miR825‐5p‐regulated TNLs govern Arabidopsis resistance to Tetranychus urticae and Pieris brassicae

Summary

• MicroRNAs are essential regulators in plant resistance to biotic stresses, but their specific roles in the plant–herbivore context require deeper investigation.

• Here, we studied how the Arabidopsis miR825‐5p differentially modulates certain TNLs (MRT1, MRT2, and MIST1), triggering defensive responses against the sucking acari Tetranychus urticae or the chewing insect Pieris brassicae.

• We demonstrated that the expression of miR825‐5p is downregulated following T. urticae and P. brassicae infestation and identified miRNA TNL targets whose expression is induced by both feeders. miR825‐5p downregulates MRT1 and MRT2, acting as a negative modulator of Arabidopsis basal resistance against T. urticae. A similar miR825‐5p‐mediated regulation of basal resistance, including MRT1 and MRT2 but also MIST1 silencing, is involved in the Arabidopsis response against P. brassicae. Moreover, miR825‐5p triggered the production of MIST1‐derived secondary interfering RNAs (phasiRNAs) and amplified silencing of MRT1 and MRT2.

• Taken together, our findings reveal the role of the miR825‐5p/TNL module in controlling Arabidopsis response and adapting plant defenses based on the specific threat.

Introduction

Plants require a rapid and appropriate defense system to minimize the damage caused by phytophagous arthropods, which depends on the effectiveness of the host and must be specific against the herbivore. The plant defense response is mediated by the reprogramming of functionally related genes, which are transcriptionally coordinated to optimize strategies that prevent a severe impact on plant growth and development. The molecular events involved in this complex process demand a fine‐tuned regulation at transcriptional and post‐transcriptional levels, with microRNAs (miRNA) playing key roles and mostly linked to a hormonal control (Mostafa et al., ²⁰²²).

The defense response starts with the recognition of herbivore/damage‐associated molecular patterns (HAMPs/DAMPs) through cell‐surface pattern recognition receptors (PRRs), activating a cascade of signal transduction pathways, finally leading to the activation of immunity (Ngou et al., ²⁰²²). This set of defensive reactions constitutes a primary immune system called pattern‐triggered immunity (PTI; Santamaria et al., ²⁰¹⁸; Stahl et al., ²⁰¹⁸). In turn, herbivores deliver effectors into host cells to manipulate and overcome plant defenses. The presence or activity of these protein effectors may be specifically detected by intracellular receptors, triggering a faster and more robust plant response known as effector‐triggered immunity (ETI; Cui et al., ²⁰¹⁵; Yuan et al., ^2021a). PTI and ETI are interdependent and are activated almost simultaneously. Although they require different early signaling factors and receptors, both converge into similar downstream responses, and their crosstalk can stimulate complementary and novel strategies for pest control (Yuan et al., ^2021b).

Among the intracellular receptors, nucleotide‐binding leucine‐rich repeat receptor proteins (NLRs), known as R‐genes, are key detectors mainly associated with ETI responses in the plant–herbivore context. In addition to the C‐terminal leucine‐rich repeat sequence (LRR) involved in effector recognition, these proteins contain a variable N‐terminal domain and a nucleotide‐binding domain (NB‐ARC), which are linked to perception and signal transduction functions, respectively (Hong et al., ²⁰¹⁷; Maruta et al., ²⁰²²). NLRs are classified into three major classes: (1) TNLs, the most abundant group with an N‐terminal Toll/interleukin‐1 (TIR) domain; (2) CNLs with a coiled coil (CC) domain, and (3) RPW8‐NLRs or RNLs, the least abundant class with an RPW8‐like CC‐type (resistance to powdery mildew 8) domain also known as helper NLRs (hNLRs) (Jubic et al., ²⁰¹⁹; Lopez‐Marquez et al., ²⁰²³; Wang et al., ²⁰²³). Recent studies on CNLs and TNLs have shown that NLR multimerization is required for function. Upon effector activation of the CNL ZAR1, inactive monomers form a pentameric, wheel‐like complex termed the resistosome, which triggers the release of the terminal α‐helices of the CC domain to form a funnel‐like structure that inserts into the plasma membrane. This structure is required for cell death‐inducing activity (Wang et al., ^2019a,b). Effector‐induced TNL tetramerization (Ma et al., ²⁰²⁰; Martin et al., ²⁰²⁰) leads to the formation of a holoenzyme in which the TIR domains exhibit enzymatic activity. In this complex, TIR domains cleave NAD⁺ into nicotinamide (Essuman et al., ²⁰¹⁸), and a variety of ADP‐ribose (cADPR) isomers (Manik et al., ²⁰²²), and related signaling molecules (Horsefield et al., ²⁰¹⁹; Wan et al., ²⁰¹⁹; Huang et al., ²⁰²²; Jia et al., ²⁰²²). Some of the molecules produced by TIR enzymatic activity are perceived by EDS1–PAD4 or EDS1–SAG101 heterodimeric nucleotide receptors, which bind and activate distinct helper NLRs upon nucleotide binding (Huang et al., ²⁰²²; Jia et al., ²⁰²²; Maruta et al., ²⁰²³).

TNLs, in particular, play a crucial role in regulating immunity and host cell death to prevent pathogen spread or herbivore feeding (Jubic et al., ²⁰¹⁹; Ngou et al., ²⁰²²; Jacob et al., ²⁰²³). To our knowledge, TNLs are the only NLRs activated in the ETI response to herbivore infestation (Garcia et al., ²⁰²¹; Santamaria et al., ²⁰²¹). To date, there is no information about CNLs or RPW8‐NLRs in the plant–herbivore scenario. The activation of TNLs is mainly controlled by threat species, with the receptor gene expression strictly regulated to maintain an appropriate balance between defense and growth (Garcia et al., ²⁰²¹). Within these precise mechanisms of control, miRNAs play an important role in modulating TNL production at the post‐transcriptional level, preventing excessive activation and limiting the costs of plant resistance (Lopez‐Marquez et al., ²⁰²³; Rodrigues et al., ²⁰²⁴). miRNAs are endogenous noncoding RNAs of 21–22 nt long, derived from single‐stranded miRNA precursors (primary miRNAs) that fold into hairpin stem‐loop secondary structures. Primary miRNAs are processed by the RNAse III DICER‐like1 (DCL1) into a double‐strand miRNA–miRNA* duplex, which is transported from the nucleus to the cytoplasm and loaded into ARGONAUTE (AGO) proteins to form a functional miRNA‐induced silencing complex (miRISC). As a result, the mature miRISC complex induces post‐transcriptional gene silencing by direct cleavage of messenger RNAs or its translational repression through a base pairing mechanism. In addition to directly targeting specific genes (silencing in cis), 22‐nt miRNAs may also indirectly regulate gene expression by promoting the production of a phased array of secondary interfering RNAs (which can be phased and then known as phasiRNAs) from primary target transcripts. These secondary siRNAs, which act in a homology‐dependent manner, can amplify the silencing effects on the primary targets (reinforcing cis silencing) or on additional target genes (silencing in trans) (Liu et al., ²⁰²⁰; Kumar et al., ²⁰²²). Expression of NLRs can be regulated by miRNA silencing alone (e.g. miR158a regulation of TNL7; Li et al., ²⁰²⁵) or by secondary siRNAs that can be generated from NLR transcripts (Li et al., ²⁰¹²; Shivaprasad et al., ²⁰¹²; Boccara et al., ²⁰¹⁴; Deng et al., ²⁰¹⁸; Cui et al., ²⁰²⁰; Lopez‐Marquez et al., ²⁰²¹). Secondary siRNAs generated from NLR mRNAs have been shown to act in cis, targeting transcripts from the miRNA target gene from which they are generated (e.g. miR472‐generated phasiRNAs; Boccara et al., ²⁰¹⁴), or in trans, targeting transcripts from sequence‐related genes not silenced by the miRNA (e.g. miR825‐5p‐generated phasiRNAs; Lopez‐Marquez et al., ²⁰²¹).

