Stomatal morphology and physiology explain varied sensitivity to abscisic acid across vascular plant lineages

Abstract

Abscisic acid (ABA) can induce rapid stomatal closure in seed plants, but the action of this hormone on the stomata of fern and lycophyte species remains equivocal. Here, ABA-induced stomatal closure, signaling components, guard cell K⁺ and Ca²⁺ fluxes, vacuolar and actin cytoskeleton dynamics, and the permeability coefficient of guard cell protoplasts (P_f) were analyzed in species spanning the diversity of vascular land plants including 11 seed plants, 6 ferns, and 1 lycophyte. We found that all 11 seed plants exhibited ABA-induced stomatal closure, but the fern and lycophyte species did not. ABA-induced hydrogen peroxide elevation was observed in all species, but the signaling pathway downstream of nitric oxide production, including ion channel activation, was only observed in seed plants. In the angiosperm faba bean (Vicia faba), ABA application caused large vacuolar compartments to disaggregate, actin filaments to disintegrate into short fragments and P_f to increase. None of these changes was observed in the guard cells of the fern Matteuccia struthiopteris and lycophyte Selaginella moellendorffii treated with ABA, but a hypertonic osmotic solution did induce stomatal closure in fern and the lycophyte. Our results suggest that there is a major difference in the regulation of stomata between the fern and lycophyte plants and the seed plants. Importantly, these findings have uncovered the physiological and biophysical mechanisms that may have been responsible for the evolution of a stomatal response to ABA in the earliest seed plants.

Physiological and biophysical evidence for insensitivity of stomata to abscisic acid in ferns and lycophytes supports stomatal responsiveness to abscisic acid evolved after the divergence of ferns.

Introduction

Stomata are pores on the leaf surface, surrounded by two guard cells, which regulate the exchange of carbon dioxide and water vapor between the leaf interior and exterior (Turner, 1974; Berry et al., 2010; Chater et al., 2011; Brodribb and McAdam, 2011; Henry et al., 2019; Sussmilch et al., 2019). The pores engage in two opposing functions: opening in response to light to facilitate CO₂ uptake from the atmosphere for photosynthesis and to deliver mineral nutrients from soil to plant tissues through the transpiration stream, and closing in response to detrimental environmental conditions, such as water deficits, to limit transpiration and thus prevent excessive loss of water to the atmosphere (Turner, 1975; Fang et al., 2010; Brodribb and McAdam, 2011, 2017; Hannes et al., 2014; Assmann and Jegla, 2016; Henry et al., 2019; Waseem et al., 2021). Functional stomata represent a crucial physiological and ecological adaptation to dry terrestrial environments and have been linked to the evolution of plants on land (Hetherington and Woodward, 2003; Pittermann, 2010; Brodribb and McAdam, 2017; Sussmilch et al., 2019; Brodribb et al., 2020). Therefore, stomatal movement is a major topic of research focus of biologists and considerable research has been conducted on stomatal traits and the molecular and physiological mechanisms associated with stomatal movement (Berry et al., 2010; Brodribb and McAdam, 2011; Ruszala et al., 2011; Zhu, 2016).

Over the last few decades, the focus has been on the stomatal regulation in seed plants, both gymnosperms and angiosperms (Bauer et al., 2012; Brodribb and McAdam, 2013; Abdul-Awal et al., 2016; Henry et al., 2019; Yao et al., 2021a, 2021b). The research has shown that stomatal closure of seed plants is controlled by both active (metabolic) processes and passive (hydraulic) processes (Pei et al., 2000; Brodribb and McAdam, 2011, 2017; McAdam and Brodribb, 2014; McAdam et al., 2016a; Sussmilch et al., 2017a). Active stomatal regulation efficiently closes stomata triggered by complex metabolic processes, especially ion trafficking (Pei et al., 2000; Geiger et al., 2010; Bauer et al., 2012; Abdul-Awal et al., 2016; McAdam et al., 2016b). In contrast, passive closure of stomata is controlled only by leaf water status, in parallel with the decrease in guard cell turgor pressure, as evidenced by exogenous ABA or endogenous ABA synthesized by ferns and lycophytes not inducing stomatal closure (Brodribb and McAdam, 2011; McAdam and Brodribb, 2012; Cardoso et al., 2019), and the linear relationship between leaf water potential and stomatal conductance (Cardoso et al., 2019).

Abscisic acid (ABA) is an important phytohormone that has been widely shown to participate in active stomatal regulation when seed plants suffer various biotic and abiotic stresses (Mittelheuser and Van Steveninck, 1969; Cowan et al., 1982; Schurr et al., 1992; Trejo et al., 1995). It can activate metabolic processes through a series of signaling transduction mechanisms, listed as follows: cytosolic ABA binds to a receptor that consists of Pyrabactin resistance (PYR)/PYR-like/regulatory component of the ABA receptor that inhibits protein phosphatase 2C, that in turn negatively regulates the SNF1-related protein kinases 2 (SnRK2s; Vlad et al., 2009). SnRK2s then activates nicotinamide-adenine dinucleotide phosphate (NADPH) oxidase (AtrbohD/F)-dependent reactive oxygen species (ROS) production of hydrogen peroxide (H₂O₂; Mittler et al., 2004; Raghavendra et al., 2010), ROS triggers nitric oxide (NO) production, and NO upregulates mitogen-activated protein kinase (MAPK) activation and activates Ca²⁺ permeable channels (Pei et al., 2000; Bright et al., 2006). Increased cytosolic Ca²⁺ activates the slow-type anion channel-associated 1 (SLAC1), thereby activating outwardly rectifying K⁺ channels and the efflux of K⁺ (Hosy et al., 2003; Geiger et al., 2010). The loss of osmotically active ions lead to water efflux and turgor loss (Hosy et al., 2003), resulting in stomatal closure. Recent studies showed ABA- and NO-induced stomatal closures require the synthesis and action of cyclic ADP-ribose (cADPR), a signaling molecule downstream of NO, and NO increases cytosolic free Ca²⁺ ([Ca²⁺]_cyt) through a pathway that includes the activation of cADPR cyclase (Neill et al., 2002; Abdul-Awal et al., 2016).

