given that only ETR1 and ERS1 possess fully conserved histidine kinase domains. The role of the histidine kinase activity in receptor signaling could not be fully addressed in previous studies because no loss-of-function ERS1 mutant had been identified. However, in support of such a role for subfamily I receptors, histidine kinase activity has been reported for the ETR1 kinase domain expressed in yeast (15). In addition, evidence for signaling through a canonical phosphorelay has been reported for other eukaryotic two-component systems, including the SLN1 osmosensor from yeast (16) and the CRE1 cytokinin receptor from Arabidopsis (17). On the other hand, the evidence for direct interaction of the ETR1 and ERS1 transmitter domains with the presumed regulatory domain of CTR1 (18) opens the possibility that ethylene receptor signaling occurs through a novel mechanism that does not involve a response regulator intermediate.

To further evaluate the role that histidine kinase-mediated phosphotransfer plays in ethylene receptor signaling, we report here on the isolation of ers1;etr1 double loss-of-function mutants that are deficient in histidine kinase-containing receptor isoforms. Analysis of these double mutants indicates that, although subfamily I members may play a special role in ethylene signal transmission, this role does not require canonical histidine kinase activity.

Materials and Methods

Screening of the ers1-2 Mutant.

The ers1-2 T-DNA insertion mutant was isolated from Wisconsin T-DNA-tagged Arabidopsis population (19). Pools of DNA from the population were screened by a PCR-based method using ERS1-specific primers and T-DNA left border-specific and right border-specific primers as described (ref. 19, www.biotech.wisc.edu/NewServicesAndResearch/Arabidopsis/default.htm). The primer sequences are provided in Supporting Text, which is published as supporting information on the PNAS web site, www.pnas.org.

Genetic Analysis.

The ers1-2 allele was isolated in a Wassilewskija (Ws) background and was backcrossed to WT Ws three times. To construct ers1;etr1 double mutants, the ers1-2 mutant was crossed to the etr1-6 and etr1-7 single mutants (13), respectively. For genotyping, DNA was isolated as described (20). PCR was performed to identify the ers1-2 allele with the T-DNA left-border primer and an ERS1-specific primer and the WT ERS1 allele with two gene-specific primers. For genotyping etr1 mutations, ETR1 DNA was amplified with two primers flanking the mutations. PCR products were purified by using the PCR Purification kit (Qiagen, Valencia, CA) and sequenced. The primers for genotyping and their sequences are provided in the Supporting Text.

Plasmid Constructions and in Vitro Site-Directed Mutagenesis.

The 5′ flanking region (1,448-bp fragment upstream of the start codon) and the 3′ flanking region (1,362-bp fragment downstream of the stop codon) of the genomic ETR1 were amplified by using PCR and subcloned into pBluescipt II SK⁻. The nucleotides AA just before the start codon and AG after the stop codon were changed into CC and GCTA, respectively, to introduce the NcoI and NdeI restriction sites. The PCR-amplified 2,217-bp cDNA coding region of ETR1 was inserted between the 5′ and 3′ flanking regions to place the ETR1 cDNA under control of its own promoter. This construct was designated as {cETR1}. The ETR1 cDNA coding region in {cETR1} was replaced with the cDNA coding regions of ERS1 (1,847 bp), ETR2 (2,322 bp), EIN4 (2,301 bp), and ERS2 (1,938 bp) to create the constructs {cERS1}, {cETR2}, {cEIN4}, and {cERS2}. A few single nucleotide mutations were introduced to eliminate restriction sites to facilitate cloning without changing their amino acid sequences. These restriction sites were also used to distinguish transgenic and endogenous transcripts of the receptors in RT-PCR analysis (see below). Site-directed mutagenesis was performed by using the QuikChange Site-Directed Mutagenesis kit (Stratagene). These constructs were subsequently cloned into the pPZP221 binary vector for plant transformation (21).

