Spider mite herbivory induces an ABA-driven stomatal defense

Abstract Arthropod herbivory poses a serious threat to crop yield, prompting plants to employ intricate defense mechanisms against pest feeding. The generalist pest 2-spotted spider mite (Tetranychus urticae) inflicts rapid damage and remains challenging due to its broad target range. In this study, we explored the Arabidopsis (Arabidopsis thaliana) response to T. urticae infestation, revealing the induction of abscisic acid (ABA), a hormone typically associated with abiotic stress adaptation, and stomatal closure during water stress. Leveraging a Forster resonance energy transfer (FRET)-based ABA biosensor (nlsABACUS2-400n), we observed elevated ABA levels in various leaf cell types postmite feeding. While ABA's role in pest resistance or susceptibility has been debated, an ABA-deficient mutant exhibited increased mite infestation alongside intact canonical biotic stress signaling, indicating an independent function of ABA in mite defense. We established that ABA-triggered stomatal closure effectively hinders mite feeding and minimizes leaf cell damage through genetic and pharmacological interventions targeting ABA levels, ABA signaling, stomatal aperture, and density. This study underscores the critical interplay between biotic and abiotic stresses in plants, highlighting how the vulnerability to mite infestation arising from open stomata, crucial for transpiration and photosynthesis, reinforces the intricate relationship between these stress types.


Introduction
Plants and phytophagous arthropods have coevolved in an "arms race" of mutual antagonism between herbivory strategies and plant defenses.Successful feeding by insect and acari pests on compatible crop plants causes substantial yield losses and represents a threat to global food security.Successful plant defense starts in part with the molecular recognition of herbivore-associated molecular patterns (HAMPs) and damage-associated molecular patterns (DAMPs).Through these early warning cues, plants prompt signal transduction pathways and transcriptional reprogramming, resulting in chemical and physical defenses (Stahl et al. 2018;Garcia et al. 2021).This cascade of events requires fine-tuned control mediated by multiple regulatory factors to induce defense regimes appropriate to herbivore species, plant host, and environmental conditions.Common chemical responses to herbivore damage include the accumulation of reactive oxygen species (ROS), calcium, and several hormones considered to be core modulators of immune signaling, i.e. jasmonic acid (JA), jasmonoyl-isoleucine (JA-Ile), and salicylic acid (SA; Erb and Reymond 2019).These responses are also shared with many plant-pathogen interactions (Bari and Jones 2009).Physical defenses against pests vary considerably; for example, lignified and suberized root and stem barriers, aerial trichomes, and epidermal cuticles can all restrict access to feeding (War et al. 2012;Holbein et al. 2019).Induced closure of stomata, as natural entry sites to the leaf interior, has received considerable attention for foliar pathogen responses (Melotto et al. 2008(Melotto et al. , 2017)).On the other hand, pathogen-induced closure of stomata can also contribute to water soaking and microbial pathogenesis later in infection (Hu et al. 2022;Lajeunesse et al. 2023).Stomatal closure has also been observed in response to phytophagous arthropods (Sances et al. 1979;Pincebourde and Casas 2006;Schmidt et al. 2009), though whether closure constituted a defense mechanism or contributed to the infestation remained unclear.
Abscisic acid (ABA) is the key hormone regulating stomatal aperture, which gates gas exchange for transpiration, photosynthesis, and responses to abiotic stimuli (Kuromori et al. 2018) and is accumulated in response to a number of environmental stimuli, including drought, salt, and wounding (Christmann et al. 2006).Both chewing and piercing-sucking arthropod pests provoke loss of water, an increase in transpiration, and a reduction in stomata conductance.All these phenomena are associated with water stress-induced accumulation of ABA and subsequent stomatal closure (Lim et al. 2015).ABA-deficient mutants are more susceptible to chewing herbivory by beet armyworm (Spodoptera exigua; Thaler and Bostock 2004) and African cotton leafworm (Spodoptera littoralis) caterpillars (Bodenhausen and Reymond 2007), suggesting ABA signaling and potentially stomatal closure could contribute to plant defense.Open stomata are natural apertures where piercing-sucking herbivores like aphids and mites insert their stylets, specialized feeding structures, to access nutrient-rich subepidermal cells.Although stylets can also be inserted between epidermal cells, feeding via opening stomata is potentially faster and avoids the damage of cell walls, which can trigger rapid plant defenses (Mathen et al. 1988;Bensoussan et al. 2016).In some cases, stomata are even preferred places to lay herbivore eggs (DeClerck and Steeves 1988).On the other hand, the caterpillar Helicoverpa zea was found to actively induce stomatal closure to reduce the release of herbivore-induced plant volatiles that can recruit H. vea predators (Lin et al. 2021).More direct regulation of ABA by piercing-sucking herbivores is also possible as the saliva of the aphid Myzus persicae upregulated genes that also responded to ABA treatment (Hillwig et al. 2016).In this interaction, ABA contributes to susceptibility to aphid infestation, indicating that the role of ABA signaling in herbivory defense could also be negative.
ABA is perceived by PYRABACTIN RESISTANCE (PYR)/PYR-like (PYL) or REGULATORY COMPONENT OF ABA RECEPTOR (RCAR) family receptor proteins (Gonzalez-Guzman et al. 2012).The accumulation of ABA in stomatal guard cells, for example, during water stress, is transduced via PYR/PYL/RCAR receptors to a complex signaling network (Ma et al. 2009).Core components include inhibitory PROTEIN PHOSPHATASE 2C (PP2C)-type proteins as well as positive regulators of ABA signaling including SUCROSE NON-FERMENTING 1-RELATED PROTEIN KINASE 2 (SnRK2)-type protein kinases as well as ABRE-BINDING FACTOR (ABF) and ABA Insensitive (ABI)-type transcription factors (Hsu et al. 2021).ABA effects comprise changes in redox homeostasis, guard cellspecific kinase regulation, and cytosolic Ca 2+ levels.These events lead to stomatal closure while regulating numerous nonstomatal functions (Neill et al. 2008;Dinh et al. 2013;Ng et al. 2014).ABA regulation of these components also occurs in many plantpathogen interactions, and as for plant-herbivore interactions, ABA has been found to both promote and antagonize microbial defenses (Bari and Jones 2009;Lim et al. 2015).
One explanation for contrasting roles for ABA in diverse plant defense responses is ABA crosstalk with ROS, calcium, and defense hormone signaling, for example, through inhibition of SA signaling (De Torres-Zabala et al. 2007;Ton et al. 2009).ABA activation of primed JA-regulated defense responses in Arabidopsis (Arabidopsis thaliana) contributes to induced resistance against small white (Pieris rapae) caterpillar herbivory (Vos et al. 2013).ABA can also regulate JA responses via direct interaction of PYL5/6 ABA receptors with MYC transcription factors (Aleman et al. 2016), and ABA amplifies JA-dependent defense responses as part of the signal transduction pathway in plants elicited by oral secretions of insects (Dinh et al. 2013).Thus, ABA-JA and ABA-SA hormonal crosstalk could also be important for ABA responses in the plant-herbivore interplay.With respect to stomatal aperture, SA and methyl-JA can promote stomatal closure while other jasmonates, particularly the bioactive JA-Ile, can induce stomatal opening (Munemasa et al. 2011), though the contribution of SA and JA signaling might be secondary to ABA signaling (Zamora et al. 2021).
Most reports on the interaction between plants and arthropod herbivores have focused on insects, but mites are also highly economically relevant agricultural pests (Vacante 2016).Among them, the 2-spotted spider mite (Tetranychus urticae), an acari of the Tetranychidae family, is one of the most polyphagous pests found worldwide, feeding on more than 1,100 documented host plants, of which about 150 are crops (Migeon and Dorkeld 2023).T. urticae pierces individual parenchymatic cells with stylets specialized for sucking their content.Infestation with T. urticae causes leaf chlorosis, and severe infestation causes substantial crop losses (Bensoussan et al. 2016).In addition to its economic importance, T. urticae has a sequenced genome and a number of available tools and protocols (Grbic ́ et al. 2011;Cazaux et al. 2014;Suzuki et al. 2017;Ojeda-Martinez et al. 2020).These characteristics and its ability to feed on the A. thaliana model plant led to T. urticae becoming a model herbivore for plant-pest interactions.However, potential positive or negative roles for ABA accumulation or signaling, including stomatal closure, in the context of plant defense against mite herbivores remain largely unexamined.How this key plant hormone that integrates a myriad of environmental cues impacts a generalist plant herbivore is relevant for a wider understanding of the plant-pest interplay.
Here, we have analyzed the capacity of T. urticae to trigger stomata closure at the feeding site and on nearby undamaged tissues using A. thaliana as a host plant.We also studied stomatal behavior and hormone crosstalk during mite infestation.We have demonstrated the accumulation of ABA and characterized this accumulation at cellular resolution using the nuclear-localized Forster resonance energy transfer (FRET) biosensor for ABA nlsABACUS2-400n (Rowe et al. 2023).Mite feeding bioassays in mutants for ABA production and perception, as well as in lines with altered stomata density, have established the biological relevance of stomatal control, which are entry gates for mite stylets, for plant defenses in the plant-mite interaction.

