Fruit ripening involves a complex interplay between ethylene and ripening-associated transcriptional regulators. Ethylene Response Factors (ERFs) are downstream components of ethylene signaling, known to regulate the expression of ethylene-responsive genes. Although fruit ripening is an ethylene-regulated process, the role of ERFs remains poorly understood.
The role of Sl-ERF.B3 in tomato (Solanum lycopersicum) fruit maturation and ripening is addressed here using a chimeric dominant repressor version (ERF.B3-SRDX).
Over-expression of ERF.B3-SRDX results in a dramatic delay of the onset of ripening, enhanced climacteric ethylene production and fruit softening, and reduced pigment accumulation. Consistently, genes involved in ethylene biosynthesis and in softening are up-regulated and those of carotenoid biosynthesis are down-regulated. Moreover, the expression of ripening regulators, such as RIN, NOR, CNR and HB-1, is stimulated in ERF.B3-SRDX dominant repressor fruits and the expression pattern of a number of ERFs is severely altered.
The data suggest the existence of a complex network enabling interconnection between ERF genes which may account for the pleiotropic alterations in fruit maturation and ripening. Overall, the study sheds new light on the role of Sl-ERF.B3 in the transcriptional network controlling the ripening process and uncovers a means towards uncoupling some of the main ripening-associated processes.
The maturation and ripening of fleshy fruits are developmental processes unique to plants. Although specific fruit ripening characteristics vary among species, fruit ripening can be generally described as a complex, genetically programmed process that culminates in dramatic changes in color, texture, flavor, aroma and nutritional characteristics (Carrari & Fernie, 2006). In the case of fleshy fruits, these changes not only make fruit attractive for seed dispersal organisms, but also provide essential vitamins, minerals and antioxidants (phenolics, folate, lycopene and β-carotene) for human diet (Seymour et al., 1993; Fraser et al., 2009; Chung et al., 2010).
Fruits have been classically categorized into climacteric and non-climacteric based on increased ethylene synthesis and a concomitant rise in the rate of respiration during ripening. Climacteric fruits display a burst in respiration at the onset of ripening, in contrast with non-climacteric fruits. Climacteric fruits, such as tomatoes, bananas and apples, also show increased biosynthesis of the gaseous hormone ethylene, which is a fundamental signal in climacteric fruit ripening (Vrebalov et al., 2002; Alba et al., 2005). In the ethylene biosynthetic pathway, 1-aminocyclopropane-1-carboxylic acid (ACC) synthase (ACS) and ACC oxidase (ACO) catalyze the conversion of S-adenosyl-l-methionine (SAM) to ACC and of ACC to ethylene, respectively (Adams & Yang, 1979; Bleecker & Kende, 2000). Autocatalytic ethylene synthesis at the onset of tomato fruit ripening is mainly mediated through ethylene-stimulated expression of ACS2, ACS4, ACO1 and ACO4 (Barry & Giovannoni, 2007). Unraveling the regulation of the ethylene signaling pathway is important to understanding the processes of fruit ripening.
Tomato possesses many favorable genetic characteristics, such as simple diploid genetics, relatively small genome size, short generation time, efficient genetic transformation and distinct ripening phenotypes, making it a primary model system for the study of the molecular basis of fleshy fruit development and the role of ethylene in climacteric fruit ripening. The adaptation of a range of technological tools and the generation of new biological resources on tomato (e.g. expressed sequence tag (EST) databases, TILLING (targeting-induced local lesions in genomes) resources, genetic and physical maps) have led to significant progress in our understanding of the molecular mechanisms underlying the ripening process through the identification of the associated key regulatory genes (Pirrello et al., 2009). Indeed, tomato fruit ripening has been thoroughly characterized with regard to metabolic changes impacting softening, the accumulation of sugars, acids and lycopene, chlorophyll degradation and dramatic increases in ethylene and flavor volatiles (Chung et al., 2010).
