RBOHF activates stomatal immunity by modulating both reactive oxygen species and apoplastic pH dynamics in Arabidopsis

SUMMARY Stomatal defences are important for plants to prevent pathogen entry and further colonisation of leaves. Apoplastic reactive oxygen species (ROS) generated by NADPH oxidases and apoplastic peroxidases play an important role in activating stomatal closure upon perception of bacteria. However, downstream events, particularly the factors influencing cytosolic hydrogen peroxide (H2O2) signatures in guard cells are poorly understood. We used the H2O2 sensor roGFP2‐Orp1 and a ROS‐specific fluorescein probe to study intracellular oxidative events during stomatal immune response using Arabidopsis mutants involved in the apoplastic ROS burst. Surprisingly, the NADPH oxidase mutant rbohF showed over‐oxidation of roGFP2‐Orp1 by a pathogen‐associated molecular pattern (PAMP) in guard cells. However, stomatal closure was not tightly correlated with high roGFP2‐Orp1 oxidation. In contrast, RBOHF was necessary for PAMP‐mediated ROS production measured by a fluorescein‐based probe in guard cells. Unlike previous reports, the rbohF mutant, but not rbohD, was impaired in PAMP‐triggered stomatal closure resulting in defects in stomatal defences against bacteria. Interestingly, RBOHF also participated in PAMP‐induced apoplastic alkalinisation. The rbohF mutants were also partly impaired in H2O2‐mediated stomatal closure at 100 μm while higher H2O2 concentration up to 1 mm did not promote stomatal closure in wild‐type plants. Our results provide novel insights on the interplay between apoplastic and cytosolic ROS dynamics and highlight the importance of RBOHF in plant immunity.


INTRODUCTION
Plants are continuously exposed to diverse microorganisms and must develop strategies to avoid infection by restricting the entry and multiplication of pathogens.The main first line of plant defences against foliar bacteria is the closure of stomatal pores at the leaf epidermis.Stomata are formed by a pair of guard cells that mediate gas exchange between the plant and the environment and are critical during the plant's innate immune response (Melotto et al., 2006;Melotto et al., 2008).Stomatal immunity is initiated by the recognition of pathogen-associated molecular patterns (PAMPs) by pattern recognition receptors at the plasma membrane.In Arabidopsis thaliana, the receptor FLAGELLIN-SENSITIVE2 (FLS2) that recognises the peptide flg22, the biologically active epitope of the bacterial PAMP flagellin, plays a prominent role in stomatal immunity (Melotto et al., 2006;Zeng & He, 2010).Downstream of PAMP perception, calcium, reactive oxygen species (ROS), nitric oxide and apoplastic alkalinisation function as secondary messengers to promote stomatal closure (Arnaud & Hwang, 2015).Stomatal immunity was also shown to be modulated by hormonal pathways, mainly the abscisic acid (ABA) and salicylic acid (SA) signalling pathways (Arnaud & Hwang, 2015;Melotto et al., 2008).To counteract PAMP-mediated stomatal closure, pathogens have evolved virulence factors, such as the phytotoxin coronatine (COR) secreted by the bacterial pathogen Pseudomonas syringae pv tomato (Pst) strain DC3000, to counteract host stomatal defences by promoting stomatal reopening (Melotto et al., 2006).
During the immune response, the apoplastic ROS burst is mainly generated by the NADPH oxidase isoform RBOHD and to a lesser extent RBOHF (Torres et al., 2002;Zhang et al., 2007).While RBOHD participates in stomatal (pre-invasive) defence responses, its contribution to plant resistance against Pst bacteria is not clear (Chaouch et al., 2012;Kadota et al., 2014;Zhang et al., 2007).Moreover, the role of RBOHD in PAMP-triggered stomatal closure has been questioned (Guzel Deger et al., 2015;Khokon et al., 2010).On the contrary, RBOHF is required for full post-invasive (late apoplastic) resistance against virulent Pst bacteria (Chaouch et al., 2012) and no role in stomatal defence has been described until now.The type III cell wall peroxidases PRX33 and PRX34 also participate in PAMP-mediated ROS production and play an important role in both pre-and post-invasive defences against Pst bacteria (Arnaud et al., 2017;Daudi et al., 2012).Although it is unknown how PAMPs activate PRXs, apoplastic alkalinisation is a prerequisite for PRX-mediated apoplastic ROS production (Bolwell et al., 2002).Apoplastic ROS, mainly hydrogen peroxide (H 2 O 2 ), diffuse into the cytosol through plasma membrane aquaporins (Rodrigues et al., 2017;Tian et al., 2016).However, recent results indicate that PAMPand bacteria-induced oxidation of the cytosolic/nuclear localised H 2 O 2 sensor roGFP2-Orp1 in leaf discs is largely independent of RBOHD, RBOHF, PRX33 and PRX34 (Arnaud et al., 2023).Moreover, it remains unclear how apoplastic ROS produced by NADPH oxidases or cell wall peroxidases upon PAMP perception influences intracellular H 2 O 2 /ROS accumulation specifically in guard cells.
In this contribution, we used three different probes to analyse oxidative events after flg22 treatment.The chemiluminescent luminol and fluorescein H 2 DCFDA probes detect, respectively, extracellular and intracellular ROS and are not specific to H 2 O 2 (Murphy et al., 2022).In contrast, the ratiometric biosensor roGFP2-Orp1 localised in the cytosol and nucleus is mainly oxidised by H 2 O 2 in vivo and can be reduced by the glutaredoxin/glutathione system permitting dynamic and real-time measurements (Arnaud et al., 2023;Nietzel et al., 2019).We show that PAMP-triggered oxidation of roGFP2-Orp1 is surprisingly stronger in guard cells of the rbohF, prx33-3 and prx34-2 mutants compared to WT plants.Despite an increase in H 2 O 2 in guard cells upon PAMP perception, we found a weak negative correlation between stomatal aperture and roGFP2-Orp1 oxidation.Moreover, roGFP2-Orp1 oxidation state in guard cells was not correlated with a stomatal aperture in control conditions.More importantly, we found that RBOHF plays a crucial role in stomatal immunity and defence against Pseudomonas syringae bacteria by a complex control of both ROS production and apoplastic pH.

