Ethylene Response Factors (ERFs) are downstream components of the ethylene signal transduction pathway, although their role in ethylene-dependent developmental processes remains poorly understood. As the ethylene-inducible tomato Sl-ERF.B3 has been shown previously to display a strong binding affinity to GCC-box-containing promoters, its physiological significance was addressed here by a reverse genetics approach. However, classical up- and down-regulation strategies failed to give clear clues to its roles in planta, probably due to functional redundancy among ERF family members. Expression of a dominant repressor ERF.B3-SRDX version of Sl-ERF.B3 in the tomato resulted in pleiotropic ethylene responses and vegetative and reproductive growth phenotypes. The dominant repressor etiolated seedlings displayed partial constitutive ethylene response in the absence of ethylene and adult plants exhibited typical ethylene-related alterations such as leaf epinasty, premature flower senescence and accelerated fruit abscission. The multiple symptoms related to enhanced ethylene sensitivity correlated with the altered expression of ethylene biosynthesis and signaling genes and suggested the involvement of Sl-ERF.B3 in a feedback mechanism that regulates components of ethylene production and response. Moreover, Sl-ERF.B3 was shown to modulate the transcription of a set of ERFs and revealed the existence of a complex network interconnecting different ERF genes. Overall, the study indicated that Sl-ERF.B3 had a critical role in the regulation of multiple genes and identified a number of ERFs among its primary targets, consistent with the pleiotropic phenotypes displayed by the dominant repression lines.
The plant hormone ethylene is involved in many developmental processes and plays a critical role in a wide range of physiological responses, including seed germination, cell elongation, flowering, fruit ripening, organ senescence, abscission, root nodulation, programmed cell death, and response to abiotic stresses and pathogen attacks (Johnson and Ecker, 1998; Bleecker and Kende, 2000; Lin et al., 2009). Ethylene Response Factors (ERFs) are known to be the last downstream components of the ethylene transduction pathway and the signal transmission cascade has been linked to the transcriptional activation of some ERF genes (Solano et al., 1998; Benavente and Alonso, 2006). According to the currently accepted model, ethylene is perceived by specific receptors, which have been shown to activate the hormone transduction pathway through release of the block exerted by CTR1 on EIN2 (Solano and Ecker, 1998; Ju et al., 2012). The release of EIN2 then activates EIN3/EIL1 primary transcription factors, resulting in the expression of secondary transcription factors, namely ERFs, which regulate the expression of downstream ethylene-responsive genes (Solano et al., 1998; Alonso et al., 2003). The receptors act as redundant negative regulators of ethylene signaling to suppress ethylene responses (Hua and Meyerowitz, 1998; Hall and Bleecker, 2003). In the absence of the hormone, the receptor actively suppresses ethylene responses and ethylene binding removes this suppression. The EIN3/EILs type of transcription factors are positive regulators of the ethylene signaling that function as trans-activating factors to trigger ethylene responses (Chao et al., 1997; Solano et al., 1998). In Arabidopsis, over-expression of EIN3 or EIL1 results in a constitutive ethylene phenotype and reduced expression of multiple LeEIL genes in the tomato results in decreased ethylene sensitivity (Chao et al., 1997; Tieman et al., 2001).
Ethylene Response Factors are plant specific transcription factors and belong to the large AP2/ERF multi-gene family (Riechmann et al., 2000). Proteins encoded by this gene family have a highly conserved DNA-binding domain known as the AP2 domain made of 58 or 59 amino acids involved in the binding to the target DNA sequences (Allen et al., 1998). ERFs from different plant species have been reported to be involved in a variety of processes such as responses to biotic and abiotic stresses, metabolic pathways, fruit ripening and ethylene response (van der Fits and Memelink, 2000; Fujimoto et al., 2000; Li et al., 2007; Trujillo et al., 2008; Lee et al., 2012). ERF proteins are known to interact with multiple cis-acting elements found in the promoter regions of ethylene-responsive genes, including the GCC box and DRE/CRT (dehydration-responsive element/C-repeat; Ohme-Takagi and Shinshi, 1995; Hao et al., 2002; Oñate-Sánchez et al., 2007). It has also been shown that Pti4, an ERF type transcription factor, regulates gene expression by interacting directly with a non-GCC element (Chakravarthy et al., 2003). Moreover, in addition to regulating the expression of ethylene-responsive genes, ERFs can regulate jasmonic acid and salicylic acid-responsive genes (Gu et al., 2000; Brown et al., 2003). ERFs can also bind the Vascular Wounding Responsive Element (VWRE) in tobacco (Sasaki et al., 2007), further demonstrating their capacity to bind a wide range of cis-regulatory elements beside the GCC and DRE/CRT boxes.
