Seedlings of tomato wild relative (S. habrochaites) LA1777 were grown under a natural light in a greenhouse with temperature regimes of 24–28°C and relative humidity of approximately 70–80%. Roots, stems, leaves, flowers and fruits were collected from these plants for RNA isolation.

T₂ generation of overexpression lines (OE7, OE8, OE9) and RNA interference (RNAi) line (Ri3), as well as the wild-type (S. lycopersicum cv. Ailsa Craig, AC), were tested for their tolerance to abiotic stresses. The abiotic stress treatments were imposed in the greenhouse. Seedlings were grown at 24–28°C and relative humidity of 70–80% of natural light. Four-week-old seedlings were used for stress tolerance assays.

The cold tolerant S. habrochaites LA1777 was used for the analysis of ATL78L expression. Four-week-old seedlings were transferred to a growth chamber at 25 ± 2°C and relative humidity of 70% with a photoperiod of 14 h light/10 h dark and 200 μmol m^-2s^-1light intensity. After 3 days, seedlings were used for abiotic stress and hormonal treatments, viz., drought, cold, salt, heat, wound, polyethylene glycol6000 (PEG6000), mannitol, MV, ETH, SA and IAA. For drought stress, seedlings of LA1777 were washed with tap water thoroughly to obliterate substrates, and then dehydrated on filter papers. Cold stress treatment was performed by transferring seedlings of LA1777 to a growth chamber at 4°C under the same light condition as described above. For salinity treatment, seedlings of LA1777 were irrigated once with 200 mM NaCl solution (200 mL per pot). For heat treatment, seedlings were subjected to a high temperature of 40°C. For wound treatment, the leaves of the seedlings were pricked once with a needle. For hormone treatments, the seedling leaves were sprayed once with either 30% PEG6000, 200 mM mannitol, 0.2 mM ETH, 0.1 mM SA, or 0.1 mM IAA. One hundred milliliters of hormone solution was used for three plants. The third fully expanded leaves from the growing point were collected at indicated time points after treatments. Samples were collected for RNA isolation.

The multiple sequence alignments were conducted using MUSCLE¹. The evolutionary history was inferred using the Neighbor-Joining method. The evolutionary distances were computed using the Poisson correction method with the units of the number of amino acid substitutions per site. All positions containing gaps and missing data were eliminated. Phylogenetic analyses were conducted in MEGA6 (Tamura et al., 2013). The sequences of the proteins were downloaded from EnsemblPlants².

The full-length cDNA, gDNA and promoter sequences of ATL78L were obtained from The Sol Genomics Network³, and the gene identifier of ATL78L is Solyc11g005280. The primer sequences used in this study were listed in Supplementary Table S1. The homologous protein sequences of ShATL78-like were found from the NCBI⁴ and aligned using the DNAman software 5.2.2 (LynnonBiosoft). The PLACE database (Higo et al., 1999) was used to identify potential motifs in the ShATL78L promoter sequence.

Total RNA was isolated using the TRIzol reagent (Invitrogen, USA). To remove genomic DNA contamination, the RNA was treated with DNase I at 37°C for 30 min. First strand cDNA was synthesized using the HiScript II 1st cDNA Synthesis Kit (Vazyme, China) according to the manufacturer’s protocol.

The expression of ATL78L under various treatments and in transgenic lines was analyzed using qRT-PCR, which was performed using SYBR^® Premix Ex Taq^TM (TaKaRa, China) and the LightCycler480 System (Roche, Switzerland) according to the supplier’s protocol. The PCR program was as follows: 95°C for 30 s; 40 cycles of 95°C for 5 s and 60°C for 25 s. Tomato actin was used as the internal control. All qPCR primers were designed by Primer5 and the primer sequences are listed in Supplementary Table S1. The qRT-PCR data of each gene were calculated using the 2^-ΔΔCT method (Livak and Schmittgen, 2001). qRT-PCR analysis was carried out with three technical replicates.

The full-length open reading frame (ORF) of ShATL78L was amplified from the cDNA of LA1777 using the primers of ATL78L (Supplementary Table S1) designed according to the S. lycopersicum full-length cDNA. For ATL78L overexpression construct, the ORF of ATL78L was incorporated into the pDONR221 vector using the Clonase BP reaction (Invitrogen), and then the recombinant plasmid was incorporated into the pMV3 vector (modified from pHELLSGATE2) by the Clonase LR reaction (Invitrogen). For RNAi construct, a 135-bp fragment of ATL78L was amplified by ATL78L-Ri primers (Supplementary Table S1), and then incorporated into the pHGRV vector using the Clonase BP reaction (Invitrogen). All the expression vector constructs were transformed into Agrobacterium strain C58 by electroporation. Cotyledons of S. lycopersicum were used for transformation, as described previously (Ouyang et al., 2005).