Vacuoles and microfilaments have been also found to be involved in stomatal closure in seed species (Hwang and Lee, 2001; Gao et al., 2005, 2009; Andrés et al., 2014; Isner et al., 2017). For example, as the stomatal close the number of vacuoles increases, and the size of the vacuoles decrease in the guard cells, factors that are crucial to achieve amplitude of stomatal movement, as confirmed by pharmacological and genetic approaches (Gao et al., 2005; Andrés et al., 2014). Studies have also revealed that guard cells show different actin filament configurations in the closed and open state. The long actin filaments of guard cells are severed into many short fragments when the stomata close in the dark, or when treated with ABA and extracellular Ca²⁺ (Hwang and Lee, 2001; Isner et al., 2017). Such reorganization of actin is considered to be a signal transducer for guard cells, and is involved in stomatal movement (Wang et al., 2011; Isner et al., 2017). For example, it has shown that casein kinase 1-like protein 2 and actin depolymerizing factor 4 are involved in reorganization of actin during stomatal closure in response to a water deficit or ABA treatment (Wang et al., 2011; Zhao et al., 2016).

While there has been considerable progress in understanding the biophysical and genetic elements of ABA-induced stomatal closure in seed plants, the role of ABA regulation of stomatal closure in ferns and lycophytes, is less clear and subject to debate (Chater et al., 2011; Ruszala et al., 2011; McAdam and Brodribb, 2012; Cai et al., 2017; Hõrak et al., 2017; Cardoso and McAdam, 2019; Cardoso et al., 2019). There are two opposing models for stomatal closure in ferns and lycophytes. One is a gradualistic model for stomatal evolution (Brodribb and McAdam, 2011, 2017; Sussmilch et al., 2017b; Cardoso and McAdam, 2019; Cardoso et al., 2019), which suggests that active (metabolic) stomatal closure in response to tissue water deficits evolved recently, just after the divergence of ferns, c. 360 million years ago (Mya). However, a contrasting single origin model, based on molecular studies, has challenged the gradualistic model. It suggests a coordinated emergence of stomata and their metabolic control by ABA in the earliest land plants (Chater et al., 2011; Ruszala et al., 2011; Cai et al., 2017; Hõrak et al., 2017). Whether the function of ABA-induced stomatal closure is present in ferns and lycophytes remains unresolved.

The differences in the stomatal responses of ferns and lycophytes have received attention in a number of studies, but little attention has been paid to the fundamental mechanistic differences in the physiological basis of ABA-induced stomatal closure, including ABA signaling, ion flux, and vacuolar and actin cytoskeleton dynamics. In this study, we investigated stomatal closure in response to ABA application in nine angiosperms and two gymnosperms, six ferns and one lycophyte from different lineages. If ABA does not induce stomatal closure in ferns and lycophytes, we hypothesize that the signaling pathway for ABA-induced stomatal closure in seed plants should be absent, and the changes in vacuoles, actin filaments, and aquaporin (AQP) activity of the guard cell membrane, which are involved in stomatal closure in seed plants, will not be induced or observed. In order to test this hypothesis, confocal and Noninvasive Micro-test Technology (NMT) was used to examine the key signaling components, fluorescence labeling, and 3D projection was used to investigate vacuolar and actin filament morphology, and the permeability coefficient of guard cell membranes was measured to evaluate the activity of AQPs in these plants. The results are critical not only to understanding a major difference in the regulation of stomata between orthologous species, ferns, and lycophytes versus the seed plant clade, but also these findings assist in understanding the physiological and biophysical mechanisms that may have been responsible for the evolution of the stomatal response to ABA in the seed plants.

Results

To explore the physiological responses of stomata to ABA in different species, leaf gas exchange was measured after adding high concentrations of ABA to the transpiration stream of a group of 11 seed plants (9 angiosperm and 2 gymnosperm species) as well as 6 ferns and 1 lycophyte. Exogenous ABA induced a rapid reduction in g_s in the seed plants (Figure 1, A–C; Supplemental Figure S1), but did not induce a reduction in g_s in ferns and lycophyte even when a concentrated solution of 15,000 ng·mL⁻¹ ABA was added for 1.5 h (Figure 1, D and E; Supplemental Figure S1). However, leaf excision caused rapid stomatal closure in all lycophyte and fern species (Figure 1, D and E; Supplemental Figure S1). Species, ABA, Ca²⁺, D-sorbitol, and the interaction between species and ABA, Ca²⁺, and D-sorbitol had highly significant effects on the stomatal aperture (Supplemental Table S1). About 50-µM ABA reduced stomatal aperture by 46, 36, and 28% in the seed plant species Arabidopsis thaliana, Sophora japonica, and Ginkgo biloba, respectively, but not in the fern Matteuccia struthiopteris and lycophyte Selaginella moellendorffii (Figure 2, A and B). Stomatal aperture responses to a wide range of concentrations of CaCl₂ (Figure 2C) were consistent with the responses to ABA (Figure 2A), while 800-mM D-sorbitol induced stomatal closure in all species (Figure 2D).

Figure 1.

Stomatal responses to exogenous ABA applied to three lineages of vascular plants. The stomata of the angiosperms A. thaliana (A), and S. japonica (B), and the gymnosperm G. biloba (C) closed rapidly in response to ABA, but the stomata of the fern M. struthiopteris (D) and lycophyte S. moellendorffii (E) did not; only leaf excision (blue arrow) caused stomatal closure in fern M. struthiopteris and lycophyte S. moellendorffii. Vertical long-dash red lines indicate the time of ABA application. Note the change of y-axis among species. Data points show means ± se (n = 3) from three individual plants of each species

Figure 2.

