Identification and Characterization of Compounds that Affect Stomatal Movements

Abstract

Regulation of stomatal aperture is essential for plant growth and survival in response to environmental stimuli. Opening of stomata induces uptake of CO₂ for photosynthesis and transpiration, which enhances uptake of nutrients from roots. Light is the most important stimulus for stomatal opening. Under drought stress, the plant hormone ABA induces stomatal closure to prevent water loss. However, the molecular mechanisms of stomatal movements are not fully understood. In this study, we screened chemical libraries to identify compounds that affect stomatal movements in Commelina benghalensis and characterize the underlying molecular mechanisms. We identified nine stomatal closing compounds (SCL1–SCL9) that suppress light-induced stomatal opening by >50%, and two compounds (temsirolimus and CP-100356) that induce stomatal opening in the dark. Further investigations revealed that SCL1 and SCL2 had no effect on autophosphorylation of phototropin or the activity of the inward-rectifying plasma membrane (PM) K⁺ channel, KAT1, but suppressed blue light-induced phosphorylation of the penultimate residue, threonine, in PM H⁺-ATPase, which is a key enzyme for stomatal opening. SCL1 and SCL2 had no effect on ABA-dependent responses, including seed germination and expression of ABA-induced genes. These results suggest that SCL1 and SCL2 suppress light-induced stomatal opening at least in part by inhibiting blue light-induced activation of PM H⁺-ATPase, but not by the ABA signaling pathway. Interestingly, spraying leaves onto dicot and monocot plants with SCL1 suppressed wilting of leaves, indicating that inhibition of stomatal opening by these compounds confers tolerance to drought stress in plants.

Introduction

Stomata in the plant epidermis play a pivotal role in regulation of gas exchange between leaves and the atmosphere. Opening of the stomata allows both CO₂ entry for photosynthesis and transpiration to promote nutrient uptake in roots. Stomata open in response to several environmental stimuli, such as blue light, red light, low CO₂ and the fungal toxin fusicoccin (FC) (Shimazaki et al. 2007, Negi et al. 2014, Inoue and Kinoshita 2017). Blue light has a pronounced effect on stomatal opening. Phototropins, which are receptors for blue light, activate the plasma membrane (PM) H⁺-ATPase by phosphorylating the penultimate residue, threonine, at the C-terminus and subsequent binding of the 14-3-3 protein to the phosphorylated C-terminus (Kinoshita and Shimazaki 1999, Kinoshita et al. 2001, Inoue et al. 2008). The blue light-activated PM H⁺-ATPase increases an inside-negative electrical potential across the PM and drives K⁺ uptake through voltage-gated inward-rectifying K⁺ channels (Schroeder et al. 1987). Accumulation of K⁺ induces swelling of guard cells, resulting in stomatal opening (Supplementary Fig. S1). Recent genetic investigations using loss-of-function mutants of PM H⁺-ATPase revealed that the PM H⁺-ATPase is essential for blue light-induced stomatal opening in plants (Toda et al. 2016, Yamauchi et al. 2016). Moreover, enhancement of stomatal opening by overexpression of PM H⁺-ATPase in guard cells increases photosynthetic activity and growth in Arabidopsis thaliana (Wang et al. 2014), indicating that stomatal aperture is a limiting factor for photosynthesis and plant growth.

Several signaling components between phototropins and PM H⁺-ATPase activation have been identified (Supplementary Fig. S1). The novel protein kinase BLUS1 binds to, and is phosphorylated by, phototropins (Takemiya et al. 2013a). The Raf-like kinase BHP interacts with BLUS1 and mediates blue light signaling between BLUS1 and PM H⁺-ATPase activation (Hayashi et al. 2017). Type 1 protein phosphatase (PP1) and its regulatory subunit PRSL1 are also reportedly involved in this signaling pathway (Takemiya et al. 2006, Takemiya et al. 2013b). The plant hormone brassinosteroid affects the expression level of inward-rectifying K⁺ channels in guard cells (Inoue et al. 2017). However, the signaling mechanism involved in blue light-induced stomatal opening is still largely unknown.

Under drought stress, the plant hormone ABA, which is synthesized in response to drought stress, promotes stomatal closure to prevent water loss (Schroeder et al. 2001, Kollist et al. 2014). The PYR/PYL/RCAR ABA receptors activate SnRK2 family protein kinases by inactivating the central negative regulator type 2 C protein phosphatases (PP2Cs) in response to ABA (Ma et al. 2009, Park et al. 2009, Santiago et al. 2009, Cutler et al. 2010), which leads to activation of the slow-type anion channel SLAC1 (Negi et al. 2008, Vahisalu et al. 2008) followed by depolarization of the PM (Lee et al. 2009, Geiger et al. 2009). Next, depolarization-dependent activation of outward-rectifying K⁺ channels in the PM induces an efflux of K⁺ from guard cells and stomatal closure (Schroeder et al. 1987, Kim et al. 2010, Joshi-Saha et al. 2011). At the same time, ABA suppresses blue light-induced activation of PM H⁺-ATPase by inducing dephosphorylation of the penultimate residue, phosphorylated threonine (Zhang et al. 2004, Hayashi et al. 2011), and the activity of voltage-gated inward-rectifying K⁺ channels (Schwartz et al. 1994, Wang et al. 2013), resulting in acceleration of stomatal closure. Because PP1, a positive regulator of blue light signaling, is inhibited by phosphatidic acid (PA), which is induced by ABA, PA may play an important role in ABA and blue light signaling cross-talk in guard cells (Takemiya and Shimazaki 2010). The PYR/PYL/RCAR ABA receptors mediate not only stomatal closure but also seed germination, root growth and gene expression (Nishimura et al. 2010, Park et al. 2009). Notably, overexpression of the Mg-chelatase H subunit, a Chl biosynthesis enzyme that enhances ABA signaling in guard cells, in the epidermis enhances ABA sensitivity and confers drought tolerance in A. thaliana (Tsuzuki et al. 2013), indicating that stomatal aperture has a marked influence on drought tolerance in plants.

