Arabidopsis thaliana Col-0, Kondara (Kon), and Bla-2 (Bla-2) accessions (Nottingham Arabidopsis Seed Collection) were used as wild-types (WT). A. thaliana T-DNA mutants (SALK_092139C, N672907, referred as mati in this article) were obtained from the Arabidopsis Biological Resource Centre (ABRC¹) through the European Arabidopsis Stock Centre (NASC²). T-DNA insertion and homozygous status and gene expression levels of the Salk lines were validated by conventional PCR and quantitative RT-qPCR assays (described below) (Supplementary Figure 1A). For growth in the soil, seeds were planted and incubated 5 days at 4°C. For in vitro growth, seeds were surface-sterilized with 75% (V/V) ethanol, dried, and plated onto Petri dishes containing 0.53 g of Murashige and Skoog salts (Sigma–Aldrich), 1% (W/V) sucrose, 0.5 g/L MES, and 0.4% (W/V) Phytagel (Sigma–Aldrich), adjusted to pH 5.7 with KOH. Plants and plates were then grown in growth chambers (Sanyo MLR-350-H) under control conditions (23°C ± 1°C, >70% relative humidity and a 16 h/8 h day/night photoperiod).

To generate overexpression lines, MATI cDNA from Col-0 plants was cloned into pGWB2 (CaMV35S, no tag) and pGWB5 (CaMV35S, C-sGFP) Gateway binary vectors (Nakagawa et al., 2007) using the specific primers included in Supplementary Table 1. The recombinant plasmids were introduced into A. thaliana Col-0 plants using Agrobacterium floral dip transformation (Clough and Bent, 1998), Col-MATI and Col-MATI-GFP plants in this article. pGWB2 plasmids were also introduced into A. thaliana Kon plants, termed Kon-MATI plants in this article. Shoots were regenerated on selective medium containing hygromycin (100 mg/L), and primary transformants (T0) were allowed to self-fertilize. Plants were then selected and self-fertilized twice more to generate the third generation lines (T3). Homozygous plants with one single copy insertion and the highest transgene expression levels coming from different transformation events were selected for our experiments (Supplementary Figure 1A).

RT-qPCR assays have been used for different purposes: (i) to validate data from transcriptomic analysis; (ii) to determine the homozygous status of the Salk mutant lines; (iii) to study MATI gene expression in major Arabidopsis tissues. A. thaliana rosettes from Col-0, Bla-2, and Kon accessions were sampled after different time of mite infestation (1, 3, 6, 12, and 24 h) to validate microarray results. Total RNA was extracted following Oñate-Sánchez and Vicente-Carbajosa (2008) and reverse transcribed using Revert Aid^TM H Minus First Strand cDNA Synthesis Kit (Fermentas). cDNAs from A. thaliana Col-0 flowers, roots, siliques, leaves from stem rosettes at 1- to 3-week-old were also prepared. RT-qPCR was performed for three samples coming from three independent experiments as previously described (Santamaria et al., 2012a) using a SYBR Green Detection System (Roche) and the CFX Manager Software 2.0 (Bio-Rad). Gene expression was referred as relative expression levels (2^-ΔCt) or fold change (2^-ΔΔCt) (Livak and Schmittgen, 2001). Similarly, gene expression levels involved in hormonal signaling pathways were analyzed in rosettes from Col-MATI_4.1, -mati and WT plants in three samples coming from three independent experiments of non-infested 1 and 24 h infested plants (20 mites/plant). mRNA quantification was expressed as relative expression levels (2^-ΔCt) normalized to ubiquitin (Livak and Schmittgen, 2001). Specific primers were designed through the Salk Institute T-DNA primer design link³ or through the PRIMER 3 program⁴. Primer sequences are indicated in Supplementary Table 1.

