Ethylene biosynthesis is directed by a family of 1-aminocyclopropane-1-carboxylic acid (ACC) synthases (ACS) that convert S-adenosyl-l-methionine to the immediate precursor ACC. Members of the type-2 ACS subfamily are strongly regulated by proteolysis with various signals stabilizing the proteins to increase ethylene production. In Arabidopsis, this turnover is mediated by the ubiquitin/26 S proteasome system, using a broad complex/tramtrack/bric-a-brac (BTB) E3 assembled with the ETHYLENE OVERPRODUCER 1 (ETO1) BTB protein for target recognition. Here, we show that two Arabidopsis BTB proteins closely related to ETO1, designated ETO1-like (EOL1) and EOL2, also negatively regulate ethylene synthesis via their ability to target ACSs for breakdown. Like ETO1, EOL1 interacts with type-2 ACSs (ACS4, ACS5 and ACS9), but not with type-1 or type-3 ACSs, or with type-2 ACS mutants that stabilize the corresponding proteins in planta. Whereas single and double mutants affecting EOL1 and EOL2 do not show an ethylene-related phenotype, they exaggerate the effects caused by inactivation of ETO1, and further increase ethylene production and the accumulation of ACS5 in eto1 plants. The triple eto1 eol1 eol2 mutant phenotype can be effectively rescued by the ACS inhibitor aminoethoxyvinylglycine, and by silver, which antagonizes ethylene perception. Together with hypocotyl growth assays showing that the sensitivity and response kinetics to ethylene are normal, it appears that ethylene synthesis, but not signaling, is compromised in the triple mutant. Collectively, the data indicate that the Arabidopsis BTB E3s assembled with ETO1, EOL1 and EOL2 work together to negatively regulate ethylene synthesis by directing the degradation of type-2 ACS proteins.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Ethylene is an important gaseous hormone that regulates numerous aspects of plant growth and development. Under normal conditions, it controls hypocotyl and root elongation, apical hook formation, root hair development, leaf and flower senescence/abscission, stem cell division in roots and fruit ripening (Abeles et al., 1992; Johnson and Ecker, 1998; Schaller and Kieber, 2002; Ortega-Martinez et al., 2007). Through its ability to readily permeate through cell membranes, ethylene also provides a mechanism for plants to rapidly and coordinately respond to adverse environments, such as pathogen attack, wounding, hypoxia, and exposure to drought and cold. Consequently, both the synthesis and perception of ethylene are tightly controlled within plants.
Extensive genetic analyses with Arabidopsis have uncovered elaborate pathways that mediate both ethylene synthesis and perception (Argueso et al., 2007; Benavente and Alonso, 2006). Ethylene perception employs an inverse agonist mechanism that begins with a family of five receptors related to bacterial two-component histidine kinases. In the absence of ethylene, they activate CONSTITUTIVE TRIPLE RESPONSE-1 (CTR1), a Raf-type mitogen-activated protein kinase kinase kinase (MAPKKK) that inhibits ethylene signaling (Chang et al., 1993; Hirayama et al., 1999; Hua et al., 1995, 1998; Kieber et al., 1993). Binding of ethylene blocks this activation of CTR1, thus allowing ethylene signaling to commence towards ETHYLENE INSENSITIVE (EIN)-2. EIN2 is a membrane-associated positive regulator related to the N-ramp family of divalent cation transporters (Alonso et al., 1999), the action of which remains unclear. It interacts with a putative CONSTITUTIVELY PHOTOMORPHOGENIC (COP)-9 signalosome component (at least in yeast) that may modulate its activity (Christians et al., 2008). Ultimately, the pathway converges towards the transcription factors EIN3 and EIN3-LIKE (EIL)-1 that induce the expression of various ethylene-regulated genes, including ERF1, which encodes a key ethylene-response transcription factor (Chao et al., 1997; Yanagisawa et al., 2003). Both the EIN3 and EIL1 proteins are continually synthesized; however, in the absence of ethylene they are rapidly degraded to prevent ethylene responses. This turnover is accomplished by the ubiquitin (Ub)/26 S proteasome system (UPS), using a pair of Skp/Cullin/F-Box (SCF) E3s assembled with the EIN3-BINDING F-BOX (EBF)-1 and EBF2 F-box proteins that coordinately ubiquitinate EIN3/EIL1 (Binder et al., 2007; Gagne et al., 2004; Guo and Ecker, 2003; Potuschak et al., 2003). Ethylene blocks this breakdown, possibly by the activation of a MAPK cascade containing MPKK9 and MPK6 that phosphorylates EIN3/EIL1, thus protecting EIN3/EIL1 from SCFEBF1/2 recognition (Yoo et al., 2008). In this way, ethylene induces the rapid accumulation of EIN3/EIL1 to promote the transcription of ethylene response genes.
In a similar fashion, ethylene production is also tightly controlled using protein degradation as a key regulatory mechanism. The first committed and generally rate-limiting step of ethylene biosynthesis involves the conversion of S-adenosyl-l-methionine to the immediate ethylene precursor 1-aminocyclopropane-1-carboxylic acid (ACC) by a family of ACC synthases (ACSs) (Schaller and Kieber, 2002). Conversion of ACC to ethylene is then catalyzed by members of the ACC oxidase (ACO) family (Argueso et al., 2007). Consistent with their importance to ethylene production, multiple forms of both ACSs and ACOs are synthesized with unique expression patterns, and post-transcriptional controls. For example, eight and 10 loci express catalytically active ACS proteins in Arabidopsis and tomato, respectively, and each has distinct transcriptional profiles in response to cell/tissue type, developmental state and hormone treatments (Liang et al., 1995; Tsuchisaka and Theologis, 2004; Wang et al., 2005; Yamagami et al., 2003).
