Leaf apoplastic alkalization promotes transcription of the ABA-synthesizing enzyme Vp14 and stomatal closure in Zea mays

Abstract The chloride component of NaCl salinity causes the leaf apoplast to transiently alkalinize. This transition in pH reduces stomatal aperture. However, whether this apoplastic pH (pHapo) transient initiates stomatal closure by interacting with other chloride stress-induced responses or whether the pH transient alone initiates stomatal closure is unknown. To clarify the problem, the transient alkalinization of the leaf apoplast was mimicked in intact maize (Zea mays L.) by infiltrating near-neutral pH buffers into the leaf apoplast. Effects of the pHapo transient could thus be investigated independently from other chloride stress-derived effects. Microscopy-based ratiometric live pHapo imaging was used to monitor pHapoin planta. LC-MS/MS and real-time quantitative reverse transcription–PCR leaf analyses showed that the artificially induced pHapo transient led to an increase in the concentrations of the stomata-regulating plant hormone abscisic acid (ABA) and in transcripts of the key ABA-synthesizing gene ZmVp14 in the leaf. Since stomatal aperture and stomatal conductance decreased according to pHapo, we conclude that the pHapo transient alone initiates stomatal closure. Therefore, the functionality does not depend on interactions with other compounds induced by chloride stress. Overall, our data indicate that the pH of the leaf apoplast links chloride salinity with the control of stomatal aperture via effects exerted on the transcription of ABA.