Mite Attack Triggered Immunity sequences were downloaded from the TAIR website⁵. Amino acid sequence was subjected to a sequence search in the Pfam database v28.0⁶ and SignalP 4.1 program⁷ to identify possible domains and a signal peptide within the protein, respectively. MATI proteins from different plant species were compiled using GreenphylDB v4⁸ and aligned by MUSCLE v3.8⁹. A phylogenetic tree was constructed by the maximum likelihood PhyML v3.0 method¹⁰, using a BIONJ starting tree and applying the approximate likelihood-ratio test (aLRT) as statistical test for non-parametric branch support. Displayed trees were visualized in the program MEGA 6.0¹¹. Conserved sites were analyzed by drawing sequence logos representing profile hidden Markov models using the Skytign tool¹². To analyze MATI gene product divergences among Arabidopsis accessions, its cDNA from Col-0, Bla-2, and Kon rosettes was cloned and sequenced using pGEM^®-T Easy Vector Systems. The three sequences were aligned using Clustal W 1.83¹³.

Mite Attack Triggered Immunity protein location within cell compartments were analyzed by two experimental approaches, using MATI gene fusions to the green fluorescent protein (GFP). First, transient expression assays were performed by particle bombardment of onion epidermal layers and second, stable transformation of Arabidopsis plants. The open reading frame (ORF) of the MATI gene translationally fused to the N-terminus of the whole reporter gene was cloned into the pGWB5 binary vector following Gateway technology instructions. As controls, the psmRS-GFP plasmid containing the cauliflower mosaic virus 35S promoter (Davis and Vierstra, 1998) and the pRTL2ΔNS/ss-RFP-HDEL plasmid containing the Arabidopsis chitinase signal sequence and the C-terminal HDEL ER retrieval signal, whose protein specifically localizes in the endoplasmic reticulum (ER) (Shockey et al., 2006), were used. Transient transformation of onion (Allium cepa) epidermal cells was performed by particle co-bombardment with a biolistic helium gun device (DuPont PDS-1000/Bio-Rad) as previously described (Diaz et al., 2005). Fluorescent images were acquired after 24 h of incubation at 22°C in the dark, under the LEICA SP8 confocal microscope (Leica). For the subcellular localization of the MATI-GFP fusion protein in A. thaliana, seedlings from Col-MATI_4.1-GFP line were observed under the LEICA SP8 confocal microscope (Leica) using the appropriated filters.

To test the involvement of MATI in plant responses to herbivory, T. urticae and Spodoptera exigua feeding bioassays were performed in MATI overexpressing, mati-mutant ans Col-0 Arabidopsis lines. A colony of T. urticae, London strain (Acari: Tetranychidae) provided by Dr. Miodrag Grbic (UWO, Canada), was reared on beans (Phaseolus vulgaris) and maintained in growth chambers (Sanyo MLR-350-H) at 23°C ± 1°C, >70% relative humidity and a 16 h/8 h day/night photoperiod. Mites were synchronized by inoculating 100 adult females (random age) on one leaf of bean confined in a closed system under water-soaked cotton. After 1 day, adult females were removed, and 10 days after, population on the leaf was synchronized. Spider mite development and behavior was studied on Col-MATI_4.1, _5.2, mati and Col-0 WT plants. T. urticae fecundity assay was performed on detached leaves from 3-week-old plants. The newest emerged leaf (about 1 cm long) from each plant was fit in special dishes and infested with 12 adult synchronized females. After 36 h of infestation, the number of eggs was counted. Eight replicates were used for plant genotype.

Spodoptera exigua eggs were kindly provided by Drs. F. Budia and E. Viñuela (Department of Producción Agraria, ETSI Agronomos-UPM, Spain). Larvae were maintained in boxes feeding on artificial diet (Poitout and Bues, 1970), in growth chambers (Sanyo MLR-350-H) at 25°C ± 1°C, > 70% relative humidity and a 16 h/8 h day/night photoperiod. The boxes were lined with filter paper to reduce humidity. Vermiculite was provided for pupation when larvae complete its development cycle. Adults were allowed to emerge in cylindrical containers supplied with 10% (V/V) honey solution in water. Eggs were deposited on strips of filter paper and neonate (24 h) and L1 larvae (48 h) were synchronized for the plant and herbivore behavior assays, respectively. Larval performance was tested by placing three freshly hatched (neonate) beet armyworm larvae were placed on 3-week-old plants from the different genotypes (Col-MATI_4.1, -mati and WT lines) and larval weight was measured in a precision balance Mettler-Toledo MT5 (Mettler-Toledo) after 4 days of feeding. Twelve replicates were used for each genotype.