Members of the ACS family have been further categorized into three types (types-1, −2 and −3) based on the presence of distinct consensus sequences near their C termini (Yoshida et al., 2005). For the type-1 and type-2 ACSs, it is now apparent that these C-terminal sequences mediate the stability of the corresponding isoforms. For example, the signature C-terminal extensions in type-1 ACSs from Arabidopsis (ACS1, ACS2 and ACS6) are phosphorylated by the MAPK MPK6, and possibly by a calcium-dependent protein kinase (CDPK) (Hernandez Sebastia et al., 2004; Liu and Zhang, 2004). This modification dramatically elevates type-1 ACS protein levels and ethylene production in planta, possibly by protecting the enzymes from the UPS. In support, a phosphomimic substitution of the phosphorylation site(s) in ACS6 acts as a gain-of-function mutant that elevates both ACS6 protein levels and ethylene synthesis (Joo et al., 2008). Similarly, dominant overproduction mutations, such as Arabidopsis eto2 and eto3, were found to affect a conserved C-terminal extension in the type-2 ACSs, ACS5 and ACS9, respectively. These mutants dramatically increase ethylene production, not by increasing ACS5 or ACS9 transcription, but by stabilizing the encoded proteins. This stability is mediated in part by protecting the proteins from rapid breakdown by the UPS (Chae et al., 2003; Vogel et al., 1998; Wang et al., 2004; Woeste et al., 1999).
Details about how ACS protein levels are controlled mechanistically by the UPS are now emerging. Important progress was made recently for the type-2 clade from an analysis of a set of recessive mutants affecting the ETHYLENE OVERPRODUCING-1 (ETO1) locus. The eto1 mutants were isolated based on their constitutive ethylene response phenotype in the absence of exogenous ethylene, which could be rescued by treatment of mutant seedlings with the ethylene biosynthesis inhibitor aminoethoxyvinylglycine (AVG), suggesting that the encoded protein dampens ethylene production (Guzman and Ecker, 1990). Wang et al. (2004) subsequently discovered that the ETO1 protein is a member of the broad complex/tramtrack/bric-a-brac (BTB) protein superfamily that participates in substrate recognition via the UPS. BTB proteins assemble with the Cullin-3 isoform (CUL3a and CUL3b in Arabidopsis; Gingerich et al., 2005) and the Ub-conjugating enzyme (or E2) docking protein RBX1 to form Ub-protein ligases (or E3s) (Smalle and Vierstra, 2004). Specific protein–protein interaction motifs present in the BTB proteins confer selectivity towards their substrates. The observations that: (i) ETO1 interacts with CUL3a/b and type-2 ACSs such as ACS5; (ii) that this interaction requires the C-terminal extension of ACS5, known to be important for stability; and (iii) that eto1 mutants have increased levels of ACS5 protein and increased ethylene production, whereas ETO1 overexpression lines have decreased levels of ACS5 protein and decreased ethylene production, point to the BTBETO1 complex directly participating in the ubiquitination and subsequent degradation of type-2 ACSs (Chae et al., 2003; Wang et al., 2004; Yoshida et al., 2005, 2006). Data showing that mutants affecting the overall activity of Cullin-based E3s also regulate ethylene synthesis, support this scenario (Bostick et al., 2004; Larsen and Cancel, 2004). However, the observation that ETO1 can inhibit the activity of recombinant ACS5 in vitro suggests that a direct effect on type-2 ACS activities is also possible (Wang et al., 2004).
Phylogenetic studies revealed that higher plants express a large collection of BTB proteins, with 80 and 149 members in Arabidopsis and rice, respectively (Gingerich et al., 2005, 2007). In particular, Arabidopsis contains two potential paralogs of ETO1, designated ETO1-LIKE 1 (EOL1) and EOL2, that share significant similarity to ETO1, including the presence of six tetratricopeptide repeats (TPR), and a coiled-coil motif at their C-terminal ends. With yeast two-hybrid (Y2H) screening we observed that EOL1 can also interact with ACS5, which suggests that these two BTB proteins also direct type-2 ACS ubiquitination and turnover (Wang et al., 2004; Yoshida et al., 2006). However, single homozygous-null mutants affecting EOL1 and EOL2 failed to display obvious growth defects, especially those related to ethylene overproduction, suggesting that EOL1 and EOL2 work outside of ethylene signaling, or have a more subtle effect on ethylene biosynthesis than ETO1 (Gingerich et al., 2005).
Here, we further explored the potential roles of EOL1 and EOL2 on ACS turnover and ethylene biosynthesis in Arabidopsis, via a comparative analysis of the ETO1/EOL1/EOL2 family. In agreement with its role in controlling type-2 ACSs, EOL1, like ETO1, interacts with type-2, but not type-1 and type-3, ACSs, or type-2 mutants that stabilize protein accumulation. Whereas single and double mutants affecting EOL1 and EOL2 do not show obvious ethylene-overproduction phenotypes, they do exaggerate the phenotype of etiolated seedlings defective in ETO1, and further increase ethylene production by eto1 plants. The triple eto1 eol1 eol2 mutant in particular displays a strong ethylene response phenotype, and substantially overproduces the hormone. Based on the accumulation of a Myc-tagged version of ACS5 in the mutant backgrounds, the exaggerated phenotypes of the double and triple mutants are directly caused by the effects of EOL1 and EOL2 on type-2 ACS protein levels. Collectively, the data indicate that the Arabidopsis BTB E3s assembled with ETO1, EOL1 and EOL2 work together to negatively regulate ethylene synthesis, by helping degrade type-2 ACSs.
ETO1 and EOL1 specifically interact with type-2 ACSs
Sequence homology among ETO1, EOL1 and EOL2, including related TPR and coiled-coil regions, have strongly implicated EOL1 and EOL2 in the recognition and turnover of type-2 ACSs by the UPS (Gingerich et al., 2005; Wang et al., 2004). In fact, recent studies have shown that Arabidopsis EOL1, like ETO1, can interact with the type-2 ACS ACS5, and its potential ortholog LeACS3 from tomato, suggesting that the interactions between ETO1-like BTB proteins and type-2 ACSs are conserved among higher plants (Wang et al., 2004; Yoshida et al., 2005, 2006). To further support this scenario, we used Y2H screening to test the binding specificity of EOL1 with members of the type-1 (ACS2 and ACS6), type-2 (ASC4, ASC5 and ASC9) and type-3 (ACS7) ACS clades from Arabidopsis. (Examining EOL2 by our Y2H strategy was not possible because of its strong autoactivation in the bait activation domain vector; data not shown.) As shown in Figure 1, EOL1 has a similar binding specificity as ETO1. It interacted with all members of the type-2 clade, but not with members of the type-1 and type-3 ACS clades. In general, EOL1 induced higher β-galactosidase activity relative to ETO1 in the Y2H pairings, suggesting that EOL1 either binds type-2 ACSs better than ETO1, or is expressed at a higher level in yeast.