the stomatal aperture by acting on the compartmental distribution of the guard cell-regulating phytohormone abscisic acid (ABA) between the leaf apoplast and symplast (Davies et al., 2002(Davies et al., , 2005Wilkinson and Davies, 2008).
There is, however, an ongoing debate over the site of synthesis of the ABA that accumulates in the leaf in response to soil water scarcity. Some studies conclude that the root is the major site of ABA biosynthesis during soil drying, with rootderived ABA acting as a long-distance chemical signal regulating stomatal conductance (Wilkinson, 1999;Wilkinson et al., 2007;Boyle et al., 2016). New knowledge gained from experiments on reciprocal grafts of ABA-deficient mutants and wild-type (WT) tomato (Solanum lycopersicum) plants has challenged this view. Reciprocal grafting experiments with ABA-synthesizing WT and ABA-deficient mutants revealed rootstock independence of stomatal regulation, irrespective of the ABA concentration delivered via the xylem from root to shoot (Holbrook et al., 2002;Dodd et al., 2009;Li et al., 2018). These grafting experiments indicate that rootstock ABA synthesis seems to be less important in regulating leaf ABA concentration and stomatal conductance as the soil dries than originally assumed (Holbrook et al., 2002;McAdam et al., 2016a). Instead, a production of ABA in phloem companion cells of the vasculature (Kuromori et al., 2014) might feed the apoplast of water-stressed leaves with ABA, as shown for water-stressed Arabidopsis thaliana. In addition, A. thaliana guard cell-autonomous ABA synthesis could allow the leaf to maintain hydration (Bauer et al., 2013). In maize, the 9-cisepoxycarotenoid dioxygenase (NCED) gene family ortholog viviparous 14 (Vp14) encodes the rate-limiting enzyme in ABA biosynthesis (Seo and Koshiba, 2002;Xiong and Zhu, 2003). The ZmVp14 gene was characterized from the ABA-deficient maize mutant Vp14 (Tan et al., 1997) and is known to be regulated by salinity and osmotic stress Geilfus et al., 2018).
However, this drought-induced alkalinization is not universal in all plant species, which has cast doubt on the assumption that it is a ubiquitous long-distance signal that regulates stomatal responses to soil drying (Sharp and Davies, 2009). Of 22 woody perennial species from different plant families, only four (Buddleja davidii, Dicksonia antarctica, Penstemon heterophyllus, and Rhododendron obtusum) evolutionarily unrelated species alkalized their xylem sap as the soil dried, but these species showed the greatest control of water status during water scarcity (Sharp and Davies, 2009). Moreover, within eight herbaceous species from various taxonomic groups the ability to alkalinize xylem sap in response to soil drying was not universal, and unrelated to phylogeny (Gloser et al., 2016). Overall, alkalinization was more likely in annual plant species than in woody perennial plants, but there are few insights as to why this is so.
Nevertheless, for those species that alkalinize the apoplast in response to soil drying, the following mode of action is proposed for regulating stomatal aperture. Under wellwatered conditions, ABA is released into the leaf apoplast. The apoplastically located ABA is present in its undissociated form (ABAH) because it is a weak acid and because the apoplast is slightly acidic (Hartung, 2010). By following its chemical gradient, the uncharged ABAH diffuses from the leaf apoplast through the plasma membrane (PM) into mesophyll cells. As soon as ABAH enters the slightly alkaline cytosol of the mesophyll, the weak acid dissociates to ABA − by releasing a proton (H + ). The phytohormone is now charged (ABA − ) and is trapped inside the mesophyll cells because charged compounds cannot diffuse across the PM (Hartung et al., 1988). It is then stored or metabolized in the mesophyll ( Slovik and Hartung, 1992;Wilkinson and Davies, 1997).
However, during progressive soil drying, the pH of the leaf apoplast increases above the pKa of ABA (=4.87) (Karuppanapandian et al., 2017). Thus, under conditions of water scarcity, the ABA that is increasingly delivered from the root or stem dissociates in the leaf apoplast into its charged anionic form, becoming trapped extracellularly (Wilkinson and Davies, 1997). Such a pH apo -based accumulation of ABA − in the leaf apoplast was recently demonstrated under conditions of NaCl salinity when the chloride component of NaCl salinity provoked the leaf apoplast of the field bean to alkalinize over a period of 2-3 h, before it re-acidified once again (Geilfus et al., 2015). After re-acidification of the leaf apoplast back to the initial steady-state pH, the guard cell-intrinsic ABA concentration was increased and the stomata were closed. The guard cellintrinsic ABA concentration was proposed to rise because the sudden re-acidification of the leaf apoplastic compartment established such a high free proton concentration in the apoplast that the bulk of the apoplastically trapped ABA − was associated with protons. In other words, a fraction of membrane-permeable ABAH was formed in the leaf apoplast in the vicinity of the guard cells; this ABAH followed its gradient and streamed into the guard cells. This is in good agreement with Goodger et al. (2005), who report for maize that the release rates of both ABA and chloride out of mesocotyl xylem sap that is flowing from roots into the shoot correlate with stomatal conductivity. We previously demonstrated that the chloride-induced transient leaf apoplastic alkalinization was instrumental in inducing stomatal closure during the beginning of NaCl salinity through pH apo -based effects on the compartmental distribution of ABA (Geilfus et al., 2015). Experiments with different chlorideor sodium-accompanying counter ions, and with agents that mimicked osmotic stress, revealed that the reduction of the stomatal aperture was linked to the chloride component, namely to the chloride-induced transient leaf apoplastic alkalinization (Geilfus et al., 2015). Nevertheless, stomatal aperture was not reduced when leaf concentrations of chloride increased during the experimental suppression of the formation of the pH apo transient. This was established by clamping the pH apo in the acid range by the infiltration of an acid pH buffer. Hence, the chloride-induced transient alkalinization of the leaf apoplast was inhibited, although chloride was taken up excessively into the leaf. These experiments showed that the pH apo transient mechanistically links chloride stress with a reduction in stomatal aperture (Geilfus et al., 2015).
Nevertheless, these experiments have not revealed whether the formation of the transient apoplastic alkalinization alone, (i) independently from other chloride stress-induced responses and (ii) in the absence of excessive uptake of chloride, is functional in initiating a reduction in stomatal aperture. Only an experimental set-up in which this transient leaf apoplastic alkalinization is mimicked but without stressing the plants with chloride is capable of clarifying this point. With this aim, the chloride-inducible alkalinization of the leaf apoplast was mimicked by infiltrating near-neutral pH buffers in this study. We hypothesized that the formation of this artificially induced pH apo transient reduces stomatal aperture, despite the plants not being salinity stressed, and that ABA synthesis is under control of the pH apo transient. To elaborate these hypotheses, the apoplastic pH dynamics, the transcript abundance of the ABAsynthesizing key enzyme ZmVp14, the abundance of ABA, the stomatal pore size, stomatal conductance (g s ), and the transpiration rate (E) were quantified in the leaves of hydroponically gown maize.