Quantification of plant damage after arthropod feeding was done on Arabidopsis T2 entire plants from selected homozygous transgenic lines (Col-MATI_4.1, Col-MATI_5.2; Col-mati, Kon-MATI_6.1 and Kon-MATI_7.2) and from the non-transformed Col-0 and Kon controls. Three-week-old plants were infected with 20 T. urticae adults per plant. After 4 days of feeding, leaf damage was assessed by scanning the entire rosette using a hp scanjet (HP Scanjet 5590 Digital Flatbed Scanner series), according to Cazaux et al. (2014). Leaf damage was calculated in mm², using Adobe Photoshop CS software. Six replicates were used for each genotype.

Leaf disks (7 mm diameter) of 3-week-old plants from the 3 Col-0 genotypes (WT, Col-MATI_4.1 and Col-mati) were infected with two neonate larvae of S. exigua. After 12 h, leaf disks were scanned using the hp scanjet (HP Scanjet 5590 Digital Flatbed Scanner series). Leaf damage was calculated in mm² using split channel tool from imageJ software. Thirty two replicates were used for each genotype.

To identify MATI interaction partners, protein extracts from control and Col-MATI-GFP plants were immuno-purified using GFP-Trap system and analyzed in LTQ-Orbitrap Velos. Five gram of fresh weight of Col-GFP WT and Col-MATI-GFP (Col-MATI_4.1 line) plants grown for 3 weeks on half-strength MS medium were frozen in liquid nitrogen, extracted by grinding with a mortar and pestle, and added to 15 mL of cold TAP extraction buffer (50 mM TRIS-HCl, 150 mM ClNa, 10% (V/V) glycerol, 0.8% (V/V) Triton and 0.4% (W/V) Chaps (Sigma) and complete protease inhibitor cocktail (Roche). This lysate was centrifuged at 10,000 rpm for 10 min at 4°C and the protein content of the supernatant was quantified by Nanodrop^®ND1000. Co-immunoprecipitation (Co-IP) experiments from Arabidopsis plants expressing GFP-tagged proteins were performed using GFP-Trap^® beads (ChromoTek GmbH), following the recommendations of the manufacturer. The immune-precipitated proteins were detected by immunoblot analysis on nitrocellulose membranes (GE Healthcare) using a monoclonal anti-GFP antibody (Miltenyi Biotec) at 1:5000 dilution in BSA blocking buffer containing 3% (WV) PBS and 0.05% (V/V) Tween-20. Acrylamide gels were staining with Coomassie and Oriole (Bio-Rad) to identify putative interactions.

Co-immunoprecipitated proteins from Col WT (Col-0) and Col-MATI_4.1 samples were digested and analyzed in LTQ-Orbitrap Velos (Gil-Bona et al., 2015). Peptide identification from raw data was carried out using licensed version of search engine MASCOT 2.3.0 and Proteome Discoverer 1.4 (Thermo Scientific). Database search was performed against UniProt-Swissprot and NCBInr database with taxonomic restriction to A. thaliana. The following constraints were used for the searches: tryptic cleavage after Arg and Lys, up to two missed cleavage sites allowed, and tolerances of 20 ppm for precursor ions and 0.6 Da for MS/MS fragment ions. Besides, searches were performed allowing optional Methionine oxidation and fixed carbamidomethylation of Cysteine. Search against decoy database (integrated decoy approach) was used to FDR calculate and MASCOT percolator filter was applied to MASCOT results. The acceptance criteria for protein identification were a FDR < 1% and at least one peptide identified with high confidence (CI > 95%). Proteins identified as putative MATI interactors were functionally categorized using AgriGO v1.2¹⁴ and STRING v10¹⁵. In addition, the interaction among MATI putative partners was analyzed in STRING v10. The proteomic analysis was carried out at the Proteomics Facility UCM-PCM, a member of ProteoRed network.

Membrane depolarization is rapidly produced after plant–herbivore interaction followed by alterations in cell membrane and ion imbalance that can be monitored by measuring the electrolyte leakage (EL). Arabidopsis leaf disks (1 cm diameter) from Col-MATI_4.1, -mati and WT plants were infested with 10 mites and incubated for 24 h. The EL was determined as described by Masood et al. (2006). EL measurements were performed after 2 h (C1) of incubation at 32°C using a conductometer (EC-Metro BASIC 30, CRISON). Total electrolyte content was determined in the same way after boiling for 10 min (C2). Results were expressed as percentage of EL = (C1/C2) × 100. Six replicates were used for each genotype and treatment.

