Ethylene receptor degradation controls the timing of ripening in tomato fruit

Authors


(fax +1 352 846 2063; e-mail hjklee@ifas.ufl.edu).

Summary

Fruit ripening in tomato requires the coordination of both developmental cues and the phytohormone ethylene. The multigene ethylene receptor family has been shown to negatively regulate ethylene signal transduction and suppress ethylene responses. Here we demonstrate that reduction in the levels of either of two family members, LeETR4 or LeETR6, causes an early-ripening phenotype. We provide evidence that the receptors are rapidly degraded in the presence of ethylene, and that degradation probably occurs through the 26S proteasome-dependent pathway. Ethylene exposure of immature fruits causes a reduction in the amount of receptor protein and earlier ripening. The results are consistent with a model in which receptor levels modulate timing of the onset of fruit ripening by measuring cumulative ethylene exposure.

Introduction

The plant hormone ethylene is a gaseous molecule that regulates multiple processes including germination, organ senescence, stress responses and fruit ripening (Abeles et al., 1992). The role of ethylene in fruit ripening has been intensively studied in a number of species, but most notably tomato, which has emerged as an important model for the study of fleshy fruit development. Ethylene plays a critical role in determining the timing of ripening, and thus provides an attractive point to control fruit ripening through genetic modification.

Climacteric fruits such as tomato are characterized by an increase in respiration and a concomitant increase in ethylene biosynthesis just prior to the initiation of ripening. Ethylene is essential for normal fruit ripening in these species, and blockage of either ethylene production or perception leads to improper ripening. In tomato fruits, ethylene has profoundly varying effects depending on the stage of development. There is a distinct developmental switch that occurs upon fruit maturation (Giovannoni, 2001). Exposure of immature fruits to ethylene results in so-called system 1 ethylene synthesis (Yang, 1987) characterized by induction of a set of ethylene-regulated genes (Alba et al., 2005), auto-inhibitory ethylene synthesis and a failure to ripen. In contrast, mature fruits exhibit system 2 ethylene synthesis characterized by induction of a larger set of genes, auto-catalytic ethylene synthesis and ripening. Although applied ethylene does not initiate ripening in immature fruits, it does significantly hasten the onset of maturity (Yang, 1987); the more ethylene to which an immature fruit is exposed, the earlier it ripens. Similar effects have been observed in banana, in which Burg and Burg (1962) demonstrated that treatment of immature green banana fruits shortened the time to ripening relative to untreated controls. The mechanism by which fruits measure cumulative ethylene exposure is unknown.

Plant ethylene receptors are related to bacterial two-component regulators. They are endoplasmic reticulum (ER)-associated integral membrane proteins (Chen et al., 2002) with protein kinase activities (Gamble et al., 1998; Moussatche and Klee, 2004). The receptors are disulfide-linked dimers, and ethylene binding is mediated by a copper co-factor (Rodriguez et al., 1999; Schaller et al., 1995). Genetic analysis in tomato and Arabidopsis has shown that the receptors act as negative regulators of the ethylene response pathway (Hua and Meyerowitz, 1998; Tieman et al., 2000). In the absence of the hormone, receptors actively suppress ethylene responses. Upon ethylene binding, that suppression is removed and the response occurs. In tomato, there are six known ethylene receptors (LeETR1,2,4–6 and NR; Wilkinson et al., 1995; Zhou et al., 1996; Lashbrook et al., 1998; Tieman and Klee, 1999). Based on gene and protein structures, the ethylene receptors have been divided into two subfamilies. In Arabidopsis, the subfamily I receptors, ETR1 and ERS1, are the most homologous to histidine kinases. The subfamily II members, ETR2, EIN4 and ERS2, have lost most of the amino acids critical for histidine kinase activity and instead possess serine kinase activity (Moussatche and Klee, 2004). They also contain an extra potential membrane-spanning domain at the amino terminus. Functional analyses have indicated that Arabidopsis subfamily I members have a more important role in ethylene signaling that cannot be replaced by subfamily II members (Wang et al., 2003). Further, no single loss-of-function mutation has a major effect on ethylene responses, indicating a degree of functional redundancy. However, a completely different picture emerges in tomato, where loss of a single subfamily II receptor, LeETR4, results in increased ethylene sensitivity. Antisense LeETR4 plants show phenotypes consistent with a constitutive ethylene response, including significantly earlier fruit ripening (Tieman et al., 2000). This mutant phenotype can be restored to wild-type by overexpression of the subfamily I receptor, NR. No ethylene-associated developmental effects have been observed in lines with reduced expression of NR (Tieman et al., 2000), LeETR1, LeETR2 or LeETR5 (Tieman and Klee, unpublished results).

