eer4 was isolated as an Arabidopsis mutant with an extreme response to ethylene in dark-grown seedlings that was also found to have partial ethylene insensitivity at the level of ethylene-dependent gene expression, including ERF1. Subsequent cloning of eer4 revealed an inappropriate stop codon in a previously uncharacterized TFIID-interacting transcription factor homologous to human TAF12 and yeast TAF61. Genetic and pharmacological analysis demonstrated that the eer4 phenotype is strictly ethylene dependent in seedlings, yet a double mutant with the partially ethylene-insensitive Arabidopsis mutant, ein3-1, had restored ethylene responsiveness, indicating that eer4 also regulates a previously unknown resetting or dampening mechanism for the ethylene signalling pathway. Consistent with the absolute requirement of EER4 for ERF1 expression, biochemical analysis showed that EER4 is localized to the nucleus where it probably recruits EIN3 and probably other transcription factors along with components of the TFIID complex for expression of a subset of genes required for either manifestation or subsequent dampening of the response to ethylene.
Ethylene is a simple unsaturated gaseous hydrocarbon that mediates several developmental phenomena in plants, including many that are of agricultural importance such as fruit ripening and pathogen response (Abeles et al., 1992). The mechanism underlying ethylene signalling has received considerable attention in the past, partly due to the ease with which mutants with defects in this pathway could be identified through the use of the ethylene-mediated triple response, which consists of hypocotyl shortening and thickening along with apical hook formation in dark-grown seedlings following ethylene signalling. Although this represents an extremely valuable approach that has resulted in identification of several components of this pathway, it is likely that screening for mutants with either ethylene insensitivity or constitutive ethylene response at the level of the triple response has been saturated. Consequently, novel approaches, including isolation of mutants with enhanced ethylene responsiveness (eer class), need to be taken to further our understanding of this apparently complex pathway.
At present, ethylene signalling is considered to be a relatively linear pathway that is mediated by a group of five ethylene receptors that bind ethylene and promote response (reviewed by Guo and Ecker, 2004; Benavente and Alonso, 2006). In the absence of ethylene, these receptors function to maintain a critical negative regulator of the pathway, CTR1, in an active state, thus suppressing ethylene signalling (Kieber et al., 1993; Clark et al., 1998). Progressive loss of the ethylene receptors gives an extreme ethylene-response phenotype that includes adult lethality (Hua and Meyerowitz, 1998). This supports the conclusion that the receptors function to regulate CTR1 (Clark et al., 1998; Gao et al., 2003), although at the same time shows that hyperactivation of this pathway results in a far greater response than what is normally seen for wild-type (wt) plants following ethylene treatment or for ctr1 loss-of-function mutants. This is likely to be related to a poorly defined set of negative regulators that are normally responsible for opposing or dampening the ethylene response, with loss of activation of these factors giving extreme responsiveness due to loss of capability to restrict the level of ethylene signalling.
Although the initial steps following ethylene binding have been intensively studied, there are large gaps in our knowledge regarding the subsequent factors required for progression of the ethylene-mediated signal. For example, CTR1 is speculated to be homologous to the mammalian MAP kinase kinase kinase, Raf1 (Kieber et al., 1993), yet to date there has not been a clear demonstration that a MAP kinase cascade participates in ethylene signalling. EIN2 is a factor of unknown function that is epistatic to CTR1 and is absolutely required for ethylene signalling, with loss of EIN2 function resulting in complete ethylene insensitivity (Alonso et al., 1999). EIN2, which encodes a large integral membrane protein with a hydrophilic domain, is likely to be related in function to the N-RAMP family of metal-ion transporters. Although it is not clear how EIN2 functions, overexpression of its C-terminus in an ein2-5 mutant background results in adult transgenic plants that demonstrate a constitutive ethylene-response phenotype at the visual level yet are insensitive to ethylene at the molecular level, suggesting that ethylene responses proceed not only through promotion of transcription but also by post-translational modification of an existing group of factors (Alonso et al., 1999). This as yet undefined EIN3- and EIL1-independent response to ethylene is supported by analysis of the rapid response to ethylene (Binder et al., 2004), which occurs even in Arabidopsis mutants lacking functional EIN3 and EIL1 (Chao et al., 1997; Alonso et al., 2003), both of which are transcription factors absolutely required for long-term responsiveness to ethylene, including induction of primary response transcription factors such as ERF1 (Solano et al., 1998).