Tomato T₀ generation was checked by PCR using CaMV 35S promoter primer and ATL78L specific reverse primer (Supplementary Table S1). The expression of ATL78L in transgenic lines was analyzed by qRT-PCR. Three independent homozygous overexpression lines (OE7, OE8, and OE9) and one RNAi line Ri3 of ShATL78L were selected for further analyses.

For cold stress tolerance experiments, 4-week-old uniform-sized seedlings of T₂ transgenic lines OE7, OE8, OE9, Ri3 and the wild-type (AC) were randomly chosen and grown in growth chamber with conditions of 4°C, 70% relative humidity and a 14 hr light/8 hr dark photoperiod (200 μmol m^-2s^-1 light intensity). The same conditions except at 25°C were used for control seedlings. After 7 days of cold treatments and 1 day of recovery, the third fully expanded leaves from the top were sampled.

For drought stress tolerance experiments, seedlings from the same stage of T₂ transgenic lines (OE7, OE8, OE9, Ri3) and wild-type (AC) were placed in a greenhouse at 25°C under natural light conditions. Seedlings with drought treatment were withheld from water for 8 days, while the control seedlings were watered every 2 days. After 8 days of treatments, the third fully expanded leaves from the top of the plants were collected.

For salt and oxidative stress treatments, 4-week-old seedling of T₂ transgenic lines OE7, OE8, OE9, Ri3 and wild-type AC were placed in a greenhouse at 25°C under natural light conditions, and the third fully expanded leaves from the top of the plants were collected. Fifteen leaf disks (0.8 cm) were detached from the third fully expanded leaves of each line and incubated in different concentrations of 15 ml solution with 0 mM NaCl, 100 mM NaCl and 5 μM MV for 72 h at 25°C, 70% relative humidity and a 16 h light/8 h dark photoperiod (200 μmol m^-2s^-1 light intensity) conditions before the Chl contents were measured.

For the water loss assay, the third fully expanded leaves from the top of the control and stressed plants were detached and used for relative dehydration rate determination. The leaves were placed on a filter paper with conditions of 25°C and 60% relative humidity and weighed after 0, 1, 2, 3, 6, 9, and 18 h. Ten compound leaves of each line were weighed. The relative drainage was calculated according to the formula: the relative drainage (%) = 100 × (W₀ - W_t)/W₀, where W₀ and W_t are weights after 0 h and t h, respectively.

The ORF of ShATL78L was amplified using specific primers ATL78L-BD (Supplementary Table S1) containing NdeI/PstI restriction sites, thus corresponding to the yeast expression vector pGBKT7 (Clontech, USA), to construct pGBKT7-ShATL78L. The pGADT7-CSN5B was combined by the homologous recombination method (ClonExpress^TM II One Step Cloning Kit, Vazyme, China). According to the manufacturer’s protocol, the plasmids of pGBKT7-ShATL78L+pGADT7-CSN5B, pGBKT7-Lam+pGADT7-RecT (negative control), and pGBKT7-53+pGADT7-RecT (positive control) were co-transformed into yeast Saccharomyces cerevisiae strain AH109. The positive and negative controls were provided with the BD Matchmaker library construction and screening kits (Clontech, USA). The transformants were tested on the SD/-Leu/-Trp and the SD/-Ade/-His/-Leu/-Trp media. CSN5B accession numbers in the SGN database are Solyc06g073150 and Solyc11g017300.

For the yeast one-hybrid assay, the promoter segments of ShATL78L were PCR amplified using the primers ATL78L-P (Supplementary Table S1) containing EcoRI/MluI restriction sites, so that the segments could be inserted into pHIS2 (Clontech, USA). Same as pGADT7-CSN5B, pGADT7-RAV2 was also combined by homologous recombination. Plasmids pHIS2-ATL78LP+pGADT7-RAV2, p53HIS2+pGAD-Rec2-53 (positive control) and pHIS2-ATL78LP+pGADT7 (negative control) were co-transferred into yeast strain Y187, and the binding was detected on SD/-Leu/-Trp/-His selective plates containing 90 mM 3-AT (3-amino-1,2,4,-triazole) according to the manufacturer’s instruction. SlRAV2 accession number in the SGN database is Solyc05g009790.

Statistical analysis was carried out using SigmaPlot 10, Excel, and SPSS. A significant difference between pairs of groups was analyzed by Student’s t-test.