The receptor signaling model states that the receptors act as negative regulators of the ethylene response. Experimentally, it has been shown that reduction of receptor content increases ethylene sensitivity (Cancel and Larsen, 2002; Hall and Bleecker, 2003; Hua and Meyerowitz, 1998; Tieman et al., 2000), while increased receptor content has the opposite effect (Ciardi et al., 2000). We have previously shown that NR and LeETR4 transcripts are up-regulated in ripening fruits (Tieman et al., 2000; Wilkinson et al., 1995). As fruit ripening is dependent upon ethylene action, it seems illogical to increase receptor content and thus decrease ethylene responses. To better understand the role of the tomato ethylene receptor family during fruit development, we have characterized the behavior of the receptor RNAs and proteins during fruit development. Contrary to the RNA data, protein blot analysis showed that receptor protein levels are at their highest during immature fruit development and significantly decline at the onset of ripening. This paradox is explained by observations that ethylene treatment induces a rapid degradation of receptor proteins. Here, we present data indicating an important role for LeETR4 and LeETR6 in modulating the timing of ripening. Reduced levels of these receptors mediated by either antisense RNA or protein degradation result in earlier fruit ripening.

Results

A subset of the receptor family shows ripening-associated expression and is ethylene-inducible in fruit

Expression of all six ethylene receptor genes was assayed throughout fruit development to assess stage-specific expression. Quantitative RT-PCR analysis of each receptor transcript found low expression of all receptors throughout immature fruit development, but upon maturation there was a significant increase in NR, ETR4 and ETR6 transcripts (Figure 1). This ripening-associated increase in expression resulted in a 10-fold increase in total receptor mRNA content by the breaker stage. As the receptors are negative regulators of ethylene responses, the observed increases in mRNA levels during an ethylene-dependent process seem counter-intuitive, as an increase in receptors would make the fruit less sensitive to ethylene.

Figure 1.

 Ethylene receptor family mRNA levels during fruit development.
Quantitative RT-PCR analysis for each receptor transcript in fruit tissue from different stages of fruit development. DPA, days post-anthesis; MG, mature green; breaker, first external color change; turning, approximately 30% red color. Expression levels are presented as a percentage of total RNA.

Ripening-associated gene expression may be the consequence of increased ethylene production. Previous analysis has shown that ETR4 and NR are in fact ethylene-inducible in leaf tissue (Ciardi et al., 2000). To determine whether the receptor gene family is regulated by ethylene in fruit tissue, individual fruits were treated with 10 ppm ethylene for 15 h. Expression analysis of each receptor showed 9-, 10- and 7-fold increases in NR, ETR4 and ETR6, respectively (Figure 2). Expression of ETR1, ETR2 and ETR5 changed little in response to the ethylene treatment. Based on this analysis, it appears that expression of NR, ETR4 and ETR6 is the consequence of the climacteric increase in ethylene production at the onset of ripening.

Figure 2.

 Ethylene-inducibility of each receptor mRNA in immature fruit tissue.
Quantitative RT-PCR analysis of expression of each receptor in response to 10 ppm ethylene. Values are presented as a percentage of total RNA (±SE).

LeETR6 antisense lines show phenotypes consistent with a constitutive ethylene response

Single gene knockouts of ethylene receptors in Arabidopsis show no obvious phenotypes, and only the subfamily I double mutant (Hall and Bleecker, 2003; Qu et al., 2007) or triple and quadruple mutants (Hua and Meyerowitz, 1998) show any ethylene-related phenotypes. As previously shown by Tieman et al. (2000), this is not the situation in tomato, as lines with significantly reduced LeETR4 expression show ethylene-hypersensitive phenotypes. When LeETR6 antisense lines were generated, we found similar phenotypes to those seen in LeETR4 antisense lines, including a reduction of time to ripening by as much as 7 days (Table 1). Additional ethylene-related phenotypes include epinastic leaf growth and premature flower senescence (Figure 3). These results indicate that gene-specific reductions in expression of either LeETR4 or LeETR6 but not the other four receptors (data not shown) result in hypersensitivity to ethylene, including premature fruit maturation and ripening.