Prior genetic and interaction studies have implicated a conserved patch near the C terminus of type-2 ACSs (residues 453–469 in ACS5) as important for their interactions with ETO1 (Wang et al., 2004; Yoshida et al., 2005, 2006). We provided further support for this proposal by testing the interactions of ETO1 and EOL1 with ACS5 missing the last 78 amino acids (ACS5Δ), with ACS5 bearing the eto2 mutation that alters the C-terminal 12 amino acids (Vogel et al., 1998), and with ACS9 bearing the eto3 mutation, which replaces the conserved Val-457 with an aspartic acid residue (Chae et al., 2003). Prior studies showed that the eto2 and eto3 mutations increase ethylene synthesis, with the eto2 mutation increasing the stability of the ASC5 protein (Chae et al., 2003). Here, the eto2 mutation substantially impaired binding of ACS5 to both ETO1 and EOL1, whereas the eto3 mutation impaired the binding of ACS9 to ETO1, but only modestly compromised the binding to EOL1 (Figure 1). The residual binding of ACS5eto2 to ETO1 and EOL1, as compared with the cognate controls, was further reduced by removing the entire C-terminal region in the ACS5Δ deletion (Figure 1).
It has been proposed that the CDPK-dependent phosphorylation of type-2 ACSs is a signal that interferes with ETO1 recognition, thus blocking the presumed ubiquitination and subsequent turnover of the enzymes (Hernandez Sebastia et al., 2004). One possible CDPK site is present in the C-terminal extension of ACS5 at Ser-461. However, introduction of a phosphomimic aspartic acid residue at position 461 (ACS5S461D), or its alanine control (ACS5S461A) that cannot be phosphorylated (S461A) into ACS5, had no significant effect on binding to ETO1 or EOL1 (Figure 1), suggesting that modification of this residue by itself does not strongly modify the interaction of these proteins.
Identification of combinatorial mutants affecting ETO1, EOL1 and EOL2
To assess the roles of EOL1 and EOL2 in planta, we generated various combinations of the T-DNA insertion mutations eto1-13, eol1-1 and eol2-2, previously described by Gingerich et al. (2005) in the Col-0 ecotype (Figure 2a). These insertions disrupt the respective coding regions, and block the detectable accumulation of the corresponding transcripts, strongly suggesting that they represent null alleles. To obtain each double mutant, the single eto1-13, eol1-1 and eol2-2 mutants were crossed in various combinations, and then the desired homozygous combinations were identified in F2 populations. The triple homozygous mutant (eto1-13 eol1-1 eol2-2) was created from an eol1-1 eol2-2 to eol1-1 eto1-13 cross. As can be seen in Figure 2b, genotyping of two representative triple mutant individuals by genomic PCR with gene-specific and T-DNA left-border primers confirmed that these plants were homozygous for the three insertions.
When grown on soil under either long (16-h light/8-h dark) or short (8-h light/16-h dark) days, only plants harboring the eto1-13 mutation were visibly affected (Figure 2c, d and data not shown). The single eol1-1 and eol2-2 mutants and the eol1-1 eol2-2 double mutant were indistinguishable from the wild type, both as 4-week-old rosette plants and as 6-week-old plants with mature inflorescences. The homozygous eto1-13 eol1-1 and eto1-13 eol2-2 double mutants and the eto1-13 eol1-1 eol2-2 triple mutant were morphologically similar to the eto1-13 single mutant, including a compact rosette with smaller leaves, reduced petiole lengths and shorter inflorescences. Although slightly fewer seeds were produced in the various eto1-13 combinations, even the triple mutant was fertile, with no evidence of increased seed abortion (data not shown). Seed germination of the triple mutant also appeared to be robust. The eto1-13 combinations were less affected than the constitutive ethylene signaling mutant ctr1-1, which has substantially decreased rosette size, darker color, reduced inflorescence height and compromised fertility (Figure 2c,d; Kieber et al., 1993). Taken together, it appears that ETO1 is mainly responsible for controlling type-2 ACS levels in light-grown plants, with EOL1 and EOL2 playing insignificant roles, at least under the growth conditions tested.
Ethylene responses of etiolated eto1, eol1 and eol2 mutant seedlings
Wild-type etiolated seedlings exposed to exogenous ethylene, or mutant seedlings that either overproduce ethylene or constitutively activate the ethylene response pathway, display a characteristic triple-response phenotype consisting of shorter, thicker hypocotyls, shorter roots with increased numbers of longer root hairs, and an exaggerated apical hook (Guzman and Ecker, 1990). This phenotype can be easily seen in 3-day-old etiolated ctr1-1 and eto1 seedlings (Figure 3; Kieber et al., 1993). Neither the eol1-1 and eol2-2 single mutants nor the eol1-1 eol2-2 double mutant displayed even a mild form of the triple response, suggesting that both EOL1 and EOL2 contribute little to the ethylene response of etiolated plants. However, introgression of the eol1-1 and eol2-2 mutations into the eto1-13 background appeared to accentuate the triple response of eto1-13 seedlings, as quantified by hypocotyl elongation (Figure 3b), indicating that EOL1 and EOL2 have a detectable role that only becomes apparent in the absence of ETO1.