The investigation of tomato mutants affected in fruit development and ripening (mainly ripening-deficient mutants), such as ripening inhibitor (rin; Vrebalov et al., 2002), Colorless non-ripening (Cnr; Manning et al., 2006), non-ripening (nor; Giovannoni, 2004), Green-ripe (Gr; Barry & Giovannoni, 2006) and Never-ripe (Nr; Wilkinson et al., 1995), led to the isolation of genes acting upstream of ethylene in the ripening cascade, and involved in determining the attainment of competence to ripen (Barry & Giovannoni, 2007). The RIN, CNR and NOR genes encode transcription factors regulating the expression of genes responsible for various fruit ripening processes, including ethylene and carotenoid biosynthesis (Vrebalov et al., 2002; Giovannoni, 2004; Manning et al., 2006; Martel et al., 2011; Fujisawa et al., 2013). The Gr gene encodes a still poorly defined component of ethylene signal transduction, whereas Nr encodes an ethylene receptor important for both fruit and non-fruit ethylene responses (Lanahan et al., 1994; Barry & Giovannoni, 2006). Other ripening transcriptional regulators have recently been characterized via functional studies in transgenic plants, including LeHB-1, which regulates directly ACC oxidase expression (Lin et al., 2008), and TAGL1, a MADS box transcription factor, which links early fruit fleshy expansion with downstream ripening (Itkin et al., 2009; Vrebalov et al., 2009). The putative transcription factor, SlAP2a, a member of the APETALA2/Ethylene Response Factor (AP2/ERF) superfamily, has also been described recently as a negative regulator of fruit ripening and of ethylene production (Chung et al., 2010; Karlova et al., 2011). Unraveling the transcriptional networks that regulate fruit ripening is crucial for the understanding of this complex process.
ERFs are plant-specific transcription factors, belonging to the large AP2/ERF superfamily (Riechmann et al., 2000). Proteins encoded by this family have a highly conserved DNA-binding domain, known as the AP2 domain, containing 58–59 amino acids involved in the high-affinity binding to target DNA sequences (Allen et al., 1998). A growing number of investigations have suggested that, through interactions with multiple cis-acting elements found in the promoter regions of ethylene-responsive genes, including the GCC box and dehydration-responsive element/C-repeat (DRE/CRT), ERF proteins play a critical role during plant development and adaptation to stress conditions (Ohme-Takagi & Shinshi, 1995; Wu et al., 2002; Wan et al., 2011). In different plant species, ERFs have been shown to be involved in hormonal signaling, responses to biotic and abiotic stresses, developmental processes, metabolic regulation, ethylene biosynthesis and fruit ripening (Ohme-Takagi & Shinshi, 1995; Fujimoto et al., 2000; Van der Fits & Memelink, 2000; Wu et al., 2002; Dubouzet et al., 2003; Zhang et al., 2009; Lee et al., 2012; Pirrello et al., 2012; Zhao et al., 2012). In tomato, although SlERF6, which corresponds to Sl-ERF.E4 in the new unified nomenclature by Pirrello et al. (2012), has been reported to play an important role in fruit ripening by integrating the ethylene and carotenoid pathways (Lee et al., 2012), the role of most ERF proteins in the ripening process awaits elucidation. The present study describes the critical role of Sl-ERF.B3, a member of the tomato ERF multi-family genes, in fruit ripening.
To date, no ERF-like mutants have been identified in tomato and, as reported previously (Liu et al., 2013), a classical reverse genetics approach based on the down- and up-regulation of ERF genes failed to provide sufficient clues regarding their functional significance. In an attempt to overcome the experimental limitations caused by the functional redundancy among members of the ERF gene family, we generated a dominant repressor version of ERF.B3 (ERF.B3-SRDX) using Chimeric Repressor Silencing Technology (CRES-T). Sl-ERF.B3 was selected as a target ERF in the present study because it has been shown previously to be an activator of GCC box-containing promoters and its transcripts accumulate on ethylene treatment, suggesting its putative involvement in ethylene-regulated processes (Tournier et al., 2003; Pirrello et al., 2012; Liu et al., 2013). This gene has been described previously as an important regulator of ethylene response and plant development (Liu et al., 2013). Using the CRES-T strategy, we show here that Sl-ERF.B3 plays a critical role in fruit development and ripening. Moreover, the altered expression of major regulators of fruit ripening, including RIN, CNR, NOR and HB-1, in ERF.B3-SRDX lines reveals that Sl-ERF.B3 is a new regulator involved in the regulatory network controlling the ripening process.
Materials and Methods
Plant materials and growth conditions
Tomato (Solanum lycopersicum L. cv MicroTom) plants were transferred to soil and grown under standard glasshouse conditions. Conditions in the culture chamber room were set as follows: 14 h day : 10 h night cycle, 25 : 20°C day : night temperature, 80% hygrometry, 250 μmol m−2 s−1 light intensity. For the measurement of time to ripening, flowers were tagged at anthesis and the number of days from anthesis to the breaker (Br) stage was counted. More than 15 fruits of each genotype were used for this analysis.
To generate the ERF.B3-SRDX transgenic plants, the coding sequence of Sl-ERF.B3 missing the stop codon was amplified by PCR from a tomato fruit cDNA library. This coding region was cloned via blunt-end ligation into the SmaI site of p35SSRDXG in frame to the region that encodes the SRDX repression domain (LDLDLELRLGFA) from SUPERMAN (Hiratsu et al., 2003; Mitsuda et al., 2006). The transgene cassette was transferred into the destination vector pBCKH, which was derived from the plant transformation vector pBIG-HYG (Becker, 1990) using the gateway LR reaction (Invitrogen Corp.) Agrobacterium tumefaciens-mediated transformation of tomato plants was carried out according to 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.