Experimental design
Maize (Zea mays L.) was cultivated hydroponically in nutrient solution (in 2019 and 2020). To mimic the formation of the transient alkalinization of the leaf apoplast in a way comparable (regarding duration and magnitude) with the chloride-induced alkalinization of the leaf apoplast (see Geilfus et al., 2015), the pH buffer MES (5 mM, pH 6.5) was infiltrated into the leaf apoplast. Ratiometric real-time pH apo imaging was applied to monitor the pH apo transient in planta (see below). The pH buffer was mixed with the pH indicator dye Oregon Green (OG) 488-dextran (25 µM) before infiltration. A volume of 0.4 ml of the pH buffer/dye mixture was infiltrated using a needleless syringe. A second experimental group was set up to rule out unspecific effects that may have arisen from the buffering agent (i.e. effects that did not arise from the pH apo ). For this purpose, 0.4 ml of a mixture of the pH buffer 2-hydroxy-MOPS (MOPSO) (7.5 mM, pH 6.5) and OG dye was infiltrated into the leaf. To guard against effects possibly arising from the infiltration procedure or the infiltration of water, control plants were infiltrated with 0.4 ml of a mixture of water and the OG dye. Osmotic controls were used to test for osmotic effects possibly arising from the pH buffer agents. For this, 0.4 ml of a mixture of the pH buffer MES (7.5 mM, pH 4.7) and OG dye was infiltrated into the leaf. In this fourth experimental group, the pH was set to 4.7 because this proton concentration prevails in the leaf apoplast.
In total, four experimental groups were tested. (i) Control: infiltration of 0.4 ml of an aqueous 25 µM OG solution. (ii) Osmotic control: infiltration of 0.4 ml of a mixture of 7.5 mM MES (pH 4.7) and 25 µM OG. (iii) MES buffer: infiltration of 0.4 ml of a mixture of 5 mM MES (pH 6.5) and 25 µM OG. (iv) MOPSO buffer: infiltration of 0.4 ml of a mixture of 7.5 mM MOPSO (pH 6.5) and 25 µM OG.
To investigate effects of the formation of the artificially induced pH apo transient on the transcription of the ABA-synthesizing maize gene 9-cisepoxycarotenoid dioxygenase 1 (ZmVp14; Tan et al., 2003) and on the abundance of ABA, leaves were sampled at eight key time points to reflect the situation before, during, and after the alkalinization. Microscopy-based live pH apo imaging allowed the identification of these eight time points. A separate batch of plants was necessary for sampling leaf material at each time point. Sampled leaf material was immediately frozen in liquid nitrogen. Since pH apo imaging was performed in planta by using a sizecalibrated microscope, the size of the stomatal aperture could be quantified concurrently. Transpiration was measured in parallel on separate batches of plants that were treated identically. To confirm that one-time infiltration of 0.4 ml MES or the MOPSO buffers does not disturb the leaf in the long term, we checked rates of photosynthesis and transpiration at 24, 48, and 72 h after infiltration (Supplementary Table S1). For all measurements, the number of biological replicates varied between four and six, as indicated in the figure legends or table captions. Biological replicates represented individual plants taken from different cultivations.

Quantification of leaf apoplastic pH and determination of stomatal aperture
Leaf pH apo and stomatal aperture were quantified using a calibrated Leica microscope (Leica DMi8; Leica Microsystems, Wetzlar, Germany). The entire plant was located in a cage incubator system, which allowed the precise control of the VDP (0.58 kPa; 75% relative humidity at 20 °C), and white light illumination via an LED over the entire experiment [light intensity was 400 µmol ms -1 of photosynthetic photon flux density (PPFD)]. LED illumination was automatically switched off during fluorescence image acquisition for pH apo quantification. For the in vivo quantification of leaf pH apo , a 25 μM solution of the fluorescent pH indicator dye OG 488-dextran (Thermo Fisher Scientific, Darmstadt, Germany) was infiltrated into the leaf apoplast of intact plants by using a needleless syringe (Geilfus and Mühling, 2011), together with a pH buffer. where applicable. OG 488-dextran is a sensor for pH apo because it does not enter the symplast, as shown by confocal imaging. Fluorescence images for calculation of pH apo were collected as a time series with the Leica DMi8 microscope connected to a cooled sCMOS camera (Leica DFC9000 GT; Leica Microsystems). An HXP lamp (HXP Short Arc Lamp; Osram) was used for illumination at the excitation wavelengths of 440/10 nm and 490/10 nm. The exposure time was 25 ms for both channels. Emission was collected at 535/25 nm by using a band-pass filter in combination with a dichromatic mirror. The fluorescence ratio F 490 /F 440 was obtained as a measure of pH apo on a pixel-by-pixel basis. For conversion of the fluorescence ratio data into pH apo values, an in vivo calibration was conducted as described elsewhere (Geilfus and Mühling, 2013). Stomatal aperture was quantified off-line on captured images by using the calibrated microscope.

