Introduction

Tomato (Solanum lycopersicum) is the second most highly produced vegetable crop worldwide after potato (S. tuberosum), with a production of 164 million tons, valued at ~$60 billion, in 2013 (http://faostat.fao.org/). Tomato fruit are used for both fresh consumption and in processing and they have been shown to have many beneficial effects on human nutrition and health (Adalid et al. 2010; Lin et al. 2014; Tanksley 2004). In addition, tomato is also a widely adopted experimental model plant in a wide range of plant research fields, including genetics (Chetelat et al. 2000; Tanksley 2004), evolution (Moyle 2008; Spooner et al. 2005), reproductive biology (Bedinger et al. 2011; Li and Chetelat 2010), genomics (Lin et al. 2014; Tomato Genome Consortium 2012), transcriptomics (Matas et al. 2011), proteomics (Yeats et al. 2010) and metabolomics (Alba et al. 2005; Osorio et al. 2011).

Tomato fruit quality is a major target for tomato breeders and consumers (Giovannoni 2006) and, in this regard, one aspect of particular interest is the color of the ripe fruit, which results from the accumulation of pigments such as carotenoid and flavonoids, as well as from chlorophyll degradation (Delgado-Vargas et al. 2000; Rosati et al. 2010). These pigments reduce damage from photo-oxidation and contribute to flower and fruit coloration which attracts insects and animals, thereby promoting pollination and seed dispersal. The pigments are also considered to be bioactive compounds and have been implicated in the prevention of cancer and cardiocerebrovascular disease in humans (Neto 2007; Riccioni 2009; Seeram et al. 2004; Wang et al. 2014).

Studies over the last decade, many involving mutants and transgenic lines, have helped elucidate the structural and regulatory pathways that influence mature fruit color. The fruit of some tomato genotypes appear pink, due to the reduced accumulation of the yellow flavonoid pigment naringenin chalcone in their peels. This was demonstrated by suppressing the expression of the key gene in flavonoid biosynthesis CHS (CHALCONE SYNTHASE), and in studies where CHI (CHALCONE ISOMERASE) was over-expressed, thereby reducing naringenin chalcone accumulation (Muir et al. 2001; Schijlen et al. 2007). Furthermore, a transcription factor MYB12 which regulated CHS expression has been shown to be the causal gene for the tomato colorless-peel y mutant, which has pink fruit (Adato et al. 2009; Ballester et al. 2010; Lin et al. 2014). The anthocyanin class of pigments is not typically detectable in cultivated tomato fruits, although the fruit of some wild tomato relatives have purple anthocyanin containing peels. Several genes associated with anthocyanin biosynthesis, including AFT (ANTHOCYANIN FRUIT from S. chilense), ATV (ATROVIOLACEA from S. cheesmaniae) and ABG (AUBERGINE from S. lycopersicoides), have been transferred into cultivated tomato, thereby creating purple fruited tomato lines with high anthocyanin levels (Gonzali et al. 2009). Moreover, the overexpression of SlANT1, a MYB transcription factor, resulted in purple spotting on the epidermis and pericarp of tomato fruits (Mathews et al. 2003).

Ripe fruits of the tomato mutant gf (green-flesh) remain green due to inhibited degradation of their thylakoid membranes, which hinders the developmental transition of chloroplasts into chromoplasts (Barry et al. 2008; Cheung et al.1993). In addition, the SGR (STAY-GREEN) gene is thought to be responsible for this mutated phenotype (Hörtensteiner 2009).

In addition to flavonoids and chlorophyll, carotenoid pigments are also important determinants of tomato fruit color, and each step in the carotenoid biosynthesis pathway has been studied in detail (Klee and Giovannoni 2011). Carotenoids comprise a class of isoprenoids, consisting of a 40 carbon long chain with a conjugated double bond. They are synthesized in the plastids from two geranylgeranyl diphosphates (GGPP), which are condensed into phytoene by phytoene synthase (PSY). Phytoene is then converted into all-trans-lycopene via the intermediate ζ-carotene product by phytoene desaturase (PDS), ζ-carotene isomerase (ZISO), ζ-carotene desaturase (ZDS) and cis–trans isomerase (CRTISO). Lycopene is located at the key branching point of α- and β-carotene, and is responsible for the red color of tomato fruits. The linear all-trans-lycopene can be converted into α-carotene, β-carotene and other xanthophylls through cyclization reactions and oxygenations by lycopene cyclase (LCY) and carotenoid hydroxylase (CHY) enzymes in the α-/β-carotene branches, respectively (Gao et al. 2015; Tomato Genome Consortium 2012).