RNA isolation and quantitative real-time-polymerase chain reaction (qRT-PCR)
Fruits were harvested, frozen in liquid nitrogen and stored at −80°C. Total RNA from the pericarp of at least five individual fruits at each developmental stage analyzed in this article was extracted using a Plant RNA Purification Reagent (Invitrogen, Cat. No. 12322-012) according to the manufacturer's instructions. Total RNA was then DNase-treated (Invitrogen, Cat. No. AM1906) to remove contaminating genomic DNA. First-strand cDNA was reverse transcribed from 2 μg of total RNA using an Omniscript Reverse Transcription kit (Qiagen, Cat. No. 74904) following the manufacturer's instructions. Gene-specific primers were designed by Primer Express software (PE-Applied Biosystems, Foster City, CA, USA) and were further checked using BLAST against all tomato unigenes (Tomato unigene database). Quantitative real-time PCR analyses were performed as described previously (Pirrello et al., 2006). The primer sequences used in this study are listed in Supporting Information Table S1.
LC-MS analysis of fruit carotenoids
Carotenoid extractions were performed as described previously (Fantini et al., 2013). Briefly, 5 mg of ground lyophilized tomato fruit powder were extracted with chloroform (spiked with 50 mg l−1dl-α-tocopherol acetate as internal standard) and methanol (2 : 1 by volume); subsequently, 1 volume of 50 mM Tris buffer (pH 7.5, containing 1 M NaCl) was added and the samples were kept for 20 min on ice. After centrifugation (15 000 g for 10 min at 4°C), the organic hypophase was collected and the aqueous phase was re-extracted with the same amount of spiked chloroform. The combined organic phases were then dried by speedvac and resuspended in 100 μl of ethyl acetate. For each genotype, at least five independent extractions were performed. LC-MS analyses were carried out using a Discovery LTQ-Orbitrap mass spectrometry system operating in atmospheric pressure chemical ionization (APCI) and in positive mode, coupled to an Accela U-HPLC system (Thermo Fisher Scientific, Waltham, MA, USA). LC separations were performed using a C30 reverse-phase column (100 × 3.0 mm) from YMC (YMC Europe GmbH, Schermbeck, Germany). The mobile phases used were methanol (A), water–methanol (20 : 80 by volume) containing 0.2% ammonium acetate (B) and tert-methyl butyl ether (C). The gradient was 95% A : 5% B for 1.3 min, followed by 80% A : 5% B : 15% C for 2.0 min and by a linear gradient to 30% A : 5% B : 65% C over 9.2 min. UV–visible detection was performed continuously from 220 to 700 nm with an online Accela Surveyor photodiode array detector (PDA; Thermo Fisher Scientific). All solvents used were LC-MS grade quality (CHROMASOLV® from Sigma-Aldrich). Carotenoids were quantified on the basis of the internal standard amounts, obtained through comparison with peak areas of known amounts of external standard LC-MS runs and by extinction coefficient correction. For APCI-MS ionization of xanthophylls (0–6 min of LC-MS run), nitrogen was used as sheath and auxiliary gas, set to 40 and 20 units, respectively, the vaporizer temperature and capillary temperature were 300 and 250°C, respectively, the discharge current was set to 4.0 μA, and the capillary voltage and tube lens settings were 27 V and 90 V, respectively. APCI-MS ionization of carotenes (6–14 min of LC-MS runs) was performed with the following parameters: 30 and 10 units of nitrogen sheath and auxiliary gas, respectively; vaporizer and capillary temperatures of 300 and 250°C, respectively; discharge current of 5.0 μA; capillary voltage and tube lens settings of 0 and 95 V, respectively. Identification was performed as reported previously (Fantini et al., 2013), and on the basis of the m/z accurate masses, as reported on Pubchem (http://pubchem.ncbi.nlm.nih.gov/) or Chemspider (http://www.chemspider.com) for the identification of monoisotopic masses, or on Metabolomics Fiehn Lab Mass Spectrometry Adduct Calculator (http://fiehnlab.ucdavis.edu/staff/kind/Metabolomics/MS-Adduct-Calculator/) for adduct ion detection.
The L, a and b values were measured with a Minolta chromameter (CR-200, 78903131; Ramsey, NJ, USA) on fruit at different stages during fruit ripening. The chromameter was calibrated against a standard white tile. Hue angle values were calculated according to the following equation: Hue angle = tan−1(b/a) if a > 0 or 180 + tan−1(b/a) if a < 0.