The genomic ETR1 (gETR1, 7,266 bp) was subcloned into pBluescript II SK⁻ (7). To generate the histidine kinase-deficient ETR1 construct getr1-[HGG], H353, G515, and G517 were mutated into Q353, A515, and A517, respectively, in the genomic ETR1 in pBluescript II SK⁻. To eliminate any undesired mutations, a BstXI–StyI fragment in the genomic ETR1 was replaced with the BstXI–StyI fragment containing the mutations. The WT and mutant genomic ETR1 fragments were then removed from pBluescript II SK⁻ and inserted into pPZP221. All of the constructs and mutations were confirmed by sequencing.

Generation of Transgenic Lines.

The constructs were transformed into Agrobacterium tumefaciens strain ABI and transferred into Arabidopsis plants by using the floral dipping method (22). Plants heterozygous for ers1-2 and homozygous for etr1-6 or etr1-7 were used for transformation because the ers1;etr1 double loss-of-function mutants are sterile. T₁ seeds were plated on 1/2 MSNS (1/2 strength of Murashige and Skoog salts without sucrose) agar plates containing 100 μg/ml gentamycin, and resistant seedlings were transplanted to soil. T₁ and T₂ seedlings were genotyped for ers1-2 to identify the homozygous double mutants. Multiple transgenic lines were obtained for each construct. The genetic background and the inheritance of transgenes were genotyped by sequencing (for the etr1-6, etr1-7, and transgenes), PCR (for the ers1-2 allele), and analyzing the segregation of kanamycin resistance (for the ers1-2 allele) and gentamycin resistance (for transgenes).

RT-PCR.

RT-PCR was carried out to analyze the transcript of ERS1 in the ers1-2 mutant by using the primers specific to the left border of T-DNA and ERS1 (Fig. 1). To determine the expression level of transgenes, seedlings of transgenic lines were germinated and grown on 1/2 MSNS agar plates for 8 days, and then transplanted into soil. Plants were grown in a 16-h day length in a growth chamber. Leaves were collected from plants grown for 5 weeks. Total RNA was extracted by using a RNeasy RNA Isolation kit (Qiagen). cDNA synthesis was carried out by using the Oligo(dT)₁₈+3N primer and the SuperScript First-Strand cDNA Synthesis System (Invitrogen). PCR was performed to determine the expression levels of the receptors in the transgenic lines. The primer pairs for each receptor in the RT-PCR analyses flanked an intron, and the length of introns was <15% of the length of PCR products. The primers for ETR1 and ERS1 flanked the second intron in their coding regions, and the primers for ETR2, EIN4, and ERS2 flanked the single intron in their coding regions (Fig. 5, which is published as supporting information on the PNAS web site). A dilution series of Arabidopsis genomic DNA was used as an internal control to determine the expression level of the five receptors. The PCR products from genomic DNA and cDNA were separated on an agarose gel and stained with ethidium bromide. The intensity of DNA bands was quantified by using optiquant software (Packard). The concentration of the genomic DNA standard was also determined on agarose gel. The expression level of the five receptor mRNAs was determined by comparing the intensity of the PCR products from genomic DNA control and cDNA. The analysis was performed in four replicates. We assume that 1 ng of Arabidopsis genomic DNA is equal to 1 × 10⁴ copies of genome for the conversion of genomic DNA mass into copy number.

Figure 1.

RT-PCR analysis of the ers1-2 mRNA. (A) The diagram shows the genomic DNA region of the 5′ end of the ERS1 gene and indicates the location of the T-DNA insertion in an intron of the 5′ UTR of the gene. (B) The diagram depicts the ers1-2 mRNA product as determined from RT-PCR analysis. Arrows indicate the location of the primers used in RT-PCR. Reverse transcription of the ers1-2 mRNA was performed with the Z primer specific to the ERS1 gene. PCR amplification was performed with the various combinations of the two forward primers (A and B) and the four reverse primers (W, X, Y, and Z) for 32 cycles. (C) In the RT+ lanes, reverse transcriptase was used to generate a cDNA template for the PCR. In the RT− lanes, an identical PCR was performed except that the reverse transcriptase was left out of the reaction to determine whether genomic DNA was responsible for the amplified product. M indicates the DNA size marker. Primer sequences are provided in the Supporting Text.