Ethylene Response Factors have been associated with ethylene-regulated growth control, with either a positive or a negative regulatory function (Alonso et al., 2003; Nakano et al., 2006; Pirrello et al., 2012). Strikingly, in Arabidopsis little information has been reported (McGrath et al., 2005) on ethylene-responsive phenotypes caused by silencing, mutation, or knock-out of ERFs probably due to the high level of functional redundancy among family members. Indeed, the ERF family is composed of up to 65 members in Arabidopsis (Nakano et al., 2006), many of which are regulated by the same stimuli and can potentially bind the same target promoter. Chimeric Repressor Silencing Technology (CRES-T), consisting in the expression of a dominant repressor version of a transcription factor encoding gene proved to be an efficient means to overcome experimental limitations caused by functional redundancy and this strategy has been developed to study the consequences of silencing target genes of single transcription factors (Hiratsu et al., 2003; Matsui et al., 2005; Heyl et al., 2008). Fusing the so-called SRDX repression domain to a transcription factor suppresses the expression of its target genes dominantly over the activity of endogenous and functionally redundant transcription factors and, as a result, the transgenic plants that express the chimeric repressor version exhibit phenotypes similar to loss-of-function of the alleles of the gene that encodes the transcription factor (Hiratsu et al., 2003; Heyl et al., 2008; Matsui and Ohme-Takagi, 2010).
Genome-wide study recently showed that the tomato ERF gene family comprises nine subclasses defined by distinct structural features and a new nomenclature for tomato ERFs was proposed (Pirrello et al., 2012) that complies with the most complete classification available in Arabidopsis and clarifies the correspondence between ERF subclasses in different species (Nakano et al., 2006). In the tomato, only a few ERF genes have been characterized functionally so far, most of these genes have been shown to participate in stress and/or hormonal responses (Gu et al., 2002; Pirrello et al., 2006; Li et al., 2007; Zhang et al., 2009; Lee et al., 2012; Pan et al., 2012). The tomato Sl-ERF.B3 is related to Arabidopsis factors ERF106 and ERF107, which are members of group IX according to Nakano et al. (2006). This group has been implicated in the regulation of defense responses and knock-out analysis of ORA59 (Pré et al., 2008) and AtERF14 (Oñate-Sánchez et al., 2007), prominent representatives of group IX, has revealed disease susceptibility phenotypes. Consistently, over-expression of ERF1, another member of the group, has led to enhanced resistance to necrotrophic pathogens (Berrocal-Lobo et al., 2002).
Sl-ERF.B3 has been shown previously to act as strong transcriptional activator on GCC-box-containing promoters and its transcripts accumulate upon ethylene treatment, suggesting a putative involvement in ethylene-regulated processes (Tournier et al., 2003; Pirrello et al., 2012). Because over-expressing and down-regulated lines failed to reveal the functional significance of Sl-ERF.B3, a dominant chimeric repressor version was used that resulted in phenotypes that were consistent with Sl-ERF.B3 involvement in both ethylene biosynthesis and signaling pathway. The ERF.B3-SRDX lines displayed constitutive ethylene responses in the absence of ethylene and the data identified a set of ERFs among the target genes regulated by Sl-ERF.B3, and supported the idea that the alteration of such a high number of ERFs may account for the pleiotropic phenotypes displayed by the transgenic lines.
Classical down- and up-regulation approaches failed to provide clear clues on Sl-ERF.B3 functional significance
To address the physiological significance of Sl-ERF.B3 and its potential role in mediating ethylene responses, tomato lines under- and over-expressing Sl-ERF.B3 gene were generated by stably transforming tomato plants with either sense or antisense constructs under the control of the constitutive 35S promoter. Several homozygous transgenic lines that corresponded to independent transformation events were obtained for both antisense and sense construct. Overall, 10 antisense and 12 sense independent lines were examined and the evidence for the expression of the transgene and for its ability to alter the levels of endogenous Sl-ERF.B3 transcripts in the transgenic lines was provided by qRT-PCR analysis (Figure S1a). No consistent phenotypes could be revealed in antisense lines whereas close examination of Sl-ERF.B3 over-expressing plants revealed slightly but significantly higher plants at early development stages (4-week-old) though the plant size returned to normal at 8-week-old plants (Figure S1b). No other consistent growth or reproductive phenotypes could be detected in these Sl-ERF.B3 over-expressing lines.
ERF.B3-SRDX suppresses the transactivation capacity of Sl-ERF.B3
In an attempt to overcome the experimental limitations probably due to functional redundancy among members of the ERF gene family, we generated a dominant repressor version of Sl-ERF.B3 (ERF.B3-SRDX) using the Chimeric Repressor Silencing Technology (CRES-T). The Sl-ERF.B3 coding sequence lacking the stop codon was fused to the SRDX repression domain LDLDLELRLGFA, known as the EAR motif (Mitsuda et al., 2006) and cloned downstream of the Cauliflower Mosaic Virus 35S promoter. The capacity of the ERF.B3-SRDX chimeric protein to function as a transcriptional repressor on ethylene-responsive genes was assessed in a transient transformation assay via co-transfection of protoplasts with reporter and effector constructs. The reporter construct was obtained by fusing the GFP coding sequence either to a synthetic promoter containing the ethylene-inducible GCC box, or to a native osmotin promoter containing the canonical GCC cis-acting element. The effector constructs allow the expression of either the Sl-ERF.B3 protein or its repressor version fused to the SRDX motif (ERF.B3-SRDX). Trans-activation assays indicated that Sl-ERF.B3 enhances the expression of the reporter gene driven by both the synthetic and native promoter, clearly indicating that Sl-ERF.B3 acts as a transcriptional activator of GCC box containing promoters (Figure 1). By contrast, co-transfection of the reporter constructs with the ERF.B3-SRDX results in 8-fold and 15-fold suppression of the activity of the synthetic and the native ethylene-responsive promoters, respectively (Figure 1). These data confirm that ERF.B3-SRDX retains the capacity to bind the same target promoters than Sl-ERF.B3 and to dominantly repress its transcriptional activity. These data support the hypothesis that the ERF.B3-SRDX chimeric protein can potentially be used as transcriptional repressor of Sl-ERF.B3 target genes in planta.
