Ethylene-binding activity, gene expression levels, and receptor system output for ethylene receptor family members from Arabidopsis and tomato


  • Just prior to print publication of this article, our colleague and friend, Tony Bleecker, passed away on January 30, 2005. The authors wish to dedicate this paper to the memory of Tony and also to Jeff Esch, who passed away last year.

(fax +608 262 7509; e-mail


Ethylene signaling in plants is mediated by a family of ethylene receptors related to bacterial two-component regulators. Expression in yeast of ethylene-binding domains from the five receptor isoforms from Arabidopsis thaliana and five-receptor isoforms from tomato confirmed that all members of the family are capable of high-affinity ethylene-binding activity. All receptor isoforms displayed a similar level of ethylene binding on a per unit protein basis, while members of both subfamily I and subfamily II from Arabidopsis showed similar slow-release kinetics for ethylene. Quantification of receptor-isoform mRNA levels in receptor-deficient Arabidopsis lines indicated a direct correlation between total message level and total ethylene-binding activity in planta. Increased expression of remaining receptor isoforms in receptor-deficient lines tended to compensate for missing receptors at the level of mRNA expression and ethylene-binding activity, but not at the level of receptor signaling, consistent with specialized roles for family members in receptor signal output.


Molecular genetic studies in Arabidopsis thaliana have led to the development of a model for ethylene signal transduction in plants (Guo and Ecker, 2004). A family of five receptors related to bacterial two-component signaling systems are thought to mediate the initial sensing of the ethylene signal. According to the model, the ETR family of ethylene receptors, along with a RAF-related kinase CTR1 (Kieber et al., 1993), act to negatively regulate ethylene response pathways in the absence of ethylene. Ethylene binding inhibits the activity of the receptor/CTR1 complex, leading to activation of response pathways (Hua and Meyerowitz, 1998). Downstream of the receptor/CTR1 complex is an Nramp-related membrane protein, EIN2, that is required for ethylene responses (Alonso et al., 1999). Genetic evidence indicates that at least some responses to ethylene result from an EIN2-mediated activation of a transcriptional cascade (Solano et al., 1998).

Evidence that all five members of the Arabidopsis ETR gene family are involved in mediating ethylene responses comes from the finding that dominant mutations in any single member of the family render the plant insensitive to ethylene (Chang et al., 1993; Hua et al., 1998; Sakai et al., 1998). Additionally, loss of function of two or more receptors can result in a constitutive response phenotype, indicating that all receptor isoforms contribute to signaling through the pathway (Hua and Meyerowitz, 1998; Wang et al., 2003). The receptor system discovered in Arabidopsis can also be generalized to other plant species. In particular, work in tomato has demonstrated that mutations in the homologous receptor genes can cause dominant ethylene insensitivity (Lanahan et al., 1994) and that loss of receptor gene function can result in constitutive activation of ethylene-response pathways (Tieman et al., 2000).

Evidence that the ETR gene family acts directly to perceive ethylene resulted from studies of ETR1 expressed in yeast. The yeast-expressed protein showed high-affinity ethylene-binding activity (Schaller and Bleecker, 1995). Additional studies indicated that binding is mediated by the copper cofactor-containing hydrophobic domain at the N-terminus of the protein (Rodriguez et al., 1999). This highly conserved hydrophobic domain also mediates ethylene binding in the Arabidopsis ERS1 receptor. The discovery that a homologous hydrophobic domain in the slr1212 gene from the Cyanobacterium Synechocystis also showed ethylene-binding activity (Rodriguez et al., 1999) increases the likelihood that all members of the ethylene receptor family function as ethylene-binding proteins.

While all five members of the ethylene receptor family from Arabidopsis share a common ethylene-binding domain, other features of the coding sequences are specific to receptor subtypes. Based on some distinguishing characteristics and overall homology, the family has been divided into two subfamilies. Subfamily I, composed of ETR1 and ERS1, contains three hydrophobic stretches near the N-terminus and possesses histidine kinase domains and activity (Moussatche and Klee, 2004), although histidine kinase activity does not appear to be required for signaling (Qu and Schaller, 2004; Wang et al., 2003). Subfamily II members (ETR2, EIN4, and ERS2) contain four hydrophobic stretches at the N-terminus, have degenerate histidine kinase domains, and possess serine kinase activities (Moussatche and Klee, 2004). Family members can also be classified as to whether they contain a receiver domain at the C-terminus (ETR1, ETR2, and EIN4), or lack such a domain (ERS1 and ERS2). Little is known about specific roles for any particular receptor subtype. Recent studies indicate that proper signaling requires the presence of a subfamily I member for most ethylene responses (Wang et al., 2003).