Unique endonuclease restriction sites for transgenic or endogenous genes were used to determine the ratio of transgenic and endogenous genes (Fig. 6, which is published as supporting information on the PNAS web site). PCR products amplified from the cDNA samples by using primer pairs specific to each receptor and flanking the unique restriction sites were digested with NheI (unique for transgenic ETR1 and endogenous ERS1), BstAPI (unique for endogenous EIN4), BamHI (unique for endogenous ETR2), and BspLU11I (unique for endogenous ERS2). The DNA samples were separated on agarose gel and quantified as above. The ratio was determined based on three replicates. More details about these RT-PCR analyses and the sequences of the primers are provided in the Supporting Text.

Etiolated Seedling Assays.

Seeds were plated on 1/2 MSNS agar plates containing 5 μM aminoethoxyvinylglycine to inhibit production of endogenous ethylene, kept at 4°C in the dark for 3 days, and illuminated under fluorescent lights at room temperature for 6 h. The plates were wrapped in foil, transferred into gas-tight Plexiglas chambers, and then gassed continuously with air or 10 μl/liter ethylene for 3 days. Ethylene concentrations were determined with a Perkin–Elmer 8500 gas chromatograph. Hypocotyls and roots of etiolated seedlings were measured as described (23).

Results

Isolation of an ERS1 Mutant.

In an effort to obtain a loss-of-function mutation in the ERS1 ethylene receptor gene of Arabidopsis, a search for T-DNA insertion mutants was conducted. Extensive screening of several T-DNA mutagenized populations yielded only a single disruption in the ERS1 receptor gene. Sequence analysis indicated that the T-DNA was located in an intron 235 bp upstream of the start codon within the 5′ UTR of the gene. After three backcrosses to WT Ws, the homozygous mutant, designated ers1-2, was similar in appearance to WT plants at all stages of development and showed similar response to ethylene in the etiolated seedling growth assay.

Quantitative RT-PCR analysis indicated that the ERS1 coding sequence was expressed at a reduced level in etiolated seedlings and did not show the ethylene inducibility characteristic of the WT gene (data not shown). Further RT-PCR analysis revealed that the ers1-2 coding sequence containing messages included a flanking T-DNA sequence (Fig. 1 B and C). A PCR product was not obtained from primers specific for the 5′ UTR upstream of the T-DNA insertion site and primer Z, indicating no native transcript was present.

The results shown in Fig. 1 B and C indicate that the aberrant message containing T-DNA consists of a hybrid of T-DNA sequence, a segment of unspliced UTR intron, a segment of native UTR, and the native coding sequence. One consequence of this arrangement is that at least eight incorrect ATG start sites are located upstream of the correct start codon in the mutant mRNA. Six of these start sites are out of frame with the coding sequence, whereas the two in the correct reading frame are followed by stop codons upstream of the correct ATG. Based on the presence of these false ATG start sites, it is considered highly unlikely that the native translation product can be produced from this message.

Constitutive Response Phenotype of the ers1;etr1 Double Loss-of-Function Mutants.

To assess the importance of subfamily I ethylene receptors in signaling, double mutants between ers1-2 and the two loss-of-function alleles of ETR1 (etr1-6 and etr1-7) were generated. The F₂ generation resulting from crosses of ers1-2 to either etr1-6 or etr1-7 was screened by using PCR for lines that were homozygous for the etr1 mutations and heterozygous for the ers1-2 allele. After self-fertilization, the progeny of these lines segregated for a small seedling phenotype when germinated in the light on agar plates (Fig. 2B). Genotypic analysis of this segregating population by PCR revealed that the small seedling phenotype cosegregated with the homozygous etr1;ers1 double mutants. The double mutants ers1-2;etr1-6 and ers1-2;etr1-7 showed similar phenotypes.

Figure 2.