In order to further our understanding of this pathway, a unique approach has been undertaken to identify Arabidopsis mutants with enhanced ethylene responsiveness, with these likely to represent defects in factors required for dampening or resetting the ethylene signalling pathway. To date, several EER factors have been identified including (i) ETR1, the ethylene receptor (Cancel and Larsen, 2002); (ii) EER1/RCN1, a PP2a regulatory subunit (Larsen and Chang, 2001; Larsen and Cancel, 2003); (iii) RTE1, a regulator of ETR1 function (Resnick et al., 2006); and (iv) EBF1 and EBF2, two F-box proteins (Guo and Ecker, 2003; Potuschak et al., 2003; Olmedo et al., 2006; Potuschak et al., 2006; Binder et al., 2007). For all but EBF1, loss-of-function mutants for these factors give a characteristic phenotype of increased ethylene sensitivity and exaggeration of response to ethylene in the triple-response assay. It is not clear why mutations in ETR1 or EER1/RCN1 result in enhanced ethylene response, although each of these is believed to be a key regulator of CTR1 function (Clark et al., 1998; Cancel and Larsen, 2002; Larsen and Cancel, 2003), with the possibility that loss of proper reactivation of CTR1 could result in failure to dampen ethylene response. It is unlikely that ETR1 and EER1/RCN1 solely function to reactivate CTR1 with regard to resetting or dampening the ethylene pathway, since complete loss-of-function ctr1 mutants do not display the phenotypic severity of ethylene-treated eer mutants (Kieber et al., 1993; Larsen and Chang, 2001), thus suggesting that ETR1 and EER1/RCN1 may also be required for modulation of the activity of some other factor that opposes ethylene response. EBF1 and EBF2 target the transcription factors EIN3 and EIL1 for degradation, with loss of EIN3 turnover resulting in inappropriate manifestation and perpetuation of the ethylene signal (Guo and Ecker, 2003; Potuschak et al., 2003; Binder et al., 2007). Interestingly, inappropriate accumulation of EBF1 and EBF2, which results from mutational loss of the 5′→3′ exoribonuclease EIN5/XRN4, leads to ethylene insensitivity due to increased turnover of EIN3 and EIL1 (Olmedo et al., 2006; Potuschak et al., 2006). Continued identification of factors required for regulating the amplitude of response to ethylene will not only increase our knowledge of this critically important plant signalling pathway, but will also give the means to develop novel strategies for controlling the manifestation and/or progression of ethylene response in agriculturally important phenomena such as ripening and senescence.
Materials and methods
For all seedling growth experiments, Arabidopsis seeds (Lehle Seeds) were surface-sterilized as previously described (Larsen and Chang, 2001) and cold stratified at 4 °C for 4 d in the dark to synchronize germination. Seeds were then suspended in 0.15% (w/v) agarose and sown on PNS medium (Larsen and Chang, 2001). For triple-response experiments, which were performed in the dark in static containers, the medium was supplemented with either 5 μM AgNO3 (Sigma Chemical, St Louis, MO, USA) or 5 μM AVG [(S)-trans-2-amino-4-(2-aminoethoxy)-3-butenoic acid hydrochloride] (Sigma Chemical) as required. Ethylene dose–response analysis and analysis of jasmonic acid (JA) responsiveness were done as previously described (Cancel and Larsen, 2002; Larsen and Cancel, 2003).
All adult plants in this study, with the exception of those used for JA responsiveness, were grown in soil under a 24 h light cycle at 20 °C in a plant growth room supplemented with Sylvania Gro-Lite fluorescent bulbs. For treatment of leaves for RNA extraction, adult plants were grown for 4 weeks in air and then treated with either air or 100 μl l−1 ethylene in an airtight chamber (Plas Labs, Lansing, MI, USA) for 24 h. Immediately after treatment, leaf tissue was collected and quick frozen for RNA extraction.
Measurement of ethylene production
Measurement of ethylene production was conducted as previously described (Larsen and Cancel, 2003).
RNA analysis was performed as previously described (Larsen and Cancel, 2003).
Map-based cloning of eer4
For generation of a mapping population, eer4 (male; ecotype Ws) was crossed to Col-0 wt (female), and 4-d-old etiolated F2 seedlings that displayed the eer4 phenotype in the presence of 100 μl l−1 ethylene were isolated and planted in soil for subsequent collection of leaf tissue and isolation of genomic DNA. Genomic DNA was prepared as described (Larsen and Cancel, 2003) and used as a template for PCR-based mapping.
To narrow the map position of eer4, novel CAPS markers were developed in a previously described manner (Larsen and Cancel, 2003) for the region to which eer4 was localized. CAPS markers generated and used were as follows. For the At1g17380 marker, primers used were 5′-CTGATGAGGTAGAGGGTTCGCC-3’ and 5′-GTAAGTGTGTGGAGAATTCTTTCT-3′. Mse1 digestion resulted in three DNA fragments for Col-0 and two for Ws-0. For the At1g17440 marker, primers used were 5′-GATTACTTGTTCGACACGTC-3’ and 5′-CAACAACAATTGCAAAACAAAC-3′. DdeI digestion resulted in three DNA fragments for Col-0 and two for Ws-0.