Fifteen fruits from each line were harvested at the Br stage and the firmness was assessed using Harpenden calipers (British Indicators Ltd, Burgess Hill, UK) as described by Ecarnot et al. (2013).
Fruits at each developmental stage were harvested and placed in open 120-ml jars for 2 h to minimize the effect of wound ethylene caused by picking. Jars were then sealed and incubated at room temperature for 35 min, and 1 ml of headspace gas was injected into an Agilent 7820A gas chromatograph equipped with a flame ionization detector (Agilent, Santa Clara, CA, USA). Samples were compared with reagent-grade ethylene standards of known concentration and normalized for fruit weight.
Gene ID data for the genes described in this article are listed in Table S2.
Expression patterns of tomato ERF genes during fruit ripening
Ethylene is known to play a critical role in fruit development and ripening, and ERFs are considered to be the primary actors in mediating responses to this hormone. To gain further insight into the expression of members of the tomato ERF gene family during the ripening process, the accumulation of Sl-ERF transcripts was assessed by quantitative RT-PCR at different stages of fruit ripening. Although the expression dynamics of most ERF genes suggest their involvement in the ripening process, no clear link could be established between their repressor or activator function and their pattern of expression. The data indicated (Fig. 1) that, among the 25 ERFs tested, transcript accumulation of five genes (Sl-ERF.B3, C2, E2, E4 and F2) peaked at the Br stage and then decreased at later ripening stages. The transcript levels of 11 genes (Sl-ERF.A2, A3, B1, B2, C1, C3, D1, D2, F3, F5 and G2) showed an increase, peaking at 3 d post-Br (Br + 3), and then declined. A distinct expression pattern was displayed by a group of ERF genes (Sl-ERF.A1, C6, D3, D4, E1, F1, F4 and H1) that underwent a steady increase in transcript accumulation throughout ripening.
Sl-ERF.B3 shows a fruit development- and ripening-related expression pattern
Of particular interest, Sl-ERF.B3 transcript accumulation showed a dramatic increase at the onset of ripening and maintained high levels at subsequent post-Br stages, suggesting that its expression may be continuously required throughout the ripening process. This observation motivated the further assessment of Sl-ERF.B3 transcript accumulation in vegetative and reproductive tissues by quantitative RT-PCR (Fig. 2a). Expression analysis in various plant tissues and organs, including stem, root, leaf, flower and a series of fruit developmental stages, indicated that the accumulation of Sl-ERF.B3 transcripts was relatively high in both vegetative and reproductive tissues (Fig. 2a).
The functional significance of this tomato ERF family member was first addressed by attempting to alter its expression using antisense or over-expression strategies. However, both approaches failed to provide significant clues on the physiological role of Sl-ERF.B3, which prompted the use of a dominant repressor version of the gene (ERF.B3-SRDX), relying on CRES-T. This technology has been developed to study the consequences of silencing of the target genes of single transcription factors, and has also been used to overcome the experimental limitations caused by functional redundancy of transcription factor families (Hiratsu et al., 2003). Ten transgenic ERF.B3-SRDX lines showed a characteristic phenotype with different expressivity, and three (SR1, SR2 and SR3) were selected for further molecular and physiological studies. The relative expression levels of the ERF.B3-SRDX transcript in fruit tissues of the three independent lines was assessed by quantitative RT-PCR using primers specific for the transgene (Fig. 2b).
Altered fruit development in ERF.B3-SRDX dominant repressor lines
One of the most evident phenotypes displayed by ERF.B3-SRDX transgenic lines was the altered fruit shape and reduced size (Fig. 3a). Wild-type tomato fruits (Fig. 3a) were round in shape, in contrast with the ERF.B3-SRDX fruits, which were heart shaped with bumpy areas present intermittently on the surface of the fruit (Fig. 3a). Changes in fruit anatomy also included a thicker pericarp and decreased jelly formation, with enhanced pericarp to fruit radius ratio when compared with the wild-type (Figs 3a, S1a,b). As a consequence of the smaller size, the mean weight of ERF.B3-SRDX fruits was significantly reduced (Fig. 3b). The number of seeds was also dramatically decreased in ERF.B3-SRDX fruits compared with the wild-type, and the average seed number decreased from 25 per fruit in the wild-type to six in ERF.B3-SRDX lines (Fig. S2a). In addition, the seeds showed reduced size (Fig. S2b) and, in the ERF.B3-SRDX line showing the strongest phenotype, seed weight was less than half that of the wild-type (Fig. S2c). Interestingly, although ERF.B3-SRDX lines yielded hairless seeds with altered morphology (Fig. S2b), the seeds produced viable plants.