T. urticae infestation induces stomata closure and modifies hormone content
To study time-resolved dynamics of defense and stomatal regulation upon mite infestation, the stomatal aperture on the abaxial leaf surface was analyzed in epidermal samples at four infestation time points (Fig. 1A).Leaf epidermal impressions showed open stomata during the day with subtle reductions in aperture in wild-type (WT) Col-0 noninfested plants toward the end of the light period (Fig. 1A, Supplementary Table S1).Mite infestation induced a striking stomatal closure, reaching the maximum closure at 24 to 30 h postinfestation (Fig. 1, A and B).To explore whether natural stomatal closure in darkness could alter mite feeding habits and resultant leaf cell death, plants were grown in regular conditions, infested, and incubated under a constant dark or light environment for 16 h, and then cell damage was assayed.Results demonstrated that mites fed better under light conditions, when stomata are open, as plants suffered increased cell death (Fig. 1C).Nonetheless, this could also be due to other lightdriven changes, such as plant hormone, carbohydrate, ROS and calcium levels, or changes in mite behavior.
Since hormone accumulation in response to mites might be responsible for phenotypic changes in stomatal behavior, SA, JA, JA-Ile, and ABA were quantified in mite-infested Arabidopsis Col-0 leaves at 4 postinfestation time points.The accumulation of all 4 compounds was induced by mites but with distinct temporal profiles.SA content significantly increased from the earliest infestation time, while JA, JA-Ile, and ABA required longer times to be differentially accumulated in comparison to noninfested plants (Fig. 2, A to D).The highest hormone levels were detected at 24 h of infestation, except for the JA-Ile, which reached the highest content at 30 h of infestation, in accordance with a reduction in its JA precursor at this time point.Interestingly, ABA levels were maintained at 30 h, while SA was reduced.To determine the impact of SA, JA, and ABA on leaf stomatal aperture, these hormones were exogenously applied to Arabidopsis Col-0 plants.SA and ABA treatment triggered stomatal closure in treated leaves, as expected (effect size in Supplementary Table S2), though the effect of JA was negligible (Fig. 2, E to G).