Candidate genes within the genetic window were identified and amplified by PCR with Pfu Turbo (Stratagene) and primers designed to the predicted 5′- and 3′- untranslated regions (UTRs), subcloned into pGemT-easy (Promega, Madison, WI, USA), and sequenced by the University of Florida DNA Sequencing Core Laboratory. Sequences were compared with the published Arabidopsis genomic sequence to identify the eer4 mutation.
Generation of transgenic plants
Functional complementation was performed by PCR-based generation of a genomic construct representing 1 kb of upstream sequence, the complete coding sequence for At1g17440, and its predicted 3′-UTR. This construct was subcloned into pCGN1547 and introduced into the eer4 mutant by Agrobacterium-mediated transformation. Primary transformants were selected by screening for those that were kanamycin resistant, with T3 plants subsequently analysed by growth in the dark for 4 d in 100 μl l−1 ethylene, after which they were assessed for manifestation of the exaggeration of ethylene response seen for eer4.
Double mutants were generated by crossing eer4 (male) to ctr1-3, ein2-5, or ein3-1 (females). F2 progeny from the crosses with ein2-5 and ein3-1 were grown in the dark for 4 d with 100 μl l−1 ethylene to isolate ethylene-insensitive individuals, which were homozygous for the recessive ein2-5 or ein3-1 mutations. Isolated individuals were subsequently genotyped by PCR to identify those that were homozygous for the eer4 mutation. For PCR analysis, 5′-GTTGTAGCCGAGGCAGAAATC-3’ and 5′-CACAATGTGAGAGAAAGAAG-3’ were used to amplify a portion of the EER4 genomic DNA, after which the PCR products were cut with BsmAI, with restriction digest resulting in two DNA fragments for Col-0 Ws wt and one for eer4.
For generation of the eer4;ctr1-3 double mutant, F2 seedlings were screened in the absence of ethylene for those that exhibited a constitutive triple response. Identified individuals were subsequently genotyped by PCR to identify those that were homozygous for the eer4 mutation.
GFP analysis of protein localization
An EER4:GFP translational fusion was made using an At1g17440 genomic construct representing 1 kb of upstream sequence, the 5′-UTR, and the complete coding sequence for At1g17440 with both introns and exons without the native stop codon. This construct was subcloned into pBI101, transformed into Agrobacterium, and infiltrated into Nicotiana benthamiana leaves. After 3 d, leaves were harvested and treated with air, 5 μM AgNO3, or 100 μl l−1 ethylene for 24 h, stained with 15 mg ml−1 DAPI for 30 min, and visualized using a Leica SP-2 confocal microscope. For control analysis, an unfused GFP construct under the control of the Arabidopsis ACT2 promoter (An et al., 1996) was introduced into tobacco leaves.
Yeast two-hybrid and in vitro binding assays
For the yeast two-hybrid assay, Saccharomyces cerevisiae strain L40 was used as previously described (Clark et al., 1998). Protein fusions were made either to the DNA-binding domain of the bacterial repressor LexA (pLexA-NLS) or to the Gal4 transcription activation domain (pACTII). β-Galactosidase activity was quantified using a CPRG-based assay according to the manufacturer's instructions (Clontech, Palo Alto, CA, USA). In vitro binding assays were performed as previously described using maltose-binding protein (MBP) fusion proteins produced in Escherichia coli BL21-DE3 cells (Clark et al., 1998).
Dark-grown eer4 mutant seedlings are hypersensitive to ethylene
EMS-mutagenized M2 Arabidopsis seeds (Ws ecotype; Lehle seeds) were grown in the dark in the presence of 100 μl l−1 ethylene for 4 d, after which the mutant population was screened for seedlings with extreme hypocotyl shortening. From this initial screen, putative eer mutants were identified and progeny were compared with Ws wt for capability to grow in the dark in air or in the presence of 5 μM AgNO3, 5 μM (S)-trans-2-amino-4-(2-aminoethoxy)-3-butenoic acid hydrochloride (AVG), or 100 μl l−1 ethylene. Putative mutants demonstrating hypocotyl elongation similar to wt in the presence of air, AgNO3, or AVG, but extreme shortening with saturating ethylene, were subsequently chosen as bona fide eer mutants. For this report, the recessive eer4 mutant was chosen for further study.