RNA extraction, cDNA synthesis, and real-time quantitative RT-PCR
Forward (f 5'-3': TTCTCGGAGGAGGAACAGAGGA) and reverse (r 5'-3′: CCAACTGTAACTCTGGTGTGCG) primers for amplifying ZmVp14 mRNA were taken from Geilfus et al. (2018). For cDNA synthesis, RNA was isolated from 100 mg of ground lyophilized leaf material by using phenol-chloroform extraction according to the method of Cox and Goldberg (1988). The quality of RNA was checked by OD 260 and OD 280 . A 1 µg aliquot of total RNA was digested with PerfeCTa DNAseI (Quanta Biosciences, Beverly, MA, USA) to eliminate residual genomic DNA, and reverse transcribed to single-stranded cDNA with the SuperScript VILO cDNA Synthesis Kit (Invitrogen by Life Technologies, Karlsruhe, Germany) according to the manufacturer's instructions. The SYBR-Green-based real-time quantitative reverse transcription-PCR (real-time qRT-PCR) technique was performed on a BioRad CFX96 real-time PCR system by using iTaq Universal SYBR Green Supermix (BioRad Laboratories, Inc.). For each reaction, 0.8 µl of diluted singlestranded cDNA was used in a total volume of 10 µl (0.5 µM forward and reverse primer, 5 µl of iTaq Universal SYBR Green Supermix, topped up with sterilized autoclaved H 2 O bidest ). Cycle information with regard to temperature, time, and cycle number is given in Geilfus et al. (2018). The specificity of the annealing was checked by dissociation kinetics performed at the end of the experiment (65-95 °C). The comparative ∆∆Ct (threshold cycles) method for relative quantification was used to analyse the data according to Pfaffl (2001). C t values were normalized by comparison with the two endogenous reference genes actin 1 and ubiquitinconjugating enzyme. Data are shown as the relative fold changes in transcript expression. Negative controls were carried out without templates. The specificity of the primer-template interactions was demonstrated by sequencing the real-time qRT-PCR products (GATC Biotech, Konstanz, Germany) (Supplementary Table S2). Moreover, agarose gels were run after real-time qRT-PCR to ensure that only a single PCR product was generated and to confirm the predicted PCR product size on the gel (Supplementary Fig. S1).

ABA quantification
Extraction of ABA from the maize leaf was performed by adding 1 ml of extraction solution {5 ng of [ 2 H 6 ] (+)-cis,trans ABA internal standard (Olomouc, Czech Republic) in 7:3 methanol:water} to 10 mg of dried ground material. After being shaken for 30 min, samples were centrifuged at 16 000 g at 4 °C for 5 min. The supernatant was collected and dried under reduced pressure, re-dissolved in methanol (10 µl per 1 mg DW), mixed, and centrifuged at 16 000 g and 4 °C for 10 min. The supernatant was used for the LC-MS/MS-based analysis of ABA (Almeida Trapp et al., 2014).

Gas exchange measurements
A portable gas exchange system (LI-COR 6400 XT; LI-COR) was used to measure g s (mmol H 2 O m -2 s -) and to obtain rates for E (mmol H 2 O m -2 s -1 ) and photosynthesis (μmol CO 2 m -2 s -1 ). PPFD was 500 μmol m −2 s −1 as provided by the red, green, and blue LEDs of the integrated fluorescence chamber head (6400-02B LED light source; LI-COR). The leaf area included in the chamber was recorded for each leaf. VPD was 0.58 kPa (75% relative humidity; 20 °C). CO 2 at a flow rate of 300 μmol mol −1 CO 2 was controlled by a CO 2 injection system.

