Tomato contrasting genotypes responses under combined salinity and viral stresses

Abstract

Tomato yellow leaf curl disease (TYLCD) and salinity stress adversely affect tomato production worldwide by causing extensive damages. In Tunisia, identifying TYLCD resistant cultivars selected in different environments is useful to devise counter-measures. To this end, 20 tomato commercial cultivars were screened for different Ty gene alleles’ combinations and evaluated either for TYLCD incidence or salinity constraint. We built a biological multi-layer network for integrating, visualizing and modelling generated data. It is a simple representation view linking allelic combinations to tomato cultivars behaviour under viral and salt stresses. In addition, we analyzed differential expression of transcriptions factors (TFs) belonging to WRKY and ERF families in selected resistant (R) and susceptible (S) tomato cultivars. Gene expression was evaluated for short- and long stress exposure to either TYLCSV infection or to both viral and salinity stresses. Evidence is that TFs promote resistance to abiotic and biotic stresses through a complex regulatory network.

Electronic supplementary material

The online version of this article (10.1007/s12298-020-00835-w) contains supplementary material, which is available to authorized users.

Introduction

Tomato yellow leaf curl disease (TYLCD) is a major biotic stress on Tunisian tomato crops threatening the growth and yield of cultivars. Identified firstly in Israel in 1930, the tomato yellow leaf curl virus (TYLCV) (genus Begomovirus), has progressively invaded Mediterranean areas, Asian countries, and the US (Akad et al. 2004). In Tunisia, the tomato leaf curl sardinia virus (TYLCSV) was reported to be prevalent on tomato crops (Gorsane et al. 2004; Gharsallah Chouchane et al. 2006, 2007).

One of the strategy to manage TYLCD is to introgress resistance traits that were identified in wild tomato species to the domesticated tomato, known to be vulnerable to virus infection (Ornubol Chomdej and Chunwongse 2012). So far, tomato genome mapping analysis has led to the identification of six different TYLCD resistant/tolerances loci. Ty-1 and Ty-3 are allelic and are mapped on the long arm of chromosome 6. These alleles code for DFDGD-class RNA-dependant RNA polymerases (Verlaan et al. 2013). Ty-3 is known to be a codominant-indel based marker (Nevame et al. 2018). Ty-2 and Ty-4 are situated on the long arm of tomato chromosome 11 and 3, respectively (Ji et al. 2009; Yang et al. 2014). Ty-2 encodes an NB-LRR gene (Yang et al. 2014; Yamaguchi et al. 2018) whereas Ty-4 is reported to slightly modulate TYLCV resistance (Ji et al. 2009). Ty-5 based-marker was developed on the basis of SlNAC1 gene (Wang et al. 2018). It was mapped on chromosome 4, where only a single gene, pelota, is reported. The consequent loss of function is due to a T-to-G transversion in the coding region (Lapidot et al. 2015). Ty-6 is a novel resistance gene, located on chromosome 10 (Caro et al. 2015). It complements resistance conferred by Ty-3 and Ty-5 genes (Gill et al. 2019). Chromosomal fragments harboring resistance genes have been mapped using PCR-based polymorphic DNA markers (Nevame et al. 2018). Breeding programs to manage TYLCD resistance have focused on Ty-1, Ty-2, and Ty-3 introgression into domestic tomatoes. Even so, breakdown of Ty-2 mediated resistance has been reported for (TYLSCV), a Sardinian strain of TYLCV (Barbieri et al. 2010) and (TYLCV-Mld), a mild strain of TYLCV (Ohnishi et al. 2016). Besides, Ty-1 based resistance was thought to be easily overcame under TYLCD pressure (García-Cano et al. 2008).

Plant exposure to salt stress is increasing because of the high soil salinity and the poor trait of water irrigation. Therefore, such a stress is considered as one of the most devastating factors threatening crops. In Tunisia, excessive soil salinity is a major threat for tomato agricultural productivity. Most tomato cultivars are sensitive to salinity. Expression of stress-responsive candidate genes and changes in physiological parameters and enzymes activities were previously explored (Gharsallah et al. 2016a, b).

With regard to global climate changes, stress interactions between combinations of salinity and viruses are projected to become more prevalent in agricultural lands. Reports dealing with co-occurrence of abiotic and biotic stresses are limited (Kissoudis et al. 2016). They are often pointed to a negative impact of abiotic stress, mainly salinity on pathogen resistance (Suzuki et al. 2014). Plant physiological and molecular responses under combined abiotic and biotic constraints are significantly different from those displayed by a single stress (Ramegowda and Senthil-Kumar 2015; Zhang and Sonnewald 2017).

Researches have always focused on screening pathways involved in regulatory networks to seek sources of resistance. Plant innate immune systems, facing environmental threats, involve transcription factors (TFs) as WRKY Group III and ERF (Ethylene Response Factors). WRKY family, named from a highly conserved WRKY domain, comprises at least 81 members assigned into three groups (Groups I, II, and III) on the basis of the type of zinc-finger structures and the number of WRKY domains (Rushton et al.2010).