Analysis of 4-d-old dark-grown eer4 seedlings revealed that in the presence of AgNO3, which is a highly effective inhibitor of ethylene perception, hypocotyl length of the mutant was nearly identical to that of Ws wt, indicating that functional EER4 is not required for normal growth in the absence of ethylene (Fig. 1A). With the addition of increasing amounts of ethylene in the presence of AVG, which is an effective inhibitor of ACC synthase activity and consequently ethylene biosynthesis, Ws wt seedlings were found to require approximately 2-fold more ethylene than eer4 mutants to reach 50% inhibition of hypocotyl elongation (Fig. 1B). At saturating levels of ethylene, eer4 mutants demonstrated extreme exaggeration of response compared with Ws wt, with eer4 hypocotyls being approximately 30% as tall as ethylene-treated Ws wt hypocotyls (Fig. 1C, D). Treatment with propylene, which is an ethylene agonist, resulted in a phenotypic response that was identical to what was seen with ethylene treatment for eer4 seedlings compared with Ws wt (data not shown).
In addition to the severe ethylene-dependent growth defect found in hypocotyls, eer4 roots have extreme root curling even in the absence of exogenous ethylene (Fig. 1E). Analysis of root growth of an eer4;ein2-5 double mutant revealed uncurled roots that were indistinguishable from Ws wt and ein2-5 (data not shown).
The eer4-mutant phenotype does not result from ethylene overproduction
In order to determine if the eer4 phenotype also resulted in increased ethylene production, ethylene was collected for 12 h from 4-d-old dark-grown Ws wt and eer4 seedlings. Following this, ethylene production was determined by measurement of the headspace using a gas chromatography system. eer4 mutants were found to produce approximately 30% more ethylene than Ws wt (Fig. 1F), although this cannot be considered the cause of the eer4 dark-grown seedling phenotype since the growth analyses were done in the presence of an ethylene biosynthesis inhibitor and the exaggerated phenotype for eer4 was observed following treatment with saturating levels of exogenous ethylene.
The eer4 mutation results in reduced expression of a subset of ethylene-responsive genes
Since eer4 demonstrates a profound ethylene-dependent phenotype at the visual level, it was of interest to determine if exaggeration of response was also observed at the level of gene expression. For this analysis, tissue was collected from 4-d-old dark-grown Ws wt and eer4 seedlings treated with either 5 μM AgNO3 or 100 μl l−1 ethylene, and total RNA was extracted for northern analysis. Several ethylene-inducible genes that are considered to be excellent indicators of ethylene response were used including ACO2, ETR2, ERS1, and AtEBP. None of the genes tested showed increased ethylene-dependent expression, although the ethylene-inducible transcription factor AtEBP was reproducibly expressed at significantly lower levels compared with Ws wt following ethylene treatment (Fig. 2A).
Patterns of ethylene-inducible gene expression were also determined for eer4 leaves, with ERF1, chiB, and PDF1.2 being used as molecular indicators of ethylene response. For this analysis, 4-week-old Ws wt and eer4 plants were treated for 24 h in air, 500 nl l−1 ethylene, or 100 μl l−1 ethylene, after which tissue was collected for total RNA extraction. For Ws wt, the ethylene-regulated transcription factor ERF1 was induced as previously described with saturating ethylene (Fig. 2B) (Solano et al., 1998), whereas eer4 leaves showed little to no expression of ERF1 following ethylene treatment. It was also found that in eer4, ethylene-dependent expression of chiB was greatly depressed, while PDF1.2 expression was abnormally regulated with regard to ethylene responsiveness. Combined with the results observed for ethylene-treated seedlings, EER4 apparently functions as a positive regulator of gene expression following ethylene signalling, thus resulting in partial ethylene insensitivity at the molecular level in the eer4 mutant.
Additionally, patterns of induction of ERF1, chiB, and PDF1.2 (Penninckx et al., 1998) were examined for eer4 leaves following treatment with JA, which in combination with ethylene has been shown to be required for proper expression of these particular genes. Expression of THI2.1, which is induced exclusively by JA (Epple et al., 1995), was examined in order to determine if the effect of the eer4 mutation was limited to ethylene signalling or also impacted jasmonate signalling. Two-week-old hydroponically grown eer4 and Ws wt plants were either untreated or exposed to 100 μM (+/–) JA for 24 h, after which tissue was collected for total RNA extraction. For both ERF1 and chiB, expression patterns in eer4 were similar following jasmonate treatment in comparison with ethylene treatment (Fig. 2C). In contrast, the PDF1.2 expression pattern was indistinguishable in the mutant compared with Ws wt. Expression of THI2.1 was measurably reduced but not eliminated following JA treatment in eer4 compared with Ws wt, suggesting that EER4 may also function in some capacity to facilitate JA-mediated changes in gene expression.