While the miRNA‐mediated regulation of plant defenses against pathogens has been widely studied (Lopez‐Marquez et al., ²⁰²³; Luo et al., ²⁰²⁴; Rodrigues et al., ²⁰²⁴; Li et al., ²⁰²⁵), the role of host miRNAs in response to phytophagous arthropods is less understood. Most of the studies have just analyzed the differential profile of herbivore‐responsive miRNAs in resistant and susceptible crop varieties, conducted at different time points after infestation. Some miRNA‐targeted genes linked to defense pathways have been identified in resistant plants, further supporting their regulatory role in plant responses to herbivores (Sattar et al., ²⁰¹²; Wu et al., ²⁰¹⁷; Li et al., ²⁰¹⁹; Han et al., ²⁰²²; Malhotra et al., ^2022a,b). This is the case of miR156, a master plant ontology regulator whose silencing in rice conferred resistance to the brown planthopper Nilaparvata lugens. miR156 acted as a negative regulator of resistance by modulating jasmonate levels (Ge et al., ²⁰¹⁸). Similarly, sweet potato plants overexpressing miR408 exhibited reduced resistance to Spodoptera litura feeding. In these transgenic plants, certain miR408‐target genes, including IbKCS (a‐ketoacyl‐CoA synthase) and a fatty acid elongase involved in the synthesis of cuticle wax, were repressed, suggesting a negative correlation between miR408 and its targets. The weight of S. litura larvae fed on IbKCS‐overexpressing tobacco plants was lower than that of larvae fed on control plants, demonstrating the coordination between miR408 modulation and IbKCS function (Kuo et al., ²⁰¹⁹). Yan et al. (2023) reported that sequestering alfalfa miR396 by overexpressing an artificial target mimicry form of miR396 improved resistance to S. litura larvae, along with increased levels of lignin, flavonoids, and glucosinolates. Likewise, Pradhan et al. (2017) found that AGO8‐silenced plants were highly susceptible to Manduca sexta infestation due to compromised levels of defensive metabolites such as nicotine, phenolamides, and diterpenoid glycosides. Very recently, the role of salivary miR‐7‐5P from the rice planthopper N. lugens has been demonstrated as an effector that suppresses plant immunity when secreted into host plants during insect feeding. miR‐7‐5P targeted transcripts of the immune‐associated rice bZIP43 transcription factor. Furthermore, infestation of rice plants by miR‐7‐5P‐silenced insects led to the increased expression of this factor, while the presence of miR‐7‐5P counteracts this upregulation effect. OsbZIP43‐overexpressing plants conferred rice resistance against insects (Zhang et al., ²⁰²⁴). So, miR‐7‐5P is the first miRNA described as an insect effector. These findings highlight the complexity of the miRNA‐mediated regulation of defenses against herbivores.

Some studies have identified TNL genes as targets of miRNAs linked to plant immunity to pathogens (Lopez‐Marquez et al., ^{2021, 2023}; Jacob et al., ²⁰²³; Li et al., ²⁰²⁵), but no data have yet described the specific transcriptional or post‐transcriptional regulation of TNLs mediated by miRNAs in response to herbivores. Therefore, it is highly relevant to investigate how miRNAs may differentially regulate defense responses to arthropods with different feeding modes.

In this study, we have focused on the functional characterization of miR825‐5p as a modulator of Arabidopsis responses against the generalist mite Tetranychus urticae (Tetranychidae family) and the specialist insect Pieris brassicae (Pieridae family). T. urticae is a piercing‐sucking mite that feeds along different phases of its development (larvae, nymphae, and adults) by inserting its stylet to access nutrient‐rich mesophyll (Santamaria et al., ²⁰²⁰). P. brassicae is a chewing lepidopteran insect that actively consumes plant tissues during its neonate larvae state (Younas et al., ²⁰⁰⁴). Both phytophagous species are widely spread, producing devastating consequences for agriculture, and, despite their common ability to feed on Arabidopsis, they are highly distinct in their feeding mechanisms and interactions with the plant. We find that mite and Pieris larvae infestation induces expression of TNL genes that are targets of miR825‐5p and describe how miRNA825‐5p acts as a negative regulator of plant defenses in a species‐specific manner, either through cis‐ or trans‐silencing effects, which may reflect differences in these herbivore feeding modes.

Materials and Methods

Plant material and growth conditions

Arabidopsis thaliana ecotype Col‐0, obtained from the Nottingham Arabidopsis Seed Collection (NASC; http://arabidopsis.info/BasicForm/), was used as the wild‐type in all experiments. Various A. thaliana mutant lines were acquired from different collections, specific world‐wide laboratories, or generated in our laboratory. MRT1 and MRT2 T‐DNA insertion mutants (mrt1‐1: SALK_141149, mrt1‐2: SALK_112450; mrt2‐1: SALK_050251, mrt2‐2: SALK_040476) were obtained from NASC. Double MRT mutants (mrt1‐mrt2 #1 and mrt1‐mrt2 #2) were generated for this study by crossing mrt1‐2 and mrt2‐1. miR825‐5p mutant plants were obtained from Lopez‐Marquez et al. (2021). All Arabidopsis seeds were surface sterilized with 70% ethanol for 2 min, incubated in a solution containing 5% SDS and 5% NaCl for 10 min, and finally washed with sterilized deionized distilled H₂O. Seeds were planted in peat moss and vermiculite (3 : 2) and stratified in the dark at 4°C for 5 d. Plants were grown in growth chambers (Aralab mod Fitoclima D1.200PLH‐LED; Rio de Mouro, Portugal) under controlled conditions (23 ± 1°C, > 70% relative humidity, and a 16 h : 8 h, day : night photoperiod) for c. 2 wk. The growth chamber includes a light control program to mimic sunrise and sunset conditions.

For transient expression experiments, Nicotiana benthamiana plants were grown in glasshouse facilities under controlled conditions (23 ± 1°C, > 70% relative humidity, and a 16 h : 8 h, day : night photoperiod).

Protein structure and homology

For protein visualization, homology‐based structural modeling was performed using AlphaFold DB v.2022‐11‐01, created with the AlphaFold Monomer v.2.0 pipeline (Jumper et al., ²⁰²¹). To assess the functional relevance of MRT1 and MRT2 proteins, we conducted a comparative analysis of their features and sequences using multiple sequence alignment (Muscle). Protein domain analysis was carried out using Pfam and Prosite databases. Gene Ontology analysis was performed using Gene Ontology Consortium tools (http://www.geneontology.org).

Pest maintenance and infestation assays

Tetranychus urticae London strain (Acari: Tetranychidae), provided by Dr Miodrag Grbic (UWO, Canada), was reared on Phaseolus vulgaris (beans) and maintained in growth chambers (Sanyo MLR‐351‐H; Sanyo, Osaka, Japan) at 25 ± 1°C, > 70% relative humidity, and a 16 h : 8 h, day : night photoperiod. Mite infestation was performed on 3‐wk‐old Arabidopsis rosettes or leaf disks with 20 mites per plant or 10 mites per leaf, adjusting the infestation time according to the experiment. For leaf infestation assays, leaf number 5 or 6 was selected (Merchant & Pajerowska‐Mukhtar, 2015), placed on ½‐strength Murashige & Skoog medium plates (Duchefa Biochemie, Haarlem, the Netherlands), infested, and covered but ventilated with a relative humidity of 60–80%. A P. brassicae colony, supplied by Lombrices California (Tarragona, Spain), was maintained in growth chambers (Sanyo MLR‐351‐H; Sanyo) at 25 ± 1°C, > 70% relative humidity, and a 16 h : 8 h, day : night photoperiod. Freshly hatched caterpillars were placed on 3‐wk‐old Arabidopsis rosettes with five caterpillars per plant or two caterpillars per disk, adjusting the infestation time according to the experiment.