WRKY III TFs seem to be part of the enhancing defense mechanisms against pathogens (Nakayama et al. 2013; Gong et al. 2015), including basal defense and systemic acquired resistance (Rushton et al. 2010; Jiang et al. 2015). They are also part of different plant defense signalling pathways (Kalde et al. 2003) acting as both positive and negative regulators of disease resistance (Eulgem and Somssich 2007; Zhang et al. 2008). Furthermore some WRKY genes display a set of W or W-like boxes in their own promoters, suggesting possible interaction of these TFs either with each other (Diqiu and Yuping 2009; Tao et al. 2009) or with their own promoters for self-gene expression regulation (Pandey and Somssich 2009; Xiao et al. 2013).

Plant specific AP2/ERF family corresponds to TFs that regulate target gene expression by binding to a cis-acting promoter region element known as the CRT/DRE element, or GCC-box (Mizoi et al. 2012). It is a large gene family including 122 members in A. thaliana, 139 in rice (Nakano et al. 2006), 200 in poplar (Populus tricocarpa) (Zhang et al. 2008), 149 in grapevine (Licausi et al. 2010), and 121 in barley (Guo et al. 2016). They are involved in several developmental processes, ERF genes are candidate regulators of environmental stress responses (Singh et al. 2002; Mizoi et al. 2012; Licausi et al. 2013). Overexpression of ERF genes led to enhancement of the expression of several PATHOGENESIS-RELATED (PR) genes promoting plant resistance to pathogens as viruses (Singh et al. 2002; Fischer and Dröge-Laser 2004; Zuo et al. 2007). In a previous work, we have shown that SIWRKY and SIERF isoforms expressions are closely connected to tomato salt stress behavior (Gharsallah et al. 2016b). TFs isoforms displayed distinct patterns corresponding to either constitutive or up-regulated expression, particularly in tolerant tomato genotype. We found that SIWRKY8/31/39 were up regulated within the tolerant tomato genotype under salt stress. SIERF 9/16/80 were over expressed in tomato cultivars and transcripts increased significantly in the tolerant one. We wondered if these TFs are involved the same way when tomato plants are facing biotic and abiotic stresses at a time.

We initiated this current study to explore tomato commercial cultivars responses to the combined effect of TYLCSV infection and salinity stress. Biotic and abiotic stresses can trigger several physiological and molecular responses in plants. Biological multilayer network inferred from Cytoscape software provides association graphs allowing, firstly, integration and visualization of SCAR (Ty2/3) and CAPS (Ty1/4/5) marker genotyping and phenotyping of tomato cultivars threatened with TYLCSV infection. Secondly, connections of cultivars to their behavior under salt stress were performed. Pathways facilitated genotype–phenotype associations and underlined sources of TYLCSV resistance. Evidence is that TYLCD resistance is mediated by a combination of Ty1, Ty3 and Ty5 genes. Integrating data sets allowed to easily identify contrasting genotypes and to select (R) genotype showing resistance to TYLCSV and salinity stresses and (S) cultivar exhibiting susceptibility to both of them.

Till now, the crosstalk among salinity and TYLCD has not yet been explored. Thus, we investigated the expression of ERF and WRKY factors known to participate to multiple stress tolerance, within contrasting tomato genotypes previously selected. The implication of such TFs remains unclear even though they are thought to participate to plant biological processes during biotic and abiotic stresses. Results support that conventional breeding strategies may be a feasible alternative to establish resistance to both stresses in tomato. Indeed, we evaluated the responses of resistant (R) and susceptible (S) genotypes to investigate whether salinity interferes with TYLCSV accumulation and spread.

Materials and methods

Plant material

The plant material used in the present investigation included twenty tomato genotypes commonly cultivated by Tunisian growers. Plants were growing in an environmentally controlled chamber at 25 °C/18 °C, day/night and a 16-h light/8-h dark cycle with 40–50% relative humidity. Each following experiment was carried out with three replicates.

Molecular marker genotyping

Molecular markers used in this work are PCR-based markers and were amplified using specific primers (Table 1).

Table 1.

Molecular markers associated with tomato loci related to TYLCV

Target locus	Marker	Sequence of the primer 5′–3′	DNA-fragment size (bp)		References
			R	S
Ty1	CAPS (Taq I)	C1F: TAATCCGTCGTTACCTCTCCTT	300/300	400/400	Chen et al. (2012)
		C1R: CGGATGACTTCAATAGCAATGA
Ty2	SCAR	TG0302F: TGGCTCATCCTGAAGCTGATAGCGC	600/600	450/450	Garcia et al. (2007)
		TG0302R: AGTGTACATCCTTGCCATTGACT
Ty3	SCAR	FLUW25R: CCATATATAACCTCTGTTTCTATTTCGAC	640/480	480/480	Ji et al. (2007)
		FLUW25F: CAAGTGTGCATATACTTCATA(T/G)TCACC
Ty4	CAPS (Af II)	At5g60160 F: TTCTCGCGGCCTTTTCTCCTC	–	250/400	Ji et al. (2009)
		At5g60160 R: GTGATCGCAAACATATACTCGC
Ty5	CAPS (Taq I)	C2F: TGCCTGGTTTCTGCTGTCA	300/425	350/350	Wang et al. (2018)
		C2R: TAAAGCTGAAGAAGGACTTACCCT