eer4 is a premature stop codon in a TFIID-interacting transcription factor
A mapping population was generated by crossing eer4 (Ws background) to Col-0 wt, with F2 progeny that displayed exaggerated responsiveness to saturating ethylene being selected. From this analysis, eer4 was localized to the bacterial artificial chromosome (BAC) F1L3 and subsequently to the gene At1g17440 (GenBank accession number NM_179349) (Fig. 3A–C). The eer4 mutation represents a single nucleotide change (C→T at nucleotide 967 of the coding sequence) in At1g17440, which leads to a premature stop codon at amino acid 322. Homology-based analysis indicates that EER4 encodes a glutamine- and serine-rich 684 aa transcription factor with a putative TFIID-interacting domain at its C-terminus, which begins at amino acid 530. EER4 is homologous to both human TAF12 and yeast TAF61, suggesting that EER4 functions to recruit transcriptional machinery for expression of a subset of ethylene-regulated genes. At present, no additional alleles of eer4 have been identified.
Total RNA was isolated from Arabidopsis roots, leaves, stems, and flowers in order to determine the tissue-specific expression pattern of EER4. Total RNA was also extracted from 4-d-old dark-grown Ws wt seedlings treated with either 5 μM AgNO3 or 100 μl l−1 ethylene and eer4 seedlings treated with 5 μM AgNO3. EER4 was expressed from low to moderate levels in all tissues tested, but was not found to be ethylene inducible (Fig. 3D). No evidence of alternatively spliced transcripts was found. The eer4 mutation significantly reduced the stability of the EER4 transcript (Fig. 3D), probably due to nonsense-mediated decay arising from the inappropriate stop codon.
A genomic construct consisting of 1 kb of the upstream promoter sequence, the 5′-UTR, exons and introns, and the 3′-UTR was generated in pCGN1547 and introduced into eer4 by Agrobacterium-mediated transformation. Kanamycin-resistant T2 progeny were subsequently analysed for growth capability in the presence of saturating ethylene, with dark-grown 4-d-old transgenic lines demonstrating ethylene-responsive growth that was identical to Ws wt instead of eer4 (Fig. 3E). Additionally, functional complementation eliminated the root curling phenotype that was described for eer4 mutants in Fig. 1E (data not shown). Overexpression of EER4 in Ws wt did not result in any identifiable phenotypic changes.
Double mutants between eer4 and known ethylene signalling components
The ethylene-signalling pathway has largely been elucidated by identification of Arabidopsis mutants with altered ethylene responsiveness. ctr1-3 represents a loss-of-function mutant with a constitutive ethylene-response phenotype, including a severe triple response in the dark in air. An eer4;ctr1-3 double mutant was generated and this was found to grow in the dark in air identically to eer4 treated with saturating levels of ethylene, with the double mutant demonstrating extreme hypocotyl shortening even in the absence of ethylene (Fig. 3F). Air-grown adult eer4;ctr1-3 plants were profoundly smaller than either eer4 or ctr1-3, with eer4;ctr1-3 double mutants almost exclusively dying at or near the time of flowering. The adult phenotype of the eer4;ctr1-3 double mutant was remarkably similar to that reported for the ers1-2;etr1-7 double mutant and the quadruple ethylene receptor loss-of-function mutant, which have an extreme ethylene-response phenotype in both seedlings and adults (Hua and Meyerowitz, 1998; Hall and Bleecker, 2003).
ein2-5 is a severe loss-of-function mutation in a factor of unknown function, with this mutation giving complete ethylene insensitivity in Arabidopsis. In order to demonstrate further that the eer4 phenotype is dependent on ethylene signalling, an eer4;ein2-5 double mutant was generated and its phenotype was analysed following growth in the dark in 100 μl l−1 ethylene for 4 d. As shown in Fig. 3G, the eer4;ein2-5 double mutant was identical to the ein2-5 mutant with no noticeable effect of ethylene on hypocotyl length of the double mutant. Although it could be argued that ein2-5 is epistatic to eer4, it is more likely that, as also shown with AgNO3 treatment, the eer4-mutant phenotype is dependent on ethylene signalling, which is effectively blocked in the ein2-5 mutant.