PSY1 has been shown to be the rate-limiting enzyme in the carotenoid synthesis pathway and much research in tomato has focused on this enzyme and the corresponding gene. For example, PSY1 has altered expression in the tomato mutants, r (yellow-flesh) and r y, resulting in pale yellow fruit flesh and yellow fruit skin, respectively (Fray and Grierson 1993). Two mutational events, in an intron and in the 3′ untranslated region (UTR), of tomato PSY1 (accession PI114490) led to a yellow-fruited phenotype, although the causal mechanism is still unknown (Yuan et al. 2008; Kang et al. 2014). Additionally, the mutated locus, r 2997, in a yellow-fleshed tomato was confirmed to be linked to PSY1, the transcripts of which were not detected (Kachanovsky et al. 2012). Isaacson et al. (2002) found that two orange-fruited tomato mutants, tangerine mic and tangerine 3183, accumulated pro-lycopene instead of all-trans-lycopene, and the two mutated loci were mapped to CRTISO (CAROTENOID ISOMERASE). Another example of a mutation in the carotenoid biosynthesis pathway is the orange-fruited mutant Delta, which accumulates δ-carotene at the expense of lycopene, and whose locus was shown by Ronen et al. (1999) to co-segregate with LCY-E (LYCOPENE ε-CYCLASE). Finally, the orange-fruited Beta and dark red-fruited old-gold mutants show increased β-carotene and lycopene levels, respectively, and the mutated loci were both mapped to CYC-B (CHROMOPLAST-SPECIFIC LYCOPENE β-CYCLASE) (Ronen et al. 2000).

In addition to genes involved in carotenoid and flavonoids synthesis as well as chlorophylls degradation, several other genes have been shown to influence tomato fruit color formation. For example, the high pigment mutants (hp-1, hp-2, hp-3) exhibit higher fruit pigmentation due to elevated lycopene, flavonoid and chlorophyll levels, and have been mapped to DDB1 (UV-DAMAGED DNA BINDING PROTEIN 1) (Lieberman et al. 2004), DET1 (DEETIOLATED1) (Levin et al. 2003; Mustilli et al. 1999) and ZEP (ZEAXANTHIN EPOXIDASE) (Galpaz et al. 2008), respectively. Other regulators of fruit coloration include transcription factors associated with fruit ripening such as RIN (RIPENING INHIBITOR) (Vrebalov et al. 2002), CNR (COLORLESS NON-RIPENING) (Manning et al. 2006), NOR (NON-RIPENING) (Zhu et al. 2014) and components of the ethylene signaling transduction such as NR (NEVER RIPE) (Lanahan et al. 1994) and EIN2 (ETHYLENE INSENSITIVE2) (Hu et al. 2010; Zhu et al. 2006).

Although it is well established that carotenoid, flavonoids and chlorophyll are involved in fruit coloration, the complex molecular mechanisms and regulatory networks governing their accumulation are still not well understood. In this study, we described the characterization of a yellow-fruited tomato mutant n3122, named as yellow-fruited tomato 1 (yft1), including mapping of the underlying mutation, biochemical analysis of the carotenoid profiles and an evaluation of molecular processes associated with the phenotype, to better understand the determinants of fruit color.

Materials and methods

Plant material

Seeds of n3122, M82 (S. lycopersicum) and LA1585 (S. pimpinellifolium) (Fig. 1) were kindly provided by Prof. Dani Zamir (http://zamir.sgn.cornell.edu/mutateds) and the Tomato Genetics Resource Center (University of California, Davis, CA, USA, http://tgrc.ucdavis.edu), respectively. All the materials including M82, n3122, LA1585 and the mapping population were planted and grown under standard greenhouse conditions at the Pujiang experimental base, Shanghai Jiao Tong University, Shanghai, China.