In our previous study, transcriptome profiles of cold tolerant (S. habrochaites LA1777 and its introgression line LA3969 under the genetic background of LA4024) and sensitive (S. lycopersicum LA4024) tomato genotypes were compared using microarray analysis, and one RING finger gene, ATL78L, was found to be more strongly induced by cold stress in the two tolerant genotypes than in the sensitive one (Liu et al., 2012). The complete ORFs of ATL78L were isolated from tomato Ailsa Craig (S. lycopersicum) and wild species LA1777 (S. habrochaites), as well as the wild species LA0716 (S. pennellii) that displays extremely high stress tolerance (Rick and Tanksley, 1981). The ORFs of ShATL78L and SpATL78L had the same size, consisting of 624 nucleotides and encoding 207 amino acid residues. Multiple alignments of amino acid sequences revealed that these three sequences showed a high degree of similarity to each other. Three amino acids at positions 20, 111, and 185 were different and one amino acid indel was found among the three genotypes (Figure 1). These four amino acids may functionally contribute to tomato stress tolerance.

ATL78L contains a RING finger domain (Cys-X2-Cys-X(9–39)-Cys-X (1–3)-His-X (2–3)-His-X2-Cys-X(4–48)-Cys-X2-Cys) and is classified as a RING-H2 finger protein (Figure 1). Phylogenetic analysis indicated that SlATL78L had nine paralogs in the cultivated tomato (S. lycopersicum) and showed the highest homology with PGSC0003DMT400015477 from potato (S. tuberosum). The closest homologue of SlATL78L from Arabidopsis is AT1G49230.1 (ATL78) therefore, Solyc11g005280 was named as ATL78L (Supplementary Figure S1).

To confirm the microarray results, expression levels of ATL78L in LA1777 and AC were examined under cold stress treatment. cDNAs of ShATL78L and SlATL78L were isolated to perform the expression analysis. As shown in Figure 2A, ShATL78L transcripts exhibited a significant increase after 3 h of cold stress treatment. However, no significant difference was observed in the relative expression level of SlATL78L between cold-treated and control plants. Therefore, there is a positive correlation between ShATL78L transcripts and cold tolerance in tomato. To further understand the role of the ShATL78L in different abiotic stresses, we investigated the expression pattern of ShATL78L in LA1777 under different abiotic stresses and hormonal treatments. As shown in Figure 2C, the transcript levels of ShATL78L were increased in response to drought, salinity, heat, wound, mannitol, PEG, ETH, SA, and IAA treatments. ShATL78L was strongly induced by drought, mannitol, salinity, heat, wound, and IAA treatments. A statistically similar but lower induction level was detected in response to SA, ETH, and PEG. To investigate whether the expression of ShATL78L has any tissue/organ specificity, the expression level of ShATL78L in different tissues of LA1777 was determined by qRT-PCR. As shown in Figure 2B, ShATL78L was constitutively expressed in all examined tissues with a little lower expression levels in young leaves and stem epidermis compared with other tissues.

In order to characterize the function of ShATL78L, the full-length ORF of ShATL78L was cloned into the plant expression vector pMV3 under the control of CaMV35S promoter. This construct was introduced into AC (S. lycopersicum). A total of 28 independent transformants (T₀) were generated, and 23 of them were positive as confirmed by PCR analysis using CaMV35S forward and gene-specific reverse primer sets. At the same time, the RNAi lines were also generated through knocking down the ATL78L transcripts. Transgene expression level was examined by qRT-PCR. As shown in Figure 3B, four independent lines, OE-7, OE-8, OE-9, and Ri-3 displayed expected expression levels of ATL78L.

To characterize the functional role of ShATL78L in cold stress tolerance, 4-week-old seedlings of the four transgenic lines (T₂) and the wild-type AC were examined under cold stress treatment at 4°C for 7 days, and then allowed to recover at 25°C for 1 day. After the stress treatment, seedlings of the AC line exhibited severe wilting, whereas the overexpression lines (OE-7, OE-8, OE-9) displayed much less leaf wilting symptoms, and the RNAi line Ri-3 showed higher leaf wilting symptoms compared to AC (Figure 3A). These findings demonstrated that ShATL78L plays a positive role in cold stress tolerance.