Stomatal apertures of three seed plants A. thaliana, S. japonica, and G. biloba, the fern M. struthiopteris and lycophyte S. moellendorffii in response to ABA, Ca²⁺, and D-sorbitol solution. A, ABA and (C) CaCl₂-induced stomatal closure in the seed plants, but did not in the fern and lycophyte. B, Images of epidermal peels showing ABA-induced stomatal closure in the three seed plants, but not in the fern and lycophyte (bar = 10 µm). D, Hypertonic D-sorbitol solution (800 mM) induced stomatal closure in the three seed plants, and the fern and lycophyte. Forty stomatal pores were measured in each separate replicated experiment in each individual plant (three different plants per treatment). Data are means ± se (total stomatal number per treatment = 120, n = 3). Statistically significant differences between treatments and the control within each species are denoted with asterisks (two-way ANOVA, P < 0.05)

ABA, H₂O₂, sodium nitroprusside (SNP), and cADPR induced stomatal closure of epidermal peels from three seed plants, A. thaliana, S. japonica, and G. biloba. However, ABA-induced closure was inhibited by the inhibitor of NADPH oxidase diphenyleneiodonium chloride (DPI), the H₂O₂-scavenging enzyme catalase (CAT), and NO-scavenging 2-phenyl-4,4,5,5-tetremethylimidazolinone-1-oxyl-3-oxide (cPTIO), and H₂O₂-induced stomatal closure was also inhibited in the presence of CAT and cPTIO (Figure 3, A–C). However, these signaling chemicals did not induce stomatal closure in the fern M. struthiopteris and lycophyte S. moellendorffii (Figure 3, D and E).

Figure 3.

The stomatal aperture of three seed plants A. thaliana, S. japonica, and G. biloba, the fern M. struthiopteris and lycophyte S. moellendorffii in response to a range of signaling chemicals. Peels of A. thaliana (A), S. japonica (B), G. biloba (C), M. struthiopteris (D), and S. moellendorffii (E) were treated with 50-μM ABA, 100-μM H₂O₂, 100-μM SNP, or 75-μM cyclic ADP ribose (cADPR) in opening buffer, in the absence or presence of either 20-μM DPI, 60 U catalase (CAT), or 100-μM cPTIO. Forty stomatal pores were measured in each separate replicated experiment in each individual plant (three different plants per treatment). Data are means ± se (total stomatal number per treatment = 120, n = 3). The different letters indicate significant differences between different treatments within each species (nested ANOVA, P < 0.05)

Using specific fluorescent dye and confocal laser scanning microscopy (CLSM), we visualized the core ABA signaling messengers. Results from two-way analysis of variance (ANOVA) showed that the effect of species, ABA treatment, and the interaction between species and ABA on H₂O₂, NO, and Ca²⁺ fluorescence intensity, the influx of Ca²⁺, and efflux of K⁺ on the surface of guard cells were all highly significantly different (Supplemental Table S2). After ABA treatment, increased H₂O₂-specific fluorescence was observed in the guard cells of the seed plant A. thaliana, the fern M. struthiopteris, the lycophyte S. moellendorffii (Figure 4, A and B), the angiosperm S. japonica, and the gymnosperm G. biloba (Supplemental Figure S2; Figure 4B), demonstrating that ABA induced the production of H₂O₂ in intact guard cells in epidermal strips of these species. In the control and ABA-treated samples, autofluorescence was observed in the inner walls of the guard cells in some seed plants and fern and lycophyte, but NO fluorescence and Ca²⁺ fluorescence were only observed in the cytosol of seed plant guard cells, not in that of the fern or lycophyte guard cells (Figure 4, A, C, and D; Supplemental Figures S3 and S4). NMT assay showed ABA activated the influx of Ca²⁺ and efflux of K⁺ on the surface of guard cells in seed plants, but did not in the fern or lycophyte (Figure 5, A and B; Supplemental Table S3).

Figure 4.

The generation of H₂O₂, NO, and [Ca²⁺]_cyt in guard cells of three seed plants A. thaliana, S. japonica, and G. biloba, the fern M. struthiopteris and lycophyte S. moellendorffii in response to ABA. (A) Confocal images demonstrating fluorescence density of H₂O₂ (H₂DCF-DA), NO (DAF₂-DA), and Ca²⁺ (Fluo-3AM) in intact guard cells of A. thaliana, M. struthiopteris, and S. moellendorffii with and without ABA treatment (bar = 10 µm). (B) H₂O₂ synthesis increased significantly in all species, but (C) NO synthesis and (D) the content of [Ca²⁺]_cyt only increased significantly in the cytosol of seed plant guard cells, not in those of the fern or lycophyte guard cells after treatment with 50-μM ABA. Twenty stomatal pores were measured in each separate replicated experiment in each individual plant (three different plants per treatment). Data are means ± se (total stomatal number per treatment = 60, n = 3). Statistically significant differences between ABA treatment and the control within each species are denoted with asterisks (two-way ANOVA, P < 0.05)

Figure 5.

Effect of ABA on K⁺ and Ca²⁺ fluxes in the guard cells of three seed plants A. thaliana, S. japonica, and G. biloba, the fern M. struthiopteris and lycophyte S. moellendorffii. (A) K⁺ and Ca²⁺ kinetics were recorded before (black circles) and after (gray circles) the addition of 50-μM ABA solution. (B) Significant changes in K⁺ efflux and Ca²⁺ influx were observed on the surface of the guard cells of the three seed plants, but not the guard cells of the fern or lycophyte before (black bars) and after (gray bars) the addition of 50-μM ABA solution. The data represent the means ± se of four different plants of each species (n = 4). Statistically significant differences before and after ABA treatment within each species are denoted with asterisks (two-way ANOVA, P < 0.05)

To assess whether vacuoles and actin are involved in ABA-induced stomatal closure, the angiosperm species, Vicia faba, was used. Results from two-way ANOVA showed that there was a highly significant effect of species, ABA, and the interaction between species and ABA on vacuole and actin traits (Supplemental Tables S4 and S5). In open guard cells, one to two large vacuoles occupied a considerable proportion of the intracellular volume, but after ABA, H₂O₂, and SNP treatment for 30 min, the one or two large vacuoles split into a larger number of small vacuoles in parallel with the reduction in stomatal aperture in V. faba (Figure 6); mean vacuolar size (diameter) decreased by 78%, 59%, and 63%, mean vacuolar number showed a 8.5-, 2-, and 3.2-fold increase with ABA, H₂O₂, and SNP treatment, respectively. However, in the fern M. struthiopteris and the lycophyte S. moellendorffii, no morphological changes were observed in the vacuoles after ABA, H₂O₂, and SNP treatment (Figure 6), but many small vacuoles were observed when stomatal closure was induced by 800-mM D-sorbital solution (Supplemental Figure S5 and Supplemental Table S6). In V. faba, ABA induced reorganization of actin configuration; in the control, actin filaments presented as a reticular distribution, but after ABA treatment for 30 min, they formed a crumb structure as a result of depolymerization, with increased skewness and a decreased density (Figure 7, A–D). In the fern M. struthiopteris and lycophyte S. moellendorffii, ABA did not induce any change in actin configuration (Figure 7, A–D), but actin depolymerized with increased skewness and decreased density as a result of stomatal closure induced by hypertonic D-sorbital treatment (Supplemental Figure S6 and Supplemental Table S7).