RESULTS
PAMP-induced oxidation of the H 2 O 2 sensor roGFP2-Orp1 is higher in guard cells of the rbohF mutant Using the chemiluminescent luminol probe to detect extracellular ROS production, RBOHD was shown to be the main NADPH oxidase implicated in PAMP-induced ROS burst in the apoplast of Arabidopsis leaves (Mersmann et al., 2010;Zhang et al., 2007).Indeed, luminol assays performed on leaf discs showed that the rbohD mutants were defective in flg22-triggered apoplastic ROS burst while a non-significant decrease in ROS production was observed in the rbohF mutants compared to Col-0 WT (Figure 1).Surprisingly, while flg22 causes stomatal closure, ROS production in the cytosol of rbohD or rbohF guard cells upon flg22 activation has not been reported.Moreover, these two NADPH oxidases were not implicated in intracellular ROS increase in guard cells after treatment with yeast elicitors (Khokon et al., 2010).To analyse the contribution of apoplastic ROS burst to cytosolic H 2 O 2 accumulation during the immune response, we used Arabidopsis lines in which the H 2 O 2 sensor roGFP2-Orp1 has been introduced into the rbohD, rbohF, prx33-3 and prx34-2 mutants, respectively (Arnaud et al., 2023).We analysed PAMPtriggered roGFP2-Orp1 oxidation in pavement cells and guard cells of leaf discs after dual excitation at 405 nm and 488 nm (emission at 508-526 nm) by confocal laser scanning microscopy (Figure 2, Figure S1).No significant difference in roGFP2-Orp1 oxidation state was observed in rbohD, rbohF, prx33-3 and prx34-2 pavement cells and guard cells as compared to Col-0 in control conditions (Figure S1a,b).The rbohD, prx33-3 and prx34-2 mutants were not affected in roGFP2-Orp1 oxidation in pavement cells at 60 min after flg22 treatment.Surprisingly, the rbohF mutant exhibited a marked increase in flg22mediated roGFP2-Orp1 oxidation compared to Col-0 in pavement cells (Figure S1a) suggesting lower antioxidant capacities in this mutant as observed before (Arnaud et al., 2023).
In Col-0 guard cells, roGFP2-Orp1 oxidation started to increase from 20 min after flg22 treatment (Figure S1d).The rbohD mutant was not impaired in flg22-mediated roGFP2-Orp1 oxidation in guard cells (Figure 2a).By contrast, the rbohF mutant showed higher roGFP2-Orp1 oxidation compared to Col-0 WT at 30 and 60 min after flg22 treatment.Similarly, the prx33-3 and prx34-2 mutants which are impaired in stomatal immunity (Arnaud et al., 2017) showed a significant increase in flg22-induced roGFP2-Orp1 oxidation in guard cells as compared to Col-0 (Figure S1b).These unexpected results suggest that a defect in PAMP-mediated stomatal closure is correlated with an enhanced H 2 O 2 or oxidation state of guard cell cytosol/nucleus.A more detailed comparison of roGFP2-Orp1 oxidation and stomatal aperture of individual guard cells in Col-0 WT after control or flg22 treatment for 30 and 60 min shows that while roGFP-Orp1 is relatively reduced in control conditions (Figure 2b), there was no correlation between stomatal aperture and roGFP2-Orp1 oxidation (R = 0.07, P = 0.48).We found a negative correlation, albeit weak, between stomatal aperture and roGFP2-Orp1 oxidation across all flg22 treatments (flg22 30 min, Ó 2023 The Authors.The Plant Journal published by Society for Experimental Biology and John Wiley & Sons Ltd., The Plant Journal, (2023), 116, 404-415 R = À0.24,P = 0.02 and flg22 60 min, R = À0.24,P = 0.01) (Figure 2b).Guard cells with roGFP2-Orp1 in a reduced state (either control or flg22-treated) had a wide range of apertures, whereas flg22-treated guard cells with more oxidised roGFP2-Orp1 all exhibited small apertures.