The ein3-1 mutant represents a defect in a transcription factor that is demonstrated to facilitate long-term ethylene response including induction of ERF1, with loss-of-function mutations in this factor giving partial ethylene insensitivity. An eer4;ein3-1 double mutant was constructed in order to determine if ein3-1 could mask the enhanced ethylene response of eer4. As shown in Fig. 3H, growth of ein3-1 in 100 μl l−1 ethylene resulted in the characteristic long hypocotyl associated with its ethylene insensitivity. Exposure of the eer4;ein3-1 double mutant to saturating levels of ethylene resulted in a growth phenotype consisting of moderate hypocotyl inhibition along with generation of a pronounced apical hook, indicating that the eer4 mutation partially restores ethylene responsiveness to the ein3-1 mutant possibly through lack of induction of an ethylene-responsive resetting or dampening mechanism, which is consistent with the exaggerated ethylene response seen for eer4 mutants.
GFP localization of EER4
In order to determine the subcellular localization pattern of EER4, a translational GFP fusion with the At1g17440 genomic sequence including 1 kb of upstream sequence, introns, and exons without the native stop codon was constructed and localization was compared with an ACT2:GFP control in Nicotiana benthamiana leaves transformed using Agrobacterium. For this analysis, leaves were transformed and after 3 d were stained with DAPI, after which patterns of GFP fluorescence were determined. Fluorescence was evenly distributed throughout the cytoplasm for the GFP control (Fig. 4A). For EER4:GFP, fluorescence was exclusively localized to the nuclei of transformed cells (Fig. 4B), with this pattern being observed in the presence of both 5 μM AgNO3 and saturating ethylene.
Since several components of the ethylene-signalling pathway have been defined, it was of interest to determine which may represent interacting partners with EER4. A yeast two-hybrid approach was initially used to test for associations, with interactions confirmed by in vitro testing for capability of bacterially produced MBP fusions to associate with [35S]methionine-radiolabelled in vitro translated test proteins. For the yeast two-hybrid analysis only, a truncated form of EER4 without the putative TFIID-interacting domain (EER41–500) was employed due to an inability to express full-length EER4 in yeast. From these analyses, there were no detectable interactions with any of the five Arabidopsis ethylene receptors, EIN2's C-terminal domain, EER1/RCN1, RCE1, TIR1, COI1, EBF1, or EBF2.
As shown in Fig. 5A, EER4 self-associated in both the yeast two-hybrid assay and an in vitro binding assay, which is consistent with the function of a wide range of transcription factors and members of the TAFIID transcription complex. Additionally, analysis using the yeast two-hybrid assay and an in vitro binding assay revealed a strong interaction between two PP2a catalytic subunits, PP2a1C and PP2a5C, and EER4 (Fig. 5B). The ethylene-related transcription factors EIN3 and ERF1 could not be tested for an interaction with EER4 using the yeast two-hybrid assay, thus making it necessary to use the aforementioned in vitro binding assay to check for associations. In both instances, EIN3 and ERF1 were found to bind specifically to EER4, with the former association consistent with the loss of ERF1 expression in the eer4 mutant (Fig. 5C, D).
The mechanism underlying ethylene signalling in Arabidopsis has been studied extensively, thus resulting in identification of several biochemical factors that are required for propagation of an ethylene-dependent signal from the receptors to the point of transcriptional regulation. Even with this large body of knowledge, considerable gaps in our understanding of this pathway exist, with traditional methods for identification of signalling components likely to be saturated due to the relative ease with which mutants with severe ethylene-dependent phenotypes, such as ethylene insensitivity and constitutive ethylene response, can be identified. Because of this, non-traditional means are required to fill in the gaps that exist in this pathway.
A unique approach to identify mutants with defects that lead to an enhanced ethylene response, which is characterized by increased sensitivity and/or exaggeration of response to ethylene, has been undertaken. As an outcome of this work, eer4, which is a previously undescribed loss-of-function mutant that, in the absence of ethylene signalling, is capable of growth indistinguishable from Ws wt, yet in the presence of ethylene presents a profound increase in the amplitude of response to ethylene, has been identified. From this analysis and the present examination of the eer4;ein2-5 double mutant, the eer4 phenotype appears to be strictly ethylene dependent since blockage of ethylene signalling either genetically or pharmacologically eliminates the observed discrepancy in seedling growth between Ws wt and eer4. From this growth analysis, it can be argued that EER4 is likely to encode a factor that opposes ethylene signalling, with loss of this factor resulting in reduced capability to dampen ethylene response.