To check whether plants reconfigure their primary metabolism during herbivore feeding to cope with the increased metabolic demands, chlorophyll and carotenoids were analyzed as standard parameters of photosynthetic activities in the three different Col-0 genotypes. Chlorophyll a and b, and total carotenoids were extracted from Col-MATI_4.1, -mati, WT lines. Arabidopsis rosettes were infested with 10 mites and incubated for 24 h. One hundred milligram of leaves were ground in a mortar with liquid nitrogen and suspended in 10 ml of 80% (V/V) acetone, using photo-protected tubes. After centrifugation at 3000 rpm for 15 min (L344 Eppendorf 5810R centrifuge), the absorbance of the supernatant was measured at 470, 663, and 646 nm, for carotenoids, chlorophyll a and b, respectively, using a UV-vis spectrophotometer (UltroSpec^TM 3300pro, Amersham Bioscience). Six replicates were used for each genotype and treatment. Pigment content was calculated using the extinction coefficients and equations determined by Lichtenthaler (1987).

The accumulation of H₂O₂ was determined using the 3,3-diaminobenzidine tetrachloridehydrate (DAB) substrate (Sigma–Aldrich) which produces a brown precipitate after oxidation in the presence of H₂O₂ (Martinez de Ilarduya et al., 2003). Col-0 Arabidopsis leaf disks (1 cm diameter) from Col-MATI_4.1, -mati and WT genotypes were infested with 10 mites and incubated for 24 h. Infested and non-infested control disks were stained with DAB following Rodríguez-Herva et al. (2012) and observed under a Zeiss Axiophot microscope. DAB staining specificity was confirmed in presence of the H₂O₂ scavenger, ascorbic acid (10 mM). ImageJ was used for the image quantification analysis¹⁶ using Methyl Green DAB vector. Six replicates were done by genotype and treatment.

The abundance of thiol groups was quantified in Col-MATI_4.1, -mati and WT a lines. Arabidopsis rosettes were infested with 10 mites and incubated for 24 h. Two hundred milligram of 3-week-old rosettes was homogenized with phosphate buffer (pH 6), centrifuged at 2500 rpm at 4°C for 10 min. Supernatant was used for the fluorometric thiol group assay following manufacturer instructions (Sigma–Aldrich). Six replicates were used for each genotype and treatment.

Since hormones are important regulatory components of defense signaling, and particularly JA and SA are known to play major roles in regulating plant defense responses against spider mites, the accumulation of JA and SA as well as ABA in infested and non-infested Col-0 genotypes was also measured. Plant hormones (OPDA: 12-oxo-phytodienoic acid, JA, JA-Ile, SA, and ABA) were quantified by isotopic dilution mass spectrometry from 3-week-old rosettes of Col-MATI_4.1, -mati and WT plants after 24 h of spider mite feeding. Six rosettes were pooled per experiment and three independent experiments were performed. Isotope-labeled standards were added to plant samples (∼0.1 g) before extraction as described previously (Durgbanshi et al., 2005). Ultra-performance liquid chromatography (UPLC)-electrospray ionization-tandem mass spectrometry analyses were carried out on an Acquity SDS system (Waters) coupled to a triple quadrupole mass spectrometer (Micromass). Quantification was accomplished with an external calibration.

Statistical analyses were performed using one-way ANOVA for gene expression validation, expression in different tissues and damage analysis and T. urticae and S. exigua bioassays. Two-way ANOVA was used for gene expression validation, EL, H₂O₂, antioxidants, photosynthesis, metabolites and hormone gene expression studies. Student-Newman-Keuls multiple comparison test was applied to all the studies. In figures, significant differences (P < 0.05) among lines for different evaluated parameters, were reported with different letters. The F-value, P-value, and df (degree of fredom) obtained from each statistical analysis are compiled in Supplementary Table 2.