Plant damage assays

The chlorotic damage experiments for measuring T. urticae‐induced damage to plants were conducted as follows: Adult T. urticae females were allowed to feed on Arabidopsis plants for 4 d. Plants were then collected, and the foliar area was scanned using a conventional scanner (HP Scanjet 5590 Digital Flatbed Scanner series, Madrid, Spain). The total leaf area of each rosette and the damaged portion (in mm²) were measured using Adobe Photoshop cs software and analyzed using Ilastik and Fiji, essentially as described by Ojeda‐Martinez et al. (2020). Nine biological replicates from independent rosettes were assayed for each genotype.

To assess the feeding activity of P. brassicae larvae, the extent of unconsumed leaf area on the disk was measured following an 8‐h feeding period. Each disk's area was captured through scanning using a conventional scanner (HP Scanjet 5590 Digital Flatbed Scanner series). The leaf area of individual rosettes was also quantified. For each genotype, nine biological replicates derived from independent rosettes were included in the assay.

T. urticae pest experiments performance

Spider mite development was studied on detached leaves from 3‐wk‐old plants. Small dishes (35 mm diameter) filled with some water and covered with Parafilm were used for these assays. The newest emerged leaf (c. 1 cm long) from each plant was fitted in the plate by introducing its petiole across the Parafilm to keep in contact with water. For the fecundity, 12 synchronized females were used to infest each leaf, and the number of eggs laid was counted under a stereoscope after 36 h of infestation according to Arnaiz et al. (2022). Six biological replicates from independent rosettes were used for each genotype. For the mortality, leaves were infested with 25 neonate larvae mites (24 h). After infestation, the plates were covered with a lid with ventilation and Parafilm to avoid possible escapes. Every day, the number of larvae that developed into protonymphs and those that died was counted to calculate developmental stages and mortality rates. Every 2 d, a new leaf from a new plant was added. Results were represented as percentages of mortality. Eight replicates from eight independent plants were used for each plant genotype.

P. brassicae larvae weight

Five freshly hatched neonate P. brassicae were placed on 3‐wk‐old plants, allowing continuous feeding for 24 h, in growth chambers (Sanyo MLR‐351‐H; Sanyo) at 25 ± 1°C, > 70% relative humidity, and a 16 h : 8 h, day : night photoperiod. After all larvae were removed, the weight of each larva was measured by an analytical balance (Radwag AS220 R2 PLUS; Biogen Cientifica SL, Madrid, Spain).

Cell death quantification

Cell death quantification was performed by trypan blue staining after 16 h of infestation in light and dark conditions. Leaf disks were boiled in trypan blue solution, followed by a clarification process with 2.5 g ml⁻¹ of chloral hydrate (Sigma) solution according to Sanchez‐Vallet et al. (2010). Disks were placed onto glass slides in 50% glycerol and observed under an epifluorescence stereomicroscope using UV filters. Quantification was performed using Adobe Photoshop (Luna et al., ²⁰¹²).

DAPI nuclei/damage per cell

To assess chlorotic damage caused by T. urticae in PDC from HR, we performed DAPI staining on 3‐wk‐old leaf disks (0.8 mm diameter each) after 24 h of feeding. This method involved counting total cells using ImageJ (Fiji). Leaf disks were stained with a DAPI solution (0.2 mg l⁻¹) and incubated for 10 min under vacuum. Afterward, the leaf disks were washed three times with distilled water. An inverted SP8 confocal microscope (Leica Microsystems, Wetzlar, Germany) was employed to count the total cells, chlorotic cells, and cells with DAPI‐stained nuclei. The resulting overlapping images, saved as 8‐bit TIFF files, were analyzed using ImageJ (Fiji) to calculate the ratios between DAPI nuclei and chlorotic cells.

H₂O₂ determination

H₂O₂ accumulation and callose deposition were determined in 9 mm disks of Arabidopsis plants after 24 h of being infested with 10 mites per disk. Reactive oxygen species production was quantified using 3,3‐diaminobenzidine tetrachloride hydrate (DAB) as a substrate (Sigma‐Aldrich) as described by Martinez de Ilarduya et al. (2003).

Electrolyte leakage

We determined electrolyte leakage in 0.8 cm diameter leaf disks after 24 h of mite infestation. Mites and eggs were removed, disks washed with 10 ml ultrapure water, and dried. Next, the samples were shaken in 10 ml ultrapure water for 3 h at 25°C. Conductivity (L ₁) was measured with an MPC227 meter (Mettler Toledo, Greifensee, Switzerland). After heating at 95°C for 20 min, the final conductivity (L ₂) was measured. Electrolyte leakage (EL) was calculated as EL (%) = (L ₁/L ₂) × 100.

Nucleic acid analysis

To confirm the genotypes of the Arabidopsis lines used in this study and to analyze gene expression patterns, we employed various molecular techniques. For plant genotyping, genomic DNA was isolated from Arabidopsis T‐DNA insertion lines and Col‐0 plants. T‐DNA homozygous status was validated by conventional PCR using ExTaq DNA Polymerase (TaKaRa, Saint‐Germain‐en‐Laye, France) under the following conditions: 3 min at 98°C, followed by 30 cycles of 30 s at 94°C, 40 s at 55°C, 20 s at 72°C, and a final extension of 7 min at 72°C. For gene expression analysis, total RNA was extracted from Arabidopsis rosettes using the phenol/chloroform method and precipitated with 8 M LiCl as described by Oñate‐Sanchez & Vicente‐Carbajosa (2008). Complementary DNAs were synthesized from 2 μg of RNA using the Revert Aid™ H Minus First Strand cDNA Synthesis Kit (Fermentas, Thermo Fisher, Waltham, MA, USA). RT‐qPCR was performed using LightCycler® 480 SYBR® Green I Master (Roche), a SYBR Green Detection System (Roche), and the LightCycler®480 Software release v.1.5.0 SP4 (Roche). mRNA quantification was expressed as relative expression levels ($2^{-\mathrm{\Delta }{C}_{\mathrm{t}}}$) or as fold change ($2^{-\mathrm{\Delta \Delta }{C}_{\mathrm{t}}}$) (Livak & Schmittgen, 2001). Arabidopsis ubiquitin 21 was used as a housekeeping control. For miRNA expression analysis, we employed a stem‐loop primer method to quantify miRNAs (Varkonyi‐Gasic, 2017). A pulsed reverse transcription reaction was conducted, involving steps at 16°C for 30 min, followed by 60 cycles at 30°C for 30 s, 42°C for 30 s, and 50°C for 1 min, and a final step at 85°C for 5 min. This procedure utilized the Revert Aid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA, USA) along with specific miRNA primers and oligo(dT). Primer sequences for both mRNA and miRNA analyses can be found in Supporting Information Table S1.