To gain insight on the physiological function of Sl-ERF.B3, transgenic tomato lines (Microtom cv) that expressed the ERF.B3-SRDX dominant repressor construct were produced. Ten independent homozygous 35S:ERF.B3-SRDX lines were generated, all of them displayed similar pleiotropic alterations. Three representative lines, SR1, SR2 and SR3, that showed a characteristic phenotype with different expressivity were selected for further studies. The relative expression level of ERF.B3-SRDX transcript in these three lines was assessed using primers that were specific for ERF.B3-SRDX (Figure S2). Accumulation of the endogenous Sl-ERF.B3 assessed by qRT-PCR was similar in the transformed and non-transformed plants and ruled out the eventuality of a feedback regulation of Sl-ERF.B3 in the transgenic lines (Figure S2).
Dark grown ERF.B3-SRDX seedlings exhibited exaggerated apical hook formation and inhibited hypocotyl elongation in the absence of exogenous ethylene treatment (Figure 2a). Hypocotyl length of 7-day-old etiolated seedlings was 50% lower in ERF.B3-SRDX lines compared with wild-type (WT; Figure 2b). Interestingly, application of 1-MCP, the ethylene perception inhibitor, reversed the triple-response phenotype of ERF.B3-SRDX dominant repressor lines (Figure 2a) and led to a complete loss of the exaggerated apical hook and recovery of hypocotyl length similar to that of WT (Figure 2a,b). Treatment with 10 μl l−1 ethylene resulted in a more pronounced ethylene triple response in ERF.B3-SRDX lines than in WT (Figure 2a,b), suggesting a higher sensitivity to the hormone for the transgenic lines.
Because Sl-ERF.B3 over-expressing plants displayed some, although very mild, growth phenotype at early stages (4-week-old) of plant development, these lines have been tested for the ethylene-response phenotype. While the over-expressing lines cannot be discriminated from WT plants when dark-grown in air, upon exogenous ethylene treatment some of the transgenic lines show a slightly lower reduction in hypocotyl length than in the WT, thus suggesting a reduced response to the hormone (Figure S3).
35S:ERF.B3-SRDX plants show a suite of ethylene hypersensitive phenotypes
Several developmental processes known to be regulated by ethylene were altered in the dominant repressor lines among which were leaf and petiole epinasty (Figure 3). Additional ethylene-related phenotypes displayed by ERF.B3-SRDX plants included premature flower senescence and early fruit abscission (Figure 3). Most flowers in ERF.B3-SRDX plants undergo premature senescence and abscission before full opening of the petals (Figure 3). Moreover, the ERF.B3-SRDX fruits display early abscission compared with WT fruit (Figure 3). Approximately 2 weeks after the breaker stage, the fruit abscission zone starts to dehisce in the ERF.B3-SRDX lines, whereas this event occurs at later stages in wild-type lines (Figure 3). Collectively, these ethylene-related phenotypes are consistent with ethylene hypersensitivity of the ERF.B3-SRDX dominant repressor lines.
Dominant repressor plants display pleiotropic vegetative and reproductive phenotypes
35S:ERF.B3-SRDX plants showed a stunted phenotype from early developmental stages and the size of adult plants was severely reduced (Figure 4a) with the average height being less than one-third of that of WT plants after 80 days (Figure 4b). Noteworthy, the transcript level of two gibberellic acid (GA) oxidase biosynthetic genes, Sl-GA20ox1 and Sl-GA20ox2, was found to be significantly lower than the transgenic plants (Figure 4c). A reduced GA synthesis may therefore account for the dramatic dwarf phenotype displayed by ERF.B3-SRDX plants. Consistent with this hypothesis, application of GA3 to 10-day-old transgenic plants partially rescued the dwarf phenotype (Figure 4d). Nevertheless, in silico analysis of the promoter region of the two GA biosynthesis genes did not reveal the presence of any canonical ethylene-response elements.
Leaf morphology is remarkably altered in the transgenic lines (Figure S4a) with a severe reduction in leaflet size, ranging from 51 to 32% in length and 47 to 22% in width (Figure S4b). The leaf margins of the ERF.B3-SRDX plants are twisted and the lamina is often wrinkled (Figure S4a). Scanning electron microscopy revealed smaller epidermal cells in the transgenic leaves (Figure S4c); the strongest ERF.B3-SRDX expressing line showed epidermal cell size less than one-third of that in wild-type (Figure S4d).