Zhurov et al. (2014) highlighted the natural genetic variation of resistance between Arabidopsis accessions to T. urticae, identifying Bla-2 and Kon as the accessions at the opposing ends of the spectrum. Genome-wide analysis of transcriptional responses of these accessions demonstrated that they are very similar, which led to the conclusion that constitutive differences in levels of JA metabolites may underlay their differential resistance to mite feeding. Among otherwise similar responses, the AT3G14395 gene showed consistently higher expression levels in the resistant Bla-2 strain relative to Kon (Figure 1A). These data were validated by RT-qPCR assays (Figure 1B). The AT3G14395 gene, named MATI, encodes a specific plant protein of unknown function. The predicted protein, of 8.35 kDa, contains 75 amino acids with no known domains and no signal peptide. Searches in the GreenPhylDB discovered that homologs of the MATI protein were widely and uniquely present in angiosperms, including both monocots and dicots. A phylogenetic tree was constructed using protein sequences of MATI homologs from the 27 different species present in GreenPhylDB (Figure 2A). The phylogenetic analysis clearly identified three different groups, one formed by proteins from monocot species and two groups containing proteins from dicotyledonous plants (Figure 2). The MATI sequence was classified in the dicot group B together with other additional five protein sequences. Multiple sequence alignment revealed that MATI gene product was identical in Col-0, Bla-2, and Kon accessions (data not shown). A distant paralog, the AT1G30260 from A. thaliana was present in the dicot group A (Figure 2A), but was not up-regulated after spider mite feeding in Bla-2 (Supplementary Figure 2).

Multiple sequence alignments of MATI homologs revealed the presence of a conserved N-terminal domain found in all protein sequences. A threonine-isoleucine pair followed by several negatively charged amino acid residues at the C-end of the conserved domain was invariably detected in all proteins. In addition, two proline residues toward the middle of the domain and several charged amino acids scattered along the conserved N-terminal domain were also identified in most cases (Figure 2B). Thus, MATI is a gene of an unknown function, but conserved among both mono- and di-cotyledon plants, potentially associated with plant responses to spider mite feeding.

To determine the expression of MATI gene, RT-qPCR studies were carried out in the major Arabidopsis tissues. MATI mRNA expression levels were mainly detected in flowers, seeds, siliques, and leaves from stem or rosette at different developmental stages, while it was scarcely detectable in roots (Figure 3A). Transient expression assays in onion epidermal cell layers were performed to determine MATI subcellular localization. The ORF of MATI translationally fused to the GFP was detected throughout the entire ER network, which was continuous with the nuclear membrane. This subcellular location was revealed by its co-location with the fluorescence emitted by 35S-RFP-HDEL plasmid, used as a control of endomembrane system location (Figure 3Bb,f,j,n). The MATI protein was also found in the plasma membrane, cytoplasm and within the nucleus (Figure 3Ba,e). Plasmolysis, induced by 1 M mannitol treatment confirmed the MATI location pattern in the cell-detached plasma membrane, with no fluorescence observed either in the cell wall or in the apoplast (Figure 3Bm,o). As expected, the 35S-GFP control showed fluorescence throughout the whole cell (Figure 3Bi,k). Furthermore, Col-MATI-GFP overexpressing lines were created to corroborate MATI subcellular location in the whole plant. The stable expression of 35S-MATI-GFP plasmid in transgenic Arabidopsis Col-0 plants confirmed the protein location in the endomembrane system, cytoplasm and nucleus as shown transgenic in roots of Arabidopsis transgenic plants (Figure 3C).

To investigate the role of MATI protein in plant defense, silenced lines (mati Salk) and overexpressing Arabidopsis Col-0 and Kon lines (Col-MAT and Kon-MATI, respectively) were generated. The characterization of the homozygous mati Salk line revealed a loss-of-function allele generated by the insertion of the T-DNA at the first part of the exon (Supplementary Figure 1A). On the other hand, gene expression analysis of transgenic plants expressing MATI gene constitutively in Col-0 and Kon backgrounds allowed the selection of lines that overexpressed the MATI ORF for further studies (Supplementary Figure 1Ba,b).