Chemical-based analysis enables elucidation of the molecular mechanism of signaling. It can overcome the obstacles of lethal genes or high redundancy in gene families, which are critical problems in classical genetics approaches, and can also facilitate development of useful compounds, such as agrochemicals (Hicks and Raikhel 2014, Dejonghe and Russinova 2017). Pyrabactin, a small-molecule ABA agonist, is the best breakthrough example whuch led to the identification of ABA receptors (Park et al. 2009) and development of drought tolerance agents using engineered ABA receptors (Park et al. 2015). Analyses of the activity of the fungal toxin FC allowed clarification of the regulatory mechanism of PM H⁺-ATPase, a key enzyme for stomatal opening (Haruta et al. 2015). FC stabilizes binding of the phosphorylated penultimate threonine residue in the PM H⁺-ATPase and the 14-3-3 protein, resulting in inhibition of dephosphorylation of PM H⁺-ATPase and accumulation of phosphorylated PM H⁺-ATPase in stomatal guard cells (Kinoshita and Shimazaki 2001). Furthermore, chemical screening revealed that the small molecule [5-(3,4-dichlorophenyl)-2-furyl](1-piperidinyl)methanethione down-regulates ABA-dependent gene expression and inhibits ABA-induced stomatal closure (Kim et al. 2011). Small molecules that increase the number of stomata in A. thaliana have been identified by screening of chemical libraries (Sakai et al. 2017, Ziadi et al. 2017). However, to our knowledge, chemical screening for compounds that regulate stomatal movements has not been performed.

In this study, we performed chemical screening for compounds that regulate stomatal movements using Commelina benghalensis. We identified nine compounds that suppress light-induced stomatal opening and were designated stomatal closing compounds (SCLs), and two compounds that induce stomatal opening in the dark. Furthermore, we investigated the molecular mechanisms of action of these compounds, and found that SCL1 may confer drought stress tolerance in plants.

Results

Screening of compounds that affect stomatal movements

To identify compounds that affect stomatal movements, we screened 90 compounds from the LOPAC^® Pfizer chemical library (Sigma-Aldrich), 20,101 compounds from the ITbM Original Library and 84 compounds from the redox library (ENZO Life Sciences). Fig. 1A shows a schematic of the chemical screening using C. benghalensis leaf discs. Commelina benghalensis has almost 2-fold larger stomata than the model plant A. thaliana (Fig. 1B), which facilitates measurement of the stomatal aperture at low magnification. In addition, C. benghalensis shows normal stomatal responses to light and ABA like A. thaliana (Murthy et al. 1984, Hayashi et al. 2011). First, we excised a leaf disc or peeled epidermis from dark-adapted C. benghalensis leaves, and incubated the samples with basal buffer containing the test compounds in a multi-well plate under light or in the dark for 4 h. Then, we qualitatively assessed the stomatal aperture under a microscope and identified candidate compounds (first screening). Next, we confirmed the reproducibility of the first screening by performing a second screening, and quantified stomatal aperture diameter by capturing images in a third screening (see the Materials and Methods). This led to identification of nine compounds that inhibited light-induced stomatal opening; these were designated SCLs. SCL1 was identified from the redox library, and SCL2–SCL9 were identified from the ITbM Original Library. We also identified two compounds from the LOPAC library, temsirolimus (Tem) and CP-100356 (CP), which induce stomatal opening in the dark (Supplementary Table S1). Note that no compound had an effect on the viability of guard cells in C. benghalensis, because epidermal fragments that had been incubated with and without compounds for 4 h were stained similarly by fluorescein diacetate (FDA) (Supplementary Fig. S2).

Fig. 1

Screening for compounds that affect stomatal movement. (A) Schematic representation of the phenotype-based screen. Excised leaf discs or epidermis from C. benghalensis were immersed in wells of a multi-well plate containing basal buffer with the test compounds and incubated under light or in the dark for 4 h. We identified compounds that inhibit light-induced stomatal opening or promote stomatal opening in the dark. In the first and second screenings, we qualitatively evaluated the bioactivity of candidate compounds to maximize throughput. Measurement of stomatal aperture (red bidirectional arrows) was performed during the third screening to quantify the bioactivity of the compounds. The protocol is provided in the Materials and Methods. (B) Photographs of epidermis from A. thaliana and C. benghalensis. Arrowheads indicate the positions of stomata. Scale bars = 50 µm.

Suppression of light- and FC-induced stomatal opening by the nine compounds

The structures of SCL1–SCL9 were diverse (Fig. 2A). All of the compounds (50 μM) suppressed light-induced stomatal opening by >50% in C. benghalensis (Fig. 2B). Among them, SCL1 at 20 µM suppressed light-induced stomatal opening most effectively. Next, we investigated the half-maximal inhibitory concentrations (IC₅₀) of the top four compounds for light-induced stomatal opening, in addition to ABA (Fig. 2D). The IC₅₀ values were: 0.36 µM for ABA, 4.62 µM for SCL1, 12.50 µM for SCL2, 20.24 µM for SCL3 and 9.00 µM for SCL4.