Figure 6.

Changes in the vacuoles of guard cells of the seed plant V. faba, the fern M. struthiopteris, and the lycophyte S. moellendorffii in response to ABA, H₂O₂, and SNP. (A) Confocal images demonstrating the changes in vacuoles of the guard cells when treated with opening buffer for 2 h followed by the addition of opening buffer (Control, black bars), ABA, H₂O₂, and SNP (gray bars) for 30 min (bar = 10 µm). (B and C) Vacuoles compartmentalized to form a number of small vacuoles in the seed plant, but not in the fern and lycophyte. (D) Treatment with ABA, H₂O₂, and SNP reduced the stomatal aperture of V. faba, but not the stomatal aperture of the fern and lycophyte. Only the central optical sections of guard cells were used to calculate the average number and diameter of vacuoles. A series of Z-axis optical sections were obtained by 3D projection. Thirty stomatal pores were measured in each separate replicated experiment in each individual plant (three different plants per treatment). Data are means ± se (total stomatal number per treatment = 90, n = 3). The asterisks indicate significant differences between treatments and control within each species (two-way ANOVA, P < 0.05)

Figure 7.

Changes of actin filaments in guard cells of the seed plant V. faba, the fern M. struthiopteris, and lycophyte S. moellendorffii. (A) Confocal images showing the effects of ABA on the changes in actin filaments of guard cells (bar = 10 µm). (B) Actin filament density (occupancy), (C) actin filament bundling (skewness), and (D) stomatal aperture of guard cells of the seed plant, fern and lycophyte. Actin filament images were taken and stomatal apertures were measured after treatment with ABA (gray bars), or only opening buffer (control, black bars). Twenty stomatal pores were measured in each separate replicated experiment in each individual plant (three different plants per treatment). Data are means ± se (total stomatal number per treatment = 60, n = 3). Statistically significant differences between ABA treatment and control within each species are denoted with asterisks (two-way ANOVA, P < 0.05)

Results from two-way ANOVA showed that there were significant effects of species, ABA, and the interaction between species and ABA on osmotic water permeability (Supplemental Table S8). ABA treatment induced an increases in both the osmotic water permeability for water outflow [P_f (out)] and for water inflow [P_f (in)] in V. faba, but had no effect on P_f (in) and P_f (out) in M. struthiopteris and S. moellendorffii, respectively (Figure 8; Supplemental Figure S7). When exposed to 50-μM ABA for 60 min, P_f (out) was 1.2-fold higher than P_f (in) in V. faba (t test, P = 0.040), while P_f (out) was similar to P_f (in) in M. struthiopteris (t test, P = 0.558) and S. moellendorffii (t test, P = 0.691; Figure 8).

Figure 8.

Effects of ABA on water permeability of guard cell protoplasts of the seed plant V. faba, the fern M. struthiopteris, and lycophyte S. moellendorffii. (A) Mean osmotic water permeability for water inflow [P_f (in)] and (B) mean osmotic water permeability for water outflow [P_f (out)] of guard cell protoplasts treated with the control solution (black bars) or 50-μM ABA (gray bars) in V. faba, M. struthiopteris, and S. moellendorffii. Around 14–19 guard cell protoplasts were measured in each separate replicated experiment in each individual plant (three different plants per treatment). Data are means ± se (total stomatal number per treatment = 42–57, n = 3). Statistically significant differences between ABA treatment and control within each species are denoted with asterisks (two-way ANOVA, P < 0.05)

Discussion

Insensitivity of fern stomata to ABA

In vascular plants, one of the primary functions of stomata is to control water loss and maintain plant water balance (Turner, 1974; Hetherington and Woodward, 2003; Berry et al., 2010; Pittermann, 2010; Chater et al., 2011; Brodribb and McAdam, 2011; Henry et al., 2019; Cardoso et al., 2019). In seed plants, it is generally recognized that stomatal closure to reduce transpiration occurs primarily by active metabolic regulation, mainly as a result of the activity of the phytohormone ABA, rather than by passive responses of stomata to leaf hydration (Brodribb and McAdam, 2011, 2013; Chater et al., 2011; Ruszala et al., 2011; Sussmilch et al., 2017a, 2017b). However, there are two opposing opinions about stomatal closure in ferns and lycophytes. A gradualistic model suggests that active (metabolic) stomatal closure in response to tissue water deficit evolved recently, just after the divergence of spermatophytes (seed plants) from ferns, c. 360 Mya (Brodribb and McAdam, 2011, 2017; McAdam and Brodribb, 2012; McAdam et al., 2016a; Sussmilch et al., 2017a, 2017b; Voss et al., 2018; Cardoso et al., 2019). Three lines of evidence support this view: (1) step transitions in humidity caused the stomata of ferns and lycophytes to respond according to exponential decay or rise functions very close to the modeled kinetics under passive hydraulic control (Brodribb and McAdam, 2011); (2) the addition of ABA to the leaf transpiration stream led to rapid stomatal closure of seed plants, but failed to close the stomata of fern and lycophyte species (Brodribb and McAdam, 2011); foliar ABA levels increased under drought in three seedless vascular plants, ferns (Pteridium esculentum, Dicksonia antarctica, and Nephrolepis exaltata) and a lycophyte (Selaginella kraussiana), but stomata did not respond to the increasing ABA (McAdam and Brodribb, 2012; Cardoso and McAdam, 2019; Cardoso et al., 2019), suggesting that the rate of ABA synthesis under water stress is not the problem, but rather the insensitivity of the stomata of ferns and lycophytes to ABA; (3) when the leaves with high endogenous levels of ABA were quickly rehydrated from a water deficit, the stomata of the ferns and lycophyte rapidly reopened to maximum aperture, while the stomata of seed plants reopened slowly (McAdam and Brodribb, 2012, 2013, 2014).