To better understand the roles that ethylene receptor subtypes play in signaling, we examined the ethylene-binding activity of the members of the ethylene receptor family from Arabidopsis and tomato. We addressed the question of whether receptor subtypes display major differences in ethylene-binding affinity. The relationships among receptor gene expression levels, binding activities, and signal output capacities in wild type and receptor-deficient lines were also investigated.


Ethylene-binding activity of members of the Arabidopsis ethylene receptor family

Functional analysis of a protein containing the first 128 residues of ETR1 fused to glutathione-S-transferase (ETR1[1-128]-GST) indicated that the N-terminal domain contains the ethylene-binding site (Rodriguez et al., 1999). To determine whether other members of the ethylene receptor family in Arabidopsis show ethylene-binding activity, we made equivalent GST translational fusion constructs for expression of the putative ethylene-binding domains of AtETR1, AtETR2, AtERS1, AtERS2, and AtEIN4. This strategy provided for the detection of all the proteins with a common antibody reagent, allowing a comparison of their relative expression levels. Western blots probed with anti-GST serum showed the presence of proteins with the expected molecular weights (approximately 45 kDa) in samples from cells containing receptor-GST expression constructs, but not in samples from cells transformed with empty vector (pYcDE2) (Figure 1). Control cells containing empty vector and cells expressing AtETR-GST fusion proteins were exposed to 0.1 μl l−114C2H4 (gas) in the presence or absence of 1000 μl l−1 non-labeled ethylene (12C2H4) to measure specific binding of ethylene. The results showed that all of the AtETR-GST constructs generated ethylene-binding activity in the cells, in contrast to control cells containing empty vector (pYcDE2, Figure 1). Cells expressing AtETR1[1-128]-GST showed the highest level of ethylene binding and AtETR[1-128]-GST protein, while cells expressing AtERS2[1-160]-GST showed the lowest level of ethylene-binding activity and immunodetectable protein.

Figure 1.

 Ethylene-binding activity in members of the ethylene receptor family in Arabidopsis (AtETRs). Intact yeast cells containing the constructs indicated were analyzed for 14C2H4 bound in the presence of 0.1 μl l−114C2H4 (white bars) or 0.1 μl l−114C2H4 + 1000 μl l−112C2H4 (black bars). The average dpm per gram of yeast ± standard deviation is shown. Equal amounts of membranes from yeast used in the binding assays were analyzed with Western blots probed with anti-GST antibodies. Protein levels were normalized to At-ETR1-GST as determined by densitometric scan. The percentage receptor occupancy was estimated based on the previous dose-binding curve for yeast-expressed ETR1 where 59% of the receptors were occupied at 0.1 μl l−1 ethylene (Schaller and Bleecker, 1995). Occupancy values were calculated based on the measured binding activity and relative protein expression levels.

To obtain an estimate of the relative affinities for ethylene of the different receptor isoforms, an estimate of the percentage of total receptor that was occupied by ethylene (receptor occupancy) was made. Based on the previous dose-binding curve for yeast-expressed ETR1 (Schaller and Bleecker, 1995), we assigned a value of 59% for the receptor occupancy of AtETR1[1-128]-GST treated with 0.1 μl l−1 ethylene (Figure 1). Assuming the same stoichiometry of ethylene binding to the other receptor isoforms, receptor occupancy values were calculated for each isoform based on the binding activity and protein expression level relative to AtETR1[1-128]-GST using antibodies against the common GST domain. The values obtained ranged from 48 to 41%, indicating that ethylene affinities for all receptor isoforms were very similar.