ERF.B3-SRDX fruits fail to display a red-ripe phenotype
In addition to the altered fruit shape and size, ERF.B3-SRDX lines exhibited distinct ripening features. The ripening-related phenotypes were investigated in wild-type and ERF.B3-SRDX lines using fruits at different developmental stages sampled from the same truss. Dramatic changes were revealed with regard to both the time at which the ripening started and the speed at which it proceeded in the dominant repressor lines. The onset of ripening occurred with at least 2 wk delay in ERF.B3-SRDX lines (57 d post-anthesis) compared with the wild-type (41 d post-anthesis), suggesting that the attainment of competence to ripen was dramatically delayed in the transgenic lines (Table 1). Moreover, once the ripening process started at the Br stage, the color change proceeded much more slowly in ERF.B3-SRDX fruits compared with the wild-type (Fig. 4a). Indeed, in contrast with wild-type fruit, which reached the red-ripe stage 5 d post-Br (Br + 5), ERF.B3-SRDX fruits remained orange at the Br + 10 stage (Fig. 4a). The assessment of color change via measurement of the evolution of hue angle values, indicative of color saturation, further emphasized the difference between the wild-type and dominant repressor lines throughout the ripening process (Fig. 4b). The value of the hue angle was even higher for ERF.B3-SRDX fruit at Br + 10 than for the wild-type at Br + 5, thus confirming the orange-ripe phenotype observed visually (Fig. 4b).
Table 1. Time period from anthesis to the breaker stage in wild-type and ERF.B3-SRDX tomato (Solanum lycopersicum) lines
Values represent means ± SD of at least 15 fruits for each line. **, 0.001 < P <0.01; ***, P <0.001 (Student's t-test).
ERF.B3-SRDX fruits show fast softening and elevated ethylene production
To uncover whether the failure to reach the red-ripe stage in dominant repressor fruits resulted from the incapacity to enter a ripening process, other major ripening-associated features, such as softening and climacteric increase in ethylene production, were investigated. The evolution of firmness determined from the Br stage to 20 d post-Br (Br + 20) clearly showed that ERF.B3-SRDX transgenic fruits underwent significantly faster softening than control fruits (Fig. 4c, Table S3). Given that fruit softening is highly regulated by ethylene, the production of ripening-associated ethylene was assessed in ERF.B3-SRDX fruits. As shown in Fig. 4(d), the accelerated softening observed in transgenic fruits was associated with a dramatic increase in climacteric ethylene production (Fig. 4d), which reached a maximum of three to four times higher than that in wild-type fruit. Altogether, these data indicate that, once the ripening process is triggered, it proceeds more rapidly in the ERF.B3-SRDX repressor than in wild-type fruit.
Decreased lycopene and increased β-carotene levels are responsible for the orange-ripe phenotype in ERF.B3-SRDX fruits
To investigate the cause of the altered pigmentation in ERF.B3-SRDX fruits, LC-PDA-MS analysis of carotenoid levels was performed on wild-type and ERF.B3-SRDX fruits at both the Br and Br + 7 stages. Total carotenoid content was reduced by 38–51% in dominant repressor lines at Br and post-Br stages (Fig. 5, Table S4). Notably, levels of lycopene and its precursors, phytoene, phytofluene, ζ-carotene and neurosporene, were decreased significantly in ERF.B3-SRDX fruits at the ripe stage (Br + 7) and, concomitantly, a sharp increase in β-carotene content was observed (Fig. 5, Table S4), in keeping with the orange-ripe phenotype. Moreover, dominant repressor fruits accumulated higher levels of α-carotene than did wild-type fruits (Fig. 5, Table S4).
To uncover the molecular basis of the altered carotenoid composition in ERF.B3-SRDX lines, we examined the transcript levels of genes involved in the carotenoid biosynthesis pathway at different stages of fruit ripening by quantitative RT-PCR (Fig. 6). Although the phytoene synthase (PSY1) transcript showed, in ERF.B3-SRDX fruits, a ripening-regulated pattern similar to that of the wild-type, its levels were dramatically reduced at all ripening stages (Fig. 6). PSY1 is a key regulator of flux through the carotenoid pathway and its repression is consistent with the reduction in lycopene and total carotenoids in Br stage fruits (Fig. 5). A decrease in phytoene desaturase (PDS) expression levels was also observed in ERF.B3-SRDX fruits (Fig. 6). By contrast, transcript accumulation of all three lycopene β-cyclases (β-LCY1, β-LCY2, CYC-β) was markedly elevated in ERF.B3-SRDX fruits compared with the wild-type (Fig. 6), probably accounting for the increased α- and β-carotene content in ERF.B3-SRDX lines (Fig. 5). The data indicate that the dominant repressor version of ERF.B3 leads to decreased expression of PSY1 and PDS, and increased expression of lycopene β-cyclases, thus resulting in a modified lycopene to β-carotene ratio.