Homozygous knock-out (Col-mati), overexpressing lines (Col-MATI_4.1, _5.2 and Kon-MATI_6.1, _7.2) as well as the corresponding WT plants were infested with spider mites and plant damage was quantified 4 days after mite feeding (Figures 4A,C). Col-MATI overexpressing lines showed approximately three times less damage than non-transformed Col WT infested plants (exactly 2.83- and 2.68-fold for Col-MATI_4.1 and _5.2, respectively). Similarly, overexpression of MATI in Kon (Kon-MATI_6.1 and _7.2) resulted in a significant decrease in damage (1.25- and 2.7-fold for Kon-MATI_6.1 and _7.2, respectively) in comparison to infested Kon WT. In contrast, damaged area in the knock-out Col-mati lines was about 2.5-fold greater than the damage quantified in Col-0 WT lines after mite feeding. The damage intensity measured as the chlorotic area on infested leaves was proportional to the MATI expression levels. To ensure that chlorotic area correlates with mite enhanced feeding and is not a consequence of greater cell death, mite performance were determined after feeding on plants with different levels of MATI expression. Fecundity assays were carried out on leaves from different Col-0 genotypes and showed that synchronized mites feeding on WT and mati plants had higher fecundity rates than the ones feeding on overexpressing MATI lines (Figure 4D). Thus, greater leaf damage reflects a mite’s ability to feed more intensely and to be more fecund, indicating that MATI overexpression correlates with plant’s capability to defend against herbivory.

To evaluate whether the plant defense mediated by MATI gene was specific to acari infestation, leaf disks from different Col-0 genotypes were infested with a generalist lepidopteran, the beet armyworm S. exigua, and the leaf damage was quantified. As shown in Figures 4B,C, the damaged area was significantly higher in Col-mati and Col-0 WT lines than in the overexpressing Col-MATI_4.1 line. In addition, the larval weight after 12 h of feeding on Col-mati and WT plants was twice and 1.5-fold greater, respectively, relative to the weight of the beet armyworm larvae fed on Col-MATI_4.1 (Figure 4D). Therefore, MATI overexpression confers plant protection against multiple herbivores.

Using NCBInr database after Co-IP assays, 94 proteins were specifically found in MATI sample but not in the control WT. Most of these proteins were recognized by the identification of one peptide (Supplementary Table 3). The identified proteins were classified in five different over-represented categories based on their GO biological function related to sulfur compound metabolism, generation of precursor metabolites and energy, photosynthesis and Serine metabolism (Supplementary Figure 3A). When the identification was performed with a singular enrichment analysis in AgriGO, the Go biological process categories overrepresented included oxidation-reduction and metabolic processes, photosynthesis, and generation of precursor metabolites and energy (Supplementary Figure 3B). Additionally, the interactome network for the MATI putative partners revealed an enrichment of interactions between MATI putative interactors, with 247 observed interactions over the 100 expected for a total of 94 proteins. Among those partners putatively involved in sulfur compound metabolic process, 58 interactions were detected over the 6 expected (Supplementary Figure 3C). Since the two main overrepresented functional categories were related with sulfur metabolism and photosynthesis, different parameters involved in these physiological processes were further analyzed.

Since MATI expression was induced early upon mite feeding on Arabidopsis leaves, it was important to determine how MATI might modulate plant responses to herbivore feeding. Thus, the EL was determined in Col-0 genotypes after mite feeding. Results showed a significant higher leakage either constitutively or mite-induced leakage in Col-mati plants than in Col-MATI overexpressing or in WT plants (Figure 5A). As these changes in conductance are usually accompanied by ROS generation, H₂O₂ concentrations were determined in the three Col-0 genotypes. The quantification of H₂O₂ in infested plants, expressed as DAB units, demonstrated that Col-0 WT and Col-mati plants accumulated about twice and ten-times more DAB, respectively, than their corresponding non-infested genotypes. However, the H₂O₂ levels in infested Col-MATI_4.1 line changed only about 1.5-fold the DAB values relative to the non-infested line (Figure 5B). As thiol-type compounds are an important class of antioxidants that quench reactive oxidant species, variations in the total thiol group’s content among different Col-0 genotypes were also quantified. T. urticae feeding caused a significant decrease in the thiol’s group concentration in WT and Col-mati plants, while the thiol accumulation pattern was not significantly altered in Col-MATI samples (Figure 5C).