Contrary to the profound growth phenotype observed for eer4, analysis of ethylene-regulated transcription in this mutant revealed that at the molecular level, eer4 is partially ethylene insensitive, with reduced expression of AtEBP in seedlings and no induction of ERF1 in leaves following ethylene treatment. To date, extensive analysis has not revealed any ethylene-regulated genes that are overexpressed in the eer4 mtuant following ethylene treatment (LM Robles, PB Larsen, unpublished results). So far, this type of phenotypic discrepancy is unique for ethylene mutants since loss of gene expression coupled with increased amplitude of ethylene response has not been reported for other mutants with enhanced ethylene response. This includes the loss-of-function mutants representing the two F-box factors, EBF1 and EBF2, both of which encode factors required for proper degradation of the ethylene-responsive transcription factor, EIN3. In the case of ebf1 and ebf2, loss of EIN3 turnover actually results in increased expression of ethylene-responsive target genes (Gagne et al., 2004), thus indicating that the enhanced ethylene-response phenotype of these mutants arises in a manner that is mechanistically distinct from that of eer4. Consistent with this, there is no evidence that there is an effect of the eer4 mutation on expression of EBF1 or EBF2 (LM Robles, PB Larsen, data not shown).
Positional cloning of the eer4 mutation revealed that it represents an inappropriate stop codon in a previously uncharacterized transcription factor that is homologous to human TAF12 and yeast TAF61. Although unusual, tissue- or process-specific TAFs have been reported to regulate specific transcription programmes in Drosophila testis (Hiller et al., 2004; Chen et al., 2005; Falender et al., 2005), supporting the likelihood that Taf12b may play a distinct and specific role in Arabidopsis. This is consistent with the existence of two Taf12 homologues in Arabidopsis. EER4, also known as Arabidopsis TAF12b, has a serine- and glutamine-rich N-terminus and possesses a TFIID-interacting domain at its C-terminus. The predicted function of EER4 as a transcription factor is consistent with the severe effects on expression of a subset of ethylene-inducible genes in the eer4 mutant. From the present double-mutant analyses, it is likely that EER4 is required not only for expression of genes required for promotion of ethylene response, such as ERF1, but also for induction of genes that are critical for a previously undescribed mechanism responsible for resetting or dampening the ethylene signal transduction pathway following ethylene signalling. This is demonstrated by the extreme response of both the ethylene-treated eer4 mutant and air-treated eer4;ctr1-3, with neither of these phenotypes being accounted for by loss of transcription of genes that promote ethylene response such as ERF1, which when overexpressed in an ein3 mutant background gives a constitutive ethylene response (Solano et al., 1998).
The existence of the proposed resetting or dampening mechanism is further supported by the eer4;ein3-1 double mutant, which unlike ein3-1 presents a clear ethylene-response phenotype including full apical hook formation. In a wt plant, EER4 is likely to work in conjunction with other as yet unidentified ethylene-responsive transcription factor(s) to induce genes required for dampening ethylene response. This suggests that in the eer4;ein3-1 double mutant, loss of EER4 function and consequently the resetting or dampening mechanism amplifies the limited response to ethylene signalling that occurs through EIL1 in the absence of functional EIN3, thus causing the partial restoration of ethylene responsiveness that is seen for the eer4;ein3-1 double mutant. It is unlikely that EIL1 represents a transcription factor required for mediation of this reset response since eil1 loss-of-function mutants give a weak ethylene-insensitivity phenotype (Alonso et al., 2003) rather than the enhanced ethylene response associated with eer4 and eer4;ein3. Alternatively, EER4 may be required for expression of an undefined subset of ethylene-regulated genes that are responsible for dampening the EIN3- and EIL1-independent response described by Binder et al. (2004). Loss of modulation of this EIN3- and EIL1-independent pathway would subsequently result in the extreme exaggeration of response seen for eer4 and other enhanced ethylene response mutants such as eer1, etr1-7, ers1-2;etr1-7, and the quadruple ethylene receptor loss-of-function mutant due to a greater contribution by this pathway to the overall level of ethylene response.
Interestingly, the phenotypic similarities between eer1 and eer4 are strikingly similar, with both giving profound exaggeration of response by ethylene-treated seedlings. Molecular cloning of the eer1 mutation revealed that it represents a loss-of-function mutation in a PP2a A regulatory subunit, RCN1, which along with a catalytic subunit and a B regulatory subunit, forms a functional holoenzyme (Larsen and Chang, 2001; Larsen and Cancel, 2003). Although it is not clear how RCN1 works to modulate the level of ethylene response, it is noteworthy that there is a strong interaction observed between EER4 and two Arabidopsis PP2a catalytic subunits, suggesting that PP2a activity may be required for proper function of this proposed resetting or dampening mechanism, which is probably through dephosphorylation of EER4 following activation of the ethylene-signalling pathway.