ABA accumulation was detected upon mite infestation
Since ABA accumulates in leaves during mite infestation concomitantly with stomatal closure, we sought to identify in which cells ABA could be functionally relevant for Arabidopsis defense using the high-resolution nlsABACUS2-400n FRET-based biosensor for ABA (Rowe et al. 2023).In leaves of Arabidopsis plants expressing the biosensor, nuclei of epidermal and internal cell types showed sufficient biosensor expression for segmentation (Fig. 3A, upper images).In all detected cell types, nuclear ABACUS2-400n fluorescence emission ratios were higher in infested plants than in controls (Fig. 3, A, bottom images, and B), indicating higher ABA concentrations.Mean emission ratio for all cells was significantly higher for mite-infested leaf images (Fig. 3B).To better understand the cellular distribution of ABA accumulation, we used a nuclear morphology classifier based on the "FRETENATOR" tool to analyze cell type emission ratios.Nuclei were classified into 5 cell types: stomata, pavement, spongy mesophyll, bundle sheath, and vascular bundle cells.With infestation, the signal increased in all cell type groups, but stomatal and  S5.Numbers indicate significant differences between control and mite treatment at each point of infestation.Data are means ± SE of 28 biological replicates.B) Image of stomata closure 24 h after mite infestation.Bars = 8 µm.C) Cell death, measured in leaf disks by trypan blue staining after 16 h of mite infestation under dark and light conditions, is expressed in square millimeters.A two-tailed t test was used to test differences due to light treatments (P < 0.05).Data are means ± SE of 10 biological replicates.
vascular bundle cell groups showed the highest emission ratios (Fig. 3, C and D).Because T. urticae cause cell death by inserting their stylets either in between epidermal pavement cells or through stomatal pores to suck mesophyll cell contents (Supplementary Fig. S1A; Bensoussan et al. 2016), it was important to get information about the viability status of leaf cells after the infestation.The ABA biosensor allowed us to detect cell viability since nuclei detected by biosensor fluorescence reflect viable cells.The comparison between infested and noninfested plants showed significant changes in the total number of quantified nuclei (Supplementary Figs.S1B and S2A).However, when this comparison was independently done in each of the 5 cell type groups, the mesophyll cell type presented a lower number of nuclei in the infested plants than in noninfested (Supplementary Fig. S1, C to G).Taken together, these results corroborated that spongy mesophyll cells were the target tissue where mites suck nutrients upon their feeding and indicated that nlsABACUS2-400n biosensor was able to detect local ABA Quantification of hormone content in Arabidopsis Col-0 upon mite infestation and stomata aperture after exogenous hormonal treatments.A) JA, B) JA-Ile, C) SA, and D) ABA accumulation in whole plants was quantified at 3, 8, 24, and 30 h after mite infestation.Values are expressed as nanograms of hormone per gram of fresh weight (FW).Significant factors (SF) indicate whether the 2 independent factors, R (infestation time) and C (mite treatment), and/or their interaction I (R × C) were statistically significant (two-way ANOVA followed by Tukey's test, P < 0.05).Detailed ANOVA results are available in Supplementary Table S5.Numbers indicate significant differences compared to control conditions.Data are means ± SE of 3 pools of 6 biological replicates.Stomata aperture measured in detached leaves at 3 h after spraying on the aerial part of the plant with E) 1 mM SA, F) 1 mM JA, and G) 10 µM ABA.Results represented as width/length ratio.Student's t test was used to assess differences due to hormone treatments (P < 0.05).Data are means ± SE of 10 biological replicates.
ABA's protective role against spider mites | 2973 increases in response to mites in leaf cell types directly involved in feeding as well as in distal cells and cell types not directly involved in feeding, i.e. bundle sheath and vascular cells.Interestingly, vascular cells have been proposed to be the major sites of foliar ABA biosynthesis (Endo et al. 2008), and thus, our data provide support for a model in which mite feeding triggers ABA synthesis in the vasculature and this results in the translocation of ABA across the leaf cell types.

ABA is involved in plant defense against mites
To study how changes in endogenous ABA levels affect mite infestation, Arabidopsis mutants in ABA biosynthesis and catabolism, aba2-1 and cyp707a1cyp707a3, respectively, which result in lower and higher ABA concentrations (González-Guzmán et al. 2002;Okamoto et al. 2009), were selected to perform mite infestation.aba2-1 lines showed significantly greater damage than control plants upon mite feeding for 4 d while the damaged area in cyp707a1cyp707a3 line was lower than in Col-0 plants (Fig. 4A).Likewise, fecundity rates, measured as cumulative number of eggs after 36 h, were higher when mites fed on the mutant aba2-1 mutant line than when they fed on Col-0 plants and were significantly reduced in the cyp707a1cyp707a3 line (Fig. 4C).These data indicated that ABA promotes plant defenses against mite damage and antagonizes mite oviposition.Additionally, 2 Arabidopsis ABA insensitive mutants, the 112458 ABA receptor sextuple mutant and ost1-3 SnRK2 kinase mutant, were also examined to determine if the canonical ABA signaling pathway A) Z-Projections of analyzed images.Upper images correspond to nuclei expressing the biosensor and bottom images to FRET emission ratios in detached leaves.Black areas represent areas excluded from analysis during segmentation.B) Quantification of nlsABACUS2-400n emission ratios in multiple leaves.Student's t test was used to assess differences due to control and mite conditions (P < 0.05).Data are means of 8 biological replicates.C) Emission ratios of nlsABACUS2-400n biosensor in cells of stomatal, pavement, spongy mesophyll, bundle sheath, and vascular bundle cells, after 24 h of mite infestation in detached leaves.Significant factors (SF) indicate whether the 2 independent factors, R (mite treatment) and C (nucleus type), and/or their interaction I (R × C) were statistically significant (two-way ANOVA followed by Tukey's test, P < 0.05).Detailed ANOVA results are available in Supplementary Table S5.Data are means of 15 biological replicates.D) Increased levels of emission ratio in the different 5 cell types after infestation when compared to noninfested plants.
participated in ABA-dependent mite responses.The OST1 gene is preferentially expressed in guard cells and the vasculature (Mustilli et al. 2002), cell types that exhibited the most pronounced increase in ABA levels (Fig. 3).While mite females laid a similar number of eggs in mutants as in control plants, both mutant lines showed more damage than Col-0 plants after infestation (Fig. 4, B and D).Overall, these results revealed that ABA signaling, including in the vasculature or stomata, is important for mite defense.Because all cells respond with increased ABA, we sought to determine whether ABA accumulation further impacted SA and JA, the core hormones involved in plant defenses.We examined the expression of PR1 and MYC2, marker genes of SA and JA signaling, respectively, in the 5 Arabidopsis genotypes at 2 infestation times.The expression of the PR1 gene was induced at 8 h postinfestation and decreased 24 h after mite feeding in all studied genotypes.However, PR1 levels were lower in the aba2-1 line and higher in the cyp707a1cyp707a3 line than in Col-0 plants at the 8-h time point and displayed the opposite expression pattern at 24 h (Supplementary Fig. S3A).No differences in PR1 behavior were found between 112458, ost1-3 mutants, and WT plants (Supplementary Fig. S3B).Together, these results indicate that ABA could crosstalk with SA signaling positively early in mite infection and negatively later in infection.In contrast to PR1, which peaked at 8 h, MYC2 presented the maximum induction at 24 h after infestation (Supplementary Fig. S3, C and D).In the 5 tested genotypes, only the aba2-1 mutant line showed altered MYC2 expression, i.e. elevated at 24 h (Supplementary Fig. S3, C and D).The observation that the reduced ABA levels in the aba2-1 mutant increased expression of both PR1 and MYC genes at 24 h could indicate hormone crosstalk but could also be an indirect result of increased mite infestation in aba2-1.Certainly, the overaccumulation of MYC2 transcript suggests that positive crosstalk of ABA with JA signaling is unlikely to explain the role of ABA in mite resistance.
One physiological response associated with ABA is the production of ROS, particularly H 2 O 2 , a signaling molecule that regulates plant stress response genes (Li et al. 2022).To further confirm such an association, the H 2 O 2 content was determined in detached leaves of the 4 Arabidopsis mutants and Col-0 plants after 24 h of mite infestation (Supplementary Fig. S4).The measurement of H 2 O 2 concentration is represented as relative DAB staining units (Supplementary Fig. S4A) and demonstrated that aba2-1 and cyp707a1cyp707a3 infested lines accumulated higher and lower levels of H 2 O 2 , respectively, than Col-0 plants.In contrast, the number of DAB deposits in the ost1-3 mutant was only slightly reduced and was not altered in 112458 after mite feeding (Supplementary Fig. S4).As with JA signaling, a negative association between ABA levels and ROS production suggests ROS accumulation is unlikely to explain the role of ABA in resistance to mite infestation.