RBOHF is required for PAMP-triggered ROS production in guard cells
As a complementary approach, we used the fluorescein probe H 2 DCFDA, which is not specific to H 2 O 2 (Murphy et al., 2022), to analyse PAMP-induced intracellular ROS production in guard cells of NADPH oxidase mutants.H 2 DCFDA assays indicate that although the basal level of ROS was decreased in both rbohD and rbohF mutants, rbohF has a significantly smaller increase in ROS compared to Col-0 by 30 min after flg22 treatment (Figure 2c).These results suggest that only RBOHF contributes to flg22-mediated ROS production in guard cells but RBOHD does not.Importantly, based on H 2 DCFDA assays, PRX33 and PRX34 are also implicated in PAMP-mediated ROS accumulation in guard cells (Arnaud et al., 2017).Therefore, across these mutants, ROS measurement by H 2 DCFDA shows a correlation between defects in ROS production and defects in flg22-induced stomatal closure while, particularly in the case of rbohF, cytosolic roGFP2-Orp1 is more oxidised.
Because a defect in stomatal defence often results in enhanced susceptibility to bacteria (Melotto et al., 2006), leaves were spray-inoculated with Pst COR À bacteria and bacterial growth was evaluated 3 days later (Figure 3c).The rbohF mutant was more susceptible than the Col-0 control to surface-inoculated bacteria, with a 24-fold increase in the bacterial population.As observed before (Kadota et al., 2014, Macho et al., 2012) the rbohD mutant was as susceptible as Col-0 to surface-inoculated Pst COR À corroborating the absence of defects in bacteria-mediated stomatal closure.To verify any deficiency in post-invasive defences, plants were also inoculated with Pst COR À by syringe infiltration.Similar to previous results with the WT Pst DC3000 bacteria (Chaouch et al., 2012), the rbohD mutant exhibited WT susceptibility to bacteria while rbohF was slightly more susceptible to syringe-inoculated Pst COR À with a threefold increase in bacterial titre as compared to Col-0 (Figure 3c).These results indicate that RBOHF is important for resistance to Pst DC3000 bacteria and contributes more to pre-invasive than post-invasive defence responses.
The rbohF mutant is partially defective in H 2 O 2 -mediated stomatal closure We also tested the stomatal response of rbohD and rbohF mutants to H 2 O 2 which closes stomata through the activation of Ca 2+ channels and inhibition of the inward K + channels (Pei et al., 2000;Zhang et al., 2001).The results indicate that rbohF mutants were partially impaired in H 2 O 2 -mediated stomatal closure while WT stomatal closure was observed for the rbohD mutants (Figure 3d, Figure S3c).Therefore, RBOHF may also act downstream of H 2 O 2 during PAMP-mediated stomatal closure.While H 2 O 2 at 100 μM is widely used for inducing stomatal closure, this effect of H 2 O 2 at this concentration is dependent upon the experimental setup (Li et al., 2021;Pei et al., 2000).Moreover, the rbohF mutant showed increased oxidation of the H 2 O 2 sensor roGFP2-Orp1 in guard cells after PAMP treatment (Figure 2a) and in whole leaf discs after H 2 O 2 treatment (Arnaud et al., 2023).Therefore, we also tested the effect of different H 2 O 2 concentrations on stomatal aperture in our experimental conditions using Col-0 WT leaf discs in stomatal buffer.Surprisingly, while H 2 O 2 at 10 and 100 μM effectively induced significant stomatal closure, we found that stomatal apertures at a higher H 2 O 2 concentration (1 mM) were similar to the control treatment (Figure 3e).These results suggest that excess H 2 O 2 accumulation in rbohF guard cells above a certain threshold is unable to induce stomatal closure.
The rbohF mutant is affected in flg22-mediated apoplastic alkalinisation Apoplastic alkalinisation is a well-characterised early event induced upon PAMP perception (Felix et al., 1999;Felle et al., 2004) and constitutively active plasma membrane H + ATPase OST2/AHA1 impedes PAMP-mediated stomatal closure and apoplastic ROS burst (Keinath et al., 2010;Liu et al., 2009).Otte and collaborators showed that elicitorinduced alkalinisation of the apoplast was inhibited by DPI in chickpea (Otte et al., 2001) and NADPH oxidases could contribute to extracellular pH changes through their electrogenic activities and/or the consumption of protons during superoxide dismutation (Segal, 2016).To test this hypothesis, we used the ratiometric Oregon green fluorescent dye which has been successfully used in Arabidopsis (Geilfus & Muehling, 2011;McLachlan et al., 2016) and adapted the method for multi-well fluorimetry on leaf discs.Interestingly, assays with Oregon green indicate that the apoplastic pH was higher in the rbohF mutant in unstressed conditions while it remained unchanged in the rbohD background (Figure 4a).This result nicely correlates with the impaired growth of rbohF rosette leaves observed previously (Chaouch et al., 2012;Torres et al., 2002)  and rbohF showed WT increase in absolute pH after flg22 treatment (Figure S3d), the flg22-triggered increase in pH relative to the initial pH in unstressed conditions was significantly lower in the rbohF mutant compared to Col-0 WT (Figure 4b).These results suggest that the rbohF mutant is partly defective in PAMP-induced apoplastic alkalinisation.Therefore, as observed for the constitutive active ost2-1D and ost2-2D mutants (Liu et al., 2009;Merlot et al., 2007),