Table 1.   Days from anthesis to breaker for the LeETR6 antisense lines
LineDaysReduction in LeETR6 mRNA (%)
  1. Values represent means ± SE for at least 15 fruit for each line.

  2. *value < 0.001, for comparison with wild-type, based on Student’s t-test.

Wild-type43.33 ± 0.71 
LeETR6 AS-138.42 ± 0.90*85.1 ± 2.4
LeETR6 AS-237.00 ± 1.46*75.6 ± 6.4
LeETR6 AS-335.83 ± 0.78*72.8 ± 5.8
Figure 3.

 Constitutive ethylene-response phenotypes of LeETR6 antisense lines.
Epinastic leaf growth (a) and early flower senescence (b) of LeETR6 antisense lines. Equivalent-aged wild-type flowers are shown for comparison (c).

Receptor protein levels are distinctly different from transcript levels during fruit development

A wealth of recent work has demonstrated that post-translational control is an important component of hormone pathway regulation. In order to uncover any potential post-translational regulation of ethylene receptors, antibodies against NR, ETR4 and ETR6 were produced. Tissues were collected for a comprehensive study of mRNA and protein expression during fruit development. Measurement of receptor mRNA expression showed an increase in transcript levels at the onset of ripening, and these levels often remained high until fruits were completely red (Figure 4a). Microsomal membranes were isolated to enrich for the receptor proteins, and were used for protein quantification. Analysis of protein levels throughout fruit development revealed an unexpected result; levels were highest during immature fruit development and significantly declined at the onset of ripening (Figure 4b). Data from cv. Flora-Dade are presented, although identical results were obtained in the Pearson and Micro-Tom cultivars. This reduction in protein occurred despite increased RNA content (Figure 4c). The results indicate that RNA levels are not predictive of receptor protein content or the signaling state of the tissue. Rather, there must be an additional level of control of ethylene perception. Because the drop in receptor content coincided with the onset of auto-catalytic ethylene synthesis, we subsequently examined whether ethylene binding induces receptor turnover.

Figure 4.

 Receptor gene expression and protein levels show distinct differences during fruit development.
(a) Quantitative RT-PCR analysis of gene expression expressed as a percentage of total RNA (± SE), and (b) protein blot analysis throughout fruit development in S. lycopersicum cv. Flora-Dade (WT). Levels of RNA and protein are also shown for independent LeETR4 (4AS-1, 4AS-2) and LeETR6 (6AS-1, 6AS-2) antisense lines. Values below each receptor protein blot represent the amount of protein in each lane relative to the IMG stage. BiP antibody was used as a loading control and to normalize protein values.
(c) Ratio of protein to mRNA.
IMG, immature green stage. Protein quantification was performed by densitometric analysis of Western blots using NCBI software ImageJ.

Treatment of leaf and fruit tissue with ethylene causes a rapid degradation of receptor proteins that probably occurs through a proteasome-dependent pathway

To determine whether ethylene binding induces receptor degradation, immature fruits and vegetative tissues were exposed to exogenous ethylene. Ethylene treatment of immature fruits resulted in 4-, 5- and 8-fold increases in NR, ETR4 and ETR6 mRNA, respectively (Figure 5a). Concomitant with this increase in transcripts, there were reductions of 60%, 60% and 40% in NR, ETR4 and ETR6 proteins, respectively, within 2 h, and this reduction was sustained throughout the treatment (Figure 5b). Removal of ethylene after 8 h of treatment lowered transcripts to pre-treatment levels, but levels of receptor proteins remained low even 24 h after treatment ceased (Figure 5a). Ethylene-mediated receptor degradation was also observed in vegetative tissues. Treatment of seedlings with 50 ppm ethylene for 2 h resulted in 10-, 5- and 13-fold increases in NR, ETR4 and ETR6 mRNA, respectively (Figure 6a). Similar to the data collected from immature fruit there were 60%, 40% and 50% reductions in NR, ETR4 and ETR6 protein levels, respectively (Figure 6b). Taken together, the results indicate that ethylene exposure in both vegetative and reproductive tissues results in an immediate drop in receptor protein levels that is independent of transcript levels.

Figure 5.