Fig. 2

Stomatal closing compounds (SCLs) suppressed light- and fusicoccin (FC)-induced stomatal opening in C. benghalensis. (A) Chemical structures of SCL1–SCL9. (B) Inhibition of light-induced stomatal opening. The test compounds (50 μM) were added 30 min prior to light exposure. (C) Competitive inhibition of FC-induced stomatal opening by SCLs. SCLs and FC were added simultaneously, and samples were incubated in the dark for 3 h. (B and C) Means±SE (n =4; 30 stomata per four leaf discs per replicate). Light treatment, 150 µmol m^–2 s^–1 red light and 50 µmol m^–2 s^–1 blue light. ABA, 20 µM ABA; FC, 10 µM FC. (D) IC₅₀ values of ABA, SCL1, SCL2, SCL3 and SCL4 for light-induced stomatal opening (n≥4; 30 guard cells per leaf disc). Asterisks indicate significant differences compared with DMSO treatment (Student’s t-test, P < 0.05).

In addition, we investigated the effect of the nine compounds on FC-induced stomatal opening (Fig. 2B). SCL1 markedly suppressed FC-induced stomatal opening; in contrast, the other eight compounds only partially suppressed FC-induced stomatal opening.

Effects of SCL1 and SCL2 on phototropin kinase activity, KAT1 activity and PM H⁺-ATPase phosphorylation

To clarify the modes of action of SCL1 and SCL2 in guard cells, we first investigated the effect of SCL1 and SCL2 on the kinase activity of the blue light receptor phototropin in guard cell protoplasts (GCPs) of Vicia faba. Phototropin is autophosphorylated in response to blue light; this is followed by binding of a 14-3-3 protein to autophosphorylated phototropin (Kinoshita et al. 2003, Inoue et al. 2008). Blue light induced autophosphorylation of phototropin, as detected by binding of 14-3-3 protein by Western blot (Fig. 3A). SCL1 and SCL2 at 50 µM had no effect on autophosphorylation of phototropin. The kinase inhibitor staurosporin markedly suppressed the blue light-induced autophosphorylation of phototropin (Kinoshita et al. 2003). Next, we investigated the effect of the SCLs on the activity of an inward-rectifying PM K⁺ channel, KAT1, in Xenopus oocytes (Uozumi et al. 1995). All nine SCLs at 20 µM had no effect on KAT1 activity (Fig. 3B;Supplementary Fig. S3A). These results suggest that SCL1 and SCL2 suppress light-induced stomatal opening without affecting phototropin and KAT1 activities in guard cells.

Fig. 3

Characterization of SCLs. (A) Effect of SCL1 and SCL2 on blue light-induced autophosphorylation of phototropin in guard cell protoplasts (GCPs) of V. faba. GCPs were treated with dimethylsulfoxide (DMSO), SCL1 or SCL2 and light irradiated. Autophosphorylation of phototropin was evaluated by Western blotting using 14-3-3 protein as a probe. The 14-3-3 protein was used as a loading control. RL, 80 µmol m^–2 s^–1 red light for 30 min; RB, 20 µmol m^–2 s^–1 blue light for 1 min superimposed on RL. (B) Effect of SCL1 and SCL2 on KAT1 activity. KAT1 activity in Xenopus oocytes was determined. (C) Relative intensities of phosphorylated plasma membrane (PM) H⁺-ATPase in guard cells of A. thaliana, as determined by immunohistochemical staining (see the Materials and Methods for details). Arabidopsis epidermis was treated with DMSO (DMSO B) or 50 µM SCLs (SCL1 and SCL2) and irradiated with 15 µmol m^–2 s^–1 blue light plus 600 m^–2 s^–1 red light for 2.5 min. DMSO D, dark-incubated DMSO sample. Means±SE (n =3). (D) Competitive inhibition of FC-induced PM H⁺-ATPase phosphorylation in A. thaliana by SCLs. Arabidopsis epidermis was treated with the indicated compounds and incubated in the dark, and the PM H⁺-ATPase phosphorylation level was quantified as described in (C). FC, 10 µM FC; SCL1, FC + 50 µM SCL1; SCL2, FC + 50 µM SCL2. Asterisks indicate significant differences compared with DMSO B or DMSO + FC (Student’s t-test, P < 0.02).

Next, we investigated the effect of SCL1 and SCL2 on blue light-induced phosphorylation of PM H⁺-ATPase in guard cells of A. thaliana (Fig. 3C). Blue light-induced phosphorylation of PM H⁺-ATPase was assayed immunohistochemically using an antibody against the penultimate residue, phosphorylated threonine (Hayashi et al. 2011). Interestingly, blue light-induced phosphorylation of PM H⁺-ATPase was completely suppressed by SCL1, and reduced by 50% by SCL2. These results suggest that SCL1 and SCL2 suppress light-induced stomatal opening, at least in part by inhibiting blue light-induced activation of the PM H⁺-ATPase. Consistent with the inhibition of FC-induced stomatal opening, FC-induced phosphorylation of PM H⁺-ATPase was also suppressed by SCL1 and SCL2 (Fig. 3D).