Results
Infiltrating pH buffers set to pH 6.5 mimics a transient alkalinization of the leaf apoplast Infiltrating 0.4 ml of a 5 mM MES solution that was set to a pH of 6.5 into the leaf apoplast immediately increased the leaf pH apo from 4.7 to 6.5 (Fig. 1, green circles). After remaining at pH 6.5 for a period of 80 min, the pH apo started to re-acidify back to the initial steady-state pH of 4.7 over a period of 60 min. A similar transient alkalinization of the leaf apoplast was seen when 0.4 ml of a 7.5 mM MOPSO solution set to a pH of 6.5 was infiltrated into the leaf apoplast (blue diamonds). Control plants that were infiltrated with 0.4 ml of water showed a stable pH apo of 4.7 over the entire experiment (black triangles). The pH apo was also stable at 4.7 when 0.4 ml of a 7.5 mM MES solution set to a pH of 4.7 was infiltrated into the leaf (grey squares). The latter experimental group served as a control to exclude osmotic effects possibly arising from the infiltration of buffering agents (Fig. 1).

Transient alkalinization of leaf apoplast increases abundance of Vp14 mRNA
Infiltration of either MES or MOPSO buffer solutions each set to a pH of 6.5 into the leaf apoplast resulted in a steady increase of the abundance of Vp14 mRNA, relative to control leaves (Fig. 2). At 210 min after buffer infiltrations (this time point is shown at minute 250 on the x-axis), the increases were the highest (52-fold in response to infiltration with MES at pH 6.5 and 68-fold in response to infiltration with MOPSO at pH 6.5). Subsequently, the alkalinization-induced increase in Vp14 mRNA abundance decreased. Infiltration of the MES buffer set to a pH of 4.7 (osmotic control) resulted in a slight increase of Vp14 mRNA abundance. In contrast to the alkalinizationinduced increase, the osmotically caused increase was detected 120 min later, reaching a much lower maximum that occurred at 280 min after buffer infiltrations (this time point is shown at minute 320 on the x-axis) (Fig. 2).

Leaf pH apo transient causes leaf ABA concentrations to increase
Infiltration of either MES or MOPSO buffer solutions each set to a pH of 6.5 into the leaf apoplast resulted in a steady increase of the leaf ABA concentrations, whereas the ABA concentration in the control leaves that were only infiltrated with water remained stable at 20±2 (mean ±SE) ng ABA g -1 DW over the entire experiment (Fig. 3A). At 280 min after buffer infiltrations (this time point is shown at minute 320 on the x-axis), the increases were at their highest (943±86 ng ABA g -1 DW in response to infiltration with MES at pH 6.5 and 1481±59 ng ABA g -1 DW in response to infiltration with MOPSO at pH 6.5). Infiltration of MES buffer set to a pH of 4.7 (osmotic control) did not cause a significant increase in ABA concentration (Fig. 3).

Transient leaf apoplastic alkalinization reduces stomatal aperture, stomatal conductance, transpiration, and photosynthesis
At 60 min after infiltration of the pH buffers (this time point is shown at minute 100 on the x-axis of Fig. 1) or immediately after the re-acidification (this time point is shown at minute 200 on the x-axis of Fig. 1), neither the stomatal aperture ( Fig. 3B) nor the stomatal conductance or the transpiration rate (Supplementary Table S3) was influenced by the infiltration of the MES and MOPSO buffer solutions set to a pH of 6.5 (when compared with the control that was infiltrated with water). At 2 h after the apoplast had re-acidified to a pH of 4.7 in response to the infiltration of the MES buffer at pH 6.5 (this time point is shown at minute 320 on the x-axis of Fig. 1), however, both the stomatal conductance (128±25 mmol H 2 O m -2 s -1 ) and the transpiration rate (2.3±0.4 mmol H 2 O m -2 s -1 ) were significantly lower compared with plants from the control group. In controls, g s mmol H 2 O m -2 s -1 was 283±45 and E was 5.8±1.1 mmol H 2 O m -2 s -1 , (Supplementary Table  S3). At the same time point, the stomatal aperture significantly decreased from 108.2±12.8 µm 2 to 34.8±8.1 µm 2 in response to the MES-induced apoplastic alkalinization (Fig. 3B). Thus, it was significantly lower compared with the control. At 4 h after re-acidification (this time point is shown at minute 440 on the x-axis of Fig. 1), the transpiration rate (1.8±0.3 mmol H 2 O m -2 s -1 ), the stomatal conductance (97±26 mmol H 2 O m -2 s -1 ), and the stomatal aperture (26.0±5.3 µm 2 ) were still significantly lower in plants from the group 'MES pH apo 6.5' when compared with plants from the control (in the control, E was 6.1±0.9 mmol H 2 O m -2 s -1 , g s was 294±48 mmol H 2 O m -2 s -1 , and stomatal aperture was 123.4±14.2 µm 2 ). Similar responses were observed when the leaf apoplast was transiently alkalinized by using the MOPSO buffer set to pH 6.5, when compared with the control. The transpiration rate significantly decreased to 1.7±0.1 mmol H 2 O m -2 s -1 or 1.9±0.3 mmol H 2 O m -2 s -1 at 2 h or 4 h after re-acidification. Stomatal conductance significantly decreased to 85.3±14 mmol H 2 O m -2 s -1 or 102.26±26 mmol H 2 O m -2 s -1 at 2 h or 4 h after re-acidification (Supplementary Table S3). Stomatal aperture significantly decreased to 24.3±6.7 µm 2 or 23.1±4.4 µm 2 at 2 h or 4 h after re-acidification (Fig. 3B). In good agreement with the stomatal closure that was elicited by the apoplastic alkalinization, the rate of photosynthesis also decreased in response to the alkalinization at the same time points (Supplementary Table S3). Use of osmotic controls revealed that all these effects were not related to the osmotic properties of the buffer solutions: infiltration of a MES buffer set to a pH of 4.5 did not significantly affect stomatal aperture (Fig. 3B) or stomatal conductance and rate of transpiration (Supplementary Table S3).