R resistant, S susceptible

SCAR (Sequence characterized amplified region) markers are associated with Ty-2 and Ty-3 genes. CAPS (Cleaved amplified polymorphic sequence) markers are associated with Ty-1, Ty-4 and Ty-5 genes. Ty-1 and Ty-5 corresponding PCR products were digested with TaqI, while Ty-4 amplicon was digested with AfII. Total DNA was extracted from 0.2 g of mature leaves according to the DNAeasy Plant Mini Kit (Qiagen). DNA quantification was performed with ND-1000 spectrophotometer (Nanodrop Technologies, USA). PCR amplifications were performed in 25 µl reaction volume containing 25 ng of template DNA, 2.5 µl of PCR buffer (10×), 1.5 µl of MgCl₂ (25 mM), 1 µl of dNTPs (10 mM), 10 nM of each primer, 0.5 U of Taq DNA polymerase (5 U/µl). PCR amplifications were performed on a T professional trio system thermal cycler (Biometra, Germany) using the following cycling program: template DNA was first denatured at 94 °C for 3 min; followed by 35 cycles of 94 °C for 45 s, 53–55 °C for 45 s, and 72 °C for 2 min; followed finally by 7 min at 72 °C and a cooling to 4 °C. DNA fragments corresponding to PCR or cleaved products were separated by electrophoresis on 2% gel agarose and visualized with ethidium bromide.

Agro-inoculation of a Tunisian TYLCSV infectious clone

An infectious clone corresponding to a 1.7 mer of the Tunisian TYLCSV isolate (AY736854) clustered in the Sicily strain was cloned in the pCB301 (Gharsallah Chouchane et al. 2006). Agrobacterium tumefaciens GV3101 cells harboring the molecular construction were cultured in LB, pelleted and re-suspended in infiltration medium at an OD₆₀₀ of 0.5. Two-leaf-stage tomato cultivars were challenged by the infectious TYLCSV clone by pressure in the leaves with a needleless syringe. The efficiency of the agro-infiltration method was assessed by virus detection on treated leaves one week dpi using qRT-PCR for detection of the Coat-Protein (CP)-coding gene.

Evaluation of TYLCSV resistance

Based on responses to TYLCSV infection, the twenty commercial genotypes were clustered into different classes according to Friedmann et al. (1998): 0 = no visible symptoms, inoculated plants show same growth and development as controls; 1 = very slight yellowing of leaflet margins on apical leaf; 2 = some yellowing and minor curling of leaflet ends; 3 = a wide range of leaf yellowing, curling and cupping, with some reduction in size, yet plants continue to develop; and 4 = very severe plant stunting and yellowing, pronounced leaf cupping and curling, and plant growth stops. Symptoms of each plant were recorded 3 weeks dpi.

Evaluation of salt tolerance

We examined the effect of salt treatment on tomato genotypes according to the method developed by Gharsallah et al. (2016b). We used plants with four fully developed true leaves, individually transferred into plastic pots (30 cm of diameter) containing a mixture of peat and sand. Then, they were irrigated with one-half Hoagland solution supplemented with 150 mM NaCl (15 dS/m, pH 7.5). Salt treatment was initiated with 50 mM of NaCl solution (6 dS/m), increased to 100 mM (12 dS/m) on day two, and finally to 150 mM (15 dS/m) on day three.

We used three biological replicates for each of the 20 varieties. Each replicate consisted of a pool of 10 plants. A set of three plants for each genotype was grown in non-saline conditions and watered with the nutrient solution. Three weeks later, salt-treated plants were evaluated for salt tolerance, based on their visual phenotypes compared to control plants. Plants were rated for severity of salt susceptibility on a 1–5 scale (Dasgan et al. 2002).

Biological network

A network was generated by Cytoscape v3.0.2 (Cline et al. 2007) to set up connections between different Ty  genes and corresponding alleles with tomato cultivars displaying variable responses to TYLCSV infection. To perform enrichment of the network, we associated cultivars’ behavior under salt stress. The algorithm MCODE28 from the Cytoscape Plug in Cluster Maker was then used to partition the network into layers to view the dynamic behavior of cultivars in the context of TYLCSV and salt stresses and to facilitate the discrimination of alleles which contributes to virus tolerance. Interaction network of WRKY proteins was conducted by STRING software (Franceschini et al. 2012).

Viral DNA isolation and TYLCSV quantification

Total DNA was extracted from 0.2 g of mature leaves according to the DNAeasy Plant Mini Kit (Qiagen). DNA quantification was performed with ND-1000 spectrophotometer (Nanodrop Technologies, USA). Reaction included 100 ng of DNA sample as a template, 300 nmol/l of each primer (TYLCSV CP gene), and Igreen qPCR master Mix-Rox (BIOMATIK, USA) (Table 2). ABI A Prism 7000 sequence detection system (Applied Biosystems, USA) was used for quantitative real-time PCR (qPCR) under the following cycle conditions: 10 min at 95 °C followed by 40 cycles of 15 s at 95 °C, 1 min at 60 °C. The β actin was used as internal reference gene (Løvdal and Lillo 2009). Each PCR assay was run with a negative control corresponding to non-inoculated plants.

Table 2.