Fig. 1
figure 1

Fruit color of the parental lines and their progeny. a M82; b n3122; c LA1585; d F1 progeny of n3122 × LA1585; e the red-fruited F2 progeny of n3122 × LA1585; f the yellow-fruited F2 progeny of n3122 × LA1585; g F1 progeny of M82 × n3122; h the red-fruited F2 progeny of M82 × n3122; i the yellow -fruited F2 progeny of M82 × n3122

Full size image

Identification of the candidate gene

The two crosses, M82 × n3122 and n3122 × LA1585, were created for genetic analysis and the color of fully ripe fruit from each plant of the F1 and F2 generations was evaluated. The n3122 × LA1585 F2 population was also used to identify the candidate gene by map-based cloning. Eighty-nine plants, including 25 yellow-fruited and 64 red-fruited plants, were genotyped using 48 markers, which collectively covered the entire tomato genome, to create linkage groups with the mutated loci. These markers were chosen from the Sol Genomics Network (http://solgenomics.net/) or designed according to the sequence of the tomato genome (Table S1). A total of 1900 plants from the F2 population were used to fine map the candidate gene using newly designed CAPS (cleaved amplified polymorphic sequences) or dCAPS (derived cleaved amplified polymorphic sequences) markers (Table S1). Data were analyzed using R/qtl software (Broman et al. 2003).

Genes located in the mapped target region were identified using the tomato genome annotation (http://solgenomics.net/), and the candidate gene, EIN2, was amplified using LA Taq DNA polymerase (Takara, China) with the gene-specific primers 5′-ATGGAGTCTGAAACTCTGACTAGAG-3′ (forward) and 5′-TTACAAGACGAAAGGGGGTGATG-3′ (reverse) for the gene sequence from the start codon ATG to the stop codon TAA, 5′-GATGTTATCGGGGAGGAAGA-3′ (forward) and 5′-TATTTGAT TTAAATGGTTAG-3′ (reverse) for the promoter sequence before the start codon ATG, and 5′-CTTGTTCATGTCAAATGGGTTG-3′ (forward) and 5′-AACCACAGCATCAGACAAAGAC-3′ (revere) for the nucleotide sequence around the start codon ATG.

EIN2 transcript levels were evaluated using quantitative real-time PCR (qRT-PCR). Total RNA was extracted using an RNAprep pure Plant Kit (Tiangen, China) from fruit pericarps of both M82 and n3122 at 35 days post-anthesis (DPA), 47 DPA, 50 DPA and 54 DPA which corresponded to the MG (mature green), BR (breaker), BR + 3 (3 days after breaker) and RR (red ripe) stages of M82 as well as n3122 at 60 DPA which corresponded to the BR stage in this genotype (Fig. S1). One μg total RNA was used as a template to synthesize cDNA using the PrimeScript™ RT Master Mix kit (Takara, China). qRT-PCR was performed in a 20 μL reaction volume containing 2 ng of cDNA and the SYBR premix Ex-Taq (Takara, China) with a qRT-PCR machine (Bio-Rad, UK). The PCR amplification program was 37 °C for 2 min and 94 °C 3 min, followed by 40 cycles of 94 °C for 20 s, 60 °C for 20 s and 72 °C for 20 s. Transcript levels were calculated using the 2−ΔCT equation (Kilambi et al. 2013), and ACTIN (Genbank Accession number BT013524) was used as a reference gene for gene expression normalization. EIN2 gene-specific primers were 5′-ATGACAGGGATGATGGAGATTCG-3′ (forward) and 5′-TATGACCCCGGACCATCAGA-3′ (reverse) and ACTIN gene-specific primers were 5′- TTGCTGACCGTATGAGCAAG-3′ (forward) and 5′-GGACAATGGATGGACCAGAC-3′ (reverse).

Carotenoid measurements

Tomato fruit were collected from both M82 and n3122 at 35 DPA, 47 DPA, 50 DPA and 54 DPA as well as n3122 at 60 DPA. The equatorial region of the pericarp of the collected fruits was sampled in three biological replicates and ground to a fine powder in liquid nitrogen prior to high-performance liquid chromatography (HPLC) analysis. Extraction and quantification of carotenoid were carried out as previously described (Fraser et al. 2000) with some modifications. 500 mg fine powder sample was used for carotenoid extraction. The dried extracted residues were dissolved in 1 mL hexane and then filtered by a 0.22-μm filter. For carotenoid separation and detection, 60 μL of filtered solution was injected into a Hitachi L-2000 (Tokyo, Japan) HPLC system coupled with a reverse-phase C30 column (5 μm, 250 × 4.6 mm, YMC Inc., USA) and a diode-array detector monitoring continuously from 220 to 500 nm. The elution programme using methanol (A), water/methanol (v:v = 20:80) with 0.2 % ammonium acetate (B) and tert-methyl butyl ether (C) as the mobile phases with 1 ml·min-1 flow rate was carried out as follows: initially an isocratic step of 95 % A, 5 % B for 12 min; then three linear gradient steps: firstly, the mobile phases changed to 80 % A, 5 % B, 15 % C by 5 min, secondly to 50 % A, 5 % B, 45 % C by 10 min, thirdly to 25 % A, 5 % B, 70 % C by 10 min; subsequently isocratically for 13 min. One more linear gradient step of 5 min was performed to return to the initial concentrations of the mobile phase, then maintaining 15 min for equilibrating the column. Carotenoid identification was conducted by the comparison of characteristic absorption spectra and retention time with authentic standards. Carotenoid quantification was calculated using a dose–response curve created by standards. Lycopene, β-carotene, zeaxanthin and lutein standards were purchased from the Sigma Chemical Co. (Sigma, USA) and α-carotene from Santa Cruz Biotechnology, Inc. (Santa Cruz, USA).