Cloning

MRT1 and MRT2 CDS fragments were amplified using conventional PCR with Phusion Hot Start II High‐Fidelity PCR Master Mix (Thermo Scientific™, Madrid, Spain). The PCR conditions for cDNA amplification were as follows: an initial denaturation at 98°C for 30 s, followed by 30 reaction cycles consisting of denaturation at 98°C for 10 s, primer annealing for 20 s, elongation at 72°C, and a final extension at 72°C for 5 min. Site‐directed mutations at the same position of the MRT1 and MRT2 TIR domains were carried out using the QuikChange II XL (Agilent, Madrid, Spain) site‐directed mutagenesis kit, following the manufacturer's instructions. The procedure alters nucleotide bases but preserves the original amino acid sequence. Once the desired sequences were obtained, they were cloned using NEBridge® Golden Gate Assembly in the pICH86966 backbone. To clone artificial microRNAs, the designer WMD3 provides four oligonucleotide sequences (I–IV). These sequences are employed to engineer the artificial microRNA into the endogenous miR319a precursor using site‐directed mutagenesis. The plasmid pRS300 and specific primer pairs for each phased secondary siRNA (phasiRNA) were used as templates for the PCRs. After obtaining the phasiRNA sequence within the pRS300 vector, it is transferred to the PBSK vector through PCR cloning with the assistance of restriction enzymes. The primers used for cloning are listed in Table S2.

Statistical analysis

Statistical analysis of the results was conducted using GraphPad Prism v.9.5.0; Dotmatics, Boston, MA, USA. Before performing the statistical analysis, we assessed the normality and homoscedasticity of the data. When the data met both assumptions, we employed a one‐way ANOVA followed by Tukey's multiple comparisons test. In experiments where both Row (R) and Column (C) were simultaneously analyzed, a two‐way ANOVA was performed, and Tukey's multiple comparisons test was utilized when the interaction (R × C) was significant. To assess the relationship between RNA‐seq and RT‐qPCR, we performed a Pearson product‐moment correlation test, drawing a linear regression line and calculating the R ² value to evaluate the goodness of fit. Student's t‐test was applied to compare two data sets. All statistical analyses are included in Table S2.

Results

Selection and characterization of plant receptor encoding genes upregulated by mites and insect infestation

Since receptors, either PRRs or NLRs, have a crucial role in the initial step of PTI and ETI immune responses, we aimed to identify and functionally characterize receptor genes involved in triggering defense responses to T. urticae, a mite that has been the focus of our research in recent years. We reanalyzed previous RNA‐seq data from mite‐infested Arabidopsis plants at 0.5, 1, 3, and 24 h post‐infestation (Garcia et al., ²⁰²¹; Santamaria et al., ²⁰²¹) and found a substantial set of upregulated genes encoding PRR and NLR receptors for each time point (Table S3). The highest number of induced genes for all receptor types was detected at 0.5 h of infestation. Of particular interest were the 21 TNL genes containing a TIR domain at the N‐terminal of the protein, which were induced by mites at the earliest post‐infestation time. The total number of TNLs upregulated was then reduced to 7, 6, and 8 TNL mite‐induced genes at 1, 3, and 24 h post‐infestation, respectively (Fig. 1a). Among these, we selected AT4G14370 and AT5G41550 genes, referred to from now on as MRT1 and MRT2 (mite‐related TNL1 and TNL2), respectively, for further assays. This selection was based on: (1) their early and strong upregulation in response to mites; (2) the TIR domain contains a highly conserved NAD⁺ catalytic site, where a glutamic acid residue serves as the key catalytic amino acid for the NAD⁺‐cleaving enzymatic activity, which is essential for TNL‐mediated immune function (Wan et al., ²⁰¹⁹); and (3) the presence of a miR825‐5p target sequence within the TIR motif, indicating a potential role for this miRNA in the regulation of mite‐related TNLs (Lopez‐Marquez et al., ²⁰²¹; Figs 1b, S1c).

Fig. 1.

Expression of receptor genes in Arabidopsis Col‐0 plants in response to Tetranychus urticae infestation at 0.5, 1, 3, and 24 h post‐infestation. (a) PRR, NLR, and TNL upregulated genes at different mite infestation times detected in RNA‐seq data from Santamaria et al. (2021). (b) Nucleotide sequence of the binding sites to miR825‐5p in the TIR domain of MRT1 and MRT2 genes. (c, d) MRT1 and MRT2 expression at different mite infestation times. The r value indicates the Pearson correlation between RNA‐seq and RT‐qPCR data.

A comparative analysis of the predicted MRT1 and MRT2 structure, general features, and sequences was done. We found that they belong to the Disease R‐genes family, with three primary domains: a Toll/interleukin‐1 receptor homology domain (TIR), an NB‐ARC domain (NBS), and a leucine‐rich repeat domain superfamily domain (LRR) (Fig. S1a,b). A closer look at their LRR domain revealed differences between MRT1 and MRT2. MRT1 had three LRR_8 subdomains and two LRR_4 subdomains, while MRT2 presented the same three LRR_8 subdomains and an additional subdomain labeled LRR_5. When comparing the full‐length proteins, MRT1 and MRT2 shared 50% identity, but when aligning both TIR domains, they exhibited over 80% identity (Fig. S1c). Additionally, as shown in Fig. 1(b), MRT1 and MRT2 shared a putative binding site to the miR825‐5p, located within their TIR domains, with a potential regulatory role in the post‐transcriptional regulation of these genes. Given the significant upregulation of MRT1 and MRT2 in response to mite infestation in RNA‐seq data, their expression patterns were validated by RT‐qPCR assays, as shown in Fig. 1(c,d).

Previous studies had identified a group of TNLs as putative targets of miR825‐5p (Niu et al., ²⁰¹⁶; Lopez‐Marquez et al., ²⁰²¹). Six of these TNLs – MRT1, MRT2, AT5G38850 (previously termed MIST1), AT4G08450, AT1G63740, and AT1G63730 – were highlighted due to their lower hybridization energies and better pairing than the others. The expression profiles of these six genes at 0.5 h post‐infestation by T. urticae and P. brassicae were analyzed using RT‐qPCR assays (Fig. 2a,b). After mite T. urticae infestation, only the MRT1 and MRT2 genes were significantly induced, while the expression patterns of the other four genes remained unchanged (Fig. 2a). Conversely, when analyzing the response to P. brassicae infestation, RT‐qPCR assays revealed that in addition to MRT1 and MRT2 being upregulated at 0.5 h post‐infestation, MIST1 and AT1G63740 were also significantly induced by this insect (Fig. 2b). In parallel, we investigated the level of miR825‐5p and that of its precursor (Pri‐miR825) at different post‐infestation time points. RT‐qPCR assays revealed that both the primary and mature forms of miR825‐5p were downregulated by T. urticae and P. brassicae infestation at 0.5 h (Fig. 2c,d). The levels of miR825‐5p and its precursor exhibited very similar profiles. These expression patterns were opposite to the TNL gene expression, showing a negative correlation between miR825‐5p levels and MRT1 and MRT2 mRNA accumulation throughout the infestation period. Specifically, miR825‐5p and Pri‐miR825 exhibited their lowest expression levels at 0.5 h after herbivore infestation and gradually increased at later time points, which contrasted with the early peak in MRT1 and MRT2 expression. These findings suggested the participation of these TNL receptors as potential targets of miR825‐5p in plant defense against diverse herbivores.

Fig. 2.

Gene expression profiles of TNL‐target genes and the miR825‐5p and PrimiR825 after herbivore infestation. (a) Relative expression levels of TNL genes in Col‐0 Arabidopsis plants after 0.5 h of mite infestation. (b) Relative expression levels of TNL genes in Col‐0 plants after 0.5 h of Pieris larvae infestation. TNL genes: MRT1 isoforms, AT4G14370.1, AT4G14370.2, and AT4G14370.3; MRT2: AT5G41550; MIST1, AT5G38850.1, and AT4G08450.1; AT4G08450.2; AT1G63730.1; AT1G63740.1; and AT1G63740.2. Data are mean ± SE of three replicates. Asterisks indicate significant differences with respect to control plants (P < 0.5, Student's t‐test). (c, d) Relative expression levels of miR825‐5p and its precursor PrimiR825 at different Tetranychus urticae (c) and Pieris brassicae (d) infestation time. Data are mean ± SE of three replicates. Asterisks indicate significant differences with respect to time 0 h (P < 0.5, two‐way ANOVA followed by Tukey's multiple comparison test).