The pH apo -induced increase in ABA correlates positively with the decrease in g s
The Pearson correlation coefficient shows a strong statistical linear relationship between the increasing leaf ABA concentration and the decreasing stomatal conductance only when the apoplast was alkalinized (Pearson r= -82 for 'MES pH apo 6.5'; Pearson r= -95 for 'MOPSO pH apo 6.5'; data not shown). There was a much weaker (Pearson r= -59) or no (Pearson r=0.06) statistical relationship between both parameters in plants of the osmotic control or the water control, respectively (data not shown). The pH apo effect on both stomatal conductance and ABA concentration becomes evident when fitting a single relationship between data points from all four experimental groups: Pearson r is -97 (Fig. 4) whereby plants from experimental groups with a re-acidified leaf apoplast ('MES pH apo 6.5', 'MOPSO pH apo 6.5') clearly clustered away from groups where pH apo did not change over the experiment ('control' or 'osmotic control").

Discussion
Soil salinization is a big constraint for plants because, among other mechanisms, it reduces the availability of water in the soil. Thus, salt and water stress have much in common (Munns, 2002). To maintain turgor and tissue hydration, the plant adjusts its transpiration. With regard to this, the pH of the apoplast is considered to act as a long-distance signal that transmits information about the decreasing soil water availability from root to shoot (Wilkinson, 1999;Wilkinson and Davies, 2008;Geilfus et al., 2015). Upon arrival in the shoot, it regulates stomatal aperture by acting on the compartmental distribution of ABA between the shoot apoplast and symplast (Wilkinson, 1999;Davies et al., 2002). In addition to this Henderson-Hasselbalch-regulated partitioning of ABA (Sharp and Davies, 2009), the present study demonstrates for the first time that the stomatal closure, elicited by the apoplastic alkalinization, is also due to a pH apo -mediated increase in the transcription of the NCED gene Vp14 in the leaf. ZmVp14 encodes the ratelimiting enzyme in ABA biosynthesis (Seo and Koshiba, 2002).
First of all, the study aimed at clarifying whether this leaf pH apo transient is only functional with regard to the initiation of stomatal closure when interacting in concert with chloride ions or other chloride stress-related responses or variables that are induced by chloride salinity. These could include changes in [Ca 2+ ], [H 2 O 2 ], [ATP], or membrane potential. Following the concept of network signalling (Jordan, et al., 2000;Plieth, 2016), all these variables (signals) would be integrated at downstream junctions to modulate the output (i.e. stomatal closure). Alternatively, the pH apo transient alone might initiate a cascade of events that regulates aperture following the concept of single-file signalling (see Plieth, 2016). The latter proposition suggests that the pH apo transient acts independently from other chloride stress-associated responses that are induced by chloride salinity.
To clarify this, we mimicked the transient alkalinization of the leaf apoplast, as occurs under conditions of chloridestress (of course, without adding chloride stress), by infiltrating nearneutral pH buffers (pH 6.5) into the leaf apoplast. By these means, we investigated the effect of the pH apo transient independently from any chloride stress-derived responses; that is, unrelated to the excessive accumulation of chloride (or its accompanying cation). If the hypothesis that the pH apo transient alone is able to initiate stomatal closure without interacting with other chloride stress-related responses is true, then we would expect that guard cell aperture and stomatal conductance would decrease in response to mimicking the chlorideinduced pH apo transient.