To confirm that the accentuated effect is related to ethylene synthesis, and not to ethylene perception or unrelated responses, we examined the response of the mutant combinations to the potent ethylene biosynthesis and perception inhibitors AVG and silver (AgNO3), respectively. Silver locks the ethylene receptor in an active state that keeps CTR1 functional to repress signaling (Schaller and Kieber, 2002). Although silver is effective in wild-type seedlings, even high concentrations cannot block the constitutive signaling of the ctr1-1 mutant (Figure 3a, c; Kieber et al., 1993). When 3-day-old eto1-13, eol1-1 and eol2-2 mutant combinations were treated with silver, the triple response was mostly eliminated, indicating that ethylene responsiveness was normal, even for the triple mutant. The mutant combinations were also rescued phenotypically upon treatment with 10 μm AVG. In contrast to the short hypocotyls and exaggerated apical hooks of ctr1-1 seedlings, the hypocotyl lengths and apical hooks of the double and triple mutant seedlings exposed to AVG were comparable with the wild type (Figure 3a, b). Taken together, it appears that the eol1-1 and eol2-2 mutations accentuate the triple response of eto1-13 seedlings solely by increasing ethylene synthesis.
Based on the likely roles of ETO1, EOL1 and EOL2 in directing the turnover of type-2 ACSs, the higher-order mutant combinations assembled with eto1-13, eol1-1 and eol2-2 should have higher ACS activities than the eto1-13 single mutant, and thus require higher AVG concentrations to rescue their triple-response defects. The AVG dose–response data in Figure 3b supports this notion. As an example, hypocotyl lengths of the eto1-13 seedlings in 0.25 μm AVG were comparable with wild-type Col-0 (1.0 versus 1.1 cm), whereas the hypocotyl lengths of the eto1-13 eol1-1 and eto1-13 eol2-2 double mutants, and the eto1-13 eol1-1 eol2-2 triple mutant, were only partially rescued (0.8, 0.7 and 0.4 cm, respectively). Even at the highest AVG concentration used (20 μm), several of the combinatorial mutants were shorter than the wild-type and single-mutant seedlings.
To further demonstrate that ethylene synthesis, and not perception, is affected upon elimination of ETO1, EOL1 and EOL2, we attempted to phenocopy the growth inhibition response of the mutants by exposing etiolated seedlings to ethylene or ACC. As shown in Figure 4a, wild-type seedlings, the eol1-1 and eol2-2 single mutants, and the eol1-1 eol2-2 double mutant, became progressively shorter as the concentrations of ethylene or ACC increased. At saturating concentrations of each (100 μL L−1 and 10 μm, respectively), the hypocotyl lengths of all single, double and triple mutants were similar to that of ctr1-1 seedlings in air. We also tested the response of the eto1-13, eol1-1 and eol2-2 mutant combinations to exogenous ethylene under growth conditions in which endogenous ethylene synthesis was inhibited by AVG (10 μm), to further confirm that ethylene perception was normal. The inhibition of hypocotyl growth of the mutants in response to 100 μL L−1 ethylene was nearly indistinguishable from each other, and from the wild type (data not shown).
The kinetics of ethylene perception were also similar for the triple mutant compared with the wild type and the eto1-13 single mutant. As described by Binder et al. (2004a,b), using a time-lapsed imaging system, 3-day-old etiolated wild-type seedlings treated with 10 μm AVG respond rapidly to exogenous ethylene, with 10 μL L−1 inducing a detectable growth inhibition within 10 min (Figure 4b). The elongation rate reaches a transient plateau within 30 min (phase I), and then decreases further to an even lower elongation rate (phase II), which can be maintained for many hours by continuous ethylene exposure. Upon removal of ethylene, the elongation rate rapidly rises close to the pre-treatment rate, after a lag of approximately 30 min, followed by a dampening oscillation as the elongation rate re-equilibrates. A variety of mutants can substantially affect this response, including mutants in ethylene signaling components (ein2 and ein3), and factors that that affect their stability (ebf1, ebf2 and ein5/xrn4) (Binder et al., 2004a,b, 2007; Potuschak et al., 2006). eto1-13 seedlings displayed similar growth kinetics to the wild type, indicating that removal of ETO1 does not perceptibly affect the various components of ethylene signaling (initiation, phase I, phase II and recovery) (Figure 4b). The eto1-13 eol1-1 eol2-2 triple mutant also behaved similarly with indistinguishable onset times for initiation and recovery (Figure 4b). The only substantial differences were the magnitudes of the initial elongation rate before ethylene exposure, and the elongation rate recovery after ethylene removal. These differences presumably reflect the high levels of type-2 ACSs in the triple mutant background, which cannot be adequately blocked by AVG even with 10 μm concentrations (see Figure 3b).
Ethylene production is regulated by the entire ETO/EOL1/EOL2 gene family
Both the Y2H screening and growth assays with etiolated seedlings are consistent with EOL1 and EOL2 working in conjunction with ETO1 to direct the ubiquitination and subsequent degradation of type-2 ACSs, with ETO1 being the major contributor. Consequently, we predicted that the eol1-1 and eol2-2 mutations would enhance ethylene production when introduced into the eto1-13 mutant background. To test this possibility, we measured ethylene production from etiolated seedlings harboring the various mutant combinations. Consistent with previous studies and our phenotypic data from etiolated and green plants, wild-type Col-0 seedlings produce little ethylene, whereas the eto1-13 seedlings synthesized high levels of ethylene (Figure 5). The eol1-1 and eol2-2 single mutant seedlings, and the eol1-1 eol2-2 double mutant seedlings, also released little ethylene. A slight but repeatable increase in emission over that seen from the wild type was detected for plants harboring the eol2-2 mutation, suggesting that EOL2 contributes some to type-2 ACS turnover, even in the presence of ETO1. However, as suspected from the phenotypic data, a substantial increase in ethylene release was observed when the eol2-2 mutation was added to etol1-13, and then when eol1-1 was added to the eol2-2 eto1-13 double mutant (Figure 5). The triple mutant released the most ethylene, which was almost twice that of eto1-13 eol2-2 and more than 140-fold higher than that of the wild type. The effect of EOL1 was observed only when both ETO1 and EOL2 were eliminated, suggesting that this BTB protein contributes the least to directing type-2 ACS proteins towards breakdown in etiolated seedlings.