The effects of exogenous ABA application on mite infestation
As demonstrated above, exogenous ABA application induces stomata closure (Fig. 2G).To elucidate if ABA treatment affected plant defenses, we pretreated nlsABACUS2-400n plants by spraying ABA on the leaves 3 h before mite infestation and analyzed cellular ABA levels and leaf responses to mites.As expected, ABA pretreated and mite-infested plants displayed higher nlsABACUS2-400n biosensor emission ratios, indicating higher ABA levels, and the highest emission ratios were observed in those plants that were first pretreated and then infested (Fig. 5A).Both ABA pretreatment and mite infestation triggered stomata closure, with no synergistic or additive effects observed when the treatments were combined (Fig. 5B).The exogenous application of ABA before mite infestation helped plant defenses since they showed less damage than nonpretreated plants and provided positive effects on plant growth, as shown in measurements of the rosette area (Fig. 5, C to E).Thus, ABA pretreatment enhanced mite defenses and favored plant growth during infestation.

Stomatal aperture determines the mite infestation outcome
As ABA accumulation and stomata closure are induced in response to mites, we questioned whether increased stomata aperture could provide any advantage for mite infestation.To clarify this point, Col-0 plants were pretreated with fusicoccin (FC), a  S5.
compound that promotes stomata opening (Hunt et al. 2010), with ABA to close stomata, or both.After treatments, stomatal aperture and plant damage were measured, and results were compared with mock-treated and noninfested plants.Plants pretreated either with ABA or FC presented closed and opened stomata, respectively, and the combination of both produced a stomatal aperture comparable to mock plants (Fig. 6, A and B).An experiment using different FC + ABA ratios demonstrated that very low concentrations of FC (0.5 µM) were sufficient to keep stomata open (Supplementary Fig. S5).Upon mite infestation, stomata remained closed in ABA-treated plants, remained open in FC-treated plants, and were actively closed similarly to mock when ABA and FC were previously combined (Fig. 6B).Plant damage results showed that more open stomata in the FC treatment correlated with higher leaf damaged area, indicating that open stomata facilitated mite feeding (Fig. 6C).As leaf damage increased in plants subjected to combined ABA and FC treatments compared with the ABA treatment alone, the control over stomatal aperture function of ABA is likely key for resistance to mite infestation.

Stomatal density affects mite infestation
Given the importance of stomatal closure as a defense mechanism against mite infestation, we investigated the potential involvement of stomatal density in the leaf as a contributing factor using Arabidopsis lines with higher and lower numbers of stomata than Col-0 plants.The selected plants were mutants in EPIDERMAL PATTERNING FACTOR (EPF), a family of secreted peptides that inhibit stomatal development (Hepworth et al. 2015).In particular, we selected the epf1epf2 mutant line and EPF2OE, an EPF2 overexpressing line.First, stomatal density was measured prior to mite infestation (Fig. 7A), and as expected, the epf1epf2 mutant exhibited an elevated stomatal count, while the EFF2OE line displayed a reduced number of stomata.Then, we evaluated stomatal aperture in the 3 Arabidopsis genotypes with and without mite infestation and found that stomatal aperture in EPF2OE plants was higher than in epf1epf2 or Col-0 plants (Fig. 7B).Interestingly, infested epf1epf2 mutant exhibited somewhat lessened stomatal closure than Col-0, though infestation closed stomata in all genotypes (Fig. 7B).As stomata allow gas exchange between leaf mesophyll cells and atmosphere, contributing to leaf cooling (Chowdhury et al. 2021), we also analyzed whether the leaf temperature was dependent on the stomata number and if temperature variations were produced during mite infestation.Uninfested temperature of epf1epf2 leaves was 2.2°C less than Col-0 and 3.1°C less than EPF2OE leaves (Supplementary Table S3, Fig. 7C), consistent with stomatal density measurements.After mite infestation, an expected increase in leaf temperature was detected in Col-0 (+2.2°C) and, to a lesser extent, in EPF2OE (+1°C) plants that have fewer stomata.Mite infestation augmented leaf temperature in ep-f1epf2 (+3.0°C), the mutant line with more stomata, although infested leaves remained cooler than in infested Col-0 and EPF2OE (Supplementary Table S3, Fig. 7C).These temperature data corroborated the interrelationship between stomatal aperture and density in the plant response to mite infestation.Finally, the damage produced by mites (Fig. 7D) was quantified in these Arabidopsis plants.The epf1epf2 line displayed more damage than Col-0 plants, while mite damage was significantly reduced in the EPF2OE line.The close correlation between stomatal density and damage (R 2 = 0.9920; Fig. 7E) indicates that the density of stomata in leaves is an essential feature that determines the success of mite infestation.S5.Data are means of 5 biological replicates.C) Plant damage in whole plant of ABA and/or FC pretreated plants after 4 d of mite infestation.Numbers indicate significant differences compared to mock treatment.One-way ANOVA followed by Tukey's multiple comparison test, P < 0.05.Detailed ANOVA results are available in Supplementary Table S5.Data are means of 8 biological replicates.