ERF.B3-SRDX plants also showed severely delayed reproductive growth (Figure 5a). The time from germination to flower bud setting was delayed by 14–20 days in transgenic lines compared with the reference WT lines (Figure 5b). Likewise, flower anthesis in ERF.B3-SRDX plants occurred 29–34 days later than in WT (Figure 5b). Moreover, compared with WT, transgenic plants produced significantly smaller flowers (Figure S5a) with up to 30% reduction in anther length. A reduction in fruit size was also observed in the ERF.B3 dominant repressor lines, which produced heart-like shaped fruit (Figure S5b) and small seeds with aberrant shape (Figure S5c). The ERF.B3-SRDX lines also displayed dramatic reduction in fruit set, leading to markedly lower fruit number per plant at maturity (Figure 5d). Up to 91% of successful fruit set was achieved in WT, while in the same growing condition, the fruit set rate reached 10–18% in the ERF.B3-SRDX lines (Figure 5d). Cross-fertilization assay was performed to examine fertility of transgenic flower. Using WT flowers as female recipient and ERF.B3-SRDX plants as pollen donor, 87% of successful fruit set was achieved. Notably, all the developed fruits were seeded and, when germinated, all the seeds were hygromycin resistant (Table 1), a finding that indicated that ERF.B3-SRDX pollen is viable and fertile. Using WT as pollen donor, pollinated ERF.B3-SRDX flowers also showed 80% success of fruit set (Table 1). The reciprocal crossing indicated that both ovule and pollen are fertile in the ERF.B3-SRDX dominant repressor lines (Table 1). Pollen viability of transgenic lines was further confirmed by Alexander's staining assay (Figure S5d). A closer examination of the flower organ structure revealed that ERF.B3-SRDX flowers display exerted stigma positioned beyond the tip of the anther cone, in contrast to WT flowers where the stigma is slightly inserted within the anther cone (Figure 5c). The stigma to anther length ratio is significantly higher in the transgenic lines (Figure 5e) which may consequently prevent efficient self-pollination thus resulting in poor fruit set.
Table 1. Cross-fertilization assay
Fruit set/crossed flowers
Fruit set (%)
F1 hygromycin resistance (%)
Emasculated wild-type flowers were fertilized with ERF.B3-SRDX pollen and the number of fruit was assessed at the ripe stage. Conversely, tomato pollen from wild-type flowers was used to fertilize emasculated ERF.B3-SRDX flowers. In the control assay, wild-type emasculated flowers were fertilized with wild-type pollen. For each cross-fertilization assay, the capacity of the F1 seeds to grow on hygromycin-containing medium was assessed. Results are representative of data from three independent ERF.B3-SRDX lines (SR1, SR2, and SR3).
Expression of ERF.B3-SRDX leads to reduced ethylene production
To investigate the role of Sl-ERF.B3 in regulating ethylene biosynthesis, the level of ethylene production was assessed in etiolated seedlings revealing that ERF.B3-SRDX seedlings produce significantly less ethylene than wild-type (Figure 6a). Accordingly the dominant repressor lines displayed reduced accumulation of transcripts corresponding to Sl-ACS and Sl-ACO ethylene biosynthesis genes (Figure 6b), which accounted for the decreased ethylene production in the ERF.B3-SRDX lines. In silico analysis of the promoter regions of Sl-ACS and Sl-ACO genes using three software packages (place (http://www.dna.affrc.go.jp/PLACE/) plantcare (http://bioinformatics.psb.ugent.be/webtools/plantcare/html) and plantpan (http://plantpan.mbc.nctu.edu.tw/seq_analysis.php) revealed the presence of cis-acting elements that can serve as putative targets for ERFs, including a GCC box (GCCGCC) and DRE/CRT (CCGAC) in Sl-ACO3 promoter and a conserved DRE/CRT (CCGAC) motif in Sl-ACS1 promoter (Table S1).