However, a contrasting single-origin model, based on physiological and molecular studies, has challenged the gradualistic model. It suggests a coordinated emergence of stomata and their metabolic control by ABA in the earliest land plants (Chater et al., 2011; Ruszala et al., 2011; Cai et al., 2017; Hõrak et al., 2017). For example, the stomata on the sporophytes of the moss Physcomitrium patens respond to environmental signals in a similar way to those of flowering plants and a homolog of ABA is involved in stomatal control in mosses (Chater et al., 2011). Stomatal responses in the lycophyte S. moellendorffii to ABA, and the underlying intracellular signaling pathways responsible for the control of stomatal aperture, are similar to those in modern vascular plant lineages, suggesting that physiologically active stomatal control originated at least as far back as the emergence of the lycophytes (c.420 Mya; Ruszala et al., 2011). Stomatal aperture assay showed N. exaltata, which exhibited a 25% reduction in a stomatal aperture in the isolated epidermis in response to exogenous ABA (Cai et al., 2017), comparative transcriptomic analysis has shown that Polystichum proliferum retain ABA-responsive genes that encode proteins associated with ABA biosynthesis, transport, reception, transcription, signaling, and ion and sugar transport that fit the general ABA signaling pathway constructed from Arabidopsis (Cai et al., 2017). Furthermore, the stomatal apertures of Athyrium filix-femina and Dryopteris filix-mas were reduced in response to spraying with ABA in the growth cabinet at 70% relative humidity (Hõrak et al., 2017). However, recent studies have shown that the stomata of A. filix-femina and N. exaltata did not close in response to endogenous ABA accumulation during drought (Cardoso and McAdam, 2019; Cardoso et al., 2019).

In this study, to determine the divergence in the stomatal responses of angiosperms, gymnosperms, lycophytes, and ferns to ABA, we evaluated the stomatal conductance (g_s) of 11 representative seed plants, 6 fern species, and 1 lycophyte when exogenous ABA was introduced into the transpiration stream. The results showed that the stomata of the lycophyte and ferns were insensitive to ABA, and that the stomata only closed in response to hypertonic solution and the rapid dehydration produced by leaf excision (Figures 1, D, E and 2, D; Supplemental Figure S1). These data suggest that the stomata of the ferns and lycophyte in this study mainly responded to the leaf water status, more precisely, passive stomatal control (Figure 1, D and E; Supplemental Figure S1). The results are strongly consistent with a gradualistic model that the stomatal response to ABA is a trait exclusive to seed plants and not present in ferns and lycophytes (Brodribb and McAdam, 2011; McAdam and Brodribb, 2012; Sussmilch et al., 2017a, 2017b; Brodribb et al., 2020). Furthermore, the most important finding that we uncovered is the difference in physiological and biophysical infrastructure in the ABA signaling pathway in seed compared with ferns and lycophytes, as discussed below.

The physiological differences accounting for stomatal responses to ABA

Open stomata1 (OST1)/SLAC1 are core components in the ABA-stomatal closure pathway in seed plants (Pei et al., 2000; Hosy et al., 2003; Bright et al., 2006; Abdul-Awal et al., 2016; Lv et al., 2017), but except in the model plant A. thaliana, it is difficult to get a mutant line to analyze the function of OST1/SLAC1 in nonmodel species, such as the S. japonica, G. biloba, M. struthiopteris, and S. moellendorffii used in this study. Therefore, we tested and compared the signaling components, i.e. H₂O₂, NO, Ca²⁺, and K⁺, involved in the pathway of signal transduction of ABA-induced stomatal closure between seed and ferns and lycophytes. Furthermore, it has been shown that the ABA–SnRK2 signaling pathway is engaged in regulating spore dormancy and sex determination, not for stomatal closure, in the fern Ceratopteris richardii and lycophyte S. moellendorffii (McAdam et al., 2016a; Figure 9). Thus, we focused on signaling components downstream of SnRK2, including the function of H₂O₂, NO, Ca²⁺, and K⁺ in seed, fern and lycophyte plants. ABA-, H₂O₂-, and NO-induced stomatal closure in epidermal strips of seed plants, and ABA-induced closure was inhibited by the H₂O₂-scavenging enzyme CAT, NO-scavenging cPTIO, and the inhibitor of NADPH oxidase DPI, indicating that ROS and NO were required for ABA-induced stomatal closure. Furthermore, pretreatment with cPTIO greatly diminished the stomatal response to H₂O₂, indicating that NO synthesis is required for H₂O₂-induced stomatal closure, and ROS is an upstream component of the NO-signaling network involved in stomatal closure in response to ABA (Figure 3), as observed previously (Bright et al., 2006; Lv et al., 2017). NO synthesis subsequently activates other components, such as the Ca²⁺ channel and K⁺ channel in the ABA-stomatal closure pathway in seed plants (A. thaliana, S. japonica, G. biloba). However, stomata did not close in response to ABA, H₂O₂, SNP, Ca²⁺, or cADPR treatment in fern and lycophyte plants. Therefore, we tried to identify why the ABA signaling pathway was not effective in ferns and lycophytes.

Figure 9.

ABA functional pathway showing the proposed divergence in signaling of stomatal closure in different plant lineages. The left branch denotes the ABA pathway in fern and lycophyte species, and the right branch denotes the ABA pathway in seed plants. The left pathway shows that ABA is involved in spore germination and sex determination in ferns and lycophytes (McAdam et al., 2016a), while ABA regulation of stomatal closure is absent owing to the lack of NO synthesis and the ABA-signaling pathway downstream of NO. The right pathway indicates that signal transduction for ABA-induced stomatal closure has evolved in seed plants (Park et al., 2009; Raghavendra et al., 2010; Zhu, 2016).