MRT1 and MRT2 transcripts are targets of miR825‐5p

To determine whether miR825‐5p regulated MRT1 and MRT2 expression by targeting the sequence located at the end of the TIR domain of these TNL transcripts, we generated translational fusions of the two MRT genes to the YFP‐encoding reporter gene controlled by the strong and constitutive PvUbi1 promoter. Additionally, mutated versions lacking the putative miRNA cleavage site, without altering amino acid sequences, were created.

Transient co‐expression assays were performed by agroinfiltration in N. benthamiana leaves, expressing these constructs either alone or in combination with miR825‐5p or the unrelated miR319, as schematized in Fig. 3(a). As shown in Fig. 3(b,c), YFP fluorescence, quantified every 12 h during the first 48 h post‐infiltration, displayed a constant emission when MRT1 and MRT2 were expressed alone or in combination with the nontarget miR319, while the co‐expression of MRT1 or MRT2 with miR825‐5p exhibited a clear decrease in YFP fluorescence from the earliest tested time point (12 h). YFP emission was drastically reduced at longer incubation times. This reduction in YFP levels was not observed in leaves co‐expressing the mutated MRT1 and MRT2 constructs with miR825‐5p or with the unrelated miR319 (Fig. 3d–f). In addition, transcript levels of the fluorescent protein YFP decreased when miR825‐5p was co‐infiltrated with MRT1 and MRT2 constructs (Fig. S2a,b).

Fig. 3.

Quantification of fluorescence of MRT1‐YFP and MRT2‐YFP constructs in Nicotiana benthamiana plants expressing them alone or with miRNAs. (a) Scheme of the experimental setup. (b, c) Quantification of fluorescence emitted by N. benthamiana agroinfiltrated with MRT1 (wt‐MRT1) and MRT2 (wt‐MRT2) alone or combined with miR825‐5p or miR319 at different expression times. (d) Nucleotide sequence of the mutated binding site to miR825‐5p in MRT1 and MRT2. (e, f) Quantification of fluorescence emitted by N. benthamiana agroinfiltrated with mutated MRT1 and MRT2 (mut‐MRT1 and mut‐MRT2) alone or combined with miR825‐5p or miR319 at different expression times. Data are mean ± SE of 24 biological replicates. Asterisks indicate significant differences with respect to control at each point (P < 0.5, two‐way ANOVA followed by Tukey's multiple comparison test).

These results demonstrate that miR825‐5p recognizes the complementary sequence in MRT1 and MRT2 transcripts, thereby modulating TNL protein accumulation. The specificity of this interaction is further supported by the lack of effect when using mutated constructs or an unrelated microRNA.

Regulatory mechanisms of miR825‐5p and MIST1‐derived phasiRNAs on MRT1 and MRT2 expression

According to Lopez‐Marquez et al. (2021), miR825‐5p possesses the ability to produce phasiRNAs derived from MIST1 transcript processing, to act in a homology‐dependent manner, targeting other TNLs to amplify the silencing effects. Therefore, an additional negative feedback mechanism mediated by these phasiRNAs could be involved in the Arabidopsis–Pieris interplay, as is indicated in the scheme presented in Fig. 4(a).

Fig. 4.

Quantification of fluorescence of MRT1‐YFP and MRT2‐YFP constructs and their mutated versions in Nicotiana benthamiana plants expressing alone or with phasiRNAs derived from miR825‐5p‐MIST1 transcripts. (a) Scheme of the participation of MST1‐derived phasiRNAs processing in the post‐transcriptional silencing of MRT1, MTR2, and MIST1 to control plant defense against Pieris brassicae. (b, c) Quantification of fluorescence emitted by N. benthamiana agroinfiltrated with MRT1 (wt‐MRT1) and MRT2 (wt‐MRT2) alone or combined with their specific phasiRNAs (phasiR263‐1, ‐2, and ‐9) at different expression times. (d, e) Quantification of fluorescence emitted by N. benthamiana agroinfiltrated with phasi‐mutated MRT1 and MRT2 (phasimut‐MRT1 and phasimut‐MRT2) alone or combined with specific phasi (phasiR263‐1, ‐2, and ‐9) and miR825‐5p at different expression times. Data are mean ± SE of 24 biological replicates. Asterisks indicate significant differences with respect to control at each point (P < 0.5, two‐way ANOVA followed by Tukey's multiple comparison test).

In this context, it is important to recall that MIST1 was activated, besides MRT1 and MRT2, in Arabidopsis in response to this feeder. Thus, we proceeded to clone three phasiRNA specific sequences derived from MIST1, termed phasiR263‐1, phasiR263‐2, and phasiR263‐9 (Lopez‐Marquez et al., ²⁰²¹). These phasiRNAs were selected because they were predicted to target MRT1 and MRT2 and were identified through sequencing of small RNAs co‐immunoprecipitated with AGO1/2, which supports a functional role (Lopez‐Marquez et al., ²⁰²¹).

Transient co‐expression assays in N. benthamiana leaves were done with the constructs fused to the fluorescent YFP reporter gene, monitoring the expression dynamics of YFP fluorescence at 12 h intervals during the first 48 h post‐infiltration.

A continuous level of fluorescence over time was detected when MRT1 and MRT2 were expressed individually, either in their nonmutated or mutated forms (Fig. 4b–e). Notably, in leaves co‐expressing MRT1 or MRT2 with each one of the three selected MIST1‐derived phasiRNAs, YFP fluorescence significantly diminished at different time points (Fig. 4b,c). This decrease in YFP levels was not observed in leaves co‐expressing MRT1 and MRT2 with the mutated target sequence for each of the phasiRNAs used in the corresponding co‐expression assay, showing only a fluorescence reduction when they were combined with miR825‐5p (Fig. 4d,e). These results demonstrate that phasiRNAs derived from MIST1 transcripts specifically target the complementary sequence in MRT1 and MRT2 and consequently reinforce silencing and regulate TNL levels.

MRT1 and MRT2 are involved in Arabidopsis defense against T. urticae and P. brassicae infestation

To investigate the biological role of MRT1 and MRT2 in the plant response to both T. urticae and P. brassicae, Arabidopsis Col‐0 and two T‐DNA insertion lines for MRT1 and two for MRT2 genes, mrt1‐1, mrt1‐2, and mrt2‐1, mrt2‐2, respectively (Fig. S3a), were selected for further studies. All mutant lines were analyzed, and results from RT‐qPCR assays showed a significant reduction of the transcript content of MRT1 and MRT2 genes in the corresponding mutant lines, indicating that the four mutants were knockdown lines (Fig. S3b). Moreover, after 0.5 h of mite infestation, mRNA levels of each gene were much higher in the infested complementary mutant lines than in infested Col‐0 plants (Fig. S3c). These data indicated the functional redundancy of both genes and how expression of each MRT gene is upregulated to compensate for the absence of the other. Likewise, the expression levels of miR825‐5p were also determined in infested mrt1, mrt2, and Col‐0 plants, but no significant alterations in any of the genotypes were observed (Fig. S3d).