Ethylene receptor levels are down-regulated in ERF.B3-SRDX plants
In order to determine whether the expression of ethylene receptor genes may contribute to the ethylene hypersensitivity of the 35S:ERF.B3-SRDX lines, we assessed the transcript accumulation of six tomato ethylene receptor genes in the leaves of transgenic plants. While no significant change was found for the expression of Sl-ETR1 and Sl-ETR4, the four remaining ethylene receptor genes (Sl-ETR2, Sl-ETR5, Sl-ETR6 and NR) were substantially down-regulated in the ERF.B3-SRDX lines (Figure 7a). Notably, the expression of Sl-ETR5 was decreased by 84% in the strongest ERF.B3-SRDX line (Figure 7a). The expression of Sl-ETR2 was reduced by 52–65% in three independent lines (Figure 7a) while that of NR was decreased by 46–61% (Figure 7a). The transcript levels of Sl-ETR6 showed 35–50% reduction compared with wild-type (Figure 7a). An in silico search revealed the absence of conserved GCC box in the promoter regions of all four ethylene receptor genes displaying altered expression in the transgenic lines (Table S1) – in contrast with NR and Sl-ETR5 promoters, which contain GCC-box-like and DRE/CRT consensus sequences. However, because Sl-ETR6 receptor has been shown to play a prominent role in regulating ethylene response (Tieman et al., 2000; Kevany et al., 2007), the ability of the native Sl-ERF.B3 and the chimeric ERF.B3-SRDX proteins to regulate the Sl-ETR6 promoter activity was tested. Transactivation assays show that Sl-ERF.B3 induced more than a 2-fold increase of the Sl-ETR6 promoter activity whereas ERF.B3-SRDX strongly suppressed this activity (Figure 7b) indicating that Sl-ERF.B3 and its dominant repressor version can both regulate the expression of Sl-ETR6 in despite of the absence of a typical ethylene-responsive element in its promoter region. Given that ERF.B3-SRDX down-regulates the expression of the ethylene receptor genes in vivo and that both Sl-ERF.B3 and its repressor version strongly impact the transcriptional activity of Sl-ETR6 in the transactivation assay, we then looked at the expression of ethylene receptor genes in tomato over-expressing lines. Among all six receptor genes present in the tomato genome, ETR1, NR and ETR6 are up-regulated in the Sl-ERF.B3 over-expressing lines, consistent with the activator function of the Sl-ERF.B3 protein (Figure 7c).
EIN3-Like genes are up-regulated in ERF.B3-SRDX transgenic plants
EIN3/EILs are positive regulators of ethylene signaling by acting as transactivation factors to trigger ethylene responses. The expression of the four EIN3-like genes (Sl-EIL1, 2, 3 and 4) present in the tomato genome was examined at the transcript level showing a 2-fold increase in transcript accumulation for all four Sl-EIL genes in the ERF.B3-SRDX lines (Figure 8). However, none of the EIN3-like genes gathers a consensus ethylene-response element in the promoter. Transactivation assays performed revealed that neither Sl-ERF.B3 nor ERF.B3-SRDX proteins are capable to modulate transcription driven by any of the four Sl-EILs promoters (Figure S6), suggesting that Sl-EILs do not serve as direct target genes for Sl-ERF.B3.
Sl-ERFs are among the target genes of Sl-ERF.B3
Considering the putative role of ERFs in mediating ethylene responses, we examined the transcript levels of Sl-ERF genes in both WT and the ERF.B3-SRDX lines. A dramatic change in the transcript levels for a number of ERF genes was revealed in the dominant repressor lines (Figure 9a). That is, among the 19 Sl-ERFs that showed detectable transcript accumulation, 14 were significantly down-regulated in the ERF.B3-SRDX dominant repressor lines while 4 Sl-ERFs displayed similar expression in transgenic and WT lines. Notably, the expression of Sl-ERF.G1 displayed dramatic up-regulation in transgenic lines (Figure 9a). To gain further insight on the mechanisms underlying the regulation of Sl-ERF genes in the transgenic lines, the promoters of down- and up-regulated ERFs genes were cloned to examine the ability of Sl-ERF.B3 and ERF.B3-SRDX proteins to regulate their activity in a single cell system. The data indicate that Sl-ERF.B3 protein acts as activator on Sl-ERF.C3, Sl-ERF.D2, Sl-ERF.F5 and Sl-ERF.F4 promoters while it is inactive on Sl-ERF.G1. The ERF.B3-SRDX repressor version retains the capacity to recognize the same target genes than Sl-ERF.B3 as demonstrated by its repressing activity on the promoters activated by Sl-ERF.B3 (Figure 9b). By contrast, neither Sl-ERF.B3 nor ERF.B3-SRDX proteins were able to modulate the activity of the Sl-ERF.G1 promoter. Taking advantage of the available Sl-ERF.B3 up-regulated lines, we also examined the expression level of Sl-ERF genes in these over-expressing lines. Opposite to the situation prevailing in the ERF.B3-SRDX lines, most ERF genes are up-regulated in the Sl-ERF.B3 over-expressing lines (Figure 9c) with the most significant up-regulation found in the lines that displayed a reduced ethylene response (Figure S3). Of particular note, Sl-ERF genes (Sl-ERF.C3, Sl-ERF.D2, Sl-ERF.F5 and Sl-ERF.F4) shown to be direct target for Sl-ERF.B3 in the transactivation assay are all up-regulated in the Sl-ERF.B3 over-expressing lines. Moreover, ERF genes that show regulation by Sl-ERF.B3 in the single cell system (Figure 9b) harbor cis-acting elements (GCC box and DRE/CRT) known to be putative binding site for ERFs whereas the Sl-ERF.G1 promoter lacks any of these typical cis-elements (Figure 9d and Table S2).