Chromoplast ultrastructure visualization

The ultrastructure of plastids in the pericarp of 35 DPA, 47 DPA, 50 DPA and 54 DPA fruit of both the M82 and n3122 genotypes was examined, as well as those from 60 DPA n3122 fruit, using transmission electron microscopy (TEM), as previously described (Schweiggert et al. 2011) with some modifications. Small pieces (1–2 mm2) were cut from the mesocarp and immediately immersed into 2.5 % glutaraldehyde in 0.1 M phosphate buffer (17.55 g Na2HPO4·12H2O and 7.95 g NaH2PO4·2H2O per L, pH 6.8) for 5 h at 4 °C and a vacuum was applied for 30 min to ensure efficient fixation. The samples were washed three times for 20 min in a 0.1 M phosphate buffer. A post-fixation step with 2 % osmium tetroxide in 0.1 M phosphate buffer for 3 h followed by three washes in the 0.1 M phosphate buffer. Subsequently, after dehydration with 50 % ethanol for 20 min, the samples were stained with 2 % uranyl acetate in 70 % ethanol for 12 h. Sequential dehydration in a series of ethanol and acetone grades as 70 % ethanol, 90 % ethanol, 90 % ethanol and 90 % acetone (v:v = 1:1), 90 % acetone, 100 % acetone was then performed for 20 min per grade. The dehydrated samples were immersed in acetone and epoxy resin 812 (Hede Biotechnology Co. Ltd., China) (v:v = 1:1 for 6 h, v:v = 1:2 for 12 h) and then twice in epoxy resin 812 for 24 h. After that, the samples were again immersed in epoxy resin 812 and kept at 65 °C for 48 h for polymerization. Finally, 70 nm thick sections were isolated from the samples using a Leica UC6-FC6 microtome with a diamond knife (Leica Microsystems Inc., Germany). These ultrathin sections were stained with lead citrate and imaged using TEM (Tecnai G2 Spirit Biotwin, USA).

Responses to ethylene

Tomato fruit were collected from both M82 and n3122 at 35 DPA, 47 DPA, 50 DPA and 54 DPA, as well as at 60 DPA from n3122. The fruit were weighed and kept at 25 °C for 2 h prior to ethylene measurements to allow the harvesting-induced burst of ethylene synthesis to subside. To measure endogenous ethylene levels, fruits from each time point (n = 5) were sealed in 500 mL airtight bottles and kept at 25 °C for 4 h, before 1 mL headspace gas was taken from the bottle and injected into a gas chromatograph (GC-2010, Shimadzu Corporation, Japan) equipped with an HP-PLOT/Q column and a flame ionization detector. The injector temperature was set at 200 °C and the flame ionization detector at 250 °C. The oven temperature was initially set at 40 °C for 5 min and then increased to 220 °C at a rate of 30 °C/min. The endogenous ethylene concentration (nL g−1 h−1) was calculated using a dose–response curve created with a 10 ppm ethylene standard.

The responses of tomato fruit and seedlings to exogenous ethylene were measured as follows. The MG stage fruit (35 DPA, n = 3) from M82 and n3122 were separately sealed in 500 mL airtight bottles with 20 ppm ethylene. The fruit were held at 25 °C for 24 h, before the fruits were removed and then left at 25 °C to evaluate their ripening. Seeds from both M82 and n3122 were sterilized with 75 % ethanol for 2 min and 25 % NaClO for 10 min, and then rinsed three times in sterile water. The sterilized seeds were germinated on Murashige and Skoog (MS) medium (Sigma, USA) with or without 10 μM 1-aminocyclopropane-1-carboxylic acid (ACC) (Sigma, USA) in petri dishes at 28 °C in the dark, and the lengths of the hypocotyls and the roots were measured on day 8 after sowing.