Three-week-old mati-mutant, overexpressing and control Col lines were used to determine chlorophyll and carotenoid content in the presence/absence of spider mites. While reduced levels of both pigments were observed in infested WT and Col-mati lines, high chlorophyll and carotenoid accumulation was detected in overexpressing Col-MATI_4.1 lines after mite feeding in comparison to non-infested genotypes (Figures 5D,E). Consequently, MATI helps to maintain moderate levels of ROS necessary for defense signaling to control redox homeostasis, and to retain high levels of pigments to avoid photosynthetic cost.

The hormonal analyses demonstrated that OPDA, JA, and its bioactive form JA-Ile, accumulated in leaf tissue during the initial 24 h in response to spider mite feeding, independently of the genotype (Figures 6A–C). The free OPDA content increased mainly in the spider mite-treated Col-MATI and -mati lines which contained about twice the JA precursor in infested WT (Figure 6A). The JA and JA-Ile accumulation also increased in Col-mati and WT genotypes, but the highest hormone levels were detected in the Col-MATI overexpressing lines after mite infestation. It was detected as an increase of approximately 45- and 16-fold in the JA and JA-Ile content, respectively (Figures 6B,C). Simultaneously, SA and ABA concentrations were also determined in the same rosette samples. Interestingly, high levels of SA were accumulated in non-infested Col-MATI lines but they were significantly reduced after mite feeding. In contrast, SA increased in response to infestation in Col-mati and WT lines (Figure 6D). Regarding ABA levels, they exclusively increased in infested Col-mati lines (Figure 6E). Thus, MATI is involved in the modulation of different hormonal signaling pathways in response to spider mite attack.

To further analyze the molecular basis of the MATI gene function, the expression pattern of some JA-, SA- and ABA-related genes located at different levels in the hormonal signaling pathways was analyzed 24 h after mite feeding. Lipoxygenase 3 (LOX3), Allene Oxide Synthase (AOS), MYC2 transcription factor (MYC2), vegetative storage protein 2 (VSP2), and plant defensin (PDF1.2) were the selected genes to dissect JA pathway. Non-infected Col WT plants showed low levels of constitutive expression of all JA-related genes but their expression increased after spider mite infestation. Interestingly, most of the selected defense genes presented higher mRNA levels in the infested overexpressing MATI than in WT or mati lines. Only the expression values of the PDF1.2 gene related to the JA/Et defense branch were lower in Col-MATI_4.1 overexpressing line than in Col-mati or WT plants (Figures 7A–E). Similarly, Isochorismate Synthase 1 (ICS1), Non-expresser of PR gene1 (NPR1) and Pathogenesis Related (PR1) genes involved in SA-biosynthesis, regulation and final defense product, respectively, were selected to study SA implications. The three genes showed lower expression in infested than in non-infested Col-MATI lines while in Col-mati and WT lines the expression of some genes were increased, some decreased and other did not present differences in response to mites (Figures 7F–H). Additionally, the expression of the same JA- and SA-regulated genes was analyzed in early mite infestation assays, just 1 h after mite attack. Regarding the genes involved in the JA pathway, LOX3, AOS, and MYC2 were up-regulated after mite feeding in WT, MATI, and mati lines while the VSP2 gene was exclusively and highly induced in infested overexpressing MATI lines. Curiously, the PDF1.2 gene was only induced in infested WT lines since its basal expression levels in non-infested MATI lines was as higher as the induced levels detected in WT plants after mite feeding. This gene was not induced in mati lines either infested or not infested (Supplementary Figures 4A–E). The expression of the genes involved in SA-biosynthesis, ICS1, NPR1, and PR1, was similar or lower in infested and non-infested Col-MATI lines but their expression levels in mati lines was not altered in response to mites. In contrast, ICS1 and NPR1 genes were up-regulated in infested WT lines and the PR1 gene did not vary (Supplementary Figures 4F–H).

Besides, three genes related to ABA pathway, ABA deficient 1 (ABA1), Regulatory Component of ABA Receptor 3 (RCAR3), and Responsive to Dessication 22 (RD22), were also checked. The ABA1 gene decreased after mite infestation following the same expression pattern in all genotypes, but RCAR3 and RD22 genes exclusively increased in mati knock-down lines after mite infestation (Supplementary Figure 5). All these results are in agreement with the hormonal content measured in the MATI overexpressing and knock-down genotypes and reinforce the importance of MATI in the regulation of different hormonal signaling pathways.