After ABA treatment, 2′,7′-dichlorodihydrofluorescein diacetate (H₂DCF-DA; H₂O₂) fluorescence density increased by 2.8-fold and 2.9-fold in the cytoplasm of guard cells of M. struthiopteris and S. moellendorffii, in addition to large increases in the three seed plants, A. thaliana, S. japonica, and G. biloba (Figure 4, A and B;  Supplemental Figure S2), indicating that the ABA receptor functions normally and ABA activates a plasma membrane-localized NADPH oxidase that generates H₂O₂ in both seed plants and fern and lycophyte. However, no increase in DAF-2DA (NO) fluorescence was observed in guard cells of M. struthiopteris and S. moellendorffii that contrasted with significant increases in the three seed plants (Figure 4, A and C; Supplemental Figure S3), indicating that H₂O₂ fail to induce NO production in the fern and lycophyte. Furthermore, SNP, the NO supplier inducing stomatal closure in seed plants, did not induce stomatal closure in the fern or lycophyte (Figure 3), suggesting ABA-induced stomatal closure via NO synthesis and the pathway downstream of NO is conserved in seed plants, but is lacking in the fern and lycophyte. cADPR has been shown to be involved in ABA-induced stomatal closure (Neill et al., 2002; Abdul-Awal et al., 2016), and nicotinamide, an antagonist of cADPR production inhibited the effects of both ABA and NO in inducing stomatal closure (Leckie et al., 1998), suggesting that inhibition of ABA responses by nicotinamide is, at least partly, due to the inhibition of cADPR synthesis following NO generation. In this study, cADPR induced a significant decrease in stomatal aperture of seed plants, but not in the fern and lycophyte (Figure 3). In addition, ABA-induced cytosolic Ca²⁺ increased in seed plants, but not in the fern and lycophyte. NMT assays indicated that ABA induced K⁺ efflux and Ca²⁺ influx, but did not trigger the activation of either channel in the fern and lycophyte, as shown by no change in K⁺ efflux and Ca²⁺ influx in M. struthiopteris and S. moellendorffii (Figure 5). In brief, all the above results suggest that the breakdown in ABA-mediated stomatal closure and ABA-signaling transduction of stomatal regulation in ferns occurs as a result of the lack of NO synthesis and the pathway downstream of NO generation. Therefore, we boldly speculate that the reason why fern and lycophyte stomata are insensitive to ABA is because of the lack of NO synthesis and its effect on the ABA-signaling pathway downstream of NO (Figure 9). The ABA-induced increase in H₂O₂ may play other functions, such as degradation of chlorophyll in response to drought or cold stress. Further work is necessary using molecular protein approaches to fully understand the roles and functions of ABA in ferns and lycophytes.

The biophysical differences accounting for stomatal responses to ABA

With stomatal movement induced by ABA and extracellular Ca²⁺ in the light, the vacuoles and actin filaments are in a state of dynamic change (Gao et al., 2005, 2009; Andrés et al., 2014; Zhao et al., 2016; Eisenach and de Angeli, 2017). In A. thaliana, the closure of stomata always occurs in parallel with the splitting of large vacuoles into smaller vacuoles and the appearance of numerous intra-vacuolar membrane structures in the guard cells (Andrés et al., 2014). Similar results were also observed in the closed stomata of in the guard cells of V. faba that had a large number of small vacuoles and various vacuolar membrane structures (Gao et al., 2005, 2009). The vacuole fusion inhibitor, (2s,3s)-trans-epoxysuccinyl-L-leucylamido-3-methylbutane ethyl ester, significantly inhibits stomatal opening. Furthermore, these small vacuoles (or vesicles) can fuse with the plasma membrane by increasing its surface area as guard cell volumes increase (Gao et al., 2005). All these results suggest that dynamic changes of the tonoplast of the guard cells are essential for enhancing stomatal movement. Actually, guard cells can adjust their volume by up to 40% within minutes (Woods and Turner, 1971), with concomitant reshuffling of their vacuolar apparatus (Jezek and Blatt, 2017).

Actin filaments (AFs) are a major component of the cytoskeleton and play important roles in organelle motility and/or positioning, vesicle trafficking to the vacuole, and vacuolar morphogenesis (Gao et al., 2009). Indeed, microfilament bundles encompass vacuoles and connect adjacent vacuoles (Wang et al., 2011). It has been shown that OST have a parallel radial array or hoop- and ring-like microfilaments, while closed stomata have randomly oriented filaments (Zhao et al., 2016). The microfilaments have been reported to co-localize with the vacuolar membrane in BY-2 cells (Toyooka and Matsuoka, 2006), and the trichome cells of Arabidopsis arp2/3 mutants (wrm1-1 and dis1-1, respectively) exhibit abnormal AF aggregation and many unfused miniature vacuoles near the central vacuole, suggesting that vacuolar membrane fusion is impaired by the defective AF organization (Mathur, 2003). Experiments with the guard cells of the arp2 and arp3 mutants demonstrate that the reorganization of AFs in guard cells regulates vacuolar fusion and function during light-induced stomatal opening. These short AFs might be involved in vacuolar dynamics during stomatal movement by activating the tonoplast ion channels, as it has been shown that AF dynamics may play a role in the regulation of K⁺ and Ca²⁺ channel activities in guard cells (Hwang and Lee, 2001; Isner et al., 2017). In the present study, similar results were observed with the depolymerization of AFs into either an actin patch or crumb structure in V. faba, the compartmentalization of big vacuoles to form a number of small vacuoles, and K⁺ efflux and Ca²⁺ influx in ABA-induced stomatal closure. However, in the fern M. struthiopteris and lycophyte S. moellendorffii, ABA did not induce any changes in vacuolar volume, AF depolymerization, or ion activation in the guard cells.

In A. thaliana, ABA induced a two-fold increase in osmotic water permeability (P_f) of guard cell protoplasts and an accumulation of ROS in guard cells, both of which were abrogated in plasma membrane intrinsic protein 2;1 plants with knockout of the plasma membrane intrinsic protein 2;1 AQP without affecting stomatal closure (Grondin et al., 2015). This suggests AQP involvement in ABA-induced stomatal closure by facilitating guard cells transmembrane water flow in A. thaliana guard cells (Hwang and Lee, 2001), and in facilitating hydrogen peroxide entry into guard cells (Rodrigues et al., 2017). In the present study, P_f (out) increased significantly more than P_f (in) in V. faba guard cells after ABA treatment, indicating that P_f (out) was a measure of the efflux of guard cell water and high AQP activity induced by ABA that facilitates stomatal closure. However, ABA has no effect on the AQPs of the fern M. struthiopteris and lycophyte S. moellendorffii that in an evolutionary sense suggests that the guard cell AQPs of ferns and lycophytes have not evolved functionally to respond to ABA.