Ethylene binding by members of the ethylene receptor family from tomato (LeETRs)

We carried out experiments to determine whether the LeETR proteins possess ethylene-binding activity. Based on the region known to contain the ethylene-binding domain of AtETR1, constructs were made for the expression of the corresponding regions of the members of the LeETR proteins fused to a GST tag. Analysis of ethylene binding in yeast containing empty vector (pYES2), LeETR1[1-141]-GST, LeETR2[1-125]-GST, LeETR3[1-125]-GST, LeETR4[1-141]-GST, and LeETR5[1-143]-GST revealed that all LeETRs possess ethylene-binding activity (Figure 2). The samples from cells expressing LeETR3[1-125]-GST showed the highest levels of ethylene binding and protein, while cells expressing LeETR5[1-143]-GST showed the lowest level of activity and protein. Estimations of the percentage receptor occupancy for the tomato receptor-binding sites varied more than the values obtained from Arabidopsis isoforms, although still remained within an order of magnitude of AtETR1[1-128]-GST. While these findings may indicate real differences in affinities, some values, particularly those for LeETR4[1-125]-GST and LeETR5[1-143]-GST, may reflect quantification errors due to the very low protein expression levels relative to AtETR1[1-128]-GST.

Figure 2.

 Ethylene-binding activity in members of the ethylene receptor family in tomato (LeETRs). Intact yeast cells containing the constructs indicated were analyzed for 14C2H4 bound in the presence of 0.1 μl l−114C2H4 (white bars) or μl l−114C2H4 + 1000 μl l−112C2H4 (black bars). The average dpm per gram of yeast ± standard deviation is shown. Equal amounts of yeast membranes from yeast used in the binding assays were analyzed with Western blots probed with anti-GST antibodies. Proteins levels were normalized to At-ETR1-GST as determined by densitometric scan. Estimates of the percentage receptor occupancy were calculated as in Figure 1.

Kinetics of dissociation of ethylene from AtETR1[1-128]-GST and AtETR2[1-157]-GST

To further investigate the relative binding properties of ethylene receptor isoforms, we compared the dissociation kinetics of 14C2H4 bound to yeast transformed with AtETR1[1-128]-GST, AtETR2[1-157]-GST and control (pYcDE) constructs (Figure 3). The results showed that the dissociation of ethylene bound to AtETR1[128]-GST is a slow process with a half-life of 12 h, as previously reported for full-length AtETR1 (Schaller et al., 1995). Cells expressing AtETR2[1-157]-GST protein showed slow-dissociation kinetics requiring approximately 10 h for the dissociation of 50% of the 14C2H4 originally bound to the cells (887 DPM g−1 yeast). In contrast, cells containing empty vector showed the loss of more than 90% of the 14C2H4 bound in <1 h. Taken together, the data indicate that the dissociation rate of ethylene from the subfamily II representative, AtETR2, is similar to that of the subfamily I representative, AtETR1.

Figure 3.

 Dissociation of ethylene from AtETR1[1-128]-GST (□) and AtETR2[1-157]-GST (○) proteins. Yeast cells containing the constructs indicated were exposed to 0.1 μl l−114C2H4, aired for 10 min, and analyzed for 14C2H4 bound after 0.1, 1.0, 2.0, 4.0, 8.0, and 15 h in a chamber with continuous flow of humidified air (10 ml min−1). The average level (±standard deviation) of 14C2H4 remaining is shown.

The relationship between receptor gene expression and total ethylene-binding activity in receptor-deficient plants

To determine the relative contributions of each gene to the total receptor expression level, a semiquantitative RT-PCR analysis of all members of the family was undertaken using rosette leaf tissues from wild type and various receptor-deficient backgrounds. Leaf samples were taken from the same plantings that were used for subsequent ethylene-binding studies (see below).

As shown in Figure 4(a), the ETR1 message level in untreated rosette leaves of wild-type plants accounts for more than 50% of the total receptor message, while ETR1 and ERS1 together account for approximately 80% of the total message in wild-type plants. In the ctr1-2-constitutive ethylene response mutant plants, the total receptor message level was 1.8-fold higher than wild type due primarily to increased expression of the ethylene-inducible receptor genes, ERS1, ETR2, and ERS2. In the etr1;etr2;ein4 receptor-deficient line, the message levels of the remaining ERS1 and ERS2 receptors were elevated so that total receptor message level was comparable to wild type. Similarly, an elevated level of ERS1 in the etr2;ein4;ers2 receptor-deficient line compensated for the loss of subfamily II messages so that total message level was comparable to wild type. By contrast, the etr1-7 mutant, which is null for ETR1 protein expression, shows a 50% reduction of total message owing to a slight increase in ERS1 expression associated with the weak constitutive response phenotype of this single gene mutant (Hua and Meyerowitz, 1998). The etr1-1 mutant line shows an even greater reduction in total receptor message level resulting from the lack of induction of remaining receptor genes in this ethylene-insensitive plant.