Effects of SCL1 and SCL2 on ABA-related responses

ABA reportedly inhibits light-induced stomatal opening and blue light-dependent phosphorylation of the penultimate residue, threonine, in PM H⁺-ATPase (Hayashi et al. 2011, Wang et al. 2014). Therefore, it is possible that SCL1 and SCL2 suppressed light-induced stomatal opening by affecting the ABA signaling pathway. Next, we investigated the effect of SCL1 and SCL2 on ABA-related responses, such as germination of A. thaliana seeds (Walton 1980), expression of ABA-induced genes in A. thaliana (Tomiyama et al. 2014) and the phosphorylation status of a 61 kDa protein, which is phosphorylated by ABA, in V. faba GCPs (Takahashi et al. 2007). ABA at 50 µM completely suppressed germination of A. thaliana seeds, but the nine SCLs at 50 µM did not (Fig. 4A;Supplementary Fig. S3B). Furthermore, SCL1 and SCL2 did not induce expression of the ABA-induced genes, RAB18 and RD29B, in Arabidopsis seedlings or phosphorylation of the 61 kDa protein in GCPs from V. faba, as determined by Western blot of 14-3-3 protein binding (Fig. 4B, C). These results suggest that suppression of light-induced stomatal opening by SCL1 and SCL2 is not mediated by the ABA-dependent pathway.

Fig. 4

Effects of SCL1 and SCL2 on ABA-related responses. (A) Effects of SCL1 and SCL2 on germination of A. thaliana seeds. Seeds were treated with water containing 50 µM ABA, SCL1 or SCL2, and incubated in the dark. Germination ratios of seeds at 7 d post-treatment. Means±SD (n =3; 30 seeds per replicate). (B) Effect of SCL1 and SCL2 on the expression level of the ABA-responsive genes, RAB18 and RD29B, in Arabidopsis seedlings. The seedlings were treated with 50 µM ABA, SCL1 or SCL2 and an equal volume of DMSO at 24°C for 3 h. Means±SD (n =3; three whole seedlings per replicate). (C) Effects of SCL1 and SCL2 on the phosphorylation status of the 61 kDa protein in GCPs of V. faba. The 14-3-3 protein was used as a loading control. The GCPs were treated with 20 µM ABA, or 50 µM SCL1 or SCL2, at 24°C for 20 min. Asterisks indicate significant differences compared with DMSO treatment (Student’s t-test, P < 0.01).

Effects of SCL1 and SCL2 derivatives on light-induced stomatal opening

To investigate the structure–activity relationship of SCL1 and SCL2, we obtained commercially available SCL1 and SCL2 derivatives and investigated their effects on light-induced stomatal opening (Fig. 5). SCL1-1 (2,5-di-tert-butylbenzene-1,4-diol) markedly suppressed light-induced stomatal opening as efficiently as SCL1, and SCL1-2 (hydroquinone) partially suppressed light-induced stomatal opening. These results indicate that a tertiary-butyl substituent may be important for target binding, but di-tertiary-butyl is too bulky. Also, the combination of a hydroxyl group and tertiary-butyl substituent in SCL1 is important for its inhibition of light-induced stomatal opening.

Fig. 5

Comparisons of SCL1 and SCL2 derivatives. (A) Chemical structures of SCL1 and SCL2 derivatives. Moieties unique between related compounds are highlighted. (B) Inhibition of light-induced stomatal opening in C. benghalensis. The compounds (50 μM) were added 30 min prior to light exposure for 4 h. Stomatal aperture diameters are relative to that caused by DMSO treatment. Light treatment, 150 µmol m^–2 s^–1 red light and 50µmol m^–2 s^–1 blue light. ABA, 20 µM ABA. Means ± SE (n = 3; 30 stomata per leaf disc per replicate). Asterisks indicate significant differences compared with DMSO treatment (Student’s t-test, P<0.01).

SCL2 possesses 4-bromo-pyridazinone (A-ring) and 2,6-dimethyl-piperidine (B-ring). SCL2-1 (4,5-dibromopyridazinone) has only the A-ring part of SCL2 with the addition of a 5-methyl group, and showed no inhibition of light-induced stomatal opening. Furthermore, 4-chloro-5-morpholinopyridazinone (SCL2-2), 4-bromo-5-(2-ethylmorpholino)-2-methylpyridazinone (SCL2-3) and 4-chloro-2-methyl-5-morpholinopyridazinone (SCL2-4) had no effect on light-induced stomatal opening (Fig. 5). These results suggest that the pyridazinone and piperidine rings are required for the activity of SCL2.

SCL1 confers drought resistance on plants

SCLs suppressed light-induced stomatal opening (Fig. 2), suggesting that they confer drought resistance on plants by suppressing light-induced stomatal opening and enhancing stomatal closure. To investigate whether SCLs confer drought resistance, we sprayed SCL1 onto rose leaves, and stored detached rose leaves for a short period (Fig. 6A). Interestingly, wilting of SCL1-sprayed rose leaves was suppressed. Furthermore, spraying of SCL1 onto oat leaves, the guard cells of which are dumbbell shaped, also suppressed wilting of the detached leaves (Fig. 6B). These results suggest that SCL1 enhances the drought tolerance of both dicot and monocot plants.

Fig. 6

Effect of SCL1 on leaf wilting. Rose leaves in a bouquet (A) and 7-day-old oat seedlings (B) were sprayed with 0.5% DMSO or 100 µM SCL1 in 0.02% Silwet L77 and 0.05% Approach BI, and incubated at 24°C under 50 µmol m^–2 s^–1 white fluorescent light and 70% relative humidity for 3 h. Then, the leaves were excised and incubated for 6 h (rose) and 20 min (oat) at 24°C under 50 µmol m^–2 s^–1 white fluorescent light and 35–50% relative humidity. All experiments were repeated in triplicate and yielded similar results. Scale bars = 1 cm.