ACS5 accumulation is affected by the entire ETO1/EOL1/EOL2 family
To confirm that the stability of type-2 ACS proteins is regulated, in part, by EOL1 and EOL2, the steady-state levels of an epitope-tagged ACS5 protein were assessed in the single, double and triple mutants. The homozygous eto1-13 eol1-1 eol2-2 line was crossed with a Col-0 line expressing an N-terminal 6X Myc-tagged ACS5 protein, under the control of a dexamethasone (Dex)-inducible promoter (Chae et al., 2003; JJK, unpublished data). Lines containing the MYC-ACS5 transgene, together with the various combinations of the homozygous eto1-13, eol1-1 and eol2-2 mutations, were then isolated from a selfed F2 population, and were tested for Myc-ACS5 protein accumulation with anti-Myc antibodies. In the absence of Dex, the expression of the MYC-ACS5 transgene was not detected by semi-quantitative reverse-transcription (RT)-PCR of total seedling RNA, using a primer specific for the appended region encoding the Myc epitope, in combination with a primer within the ACS5 coding region (Figure 6a). However, when grown on 12 nM Dex (a concentration sufficiently low to avoid saturating the expression of the MYC-ACS5 transgene in the wild-type seedlings; Chae et al., 2003), relatively similar levels of expression (compared with the control UBC21 transcript) were evident in all wild-type and mutant plants harboring the MYC-ACS5 transgene (Figure 6a).
When crude extracts from Dex-treated seedlings were subjected to immunoblot analysis with anti-Myc antibodies, the Myc-ACS5 protein at the expected mass of approximately 62 kDa could also be detected in plants harboring the transgene (Figure 6b). Whereas the level of the Myc-ACS5 protein was low in wild-type plants, its levels were increased in all backgrounds harboring mutations in ETO1, EOL1 and EOL2, either singly or in combinations. In accord with the ethylene synthesis measurements, the eto1-13 mutation had the strongest effect on Myc-ACS5 levels, with the triple mutant and eto1-13 eol2-2 double mutant containing the highest protein levels. However, whereas the effects of eol1-1 and eol2-2 mutations by themselves were not significant, with respect to ethylene synthesis, we could detect a small increase in Myc-ACS5 protein levels in the backgrounds containing the eol1-1 and eol2-2 mutations alone, or in combination (Figure 6b). These increases indicate that EOL1 and EOL2 can affect type-2 ACS protein levels, even in the presence of ETO1, but that these increases are not enough to substantially alter ethylene production.
Expression patterns of the ETO1/EOL1/EOL2 family
There are a number of possible reasons why ETO1, EOL1 and EOL2 might contribute differentially to ethylene synthesis, including the preference of each for specific members of the type-2 ACS family, different expression patterns for the corresponding genes and/or the possibility that each BTB protein accumulates to different levels. To examine the expression patterns, we assayed the abundance of each transcript by analysis of various transcriptome databases, RNA gel blot analysis and quantitative real-time PCR. The expressed sequence tag (EST) database (http://www.arabidopsis.org) currently has 31, 23 and 17 cDNA representatives for ETO1, EOL1 and EOL2, respectively (as of 10 May 2008), indicating that all three genes are expressed at reasonably similar low levels in Arabidopsis. RNA gel blot analysis of the total RNA showed that ETO1 and EOL1 have similar expression patterns in a range of tissues and developmental stages, with the highest levels found in flowers (Figure 7a). Levels of both were near the detection limit in 10-day-old green seedlings, and were not induced upon exposure to 10 μm ACC, as compared with ERF1, which is known to be ethylene induced (Solano et al., 1998). Transcript levels for EOL2 were noticeably different from ETO1 and EOL1, and appeared to be more constitutively expressed in green and etiolated seedlings (Figure 7a).
Real-Time PCR analysis qualitatively supported these expression patterns. For example, ETO1 and EOL1 displayed similar patterns of transcript abundance, with the highest levels of each being found in flowers, whereas EOL2 was more constitutively expressed over a broad range of tissues (Figure 7b). The abundance of the ETO1, EOL1 and EOL2 transcripts was not affected by ACC, indicating that the degradation pathway requiring these BTB proteins is not upregulated by ethylene. Analysis of publicly available microarray databases (Zimmermann et al., 2004) also indicated that the abundance of the ETO1 and EOL1 transcripts is not substantially altered by various stresses (except by hypoxia, where a greater than twofold decrease was reported for EOL1) or exposure to various plant hormones (data not shown). In contrast, transcript levels for EOL2 did show some significant changes, including a decrease upon treatment of seedlings with hydrogen peroxide (greater than fivefold), infection with Botrytis cinerea (greater than twofold), and low potassium and glucose levels (greater than two- and greater than fourfold, respectively). Conversely, several hormones increased EOL2 mRNA levels, including methyl jasmonate (greater than sixfold), gibberellin (greater than threefold), and auxin and zeatin (greater than fourfold) (data not shown). Taken together, the data suggest that expression of EOL2 is controlled by a number of factors, whereas the expression of ETO1 and EOL1 are under more developmental control, and are less affected by the external environment.
Prior studies have implicated the BTB proteins EOL1 and EOL2, along with ETO1, in the proteolytic control of type-2 ACSs, through the direct interaction of the BTB proteins with the C-terminal ends of the ACSs (Chae et al., 2003; Wang et al., 2004; Yoshida et al., 2005). Here, we provide further support for this mode of action by showing that EOL1 and EOL2 contribute to the control of the ethylene-associated phenotype of etiolated seedlings, ethylene biosynthesis and the accumulation of a type-2 ACS (ACS5) protein through the analysis of single, double and triple mutant combinations, affecting ETO1, EOL1 and EOL2. Whereas homozygous eol1-1, eol2-2 and eol1-1 eol2-2 plants are phenotypically indistinguishable from the wild type when grown in the dark or in the light, the triple mutant combination with eto1-13 displays a stronger ethylene overproduction phenotype in etiolated plants, produces significantly more ethylene and accumulates more of Myc-ACS5 reporter, compared with the eto1-13 single mutant. Like ETO1, EOL1 specifically interacts via Y2H with the Arabidopsis type-2 ACSs ACS4, ACS5 and ACS9, and not with members of the type-1 (ACS2 and ACS6) and type-3 (ACS7), clades. This interaction is largely abolished by removal of the C-terminal domain of ACS5, or by C-terminal mutations known to stabilize ACS5 and ACS9 in planta.