Figure 4.

 The relationship between transcript levels and ethylene-binding activity in wild-type and mutant Arabidopsis rosettes.
(a) Transcript levels of ethylene receptor isoforms present in rosette tissues of wild type and mutant plants. Only transcripts from wild-type alleles are shown.
(b) Relative ethylene-binding activity in rosettes of wild type and mutant plants.
(c) Correlation of transcript levels and ethylene-binding activity in wild type and mutant plants. The linear regression has an R2 value of 0.944.

To investigate the relationship between receptor message levels and levels of functional receptor proteins, in vivo ethylene binding assays were performed on leaves of wild type and mutant plants (Figure 4b). The ethylene binding assay involved treatment of leaf tissues with 0.1 μl l−1 ethylene for 4 h. Control experiments indicated that this treatment did not alter the total receptor expression level by more than 15% in wild-type leaves (data not shown), which is within the margin of error for the assay. Ethylene-binding activity was 1.5-fold higher in ctr1-2 than in wild type, consistent with the increased total receptor message. Levels of ethylene binding activities in the two receptor-triple-mutants were also consistent with the total receptor message levels in those lines. Interestingly, both triple mutants show total receptor mRNA abundance and total ethylene-binding activity similar to wild type. Finally the etr1-7 and etr1-1 mutant lines showed reductions in ethylene-binding activity commensurate with the reduced levels of total wild-type receptor message levels in these lines. These results indicate a simple relationship in which total functional binding sites are correlated with total mRNA abundance (Figure 4c).

Increased expression of remaining receptors does not completely compensate for signaling defects in receptor-deficient lines

The finding that the leaf tissues of receptor triple mutants have wild-type levels of ethylene-binding activity raised the question of whether the altered proportions of receptor isoforms in the triple mutants affect signaling capacity. The diameter of the adult rosette provides a measure of signaling capacity of the receptor/CTR1 complexes; reduced signaling capacity being associated with reduced leaf size (Hall and Bleecker, 2003). As shown in Figure 5, rosette diameter is greatly reduced in both the receptor triple mutant lines and the ctr1-1 mutant. These results imply that, while triple mutants compensate for receptor deficiency with respect to number of ethylene-binding sites in leaves, the altered composition of receptor isoforms does not completely compensate in terms of signal output. Specifically, increased expression of subfamily I receptors in the etr2;ein4;ers2 triple mutant did not completely restore downstream signaling in the absence of subfamily II receptors. Likewise, compensatory increases in receptors lacking receiver domains could not restore signaling in the receiver domain-containing receptor-deficient etr1;etr2;ein4 triple mutant.

Figure 5.

 The effects of various ethylene-sensing mutations on rosette diameter. Rosette diameters were measured in 3-month-old Arabidopsis plants. The average diameter ± standard deviation is shown for each line.


A survey of ethylene receptor isoforms related to AtETR1 from Arabidopsis and tomato revealed that all sequences with the conserved ethylene sensor domain display high-affinity ethylene binding when expressed in yeast. This finding is consistent with the fact that all known homologs of AtETR1 show conservation of amino acid residues known to be essential for ethylene binding in AtETR1 (Rodriguez et al., 1999). This conservation extends to the cyanobacterium Synechosystis (Rodriguez et al., 1999) and by inference to the large number of ethylene receptor genes identified across angiosperms and in the cyanobacteria (Mount and Chang, 2002), although a biological role for ethylene in the latter has not been established.

One possible function for multiple ethylene receptor isoforms in a single species could be that different members of the family could operate over significantly different concentrations of ethylene. This would provide an explanation for the fact that, collectively, ethylene responses in Arabidopsis occur over seven orders of magnitude in ethylene concentration (Binder et al., 2004a; Chen and Bleecker, 1995). However, this survey of receptor isoforms revealed that similar binding activity per unit expressed protein was obtained for all receptor isoforms using a concentration of ethylene near the calculated Kd for ethylene binding to AtETR1 (Schaller and Bleecker, 1995). Assuming that all receptor isoforms conform in general to Michaelis–Menten type kinetics, the similar stoichiometry of ethylene binding per unit protein for the other Arabidopsis receptor isoforms indicates that they are also near 50% occupancy at the applied ethylene concentration, implying an affinity for ethylene similar to AtETR1. The more variable affinities calculated for the tomato isoforms will need to be reevaluated with constructs that express better in yeast.