Temsirolimus and CP-100356 induced stomatal opening in the dark

We identified Tem, an inhibitor of mammalian target of rapamycin (mTOR) (Peralba et al. 2003) and CP, a high-affinity P-glycoprotein/MDR-1 inhibitor (Kajiji et al. 1994), from the LOPAC^® Pfizer library as inducers of stomatal opening in the dark (Fig. 7A). The magnitude of opening of stomatal apertures in the epidermis of C. benghalensis induced by Tem and CP at 50 µM was about half that induced by FC at 10 µM (Fig. 7B, C). Tem- and CP-dependent stomatal opening was significantly suppressed in the presence of ABA at 20 µM, but FC-dependent stomatal opening was insensitive to ABA (Fig. 7D). It is possible that Tem and CP induce stomatal opening in different manners from FC. Further investigations will be needed to clarify the molecular mechanisms of these compounds.

Fig. 7

Temsirolimus (Tem) and CP-100356 (CP) promote stomatal opening in C. benghalensis in the dark. (A) Chemical structures of Tem and CP. (B) Representative images of stomata in epidermis treated with 50 µM Tem or CP in the dark. Scale bars = 50 µm. (C) Tem- and CP-induced stomatal opening. Epidermal peels were treated with 50 µM Tem, 50 µM CP or 10 µM FC in the dark for 4 h. Asterisks indicate significant differences compared with DMSO treatment (Student’s t-test, P < 0.01). (D) Competitive inhibition of ABA-induced stomatal closure by Tem and CP. Epidermal peels were treated with the indicated compounds in the absence or presence of 20 µM ABA. Tem and CP, 50 µM; FC, 10 µM. Means ± SE (n=3; 45 stomata were measured per replicate). Asterisks indicate significant differences compared with ABA treatment (Student’s t-test, P < 0.01).

We investigated the effects of analogs of the mTOR inhibitors rapamycin (Rap), everolimus (Eve), ridaforolimus (Rid) and umirolimus (Umi) on stomatal aperture (Ballou and Lin 2008, Guertin and Sabatini 2009) (Fig. 8A). These analogs have different side chains at the C40 position, and have improved pharmacokinetics and reduced immunosuppressive effects in vivo. Eve and Umi, but not Rap and Rid, at 50 µM induced stomatal opening in the dark, similar to Tem (Fig. 8B, C).

Fig. 8

Effect of other mammalian target of rapamycin (mTOR) inhibitors on stomatal aperture in C. benghalensis. (A) Chemical structures of mTOR inhibitors. Temsirolimus (Tem), everolimus (Eve), ridaforolimus (Rid) and umirolimus (Umi) differ from rapamycin (Rap) only in the red dotted region in Rap. Unique moieties are highlighted. (B and C) Effect of (B) Rap and (C) Eve, Rid and Umi on stomatal apertures in the dark. Epidermal peels were treated with the indicated inhibitors at 50 µM and incubated in the dark for 4 h. Abbreviations are as in (A). FC, 10 µM FC. Means ± SE (n=3; 45 stomata were measured per replicate). Asterisks indicate significant differences compared with DMSO treatment (Student’s t-test, P < 0.01).

Discussion

Chemical screening for compounds that affect stomatal movements

In the present study, we screened for compounds that affect stomatal movements using chemical libraries containing a total of 20,275 compounds. We identified nine compounds (SCL1–SCL9) that suppress light-induced stomatal opening (Fig. 2), and two compounds (Tem and CP) that induce stomatal opening in the dark (Fig. 7). To our knowledge, this is the first study to screen comprehensively for compounds that affect stomatal movements. Our analysis of these compounds provides novel insight into regulation of stomatal movements. Such chemical screenings overcome gene essentiality and the high redundancy in gene families, which are critical problems with classical genetics approaches.

In order to maximize the screening throughput, we devised the chemical screening as follows. First, we used C. benghalensis as a plant material not only because of its large stomatal size, but for its fast growth speed and its ability to regenerate from stem cuttings. These unique characteristics made it possible to use leaves for daily assay that all have the same genetic background and are at a similar developmental stage. In addition, we introduced a qualitative evaluation of stomatal opening, as well as treating two chemicals per sample in the first screen to scan through the chemical library as quickly as possible (Fig. 1). As a result, we were able to obtain multiple leading molecules within a short period of time. Notably, in our current screening method, image acquisition and quantification of stomatal aperture from such images in the third screen is the limiting step. We are currently developing a quantification program that enables automation of this process. Using such a program is expected to enhance the screening throughput further.

Suppression of light-induced stomatal opening by SCLs

The SCLs had diverse structures (Fig. 2A), suggesting that these compounds have different targets in guard cells. SCL1 is a hydroquinone analog and may influence the redox reaction in guard cells. Indeed, redox homeostasis in guard cells is important for optimizing stomatal aperture (Murata et al. 2015). The target of SCL1 may be redox sensitive and play an important role in light-induced stomatal opening. In addition, SCL6 is similar to caffeine in structure. Caffeine inhibits blue light-dependent H⁺ pumping in GCPs and light-induced induced stomatal opening in V. faba (Shimazaki et al. 1999). Caffeine probably inhibits these responses by emptying the intracellular Ca²⁺ stores in guard cells, as it does in animal cells (Golovina and Blaustein 1997). Inhibition of blue light-dependent H⁺ pumping and light-induced induced stomatal opening requires millimolar concentrations of caffeine (Shimazaki et al. 1999); however, 50 µM SCL6 markedly suppressed light-induced stomatal opening (Fig. 2B), suggesting that the side chains at C7 and C8 in SCL6 increase cell permeability.