Discussion
the content of ABA, considered the key hormone that closes stomata, although other hormones also participate in governing stomata responses (Wei et al. 2021).Under a subset of biotic stresses, ABA levels and stomata behavior play an essential plant defense role by hindering adversary access to the leaf interior (Lim et al. 2015;Melotto et al. 2017).In the plant-mite context, stomata are target sites where mites insert their stylets to access nutrient-rich mesophyll cells.Once the plant perceives the mites, a signal transduction cascade activates the synthesis of JA and SA, besides other chemical responses, to generate defenses (Zhurov et al. 2014;Santamaría et al. 2019).Our results confirmed that mite infestation induced JA and SA, demonstrated the timeline of this response, and revealed a mite-induced increase of ABA.Previous studies have established connections between stomatal movement, light response, circadian clock pathways, and pathways influenced by light changes, such as ROS (Rossel et al. 2002;Tallman 2004).We found that T. urticae caused less cell damage by feeding under dark conditions when stomata were closed.Concomitantly to stomatal closure, ABA accumulates in response to herbivory during the day.Additionally, the temporal pattern of the phenotypic changes in stomatal behavior in response to mite infestation during the day matched the ABA accumulation more than the JA or SA content.Together, these data pointed to ABA accumulation and stomatal closure as a plant response to herbivory acting to limit leaf cell damage associated with mite infestation.Stomatal closure has been previously described as a plant response to other phytophagous species (Sances et al. 1979;Pincebourde and Casas 2006;Schmidt et al. 2009), though whether closure constituted a defense mechanism or contributed to the infestation remained unclear.DNA microarray experiments of

D E
7. Stomatal density in epf1epf2, an EPF2 overexpressing line (EPF2OE), and Col-0 plants and determination of damage, stomata aperture, and temperature in the 3 genotypes after mite infestation.A) Stomatal density of Col-0, epf1epf2, and EPF2OE expressed per square millimeter from detached leaves.Numbers indicate significant differences compared to Col-0 genotype.Data are means of 5 biological replicates.One-way ANOVA followed by Tukey's multiple comparisons test, P < 0.05.Detailed ANOVA results are available in Supplementary Table S5.B) Stomatal aperture represented as width/length ratio in detached leaves of Col-0, epf1epf2, and EPF2OE under mite infestation or control.Data are means of 6 biological replicates.One-way ANOVA followed by Tukey's multiple comparisons test, P < 0.05.Detailed ANOVA results are available in Supplementary Table S5.C) Temperature in degree Celsius measured in whole plants of Col-0, epf1epf2, and EPF2OE under mite infestation or control.Significant factors (SF) indicate whether the 2 independent factors, R (mite treatment) and C (genotype), and/or their interaction I (R × C) were statistically significant (two-way ANOVA followed by Tukey's multiple comparison test, P < 0.05).Detailed ANOVA results are available in Supplementary Table S5.Data are means of 6 biological replicates.D) Foliar damage in whole plant, expressed in square millimeters, was quantified 4 d after mite infestation of Col-0, epf1epf2, and EPF2OE.Data are means of 7 biological replicates.One-way ANOVA followed by Tukey's multiple comparisons test, P < 0.05.Detailed ANOVA results are available in Supplementary Table S5.E) Correlation between stomatal density and damaged area for the 3 Arabidopsis genotypes (Pearson's product moment R 2 = 0.992).
Arabidopsis infiltrated with M. persicae saliva allowed the identification of a number of aphid saliva upregulated genes that also responded to ABA treatment (Hillwig et al. 2016).Aphid infestation induced ABA production in Arabidopsis; however, aphids preferred and performed better on WT plants than on ABA-deficient mutants, suggesting that ABA might not be a plant defense response.These authors also demonstrated that ABA increased the synthesis of some glucosinolates with defensive properties.We demonstrated a positive role for ABA in the plant defense to the mite T. urticae since mutant lines in ABA biosynthesis and catabolism showed higher and lower leaf damage, respectively, than Col-0 infested plants.These data directly corresponded to the higher and lower number of eggs accumulated in these mutants.Thus, mite infestation results in ABA accumulation as a plant defense response that closes stomata, reduces mite feeding ability, lessens leaf damage, and lowers mite fecundity rates.
Which mite-produced signal is detected and how this promotes protective ABA accumulation remains an open question.Plant damage, mite behavior, feeding vibrations, and microbes associated with the herbivore may elicit defense responses (Santamaria et al. 2020), and wounding has previously been shown to induce ABA accumulation (Zhang et al. 2021).Salivary proteins, feces, egg silks, and molts may provide HAMPs detected by plant pattern recognition receptors (PRRs).Of the characterized HAMPs, 2 proteins, Tetranin1 (Tet1) and Tetranin2 (Tet2), are candidates expressed in the salivary glands of T. urticae (Iida et al. 2019).Tet1 and Tet2 infiltration protects multiple plant species against herbivory, activating ROS and calcium responses and, in Phaseolus vulgaris, accumulation of several plant hormones (Iida et al. 2019).
Several possible biochemical mechanisms could integrate defense signaling into the ABA accumulation we observe during T. urticae infestation.De novo biosynthesis or deconjugation of the ABA glucosyl ester (ABA-GE) into bioactive ABA could increase the bioactive ABA pool, as could decreased catabolism or ABA repartitioning via mite-induced transporter dynamics.Our previous transcriptomics study of a mite infestation time series demonstrated early induction of several ABA-related enzymes and transporters (Santamaria et al. 2021), including 2 biosynthetic 9-CIS-EPOXYCAROTENOID DIOXYGENASE enzymes (NCED5 at 1 h and NCED3 at 3 h; Tan et al. 2003) and an endoplasmic reticulum (ER)-localized ABA-GE deconjugation BETA-GLUCOSIDASE enzyme (BGLU18 at 1, 3, and 24 h; Lee et al. 2006).Wound-induced leaf ER bodies containing BGLU18 have previously been shown to be critical for antiherbivory compound production, suggesting a possible dual role in mite defenses (Nakazaki et al. 2019).Key ABA catabolic enzymes, CYTOCHROME P450 707A and CYP707A1-A3, were induced by herbivory at all time points, suggesting that reduced catabolism is unlikely to drive ABA defense.Rather, as CYP707A enzymes are induced by ABA in the absence of water stress (Kushiro et al. 2004), their induction is consistent with negative feedback on ABA accumulation.Several ABA transporters also show a marked increase in expression during mite infestation at 1 or more time points.These include DETOXIFICATION EFFLUX CARRIER 50 (DTX50) and ABCG25, efflux carriers expressed in the vasculature (Kuromori et al. 2010;Zhang et al. 2014), and importers ABCG40 and ABCG18 that are expressed in the guard cells and mesophyll, respectively (Kang et al. 2010).Together, these results indicate that biosynthesis, deconjugation, and repartitioning from veins to other leaf cells could all contribute to the cellular ABA dynamics we observe in mite defense.
Vos et al. ( 2013) also reported that ABA is a crucial regulator of induced resistance against P. rapae by activating primed JA-regulated defense responses in Arabidopsis.Additionally, ABA has been associated with redox homeostasis (Li et al. 2022) and ROS accumulation, which is also essential in plant defense against mites (Santamaría et al. 2017;Arnaiz et al. 2021).Our results showed increased levels of JA signaling and H 2 O 2 production in ABA-deficient mutants after mite infestation, where leaf damage was increased.Based on the current observations, it can be concluded that ABA signaling, independent of positive crosstalk with ROS and JA signaling, plays a critical role as a defense determinant against mite infestation.It remains to be seen how calcium signaling, which is often associated with ROS production (Takahashi et al. 2011), stress responses (Gilroy et al. 2014), and ABA (Edel and Kudla 2016) might interact with ABA-induced stomatal defense during T. urticae infestation.
To probe the regulation and function of ABA accumulation, we used nlsABACUS2, a FRET-based biosensor for ABA, as a tool to quantify in vivo ABA at high spatio-temporal resolution in response to mite infestation.Elevated levels of ABA were broadly detected in the nuclei of a range of leaf cell types, with slightly stronger responses in stomata and vascular tissues.These ABA dynamics are consistent with an ABA regulation model in which the leaf vasculature is the key site of biosynthesis triggered by mite infestation, with ABA subsequently transported to guard cells for stomatal closure, similar to that proposed for ABA during water stress by Kuromori et al. (2018).However, we also show induction of ABA in all leaf cells involved in feeding by mites, including pavement cells and the mesophyll cells on which they feed.
Given the pronounced stomatal closure observed during infestation, an additional significant aspect of this study focused on clarifying the impact of stomatal status versus ABA signaling more generally.This was achieved by administering exogenous ABA or FC to induce or repress stomatal closure.When pretreated with ABA and subsequently infested with mites, plants were protected without sacrificing their growth.Pretreatment with FC increased leaf damage, consistent with a protective role for stomatal closure.In an ABA + FC treatment, mite damage correlated with stomatal aperture rather than ABA levels, providing further support for the role of stomata in ABA-mediated mite defense.The closure of stomata likely hindered mite feeding through the natural openings on the leaf, compelling them to utilize their stylets for penetrating closed stomata or between the epidermal pavement cells.This mode of stylet penetration could trigger earlier damage and possibly defense signaling and likely reduces feeding success since mites try to avoid epidermal cell damage (Bensoussan et al. 2016), and we observed ABA to reduce overall leaf damage and defense signaling at 24 h infestation.
We also investigated the importance of leaf stomatal density upon infestation based on the hypothesis that a higher number of stomata could facilitate mite feeding.In mutants with more and fewer stomata, we observed lower and higher leaf temperatures that increased with mite feeding, indicating that mite-induced stomatal closure remained functional in leaves with altered stomatal density.Indeed, stomatal density was tightly correlated with mite-induced leaf damage, suggesting that the number of stomata, in conjunction with their aperture, is a highly relevant trait for plant resistance to mite infestation.
This study provides valuable insights into the important role of stomata and its regulation mediated by ABA in the Arabidopsis defense against T. urticae.We demonstrate that ABA, possibly of vascular origin, accumulates in the guard cells of the stomata as a result of mite feeding and triggers stomatal closure as a structural defense response.We also show that stomatal aperture and abundance play a crucial role in determining the success of an infestation.As both traits are regulated by ABA, the level of this phytohormone, which also integrates myriad other biotic and abiotic stress cues, is an important determinant of mite infestation that could be considered in breeding programs for pest control.