To investigate whether ShATL78L has an effect on drought stress tolerance, 4-week-old seedlings of the three overexpression lines (OE7, OE8, and OE9), the RNAi line (Ri-3) and the wild-type AC were withheld from water for eight consecutive days. After the drought stress treatment, obvious dehydration symptoms (leaf wilting) were found both in the RNAi line and wild-type plants compared to overexpression lines, which exhibited inconspicuous wilting. The stem of wild-type and RNAi seedlings bent due to loss of turgor pressure while the overexpression plants remained turgid and erect (Figure 3C). Water loss assay was conducted under the dehydration stress. The detached leaf water loss was significantly lower in the overexpression lines compared to wild-type plants while RNAi line showed the maximum water loss among all the treated lines (Figure 3D). These findings demonstrated that ShATL78L plays a positive role in drought stress tolerance.

The potential role of ShATL78L under salt and oxidative stress was further evaluated by incubating the leaf disks of 4-week-old seedlings in 100 mM NaCl and 5 μM MV solutions, respectively. After 72 h treatment, the leaf disks of the transgenic plants were similar to that of the wild-type AC in H₂O and NaCl solutions. In the MV solutions, much more severe bleaching was observed in the RNAi line than overexpression plants, and wild-type plants showed an intermediate phenotype (Figure 3E). The Chl contents further confirmed the difference of salt and oxidative damages between transgenic and wild-type AC plants (Figure 3F). Taken together, these results supported that ShATL78L plays a positive role in tolerance to oxidative stress.

The induction of ShATL78L gene under various abiotic stresses prompted us to analyze the promoter of this gene (approximately 1000 bp upstream the transcription start site). The promoter sequence was isolated from LA1777 and fused with pHIS2 to construct the bait vector (pHIS2-ATL78LP). As shown in Figure 4B, 90 mM inhibitory 3-amino-1,2,4-triazole (3-AT; Sigma) was screened from 0 to 90 mM concentrations of 3-AT for the yeast one-hybrid system. Bait constructs and tomato cDNA library fusion plasmid were co-transformed into yeast strain Y187 and a positive clone of SlRAV2 was selected from 152 clones for further confirmation (Supplementary Table S2). As shown in Figure 4D, all yeast cells harboring different constructs could grow on SD/-Trp/-His/-Leu without 3-AT, which indicated that all the recombinant plasmids were introduced successfully into yeast. However, co-transformed cells with pGADT7-SlRAV2 and pHIS2-ATL78LP could grow in the presence of the 90 mM 3-AT. Moreover, the growth of transformants containing constructs but lacking SlRAV2 was completely inhibited (Figure 4E), which suggested that SlRAV2 can bind to the promoter of ATL78L in yeast.

To further investigate the molecular mechanism of the ATL78L-mediated pathway in tomato stress tolerance, a yeast two-hybrid screening was performed to identify the interacting proteins with ATL78L. As shown in Figure 4A, ATL78L does not contain self-activation due to that the yeast cells pGBKT7-ShATL78L could not grow in the SD/-His/-Trp or SD/-Ade/-Trp media while could grow in the SD/-Trp medium. In the Y2H assay, the full-length ORF of ShATL78L was used as the bait to screen a tomato cDNA Y2H library, which was constructed with the prey vector pGADT7-Rec. Two positive clones of COP9 signalosome (CSN) complex subunit CSN5B were selected from 138 clones for further analysis (Supplementary Table S3). The amino acid sequences of these two genes showed high homology (Supplementary Figure S2). As shown in Figure 4C, the yeast cells co-transformed with pGBKT7-ShATL78L and pGADT7-CSN5B could grow in the SD/-Ade/-His/-Leu/-Trp media, same as the positive control (pGBKT7-53+pGADT7-RecT). In addition, negative control transformants (pGBKT7-Lam+pGADT7-RecT) did not grow in the same media. These results indicated that ShATL78L interacts with SlCSN5B in yeast.

RING finger family proteins have been widely reported to play an important role in abiotic and biotic stresses (Lee et al., 2001; Cheung et al., 2007; Hong et al., 2007; Qin et al., 2008; Zhang et al., 2015). According to the motifs, ATL78L can be classified into RING-H2, a subgroup of RING finger proteins (Borden and Freemont, 1996). RING-H2 finger proteins are involved in different abiotic stress signaling pathways in plants (Ko et al., 2006; Zhang et al., 2007). However, the functions of RING-H2 proteins in tomato remain largely unknown (Hondo et al., 2007; Qi et al., 2016). Particularly, the molecular mechanism of RING-H2 finger proteins in response to stress tolerance is still unclear. ShATL78L has not been functionally characterized, though it was screened in our previous microarray experiment under cold stress (Liu et al., 2012). In this study, we showed that ShATL78L exhibited a strong response to cold, drought, salt, heat and wound stresses (Figure 2C).