Although ERFs are generally considered as important components of the ethylene-response mechanism, direct evidences for the involvement of these transcription factors in this process are still scarce. So far, classical approaches of forward and reverse genetics aiming at up- or down-regulating the expression of ERF genes failed to provide sufficient clues on the physiological significance of different members of this gene family probably due to functional redundancy among family members. In the present study, the ectopic expression of a dominant repressor form of the Sl-ERF.B3 protein provided a mean towards altering the activity of the native Sl-ERF.B3 protein. This strategy allowed revealing vegetative and reproductive growth phenotypes that could not be uncovered by the expression of neither sense nor antisense constructs of Sl-ERF.B3. Notably, the ERF.B3-SRDX plants display enhanced ethylene responses that tend to phenocopy the Arabidopsis ctr1 mutant as well as the transgenic tomato lines deficient in receptors, exhibiting all hallmarks of exposure to ethylene (Kieber et al., 1993; Tieman et al., 2000). Although, the opposite effect would have been intuitively expected from blocking the action of an ERF, the physiological and molecular characterization clearly indicated that the phenotypes are consistent with enhanced ethylene sensitivity due to depletion of ethylene receptor pools but not to ethylene over-production.
The 35S:ERF.B3-SRDX lines displayed enhanced ethylene responses and pleiotropic ethylene-related alterations, probably resulting from the transcriptional repression of ethylene-responsive genes that are natural targets of the native protein. Indeed, Sl-ERF.B3 and ERF.B3-SRDX are shown to modulate the activity of the same promoters harboring ethylene-responsive elements, indicating that ERF.B3-SRDX has the ability to interfere with the regulation of Sl-ERF.B3 target genes. ERF.B3-SRDX fusion protein is a strong repressor of both synthetic and native ethylene-responsive promoters whereas the native Sl-ERF.B3 protein enhances the activity of these promoters. The eventuality that the pleiotropic phenotypes displayed by the ERF.B3-SRDX dominant suppressor plants may arise from a co-suppression of the endogenous Sl-ERF.B3 is ruled out since the levels of Sl-ERF.B3 transcripts are not altered in the transgenic lines. Notably, the higher the ERF.B3-SRDX transgene expression the more severe was the phenotypic abnormality, indicating that the phenotypic effects were directly related to the expression levels of the ERF.B3-SRDX transgene. Therefore, the ERF.B3-SRDX tomato lines proved to be a valuable tool to uncover at least some of the processes controlled by Sl-ERF.B3 and to reveal roles for ERF genes that have not been described previously.
Dark-grown ERF.B3-SRDX seedlings displayed a constitutive ethylene response-like phenotype with inhibited hypocotyl elongation and exaggerated apical hook formation in the absence of exogenous ethylene. Moreover, adult plants show typical constitutive ethylene responses including leaf epinasty, premature flower senescence and accelerated fruit abscission. These phenotypes may arise from: (i) a constitutive ethylene response, (ii) an increased sensitivity to endogenous ethylene, or (iii) an ethylene over-production. Noteworthy, the ethylene-response phenotypes displayed by ERF.B3-SRDX etiolated seedlings can be reversed by the inhibition of ethylene perception (Figure 2a) and treatment with exogenous ethylene resulted in a more pronounced ethylene triple response compared with wild-type. Taken together with the reduced ethylene production, these results indicate that the ethylene-response phenotypes displayed by ERF.B3-SRDX lines are not due to constitutive activation of ethylene signaling pathway but rather to enhanced ethylene sensitivity. This hypothesis is further supported by the reversion of the ethylene-response phenotype upon treatment with 1-MCP, a potent inhibitor of ethylene receptors. That is, endogenous ethylene levels inactivate the residual amounts of the receptors; blocking ethylene receptors by 1-MCP reverts the phenotype. It is well accepted that ethylene receptors act as negative regulators and function redundantly in ethylene signaling with a decreased expression of ethylene receptor genes that results in increased sensitivity to the hormone (Hua and Meyerowitz, 1998; Kevany and Klee, 2007; Wuriyanghan et al., 2009). The reduced transcript levels of the receptors and the ethylene hypersensitivity of ERF.B3-SRDX lines are consistent with this model. In tomato, although gene-specific antisense reductions in Sl-ETR1, Sl-ETR2, NR or Sl-ETR5 do not affect ethylene sensitivity, transgenic lines with single reduction in Sl-ETR4 or Sl-ETR6 expression display phenotypes consistent with enhanced ethylene response (Tieman et al., 2000; Kevany et al., 2007) – a finding that indicates that these two receptors may act as a special component in regulating ethylene response. The down-regulation of Sl-ETR6 in the ERF.B3-SRDX lines may therefore account for the increased ethylene sensitivity. Interestingly, opposite to its down-regulation in the dominant repressor lines, ETR6 shows a net up-regulation in the Sl-ERF.B3 over-expressing plants and suggests that this receptor gene may represent a direct target for Sl-ERF.B3 protein in vivo.