The IC₅₀ values of SCL1–SCL4 for inhibiting light-induced stomatal opening were >10-fold higher than that of ABA (Fig. 2D), although our results suggested that SCLs suppress light-induced stomatal opening in a manner not involving ABA signaling (Fig. 3; Supplementary Fig. S3B). We investigated the effects of SCL1 and SCL2 derivatives on light-induced stomatal opening (Fig. 5), and found that the combination of a hydroxyl group and tertiary-butyl substituent in SCL1, and of pyridazinone and piperidine rings in SCL2, is required for their activities. SCLs thus have potential as lead compounds for the development of high-affinity compounds and affinity probes for target purification.

Molecular mechanisms of the effects of SCL1 and SCL2 in guard cells

SCL1 and SCL2 inhibited blue light- and FC-induced phosphorylation of PM H⁺-ATPase in guard cells, but had no effect on phototropin and KAT1 activities, or ABA-related responses (Figs. 3, 4). Thus, SCL1 and SCL2 probably affect the components of the blue light signaling pathway between downstream of phototropin and PM H⁺-ATPase activation (phosphorylation), including BLUS1, BHP1 and PP1, in stomatal guard cells (Supplementary Fig. S1). In addition, SCL3–SCL9 had no effect on KAT1 activity (Supplementary Fig. S3A), suggesting that these compounds affect the components of the blue light signaling pathway between phototropin and PM H⁺-ATPase activation (phosphorylation) in stomatal guard cells (Supplementary Fig. S1).

Phosphorylation of the threonine at the penultimate position of PM H⁺-ATPase is catalyzed by an unidentified protein kinase, which directly phosphorylates PM H⁺-ATPase not only in guard cells but also in other tissues and cells (Haruta et al. 2015, Falhof et al. 2016, Inoue and Kinoshita 2017). Indeed, protein kinase activity of the threonine residue of PM H⁺-ATPase in vitro was found in the PM of spinach leaf cells, microsomes from guard cell protoplasts of V. faba and the PM of cells in etiolated seedlings of A. thaliana (Svennelid et al. 1999, Hayashi et al. 2010). These results suggest that the unidentified protein kinase is a strong candidate to be the target of SCL1 and SCL2. Dephosphorylation of the phosphorylated threonine of PM H⁺-ATPases may be mediated by the membrane-localized Mg²⁺/Mn²⁺-dependent PP2C-like activity in Vicia guard cells and Arabidopsis etiolated seedlings (Hayashi et al. 2010). D-clade PP2Cs are involved in the dephosphorylation of PM H⁺-ATPase in etiolated seedlings (Spartz et al. 2014, Ren and Gray 2015). Furthermore, the SMALL AUXIN-UP RNAs (SAURs), a large multigene family of early auxin-responsive genes, inhibit D-clade PP2C activity through physical interaction (Spartz et al. 2014, Sun et al. 2016). SAUR19-overexpressing plants display enhanced water loss in detached leaves and exhibit delayed stomatal closure (Spartz et al. 2014, Spartz et al. 2017). Therefore, SCL1 and SCL2 may activate D-clade PP2C or inhibit the interaction between D-clade PP2C and SAURs. Identification of the target(s) of SCL1 and SCL2 in guard cells may reveal the molecular basis of blue light-induced stomatal opening.

Potential advantages of ABA-independent action of SCLs in application

In addition to promotion of stomatal closure, ABA affects various physiological responses of plants, such as acceleration of leaf senescence, down-regulation of plant growth and induction of seed dormancy (Cutler et al. 2010). This pleotropic property is a disadvantage that restricts the use of potential drought tolerance-conferring agrochemicals that strongly depends on activation of ABA signaling, since it greatly affects the productivity and quality of crops. We have previously reported that within the numerous physiological responses that plants use to adapt to drought stress, promotion of stomatal closure was sufficient to confer drought tolerance (Tsuzuki et al. 2013). Hence, we expected that treatment of compounds that promote stomatal closure via an ABA-independent pathway will side step the above-mentioned problems. Although further analysis is awaited to clarify the mode of action of the respective SCLs, our results suggested that they all are involved in an ABA-independent process (Fig. 4; Supplementary Fig. S3B). Together, our results raised the possibility of development of novel drought tolerance-conferring agrochemicals that are differentiated from previously reported ABA agonists (Park et al. 2009, Park et al. 2015).

Indeed, spraying of SCL1 onto rose and oat leaves inhibited wilting (Fig. 6). To develop SCL1 as a drought tolerance-enhancing agrochemical, the duration of its effect and long-term toxicity must be determined. In addition, whether other SCLs have similar effects on plants should be investigated.

Molecular actions of temsirolimus and CP-100356 in guard cells

The mTOR inhibitor, Tem, induced stomatal opening even in the dark. Indeed, two analogs of mTOR inhibitors, Eve and Umi, induced stomatal opening, but Rap and Rid did not (Fig. 8). At present, the reason why only some mTOR inhibitor analogs exert a positive effect is unclear. In addition, the effects of other mTOR inhibitors, such as specific inhibitors of mTOR kinase activity (Guertin and Sabatini 2009), on stomatal aperture should be investigated.