List of primers used for qRT-PCR analysis

Genes names	Primers 5′–3′	References
CP gene (TYLCSV)	F: GCTACGGATGTACAGAATGACAAAA	Gharsallah Chouchane et al. (2006)
	R: CAAAAGCTACGGATGTACAGAATGAC
SIWRKY 8	F: TAATTCTGCCGGAAAGCCTC	Huang et al. (2012)
	R: ATGCTTATTGCCGGTACTCGA
SIWRKY 31	F: ACAACCTATGAAGGGAAGCACA
	R: AGGGTGCTCCCATTTCAGAC
SIWRKY 39	F: GCGGTAATGCCAAGACAAAC
	R: TCAGTTCCTGGTGATTTACGC
SIERF 9	F: TGGAAGGAGTATAATGAGAAACTAGACAA	Sharma et al. (2010)
	R: CCT TCTTTGAACCTTTAGCAGGAA
SIERF 16	F: GCGAATAATACAGAACCCGAACTT
	R: TGAGGAAGAAGAAAGATCCGAATT
SIERF 80	F: TTTCAATCATGGTTGCTGCTTT
	R: AAGGGCGGCGACATACC
β actin	F: GAAATAGCATAAGATGGCAGACG	Løvdal and Lillo (2009)
	R: ATACCCACCATCACACCAGTAT

WRKY and ERF quantitative real-time PCR analysis

Two contrasting TYLCSV tolerant and sensitive genotypes were firstly TYLCSV inoculated and then subjected to salt stress. Tomato plants were divided into two groups. The first one corresponds to watered plants. The second set of plants was subjected to salt treatment. Each assay was carried at 4 weeks post-TYLCSV inoculation (p.i.) in three replicates. A set of uninfected and watered plants were used as a control.

Total RNA was isolated from inoculated or from upper leaf tissues using the TRIzol VR LS Reagent (Trizol RNA stabilization solution, Invitrogen; Life Technologies, Carlsbad, CA, USA) according to the manufacturer’s instructions. Total RNA was quantified by ND-1000 spectrophotometer (Nanodrop Technologies, USA). First-strand cDNA was synthesized from 2 µg of total RNA with oligo (dT) and MMLV reverse transcriptase (200 U/µl, Invitrogen) according to the manufacturer’s instructions. ABI A Prism 7000 sequence detection system (Applied Biosystems, USA) was used for quantitative real-time PCR (qPCR) under the following cycle conditions: 10 min at 95 °C followed by 40 cycles of 15 s at 95 °C, 1 min at 60 °C (Table 2). The β actin tomato gene was used as internal reference gene (Løvdal and Lillo 2009). Each PCR assay was run with a negative control corresponding to watering, non-inoculated plants.

PCR reactions were carried out in 96-well optical reaction plates (Applied Biosystems, USA). Reaction included 50 ng of cDNA sample as a template, 400 nM reverse primers, and Igreen qPCR master Mix-Rox (BIOMATIK, USA). Relative quantification was performed by applying the 2^−ΔΔCt method (Livak and Schmittgen 2001).

Statistical analysis

Analyses of gene expression were performed with DataAssist TM v3.0 Software (Applied Biosystems, USA). Data were analyzed using two-way ANOVAs with salt stress and TYLCSV infected varieties as the two predictor variables. Differences at Tukey’s test HSD P = 0.05 were considered statistically significant. Analyses were performed using GraphPad Software (version 6.0, CA, USA).

Results

SCAR and CAPS markers genotyping

To explore allele-specific polymorphism, tomato cultivars were screened through SCAR markers for Ty-2 and Ty-3. Upon PCR amplification, Ty-3 gave two distinct patterns corresponding to an heterozygous genotype 640/480 and to an homozygous genotype 480/480. Ty-2 showed a monomorphic profile with a unique genotype 450/450. Besides, CAPS markers were performed to screen Ty-1, Ty-4 and Ty-5. Ty-1 PCR-(Taq I) digestion created three different patterns. The first included an homozygous genotype 300/300, the second an heterozygous genotype > 300/400 and the third an homozygous genotype 400/400. Results for Ty-4 PCR-(AfII) digestion revealed a monorphic pattern corresponding to a 400/250 heterozygous genotype. Results for Ty-5 PCR-(TaqI) digestion produced two discernable patterns corresponding to an heterozygous genotype 300/425 and a homozygous genotype 350/350 (Supplementary files S1/2/3/4/5).

Phenotyping for TYLCSV resistance

TYLCSV infection of 20 tomato commercial cultivars was performed using Agrobacterium tumefaciens-mediated inoculation (Zhang et al. 2009) (Supplementary file S6). An infectious clone was previously engineered from a Tunisian TYLCSV strain and was attested to be infectious (Gharsallah Chouchane et al. 2006). To assess the level of resistance to TYLCSV, symptoms were meticulously scored three weeks pi (post inoculation). Evaluation for viral resistance showed that tomato cultivars are ranging from fully susceptible to highly resistant. Plants exhibited various symptoms of dwarfing with yellow and curly leaves allowing them to be assorted in the corresponding scale class according to Friedmann et al. (1998). 10% of the genotypes were clustered to the resistant scale-class 0; 5% to the scale-class 1; 25% to the scale class 3 and the majority, 60%, to the susceptible scale-class 4.