The increased expression of transcription factors belonging to the EIN3 gene family may also contribute to enhanced ethylene responses. Over-expression of EIN3 or EIL1confers constitutive ethylene phenotypes in Arabidopsis, while reduced Sl-EILs expression in transgenic tomato decreases ethylene sensitivity (Chao et al., 1997; Tieman et al., 2001). Four EIN3-like genes were isolated in tomato and designed as Sl-EIL1, Sl-EIL2, Sl-EIL3 and Sl-EIL4 (Tieman et al., 2001; Yokotani et al., 2003). As it is well documented that EIN3/EIL proteins act as transactivation factors to trigger ethylene responses, up-regulation of all four Sl-EIL genes in the ERF.B3-SRDX plants may contribute to their ethylene hypersensitivity. However, because the promoter of EIN-like genes are devoid of consensus ethylene-response elements and as transactivation assays indicated that Sl-ERF.B3 and ERF.B3-SRDX proteins are unable to modulate transcription driven by any of the four Sl-EILs promoters, it is likely that the up-regulation of Sl-EIL genes in the dominant repressor lines is due to intermediate factor(s) whose expression/activation is regulated by ERF.B3-SRDX.
Previous studies have already shown that ERF proteins are involved in a feedback regulation of ethylene production by modulating the expression of ethylene biosynthesis genes (Zhang et al., 2009; Lee et al., 2012). Our data show that ectopic expression of the ERF.B3-SRDX dominant repressor results in reduced ethylene production associated with the down-regulation of ACS and ACO ethylene biosynthesis genes. The presence of conserved GCC box and DRE/CRT motifs in ACS and ACO promoters that can serve as binding sites for ERF proteins supports the hypothesis that these ethylene biosynthesis genes can directly be regulated by Sl-ERF.B3. Together, the reduced ethylene production and enhanced ethylene sensitivity in the ERF.B3-SRDX lines suggest the presence of a feedback loop regulating both ethylene biosynthesis and a signal transduction pathway and involving ERF proteins.
Strikingly, the expression of a considerable number of Sl-ERF genes, 15 out of 19 monitored in our study, was found to be markedly altered in ERF.B3-SRDX tomato lines, a finding that suggested intense inter-regulation among ERF family members. Consistent with the dominant repressor function of the ERF.B3-SRDX protein, most ERF genes were down-regulated while Sl-ERF.G1 alone displayed higher transcript levels in the dominant repressor lines. By contrast, in Sl-ERF.B3 over-expressing lines, most ERF genes tested displayed enhanced transcript levels. In particular Sl-ERF.C3, Sl-ERF.D2, Sl-ERF.F5 and Sl-ERF.F4, shown to be direct target for Sl-ERF.B3 in the transactivation assay, displayed enhanced expression in the Sl-ERF.B3 sense lines. While these data support the idea that these ERFs can serve as direct target for both the native and chimeric Sl-ERF.B3 proteins, the up-regulation of Sl-ERF.G1 in the dominant repressor lines probably requires an additional mediating factor. An in silico search revealed that all ERF genes down-regulated in the transgenic lines harbor cis-acting elements that are known to be putative binding targets for ERFs. The down-regulation of such a high number of Sl-ERFs supports a model that implies that a single ERF can affect the expression of other members of the gene family. This inter-connected regulation among ERF genes may therefore account for the pleiotropic alterations in the ERF.B3-SRDX lines and for the diversity of responses displayed by the dominant repressor lines.
Phenotypes such as stunted plant development, reduced leaf size and late-flowering time are reminiscent not only of constitutive ethylene-response mutants but also of GA-deficient Arabidopsis plants (Kieber et al., 1993; Hua and Meyerowitz, 1998; Hall and Bleecker, 2003; Magome et al., 2004). The partial rescue of the dwarf phenotype in the ERF.B3-SRDX lines by exogenous application of GA suggests that these alterations are partly due to GA deficiency. In line with the model supporting that idea that ethylene regulates plant growth and floral organ differentiation via modulating GA levels (Achard et al., 2007), ethylene hypersensitivity in the ERF.B3-SRDX dominant suppressor lines is associated with reduced plant size and substantially delayed flowering time. The reduced expression of GA oxidase genes in the transgenic lines sustains the idea of altered GA metabolism and suggests that ERFs may represent a potential molecular link between ethylene and GA. In agreement with this idea, it has been recently reported that transcriptional activation of some genes involved in GA metabolism is mediated by ERF6 in Arabidopsis leaves (Dubois et al., 2013). Because the study has been carried out with Micro-Tom, a dwarf genotype, it is important to mention that the dwarfing mutations in this genotype do not seem to impact the phenotype displayed by ERF.B3-SRDX plants, as the dwarf phenotype is well reproduced in Ailsa Craig tomato, a non-dwarf variety. Altogether, the data suggest that ethylene hypersensitivity is likely to be the fundamental cause of the severe dwarf and late-flowering phenotypes in the ERF.B3-SRDX plants.
As ectopic expression of transcription factors might influence target genes that are normally not under the control of this regulator, it cannot be totally ruled out that at least part of gene regulation caused by ERF.B3-SRDX are off-target effects due to interference with other related transcription factors. However, the data support the idea that Sl-ERF.B3 is part of an intricate web of regulation in which multiple transcription factors are competing for promoters to control the expression of genes that are essential for a wide range of plant responses to ethylene. As depicted in the tentative regulation model presented in Figure 10, Sl-ERF.B3 is shown to modulate ethylene responses at four different levels: (i) ethylene biosynthesis, (ii) ethylene receptor, (iii) primary ethylene transcription factors (EIL genes), and (iv) downstream ERF genes. The high number of ERF genes regulated by Sl-ERF.B3 is consistent with the pleiotropic phenotypes displayed by the dominant repressor lines and suggests that ERFs form a complex network with a subset of the family members that function in an inter-connected manner. Such level of complexity matches the high level of plasticity needed for the implementation of plant growth and developmental processes that require continuous fine-tuning through the integration of different cues and signaling pathways.