Ethylene- and ripening-related genes are highly induced in ERF.B3-SRDX-expressing fruits
To gain some insight at the molecular level into the ripening of ERF.B3-SRDX fruits, we examined the transcript accumulation of a set of ripening-related genes. Once the ripening process had started, the expression of ethylene biosynthesis genes, such as ACS2, ACS4 and ACO1, was significantly higher in ERF.B3-SRDX-expressing fruits than in the wild-type (Fig. 7). Transcript accumulation of these genes was similarly low in transgenic and control fruit at the mature green stage, but was more strongly induced after the Br stage in the dominant repressor lines, concomitant with the rise in ethylene production. In addition, mRNA accumulation of ethylene-inducible genes, such as E4 and E8, was also increased in ERF.B3-SRDX lines (Fig. 7), consistent with the elevated ethylene production. The transcript accumulation of a major fruit polygalacturonase gene, PG2A, involved in ripening-related cell wall metabolism, was significantly induced in ERF.B3-SRDX fruits (Fig. 7), in line with the enhanced softening phenotype. Similarly, the expression of key regulatory genes of the ripening process, such as RIN, NOR and CNR, was increased at post-Br stages compared with the wild-type, although their induction took place later than in control fruit (Fig. 8). The altered expression pattern of these genes in the ERF.B3-SRDX fruits is consistent with the dramatically delayed onset of ripening in the transgenic fruits. Moreover, the mRNA levels of LeHB-1, another ripening regulator gene, were higher in ERF.B3-SRDX lines at all ripening stages (Fig. 8). By contrast, the expression of TAGL1, a tomato SHATTERPROOF gene, and AP2a, an AP2/ERF family gene acting as a negative regulator of fruit ripening, did not display significant changes in ERF.B3-SRDX dominant repressor fruits compared with the wild-type (Fig. 8).
A number of ERF gene family members show altered expression in the ERF.B3-SRDX lines
Considering the putative role of ERFs in mediating ethylene responses, and given the major role devoted to ethylene in regulating the ripening process, we examined the transcript levels of 25 Sl-ERF genes in both wild-type and ERF.B3-SRDX fruits. A dramatic change in the transcript levels for a number of ERF genes was revealed in the dominant repressor lines (Fig. 9). That is, among the 25 Sl-ERFs that showed detectable transcript accumulation, 10 were significantly down-regulated in the ERF.B3-SRDX dominant repressor lines, and eight Sl-ERFs displayed up-regulation in the transgenic lines (Fig. 9). It is noteworthy that the accumulation of transcripts corresponding to Sl-ERF.A1, whose expression was strongly induced during ripening (Fig. 1), was dramatically enhanced in the ERF.B3-SRDX-expressing lines.
So far, only a limited number of studies have reported the direct involvement of ERF genes in fleshy fruit development and ripening (Li et al., 2007; Lee et al., 2012). This is rather striking given the well-accepted role of ERFs in mediating ethylene responses and the major role played by this phytohormone during the ripening of climacteric fruit. In this regard, the present study, by shedding new light onto the physiological significance of Sl-ERF.B3, supports the idea that members of the tomato ERF family of transcription factors are active players in the control of fruit ripening in tomato. Sl-ERF.B3 has been shown recently to play an important role in controlling pleiotropic ethylene responses via the feedback regulation of genes encoding components of ethylene signaling and other ERFs (Liu et al., 2013). Here, we show that the expression of this dominant repression version of Sl-ERF.B3 (ERF.B3-SRDX) broadly impacts tomato fruit development and ripening. The ripening-related phenotypes displayed by the ERF.B3-SRDX lines are supportive of a prominent role for Sl-ERF.B3 in regulating important aspects of fruit ripening in tomato, including the attainment of competence to ripen and subsequent changes in color and texture. Although Sl-ERF.B3 expression is high in stem tissue, the induced accumulation of its transcripts at the Br stage is indicative of an active role for this ERF gene in triggering the ripening process and, in this regard, it is not surprising that the onset of ripening is delayed in the dominant repressor lines. Compared with the wild-type, the time period from anthesis to Br was extended by c. 2 wk in the transgenic lines, indicating that the dominant repression activity of Sl-ERF.B3 impacts tomato early fruit development and, particularly, the attainment of competence to ripen. Another obvious effect of Sl-ERF.B3-SRDX on fruit development is the reduced fruit size and bumpy shape, as a result of a reduction in epidermal cell size and a defect in the normal coordinated expansion of the pericarp (Fig. S3). This is illustrated by a thicker pericarp, smaller volume of jelly and dry/crumbly appearance of the pericarp, suggesting a defect in the expansion or elasticity of the epidermis.