Results

Characterization of the n3122 phenotype

The n3122 mutant, which was identified from a fast-neutron irradiation treated M82 population (Menda et al. 2004), produced yellow fruit whose color was readily distinguishable from the red-fruited wild type M82 (Fig. 1). We determined that the n3122 fruit showed delayed ripening and reached the BR stage at 60 DPA, in contrast to the M82 fruit which reached the BR stage at 47 DPA (Fig. S1).

A single recessive yft1 allele is associated with the n3122 yellow fruit phenotype

To investigate the genetic basis of the n3122 yellow fruit phonotype, two crosses (n3122 × LA1585 and M82 × n3122) were made and the fruit color of the progeny was evaluated. All fruits from both crosses appeared red (18 F1 plants of n3122 × LA1585 and 12 F1 plants of M82 × n3122) in the F1 generation, whereas both red and yellow fruit were produced by plants of the F2 generation, with a segregation ratio of 79:25 and 58:13 (red:yellow) for n3122 × LA1585 and n3122 × M82, respectively (Table 1). And the Chi square (χ 2) values of red-fruited plants: yellow-fruited plants were 0.0128 for n3122 × LA1585 (\(\chi^2_{0.05}\) = 3.84, p > 0.05) and 1.357 for n3122 × M82, (\(\chi^2_{0.05}\) = 3.84, p > 0.05) (Table 1), which fits the theoretical Mendelian ratio of 3:1, suggesting that the phenotype is controlled by a single recessive allele, yft1.

Table 1 Segregation of fruit color in the segregating populations
Full size table

Fine mapping and identification of a candidate gene underlying yft1

To identify the yft1 candidate gene underlying the mutant phenotype, the F2 generation (n3122 × LA1585) was used as a mapping population. A total of 89 individuals (25 yellow- and 64 red-fruited) were used to construct linkage groups with 48 molecular markers that were evenly distributed among the 12 tomato chromosomes. yft1 was initially mapped to a position between the CAPS markers 9-29 and cLEX-3-N24, which is a ~1 Mb region at the top of chromosome 9. Subsequently, 1900 F2 plants were used for fine mapping, and the yft1 gene was finally assigned to an 88.2 kb region between CAPS markers 9-S and 9-R (Fig. 2a). According to the tomato genome annotation (ITAG2.3, http://solgenomics.net), a total of 12 genes are located in this target region (Table 2); however, of these, only one, Solyc09g007870, which is annotated as encoding the ETHYLENE INSNSITIVE2 (EIN2) protein, has been shown to be essential for the ethylene signal transduction pathway and to be involved in regulating fruit ripening (Zhu et al. 2006; Hu et al. 2010). Nucleotide sequencing revealed that there were no differences in the DNA-coding sequences of the EIN2 gene from n3122 and wild type M82 (Fig. S2). However, a 13 bp deletion and a 573 bp insertion were found at -318 bp upstream of the translation initiation codon of EIN2 in n3122 (Fig. 2b, Fig. S3). Moreover, EIN2 expression was significantly down-regulated in n3122 at all tested time points between 35 DPA and 60 DPA (Fig. 2c), indicating that EIN2 is a candidate gene for yft1.

Fig. 2
figure 2

EIN2 is a candidate gene for the n3122 mutation. a Schematic view of the map-based cloning; b mutation events, a 13 bp deletion and a 573 bp insertion at position −318 bp upstream of the translation initiation site (ATG) of EIN2 in n3122; c transcriptional expression of EIN2 in M82 and n3122 fruits at different developmental stages

Full size image
Table 2 Twelve predicted genes in the 88.2-kb mapping region at the chromosome 9 in tomato
Full size table

n3122 fruit show reduced carotenoid accumulation and delayed chromoplast development

The carotenoid profiles of fruit from both M82 and n3122 were evaluated by HPLC to determine the cause of the yellow fruit color. Both the composition and content of carotenoid were significantly different between n3122 and M82, such that M82 showed a substantial increase in total carotenoid levels during fruit development from 35 DPA (MG) to 54 DPA (BR + 7), with the highest concentration of 102.8 μg g−1 fresh weight (FW) at 54 DPA, which was an approximately 20-fold increase compared to 35 DPA (5.1 μg g−1 FW). In contrast, the total carotenoid levels were consistently low (6.3 μg g−1 FW avg.) in n3122, and the content was only 6 % of that of M82 fruit at 54 DPA (Fig. 3a).