Plant materials and growth conditions
Tomato plants (Solanum lycopersicum cv. Micro-Tom) were grown under standard greenhouse culture conditions. The culture chamber rooms were set as follows: 14 h-day/10 h-night cycle, 25/20°C day/night temperature, 80% hygrometry, 250 μmol m−2 sec−1 intense luminosity.
Constructs and plant transformation
To generate the chimeric repressor transgene, the coding sequence of Sl-ERF.B3 without the stop codon was cloned via blunt-end ligation into the SmaI site of p35SSRDXG in frame with the SRDX repression domain (LDLDLELRLGFA) from SUPERMAN (Hiratsu et al., 2003; Mitsuda et al., 2006). Agrobacterium tumefaciens-mediated transformation of tomato plants was carried out in accordance with Wang et al. (2005) and transformed lines were selected on a hygromycin-containing medium. All experiments were carried out using homozygous lines from F3 or later generations.
Transient expression using a single cell system
Protoplasts used for transfection were isolated from suspension-cultured tobacco (Nicotiana tabacum) BY-2 cells in accordance with (Leclercq et al., 2005). The synthetic reporter construct (4xGCC-GFP) was generated by fusing the synthetic GCC-box promoter to the coding region of the GFP (Pirrello et al., 2012). Reporter constructs were also generated with native promoters, Sl-osmotin (C08HBa0235H18.1) and Sl-ERFs (ERF.C3, ERF.D2, ERF.F4, ERF.F5 and ERF.G1), fused to GFP. Protoplast co-transfection assays was performed using the reporter plasmids and effector vectors carrying 35S:ERF.B3 or 35S:ERFB3-SRDX. GFP expression was analyzed and quantified by flow cytometry (FACS Calibur II instrument; BD Biosciences, San Jose, CA, USA, http://www.bdbiosciences.com/eu/index.jsp) 16 h following protoplast transfection. For each sample, 100–1000 protoplasts were gated on forward light scatter; GFP fluorescence per population of cells corresponds to the average fluorescence intensity of the population of cells above the background. The data were analyzed using cell quest software (BD Biosciences) and were normalized using an experiment with protoplasts transformed with the reporter vector in combination with the vector used as effector but lacking the Sl-ERF.B3 coding sequence.
RNA isolation and qRT-PCR
Total RNA from 4-week-old plants was extracted using a Plant RNA Purification Reagent (Invitrogen, cat. no. 12322-012, Carlsbad, CA, USA, http://www.lifetechnologies.com/fr/fr/home.html). Total RNA was DNase-treated (Invitrogen, cat. no. AM1906) and first-strand cDNA was reverse transcribed from 2 μg of total RNA using an Omniscript Reverse Transcription kit (Qiagen, cat. no. 74904, Valencia, CA, USA, http://www.qiagen.com/). Gene-specific primers were designed by primer express software (PE-Applied Biosystems, Foster City, CA, USA, http://www.appliedbiosystems.com/absite/us/en/home.html) and were further checked using BLAST against all tomato unigenes (Tomato unigene database). qRT-PCR analyses were performed as described previously (Pirrello et al., 2006). The primer sequences used in this study are listed in Table S3.
For application of gibberellin to young plants growing on soil, 10−5m of gibberellic acid (GA3) was sprayed on the plants twice a week starting on the 10th day post-germination. After 2 weeks of treatment, the treated plants were compared with the control ones.
Sterilized seeds were first put on MS/2 medium plates and placed at 4°C for 3 days and then transferred to 25°C for germination in the dark for another 5 days. The seedling triple response was scored by assessing hypocotyl length and apical curvature. At least 50 seedlings were scored for each measurement. For ethylene treatment, Petri dishes were enclosed in wide mouth Mason jars sealed with a lid containing a rubber syringe cap. Ethylene (10 μl l−1) was then injected into the Mason jars using a syringe. For 1-MCP treatment, 1 μl l−1 was applied into the Mason jars and kept in the dark for 1 week. At least 50 seedlings were used for each experiment and three independent biological replicates were performed.
Gene ID data for the genes described in this article are listed in Table S4.
This work forms part of the requirement for the degree of Ph.D. for M.C Liu supported by the China Scholarship Council. The authors are grateful to L. Lemonnier, D. Saint-Martin, O. Berseille and L. Gress for genetic transformation, culture and phenotyping of tomato plants. This research was supported by funds from the ‘Laboratoire d'Excellence’ (LABEX) entitled TULIP (ANR-10-LABX-41) and benefited from the networking activities within the European COST Action FA1106.