Spider mite growth, maintenance, and plant infestation
T. urticae London strain (Acari: Tetranychidae) population, provided by Dr. Miodrag Grbic (UWO, Canada), was reared on P. vulgaris (beans) and maintained in growth chambers (Sanyo MLR-351-H, Sanyo, Japan) at 25°C ± 1°C, >70% relative humidity, and a 16-h:8-h (light:dark) photoperiod.Mite infestation was performed in 3-wk-old Arabidopsis rosettes or detached leaves collected from the corresponding Arabidopsis genotypes with 20 mites/plant or 10 mites/leaf, adapting the infestation time according to the experiment.For leaf infestation assays, leaf number 5 or 6 was selected (Merchant and Pajerowska-Mukhtar 2015), placed on a one-half-strength MS (Duchefa Biochemie) medium plates, infested, and covered but ventilated with a relative humidity of 60% to 80%.

Hormone analysis
Frozen plant material (ca.50 mg) was used for the extractions.Plant samples were spiked with 25 µL of an internal standard mixture (containing ABA-d 6 , DHJA, and C 13 -SA) to correct for analyte losses (De Ollas et al. 2021).Extraction was carried out in ultrapure water in a ball mill at room temperature using 2-mm glass beads.Homogenates were centrifuged at 10,000 rpm for 10 min at 4°C, and supernatants recovered.The resulting solutions were partitioned twice against an equal volume of diethyl ether after adjusting pH to 3.0 with a 30% acetic acid solution.The combined organic layers were evaporated under vacuum in a centrifuge concentrator (Jouan, Sant Germaine Cedex, France), and the dry residues were reconstituted in a 10% (w/v) aqueous methanol solution.Prior to injection, extracts were filtered through 0.20-µm PTFE syringe membrane filters, and filtrates were recovered in chromatography amber glass vials.Samples were analyzed by tandem LC/MS in an Acquity SDS UPLC system (Waters Corp., United States) coupled to a TQS triple quadrupole mass spectrometer (Micromass Ltd., United Kingdom) through an electrospray ionization source.Separations were carried out on a C18 column (Luna Omega Polar C18, 50 × 2.1 mm, 1.6 µm particle size, Phenomenex, United States) using a linear gradient of ultrapure acetonitrile and water, both supplemented with formic acid to a 0.1% (v/v) concentration, at a constant flow rate of 0.3 mL/min.During analyses, column temperature was maintained at 40°C and samples at 10°C to slow degradation.Plant hormones were detected in negative electrospray mode following their specific precursor-to-product ion transitions and quantitated using an external calibration curve with standards of known amount.