DISCUSSION
The role of RBOHF in stomatal defences PAMP-mediated stomatal closure is considered to involve ROS production (Arnaud & Hwang, 2015).Previous reports suggest that the RBOHD isoform is required for closure, presumably mediated by ROS production initially in the apoplast (Kadota et al., 2014;Li et al., 2014;Macho et al., 2012;Mersmann et al., 2010).However, the defect in PAMPmediated stomatal closure observed in the rbohD mutant was not correlated with increased susceptibility to Pst bacteria (Kadota et al., 2014;Macho et al., 2012).A more recent study showed that RBOHD and RBOHF are not involved in flg22-triggered stomatal closure (Guzel Deger et al., 2015).Contrary to previous work, and using multiple mutant alleles, we found that stomatal closure in response to flg22 and to Pst COR À bacteria depends on RBOHF but not on RBOHD.Correspondingly, there was increased bacterial growth in the rbohF mutant but not rbohD.It should be noted that the rbohD and rbohF mutants showed the expected responses for the apoplastic burst in response to flg22.Our results and previous data suggest that the relative contribution of RBOHD and RBOHF in stomatal immunity likely depends on experimental conditions.While unquantified differences in growth conditions may be involved, it should also be noted that immune responses are strongly dependent on the circadian clock which could, for example, mean that different NADPH oxidase isoforms influence responses at different times of day (Zhang et al., 2013).
Is cytosolic H 2 O 2 important for stomatal closure?
The relationship between ROS production and stomatal closure was complex.The widely used probe H 2 DCFDA, which it is not specific for any particular ROS and not oxidised directly by H 2 O 2 (Murphy et al., 2022) showed decreased ROS production in rbohF guard cells in response to flg22, which is commensurate with defects in stomatal closure.roGFP2-Orp1 and GRX1-roGFP2 have been used to demonstrate oxidation in guard cell cytosol during NaHS-induced stomatal closure (Scuffi et al., 2018).However, during the flg22 response, roGFP2-Orp1 oxidation state did not show a strong relationship with the aperture and the probe was more oxidised in rbohF guard cells despite a lower H 2 DCFDA oxidation.As suggested previously for whole leaves (Arnaud et al., 2023;Chaouch et al., 2012), a lower antioxidant capacity in rbohF guard cells may explain the higher roGFP2-Orp1 oxidation upon PAMP activation.This could reduce the thiol-reducing capacity affecting roGFP2-Orp1 regeneration without influencing H 2 DCFDA oxidation.Added to this, stomatal aperture in Col-0 had a poor correlation with roGFP2-Orp1 oxidation state in control conditions.Using epidermal strips incubated in stomatal opening buffer, the stomatal aperture seems not to be affected by high roGFP2-Orp1 oxidation in guard cells (Scuffi et al., 2018).A possible explanation is that roGFP2-Orp1 is specifically reporting cytosolic H 2 O 2 which could be derived from multiple compartments.For example, mitochondrial ROS contribute to ABA-mediated stomatal closure (Postiglione & Muday, 2023).The use of catalase mutants has shown that, despite higher intracellular H 2 O 2 , stomatal aperture is unaffected although still responsive to exogenous H 2 O 2 (Jannat et al., 2011).This may explain why stomata remain open despite photosynthetic ROS production in the light.Our results, therefore, support the uncoupling of cytosolic H 2 O 2 from stomatal aperture and that the extracellular production of superoxide or H 2 O 2 by RBOHF is critical for flg22-induced closure.
The LRR receptor kinase HPCA1 is a proposed extracellular H 2 O 2 sensor which, when oxidised on extracellular cysteines, activates calcium channels leading to increased intracellular Ca 2+ (Wu et al., 2020).This sensor provides an explanation for the requirement of apoplastic H 2 O 2 for stomatal closure.Intriguingly, our results suggest that the ROS probe H 2 DCFDA could indicate that ROS other than H 2 O 2 are involved in triggering closure while roGFP2-Orp1 indicates cytosolic conditions that are at least partly independent of stomatal aperture.The results with guard cells therefore partly mirror the recent findings with whole leaves (Arnaud et al., 2023) where the RBOHD-dependent apoplastic oxidative burst measured by luminol is independent of subsequent prolonged oxidation of cytosolic roGFP2-Orp1.(a) Stomatal apertures in WT Col-0, rbohD and rbohF leaf discs exposed to mock control (10 mM MgCl 2 ), 10 8 cfu/ml COR-deficient Pst DC3000 (Pst COR