 Ethylene binding induces degradation of receptors in detached immature fruits.
Fruits were exposed to 10 ppm ethylene for 8 h. The 32 h time point represents fruits that were treated for 8 h and left in air for a further 24 h. (a) Quantitative RT-PCR analysis of gene expression and (b) protein blot analysis of ethylene-treated immature fruits. Values below the protein blots represent the amount of protein in each lane relative to the 0 h time point. BiP antibody was used as a loading control and to normalize protein values. Data represent the results of two independent experiments (± SE).

Figure 6.

 Ethylene binding induces degradation of receptor proteins in vegetative tissue.
(a) Quantitative RT-PCR analysis of gene expression and (b) protein blot analysis of S. lycopersicum cv. Micro-Tom and Never-ripe (Nr) seedlings after treatment with 50 ppm ethylene for 2 h. Data represent the results of two independent experiments (± SE). Values below the protein blots represent the amount of protein in each lane relative to the 0 h time point. BiP antibody was used as a loading control and to normalize protein values.

The 26S proteasome-dependent degradation pathway has emerged as a key point of regulation in many phytohormone signaling pathways (Dharmasiri et al., 2005; Dill et al., 2004; Gagne et al., 2004; Guo and Ecker, 2003; Kepinski and Leyser, 2005). To determine whether this pathway is responsible for the turnover of ethylene receptors, seedlings were treated with the proteasome inhibitor MG132 prior to ethylene treatment. Following ethylene treatment, levels of each protein actually increased, probably because of ethylene-induced increases in transcription/translation (Figure 6b). Very little is known about mechanisms of ER-associated protein degradation in any system (Meusser et al., 2005). Presumably ubiquitinated proteins are rapidly extracted from the membrane and degraded by the cytoplasmic 26S proteasome complex. We did not observe larger ubiquinated forms of immuno-reactive receptors in the microsomal membrane fractions. Even after several-fold concentration, no receptors could be detected in the soluble fraction (data not shown). Nonetheless, the MG132 results are consistent with ubiquitin-mediated receptor degradation.

In order to demonstrate that ethylene binding is necessary for degradation, seedlings were pre-treated with the ethylene action inhibitor 1-methylcyclopropene (1-MCP) prior to ethylene treatment. 1-MCP is a competitive inhibitor of ethylene and its attachment to the receptor is essentially irreversible (Sisler, 2006). If ethylene binding is essential for degradation of the receptor, 1-MCP should stabilize the receptor protein. Pretreatment of tomato seedlings with 1-MCP prevented ethylene-induced receptor degradation (Figure 6b) as well as the ethylene-induced increase in mRNA (Figure 6a), indicating that ethylene binding is essential for receptor degradation. To further confirm that ethylene binding is necessary for protein degradation, we utilized the semi-dominant Never-ripe (Nr) mutant that has a greatly reduced ethylene response. The mutant Nr protein is unable to bind ethylene when heterologously expressed in yeast (Klee and Bleecker, University of Wisconsin - Madison, deceased unpublished data). Treatment of Nr seedlings with 50 ppm ethylene for 2 h caused 50 and 62% decreases in ETR4 and ETR6 proteins, respectively (Figure 6b), but caused significantly less change in the level of NR protein. Taken together, the results are consistent with enhanced receptor degradation following ethylene binding. However, we cannot completely exclude the existence of an ethylene-induced receptor degradation machinery.

Receptor levels in developing fruit determine the timing of ripening

To determine whether ethylene-induced receptor depletion is the cause of the early-ripening phenotype seen in ethylene-treated fruit, immature fruits were exposed to ethylene while still attached to the plant and then allowed to ripen. Protein and mRNA samples were collected throughout the duration of the experiment to determine a possible correlation between lower protein levels and reduced time to ripening. Treated fruits ripened on average 3 days earlier than untreated fruits (Table 2). Receptor protein levels were lower upon treatment with ethylene at 15 days post-anthesis (DPA), and remained lower than untreated controls throughout fruit development, indicating that lower receptor levels correlate with earlier ripening (Figure 7). Transcript data show that the fruits responded to the ethylene treatment, and, upon removal of the ethylene, transcripts returned to pre-treatment levels (data not shown).

Table 2.   Days from anthesis to breaker for ethylene-treated Micro-Tom fruit
TreatmentDays
  1. Values represent means ± SEM for at least 10 fruit for each treatment. The experiment was repeated with similar results.