Fig. 3
figure 3

Inhibition of carotenoid accumulation in n3122. a Carotenoid contents including lycopene, α-carotene, β-carotene and lutein, value = mean ± SD (n = 3). b Plastid ultrastructure in M82 and n3122 fruit at different development stages. pg plastoglobule, gr grana, lth long linear thylakoid membrane structure, ccr carotenoid crystalloids, um undulating membrane representing the lost carotenoid crystalloids. ths thylakoid membrane with swollen compartment. Bars 500 nm

Full size image

Based on HPLC analysis, levels of the carotenoid lycopene, α-carotene, β-carotene and lutein, were significantly different between extracts form n3122 and M82 fruit, while zeaxanthin was not detectable at any stage. Lycopene began to accumulate at 47 DPA and the levels increased sharply with fruit ripening, reaching a peak of 86 μg g−1 FW at 54 DPA, thereby contributing up to 83 % of the total carotenoid in M82. However, lycopene was completely absent from all developmental stages of n3122. β-carotene accumulation gradually increased in M82 from 35 DPA to 54 DPA, where it reached a peak of 16.2 μg g−1 FW, equal to 16 % of the total carotenoid concentration. In contrast, β-carotene levels in n3122 were relatively consistent (5.3–6.9 μg g−1 FW) and represented up to 90 % of the total carotenoid. The concentration of α-carotene decreased with fruit development in M82 from 47 DPA to 54 DPA, while lutein content increased, suggesting that lutein accumulated at the expense of α-carotene. The levels of both these compounds remained at relatively low levels in n3122.

We next examined the ultrastructure of the plastids using TEM, to explore carotenoid storage and chromoplast development in n3122. Typical lens-shaped chloroplasts were observed at 35 DPA in both M82 and n3122. In the later BR stages of M82 fruit (47 DPA), the tightly stacked grana organization was rapidly disrupted and replaced by the linear thylakoid membrane structures with swollen compartments which also disappeared at 50 DPA. The number of plastoglobules gradually increased from 35 DPA to 50 DPA, and crystalloid structures and undulating membranous structures appeared at 50 to 54 DPA. In contrast, the lens-shaped chloroplasts were constantly observed in 35 DPA to 50 DPA n3122 fruit and the grana organization was gradually disassembled during fruit development. Although the linear thylakoid membrane structures and even some swollen thylakoid compartments were observed in later stage, the crystalloids or undulating membranous structures were not observed all along (Fig. 3b). These observations indicated that chromoplast development was delayed in n3122 fruit, coincident with the abnormal production of carotenoid accumulation.

n3122 shows impaired ethylene synthesis and perception

The yft1 candidate gene, EIN2, is known to be a key component of ethylene signaling transduction (Qiao et al. 2012), and so we hypothesized that the reduced expression of EIN2 in n3122 might impair ethylene synthesis and perception. To test this, we measured ethylene production from both M82 and n3122 fruit. Ethylene biosynthesis by fruit of both genotypes was relatively low at 35 DPA, but subsequently increased sharply in M82, reaching a peak of 7.33 nL g−1 h−1 FW at 47 DPA (approximately, 14-fold greater than at 35 DPA), before gradually decreasing to 4.95 nL g−1 h−1 FW at 54 DPA. In contrast, n3122 fruit showed a slow increase in ethylene production from 35 DPA to 54 DPA (1.72 nL g−1 h−1 FW), and the amounts of ethylene produced by 47 DPA, 50 DPA, and 54 DPA fruit were significantly lower than those of M82 (p < 0.001). Although a rapid increase in ethylene levels occurred at 60 DPA, reaching 5.20 nL g−1 h−1 FW, they were still significantly lower than levels produced by M82 fruit at 47 DPA (BR stage, p < 0.05 level) (Fig. 4a).