While the above assumption that all receptor isoforms conform to Michaelis–Menten kinetics needs to be tested with more thorough kinetic analyses, the possibility that receptor isoforms do not differ appreciably in affinities for ethylene is consistent with the linear relationship between total receptor message level and ethylene-binding activity in planta that is independent of the particular receptor isoforms expressed. Our findings are also consistent with previous estimates of ethylene-binding affinities in various plant tissues (Hall et al., 1990; Sisler, 1991). These tissues presumably express multiple ethylene receptor isoforms, yet the calculated Kd values for the slow-release components are very similar to those reported here for ETR1 and the other receptor isoforms. If there is indeed little difference in the affinities of the ethylene receptor isoforms for ethylene, some alternative explanation for the ability of the receptor system to operate over such a wide ethylene concentration range must be sought.

Previous studies of ethylene binding in plant tissues led to the suggestion that plants have a fast-release and a slow-release binding activity (Hall et al., 1990; Sisler, 1991). However, release kinetics of two receptors representing the two defined receptor subfamilies did not indicate any large difference in release kinetic. Both the AtETR1 and the AtETR2 isoforms showed characteristics of the slow-release forms identified in these previous studies. It remains an open question whether as yet unidentified ethylene-binding sites account for the fast release activities previously reported, or whether receptor heterodimers of ETR family members could produce binding activities that differ from those identified in yeast expressed homodimers. The present study provides no evidence for such interactions.

Assuming that identified ethylene receptor isoforms in Arabidopsis all show similar binding characteristics, a one-to-one correspondence between ethylene-binding activity and total receptor mRNA levels was found between wild-type and receptor-deficient lines. This finding is consistent with a direct transcriptional control of receptor level in this system. The correlation between total receptor message level and ethylene-binding activity is surprisingly high given the number of factors that could affect this relationship in Arabidopsis leaves. For example, Zhao et al. (2002) reported that ETR1 protein levels were equivalent in the etr2;ein4 ers2 triple mutant and isogenic wild type. This finding is consistent with the equivalent message levels found in both backgrounds in this study. On the contrary, etr1-1 protein levels were over twofold higher in the etr1-1 mutant background than in the isogenic wild type in the previous study (Zhao et al., 2002), while message level for etr1-1 was comparable to wild type. Thus, there are conditions in which message and protein levels may not correspond, although in the latter case there is no consequence to ethylene-binding activity because the etr1-1 mutant protein lacks ethylene-binding activity (Schaller and Bleecker, 1995). The estimates of ethylene-binding activity in leaf tissues may also be influenced by the rate of endogenous ethylene production in the different backgrounds due to competition for binding sites. This should not be a problem with the mutants showing constitutive ethylene responses given that reported rates of ethylene production in the ctr1 mutant background are similar to wild type (Kieber et al., 1993). Measurements of ethylene production by triple mutants in this study also showed no significant differences from wild type (data not shown). However, higher rates of ethylene production have been reported for dominant ethylene-insensitive mutants (Sanders et al., 1991) which may lead to an underestimate of binding activity in the etr1-1 background.

Previous work in tomato indicated that, under some circumstances, increased expression of one receptor isoform, when expression of another was suppressed, could lead to functional compensation of phenotype (Tieman et al., 2000). This is clearly not the case in the current study with triple mutants. Compensation did occur at the level of receptor gene expression, but the increased expression of a limited set of receptor isoforms did not prevent the expression of a constitutive ethylene response phenotype for these two triple mutants at the level of the seedling growth response (Hall and Bleecker, 2003) or adult vegetative development (this study).

