Mycorrhization and chemical seed priming boost tomato stress tolerance by changing primary and defence metabolic pathways

Priming modulates plant stress responses before the stress appears, increasing the ability of the primed plant to endure adverse conditions and thrive. In this context, we investigated the effect of biological (i.e. arbuscular mycorrhizal fungi, AMF) agents and natural compounds (i.e. salicylic acid applied alone or combined with chitosan) against water deficit and salinity on a commercial tomato genotype (cv. Moneymaker). Effects of seed treatments on AMF colonization were evaluated, demonstrating the possibility of using them in combination. Responses to water and salt stresses were analysed on primed plants alone or in combination with the AMF inoculum in soil. Trials were conducted on potted plants by subjecting them to water deficit or salt stress. The effectiveness of chemical seed treatments, both alone and in combination with post-germination AMF inoculation, was investigated using a multidisciplinary approach that included ecophysiology, biochemistry, transcriptomics, and untargeted metabolomics. Results showed that chemical seed treatment and arbuscular mycorrhizal symbiosis modified the tomato response to water deficit and salinity triggering a remodelling of both transcriptome and metabolome, which ultimately elicited the plant antioxidant and osmoprotective machinery. The plant physiological adaptation to both stress conditions improved, confirming the success of the adopted approaches in enhancing stress tolerance.

Methods
Results reported in Datasets from S1 to S36 were generated after bioinformatic elaboration of RNA-seq (Datasets from S1 to S30) and untargeted metabolomics (Datasets from S31 to S36) data.

Data elaboration procedures are following reported:

RNA-seq data analysis

For alignment, reads were mapped onto the reference genome GCF_000188115.5_SL3.1 (Hosmani et al., 2019) using STAR v. 2.7.10 (Dobin et al., 2013), a splice junction mapper designed for RNA-Seq reads, under default parameters. The software htseq-count v. 2.0.2 (Anders et al., 2015) was utilized to count the overlapping of reads with genes. The data were then used to identify differentially expressed genes (DEGs) using the DESeq2 package v1.34.0 (Love et al., 2014). The variance on read count was calculated based on three biological replicates per condition by applying a negative binomial distribution to model the count data, therefore identifying genes showing significant expression changes among the different tested conditions. The DEG identification was performed after normalization of the count data and correction for multiple testing, both accounted by DESeq2, through the Wald test. During DESeq2 analysis, the shrinkage estimation of effect size (LFC estimates) was used, to generate more accurate Log2 foldchange estimates and considering the variability among replicates. A cut-off of the p-adjusted value < 0.05 was used to classify a gene differentially expressed (DEG) in comparison with the reference (i.e. untreated (CTRL) in not-stressed condition (NS) and without AM fungal inoculation (NMYC). Both the identified DEGs and all transcripts of the tomato (Solanum lycopersicum L.) transcriptome were annotated through Blast2GO v5.2.5 (Conesa et al., 2005) to obtain an updated functional annotation and to assign the corresponding Gene Ontology (GO) terms. A gene class functional enrichment analysis was then conducted using Blast2GO to reveal the biological processes, pathways, or other functional categories that are enriched among the identified DEGs.

Non-targeted polar metabolite profiling

Polar metabolites were separated using hydrophilic interaction liquid chromatography (HILIC) coupled to hybrid quadrupole-time of flight mass spectrometry (QTOF-MS) according to Andrade et al. (2021). HILIC separation was performed on a 2.1×100 mm InfinityLab Poroshell 120 HILIC-Z, 1.9 µm (Agilent Technologies, Inc., Santa Clara, CA, USA) using acetonitrile:water, 95:5 (v/v) (solvent A) and acetonitrile:water, 2:98 (v/v) (solvent B), both supplemented with ammonium formate at 0.063% and 0.126%, respectively, as solvents and at a flow rate of 0.3 ml min−1. During chromatographic runs, column temperature was at 40°C. Tomato leaf samples (10 mg of dry weight) were extracted in triplicate by ultrasonication in 300 µl of 80% aqueous methanol supplemented with kinetin (1 mg l−1) as internal standard for relative quantitation. After extraction, samples were centrifuged at 10,000 rpm and 4ºC for 10 min and the supernatants recovered. Subsequently, supernatants were diluted 1:4 with acetonitrile (LC/MS grade) and filtered through 0.2 µm PTFE syringe filters directly into chromatography vials. Mass chromatographic data were acquired in positive and negative ionization modes within the 50–1000 amu mass range. Nitrogen was used both as nebulization and desolvation gas (60 and 800 l h−1 and 350°C temperature, respectively). During measurements, capillary and cone voltages were set at 3.5 kV and 30 V for positive electrospray and 2.3 kV and 30 V for negative electrospray, respectively. An additional acquisition function to obtain collision-induced dissociation (CID) information was set by performing a voltage ramp between 6–40 eV. To ensure accurate mass data acquisition, a lockmass reference (Leu-Enkephalin, [M+H]+ 556.2771 and [M-H]- 554.2614) was regularly infused during runs. Data files from each ionization mode were converted to mzML with msconvert (Chambers et al. 2012) and processed independently with XCMS (Smith et al., 2006). Mass chromatographic features were annotated and grouped with CAMERA (Kuhl et al., 2012). Peak areas were normalized to internal standard area and actual sample weight before statistical analyses. Significantly altered mass chromatographic features were subsequently identified as individual compounds by matching mz and retention time values with those of authentic standards or tentatively annotated by matching experimental mass spectra in public databases (Metlin, Massbank or HMDB).