Arabidopsis mutant lines for MRT1 and MRT2, and Col‐0 plants were used to perform infestation bioassays. Unexpectedly, after 4 d of mite feeding, the damage of the rosette was significantly lower in all mutant lines than in Col‐0 plants (Fig. 5a,c), raising the possibility that compensatory upregulation may be increasing defenses in the single‐mutant lines. These differences in leaf damage correlated with a higher number of nuclei detected by DAPI staining, which indicated cell viability (Fig. S4a,d). Moreover, higher levels of H₂O₂ deposition, expressed as DAB relative units, were detected in the mutant lines in comparison with Col‐0 plants (Fig. S4b,e). The cell death area and the electrolyte leakage determined in leaves after 24 h of mite infestation were much greater in mrt1 and mrt2 lines than in Col‐0 plants (Figs 5b,d, S4c,f). To better visualize the biological relevance of these parameters, Fig. S4(g,h) provides a summarized comparative overview across the different genotypes. Additionally, the effect of the five Arabidopsis genotypes on T. urticae was evaluated by measuring mite mortality and fecundity after feeding on mutants and Col‐0 plants. The percentage of mite mortality increased when they fed on mutant lines, except for the mrt1‐1 line, and the cumulative number of eggs was slightly greater in the knockdown lines than in Col‐0 plants (Fig. 5e).

Fig. 5.

Plant damage, cell death of Col‐0 and mrt1 and mrt2 mutant lines, and mite fitness after feeding on the five Arabidopsis genotypes. (a, c) Foliar damage quantified in Col‐0 and mrt mutant lines after 4 d of mite infestation. (b, d) Cell death quantified in Col‐0 and mrt mutant lines after 24 h of mite infestation. (e) Cumulative number of mite eggs measured 36 h after infestation with synchronized mite females, and mite mortality quantified after 10 d. Data are mean ± SE of nine (a, c), 10 (b, d), and six (e) replicates. Different letters indicate significant differences (a–d) and asterisks indicate differences from Col‐0 genotype (e) (P < 0.5, one‐way ANOVA followed by Tukey's multiple comparison test).

To broaden our understanding of plant–herbivore interactions and to investigate whether the observed effects were specific to T. urticae or extended to other herbivores, we conducted bioassays with P. brassicae larvae on the same Arabidopsis mutant lines. In this case, feeding bioassays on Arabidopsis mrt1 and mrt2 mutant lines resulted in a smaller remaining area of leaf disks compared to Col‐0 plants, indicating higher damage, as expected (Fig. 6a,b). Concomitantly, a significant increase in larval weights was observed in the mrt1‐1 and mrt2‐2 mutant lines (Fig. 6c). The damage data were opposite to those resulting from mite infestation in the same mutant lines (Fig. 5a,c) and suggested strong evidence for the different role of MRT genes in plant defense against each of these herbivores.

Fig. 6.

Plant damage and larval weight after Pieris brassicae feeding on Col‐0 and mrt1 and mrt2 single mutant lines. (a, b) Remaining leaf tissue area after 8 h of Pieris larvae feeding on Col‐0 and mrt1 and mrt2 mutant lines. (c) Pieris larval weight after 24 h of feeding on Col‐0 and mrt1 and mrt2 mutant lines. Data are mean ± SE of 12 (a, b) and 24 (c) replicates. Different letters indicate significant differences (P < 0.5, one way ANOVA followed by Tukey's multiple comparison test).

Taking into account the compensatory upregulation of the reciprocal gene in each of the mutant lines, and their relevance for the plant responses against T. urticae and P. brassicae, we were motivated to generate double mrt1‐mrt2 mutant lines by crossing single mrt1‐2 and mrt2‐1 lines to further confirm and elucidate the defense role of both MRT genes against herbivores. Once the mrt1‐mrt2 double mutant lines were genotyped (Fig. S5a), we validated that expression of both MRT1 and MRT2 genes using RT‐qPCR assays which were knocked down (Fig. S5b). Despite the significant role that these genes play in plant defense, no discernible differences in terms of growth or development were observed when comparing double mrt1‐mrt2 mutant lines with Col‐0 plants under normal growth conditions (Fig. S5c). The absence of phenotypic differences in nonstressed conditions further highlights the specific role of these genes in defense responses rather than in general plant development, underscoring the importance of studying their function in the context of herbivore attacks. Considering the compensatory effect previously observed between the MRT1 and MRT2 genes in single mrt1 and mrt2 mutants, and to rule out any potential compensatory effect on the expression of additional miR825‐5p TNL targets, we analyzed the basal expression levels of four further TNL genes (MIST1, AT4G08450, AT1G63740 and AT1G63730) in the mrt1‐mrt2 double mutants. RT‐qPCR assays showed no significant alterations in the expression of any of these genes (Fig. S5d–g). Additionally, since MIST1 is also induced by Pieris feeding, we measured MIST1 gene expression in Col‐0 and double mrt mutant lines after P. brassicae infestation. As expected, RT‐qPCR assays showed that MIST1 was induced by the chewing larvae in all genotypes, but the levels were much higher in double mutant lines than in Col‐0 plants (Fig. S6). These results confirm the regulatory role of these TNLs in the plant‐Pieris relationship.

Feeding bioassays in mrt1‐mrt2 double mutant lines displayed significantly more damage produced by mite infestation than the Col‐0 plants (Fig. 7a). Mite fecundity analysis revealed that Col‐0 plants supported the lowest egg production, while increased mite fecundity was observed after feeding on both mrt1‐mrt2 lines. Interestingly, despite differences in egg numbers, statistical analysis showed no significant variation in mortality rates across genotypes (Fig. 7b). As expected, feeding experiments with Pieris larvae revealed decreased resistance in mrt1‐mrt2 double mutant plants compared to Col‐0. mrt1‐mrt2 mutant lines showed significantly smaller remaining leaf area after Pieris feeding (Fig. 7c,d), indicating increased tissue consumption by the chewing larvae. Additionally, larvae feeding on mrt1‐mrt2 plants exhibited substantially greater weights compared to those on Col‐0 (Fig. 7e). These findings demonstrated the participation of MRT1 and MRT2 in plant defense against diverse herbivores. The higher susceptibility of mrt1‐mrt2 mutants to both mites and lepidopteran larvae suggested a synergistic effect of these genes in the defense response. MRT1 and MRT2 likely work together to enhance plant resistance against different types of herbivores.

Fig. 7.

Plant damage of mrt1‐mrt2‐double mutant lines and Col‐0 plants, and herbivore fitness after feeding on the three genotypes. (a) Foliar damage quantified in Col‐0 and double mrt mutant lines after 4 d of mite infestation. (b) Cumulative number of mite eggs measured 36 h of infestation with synchronized mite females, and mite mortality quantified after 10 d. (c, d) Remaining leaf tissue area after 8 h of Pieris larvae feeding on Col‐0 and mrt1‐mrt2 mutant lines. (e) Pieris larval weight after 24 h of feeding on Col‐0 and mrt1‐mrt2 mutant lines. Data are mean ± SE of nine (a), six (b), 12 (c, d) and 24 (e) replicates. Different letters indicate significant differences (a, d, e) and asterisks indicate differences from Col‐0 genotype (b) (P < 0.5, one way ANOVA followed by Tukey's multiple comparison test).