Stomata leaf impression and aperture quantification
Stomata leaf impressions were made using detached leaves, with and without treatment and mite infestation (after a previous mite removal).Leaves were pressed on fast-setting dental resin AquasilUltra+ (Dentsply Sirona), which was allowed to set; then, the leaf material was removed.Transparent nail polish was spread onto the resin pieces (Wang et al. 2006).After they had thoroughly dried, nail varnish impressions were peeled using Sellotape and affixed to a glass microscope slide.These slides were then imaged on a Leica DM1000 LED with an ICC50 W camera. Stomatal density and aperture were quantified (Doheny-adams et al. 2012) with Fiji ImageJ software (Schindelin et al. 2012).

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 chloral hydrate (Sigma) solution (Sanchez-Vallet et al. 2010).Disks were placed onto glass slides in 50% (v/v) glycerol and observed under an epifluorescence stereoscope using UV filters.Pixel quantification was performed using Adobe Photoshop (Luna et al. 2012).

Exogenous plant pretreatments
The stock solutions for hormone/FC treatments were dissolved in 96% ethanol and diluted in distilled water to make pretreatment solutions.Mock solutions and pretreatment solutions were matched, so plants received 0.19% ethanol.Exogenous treatments were applied to the aerial part of the whole plant by spraying with 1 mM SA, 1 mM JA, 10 µM ABA, or 1 µM FC (all from Sigma) unless stated in the figure.After 3 h of treatment, mite infestation or stomatal leaf impressions were performed.The mock treatment involved the same volume of ethanol as the pretreatment.

Confocal microscopy and image processing
An inverted SP8 confocal microscope (Leica) was used for biosensor imaging assays in detached leaves.All images were acquired as Z-stacks in 16-bit mode, with a 10× dry objective.Samples were mounted in one-fourth MS, pH 5.7.Typical settings were as follows: sequential scanning was performed with excitation lasers and HYD detectors; 442 nm excitation 3% to 10% was used with HYD1 460 to 500 nm, 100 gain, for a first acquisition to detect the donor T7edCerulean fluorescence (donor excitation, acceptor emission or DxAm).Second, 442 nm excitation 3% to 10% was used with HYD2 525 to 560 nm, 100 gain, to acquire energy transfer fluorescence (donor excitation, donor emission or DxDm).Next, 514 nm excitation 5% to 10% was used with HYD2 525 to 560 nm, 100 gain, for a third acquisition to detect the edCitrineT7 acceptor protein fluorescence (acceptor excitation, acceptor emission or AxAm).The scan speed was set at 400; line averaging was 2 to 4, and bidirectional X was on.

FRET cell and tissue classification
The image processing for fluorescence emission ratio (DxAm/ DxDm) quantification was carried out using the "FRETNATOR" tool reported by Rowe et al. (2022Rowe et al. ( , 2023) ) (Supplementary Fig. S5B).The AxAm images were used to segment the nuclei.The specific cell type emission ratio classification was done using the FretCellType extension plugin.The extension was developed using Apache Groovy and reads each series of acquisitions (Leica file format [.lif]).FretCellType uses the "FRETENATOR_Segment_ and_ratio" output, specifically the "threshold image output" and "The label map," to classify the nuclei based on the ROI shape.Cell types were defined as vascular bundle, bundle sheath, spongy mesophyll, pavement, and stomata cells.The plugin provides 2 tables (.csv) for each image: the "Result Table " and the "Summary Table ."The "Result Table " displays the number of pixels, the coordinates, and the emission ratio value for each nucleus.Supplementary Fig. S2B provides information about the nuclei classification shape validation.

Plant damage and mite performance assessment
Chlorotic damage in whole plants was quantified 4 d after mite infestation, according to Ojeda-Martinez et al. (2020), using 9 biological replicates from independent rosettes for each genotype.Mite fecundity was determined using entire leaves detached from different Arabidopsis genotypes.Each leaf was infested with 10 synchronized females, and the number of eggs was counted after 36 h (Santamaría et al. 2019).

Nucleic acid analysis
Total RNA was extracted from Arabidopsis rosettes by the phenol/ chloroform method and precipitated with 8 M LiCl as described (Oñate-Sánchez and Vicente-Carbajosa 2008).Complementary DNAs (cDNAs) were synthesized from 2 μg of RNA using the Revert Aid H Minus First Strand cDNA Synthesis Kit (Fermentas).RT-qPCR was performed using LightCycler 480 SYBR Green I Master (Roche), a SYBR Green Detection System (Roche), and the LightCycler 480 Software release 1.5.0SP4 (Roche).mRNA quantification was expressed as relative expression levels (2 −dCt ) or as fold change (2 −ddCt ; Livak and Schmittgen 2001).Arabidopsis ubiquitin 21 was used as a housekeeping control gene and PR1 (Pathogenesis-Related protein-1) and MYC2 (MYC2 transcription factor) as marker genes of SA and JA hormone signaling, respectively.Primer sequences are stated in Supplementary Table S4.

Hydrogen peroxide determination
The accumulation of H 2 O 2 was visualized in leaf disks after 24 h of mite infestation, using 3,3-diaminobenzidine tetrachloride hydrate (DAB, Sigma) as a substrate, according to De Ilarduya et al. (2003).Disks were placed onto glass slides in 50% (v/v) glycerol and observed under a stereoscope.Pixel quantification was performed with Fiji ImageJ software (Schindelin et al. 2012).