The ers1;etr1 loss-of-function mutants exhibit a constitutive ethylene response phenotype. (A) Seedlings grown in the dark on agar plates for 4 days in air. Compared with WT Ws (Left), the ers1-2;etr1-7 seedling exhibited a partial inhibition of hypocotyl growth, but a more complete inhibition of root growth relative to WT (Ws) seedlings. (B) Seedlings grown in the light for 3 days in air. A WT Ws seedling (Left) is shown. The ers1-2;etr1-7 mutant exhibited shorter hypocotyls and roots and small, dark, unexpanded cotyledons. (C) Rosette-stage phenotype of 3-week-old ers1-2;etr1-7 homozygous mutant (white arrow). A WT Ws plant (Left) is shown for comparison. The ers1;etr1 loss-of-function mutants exhibit an extremely compact rosette, delayed flowering, and stunted inflorescence stem. (D) An ers1-2;etr1-6 homozygous mutant (Right, white arrow) is compared with a WT Ws plant (Left). Both plants are 10 weeks old. (E) Close-up of the 10-week-old ers1-2;etr1-6 mutant that has bolted and produced flowers.

As both ers1-2;etr1-6 and ers1-2;etr1-7 double mutants progressed through their life cycle, they continued to display aberrant growth patterns (Fig. 2 C–E). Grown on soil, the double-mutant leaf rosettes were only ≈15% the diameter of the WT or the ERS1 heterozygous segregants in the population. Double mutants also showed a 3-fold delay in time to flowering (average 63 days vs. 23 for WT), with 25% of the double-mutant plants not flowering at all. Double-mutant plants that did flower produced stunted inflorescences with tiny sterile flowers.

The phenotypes of seedlings and adult double mutants grown in the light were more severe than the phenotypes of other ethylene constitutive-response mutants such as ctr1-1 (12) or the triple mutant loss-of-function receptor lines analyzed (13). In fact, the ers1;etr1 double loss-of-function mutants were similar to the phenotype described for the quadruple loss-of-function etr1;etr2;ein4;ers2 receptor mutant, supporting the prediction that the ers1-2 mutation causes severe loss of function for that receptor isoform.

Surprisingly, the ers1;etr1 double loss-of-function mutants displayed a much less severe phenotype in etiolated seedling hypocotyls than expected based on observations with light-grown plants (Fig. 2A). Hypocotyls, but less so roots, of dark-grown seedlings of the double mutant showed an intermediate constitutive response phenotype somewhat less severe than ctr1-1 when grown in air and still showed responsiveness to ethylene in the seedling growth assay.

ETR1-Promoter-Driven Expression of Subfamily I but Not Subfamily II cDNAs Rescues the ers1;etr1 Double Loss-of-Function Mutant Phenotype.

To investigate the possibility that differences in expression level or pattern could account for the functional requirement for subfamily I members in ethylene signaling, cDNAs of all five receptor genes were placed under control of the ETR1 promoter. The cDNAs were spliced between a 1,448-bp fragment of genomic sequence upstream of the ETR1 coding sequence and a 1,362-bp fragment of ETR1 sequence downstream of the stop site. These constructs were transferred into the ers1-2;etr1-7 double mutant background, and the phenotypes of double-mutant plants expressing transgenes were evaluated.

As shown in Fig. 3A, ETR1-promoter-driven expression of ETR1 or ERS1 cDNA completely rescued the severe ers1-2; etr1-7 double-mutant phenotype. Conversely, ETR1-promoter-driven expression of ETR2, EIN4, or ERS2 cDNAs did not detectably reduce the severe phenotypes of the double mutant. Double mutants carrying the subfamily II transgenes were indistinguishable from the nontransformed double mutant in light-grown seedling development (Fig. 3A), etiolated seedling development, adult size, flowering time, and reproductive sterility (data not shown). Quantitative PCR studies indicated that the failure of subfamily II cDNAs was not caused by a lower expression level for these transgenes (Fig. 3B). Translatability of transcripts did not seem to be a problem, because the ETR1- promoter-driven EIN4 and ETR2 cDNA constructs were capable of restoring ethylene responsiveness to the etr1-6;etr2-3;ein4-4 triple mutant (data not shown).

Figure 3.