Collectively, the data indicate that the corresponding BTBEOL1 and BTBEOL2 complexes play a minor role in controlling ethylene synthesis in etiolated seedlings, which only becomes apparent phenotypically upon elimination of the BTBETO1 complex. This accentuated phenotype is best seen upon the addition of the ACS inhibitor AVG, where substantially higher concentrations of the inhibitor were needed to reverse the triple response phenotype. This phenotype was caused by the increased ethylene production resulting from a stronger stabilization of type-2 ACSs, as is shown by the enhanced accumulation of the Myc-ACS5 protein in these mutant backgrounds. Homozygous eto1-13 eol1-1 eol2-2 seedlings treated with AVG to suppress endogenous ethylene synthesis responded normally to exogenous ethylene, and could be phenotypically rescued by the ethylene receptor antagonist silver, indicating that the enhanced phenotype was related to ethylene biosynthesis, and that aspects of ethylene perception or general growth were not altered upon loss of these BTB proteins.
Our Y2H binding studies with ETO1 and EOL1 agree with prior data showing that the interactions of ETO1-type proteins involve the recognition of a conserved sequence close to the C terminus of type-2 ACSs. More specifically, Yoshida et al. (2006) have proposed, through interaction studies with the tomato LeETO1 ortholog of Arabidopsis ETO1, and the type-2 ACS LeACS3, that the signature ‘target of ETO1’ or TOE motif, bearing a near invariant WVFRLSF/W sequence, is the site that directly interacts with the TPR repeats of the ETO1 proteins. When appended to a type-1 ACS (LeACS2) or GFP, the TOE motif appeared to induce instability in planta, strongly suggesting that it can work in the absence of other sequences to target ACSs for degradation.
The abundance of type-1 ACSs may be similarly regulated by C-terminal recognition followed by ubiquitination, but it remains unclear which E3(s) are responsible. For type-1 ACSs, such as Arabidopsis ACS6, it appears that phosphorylation by an MAPK cascade involving MPK6 is a key regulatory step that prevents recognition by the UPS, thus stabilizing the protein (Joo et al., 2008; Liu and Zhang, 2004). The site of phosphorylation has been tentatively mapped to a collection of serines near the C terminus of ACS6, thus providing a simple model where the introduction of negative charges near the E3-recognition site abrogates E3 binding (Joo et al., 2008). A similar mechanism has been proposed for regulating type-2 ACS recognition by the BTBETO1 E3, but for these ACSs, a CDPK has been implicated in the phosphorylation step (Hernandez Sebastia et al., 2004). A possible CDPK phosphorylation site near the C terminus of type-2 ACSs includes a serine that could serve as the phosphate acceptor. However, our data showing that the introduction of alanine, or the phosphomimic aspartic acid, does not detectably affect the Y2H interaction of ETO1 or EOL1 with ACS5, argues that this residue is either not the phosphorylation site or, if modified, that this phosphoserine is not the sole determinant in blocking the interaction of type-2 ACSs with BTB E3s containing ETO1, EOL1 or EOL2.
Phylogenetic studies performed by us and others have identified ETO1-type sequences in numerous plant species, including the fern Ceratopteris richardii (one gene) and the mosses Physcomitrella patens (two genes) and Selaginella moellendorfi (two genes), but not in the green algae Chlamydomonas reinhardtii and Ostreococcus lucimarinus. We also detected two genes encoding potential ACS orthologs in P. patens, although their classification into type-1, −2 or −3 clades was ambiguous. Thus, it is possible that the proteolytic control of ACSs by BTB E3s arose early in land-plant evolution (Gingerich et al., 2007; Yoshida et al., 2006; MJC and DJG, unpublished data). Furthermore, the presence of multiple ETO1-related proteins in plants, ranging from bryophytes to dicots and moncots, suggests that the collective control of ACSs by multiple BTB proteins is common within the plant kingdom.
The question then arises why plants need additional BTB proteins like EOL1 and EOL2, if, in Arabidopsis, ETO1 by itself appears to suffice in controlling type-2 ACS levels (at least phenotypically)? Differential expression patterns could explain part of this redundancy. Although the EST database revealed that ETO1, EOL1 and EOL2 are all expressed at similar levels, RNA-gel blot analysis, real-time PCR and DNA microarray studies indicate that EOL2 and the ETO1/EOL1 pair are differentially expressed, with EOL2 being more responsive to hormones and several environmental stresses. In some situations such as infection with B. cinerea, the drop in EOL2 mRNAs could in turn help increase ethylene synthesis, by slowing the degradation of the target type-2 ACSs. Transcript levels for ETO1, EOL1 and EOL2 are not significantly upregulated by ethylene, implying that the synthesis of these BTB proteins is not induced by ethylene in an effort to dampen ethylene production by a negative feedback loop.
A second possibility is that each BTB protein is specific for individual type-2 ACS targets. Whereas our Y2H data indicate that both ETO1 and EOL1 can interact with members of the type-2 clade, we noticed that the eto3 mutation in ACS9 differentially attenuated binding. Whereas the binding of ACS9eto3 to EOL1 was only partially compromised, as compared with ACS9WT, its interaction with ETO1 was strongly blocked, suggesting that ETO1 and EOL1 have different affinities for the ACS9 isoform. Clearly, more interaction studies will be needed to confirm this possibility. A third possibility is that EOL1 and EOL2 play important roles during specific ethylene responses, not studied in detail here. Although we have examined the mutant combinations throughout their life cycle in search of additional phenotypic defects, it remains possible that we missed more subtle ethylene-related processes selectively controlled by EOL1 and EOL2.