Arabidopsis thaliana possesses an ortholog of mTOR, Arabidopsis target of Rap (AtTOR) (Xiong and Sheen 2014). Therefore, we next investigated the stomatal phenotype in TOR/tor-4 heterozygous plants, as the phenotype of the homozygote is lethal (Menand et al. 2002, Deprost et al. 2007). However, the stomatal aperture under dark and light conditions, and in the presence of 10 µM FC, of TOR/tor-4 heterozygous plants was comparable with that of the wild type (Supplementary Fig. S4), indicating that TOR/tor-4 heterozygous plants show a normal phenotype in stomatal movement as well as in plant growth (Montané and Menand 2013). TOR/tor heterozygous plants reportedly show hypersensitivity to mTOR inhibitors (Montané and Menand 2013). Further investigations of the sensitivity of stomatal movements of TOR/tor-4 heterozygous plants to mTOR inhibitors are thus needed.

CP, which induces stomatal opening in the dark (Fig. 7), is a high-affinity P-glycoprotein/MDR-1 inhibitor (Kajiji et al. 1994). ABC transporters are members of the P-glycoprotein/MDR-1 family (Hwang et al. 2016). ABCB14, ABCC5 and ABCG40 are involved in regulation of stomatal opening, and particularly stomatal closure, in response to various signals (Leonhardt et al. 1999, Nagy et al. 2003, Suh et al. 2007, Lee et al. 2009, Kang et al. 2010). CP may inhibit these ABA transporters, resulting in stomatal opening even in the dark. Further investigations should focus on the effect of CP on ABC transporters in guard cells.

In conclusion, we have identified several compounds that affect stomatal movements. Detailed analyses of these compounds would lead to clarification of signaling pathways for stomatal movements and future development of agrochemicals that control drought tolerance and plant growth. In particular, drought stress severely affects crop yield and quality. Drought stress at the reproductive stage can directly result in an average yield loss of >50% (Hu and Xiong 2014). If such a negative impact can be compensated by agrochemicals that confers drought tolerance like the compounds we identified, it is possible not only to improve the yield of crops but also to cultivate crops in areas which are unsuitable due to water shortage. Furthermore, the amount of carbon dioxide absorption by terrestrial gross photosynthesis is estimated to be 440×10¹⁵ g year^–1 on a global scale (Hetherington and Woodward 2003). Therefore, applying the technology using the compounds found in this research is expected to contribute greatly to the realization of a low-carbon society by increasing carbon dioxide absorption.

Materials and Methods

Plant growth conditions

Commelina benghalensis ssp. plants were grown in soil in a greenhouse at 25 ± 3°C. Vicia faba L. plants were grown hydroponically in a greenhouse at 20 ± 3°C, as described previously (Kinoshita and Shimazaki 1999). Arabidopsis thaliana plants were grown in soil at 22°C under a photoperiod of 16 h white light (50 µmol m^–2 s^–1)/8 h dark. Columbia-0 (Col-0) was used as the background ecotype of the TOR/tor-4 mutant. Seeds of the TOR/tor-4 mutant (SALK_007654) were obtained from the Arabidopsis Biological Resource Center. Oat (Avena sativa) seedlings were grown at 22°C under a photoperiod of 16 h white light (50 µmol m^–2 s^–1)/8 h dark.

Chemical library and screens

We screened 90 compounds from the LOPAC Pfizer Chemical Library (LO5100; Sigma-Aldrich), 20,101 compounds from the ITbM Original Library (purchased from ChemDiv Inc. and Enamine Inc.) and 84 compounds from the Redox Library (Enzo Life Sciences). All compounds were dissolved in dimethylsulfoxide (DMSO) at a concentration of 10 mM. For screening, each compound was added at a 1:200 dilution to basal buffer [5 mM MES/bistrispropane (pH 6.5), 50 mM KCl and 0.1 mM CaCl₂] in a multi-well plate. Prior to the assay, C. benghalensis plants were incubated in the dark overnight to ensure complete closure of stomata. Under dim light, 6 mm diameter leaf discs and 1–2 cm² epidermal peels were excised from fully expanded leaves of 4- to 6-week-old C. benghalensis plants using a hole punch (Biopsy Punch, Kai Medical), and scissors and forceps, respectively. Samples were immersed in basal buffer containing one or two compounds at 50 µM. In the first screening for compounds that inhibit light-induced stomatal opening, or that promote stomatal opening in the dark, the samples were incubated under fluorescent white light (50 µmol m^–2 s^–1) or in the dark at 25°C for 4 h, and samples in which the stomata were uniformly closed or open were identified using a stereoscopic microscope (Stereo Discovery; Zeiss). The compounds that were qualitatively judged to be effective were reassessed in the second screening. The second screening was identical to the first; however, if positive samples in the first screening contained two compounds, the effect of each was independently evaluated. Four biological replicates per compound were prepared, and compounds that were effective in at least three replicates were included in the third screening, in which images of leaf discs and/or epidermal peels were acquired using an optical microscope (BX43; Olympus) with a charge-coupled device (CCD) camera (DP27; Olympus) with a ×10 objective lens (UPlanFL N; Olympus). Upon acquisition, the extended focus imaging function of cellSens standard software (Olympus) was used to maximize the number of analyzable focused stomata within each image. Lastly, stomatal apertures were measured to determine the bioactivity of the compounds. In the third screening, 150 µmol m^–2 s^–1 red light (LED-R; Eyela) and 50 µmol m^–2 s^–1 blue light (Stick-B-32; Eyela) were used instead of white light. We used the re-ordered compounds in the third screening and further experiments. We purchased SCL1-1, SCL1-2 and SCL2-1 from TCI, and SCL2-2, SCL2-3 and SCL2-4 from Enamine.