Conclusion

In our study, we found that the stomata of ferns and lycophytes are ABA insensitive. This may be due to a difference in substantial genetic infrastructure that we observed across the ABA-signaling pathway: seed plants have ABA-induced H₂O₂ elevation, NO production, K⁺ efflux and Ca²⁺ influx, vacuolar compartmentation, actin filament disintegration, and an increased permeability coefficient of guard cell protoplasts that are not present in ferns and lycophytes. Our results provide strong support for a gradualistic model for the evolution of ABA-mediated stomatal responses, with functional evolution of stomata from passively controlled valves in early land plants, to actively regulated pores with complex ABA-mediated control mechanisms in modern seed plants.

Materials and methods

Plant materials and growth conditions

A functionally- and phylogenetically diverse range of vascular plant species were selected, including nine angiosperms A. thaliana, Spathiphyllum kochii, Caragana boisi, Hordeum vulgare, S. japonica, Populus euphratica, Pharhiris nil, Helianthus annuus, and Caragana intermedia, two gymnosperms Metasequoia glyptostroboides and G. biloba, six ferns Eremochloa ciliaris, A. filix-femina, M. struthiopteris, Polystichum neolobatum, Coniogramme japonica, and N. exaltata, and one lycophyte S. moellendorffii, to test their responses to exogenous ABA. S. japonica and G. biloba were sampled from the gardens of the campus of Lanzhou University, Lanzhou, Gansu Province, China, while the other plants were grown in a growth chamber at Lanzhou University with 16-h/8-h day/night, 22 ± 1°C, 70% relative humidity, and 800-μmol m⁻² s⁻¹ photosynthetic photon flux density (PPFD) in the light period for seed plants and 400 μmol m⁻² s⁻¹ for ferns and the lycophyte. The plants in the growth cabinet were irrigated weekly with half strength Hoagland’s solution.

By introducing exogenous ABA to the transpiration stream, the responses of gas exchange to ABA was examined across all 18 plants. Then, one herb (A. thaliana, model plant), one tree species (S. japonica), one gymnosperm (G. biloba), one fern (M. struthiopteris), and one lycophyte (S. moellendorffii), all with an epidermis that is easy to isolate from the leaf, were chosen to measure stomatal aperture, H₂O₂, NO, and Ca²⁺ fluorescence, and Ca²⁺ and K⁺ flux in the guard cells of isolated epidermis using confocal microscopy and an NMT assay that measures the flux of ions in living cells. When morphological vacuoles and actin filaments were observed by confocal microscopy, V. faba was substituted for A. thaliana and S. japonica, as it is recognized as ideal material in which to observe these parameters (Gao et al., 2005). Then, stomatal aperture and morphological vacuoles and actin filaments of guard cells were studied in one angiosperm, V. faba, one fern, M. struthiopteris, and one lycophyte, S. moellendorffii. All sampled leaves were fully expanded and in good condition.

Stomatal conductance measurement

The stomatal conductance (g_s) was measured following a previously described method by Brodribb and McAdam (2011). For responses to ABA, three leafy shoots/stems were severed from three different individuals of each species (18 species, 3 replicates each), recut under distilled water, and the shoot/stem base immersed in a vial of distilled water. The g_s was measured using an infrared gas analyzer (Li-6400, Li-Cor, Lincoln, NE, USA) with a leaf cuvette temperature of 22°C, vapor pressure deficit of about 1.2 kPa, a CO₂ concentration of 400 µmol mol⁻¹, and PPFD of 400–1000 µmol quanta m⁻² s⁻¹. After 10 min for equilibration in the chamber, the stomatal dynamics of the leaf were measured for at least 50 min to ensure that a stable g_s had been reached, after which the distilled water was replaced by 50-µM ABA so that the ABA entered into the transpiration stream of the excised leaf, and the g_s was measured until it was stable. In lycophyte and fern species, the rachis or stem was then cut in air, severing the leaf water supply and g_s was measured. After measurement, sample was removed from the cuvette and the leaf was scanned by scanner (Epson Perfection V800 Photo Scanner, Seiko Epson Corporation, Nagano, Japan) and leaf area was measured using ImageJ software (https://imagej.nih.gov/ij/) so that the g_s data could be adjusted for the leaf area in the cuvette.

Stomatal aperture assay

Stomatal bioassays were measured following a previously described method by Cornelia et al. (2012). The abaxial (lower) epidermis was peeled from the leaves from three individuals of each species (as three replicates) with forceps, floated in glass Petri dishes containing morpholino ethane sulfonic (MES)-KCl buffer [10-mM 2-MES acid, 50-mM KCl, 20-μM CaCl₂, pH 6.15, hereafter called opening buffer] for 2 h in the light (a PPFD of 100 μmol m⁻² s⁻¹) in a growth chamber before the addition of the various compounds. To test the effect of the dose of ABA and CaCl₂ on stomatal aperture, peels were incubated in the presence of various concentrations of ABA and CaCl₂ for 2 h, and to test the effect of osmotic solutes on stomatal aperture, peels were incubated in a solution of 800-mM D-sorbitol for 30 min. In order to investigate the effect of different compounds on ABA-induced stomatal closure, H₂O₂, SNP, cADPR, the NO scavenger cPTIO, the H₂O₂ scavenger catalase (CAT), and NADPH oxidase inhibitor DPI were added one-by-one to the opening buffer, the epidermal peels incubated in the presence of these inhibitory compounds for 30 min prior to treatment with ABA, SNP, or H₂O₂. Images of the stomatal aperture were taken using a light microscope (Ex30LED; Sunny Instruments, Ningbo, China) and the aperture measured using ImageJ software (https://imagej.nih.gov/ij/). Forty stomata were measured for each independent replicate, and mean values and standard errors of stomatal aperture were calculated on the basis of three independent replicates per treatment per species.