Phenotyping for salt stress tolerance

Twenty tomato cultivars were screened for salinity tolerance using an approach developed by Gharsallah et al. (2016b). Differences among the phenotypes were noticed 3 weeks post salt stress imposition. Salinity symptoms were meticulously recorded according to salt scale susceptibility (Dasgan et al. 2002) and allowed cultivars to be arranged in five scale classes as follows: 15% showed no pronounced symptoms and were assigned to the tolerant Class 1. Remaining cultivars were clustered to scale-class 2 (10%); 3 (15%); 4 (45%) and the sensitive class 5 (15%).

Biological association network

A multi-layer network inferred from the Cytoscape v3.0.2 software was generated to integrate and visualize relationships formed between screened loci, corresponding alleles, tomato scale-classes and phenotypes engendered by either virus or salt stresses. This is a necessary step toward a complete description of the dynamic behaviour of tomato cultivars facing either salinity constraint or TYLCSV infection. Considering the first set of the network, pathways permitted to identify discriminating genotypes corresponding to Ty-1 (300/300); Ty-3 (640/480) and Ty-5 (300/425) (Fig. 1). Each of these markers represents sources of TYLCSV resistance that might be tomato species-specific. The second set of the network provided the effect of salinity stress on tomato cultivars being assorted in salt scale-classes. Integrating data sets allowed to easily identify contrasting genotypes (Gharsallah et al. 2016a). Comparative analysis of two contrasting genotypes are essential to explore stress-responsive genes regulation at the transcriptional level. Based on this network, we selected two tomato genotypes that responded differentially to both viral and salinity stresses. They correspond to San Miguel cultivar (R) showing resistance/tolerance to TYLCSV and salinity stresses and Mouna HF1 (S) cultivar exhibiting susceptibility to both of them. Differential phenotypic responses of both (R) and (S) genotypes to a combination of salt stress and TYLCSV infection were recorded (Supplementary file S7).

Fig. 1.

Genotype–phenotype association network. Black circles represent Ty markers while purple nodes represent corresponding alleles. Phenotypes are associated to corresponding genotypes and are connected to related varieties following color-code branches (red, green and blue) (color figure online)

TYLCSV accumulation

We aimed to know whether tomato defenses against TYLCSV infection was influenced by salinity stress due to salinized water irrigation. The virus may enhance salt-stress response to ensure a successful long-term infection cycle. Thus, to appreciate the interplay between abiotic (salinity) and biotic (TYLCSV) stress, we used the two previously selected contrasting tomato genotypes. TYLCSV- infected tomato plants from R or S genotype were divided into two sets of 10 plants each. The first set corresponds to watered plants and the second one was submitted to salt stress. Young leaves were harvested every 7 days for 4 weeks. Regarding S genotype subjected to salt stress, the amount of TYLCSV continuously increases during infection, being 2–3 times greater than the rate observed when plants are grown in normal conditions (Fig. 2a). R genotype contained barely TYLCSV amounts in watered plants while exhibiting an increasing level of the virus, starting already at 2 weeks after the salt stress imposition (Fig. 2b).

Fig. 2.

QPCR estimation of relative TYLCSV amounts in S genotype (a) and R genotype (b). C: TYLCSV inoculated and watered plants; SS: TYLCSV inoculated plants and salt stressed. Error bars show the standard error between three replicates performed. Bars with different letters within each panel are significantly different at P > 0.05 according to Tukey’s test

Loss of resistance is expressed by a development of viral symptoms typical of infected susceptible plants. Despite of this, the amount of the virus remained low when compared to the S genotype, during each stage of the experiment. Salinity stress led to a significant viral accumulation in tomato genotypes whether susceptible to the virus or not.

Analysis of differentially expressed ERF genes

We examined the patterns of SIERF isoforms expression in TYLCSV infected tomato R and S genotypes, 4 weeks pi, whether plants were submitted or not to salinity constraint. Two-way ANOVA followed by Tukey’s multiple comparisons test indicated that relative SIERF 16 and SIERF 9 expressions are statistically significant between genotypes under salt stress. Indeed, analysis of SIERF 16 expression showed a similar trend of transcript accumulation in S genotype with or without salt stress imposition while it was significantly upregulated in R genotype subjected either to TYLCSV infection or combined stresses (Fig. 3a). Compared to the control, salinity enhanced leaf expression level of SIERF 9 in both genotypes. Nevertheless, expression is twofold increased in S genotype compared to the R (Fig. 3b). Expression of SIERF 80 remained constant in the R genotype when subjected  to salinity. By contrast, expression was increased under viral infection and was threefold up regulated within S genotype under salt stress (Fig. 3c).

Fig. 3.