  2. *P value < 0.05, for comparison with untreated controls, based on Student’s t-test.

Without ethylene45.33 ± 1.41
With ethylene41.20 ± 0.80*
Figure 7.

 Ethylene treatment induces turnover of receptor leading to early ripening fruit.
At 15 days post-anthesis (DPA), fruit were treated with 50 ppm ethylene while attached to the plant. Relative protein expression of NR, ETR4 and ETR6 normalized to an internal control, BiP. Values are plotted relative to the pre-treatment protein level.

Discussion

Upon maturation, tomato fruits undergo a developmental transition that is defined by their response to ethylene (Lincoln et al., 1987). A number of system 1 and/or system 2-associated genes have been identified in fruits. The E4 and E8 genes are excellent examples, with E4 being ethylene-inducible throughout fruit development (both in response to system 1 and system 2 ethylene), and E8 only being ethylene-inducible in mature fruit (system 2-specific). While much is known concerning the role of ethylene during ripening, its function during the immature phase of fruit development is less well understood. When mature fruits are exposed to ethylene, a ripening program is initiated. While treatment of immature fruits does not initiate ripening, it does hasten the onset of ripening; the greater the extent of exposure of the fruit to ethylene, the earlier it ripens (Burg and Burg, 1962; Yang, 1987). How the fruit measures cumulative ethylene exposure is not known. We have provided evidence indicating a specialized role for two receptors, ETR4 and ETR6, in modulating ethylene responses, including fruit maturation. A reduced level of these receptors mediated by either antisense RNA or ethylene-mediated protein degradation results in earlier fruit ripening. Ethylene exposure also resulted in a parallel depletion of the other ethylene-inducible receptor protein, NR. Our results are consistent with a model in which ethylene receptor content is a major determinant of when fruits initiate the ripening program. As the receptors are negative regulators of ethylene signaling, depletion would lead to a progressive increase in hormone sensitivity. When a particular threshold sensitivity is reached, ripening would commence. Alternatively, receptors may act as a brake on ripening initiation. It must be noted that there are other elements independent of ethylene that must also be in place for ripening to initiate, most notably the RIN transcription factor (Vrebalov et al., 2002).

Receptor gene expression is low and constitutive throughout immature fruit development, with little difference between any of the family members (Figure 1). At the onset of ripening, there is an increase in expression of NR, ETR4 and ETR6 that results in a 10-fold increase in total receptor mRNA content. In contrast to mRNA expression, the corresponding protein levels are at their highest in immature fruits, and show a significant decrease at the onset of ripening and remain low (Figure 4b) as a consequence of ethylene exposure. Ethylene binding probably causes a conformational change in the receptors that makes them susceptible to degradation. In this context, it is interesting to note the model of Arabidopsis receptor signaling presented by Wang et al. (2006). These authors provide genetic evidence supportive of a transitional state in which a receptor continues to actively suppress downstream ethylene responses after ethylene is bound. This intermediate state subsequently transitions to a receptor-inactive state. Our results suggest that the ‘transmitter off’ state may actually represent a lack of receptor because of ligand - mediated degradation. It would be most interesting to determine whether the mutations that define this transition state stabilize the protein. The receptor degradation is dependent upon the action of the 26S proteasome. At least in some cases, ubiquitination is associated with the phosphorylation state (Hochstrasser, 1996). Although the ethylene receptors are considered to be ancestral histidine kinases, many do not possess histidine kinase activity (Moussatche and Klee, 2004). However, all of the receptors are functional kinases; those that do not have histidine kinase activity are serine kinases. In light of the degradation of receptors following ethylene binding, it is possible that the phosphorylation state of the receptor may mediate ubiquitin binding. Although ligand-induced receptor degradation has not been reported for plant hormones, it has been observed in animals, where growth hormone (GH) signaling is mediated by receptor levels (Flores-Morales et al., 2006). The GH receptor, like ethylene receptors, is a membrane-associated protein in which hormone binding also increases ubiquitin-mediated turnover (Govers et al., 1999).