Fig. 4
figure 4

Ethylene synthesis and perception is partially impaired in n3122. a Ethylene production by M82 and n3122 fruit at different developmental stages. Error bars indicate SD (n = 5), triple asterisks and single asterisk indicate that ethylene release from n3122 fruits was significantly lower than that from M82 at the p < 0.001 and p < 0.05 level, respectively (Student’s t test). FW, fresh weight. b Hypocotyl length in 8-day old seedlings. Error bars indicate SD (n = 24), triple asterisks indicate a P value <0.001. c Root length of 8-day old seedlings. Error bars indicate SD (n = 24), ***p < 0.001. d changes in tomato fruit color on the day after MG stage fruit were exposed to exogenous ethylene (20 ppm) for 24 h, bars 3 cm

Full size image

To assess ethylene perception, MG fruits were exposed to exogenous ethylene (20 ppm) for 24 h. The M82 fruits reached the BR stage on day 7 after treatment, which was 5 days faster than normally developing fruit (i.e., no treatment). However, treated n3122 fruit took 13 days to transition from MG to BR (Fig. 4d). To test the sensitivity of the n3122 seedling to ethylene, an ethylene triple response assay (Lanahan et al. 1994) was conducted. The M82 and n3122 seedlings showed no differences in hypocotyl and root length, when grown without exogenous ACC. However, in the presence of 10 μM ACC n3122 hypocotyls and roots were significantly longer than those of M82 (p < 0.001) (Fig. 4b, c). Both roots and hypocotyls were significantly shorter under these conditions in both genotypes compared to those grown without ACC (p < 0.001). These results suggested that n3122 exhibited partial insensitivity to ethylene.

Discussion

EIN2 is a candidate gene for the n3122 yft1 mutation

The tomato mutant, n3122, produced characteristic yellow colored fruit and showed perturbed ripening. Using map-based cloning, the mutated locus was determined to be located within an 88.2 kb region that was predicted to contain 12 genes. Comparing the known functions of these genes to the phenotype of the n3122 fruits, we focused on Solyc09g007870, which encodes EIN2. Two mutations were found in the promoter region of EIN2, and its transcript levels were significantly lower in n3122 than in wild type M82.

The EIN2 gene (AtEIN2) was first cloned from Arabidopsis thaliana by map-based cloning using a loss-of-function mutant that showed complete ethylene insensitivity. AtEIN2 encodes a 1294 amino acid protein with 12 transmembrane helices in the N-terminus (residues 1–461) and a hydrophilic C-terminal domain (residues 462–1294) (Alonso et al. 1999). The residue 1262–1269 region corresponds to a nuclear localization signal (NLS), which is important for nuclear import of the C-terminal domain and the formation of the EIN2-ETR1 (ETHYLENE RESPONSE1) complex (Bisson and Groth 2011, 2015). EIN2 is a key component of the ethylene signal transduction pathway as it links ethylene perception in the endoplasmic reticulum (ER) with the response in the nucleus (Wang and Wen 2015). It has been demonstrated that in A. thaliana EIN2 acts downstream of the ethylene receptors and the Raf-like protein kinase, CTR1 (CONSTITUTIVE TRIPLE RESPONSE1), and that it interacts with ETR1 and CTR1 in the ER to regulate ethylene signaling (Bisson et al. 2009). In the absence of ethylene, the receptors activate CTR1, which directly phosphorylates EIN2 to inhibit an ethylene response. In the presence of ethylene, the C terminus of un-phosphorylated EIN2 is cleaved and translocated to the nucleus to activate the ethylene response via the transcription factors, EIN3/EIL1 (ETHYLENE INSENSITIVE LIKE1) (Ju et al. 2012; Qiao et al. 2012; Wen et al. 2012).

In addition to A. thaliana, EIN2 genes have been characterized from rice (Oryza sativa) (Jun et al. 2004), petunia (Petunia hybrida) (Shibuya et al. 2004), maize (Zea mays) (Gallie and Young 2004) and tomato (Wang et al. 2007). SlEIN2 from tomato was isolated by Wang et al. (2007) using the rapid amplification of cDNA ends method (RACE); however, we note that the published sequence from that study (Genbank No. XM_010327408), derived from the tomato cultivar Ailsa Craig, is different in several nucleotides from the sequence obtained in this present study. Zhu et al. (2006) used a TRV-based VIGS system to reduce EIN2 expression in tomato fruits, resulting in EIN2-silenced fruit with distinct regions that were either green and non-ripening, or soft and red. In addition, Hu et al. (2010) constructed a LeEIN2-silenced transgenic tomato line that had hard yellow fruit at the BR + 4 stage and seedlings with partially impaired ethylene sensitivity. In this current study, impaired fruit ripening, as evidenced by a yellow fruit color, and insensitivity to ethylene were also observed in n3122. These phenotypes, taken together with the results from Zhu et al. (2006) and Hu et al. 2010), support the hypothesis that EIN2 is the candidate gene responsible for the n3122 phenotypes.