These results imply some special role for both subfamily II receptors and hybrid (receiver-domain containing) receptors in the downstream signaling capacity of the ethylene signaling system. Adding to this the previous evidence for a special role for subfamily I receptors in signaling (Wang et al., 2003), it becomes apparent that full capacity signaling by the ethylene receptor system may require a complex integration of features specific to different receptor subtypes. However, the integration of these features appears to be strictly quantitative with respect to receptor signal output, given the evidence that all receptor-deficient phenotypes are channeled through EIN2 (Hall and Bleecker, 2003).

The lack of a simple additive relationship between ligand binding and receptor output indicates synergistic effects of receptor isoform composition on receptor signal output that may be accounted for by a model like those currently favored for the related two-component chemoreceptors from bacteria (Thomason et al., 2002). According to these models, receptors exhibit amplification and adaptation through complex biochemical interactions. Chemoreceptors adapt to a given signal strength by returning to the pre-stimulus signaling state through a feedback modification of the receptor complex. Amplification is thought to result from receptors forming higher order clusters composed of receptor-dimer subunits. Through direct contact, receptor dimers can influence the signaling states of neighboring dimers so that transmitters from many receptors may be altered by a single ligand-binding event. While there is currently no direct biochemical evidence for similar adaptation or amplification systems in the ethylene signaling complex, the models from bacterial chemotaxis may provide explanations for dominant ethylene-insensitive receptor mutants (Gamble et al., 2002) as well as the effects of receptor composition on signal output (Hall and Bleecker, 2003), and other complex phenomena related to the overall behavior of the ethylene signal transduction system (Binder et al., 2004a,b).

Experimental procedures

DNA constructs, cell strains, and growth conditions

The backbone for all AtETR-GST constructs was generated by restriction digest of the BglII-BamHI fragment containing the first 1200 bp of the ETR1 cDNA from the pYcDE2-ETR[1-400]-GST construct described previously (Hall et al., 2000). The sequence encoding the ethylene-binding domain of each gene (Figure 1) was amplified by PCR from the corresponding cDNA clones (Chang et al., 1993; Hua et al., 1995; Sakai et al., 1998) with primers introducing BamH I and BglII restriction sites for cloning. In an attempt to enhance similar expression levels for all the constructs, the forward primers for all the AtETR-GST constructs put the ATG start codon on a similar context by preceding it with a fragment the sequence upstream of the ETR1 cDNA (atagtgttaaaaaattcata) included in previously characterized ETR1 constructs (Rodriguez et al., 1999; Schaller et al., 1995). ETR1[1-128] was generated by PCR amplification with the forward primer 5′cgcggatccATAGTGTTAAAAAATTCATAATGGAAGTCTGCAATTGTATTGAACCGC3′ and the reverse primer 5′ggagatctTCAGCAGCTTTATTTTTCAAGAAAAGC3′, ERS1[1-128] using the forward primer 5′cgcggatccatagtgttaaaaaattcataATGGAGTCATGCGATTGTTTTGAG and reverse primer 5′ggagatctTCTAACTCATCAGCTTTCTTCTTGAG3′, ETR2[1-157] with the forward primer 5′cgcggatccatagtgttaaaaaattcataATGGTTAAAGAAATAGCTTCTTGG3′ and reverse primer 5′ggagatctTCATGAGCTTTCTTCTTAAGCATAAAC3′, ERS2[1-160] using the forward primer 5′cgcggatccatagtgttaaaaaattcataATGTTAAAGACATTGTTAGTCCAATGGC3′ and reverse primer 5′ggagatctTCTCTGGTCTTCTTACTCAAC3′, and EIN4[1-151] using the forward primer 5′cgcggatccatagtgttaaaaaattcataATGTTAAGgTCTTTAGGTCTTGGATTGC3′ and reverse primer 5′ggagatctTCCAACACATTCTGCTTCAAGTAAAGC3′. The forward primer for AtEIN4[1-151] introduces a silent mutation (+9A to g) to eliminate a native BglII (AGATCT) site and allow for the cloning. Original bases are indicated in capital letters and non-native bases are indicated by small case. After PCR, each fragment was gel purified, digested with BamHI and BglII restriction enzymes and ligated in frame with the glutathione-S-transferase sequence in the pYcDE2-GST backbone. The sequence was confirmed by fluorescent dideoxy nucleotide sequencing at the University of Wisconsin Sequencing Facility. All constructs were propagated in Escherichia coli strain DH5α and introduced into Saccharomyces cerevisiae cells by the lithium acetate transformation method (Gietz et al., 1995).