The ethylene receptor family in tomato, like Arabidopsis, is split into two groups, with LeETR1, LeETR2 and NR belonging to subfamily I and LeETR4-6 belonging to subfamily II. The Arabidopsis results indicate that there is a distinct difference between subfamily I and II members. With the exception of a subfamily I double mutant (etr1 ers1), single and double gene knockouts in Arabidopsis show no obvious phenotypes. This is probably due to functional redundancy within the gene family. Overexpression of a subfamily II member in an etr1 ers1 double mutant cannot rescue the ethylene-hypersensitive phenotype (Wang et al., 2003). In a reciprocal experiment, overexpression of a subfamily I member in a subfamily II triple mutant was sufficient to rescue the ethylene-response phenotype. Together these data indicate that the subfamily I receptors are more important than the subfamily II receptors in determining competency to respond to ethylene. The Arabidopsis paradigm does not hold for tomato (Figure 3, Tieman et al., 2000). Plants with reduced expression of either LeETR4 or LeETR6, both subfamily II members, show phenotypes that are consistent with an exaggerated ethylene response, including epinastic growth, premature flower senescence and early fruit ripening. Overexpression of NR in a LeETR4 antisense line is able to rescue the ethylene-response phenotype, indicating functional redundancy between subfamily I and II members (Tieman et al., 2000). Apparently there is a large degree of plasticity within the ethylene signaling pathway, and different plants have adapted the signaling components as appropriate for their situation.

Plant hormones are involved in most developmental processes and are critical for abiotic and biotic stress responses. Plants can regulate hormone action through synthesis, catabolism or perception. We have shown that a significant part of the regulation of ethylene responses involves ligand-mediated receptor degradation. Ethylene responses, particularly those related to stresses, are often transitory. In order to shut down an ethylene response, synthesis of new receptors is essential. Our results for ethylene exposure of immature fruits indicate that receptor degradation is apparently an important level of developmental control. Our results also indicate that conclusions concerning receptor functions based on RNA levels must be interpreted cautiously. Whether ethylene-mediated receptor turnover and replenishment are important for other ethylene-mediated processes remains to be determined.

Experimental procedures

Plant materials and growth conditions

S. lycopersicum cv. Flora-Dade, cv. Pearson, LeETR4-AS, LeETR6-AS and LeNR-AS lines were grown in a greenhouse set at approximately 27°C. Individual plants were grown in 3 gal pots that were watered twice a day and supplemented with slow release fertilizer. LeETR6-AS lines were generated by cloning the full-length LeETR6 coding region into a vector in the antisense orientation under the control of the figwort mosaic virus 35S promoter (Richins et al., 1987) and followed by the Agrobacterium tumefaciens nopaline synthase (nos) 3′ terminator. The transgene was introduced into cv. Flora-Dade using the method described by McCormick et al. (1986), with kanamycin resistance as a selectable marker. Transgenic lines with a reduction of >70% in the level of LeETR6 transcript were identified (Table 1). The specificity of the transgene was determined by quantification of every receptor mRNA from leaf tissue. In each case, there was no effect on RNA levels of any other receptor. Data on time to ripening were collected from greenhouse-grown plants in two separate seasons. Open flowers were tagged, and the number of days from anthesis to breaker was recorded. S. lycopersicum cv. Micro-Tom and Nr plants were grown in a growth chamber under standard conditions (16 h day/8 h night). Ethylene treatments of plant material were performed in sealed 38 l tanks. Treatments were performed using either 10 or 50 ppm ethylene, as indicated. These levels are both within the linear response range for NR and LeETR4 ethylene inducibility (Ciardi et al., 2000). Proteasome inhibitor studies were performed by spraying seedlings with an 80 μm MG132 solution (8% DMSO) 4 h prior to 2 h ethylene treatment. Control seedlings were sprayed with an 8% DMSO solution. 1-MCP treatment of seedlings was performed at 1 ppm in a sealed 38 l tank for 16 h prior to 2 h ethylene treatment. Control seedlings were sealed in identical tanks for the same duration of time. All microsomal membrane preparations were obtained immediately after treatment ended.