Relative gene expression of SIERF 16 (a), SIERF 9 (b) and SIERF 80 (c) in leaves of S and R tomato genotypes. C: uninfected and watered plants; TYLCSV inoculated and watered plants; SS: TYLCSV inoculated plants and salt stressed. Total RNA was purified from tissues of tomato plants inoculated by TYLCSV and treated or not with 150 mM NaCl for 4 weeks pi. Transcript level was analyzed by qRT-PCR using primers indicated in Table 2. Tomato actin gene was used as reference gene. Error bars show the standard error between three replicates performed. Bars with different letters within each panel are significantly different at P > 0.05 according to Tukey’s test

Analysis of differentially expressed WRKY genes

In response to salt stress, the expression of WRKY genes was differentially regulated since the different isoforms gave distinct patterns. Two-way ANOVA followed by Tukey’s multiple comparisons test indicated that relative SIWRKY 8/39/31 expressions are statistically significant between genotypes. Isoforms 8 and 39 were preferentially expressed in R genotype subjected to TYLCSV infection. Indeed, SIWRKY 8 expression was enhanced several folds following TYLCSV infection, before being reduced under salt stress imposition (Fig. 4a). SIWRKY 39 transcipts were accumulated only within R genotype under TYLCSV infection and was reduced under combined stresses (Fig. 4b). Conversely, SIWRKY3 1 transcripts levels remained equal in S genotype in both normal and saline conditions while transcript accumulation was slightly but significantly up regulated in viral stress conditions within R genotype (Fig. 4c).

Fig. 4.

Relative gene expression of SIWRKY 8 (a), SIWRKY 39 (b) and SIWRKY 31 (c) in leaves of S and R tomato genotypes. C: uninfected and watered plants; TYLCSV inoculated and watered plants; SS: TYLCSV inoculated plants and salt stressed. Total RNA was purified from tissues of tomato plants inoculated by TYLCSV and treated or not with 150 mM NaCl for 4 weeks pi. Transcript level was analyzed by qRT-PCR using primers indicated in Table 2. Tomato actin gene was used as reference gene. Error bars show the standard error between three replicates performed. Bars with different letters within each panel are significantly different at P > 0.05 according to Tukey’s test

Interaction network of tomato WRKY proteins

Interaction network of WRKY TFs was built to give insights into crosstalk between proteins regulation pathways, experimentally determined or retrieved from data bases (Fig. 5). According to this network, SIWRKY 39, SIWRKY 8 and SIWRKY 33 seem to be related to each other. In addition, SIWRKY 39, SIWRKY 8 interact with MPK3 (mitogen-activated protein kinase), a positive regulator of plant defense response while SIWRKY 31 is co-expressed with MAPK7. Co-expressions were also identified between NAC1 and SIWRKY 31 and SIWRKY 39. NAC is a factor-based TY-5 marker. It is worth noticing that SIWRKY 39 and SIWRKY 8 seem to interact with CNGIC gene. CNGIC is a Cyclic nucleotide-gated ion channel 1-like gene which may be responsible for cAMP-induced calcium entry in cells, being part of the signal transduction involved in plant stress responses. SIERF participates in different functional pathways by interacting with other proteins. SIERF 16 is connected to SIWRKY 31 and MAPK3. SIERF 80 is linked to HSFA3 (Heat Stress Factor A3) while SIERF 9 is co-expressed with this Heat Stress Factor.

Fig. 5.

Interaction network of six SiWRKY TFs in response to abiotic and biotic stress. NP98A Nuclear pore complex protein NUP98A, MYC2 transcription factor MYC2-like, WRKY Transcription Factors (SIWRKY 8, SIWRKY 31, SIWRKY 33, SIWRKY 39, SIWRKY 70), ERF ethylene transcription factor (SIERF 1, SIERF 9, SIERF 16, SIERF 80), CNGIC Cyclic nucleotide-gated ion channel 1-like, MYB 44 MYB transcription factor, MPK mitogen-activated protein kinase (MPK3, MAPK), MCSC magnesium-chelatase subunit ChlH, MKS1 MAP kinase substrate 1, VQ motif protein MKS1-like, ACO4 amino cyclo propane carboxylate oxidase, Cyto450 cytochrome P450 71A9-like, ERD4 early-responsive to dehydration stress protein (ERD4), ERDs family early-responsive to dehydration stress protein, NAC1 NAC domain protein 1, HSFA3 heat stress transcription factor A3, Solyc11g039880.1.1: Annotation not available, Solyc02g077820.1.1: Annotation not available

Discussion

In the field, plants are often exposed to unfavourable environmental changes including incessantly and simultaneously combination of biotic and abiotic stresses that can impart adverse issues on their growth and development (Suzuki et al. 2014). Salinity decreased plant biomass and size leading to yield decline. In Tunisia, 25% of irrigated areas are afflicted by salinity (Bouksila et al. 2013). The plant defense responses involve interconnecting signalling pathways (Dong 2001). With regard to symptomatology, tomato commercial cultivars were salt stressed and phenotypic assortment allowed them to be classified into five scale classes (Dasgan et al. 2002). Two contrasting cultivars were selected from a salt-tolerant scale class 1 and a salt-sensitive scale class 5. Based on visual appearance and differences in sensitivity towards salinity; we have previously assigned these local cultivars into the same classes (Gharsallah et al. 2016b). Salt stress is a complex process involving many candidate genes related to biological and molecular pathways. The question is whether salinity modulated plant responses when a biotic stress occurs.