pH apo transient increases transcript abundance of the Vp14 gene in maize leaves
To stimulate transient leaf apoplastic alkalinization in a similar way to that observed for chloride salinity, a MES-based pH buffer (pH 6.5) was infiltrated into the maize leaf apoplast (Fig. 1). To link this mimicked pH apo dynamic with the synthesis of ABA, the transcript abundance of the ABA-synthesizing gene ZmVp14 (Tan et al., 2003) was analysed. Relative to the control, the ZmVp14 transcript abundance increased by up to 52-fold (Fig. 2). To demonstrate that this effect was due to changes in the pH apo and not from the buffering agent MES, a different buffer (MOPSO set to pH 6.5) was infiltrated in a repeat experiment. Relative to the control, ZmVp14 transcript abundance increased by up to 68-fold (Fig. 2). Infiltration of an osmotic control (i.e. 7.5 mM MES solution that was set to a pH of 4.7) showed that it is the apoplastic pH transient that is responsible for the fast and steep increase in the transcription of the Vp14 gene in maize, being unrelated to the osmotic properties of the buffer agents (Fig. 2). Any mechanical effects arising from the infiltration procedure or from flooding the apoplast with water can be excluded because controls were infiltrated with water and the above-mentioned fold changes were expressed relative to these controls. It has long been assumed that xylem-transported ABA, which was synthesized in the roots, is of major relevance for the shoot ABA pool when the plants experiences salt stress at the roots. The presented data on Vp14 gene transcription demonstrated that the leaf is the key site of ABA synthesis, challenging the paradigm that the vast amount of the foliar-located ABA is produced in the roots. Reciprocal grafting experiments with ABA biosynthetic mutant and WT plants suggest that there is a signal that carries the information of a declining root water potential through the xylem to the leaves (Holbrook et al., 2002;Christmann et al., 2007). Upon arrival, leaf hydration, leaf ABA status, and stomatal aperture are adjusted (Manzi et al., 2015;McAdam et al., 2016a, b, c). The leaf apoplastic alkalinization, that travels systemically and acropetally through the shoot (Geilfus et al., 2015), appears to be a candidate for transmitting information about the onset of water scarcity because it induces leaf ABA synthesis (Figs 3, 4) and elicits stomatal closure (Supplementary Table S3).