Homology analysis indicates that EER4 is closely related to Saccharomyces cerevisiae yTAFII61 (Moqtaderi et al., 1996). In yeast, yTAFII61 serves as a bridge between the transcription factor GCN4p, which is a transcriptional activator of genes required for amino acid biosynthesis, and either the TATA-binding protein directly or components of the SAGA complex (yeast Spt–Ada–Gcn5–acetyltransferase complex) in order to recruit the TATA-binding protein and RNA polymerase II for transcription of target genes (Natarajan et al., 1998). In this role, yTAFII61 forms a DNA-associated histone octameric-like structure in association with other yTAFIIs, with this structure probably serving as a protein anchor for SAGA for proper positioning of the transcriptional complex near the TATA box of the promoter (Hoffmann et al., 1996; Xie et al., 1996). Although EER4 probably functions in a manner similar to TAF61 in bridging an interaction between a specific transcription factor, such as EIN3 or ERF1 in the case of ethylene signalling, and the general TAFIID transcriptional machinery, a direct association between EER4 and any of the reported Arabidopsis TATA-binding proteins has not been found. Additionally, whereas Natarajan et al. (1998) argued that yTAFII61 is required for high-level expression of a large portion of class II genes in yeast with a concomitant severely negative impact on overall yeast growth, it is found that the effects of the eer4 mutation are relatively specific even within the scope of ethylene-dependent gene expression, with only a limited subset of genes negatively affected and few discernible effects on growth, development, or morphology in the absence of ethylene.
EER4 is also highly homologous to the much shorter human hTAF12 (hTAFII20), which represents a protein that is conserved with the TFIID-interacting domain found at the C-terminus of EER4. As with yTAFII61, hTAF12 forms a histone fold with other hTAFIIs, including hTAF4 (hTAFII135), hTAF6, and hTAF9, with this structure participating in the recruitment of the TFIID complex to specific promoters for transcriptional initiation (Gangloff et al., 2000; Werten et al., 2002). Unlike the human model for TFIID complex formation, EER4 has not been found to interact with Arabidopsis TAF4 (At5g43130 and At1g27720), TAF6 (At1g04950 and At1g54360), or TAF9 (At1g54140) homologues in vitro, suggesting that EER4 mediates transcriptional activation in a manner differently from hTAF12.
A previous study that exclusively utilized the yeast two-hybrid assay suggested that EER4 (described as Arabidopsis TAF12b in this report) specifically interacts with Arabidopsis homologues of TAF8 (At4g34340), TAF10 (At4g31720), TAF13 (At1g02680), and TAF14 (At2g18000) (Lawit, 2003). An attempt has been made to substantiate these observations using an in vitro binding assay, but it has not been possible to demonstrate these interactions under these stringent conditions. It is possible that the demonstrated interactions in the yeast two-hybrid assay are representative of what occurs in Arabidopsis with regard to the assembly of the TFIID complex, but currently it is left to speculation how EER4 serves as a bridge between phenomenon-specific transcription factors such as EIN3 and the general transcriptional machinery represented by the TFIID complex, including the TATA-binding protein, and RNA polymerase II.
Model for EER4 function
Based on the present results, it is likely that, following ethylene signalling, an as yet undefined mechanism triggers EER4 to form a complex with ethylene- and JA-responsive transcription factors, such as EIN3 and ERF1, along with target promoter sequences. Concomitant to this, EER4 would function to recruit components of the TFIID complex, including the TATA-binding protein, and RNA polymerase II in order to promote transcription of a subset of ethylene-responsive genes, which either work to induce ethylene responses or dampen the ethylene-signalling pathway. Alternatively, EER4 may function to promote transcription in an ethylene-independent manner of a subset of genes that oppose ethylene signalling. As part of the mechanism responsible for dampening the ethylene signalling pathway, EIN3 and other ethylene-regulated transcription factors are targeted for proteolytic degradation by the F-box factors, EBF1 and EBF2 (Guo and Ecker, 2003; Potuschak et al., 2003; Gagne et al., 2004; Olmedo et al., 2006). It is unclear what the fate of EER4 is following an ethylene-signalling event, although no interaction with EBF1 and EBF2 has been observed.
Currently work is being undertaken to identify other transcription factors that may work with EER4 in order to induce the ethylene response, including the mechanism for resetting or dampening the ethylene-signalling pathway. Through the identification of such factors, through screening either for EER4 interactors or for new Arabidopsis eer mutants, it is hoped to further our understanding of this critically important plant hormone signal transduction pathway and potentially devise new strategies for strictly regulating it with regard to manifestation of ethylene responses.
We thank Jesse Cancel for technical assistance with this project and Drs Jean-Denis Faure (INRA-Versailles), Daniel Gallie (UC-Riverside), and Frank Sauer (UC-Riverside) for helpful discussions with regard to this work.