ETR1-promoter-driven cDNAs of ETR1 and ERS1, but not ETR2, EIN4, and ERS2, rescue the ers1-2;etr1-7 double mutant phenotype. (A) Seedlings grown under light for 4 days. Vector indicates pPZP221. {cETR1}, {cERS1}, {cETR2}, {cEIN4}, and {cERS2} represent cDNAs of each ethylene receptor under the control of the 5′ and 3′ flanking regions of the genomic ETR1. The genetic background of all transgenic plants is ers1-2;etr1-7. A WT Columbia seedling (Left) is shown. (B) The expression level of transgenic or endogenous genes was quantified by RT-PCR. Dark gray indicates the expression of the endogenous genes, and light gray represents the transgenes.

Rescue of the ers1;etr1 Loss-of-Function Mutant Phenotype with a Kinase-Inactivated ETR1 Genomic Clone.

One possible explanation for the severity of the ers1;etr1 loss-of-function mutant phenotype is that these mutants lack the histidine kinase activity required to activate downstream signaling components such as CTR1. To test this idea, a genomic clone of ETR1 coding for a kinase-inactivated form of ETR1 (getr1-[HGG]) was used to transform populations of Arabidopsis segregating for ers1-2;etr1-6 and ers1-2;etr1-7 mutants. The getr1-[HGG] mutant transgene contains mutations in both the conserved histidine and conserved residues in the kinase catalytic domain (G515 and G517 in G1 box). These mutations have been shown to cause a loss of detectable histidine kinase activity in the yeast-expressed ETR1 kinase domain (15).

Transformation of ers1-2/ERS1;etr1-6/etr1-6 or ers1-2/ERS1;etr1-7/etr1-7 plants with the getr1-[HGG] transgene yielded T₁ plants that were genotyped as homozygous for ers1-2;etr1-6 or ers1-2;etr1-7, respectively, and carried the getr1-[HGG] transgene. These plants were WT in phenotype and produced WT-looking T₂ progeny that were genotyped homozygous for ers1;etr1. Transgenic lines homozygous for ers1-2;etr1-7 and the transgene were used for phenotypic analysis. As shown in Fig. 4 B and C, etiolated ers1-2;etr1-7;getr1-[HGG] seedlings showed WT growth of both hypocotyls and roots when grown in air and responded normally to ethylene. Development of light-grown seedlings (Fig. 4A) and adult plants (data not shown) of these lines was also indistinguishable from WT plants, showing none of the phenotypic defects characteristic of the ers1-2;etr1-7 mutants. These results lead to the conclusion that canonical histidine kinase activity of the ETR1 receptor is not essential for the transmission of signal from receptor to downstream effectors in the ethylene signal transduction pathway.

Figure 4.

Developmental defects of the ers1-2:etr1-7 mutant are rescued by a kinase-inactivated ETR1 transgene. (A) Seedlings grown under light for 4 days. (B) Hypocotyl length of etiolated seedlings. Seedlings were grown on agar plates in the dark for 3 days in either air (filled bars) or 10 μl/liter ethylene (open bars). (C) Root length of etiolated seedlings. The specified lines were transformed with either a WT (gETR1) or kinase-inactivated ETR1 genomic clone (getr1-[HGG]). At least 20 seedlings were measured for each treatment.

Discussion

The research reported here was undertaken to investigate whether the subfamily I receptors have a special role in ethylene signal transmission. A previous study with the receptor family from tomato demonstrated that overexpression of a subfamily II receptor could compensate for reduced expression of a subfamily I receptor (24). However, the tomato genome contains three subfamily I receptor isoforms, only one of which was suppressed in the study. In the present study, loss of function of the only two subfamily I receptors of Arabidopsis could not be compensated by overexpression of any of the three subfamily II receptors, supporting the idea that subfamily I receptors play a unique and necessary role in ethylene signaling.

Given that subfamily I receptors are distinguished from subfamily II receptors by the presence of a conserved histidine kinase domain, the ability of a kinase-inactivated form of ETR1 to provide a virtually complete rescue of the ers1;etr1 double mutants was unexpected and appears to rule out canonical histidine kinase phosphotransfer as the necessary mechanism by which ethylene receptors transmit signal to downstream effectors. A recent study of signaling by the dominant etr1-1 mutant receptor also failed to support, although did not rule out, a role for histidine kinase activity in signaling (25). Transgenic forms of the etr1-1 mutant receptor could confer dominant insensitivity to ethylene even when the kinase domain was rendered inactive by mutagenesis or truncated entirely, leading to the conclusion that constitutive signaling by the mutant receptor either occurred independently of histidine kinase activity, or was achieved by mutant receptors interacting with WT receptors. Further support for the latter possibility comes from another recent study demonstrating that ethylene insensitivity conferred by the dominant subfamily II etr2-1 allele partially depends on the presence of a functional ETR1 gene (26).