Putative role of RBOHF in PAMP-mediated apoplastic alkalinisation
The rbohF mutant was affected in stomatal closure induced by exogenous H 2 O 2 suggesting that, besides ROS production, RBOHF also functions in other mechanisms important for stomatal immunity.Interestingly, despite a higher apoplastic pH under normal conditions, the rbohF mutant was affected in flg22-induced apoplastic alkalinisation.In the rbohC mutant, the amplitude of apoplastic pH fluctuations during root hair elongation is affected (Monshausen et al., 2007).It was shown recently that damage-induced membrane depolarisation in roots is impaired in rbohD and rbohF mutants (Marhavý et al., 2019) and RBOHD might be important in propagating the systemic electric potential induced by wounding in leaves (Suzuki et al., 2013).These results suggest that besides superoxide production, plant NADPH oxidases like their animal counterparts (Segal, 2016) are electrogenic through electron transport across the plasma membrane which is coupled to proton efflux probably through H + -ATPases (Figure 4c).These extracellular H + ions could participate in cell wall loosening via the expansion activities in normal conditions (Cosgrove, 2005).During PAMP-triggered immunity, superoxide dismutation to H 2 O 2 , probably catalysed by unknown PAMP-activated superoxide dismutases or germin-like proteins (Smirnoff & Arnaud, 2019), consumes protons and could increase apoplastic pH (Figure 4c) which in turn activates apoplastic peroxidases (Bolwell et al., 2002).Alternatively, rbohF could have an indirect effect on the activity of unknown proton transporters or H + -ATPases at the plasma membrane (Segal, 2016).It has been suggested that flg22 inhibits AHA1 and AHA2 H + -ATPase via dephosphorylation (Elmore & Coaker, 2011;Nuhse et al., 2007).More importantly, apoplastic alkalinisation driven by RBOHF or AHA1/2 inhibition likely contributes to plasma membrane depolarisation which is well known to activate outward K + channels and S-type anion channels leading to stomatal closure (Kollist et al., 2014).The function of RBOHF, RBOHD and other NADPH oxidases in connecting extra-and intracellular ROS production, apoplastic and cytosolic pH changes and plasma membrane polarisation during plant response to environmental stresses needs to be further explored.

Statistical analysis
The experiments reported here were repeated at least three times with similar results unless otherwise mentioned.Experiments were analysed by Student's t-tests or ANOVAs followed by Tukey's honestly significant difference (HSD) post hoc test using R software (R Core Team, https://www.R-project.org).Raw data for stomatal aperture measurements are in Supplemental Table S1.The time course experiment was analysed by 2-way ANOVA.Significant differences between each treatment at each time were determined by uncorrected Fisher's LSD test using GraphPad Prism v8 (GraphPad, San Diego, California, USA).

Bacterial infection assay
The bacterial strain Pst DC3000 COR À (DB4G3) (Brooks et al., 2004) was cultivated overnight at 28°C in King's B medium supplemented with Kanamycin and Rifampicin (each at 100 μg/ml).Bacteria were collected by centrifugation at 3000 g for 5 min at room temperature and washed twice in 10 mM MgCl 2 .Plants were surface-inoculated by spraying with a bacterial solution of 10 8 cfu/ml in 10 mM MgCl 2 containing 0.02% Silwet L-77, and plants were covered to maintain high humidity until disease symptoms developed.Alternatively, rosette leaves were syringe-infiltrated with a bacterial solution of 10 6 cfu/ml in 10 mM MgCl 2 .After 3 days, bacterial growth in the apoplast of three leaves per plant and six plants per genotype was determined as previously described (Katagiri et al., 2002).