The difference in color development between wild-type and ERF.B3-SRDX fruits is obvious at the post-Br stages, and ERF.B3-SRDX fruits fail to turn red, retaining an orange color at late ripening. However, the incapacity to reach a red-ripe color is not caused by an inability to undergo ripening. Indeed, once ripening is triggered, the ERF.B3-SRDX fruits display faster softening and higher climacteric ethylene production, indicating that these aspects of ripening are disconnected from red pigment accumulation. The dramatic development of red pigmentation of ripening fruits is one of the most notable features of tomato, mainly as a result of an accumulation of the red carotene, lycopene. Lycopene accumulation during ripening is caused by the up-regulation of genes for lycopene biosynthesis, including PSY1 and PDS (Giuliano et al., 1993), and the down-regulation of lycopene cyclases, which convert lycopene into downstream compounds (Pecker et al., 1996; Ronen et al., 1999, 2000). Accordingly, the silencing of PSY1 or PDS results in a reduction in fruit lycopene (Fray & Grierson, 1993; Fantini et al., 2013), whereas over-expression of β-LCY or CYC-β cyclases results in a conversion of lycopene into β-carotene, which turns the fruit orange (Ronen et al., 2000; Rosati et al., 2000). In addition, a positive link between β-LCY and CYC-β expression and β-carotene levels was revealed by correlation analysis of fruit metabolome and transcriptome data from S. pennellii × S lycopersicum introgression lines (Lee et al., 2012). Thus, the changes in fruit carotenoid content and pigmentation observed in ERF.B3-SRDX fruits are perfectly compatible with the changes in expression of the PSY1, β-LCY and CYC-β genes observed in such fruits. It should be noted that this effect contrasts with that exerted on other ripening-regulated genes. That is, although the majority of ripening-regulated genes show an exaggerated response in ERF.B3-SRDX fruits, carotenoid genes show an opposite trend, with PSY1 and PDS induction during ripening being repressed and lycopene cyclases being up-regulated, contrasting with their behavior in wild-type fruits. These contrasting effects of ERF.B3-SRDX on carotenoid biosynthesis genes, in comparison with other ripening-inducible genes, result in the contrasting phenotypes displayed by ERF.B3-SRDX fruits: exaggerated softening vs reduced lycopene accumulation.
The effects of ERF.B3-SRDX over-expression on carotenoid gene expression are not necessarily mediated by ethylene: first, because PSY1 expression seems to be controlled by upstream ripening regulators, such as NOR, rather than by ethylene itself (Alba et al., 2005; Osorio et al., 2011), and, second, because the expression of carotenoid genes in ERF.B3-SRDX fruits shows a contrasting behavior with other ethylene responses and with ethylene itself.
The decreased total carotenoid levels and elevated ethylene production in Sl-ERF.B3 fruits at least partially resemble the phenotype of lines repressed in the SlAP2a regulatory gene, in which significantly elevated ethylene levels are associated with altered total carotenoids and a shift to β-carotene rather than lycopene (Chung et al., 2010; Karlova et al., 2011). Moreover, the phenotype also recalls that of the tomato Nr mutant, in which lycopene biosynthesis and PDS gene expression are repressed and ethylene production is increased (Alba et al., 2005). The assessment of the relative mRNA accumulation of SlAP2a failed to reveal a significant difference in its levels between ERF.B3-SRDX and wild-type fruits, suggesting that the regulation of Sl-ERF.B3 dominant repression activity in fruit ripening is probably independent of SlAP2a. It is also possible that dominant repression of Sl-ERF.B3 in tomato leads to complex alterations in the carotenoid accumulation network and impacts carotenoid biosynthesis genes through mechanisms beyond the influence of ethylene. Transcription factors impacting carotenoid accumulation in tomato include RIN (Vrebalov et al., 2002), CNR (Manning et al., 2006), HB-1 (Lin et al., 2008), TAGL1 (Vrebalov et al., 2009), SlAP2a (Chung et al., 2010; Karlova et al., 2011), Sl-ERF.E4, formerly SlERF6 (Lee et al., 2012), and SlMADS1 (Dong et al., 2013). The expression data revealed altered transcript accumulation for RIN, CNR, HB-1 and Sl-ERF.E4 in ripening ERF.B3-SRDX fruit, suggesting that these transcription factors may be involved in the regulation networks of carotenoid accumulation in the ERF.B3-SRDX lines. Interestingly, an Arabidopsis ERF transcription factor, RAP2.2, has been reported to regulate the expression of carotenoid biosynthesis genes via binding to the ATCTA cis-element in the promoter regions of PSY and PDS (Welsch et al., 2007). The presence of the ATCTA motif in the promoter regions of both tomato PSY1 and PDS genes, together with the suppression mediated by ERF.B3-SRDX on its target genes, support the hypothesis that the chimeric ERF.B3-SRDX protein represses the expression of PSY1 and PDS in ERF.B3-SRDX lines by binding directly to their promoters.