Tomato domestication through selection for valuable agricultural features as yield and fruit quality has resulted in the loss of part of gene alleles conferring resistance to viral infection. Tomato hosts more than 200 species of pathogens, some of which are controlled by R genes derived from wild accessions (Nevame et al. 2018). The molecular mechanisms underlying natural resistance to TYLCD are still unknown (Gorovits and Czosnek 2007). Mediated resistance genes, Ty-1, Ty-3 and Ty-4, have been introgressed from S. chilense accessions LA1969 and LA2779, LA1932 respectively (Zamir et al. 1994; Ji et al. 2007, 2009). Ty-6, the newest identified locus, is also originated from S. chilense, accessions LA1938 and LA2779 (Gill et al. 2019). Ty-2 was identified from S. habrochiates, accession B6013 and Ty-5 from a complex of accessions belonging to S. Peruvianum (Vidavski 2007; de Castro et al. 2007; Caro et al. 2015). Hereafter, polymorphic DNA markers were developed to tag resistance genes (Ji et al. 2007). We have scanned tomato varieties genome through SCAR markers to target Ty-2 and Ty-3 genes and CAPS markers for Ty-1, Ty-4 and Ty-5 genes. All screened genotypes exhibited monomorphic allelic patterns for both Ty-2 (450/450) and Ty-4 (400/250) suggesting that theses loci are probably not supporting TYLCV resistance. Resistance/tolerance in tomato genotypes is likely performed by genotypes 300/300 and > 300/400 (Ty-1), 640/480 (Ty-3) and 425/300 (Ty5). It is worthwhile to notice that Ty-1, Ty-2 are considered as major loci currently exploited for TYLCV resistant. Conversely, the use of Ty-5, considered as recessive, is limited in breeding programs (Bai et al. 2018). Otherwise, it has been suggested that Ty-1, Ty-3 and Ty-5 genes control broad-spectrum resistance mechanisms (Verlaan et al. 2013; Lapidot et al. 2015).

Correlations of genotype–phenotype deployed by the Cytoscape network allowed an overall view of marker-genotyping of tomato cultivars with regard to their behaviour under TYLCSV infection. This network permits to underline that genotypes tested here carry specific and discriminating alleles belonging to Ty-1, Ty-3 and Ty-5 loci.

Abiotic stress occurring in fields affects plant response and their resistance to plant pathogens (Suzuki et al. 2014; Kissoudis et al. 2017). For example, tomato resistance to powdery mildew caused by Oidium neolycopercisi may not be stable under salinity stress (Bai et al. 2018). Therefore, we focused on the stability of sources of resistances to TYLCSV infection under salt stress conditions within two contrasting varieties. Responses of these tomato genotypes to combined salinity and viral stresses were significantly different. Under salt stress, tomato sensitive genotype defense seemed to be weaker and its sensitivity against TYLCSV was significantly increased. Salt stress also attenuated TYLCSV resistance mediated by Ty-1, Ty-3 and Ty-5 loci. Salinity treatment led to an increase of TYLCSV accumulation in R genotype. It is evident that TYLCSV tomato-resistance was less effective under salinity factors. In spite of that, the viral amount exhibited by the resistant genotype was lower than what was recorded in a sensitive genotype. TYLCSV tomato-response was modulated by salinity in a genotype-dependent manner. In line with these findings, elevated temperatures interfere with plant virus accumulation. Both resistant and susceptible tomato plants displayed increased TYLCV amounts when faced with heat stress (Anfoka et al. 2016). With regard to fungal potato disease, salinity influences mycelial growth, sporulation and spore germination (Mills et al. 2004). This is likely due to Na⁺ and Cl⁻ toxicity effect on the fungus (Kissoudis et al. 2015, 2016). Tomato and powdery mildew interaction was explored under salinity conditions (Kissoudis et al. 2016, 2017). Salt stress has a negative impact only on incomplete resistance but not on mediated complete resistance (Bai et al. 2018).

The Biological network outlines that contrasting salt-tolerant and salt-sensitive varieties are clustered into the same type of scale-classes with regard to TYLCSV infection. This permits to speculate that plant responses to both types of stresses consist of broad and spatially coordinated signaling networks with an intense cross-talk between them.

In response to abiotic and biotic stresses, plants are able to modulate gene transcription through a regulated and dynamic regulatory network. Because of their critical roles in biotic and abiotic stress responses (Huang et al. 2016; Bai et al. 2018), we focused on WRKY and ERF isoforms. a time. In a previous work (Ghrasallah et al. 2016b), we reported that SIWRKY8/31 displayed constitutive expression within the salt tolerant genotype while SIWRKY39 was highly induced at the end stage of the stress imposition. Concomitantly, SIERF9/16/80 genes showed transcripts abundance mainly in the tolerant tomato genotype. TFs are currently explored under a single stress, therefore, we examined SIWRKY8/31/39 and SIERF9/16/80 regulation in tomato cultivars under salt and virus challenge.