Luminol assay
Leaf discs of 6 mm diameter were cut into four equal pieces, immersed in distilled water and incubated for 3 h minimum at room temperature for recovery after wounding.Before starting the assay, water was exchanged by a solution containing 100 μM luminol and 10 μg/ml horseradish peroxidase (≥ 250 units/mg solid).After adding a control solution or 1 μM flg22, the luminescence was measured immediately using a CLARIOstar plate reader (BMG Labtech, Aylesbury, UK) with a reading time of 2 sec.

Confocal imaging and image analysis
Leaf discs from rosette leaves were immersed in 10 mM MES-KOH pH 6.15, 30 mM KCl, incubated for 2 h at 21°C under laboratory lighting (PPFD ∼ 10 μmol m À2 sec À1 ) for recovery after wounding and subsequently treated with control solution or 1 μM flg22 for 30 and 60 min.Leaf discs were mounted under a Zeiss confocal microscope LSM 880.Images were collected with a 20X lens (Plan-Apochromat, 0.8 numerical aperture) and roGFP2-Orp1 was excited sequentially at 405 and 488 nm and emission was detected at 508-526 nm, with the pinhole set to 1 airy unit.Single-plane images were processed with ImageJ software.Background fluorescence was insignificant and not subtracted.Fluorescence intensity values of individual guard cells selected as region of interest were quantified on 32-bit converted images and the ratio 405/488 nm was calculated.

Measurements of stomatal aperture
Stomatal experiments were conducted as previously described (Desclos-Theveniau et al., 2012).Epidermal peels collected from the abaxial side of young fully expanded leaves or leaf discs were floated in stomatal buffer (10 mM MES-KOH pH 6.15, 30 mM KCl) for 2.5 h under light (100 μmol m À2 sec À1 ) to ensure that most stomata were opened before treatments.Solutions of chemicals or bacterial suspension at 10 8 cfu.mlÀ1 in 10 mM MgCl 2 were directly added to the stomatal buffer.Monitoring ROS in guard cells 2 0 ,7 0 dichl orofluorescein diacetate (H 2 DCFDA, Sigma) was used to measure ROS in guard cells (Pei et al., 2000).After 2.5 h incubation in stomatal buffer, epidermal peels were incubated with 50 μM H 2 DCFDA in 0.1% DMSO for 15 min.Excess H 2 DCFDA was then removed by washing 3 times for 20 min with stomatal buffer.Then, 5 μM flg22 or control solution was added.After 30 min incubation, H 2 DCFDA fluorescence was observed with a fluorescence microscope (Carl Zeiss, Axioplan 2) and fluorescence intensity of guard cells was analysed using ImageJ software.

Gene expression analysis in guard cell protoplasts
For each condition, about 50 young fully expanded leaves were immerged for 2 h in stomatal buffer containing 0.02% (v/v) Silwet-L77 with or without 1 μM flg22.Guard cell protoplasts were isolated as previously described et al., 2013) in the presence of transcriptional inhibitors 0.01% (w/v) cordycepin and 0.0033% (w/v) actinomycin D. For each condition, about 10 6 guard cell protoplasts were obtained and their purity was above 98%.RNA was extracted using the RNeasy Plant Mini Kit with incolumn Dnase 1 digestion (Qiagen, Hilden, Germany).Two hundred nanograms of total RNA were reverse transcribed using 500 ng of oligo(dT)15 and the ImProm-II Reverse Transcription System following the manufacturer's instructions (Promega, Madison, WI, USA).Quantitative real-time PCR reaction was performed on a Light Cycler 2 (Roche, Basel, Switzerland) using 10 μl SYBR Premix Ex Taq (Takara, Kusatsu, Japan), 2 μl of twofold diluted cDNA and 0.5 μM of primers in a total volume of 20 μl per reaction.The cycling conditions were composed of an initial 20 sec denaturation step at 95°C, followed by 45 cycles of 95°C for 7 sec, 60°C for 10 sec and 72°C for 13 sec.A melting curve was run from 65°C to 95°C to ensure the specificity of the products.Data were analysed with the delta Ct method.Ubiquitin 1 (UBQ1) was used as a reference gene for normalisation of gene expression levels.Control treatment was considered as expression level = 1.As controls, the expression of the reference gene ACT2 did not change between control and flg22 conditions while the expression of the PTI marker gene PRX4 was induced by flg22 (Figure S2d).qRT-PCR primer sequences are listed in Supplemental Table S2.