A hallmark of climacteric fruit ripening, such as tomato, is the dramatic induction of respiration and ethylene production at the onset of ripening. The dominant repressor version of Sl-ERF.B3 in tomato resulted in substantially elevated levels of ethylene production (Fig. 4d), although the onset of ripening was dramatically delayed. This indicates that the altered fruit ripening phenotype in ERF.B3-SRDX lines occurs at least partly through an influence on ethylene synthesis. ERF.B3-SRDX fruits produced up to four-fold more ripening ethylene than the wild-type and, accordingly, displayed elevated transcript accumulation of ethylene biosynthesis genes, including ACS2, ACS4 and ACO1 (Fig. 7). Ethylene biosynthesis is tightly controlled by ACS and ACO multigene families during fruit development and ripening (Nakatsuka et al., 1998; Barry et al., 2000; Barry & Giovannoni, 2007). Two systems of ethylene regulation have been proposed (McMurchie et al., 1972), with System 1 representing the basal level of ethylene in immature fruit and vegetative tissues, and System 2 corresponding to high levels of ethylene production associated with fruit ripening (Oetiker & Yang, 1995). In contrast with System 1, where ethylene auto-inhibits its own biosynthesis, System 2 ethylene has a stimulatory effect on its own synthesis, the so-called autocatalytic ethylene production. Tomato ACS1 and ACS6 have been shown to mediate System 1 ethylene production in immature fruit in tomato, whereas autocatalytic ethylene biosynthesis in System 2 is mediated through ethylene-stimulated expression of ACS2, ACS4, ACO1 and ACO4 genes (Nakatsuka et al., 1998; Barry et al., 2000; Barry & Giovannoni, 2007). Tomato ACS2 and ACS4 are the predominant ACS mRNAs in ripening fruit (Barry et al., 2000; Yokotani et al., 2004). The accumulation of ACS2 mRNA is induced at the onset of ripening in an ethylene-dependent, but RIN-independent, manner (Nakatsuka et al., 1998; Barry et al., 2000), and repression of the ACS2 gene blocks fruit ripening in tomato (Oeller et al., 1991). Likewise, ACS4 is also up-regulated during ripening, but this induction is RIN dependent. The accumulation of ACO1 transcripts increases at the onset of ripening and is sustained at a high level during subsequent tomato fruit ripening (Nakatsuka et al., 1998), suggesting the critical role of this gene in controlling ripening-associated ethylene synthesis. As ACS and ACO catalyze the rate-limiting and final steps in ethylene biosynthesis, the significantly high mRNA levels of ACS2, ACS4 and ACO1 are probably responsible for the elevated ethylene levels in the ERF.B3-SRDX lines. It is noteworthy that expression of the repressor version of Sl-ERF.B3 leads to reduced ethylene production in dark-grown seedlings (Liu et al., 2013), whereas it results in elevated ethylene production in ripening fruit, in keeping with the contrasting behavior of vegetative and ripening fruit tissues, in which ethylene production is controlled by System 1 and System 2, respectively.
The expression dynamics of most Sl-ERF genes during fruit ripening supports their putative involvement in the ripening process. Interestingly, the expression pattern of a large number of Sl-ERF genes was found to be markedly altered in ERF.B3-SRDX fruits at the post-Br stage, which may account for the altered ripening phenotype displayed by the dominant repressor fruits. Of particular note, the transcript accumulation of Sl-ERF.A1, whose expression is strongly induced during ripening (Fig. 1), was dramatically enhanced in the ERF.B3-SRDX-expressing lines. Although the dramatic change in the expression of Sl-ERF.A1 suggests a particular role for this gene in the ripening process, its up-regulation in the dominant repressor lines seems to rule out the possibility that it may be a direct target of Sl-ERF.B3. By contrast, Sl-ERF.C3, Sl-ERF.D2, Sl-ERF.F5 and Sl-ERF.F4 genes, shown to be putative targets of Sl-ERF.B3 (Liu et al., 2013), are all down-regulated in ERF.B3-SRDX fruits, further supporting the model that a single ERF can impact the expression of other members of the gene family. This interconnected regulation among ERF genes may account for the complexity of the mechanisms controlling the ripening-associated processes, and therefore for the pleiotropic alterations displayed by the ERF.B3-SRDX lines.
The authors are grateful to L. Lemonnier, D. Saint-Martin and O. Berseille for genetic transformation and culture of tomato plants. This research was supported by the ‘Laboratoire d'Excellence’ (LABEX) entitled TULIP (ANR-10-LABX-41) to M.B. and the EU-DISCO project to G.G., and benefited from the networking activities within the European COST Action FA1106.