Recombinant protein expression and antibody production

Coding regions of LeETR4 (amino acids 532–684) and LeETR6 (amino acids 522–688) were amplified with primer pairs ETR4-PF, ETR4-PR, ETR6-PF and ETR6-FR (Table S1) from fruit cDNAs generated using the Clontech One-Step cDNA synthesis kit (http://www.clontech.com/). PCR products were digested with BamHI and BglII, cloned into the Invitrogen pTrcXHisA vector (http://www.invitrogen.com/) and subsequently transformed into the BL21(DE3; Invitrogen) Escherichia coli strain for recombinant protein expression. Cultures (100 ml) were grown at 30°C and induced with 1 mm IPTG for 4 h. Cells were spun down at 8000 g, resuspended in 10 ml lysis buffer (8 m urea), and pulse sonicated for 1 min. The lysate was spun down at 8000 g, and supernatant was purified with an Ni-NTA affinity column as directed by the manufacturer (http://www.novagen.com). Recombinant protein was submitted to Cocalico Biologicals for antibody production in rabbits using their standard protocol (http://www.cocalicabiologicals.com). The antiserum obtained was used to probe both antigens to determine antiserum specificity for its respective antigen.

RNA expression analysis

Total RNA extractions were performed using the Qiagen RNeasy mini kit (http://www.qiagen.com/) with subsequent DNase treatment to remove any contaminating DNA. RNA was quantified by spectroscopy and visually analyzed on ethidium bromide-stained gels to ensure equal concentrations of all RNAs. Quantitative RT-PCR assays were performed using the Applied Biosystems Taqman One-Step RT-PCR kit in an Applied Biosystems GeneAmp 5700 sequence detection system (http://www.appliedbiosystems.com/) as described previously (Tieman et al., 2001). PCR conditions were as follows: step 1, 48°C for 30 min; step 2, 95°C for 10 min; step 3, 95°C for 15 sec and 60°C for 1 min. Step 3 was repeated 40 times. Primer and probe pairs for each gene assayed are listed in Table S1. Levels of LeETR RNAs were quantified using RNAs synthesized by in vitro transcription from plasmids containing the coding region of each gene using a Maxiscript in vitro transcription kit (Ambion, http://www.ambion.com). The total quantity (μg) of in vitro-transcribed RNA was determined, and the in vitro transcription product was used to generate a standard curve in real-time RT-PCR analysis. Results are reported as percent LeETR RNA in total RNA.

Microsomal membrane isolation and protein blot analysis

Microsomal membrane fractions were isolated from fruit or seedlings with a homogenization buffer containing 30 mm Tris (pH 8.2), 150 mm NaCl, 10 mm EDTA and 20% v/v glycerol with protease inhibitors (1 mm PMSF, 10 μg ml−1 aprotinin, 1 μg ml−1 leupeptin and 1 μg ml−1 chymostatin), as described previously (Schaller et al., 1995). Tissue was homogenized at 4°C using a polytron (IKA Labortechnik; http://www.ika.net), and then centrifuged at 8500 g for 15 min. The supernatant was strained through cheesecloth, and then centrifuged at 100 000 g for 30 min. The subsequent membrane pellet was resuspended in 10 mm Tris (pH 7.5), 5 mm EDTA and 10% (w/w) sucrose with protease inhibitors. Protein concentrations were determined using Bio-Rad protein assay reagent (http://www.bio-rad.com/). A 20 μg aliquot of total protein was run out for each sample on a 12% Tris–HCl gel, and proteins were transferred to a nitrocellulose membrane using the Bio-Rad Mini Trans-Blot cell. Membranes were blocked overnight in 10% Carnation milk/TBST at 4°C (Nestle; http://www.nestle.com). Membranes were washed twice for 5 min in TBST, and then incubated with primary anti-ETR4 (1:2000) or anti-ETR6 (1:5000) antibody diluted in 5% Carnation milk/TBST for 1 h. Membranes were subsequently washed three times for 10 min in TBST, and then incubated with peroxidase-conjugated goat anti-rabbit (1:5000) secondary antibody (Kirkegaard & Perry Laboratories; http://www.kpl.com) diluted in 5% Carnation milk/TBST for 45 min. Membranes were finally washed three times for 10 min in TBST. Visualization of signal was performed using Amersham ECL detection reagents (http://www.amershambiosciences.com/) before exposure to film. Quantification of bands was accomplished by using the NCBI imaging software ImageJ (http://rsb.info.nih.gov/ij/). Values were normalized to an anti-BiP (endoplasmic reticulum immunoglobulin binding protein) antibody (generously provided by Alan Bennett, University of California, Davis, USA) which was used as an ER-localized loading control.

Acknowledgements

We gratefully acknowledge provision of the anti-NR antibody by Jack Wilkinson, Pioneer Hi-Bred International. This work was funded by grant number 2005-35304-15988 from the United States Department of Agriculture - National Research Initiative to H.K.

Ancillary