The potential role of MRT1 and MRT2 in the plant defenses against T. urticae and P. brassicae motivated us to determine the expression of key defensive genes like PR1 and VSP2, salicylic acid (SA) and jasmonic acid (JA) responsive markers, respectively, in noninfested control and double mutant plants by RT‐qPCR assays. In control conditions, the expression of both genes did not show significant differences between genotypes (Fig. S7a,b). By contrast, PR1 and VSP2 genes were induced 24 h after mite or Pieris attack, but displayed lower mRNA levels in mrt1‐mrt2 double mutant than in Col‐0 (Fig. S7c–f). These data confirm the prompting of signaling defense pathways, downstream and dependent on MRT1 and MRT2 gene levels, and corroborate the crucial role of JA and SA hormones in plants against herbivores.

miR825‐5p differentially regulates defenses against T. urticae and P. brassicae

In addition to the experiments carried out with all mrt mutant lines, we also used Arabidopsis transgenic silencing (STTM825‐5p) or overexpressing (amiR825‐5p) plants, previously generated by Lopez‐Marquez et al. (2021), to confirm the role of miR825‐5p in plant defenses. Analysis of miR825‐5p expression in these lines confirmed the overaccumulation of miR825‐5p in the overexpressing lines. However, miR825‐5p levels were significantly reduced 24 h after mite feeding (Fig. S8a). No significant alterations were detected in those conditions in infested silenced lines, likely owing to the low basal miR825‐5p levels displayed in these lines. Nonetheless, cell death quantified by trypan blue staining was significantly higher in the miR825‐5p silenced lines than in Col‐0, but was not significantly altered in the overexpressing lines (Fig. S8b). When the accumulation of MRT1 and MRT2 transcripts was measured in these lines, significantly higher levels of both these mRNAs were detected after mite feeding independently of miR825‐5p presence (Fig. S9a,b).

Mite feeding bioassays in STTM825‐5p and amiR825‐5p plants demonstrated that silenced miR825‐5p plants were more resistant to mite infestation, as shown by the reduction in damaged area and the decrease in cumulative number of eggs (Fig. 8a–c). Nonetheless, no differences in leaf damage were observed in infested overexpressing lines, although the oviposition rates were also reduced in these transgenic lines (Fig. 8a–c). Feeding experiments conducted with Pieris larvae showed that miR825‐5p‐overexpressing plants were less resistant to Pieris infestation than miR825‐5p‐silencing lines or Col‐0 plants, since the remaining leaf tissue after feeding was significantly lower (Fig. 8d,e). Surprisingly, a significant decrease in larval weights was detected after feeding on overexpressing plants (Fig. 8f). These findings highlight the complex role of miR825‐5p in regulating plant defenses against different herbivores.

Fig. 8.

Plant damage of Col‐0 and miR825‐5p silencing and overexpressing lines and herbivore fitness after feeding on the five Arabidopsis genotypes. (a, b) Foliar damage quantified in Col‐0 and miR825‐5p mutant lines after 4 d of mite infestation. (c) Cumulative number of mite eggs measured 36 h of infestation with synchronized mite females. (d, e) Remaining leaf tissue area after 8 h of Pieris larvae feeding on Col‐0 and miR825‐5p mutant lines. (f) Pieris larval weight after 24 h of feeding on Col‐0 and mutant lines. Data are mean ± SE of nine (a), six (b), 12 (c, d), and 24 (e) replicates. Different letters indicate significant differences (P < 0.5, one‐way ANOVA followed by Tukey's multiple comparison test).

Discussion

Numerous examples of miRNAs regulating gene expression in primary and secondary metabolism, processes associated with plant growth, development, and responses to abiotic and biotic stresses, have been identified (Sesic et al., ²⁰²¹; Samynathan et al., ²⁰²²; Lopez‐Marquez et al., ²⁰²³; Raza et al., ²⁰²³). By contrast, remarkably little information is available on plant miRNA roles in response to herbivores. This information is particularly focused on the plant–insect relationship, linked to the identification of differentially expressed miRNAs and their target genes, mostly derived from transcriptomic data (Li et al., ²⁰¹⁸; Razna & Gagan, ²⁰¹⁹; Kumar et al., ²⁰²²; Malhotra et al., ^2022a,b). Moreover, transcriptomic analyses of whole rosettes may dilute localized responses, as only part of the tissue is affected. This spatial variability can mask relevant expression changes, so statistical thresholds should be balanced with biological relevance. To our knowledge, no studies have been published regarding miRNA‐mediated regulatory events in the plant–acari interplay.

In this study, for the first time, R‐genes (TNL receptor type) involved in the defense response against phytophagous arthropods have been identified and characterized. We focused our attention on TNLs induced at early stages of T. urticae infestation in Arabidopsis plants. Of particular interest were 21 TNLs that were early upregulated after mite infestation, but with reduced expression at later time points, as expected for most receptor‐encoding genes. Two of these genes, AT4G14370 and AT5G41550, referred to as MRT1 and MRT2, were selected for a deep functional analysis within the plant–spider mite context, since both showed significant induction at 0.5 h post‐mite feeding and contain a binding site for miR825‐5p. The function of mR825‐5p has previously been described as a negative regulator in Bacillus cereus elicited systemic resistance to Botrytis cinerea, as well as a downregulator of basal immunity against Pseudomonas syringae by targeting MIST1 receptor, both in Arabidopsis (Nie et al., ²⁰¹⁹; Lopez‐Marquez et al., ²⁰²¹). Little is known about the function of MRT1 and MRT2, and most of the available information is putative or based on predictions. These genes have been associated with plant immunity, stress responses, and genomic regions enriched in defense‐related duplications. MRT1 has been linked to cell death across different accessions (Barragan et al., ²⁰²⁰). In particular, interactions between ACD6 alleles and AT4G14370 trigger temperature‐dependent immune activation and growth defects (Todesco et al., ²⁰¹⁴).

Still more interesting was the finding that MRT1 and MRT2 were also induced by the chewing larvae P. brassicae, but in this interaction, additional R‐genes such as MIST1 and AT1G63740 were also upregulated at this early time‐point. This suggests that distinct herbivore feeding strategies may elicit partially overlapping but specialized sets of immune receptors. This differential activation implies a context‐dependent deployment of TNLs tailored to the nature and intensity of the herbivore challenge. All mentioned TNLs had in common a target sequence of the miR825‐5p within their TIR motif, which has been demonstrated to determine TNL regulation for MIST1 in the plant‐pathogenic bacteria context, suggesting a more general role in TNL regulation (Niu et al., ²⁰¹⁶; Lopez‐Marquez et al., ²⁰²¹). TNLs are key players in immune defense, and their expression is induced upon feeder detection. Once these receptors activate due to insect/acari infestation, an ETI response is initiated. Phytophagous feeders not only prompt defenses but also trigger the accumulation of specific miRNAs to specifically modulate the plant immune response. We observed a negative correlation between miR825‐5p levels and the accumulation of MRT1, MRT2, and MIST1 mRNAs either during T. urticae or P. brassicae infestation, suggesting a finely tuned regulated mechanism that balances immune activation and growth. Upon mite attack, this regulation became more evident at time points beyond the initial 0.5 h. This regulation is particularly critical given that excessive activation of immune responses can lead to detrimental effects on plant fitness, as well as high costs for the plant. Furthermore, mutual regulatory relationships between arthropods and plants are produced, in which miRNAs from both interaction sides regulate feeder adaptation to plant hosts, and plant resistance to herbivores (Wu et al., ²⁰¹⁷; Garcia et al., ²⁰²¹; Kumar et al., ²⁰²²; Zhang et al., ²⁰²⁴).