Viability was assayed using FDA, as described previously (Regan and Moffatt 1990, Kinoshita and Shimazaki 1997).

Measurement of stomatal apertures

The stomatal apertures of dark-adapted C. benghalensis leaf discs or peeled epidermis were measured as described above for the third screening.

The stomatal apertures of A. thaliana were measured as described previously (Inoue et al. 2008) using fully expanded rosette leaves from 5- to 7-week-old plants.

Determination of phosphorylation of phototropin and PM H⁺-ATPase, and detection of the 14-3-3 protein in guard cells

Blue light-induced autophosphorylation of phototropin in GCPs from V. faba was determined by protein blot analysis using the 14-3-3 protein as a probe, as described previously (Kinoshita et al. 2003). Blue light- and FC-induced phosphorylation of PM H⁺-ATPase in guard cells from the epidermis of A. thaliana was determined immunohistochemically, as described previously (Hayashi et al. 2011). The 14-3-3 protein in GCPs from V. faba was detected by Western blotting, as described previously (Kinoshita and Shimazaki 1999).

Measurement of KAT1 activity

Capped complementary RNA encoding KAT1 was synthesized using the mMESSAGE mMACHINE T7 Kit (Ambion) and injected into Xenopus laevis oocytes, as described previously (Uozumi et al. 1995). Two-electrode voltage clamp recordings were performed using an AxoClamp 2B voltage clamp amplifier (Axon Instruments) according to the following protocol: a holding potential of −40 mV, and −20 mV increments from +10 mV to −170 mV for 500 ms. The external solutions contained 120 mM KCl, 1 mM MgCl₂, 1 mM CaCl₂, 10 mM HEPES (pH 7.3, adjusted with NaOH) and 0.2% (w/v) of the test compound(s) in DMSO.

Analyses of ABA-related responses

Seed germination tests were performed using Arabidopsis seeds, as described previously (Tsuzuki et al. 2011).

The expression levels of the ABA-responsive genes RAB18 (At5g66400) and RD29B (At5g52300) in Arabidopsis seedlings were determined by quantitative reverse transcription–PCR (RT–PCR), as described previously (Tomiyama et al. 2014). Relative quantification was performed using the comparative cycle threshold method, and the relative amount of the amplified RAB18 or RD29B product was normalized to that of TUB2, which served as an internal control. In brief, 2-week-old seedlings were incubated in liquid Murashige and Skoog (MS) medium (pH 5.8) containing 50 µM ABA, SCL1 or SCL2 and an equal volume of DMSO at 24°C for 3 h. Total RNA was extracted and first-strand cDNAs were prepared from the compound- or ABA-treated plants. Quantitative RT–PCR was performed using Power SYBR Green PCR Master Mix and a StepOne™ Real-Time PCR System (Applied Biosystems). RAB18, RD29B and TUB2 cDNAs were amplified by PCR using specific primers (Supplementary Table S2).

ABA-induced phosphorylation of the 61 kDa protein in GCPs of V. faba was assessed by Western blot analysis using the 14-3-3 protein as a probe, as described previously (Takahashi et al. 2007). The GCPs were treated with 20 µM ABA, or 50 µM SCL1 or SCL2, at 24°C for 20 min.

Leaf wilting assay

Rose leaves in a bouquet purchased from a local flower shop and 7-day-old oat seedlings were sprayed with 0.5% DMSO or 100 µM SCL1 in 0.02% Silwet L77 (Biomedical Science) and 0.05% Approach BI (Maruwa Biochemical), and allowed to stand at 24°C under 50 µmol m^–2 s^–1 white fluorescent light and 70% relative humidity for 3 h. Then, the leaves were excised and incubated for 6 h (rose) or 20 min (oat) at 24°C under 50 µmol m^–2 s^–1 white fluorescent light and 35–50% relative humidity.

Supplementary Data

Supplementary data are available at PCP online.

Funding

This work was supported by the Japan Science and Technology Agency [Advanced Low Carbon Technology Research and Development Program grant] and the Ministry of Education, Culture, Sports, Science and Technology of Japan [Scientific Research on Priority Areas grant to T.K. (15H059556) and Grants-in-Aid for Scientific Research to N.U. (16H06558 and 16H04906)].

Acknowledgments

We thank Kosuke Ariga, Naoya Kadofusa, Ryota Nakahigashi, Takayuki Nimura, Mao Sasaki and Masayuki Yasutomi of Nagoya University for their technical assistance.

Disclosures

The authors have no conflicts of interest to declare.

Abbreviations

Abbreviations
 
• CP

  CP-100356

 
• DMSO

  dimethylsulfoxide

 
• Eve

  everolimus

 
• FC

  fusicoccin

 
• FDA

  fluorescein diacetate

 
• GCP

  guard cell protoplast

 
• IC₅₀

  half-inhibitory concentration

 
• PM

  plasma membrane

 
• PP2C

  type 2 C protein phosphatase

 
• Rap

  rapamycin

 
• Rid

  ridaforolimus

 
• RT–PCR

  reverse transcription–PCR

 
• SCL

  stomatal closing compound

 
• Tem

  temsirolimus

 
• TOR

  target of rapamycin

 
• Umi

  umirolimus

Author notes

Shigeo Toh and Shinpei Inoue authors contributed equally to this work.