Thermal imaging determination
Thermal images were obtained using a FLIR infrared camera (FLIR-T600) equipped with a 16° lens 24 h after whole plant infestation.To avoid changes in leaf temperature due to environmental conditions, experiments were conducted inside a walk-in growth chamber set at 25°C, 60 ± 10% relative humidity, and light intensity of 105 μmol/m 2 /s.The camera was vertically mounted at approximately 20 cm above the setup.Images were saved as 8-bit TIFF files, and pictures were analyzed using ImageJ (Fiji), in which all pictures were set at a constant range of temperature based on pictures of Col-0 control conditions; the emissivity of the samples was 0.925.

Statistical analysis
Statistical analyses were done using GraphPad Prism v9.4.1.The normality and homoscedasticity of the data were previously analyzed to apply the proper analysis.Student's t test was used for individual analysis.One-way ANOVA followed by Tukey's multiple comparisons test was used to compare multiple data sets.Two-way ANOVA was performed in the experiments in which row data (R) and treatment with column data (C) were simultaneously analyzed, and Tukey's multiple comparisons test was used when the interaction (R × C) was significant.The correlation between damage and stomata density was analyzed by applying Pearson product moment correlation test.Statistical tests applied are listed in Supplementary Table S5.

Figure 1 .
Figure 1.Effects of mite infestation on stomata aperture and on the leaf cell death when plants are incubated under light or dark conditions.A) Stomatal aperture measured in Arabidopsis Col-0 detached leaves after 3, 8, 24, and 30 h of mite infestation.Results referred as width/length ratio.Significant factors (SF) indicate whether the 2 independent factors, R (infestation time) and C (mite treatment), and/or their interaction I (R × C) were statistically significant (two-way ANOVA followed by Sidak's multiple comparison test, P < 0.05).Detailed ANOVA results are available in Supplementary TableS5.Numbers indicate significant differences between control and mite treatment at each point of infestation.Data are means ± SE of 28 biological replicates.B) Image of stomata closure 24 h after mite infestation.Bars = 8 µm.C) Cell death, measured in leaf disks by trypan blue staining after 16 h of mite infestation under dark and light conditions, is expressed in square millimeters.A two-tailed t test was used to test differences due to light treatments (P < 0.05).Data are means ± SE of 10 biological replicates.

Figure 3 .
Figure3.ABA accumulation at the cellular resolution in leaf tissues using the nuclear ABA biosensor plants nlsABACUS2-400n, after mite infestation.A) Z-Projections of analyzed images.Upper images correspond to nuclei expressing the biosensor and bottom images to FRET emission ratios in detached leaves.Black areas represent areas excluded from analysis during segmentation.B) Quantification of nlsABACUS2-400n emission ratios in multiple leaves.Student's t test was used to assess differences due to control and mite conditions (P < 0.05).Data are means of 8 biological replicates.C) Emission ratios of nlsABACUS2-400n biosensor in cells of stomatal, pavement, spongy mesophyll, bundle sheath, and vascular bundle cells, after 24 h of mite infestation in detached leaves.Significant factors (SF) indicate whether the 2 independent factors, R (mite treatment) and C (nucleus type), and/or their interaction I (R × C) were statistically significant (two-way ANOVA followed by Tukey's test, P < 0.05).Detailed ANOVA results are available in Supplementary TableS5.Data are means of 15 biological replicates.D) Increased levels of emission ratio in the different 5 cell types after infestation when compared to noninfested plants.

Figure 4 .
Figure 4. Plant damage and mite fecundity after mite infestation of aba2-1 and cyp707a1cyp701a3 and pyr1pyl-112458, ost1-3 Arabidopsis mutants, and Col-0 plants.A) Foliar damage quantified 4 d after mite infestation in whole plants in aba2-1, cyp707a1cyp707a3, and Col-0 plants and in B) 112458, ost1-3, and Col-0 plants.Data are expressed in square millimeters.Effects on T. urticae fecundity measured 36 h after the infestation with synchronized mite females in detached leaves on C) aba2-1 cyp707a1cyp707a3 and Col-0 plants and in D) 112458, ost1-3, and Col-0 plants.Numbers indicate significant differences compare to Col-0 genotype.Data are means for 10 biological replicates for damage area and 6 biological replicates for eggs.A to D) One-way ANOVA followed by Tukey's multiple comparisons test, P < 0.05.Detailed ANOVA results are available in Supplementary TableS5.

Figure 5 .
Figure 5.Effect of ABA pretreatment of nlsABACUS2-400n plants before mite infestation.A) Emission ratio of biosensor signal for detached leaves under combined 3 h ABA pretreatment and 24 h infestation or controls; t tests with a false discovery rate multiple testing correction were used.Numeric annotations represent q-values, n = 6.B) Stomatal aperture presented as the width/length ratio in detached leaves under 3 h ABA pretreatment and 24 h infestation or controls; t tests with a false discovery rate multiple testing correction were used.Numeric annotations represent q-values, n = 6.C) Damaged area in whole plants expressed in square millimeters with and without 3 h ABA pretreatment and 24 h infestation or controls.Student's t test was used.Number annotation represents P-values, n = 5.D) Whole plant growth quantified in pixels; t tests with a false discovery rate multiple testing correction were used.Number annotations represent q-values, n = 10.E) Infested Arabidopsis plant images of either 3 h ABA pretreatment and 24 h infestation or controls.Images show the digitally extracted plant rosettes used for quantification in D).

Figure 6 .
Figure 6.Effects of 10 µM ABA and/or 1 µM FC pretreatments on stomatal aperture and on Arabidopsis plant responses to mite infestation.A) Images of stomata behavior in ABA and/or FC pretreated plants.Bars 8 µm.B) Stomatal aperture in ABA and/or FC pretreated plants after 24 h of mite infestation in detached leaves.Significant factors (SF) indicate whether the 2 independent factors, R (mite treatment) and C (pretreatment), and/or their interaction I (R × C) were statistically significant (two-way ANOVA followed by Tukey's multiple comparison test, P < 0.05).Detailed ANOVA results are available in Supplementary TableS5.Data are means of 5 biological replicates.C) Plant damage in whole plant of ABA and/or FC pretreated plants after 4 d of mite infestation.Numbers indicate significant differences compared to mock treatment.One-way ANOVA followed by Tukey's multiple comparison test, P < 0.05.Detailed ANOVA results are available in Supplementary TableS5.Data are means of 8 biological replicates.