yft1 regulates fruit coloration via ethylene action

As a climacteric berry fruit, tomato requires ethylene to ripen (Klee and Giovannoni 2011). The ripening deficient mutant Nr has a mutation in the ethylene receptor, which functions upstream of EIN2 in the ethylene signaling transduction pathway, and also has ripening-impaired fruits with a yellow/orange color (Lanahan et al. 1994; Wilkinson et al. 1995). Down-regulation of the tomato ethylene synthesis-associated genes, ACS4 (1-AMINOCYCLOPROPANE-1-CARBOXYLIC ACID SYNTHETASE 4), ACS2 and ACO1 (1-AMINOCYCLOPROPANE-1-CARBOXYLIC ACID OXIDASE 1), prevents fruit ripening (Hamilton et al.1990; Oeller et al. 1991; Picton et al. 1993), and the ripening-deficient mutants, rin (Vrebalov et al. 2002), cnr (Manning et al. 2006) and nor (Zhu et al. 2014), which all have hard yellow/orange mature fruits, have mutations in transcription factors that regulate fruit ripening at least partially via an ethylene dependent pathway (Karlova et al. 2014). Thus, blocking ethylene synthesis or perception in tomato results in ripening-deficient fruit with altered color. In this study, of the 12 genes located in the target region identified via map-based cloning, only EIN2 was annotated as being related to ethylene perception or synthesis. Based on the results reported in this study, we conclude that the yft1 is a new EIN2 allelic mutant that produces yellow fruit as a consequence of reduced ethylene biosynthesis.

Tomato fruit color in large part reflects the composition and content of carotenoid (Fray and Grierson 1993), with the carotenoid pigment, lycopene, being responsible for the red color. In fully ripe fruit of most red-fruited tomato varieties, lycopene contributes 70–90 % of the total carotenoid, while the orange pigment β-carotene accounts for 5–40 % (Ma et al. 2014). In this study, lycopene and β-carotene represented 83 and 16 % of the total carotenoid levels in M82 fruit, respectively. However, lycopene was not detectable in n3122 fruit and the total carotenoid levels were ~6 % of those in M82 fruit. The carotenoid profile of M82 was reported by Ronen et al. (2000), and showed, in agreement with our results, that the lycopene accounted for 90 %, followed by β-carotene (5 %) and then lutein. Lycopene is completely absent from the yellow-fruited tomato mutant r 2997 and the levels of total carotenoid, which were composed largely of β-carotene and lutein, were only ~4 % of those in the wild type fruit (Kachanovsky et al. (2012). Similarly, lycopene was also not detected in the yellow-fruited tomato line, PI114490 (Yuan et al. 2008), and we inferred that the yellow fruit color of n3122 reflects a decrease in carotenoid accumulation, and especially of lycopene.

Plastids play a central role in the synthesis and storage of carotenoid. Harris and Spurr (1969a, b) observed two carotenoid crystalloid substructures in normally ripening red tomato and the Beta mutant but not in the yellow-fruited r mutant. One was a high electron-dense crystalloid, and the other was an undulating membrane representing the crystalloids which were lost during dehydration. Schweiggert et al. (2011) also found carotenoid crystalloids in red papaya (Carica papaya) but not in yellow papaya, which contained only trace amount of lycopene, and similar results were obtained in studies of red- and white-fleshed loquat (Eriobotrya japonica) fruits (Fu et al. 2012). In this current study, both the highly dense crystalloids and the transparent undulating membranes were observed in M82 but not in n3122. The carotenoid profiles were consistent with the changes in chromoplast ultrastructure in both n3122 and wild type M82, and the chloroplast-to-chromoplast transition and carotenoid accumulation coincided with ethylene production in both M82 and n3122. This suggests that yft1 regulates fruit coloration via ethylene action, but how ethylene influences this biological process has yet to be determined.

Author contribution statement

LG and LXZ designed the research and wrote the manuscript; LG performed qRT-PCR, TME, HPLC and primary mapping; WHZ performed fine mapping; HOQ measured ethylene; WQS analyzed map-based cloning data.