Our studies with ETO1, EOL1 and EOL2 provide another example for the importance of ubiquitination in plants, and the combinatorial exploitation of specific E3s in target recognition. Prior studies with Arabidopsis have shown that the TIR1/AFB1-5 family (Parry and Estelle, 2006), EBF1 and EBF2 (Gagne et al., 2004; Guo and Ecker, 2003; Potuschak et al., 2003), and SNEEZY and SLEEPY F-box proteins (Dill et al., 2004; Strader et al., 2004) assemble several SCF E3 isoforms that work collectively to ubiquitinate the cognate families of target AUX/IAA, EIN3/EIL1 and Della proteins, as mechanisms to regulate auxin, ethylene and gibberellin signaling, respectively. Some of the TIR1/ABF1–ABF5 family may have differential affinities for their targets (Parry and Estelle, 2006), whereas EBF1 and EBF2 appear to work in temporally distinct manners (Binder et al., 2007). As with ETO1, these mechanisms of regulation are apparent in more ancient land plants, indicating that the complex use of ubiquitination for developmental control appeared early in plant evolution. Such complexity in turn may help explain the extremely large numbers of E3s that possibly exist in plants, and suggests that the combinatorial use of E3s to direct the turnover of families of key targets is common to the plant kingdom (Smalle and Vierstra, 2004). For the ethylene system alone, the UPS in general, and individual E3s in particular, have now been implicated in controlling the levels of the ethylene receptors, EIN3/EIL1, and type-1 and type-2 ACSs, with additional targets likely to be found in the future (e.g. type-3 ACSs, ERFs and ACOs; see above and Chen et al., 2007), indicating that proteolysis plays a pivotal role in regulating signaling by this hormone.
Yeast two-hybrid (Y2H) analysis
Arabidopsis type-1 (ACS2 and ACS6 ), type-2 (ACS4, ACS5 and ACS9 ) and type-3 (ACS7 ) ACSs (Yoshida et al., 2005), and ETO1, EOL1 and EOL2, cDNAs were cloned into the pEG202-GW bait or pJG4-S-GW prey plasmids, respectively (Gyuris et al., 1993; Holt et al., 2005), and were then transformed into the yeast strain pSH18-34 by the LiCl method (Gyuris et al., 1993). Three independent transformants for each assay were then analyzed for interactions using the β-galactosidase assay, in combination with the substrate O-nitrophenyl-β-galactoside (ONPG; Clontech Laboratories, http://www.clontech.com). Each colony was grown overnight in 1 ml of non-induction liquid drop-out medium (–histidine/uracil/tryptophan; MP Biomedicals, http://www.mpbio.com) at 28°C, diluted to an OD600 of 0.2 in 3 ml of inductive drop-out medium (2% galactose, 1% raffinose, –histidine/uracil/tryptophan), and were then grown overnight. The culture was concentrated into 1 ml of 100 mm sodium phosphate (pH 7.0), 10 mm KCl, 1 mm MgSO4, and then 4 mg ml−1 ONPG was added. After 50 min, the reaction was stopped by the addition of 1 m Na2CO3, and the absorbance of the culture was measured at 420 nm, and then standardized for cell number by OD600. The Y2H data represent the average of three separate experiments, with the omission of four spurious data points.
Mutant construction and plant growth conditions
The eto1-13, eol1-1 and eol2-2 T-DNA insertion mutants in the Arabidopsis thaliana Col-0 background were as described by Gingerich et al. (2005). Combinatorial mutants were created by genetic crosses. Homozygous lines were identified in the F2 generation by PCR-based genotyping with appropriate primer combinations (Gingerich et al., 2005). Primers included the left border T-DNA sequence (Lba1, TGGTTCACGTAGTGGGCCATCG), forward and reverse primers for ETO1 (#1 TGAGCCATCTTCTCAACCAAATCA and #2 AAAAAGATACCACAAACAAACACA), forward and reverse primers for EOL1 (#3 CAAGCCAAAACCAAAGTGGACA and #4 CTCGAGAAGGCAACCGAATTG), and forward and reverse primers for EOL2 (#5 TGTTGAGTGCCAACATTGCCT and #6 TCTCTTGCTCTCTCCGTCTCCC) (see Figure 2). The eto1-13 eol1-1 eol2-2 mutant was constructed by crossing eol1-1 eol2-2 (male) with eol1-1 eto1-13 (female). Dark-grown seedlings in the F2 generation predicted to overproduce ethylene by their ctr1-like phenotype were genotyped to identify homozygous triple mutants.
For hypocotyl growth assays, surface-sterilized seeds were planted on half-strength MS medium minus sucrose, either with or without various concentrations of AgNO3, AVG or ACC, and were stratified for 4 days in the dark at 4°C. Seeds were exposed to 1–2 h of white light before being transferred to the dark for 3 days at 22°C. Each treatment data point represents the average (± SE) of at least 30 seedlings. For ethylene treatments, seedlings were grown in a continuous gas flow-through system, in which the flow rate was regulated to maintain the various ethylene concentrations. For analysis of adult plants, plants were transferred from the agar plates to soil, and were then exposed to a long-day photoperiod (16-h light/8-h dark).
Hypocotyl elongation kinetics on etiolated seedlings were as described by Binder et al. (2004b). Surface-sterilized seeds were planted on half-strength solid MS medium (pH 5.7) containing 10 μm AVG, but no added sucrose, were stratified for 2–4 days at 4°C, and were then exposed to white light for 2–4 h prior to placing the plates vertically in the dark for 2 days at 22°C. Once the germinating seedlings reached a height of 2–4 mm, growth rates were measured using images captured from either an EDC-1000N CCD camera (Electrim Corp., http://www.electrim.com) or an Infinity 2-1M digital camera (Luminera Corporation, http://www.lumenera.com), in conjunction with light provided by an infrared light-emitting diode. Growth rates were calculated using custom software generated by Dr Edgar Spalding in LabVIEW 5.0 (National Instruments, http://www.ni.com), as previously described (Folta and Spalding, 2001). Average growth rates were calculated using at least four seedlings from a minimum of three separate experiments.
Ethylene production measurements
Surface-sterilized seeds (40–80 seeds) were germinated in 22-ml gas chromatography vials on 0.4% agar containing 3 ml of full-strength MS medium plus 1% sucrose. Following a five-day stratification at 4°C, the seeds were exposed to white light for 4–6 h. The vials were capped, incubated at 22°C for 4 days in the dark, and were then frozen to −20°C. The accumulated ethylene in the vials was measured by gas chromatography, as described in Vogel et al. (1998). Ethylene concentrations were measured three times for each of three replicate vials. Ethylene production was then calculated by dividing the ethylene concentration by the number of seedlings in the vial.