Ratiometric pH quantification using Oregon green
Fully expanded leaves were syringe-infiltrated with 25 μM of Oregon Green 488 dextran (ThermoFisher, Waltham, MA, USA) (Geilfus & Muehling, 2011;McLachlan et al., 2016).For the autofluorescence background, milliQ water was infiltrated into the apoplast.Plants were kept in the growth chamber for 2 h until excess water has evaporated.Oregon Green-treated leaf discs were placed in a 96-well plate, immersed in milliQ water and further incubated for 2 h at 21°C under laboratory lighting (PPFD ∼ 10 μmol m À2 sec À1 ) for recovery after wounding.Oregon Green was excited sequentially at 440 nm AE 8 nm and 495 nm AE 8 nm in a CLARIOstar plate reader (BMG Labtech) and emission was recorded at 525 nm AE 20 nm with a gain set at 1250 and 1000 for the 440 and 495 nm excitations, respectively.Each leaf disc was scanned from the top with the fluorescence recorded and averaged from 76 flashes per well-organised as a spiral of 5 mm diameter.The initial 495/440 ratio of the resting state of leaf discs was estimated by reading the wells for 15 min before treatment.The emission of water-infiltrated leaf discs was averaged and subtracted for all the data points to correct for background fluorescence.To convert the fluorescence 495/440 ratio into pH values, a calibration curve with pH-buffered Oregon Green solution infiltrated into the apoplast of leaves was performed according to (McLachlan et al., 2016).The Boltzmann fit was chosen for fitting sigmoidal curves to calibration data and pH was determined according to the equation: x ¼ V 50ÀSlope In TopÀy yÀBottom where x = pH, y = 495/440 ratio, Bottom = 0.4833, Top = 2.045, V50 = 4.513 and Slope = 0.6267.Changes in pH over time were normalised to the initial pHi as pH/pHi.

Figure 2 .
Figure 2. The H 2 O 2 sensor roGFP2-Orp1 is more oxidised in guard cells of the rbohF mutant.(a) Oxidation state of roGFP2-Orp1 in Col-0 WT, rbohD and rbohF guard cells in response flg22.Leaf discs were exposed to control solution or 1 μM flg22 and the ratio 405/488 nm of stomata was quantified at 30 min and 60 min after treatment from images of the fluorescence emission at 517 AE 9 nm following excitation at 488 and 405 nm.Data are means AE SE (n ≥ 50 guard cells) from a representative experiment.Different letters indicate significant differences at P < 0.05 based on a Tukey's HSD test.(b) Correlation between the stomatal aperture and the oxidation state of roGFP2-Orp1 in guard cells of leaf discs after treatment with control solution or 1 μM flg22 for 30 min and 60 min.The scatter plots show the stomatal aperture as a function of the ratio 405/488 nm for each condition (n ≥ 100).The Pearson correlation coefficient (R) is 0.0651 (P = 0.483) for control treatment and À 0.2428 (P = 0.015) and À 0.2395 (P = 0.011) for flg22 treatment at 30 min and 60 min, respectively.(c) ROS production detected by H 2 DCFDA fluorescence in guard cells of Col-0 WT, rbohD and rbohF epidermal peels 30 min after treatment with control solution or 5 μM flg22.Data are means AE SE (n ≥ 200) from 3 independent experiments.Different letters indicate significant differences at P < 0.05 based on a Tukey's HSD test.
Authors.The Plant Journal published by Society for Experimental Biology and John Wiley & Sons Ltd., The Plant Journal, (2023), 116, 404-415 defects in apoplastic alkalinisation may also contribute to the impairment of rbohF in H 2 O 2 -, ABA-and PAMPmediated stomatal closure.

Figure 3 .
Figure 3. Analysis of the function of RBOHD and RBOHF during stomatal defence responses.(a)Stomatal apertures in WT Col-0, rbohD and rbohF leaf discs exposed to mock control (10 mM MgCl 2 ), 10 8 cfu/ml COR-deficient Pst DC3000 (Pst COR À ) bacte- ria and 5 μM flg22 for 2 h.Data are means AE SE (n ≥ 100) from a representative experiment.(b) Stomatal apertures in Col-0 WT and other allelic mutants of rbohD (SALK_070610) and rbohF (SALK_034674).Epidermal peels were exposed to control solution or 5 μM flg22 for 2 h.Data are means AE SE (n ≥ 80) from a representative experiment.(c) Bacterial growth in WT Col-0, rbohD and rbohF mutants assessed at 3 days after syringe-infiltration with Pst COR À at 10 6 cfu/ml (left panel) or after sprayinoculation with Pst COR À at 10 8 cfu/ml (right panel).Values are the means AE SE (n = 6).(d) The rbohF mutant is partially defective in H 2 O 2 -mediated stomatal closure.Stomatal apertures in WT Col-0, rbohD and rbohF leaf discs exposed to Control or 100 μM H 2 O 2 for 2 h.Data are means AE SE (n ≥ 60) from a representative experiment.(e) Dose response of Col-0 WT stomata to H 2 O 2 indicate that high H 2 O 2 concentration does not close stomata.Stomatal apertures in WT Col-0 leaf discs exposed to different concentrations of H 2 O 2 (0, 10, 100 and 1000 μM) for 2 h.Data are means AE SE (n ≥ 60) from a representative experiment.Different letters indicate significant differences at P < 0.001 (A, B, D and E) and P < 0.05 (C) based on a Tukey's HSD test.