The LeETR-GST constructs were generated following the same strategy, but cloned in pYES2 vectors. All were in-frame fusions to GST with GST at the carboxyl terminus. The amino acids for the receptors in each fusion construct were LeETR1[1-141], LeETR2[1-125], LeETR3[1-125], LeETR4[1-141], and LeETR5[1-143].

Yeast cultures were grown in SD media (Sherman, 1991) lacking the appropriate amino acid at 30°C with aeration and harvested as described by Schaller and Bleecker (1995). Cell pellets were stored at −80°C until use. AtETR-GST constructs were introduced in the strain LRB520 and expressed constitutively under control of a modified ADH2 promoter (Schaller et al., 1995). The LeETR-GST constructs were expressed in the strain Sc295 + pMTL4C (Gamble et al., 1998). Expression of LeETR-GSTs was induced by addition of galactose to 20% and growth for an additional 4 h.

Ethylene-binding assays

For ethylene-binding assays of plant tissues, 10 g of soil-grown rosettes were harvested by cleaving at the base of the root. The tissue was kept moist on paper towels for 2 h before the ethylene-binding assay was initiated. Ethylene binding was assessed by previously described methods (Bleecker et al., 1988; Sisler, 1979). After duplicated samples were incubated with 0.1 μl l−114C2H4 (56.9 mCi mmol−1; American Radiolabeled Chemicals, St Louis, MO, USA) or 0.1 μl l−114C2H4 + 1000 μl l−112C2H4 for 4 h, the samples were kept moist while airing for 15 min. Samples were then transferred to a new sealed chamber containing a mercuric perchlorate trap. The chamber was heated to 65°C for 90 min and then incubated at room temperature for 24 h to trap the ethylene released from the samples. Trapped 14C2H4 was quantified by liquid scintillation counting. Saturable ethylene binding was calculated by subtracting the values obtained from samples diluted with cold ethylene from equivalent samples incubated with 14C2H4 alone. The background levels determined by 14C2H4 + 12C2H4 assay were consistent across all of the samples. After normalizing for background, the activity of each sample was expressed as percentage binding of the wild type. Values reported represent the mean of three independent experiments.

For ethylene-binding assays in yeast, 1 g samples of yeast containing control or experimental constructs were analyzed for 14C2H4 bound in the presence of 0.1 μl l−114C2H4 or 0.1 μl l−114C2H4 + 1000 μl l−112C2H4 as described in Schaller and Bleecker (1995). Ethylene dissociation measurements were made on 14C2H4-labeled samples of yeast that were aired for 10 min, then placed in a chamber with continuous flow of humidified air (10 ml min−1) for 0.5, 1.0, 2.0, 4.0, 8.0 and 16 h before the amount of 14C2H4 bound was analyzed. All samples were analyzed in triplicate and data were normalized to samples analyzed immediately after airing for 10 min.

Protein isolation and detection

Yeast pellets of 25 mg of cells were resuspended in 250 μl of 10% [w/v] trichloroacetic acid, incubated on ice for 15 min, and total protein samples pelleted by centrifugation at 12 000 g for 5 min. Pellets representing total protein isolates were resuspended with 100 μl of 2X SDS-PAGE loading buffer [125 mm Tris–Cl pH 6.8, 20% glycerol (v/v), 4% SDS (w/v), 0.01 bromphenol blue]. To obtain monomeric ETR1, disulfide links were reduced with 100 mm dithiothreitol and incubated at 37°C for 1 h (Schaller et al., 1995).

For Western blots, the samples were cleared of insoluble cell debris by brief centrifugation and 10 μl samples were fractionated via SDS-PAGE in 12% (w/v) acrylamide gels (Bio-Rad, Hercules, CA, USA). Proteins were electrotransferred to Immobilon PVDF membranes (Millipore, Bedford, MA, USA), probed with anti-GST serum (1:5000; Sigma, St Louis, MO, USA) Immunodecorated proteins were visualized by a chemiluminescent detection system (Kirkegaard-Perry, Gaithersburg, MD, USA) following instructions of the manufacturer.