The finding that histidine kinase activity of ETR1 is not required for receptor signaling presents an interesting evolutionary parallel between the ethylene receptors and the phytochromes. Some cyanobacterial forms of phytochrome appear to signal as canonical histidine kinases, whereas phytochromes of flowering plants contain degenerate histidine kinase domains (27). Ethylene receptor-like genes are also found in cyanobacteria, including slr1212 of Synechocystis, that has a conserved ethylene binding domain with ethylene binding activity, but lacks a complete histidine kinase transmitter domain (5). However, genes from the recently sequenced Anabaena genome encode proteins with both an ethylene binding domain and a conserved histidine kinase domain (28). Given the likelihood that vascular plants recruited both phytochrome and ethylene receptor functions from the cyanobacterial ancestor of the chloroplast (29), it is interesting to speculate that what may have started as histidine kinase transmitters have evolved into novel signaling mechanisms.

Despite the degenerate nature of histidine kinase domains in plant phytochromes, they retain the capacity for both autophosphorylation and transphosphorylation of substrates with a specificity for Ser/Thr residues (27, 30). In fact, histidine kinases may have evolved into Ser/Thr kinases several times independently (31). The possibility that ethylene receptor signaling occurs through a novel autophosphorylation mechanism in the subfamily II ethylene receptor isoforms and even in the WT subfamily I isoforms and the protein carrying the mutated getr1-[HGG] gene cannot be ruled out.

Alternatively, a novel signaling mechanism that does not require phosphotransfer by receptor may have evolved in the ethylene receptor family. Precedents for such evolutionary trends are found in some receptor tyrosine kinases (RTKs) in animals. A number of so-called “kinase dead” RTKs retain the capacity for signal tranduction (32). These systems may rely on heterologous RTKs or soluble kinases for transphosphorylation. Perhaps the RAF-related CTR1 fulfills this role in ethylene signaling.

The physical interaction of the ETR1 and ERS1 transmitter domains with the presumed regulatory domain of CTR1 (18) could provide for the direct propagation of conformational change from receptors to the CTR1 catalytic domain. Alternatively, conformational changes in receptors could recruit and/or release CTR1 from a receptor complex, analogous to the recruitment of RAF kinases to receptor complexes in animals (33). Further propagation of the signal by CTR1 could occur through autophosphorylation, transphosphorylation of receptors, or phosphorylation of downstream effectors.

Perhaps the physical interaction of the transmitter domains of ETR1 and ERS1 with the regulatory domain of CTR1 represents the special role that subfamily I receptors play in signaling. Physical interactions of the subfamily II ETR2 transmitter domain with CTR1 are much weaker than those reported for ETR1 and ERS1 (26). The severe phenotype of the ers1;etr1 double loss-of-function mutants implies that this weaker interaction may be insufficient to promote significant signaling, except perhaps in the case of etiolated hypocotyls (Fig. 4B). If this is the case, then subfamily II receptors could contribute to signaling through receptor/receptor interactions with subfamily I, as has been recently suggested (25, 26, 34). Recent studies with bacterial two-component systems support an important role for receptor interactions in signal output (35).

From an evolutionary perspective, all of these mechanisms can be viewed in terms of the recruitment of CTR1 as an adapter kinase that either led to or compensated for the loss of canonical histidine kinase activity as the principal mode of signal transduction by the ethylene receptors (32). The retention of histidine kinase activity in subfamily I receptors may have been selected in evolution to provide fine-tuning of the signaling pathway rather than functioning as the primary mechanism for signal transduction. More detailed studies of the activity of the getr1-[HGG] protein in planta may reveal such a function.

Abbreviation

Ws

Wassilewskija