It has been well reported that RING finger proteins can modulate various plant hormone responses, such as ABA (Ko et al., 2006; Zhang et al., 2007; Li H. et al., 2011), IAA (Gray et al., 2002; Xie et al., 2002), GA (Park et al., 2013), ETH (He et al., 2005), JA (Hondo et al., 2007), and SA (Cheong et al., 2009). Phytohormones play critical roles in plant adaptation to adverse environmental conditions. Cross-talk between different plant hormones and the environment results in synergistic or antagonist interactions that play a significant role in plant responses to abiotic stresses (Peleg and Blumwald, 2011). Our results indicated that ShATL78L was induced by different plant hormone treatments such as ETH, IAA, and SA (Figure 2C), which indicates that ShATL78L possibly functions as a balance regulator through mediating different signaling pathways between ETH, IAA, and SA in response to abiotic stresses.

In this study, we demonstrated that ShATL78L overexpression plants positively contributed to enhanced cold, drought, and oxidative stress tolerance (Figures 3A,C,E). These results are consistent with previous studies in other plant species (Hirayama and Shinozaki, 2010). On the other hand, suppression of the RING E3 ubiquitin ligase AtATL78 increased the cold stress tolerance, while decreased drought stress tolerance (Kim and Kim, 2013). Arabidopsis RING-H2 gene XERICO confers drought stress tolerance through increased ABA biosynthesis (Ko et al., 2006). In S. pimpinellifolium, the RING Finger E3 Ligase SpRing is a positive regulator of salt stress signaling (Qi et al., 2016). However, ShATL78L transgenic plants did not show significantly different response to salt stress compared to the wild type (Figure 3E).

In this study, we demonstrate that RAV2 binds to the promoter of ShATL78L and activates/alters ShATL78L transcription (Figures 4D,E). The RAV transcription factor contains AP2-like and B3-like domains, which specifically bind to DNA with bipartite sequence motifs of RAV1-A (CAACA) and RAV1-B (CACCTG; Kagaya et al., 1999). The motif of RAV1-A and RAV1-B can be detected in the promoter of ShATL78L. It has been well reported that RAV2 responds to different abiotic stresses (Li C.W. et al., 2011; Fu et al., 2014; Matias-Hernandez et al., 2014; Mittal et al., 2014). The promoter of RAV2, which contains the CRT/DRE element, can be bound by the DREB transcription factors (Yamaguchi-Shinozaki and Shinozaki, 1994; Dubouzet et al., 2003; Li C.W. et al., 2011). DREB proteins specifically bind to DRE/CRT element and activate target gene transcription in Arabidopsis (Liu et al., 1998) and tomato (Li C.W. et al., 2011; Li et al., 2012). It has been well studied that DREB family play an important role in abiotic stresses in Arabidopsis, such as DREB (Liu et al., 1998) and CBF1 (Stockinger et al., 1997). In recent years, several studies have supported that DREB family genes exhibit similar role in abiotic stresses in tomato. For example, SlDREB gene acts as a positive regulator of drought stress responses by restricting leaf expansion and internode elongation (Li et al., 2012); The dehydration-responsive transcription factor SlDREB2 mediates salt stress tolerance in tomato and Arabidopsis (Hichri et al., 2016). Tomato has a complete CBF cold responsive pathway, however, its CBF regulon differs from Arabidopsis and shows much smaller and less diverse in function (Zhang et al., 2004). Our results and these previous findings clearly indicate the responsiveness of ShATL78L to multiple abiotic stresses (Figure 5).

COP9 (constitutive photomorphogenesis) is a component of the signaling complex mediating light control of development in Arabidopsis (Wei et al., 1994). CSN consists of eight subunits (CSN1-8), and plays a critical role in the ubiquitin/26S proteasome proteolytic regulatory pathway (Rubio et al., 2005). We show that ShATL78L interacts with CSN5B (Figure 4C), which plays a crucial role in abiotic stress tolerance (Wang et al., 2013). In rat, CSN5 inhibits cardiac L-type Ca²⁺ channel activity through protein–protein interactions (Kameda et al., 2006). Taken together, we hypothesize that ShATL78L responds to multiple abiotic stresses possibly through interacting with CSN5B, which alters the Ca²⁺ channel activity. On the other hand, CSN regulates plant growth and development through mediating protein degradation (Serino and Deng, 2003). In addition, CSN plays a vital role in mediating E3 ubiquitin ligase-mediated responses (Schwechheimer et al., 2001; Cavadini et al., 2016). Altogether, our results indicate that ShATL78L enhances multiple abiotic stresses tolerance possibly through altering the ubiquitin-mediated protein degradation (Figure 5).