RNA gel blot analysis
Total RNA was extracted from various tissues using the Mini-POP method. Briefly, tissue was ground at liquid nitrogen temperatures, and then homogenized in 1 ml per ∼ 400 mg fresh weight of RNA extraction buffer (100 mm Tris-HCl, pH 8.0, 20 mm LiCl, 100 mm Na4EDTA, 100 nm 2-mercapthoethanol, 16.8 mg ml−1 aurintricarboxylic acid; Sigma, http://www.sigmaaldrich.com). After phenol/chloroform extraction, RNA was precipitated with 3 M LiCl, resuspended in water, and then re-precipitated with ethanol. RNA (10 μg) was electrophoresed in 1% agarose gels, and then blotted and baked onto Hybond-XL nylon membranes (Amersham Biosciences, http://www.amersham.com). Methylene blue staining was performed to assure uniform loading and transfer. The membranes were probed with 32P-labeled RNA produced using the SP6/T7 Riboprobe™ system (Promega, http://www.promega.com), as described in Smalle et al. (2002). cDNA probes for ETO1, EOL1 and EOL2 (500–700 bp) were generated by RT-PCR of Arabidopsis RNA using the primers: ETO1 (GGGACAGACCTCCTTTTGGT and CCTGCCACAACAATCCATAG), EOL1 (GGTCTAGCTCGTGTCCATTTCC and GGCTATTAACACGGCTATGG) and EOL2 (TGCCAACATTGCCTCAATAG and GGTAGCCCTCGAATGATGAA). These products were then cloned into the pGEMTeasy vector (Promega).
Total RNA was extracted from various tissues using the RNeasy plant kit (Qiagen, http://www.qiagen.com), treated with 1 μg of DNase (RQ1 DNase; Promega), and was then converted to cDNA using Superscript II (Invitrogen, http://www.invitrogen.com). The cDNA was treated with Rnase H and precipitated with 50% isopropanol. The cDNA at a dilution of 1:625 was subjected to real-time PCR using the iCycler (Bio-Rad, http://www.bio-rad.com) with iQ Supermix (Bio-Rad), in combination with gene-specific primers for ETO1, EOL1, EOL2, ERF1 and PP2A-1 (At1g59830, the gene encoding protein phosphatase 2A-1). PP2A-1 was used as the sample template because this transcript is evenly expressed throughout the tissues tested (Czechowski et al., 2005). The PCR products were confirmed as correct by agarose gel electrophoresis of the samples, followed by ethidium bromide staining. The relative expression level for each gene was calculated using the Delta–Delta cycle threshold (Ct) method (Wong and Medrano, 2005), using PP2A-1 as the control. The expression levels were standardized to 10-day-old seedlings grown in air (1.0, relative expression) for each gene. Primers used for amplification included: ETO1 (GGTCTTGCACGTGTTTATCACCTG and GTGGTCGTCCATAAGAACTGCGG), EOL1 (GTGCACACCAAGGTCTAGCTCGTG and CCGACTATCCATCAACACTGCGGC), EOL2 (GTGAGGAGATGACAAAGCTGATCGAGAAG and GATCATCCATCAGCACGGCAGCT), ERF1 (CAGTCCCCATTCTCCGGCTTCTC and GATCGGAAGGTCTTGACTATCCGAGTC) and PP2A-1 (GTTCCACACGAAGGACCAATGTGC and CACTGTCACTACGTTCTTTTCCTGACACC).
Immunoblot analysis of Myc-ACS5 protein
The triple mutant eto1-13 eol1-1 eol2-2 was crossed with a Col-0 transgenic plant containing the Dex-inducible MYC-ACS5 transgene system, previously described by Chae et al. (2003). F2 plants hemizygous for the transgene (+/–MYC-ACS5), and homozygous for either wild type or single, double or triple mutant combinations of eol1-1, eol2-2 or eto1-13 were identified and allowed to self cross. Approximately 100 seedlings of the progeny (segregating 1:2:1 for the MYC-ACS5 transgene) were collected and rapidly frozen after growth for 3 days in the dark on half-strength MS medium, supplemented with 12 nM Dex (Sigma-Aldrich). Total protein was isolated by homogenization in 2 μL per seedling of SDS sample buffer (125 mm Tris-HCl, pH 6.8, 4% SDS, 20% glycerol and 10% 2-mercaptoethanol), and was then clarified by centrifugation at 10 000 g. The extracts were subjected to SDS-PAGE and immunoblot analysis with anti-Myc antibodies (Roche, http://www.roche.com) and anti-PBA1 antibodies (Smalle et al., 2002).
RT-PCR of the MYC-ACS5 transcript
Total RNA was extracted using the RNeasy Plant Kit (Qiagen) from 100 3-day-old etiolated seedlings segregating for the MYC-ACS5 transgene. After treatmeant with RQ1 DNase (Promega), 2 μg of the RNA was used for cDNA synthesis in conjunction with an oligo dT(18) primer for first-strand synthesis, and the Superscript II RT enzyme (Invitrogen). Semi-quantitative RT-PCR was performed with this template using ExTaq Polymerase (Takara, http://www.takara-bio.com). The MYC-ACS5 mRNA transcript was amplified using a 5′ primer ending at the ATG initiation codon for ACS5 (TGAATTCGGTACCCCGGGAAATG), in combination with a 3′ primer within the ACS5 coding region (GTGCCTGAATAGATCTCGTCGCTAATG). The control UBC21 mRNA was amplified using primers TCATCCTTTCTTAGGCATAGCGGC and TTAGAGATGCAGGCATCAACAGCG.
We thank Joseph Walker, Melissa Wei, and Dr. Donald Salter for technical help. This research was supported by grants from the National Science Foundation to R.D.V.
(Arabidopsis 2010 Program – MCB-0115870) and J.J.K. (MCB-0541973), and a National Institute of Heath postdoctoral fellowship to D.J.G. (F32-GM68361).