Our data showed that SIWRKY 8 was upregulated within tomato R genotype following TYLCSV infection while transcripts levels remained equal in S genotype in both normal and saline conditions. In Nicotiana benthamiana, NtWRKY 8 was reported to be phosphorylated by pathogen-responsive MAPKs, so that they can bind to RBOHB gene promoter leading to ROS burst (Ishihama and Yoshioka 2012; Adachi et al. 2015). This regulation is acting through mitogen-activated protein kinases (MAPKs) (Pandey and Somssich 2009; Ishihama and Yoshioka 2012) and was outlined in the tomato protein network. SIWRKY 39 and SIWRKY 8 are supposed to be targets of MAPK3. This protein kinase is involved in Ca²⁺ signaling pathway which modulates transcriptional machinery to regulate gene expression (Mehlmer et al. 2010). SIWRKY 31, SIWRKY 39 and SIERF 16 seem also to be phosphotylated by MPK3. These TFs are serving as potential targets of MAPK cascades to regulate related-genes expression to counter TYLCSV infection. In Arabidopsis, MAPK were reported as pathogenesis responsive proteins (Meng et al. 2013). This seems to be the case for SIWRKY 8 since it is co-expressed with MAPK3 according to a regulatory TFs network. Moreover, SIWRKY 39 also displayed an interaction with MAPK3 and was found to greatly express in R genotype under TYLCSV infection. This isoform is reported to be upregulated in tomatoes when challenged with P. Syringae (Huang et al. 2012). Thus, when transgenic lines over-expressed SlWRKY 39, they displayed increased resistance against this pathogen (Sun et al. 2015). SIWRKY 31 was upregulated by TYLCSV inoculation and down regulated following salt stress imposition. This isoform, also named SlDRW1, is proposed to act as positive regulators of plant responses to biotic stresses (Liu et al. 2014). Indeed, SlWRKY 31 is an activator of plant defense against several pathogens (Liu et al. 2014; Li et al. 2015) and to drought and/or salt stresses (Huang et al. 2012). In a previous work, the same tomato varieties were challenged by a salinity stress. SIWRKY 8 and SIWRKY 31 transcripts highly accumulated, especially within the tolerant genotype even in the absence of the salt stress. Conversely, SIWRKY 39 accumulation exhibited a significant increase in the tolerant genotype (Gharsallah et al. 2016b). It is relevant to notice that SIWRKY 31 and SIWRKY 39 are involved in regulation of NAC gene, encoding a mRNA surveillance factor (Pelo) involved in ribosome recycling phase of protein synthesis (Lapidot et al. 2015). Although no data gave evidence that pelota gene play a role in TYLCV resistance, viruses always require the host cell machinery to achieve their infection cycle.

WRKY genes are known to display constitutive or induced expression profiles, which are at crossroads of plant responses to environmental stresses. In Arabidopsis, the isoform WRKY 33 is regulated through the MAPK cascades that is activated by TYLCV infection (Gorovits et al. 2007; Meng and Zhang 2013). Once phosphorylated, SIWRKY 33 regulates its own expression and binds to camalexin genes leading to defensive metabolites production, effective against plant pathogens (Birkenbihl et al. 2012). It is likely that SIWRKY 8 and SIWRKY 39 undergo a regulatory network, in which MAPK3 is connected to these 2 isoforms and to SIWRKY 33 as well. Indeed, TYLCSV viral tolerance of tomato plants is mediated by a regulatory network where SIWRKY 39/31/8 are acting through resistance gene expression regulation or following the pattern of SIWRKY 33 expression.

WRKY factors are able to act as activators or repressors of candidate genes depending whether plants are subjected to individual or to a combination of two stresses. Regardless of the way they are down or upregulated, they seem to be part of a dynamic regulatory network (Bakshi and Oelmüller 2014; Phukan et al. 2016).

Besides WRKY, Ethylene response factors (ERFs) are stress-responsive transcription factor known to be differentially modulated under abiotic and biotic stresses (Bai et al. 2018; Debbarma et al. 2019). Indeed, Overexpression of BoERF1 from Brassica oleracea enhances tolerance to salinity and to the fungal Sclerotinia pathogen (Jiang et al. 2018). Among the studied isoforms, SIERF 16 was the only isoform up regulated when the R genotype was challenged with TYLCSV and salinity. The interaction network between TFs showed that SIERF 16 is connected to MAPK suggesting that this isoform participates in a functional pathway with other proteins. SIERF 80 and SIERF 9 were connected to HSF factor and may be involved in heat stress. We previously reported a rapid and high expression of SIERF 9, SIERF 16 and SIERF 80 within salt tolerant tomato cultivar (Gharsallah et al. 2016b). So that tomato-specific ERF factors have critical functions in adaptive mechanistic responses when abiotic and biotic stresses occur simultaneously. The tomato SIERF 84 plays a paradoxical role in Arabidopsis since it enhanced tolerance to drought and salt stresses while having negative role in modulating resistance to pathogen (Li et al. 2018).

Symptom phenotyping based on scale-class assortment, marker-genotyping, association network and WRKY and ERF gene expression patterns were explored to give insights into tomato defense mechanisms. TYLCSV infected tomato contrasting genotypes were challenged to a hostile environment such salinity stress. Discriminating alleles and candidate genes, which are differentially expressed in the resistant genotype, may be involved in a complex network sustaining resistance to both abiotic and biotic stresses.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author’s contribution

CG and HF designed the experiments. HF obtained funds to carry out the work. SG provided the infectious viral clone. CG performed the experiments. SW contributed to data analysis. CG and FG wrote the paper. FG supervised the project and performed critical revision.

Funding

This study was partially supported by the Ministry of Higher Education and Scientific Research of Tunisia and the USAID Project TA-MOU-08-M28-048.

Compliance with ethical standards

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.