Genomic DNA purification

Genomic DNA was purified from 1 g of 5-day-old seedlings using the 10 columns of the DNeasy kit (Qiagen, Carlsbad, CA, USA), and recombining all the samples. DNA concentration was determined by gel electrophoresis with ethidium bromide and confirmed by fluorometry using Hoechst reagent. Gel electrophoresis indicates that the size of the genomic DNA is in the 15–10 kb range.

RNA extraction and cDNA synthesis

Total RNA was extracted with the Qiagen Rneasy kit from 200 mg of A. thaliana leaf tissue harvested at the same time as the tissues used for ethylene binding assays. The total quantity of RNA was determined by UV and gel electrophoresis. Denaturing PAGE analysis further confirmed the concentration and provided a qualitative check on the RNA integrity.

One microgram of total A. thaliana RNA from leaf tissue was digested with 1 U/μg DNase (Gibco BRL, Carlsbad, CA, USA) and used to prepare cDNA by reverse transcription (Superscript II; Gibco BRL) with 500 nm oligo(dT)18 + oligo(N)3 in a 20-μl reaction with a 50-min reverse transcription step. Both the cDNA samples from treated and untreated seedlings were diluted fivefold to 100 μl final volume and split into three equal 33.3 μl aliquots. A genomic DNA fourfold dilution series of 40 × 103, 10 × 103, and 2.5 × 103 molecules μl−1 was prepared, and 33.3 μl of each genomic DNA dilution was added to one of the cDNA aliquots.

Semiquantitative RT-PCR

Absolute message levels were calculated for each of the receptor isoforms utilizing intron flanking primers as previously described (Wang et al., 2003). The cDNA and genomic DNA mixtures were PCR amplified for 26 cycles (45 sec at 94°C, 1 min at 55°C, 1:10 min a 72°C, and a final 7 min at 72°C for extension) with each of the five ETR1 family isoform-specific primer pairs. The products were run on a 1% agarose gel (100 V, 0.75 h) which contained ethidium bromide. Products were detected by transillumination, photographed with a Kodak digital camera, and quantified with ImageQuant software (Molecular Dynamics, Sunnyvale, CA, USA). Analysis of the data was performed as described in (Pfaffl et al., 1998).

For each ETR1 isoform the gene-specific primers span the last intron and were designed to generate cDNA and genomic PCR products that differ in size between 12 and 16%. The sense primers are all around 700 bp from the 3′end of the mRNA to ensure that all primer sets are targeting similarly sized cDNA products at similar distances from the poly adenylated tail of the mRNA, avoiding effects on cDNA concentration due to incomplete reverse transcription of the mRNA.

Following are the forward and reverse primers used in the semiquantitative PCR. Included are the gene AGI identifiers and amplified product size: ETR1 (At1g66340), (5′-GGATGCACGGCTATCTTTGATGT-3′) and (5′-ATTTGTCAGTGTTACCACTGAGT-3′) (398 bp product); ETR2 (At3g23150), (5′-AAGCTTTGGTGTTTGTAAGAAAGT-3′) and (5′-TCCCACATTTCTTCATCCAA-3′) (478 bp product); ERS1(At2g40940), (5′-ACAGCCTCGGACCGGAACT-3′) and (5′-AAGATTTATATCGTTTATTT-3′), (568 bp product); ERS2(At1g04310), (5′-CAGCTGAAGTTTATAGAGCATT-3′) and (5′-CGTCAATGATCAGTGGCTAG-3′), (594 bp product); and EIN4(At3g04580), (5′-AGATTAATGAGATTCAGAA-3′) and (5′-TACTTCGCAGCCGAGTTTCTCAA-3′), (462 bp product).

Testing the semiquantitative RT-PCR assay with cDNA standards generated by PCR

To verify this methodology for the Arabidopsis ethylene receptors, PCR cDNA constructs were designed from primers that flank the primers used for receptor quantification. All the receptor PCR constructs were combined at equal concentrations, and a concentration series (40 × 103, 10 × 103, and 2.5 × 103 molecules μl−1) was prepared. Each of these dilutions was amplified against the genomic standard at the same concentration series (40 × 103, 10 × 103, and 2.5 × 103 molecules μl−1). The cDNA copy numbers quantified with this methodology were within 20% of the actual DNA concentration after correcting for product size differences. The intergenic comparison of the five receptors were within 25% of each other.


This work supported by NSF grant (MCB-0131564) to A.B.B.