Transient apoplastic alkalinization modulates ABA and increases it in maize leaves
Next, we investigated whether the artificially induced leaf apoplastic alkalinization was associated with a rise in leaf ABA concentration. A shift in leaf apoplastic pH to 6.5 by means of MES and MOPSO infiltration caused the ABA concentrations steadily to increase steadily over the entire experiment relative to the water control and the osmotic control (Fig. 3). A temporal comparison (compare Fig. 2 and Fig. 3A) reveals that the increase of the Vp14 mRNA abundance preceded the increase of the ABA concentration. This indicates that the increase in leaf ABA was caused by increased biosynthesis. However, the means by which the pH apo transient is associated with Vp14 expression awaits clarification. Is a pH sensor involved? Such a putative sensor would need a pH-responsive (apoplasticlocated) domain as part of an integral protein that spans the entirety of the PM. It would sense changes in the pH apo or the transmembrane pH gradient, passing on the information to start signal transduction via intermediate cellular transmitters that finally involve Vp14 promoter elements. Because the existence of such a sensor is unclear, the elucidation of this problem is extremely difficult. Thus, other explanations have to be taken into consideration. For instance, large changes in the pH apo are known to be able to influence the cytosolic pH (Monshausen et al., 2007). Since cytosolic pH dynamics are accompanied by cellular Ca 2+ transients (Monshausen et al., 2008(Monshausen et al., , 2009Michard et al., 2016;Dindas et al., 2018) and have been implicated in the specificity of Ca 2+ signalling (Behera et al., 2018), large transients in the pH apo might affect ABA transcription via cellular signal transduction. However, such a sequence of events remains speculative unless a clear link between pH apo transients, Ca 2+ transients, and the induction of Vp14 gene expression is established. Currently, the transcription of the NCED gene is known to be regulated by light (study on tomato by Thompson et al., 2000), salinity and osmotic stress (study on maize by Geilfus et al., 2018), water deficit (study on maize by Tan et al., 1997), and drought (study on Citrus by Agustí et al., 2007).
Apoplastic alkalinization reduces stomatal aperture and transpiration rate Wilkinson and Davies (2008) treated leaves of an ABAdeficient tomato mutant with alkaline-buffered foliar sprays and demonstrated that an increase in the leaf pH apo reduced stomatal conductance only when ABA was simultaneously sprayed onto the leaves of the ABA-deficient tomato. They and others postulated that the apoplastic alkalinization required ABA to act on stomatal aperture (Wilkinson andDavies, 1997, 2008;Sobeih et al., 2004;Else et al., 2006;Jia and Davies, 2007) and suggested a pH apo -based mechanism whereby the rise in pH apo increased the amount of ABA that penetrated into the guard cells by changing the compartmental distribution of ABA between the apoplast, the symplast, and various symplastic components.
The presented study adds knowledge about this system as it demonstrates that it is indeed the leaf apoplastic alkalinization that induces Vp14 gene expression (Fig. 2), presenting a further mechanism that causes the foliar ABA level to rise (Fig. 3A) and stomatal conductance to decline (Fig. 4). This novel finding provides a compelling explanation for the observed stomatal closure and the decline in photosynthetic rate (Supplementary  Table S3) elicited by the apoplastic alkalinization.
Apoplastic alkalinization can also be argued to induce stomatal closure directly via pH apo -based effects on guard cell K + fluxes. A rising pH in the stomatal cavity decreases the activity of guard cell-localized inwardly rectified K + channels and increases the activity of guard cell-localized outwardly rectified K + channels (Hedrich et al., 1995;Ache et al., 2000). However, two arguments oppose the assumption that stomata close because of the direct pH apo -based stimulation of outwardly rectified guard cell K + channels, highlighting the role of a pH apo -based effect of ABA biosynthesis. First, both the stomatal aperture and the transpiration rate remained reduced at 4 h after re-acidification (Supplementary Table S3). During these 4 h, the plants were continuously illuminated with white light. As light increases the activity of inwardly rectified K + channels (Roelfsema and Hedrich, 2005;Marten et al., 2007), enough time would have been available for a light-induced K + influx to mediate the re-opening of the stomata. However, the stomatal aperture remained reduced. Second, the stomatal aperture and transpiration remained reduced, although the apoplast had become acidified once again, over the 4 h, to a pH apo range that favoured K + influx into guard cells via inwardly rectified K + channels; such changes could ultimately result in guard cell re-opening, if not blocked by ABA.

Conclusions
We have shown that an artificially induced transient of the pH of the leaf apoplast initiates stomatal closure and reduces the rate of leaf transpiration (Supplementary Table S3). A pH apobased de novo synthesis of the guard cell-regulating hormone ABA in leaves (Fig. 2) and a pH apo -based increase of leaf ABA concentration (Fig. 3) may be the reason for the reduced leaf transpiration rate; that is, the reduction in stomatal conductance (Fig. 4). Our first novel finding is that the pH of the apoplast increases the transcription of Vp14, a gene that is key for ABA synthesis in maize. The second novelty is that the functionality of the pH apo transient with respect to reducing the rate of transpiration does not depend on interactions with compounds that are induced by chloride stress. These findings are relevant, since the transient alkalinization of the leaf apoplast is a widespread phenomenon that occurs not only under chloride salinity, but also in response to drought (Bacon et al., 1998) or leaf infections with Blumeria graminis (Felle et al., 2008) or Piriformospora indica (Felle et al., 2009). Future studies should compare plant species that show the apoplastic alkalization with plant species that do not, in order to understand the underlying mechanism(s) of the transient shift in the apoplastic pH.

Supplementary data
The following supplementary data are available at JXB online. Table S1. Effect of infiltration on photosynthetic rate. Table S2. Specificity of the ZmVp14 primer pair was demonstrated by sequencing the real-time quantitative RT-PCR product. Table S3. pH apo transient reduces transpiration rate, stomatal conductance, and photosynthetic rate Fig. S1. Specificity of the Vp14 primer pair.