Figure 4 .
Figure 4. RBOHF influences apoplastic pH.(a) Apoplastic pH in Col-0 WT, rbohD and rbohF mutant leaves.Apoplastic pH in untreated condition was determined by multiwell fluorimetry (sequential excita- tion at 440 AE 8 and 495 AE 8 nm; emission, 525 AE 20 nm) on Oregon green dextran-infiltrated leaf discs from 5 week-old plants (See Methods for details).Data are means AE SE (n ≥ 12) from a representative experiment.Different letters indicate significant differences at P < 0.001 based on a Tukey's HSD test.(b) Kinetics of flg22-induced leaf apoplastic alkalinisation in Col-0 WT, rbohD and rbohF mutants.Oregon green dextran-infiltrated leaf discs were exposed at t = 0 min to control solution or 1 μM flg22, the apoplastic pH was measured over time by multiwell fluorimetry and expressed relative to the mean initial pH (pHi) before treatment (pH/pHi).Data are means AE SE (n ≥ 18) from 3 independent experiments.Significant differences at P < 0.05 between flg22-treated Col-0 and rbohF were found between 10 and 60 min based on two-way ANOVA and uncorrected Fisher's LSD analyses for each time point.(c) Hypothetical model of the regulation of apoplastic pH by RBOHF during PTI activation.In normal unstressed condition, the efflux of electron (e À ) produced by the NADPH oxidase activity is compensated by an efflux of proton (H + ) probably generated by plasma membrane H + ATPases.Both the acidification of the apoplast and the production of hydroxyl radicals (OH • ) from superoxide (O 2 À ) through the Haber-Weiss and Fenton reactions contribute to cell expansion.The plant immune response may induce the inhibition of H + ATPases and the activation of RBOHF together with superoxide dismutases (SOD) or germin-like proteins which dismutates O 2 À to H 2 O 2 , a reaction that consumes H + .The resulting alkalinisation of the apoplast induces a depolarisation of the plasma membrane ultimately leading to stomatal closure via the activation of ion channels.
and in our laboratory conditions suggesting that RBOHF is important for the acidic growth of plant cells.Although rbohD Ó 2023 The Authors.The Plant Journal published by Society for Experimental Biology and John Wiley & Sons Ltd., The Plant Journal, (2023), 116, 404-415 À ) bacte- ria and 5 μM flg22 for 2 h.Data are means AE SE (n ≥ 100) from a representative experiment.(b) Stomatal apertures in Col-0 WT and other allelic mutants of rbohD (SALK_070610) and rbohF (SALK_034674).Epidermal peels were exposed to control solution or 5 μM flg22 for 2 h.Data are means AE SE (n ≥ 80) from a representative experiment.(c) Bacterial growth in WT Col-0, rbohD and rbohF mutants assessed at 3 days after syringe-infiltration with Pst COR À at 10 6 cfu/ml (left panel) or after sprayinoculation with Pst COR À at 10 8 cfu/ml (right panel).Values are the means AE SE (n = 6).(d) The rbohF mutant is partially defective in H 2 O 2 -mediated stomatal closure.Stomatal apertures in WT Col-0, rbohD and rbohF leaf discs exposed to Control or 100 μM H 2 O 2 for 2 h.Data are means AE SE (n ≥ 60) from a representative experiment.(e) Dose response of Col-0 WT stomata to H 2 O 2 indicate that high H 2 O 2 concentration does not close stomata.Stomatal apertures in WT Col-0 leaf discs exposed to different concentrations of H 2 O 2 (0, 10, 100 and 1000 μM) for 2 h.Data are means AE SE (n ≥ 60) from a representative experiment.Different letters indicate significant differences at P < 0.001 (A, B, D and E) and P < 0.05 (C) based on a Tukey's HSD test.
Ó 2023 The Authors.The Plant Journal published by Society for Experimental Biology and John Wiley & Sons Ltd., The Plant Journal, (2023), 116, 404-415 The Plant Journal published by Society for Experimental Biology and John Wiley & Sons Ltd., The Plant Journal, (2023), 116, 404-415 peels or leaf discs were further incubated under light for 2 h and observed under a light microscope (Carl Zeiss, Axioplan 2).Stomatal apertures of stomata in random areas were measured using ImageJ 1.42 software.Measurements of stomatal aperture were performed 7-8 h after starting illumination.