Results from transient co‐expression assays using YFP‐fused MRT1 and MRT2 demonstrated that miR825‐5p targets complementary sequences in the TIR domain of both genes, thereby modulating TNL‐YFP protein accumulation. This specificity was further supported by the lack of effect when using mutated constructs or an unrelated miRNA. In this scenario, it was important to decipher whether phasiRNAs brought a second wave of TNL regulation since some reports had shown that miRNA825‐5p triggers the production of phasiRNAs from MIST1 transcripts (Chen et al., ²⁰¹⁰; Niu et al., ²⁰¹⁶; Cai et al., ²⁰¹⁸; Lopez‐Marquez et al., ²⁰²¹). PhasiRNA production amplifies and enhances the silencing signal, with significant implications. These mobile secondary silencing agents enable systemic gene regulation, allowing a single miRNA to influence multiple genes across the plant. The mechanism facilitates coordinated responses to environmental stresses and potentially drives the evolution of new regulatory networks in plants. Lopez‐Marquez et al. (2021), using effective experimental approaches, confirmed that miRNA825‐5p MIST1 target site activated functional phasiRNAs capable of trans gene silencing of a second wave of TNLs in response to the bacterial flagellin peptide flg22. These data clearly established a link between MAMP's perception and TNL gene expression. Moreover, phasiRNAs can migrate, extending the silencing, cis to trans‐targeting, to distal tissues by the well‐known noncell‐autonomous mechanism (De Felippes, 2019). Here, we demonstrate that miRNA825‐5p‐triggered MIST1‐generated phasiRNAs also activate a second layer of regulation to reinforce the silencing of MRT1 and MRT2. These phasiRNAs derived from MIST1 transcripts processing amplify trans silencing of MRT1 and MRT2, having an important effect on resistance to P. brassicae but not against T. urticae because MIST1 is not induced by mite attack. Therefore, the release of miRNA/phasi‐TNL repression prompts an increase in TNL levels, contributing to enhancing ETI responses, particularly modulating defenses against specific feeders (Zhai et al., ²⁰¹¹; Lopez‐Marquez et al., ²⁰²¹). Chewing insects rapidly consume leaf tissue, producing greater and quicker plant damage than sucking mites, so they require additional control systems to avoid injury and tissue losses.

Our results showed that single mrt1 and mrt2 mutant lines presented less leaf damage after mite feeding than Col‐0 plants, but an increase in mite fecundity and mortality was detected after feeding on single mutant lines. These data seemed to be contradictory, but the comparative analysis of single mrt1 and mrt2 mutants revealed a significant functional redundancy between both genes, pushing us to generate double mrt1/2 mutant lines. As expected, suppression of MRT1 and MRT2 in double mrt1/2‐mutants increased their susceptibility to mites, reflected by a larger leaf damage, and the consequent reduction in mite mortality and increase in fecundity rates. Likewise, the leaf tissue consumed by P. brassicae larvae was greater in mrt1/2 double mutant lines than in Col‐0 plants, concomitantly with an increase in larval weight. These findings evidenced that both MRT genes were essential for a robust defense mechanism. Notably, P. brassicae triggers additional TNLs beyond MRT1 and MRT2, so double mutants lacking two defense layers were more vulnerable than miR825‐5p overexpression lines, which still retain a broader defense arsenal despite modulation. Moreover, while miR825‐5p overexpression impacts TNL‐mediated responses, other immune pathways not regulated by this miRNA remain active. The synergistic effect suggested that MRT1 and MRT2 might operate in concert to enhance resistance against a range of herbivores, reflecting the complexity of plant immune responses.

In addition, feeding experiments with T. urticae and P. brassicae on Arabidopsis silencing miRNA825‐5p lines showed reduced leaf damage, suggesting that MRT1, MRT2, and MIST1 mRNAs were not inhibited, allowing defence signaling pathways to be activated and plant defences to be mounted. These data confirmed the specificity of this miRNA on the targeted TNLs and its functional ability as a modulator of Arabidopsis defences to fight against these two herbivores. Interestingly, the amount of remaining tissue determined in miRNA825‐5p‐overexpressing plants after P. brassicae attack was reduced, indicating that plant defenses were not sufficient to control the chewer's feeds. In the Pieris‐plant scenario, since MIST1 and AT1G63740 were also induced by larval feeding, both NLs likely participate in initiating the immune signal cascade upon attack. Collectively, these results underscore the essential roles of MRT1 and MRT2, and the complementarity of other TNLs in the plant resistance to phytophagous feeders, with their absence disrupting defence gene expression – particularly in SA and JA signaling pathways – and compromising the plant's ability to defend effectively. While defences against generalist phytophagous species like the T. urticae mite are primarily regulated by jasmonates and/or SA (Lortzing et al., ²⁰¹⁷; Santamaria et al., ²⁰¹⁷; Rosa‐Diaz et al., ²⁰²³), defenses against the specialist herbivore P. brassicae are predominantly driven by SA (Lariviere et al., ²⁰¹⁵; Lortzing et al., ²⁰¹⁷).

The significance of these R‐genes to herbivores in the absence of stress requires a precise and fine regulation to balance their synthesis. Unregulated expression of R‐genes can potentially lead to autoimmunity, further impeding plant growth. It is also well known that HR is a trait associated with plant defense against pests, and often governed by individual R‐genes. In our work, we demonstrated an increase in HR, linked to cell death in single mrt mutant plants. This phenomenon appears to be regulated constantly, even in nonstress conditions. Although miR825‐5p and its TNL targets are not the only components involved in plant immunity, they form part of a finely tuned defense system. By directly regulating the intracellular receptors MRT1, MRT2, and MIST1, miR825‐5p contributes to ETI and mediates species‐specific defense responses against herbivores with distinct feeding modes. Our findings elucidate the regulatory role of this miRNA–TNL module in Arabidopsis defense and highlight the importance of understanding these networks to improve crop resilience. Further studies will be essential to unravel the herbivore‐specific signaling pathways that shape the complexity of plant immune responses.

Competing interests

None declared.

Author contributions

ID and IR‐D conceived the research. IR‐D, ID, DL‐M and CRB performed the experimental research and/or contributed to the analysis and interpretation of data. ID and IR‐D wrote the first draft of the manuscript. All authors contributed to the final version of the manuscript.

Disclaimer

The New Phytologist Foundation remains neutral with regard to jurisdictional claims in maps and in any institutional affiliations.

Supporting information

Fig. S1 MRT1 and MRT2 TNL protein features, structure, and miR825‐50 binding site.

Fig. S2 Expression levels of YFP in MRT1‐YFP and MRT2‐YFP in Nicotiana benthamiana plants.

Fig. S3 Molecular characterization of Arabidopsis mrt1 and mrt2 mutant lines.

Fig. S4 DAPI/nuclei damage relation, H₂O₂ content, and membrane depolarization quantified after 24 h of Tetranychus urticae feeding in mrt1‐ and mrt2‐ single mutant plants.

Fig. S5 Molecular characterization of Arabidopsis mrt1‐mrt2 mutant lines.

Fig. S6 MIST1 gene expression in Col‐0 and double mrt1‐mrt2 mutant lines.

Fig. S7 Gene expression analysis of PR1, VSP2 genes in mrt1‐mrt2 double mutant lines and Col‐0 plants.

Fig. S8 Gene expression analysis of miR825‐5p, MRT1, and MRT2 genes in miR825‐5p‐overexpressing and ‐silencing lines and Col‐0 plants, and cell death determination.

Fig. S9 Gene expression analysis of MRT1 and MRT2 genes in miR825‐5p‐overexpressing and ‐silencing lines and Col‐0 plants.

Table S1 Oligonucleotide sequences.

Table S2 Statistical test applied indicating the corresponding figures.

Table S3 List of upregulated TNL, NLR, and PRR receptor genes from Arabidopsis plants at 0.5, 1, 3, and 24 h of mite post‐infestation, derived from RNA‐seq analyses.

Please note: Wiley is not responsible for the content or functionality of any Supporting Information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.

Data availability

All data are available within the article and its Datasets: Tables S1–S3. Sequence data used in this article can be found in the GenBank/EMBL data libraries for A. thaliana under the following accession nos.: MRT1: AT4G14370; MRT2: AT5G41550; MIST1: AT5G38850; AT4G08450; AT1G63740; AT1G63730; AT2G14610; AT5G24770.