ROS, NO, and Ca²⁺ detection in guard cells

ROS production in the guard cells was determined using the fluorescent indicator H₂DCF-DA (Sigma, USA), and NO-specific fluorescence was visualized using the specific NO dye 4,5-diaminofluorescein diacetate (DAF₂-DA, Sigma, USA) following the method described by Bright et al. (2006). To measure and detect [Ca²⁺]_cyt dynamics in guard cells, the Ca²⁺-specific fluorescence probe Fluo 3-acetomethoxyester (Fluo 3-AM) dissolved with dimethyl sulfoxide was used, as described by Pei et al. (2000). Epidermal peels from three individuals of each species (as three replicates) were pretreated with the opening buffer for 2 h before adding 50-µM H₂DCF-DA, 15-μM DAF₂-DA, and 10-µM Fluo 3-AM for 20 min in the dark, followed by a 5-min flush in the opening buffer to remove excess dye. The strips were subsequently incubated in the opening buffer containing 50-μM ABA for 60 min. Confocal microscopy observations and images were performed on a CLSM (LSM 880 CLSM; Zeiss, Germany) with 40× objective for observation using 1% maximum laser power at 488 nm with 500 gains. The fluorescence signal was collected between 495 and 550 nm with a pinhole yielding a 4-µm cross section. All images were acquired using identical offset, gain, and pinhole settings, and analyzed using ImageJ software (https://imagej.nih.gov/ij/). Twenty guard cells were measured for each independent replicate, and mean values and standard errors of H₂O₂ fluorescence, NO fluorescence, and Ca²⁺ fluorescence were calculated on the basis of three independent replicates per treatment per species.

Net K⁺ and Ca²⁺ flux analysis in guard cells

The net K⁺ and Ca²⁺ flux was measured at Xu-Yue Research Institute (Beijing, China), using NMT (NMT100 series; Younger USA LLC, Amherst, MA, USA) and imFluxes V2.0 software (Younger USA LLC, Amherst, MA, USA) as described previously (Tang et al., 2012). The abaxial epidermis was peeled from the youngest fully expanded leaf from four individual plants of each species (as four replicates) and fixed to the bottom of a Petri dish by double-sided tape and incubated in opening buffer for 2 h in the light, equilibrated in measuring buffer (0.1-mM KCl, 0.1-mM CaCl₂, 0.1-mM MgCl₂, 0.5-mM NaCl, 0.2-mM Na₂SO₄, 0.3-mM MES, pH 6.15) for 10 min. Net K⁺ fluxes and Ca²⁺ fluxes were measured for 10 min in measuring buffer as the control, then peels were incubated for a further 10 min in the measuring buffer containing 50-μM ABA, and net K⁺ fluxes and Ca²⁺ fluxes were measured.

Vacuole fluorescent staining in guard cells

Vacuole fluorescent staining was carried out with the method described by Gao et al. (2009). Briefly, peels from the abaxial epidermis of three individuals of each species (as three replicates) were pretreated with the opening buffer for 2 h, transferred into opening buffer plus 10-μM Acridine Orange (Sigma, USA) for 10–15 min at room temperature in the dark, then washed three times with opening buffer before incubation in the presence of 50-μM ABA, 100-μM H₂O₂, and 100-μM SNP for 30 min for V. faba, M. struthiopteris, and S. moellendorffii, or incubation in the presence of 800-mM D-sorbitol solution for 30 min for M. struthiopteris and S. moellendorffii. Confocal microscopy observations and images were performed on a confocal laser scanning microscope (LSM 880 CLSM; Zeiss, Germany) with a 40× objective for observation. With an excitation laser of 1% maximum laser power at 488 nm with 600 gains, a series of optical sections were taken with a 505–550 nm bandpass filter. 3D projections were obtained from a series of 0.4- 0.6-µm interval continuous optical sections. Thirty guard cells were measured for each independent replicate, and mean values and standard errors of vacuole diameter, vacuole number, and stomatal aperture were calculated on the basis of three independent replicates per treatment per species.

Labeling of actin filaments in guard cells

Actin filaments in the guard cells were labeled following the method described by Nishimura et al. (2003). Peels from the abaxial epidermis of leaves from three individuals of each species (as three replicates) were pretreated with the opening buffer for 2 h to open the stomata, then transferred to the opening buffer containing 50-µM ABA for 30 min for V. faba, M. struthiopteris, and S. moellendorffii, or the opening buffer containing 800-mM D-sorbitol solution for 30 min for M. struthiopteris and S. moellendorffii. Then peels were prefixed for 20 min with 1% (v/v) paraformaldehyde and 0.025% (v/v) glutaraldehyde in PME buffer [100-mM Piperazine-1,4-bisethanesulfonic acid (PIPES), 5-mM MgSO₄, and 10-mM ethylene glycol-bis (2-aminoethylether)-tetraacetic acid (EGTA) at pH 6.8], immersed in 2% (v/v) paraformaldehyde and 0.05% (v/v)glutaraldehyde in PME for 20 min, then fixed in a final concentration of 4% (v/v) paraformaldehyde and 0.1% (v/v) glutaraldehyde in PME for 20 min. After washing three times in PME, epidermal strips were pretreated with 3% (w/v) Driselase (Sigma, USA) for 30 min before being stained overnight in darkness, using 0.3-µM Alexa 488-phallodin diluted in PME buffer with 5% (v/v) dimethyl sulfoxide and 0.05% (v/v) Nonidet P-40. Confocal microscopy observations and images were performed on a confocal LSM (880 CLSM; Zeiss, Germany) with a 40× objective for observation. Alexa 488-phallodin fluorescence was imaged using 1.5% maximum laser power at 488 nm with 800 gains and a 505–530 nm bandpass emission filter. 3D projections were obtained from a series of 0.4- to 0.6-µm interval continuous optical sections. The skewness and density of actin filaments were quantified by ImageJ software (https://imagej.nih.gov/ij/) according to Higaki et al. (2010). Twenty guard cells were measured for each independent replicate, and mean values and standard errors of filament occupancy, filament skewness, and stomatal aperture were calculated on the basis of three independent replicates per treatment per species.

Osmotic water permeability coefficient (P_f) measurements

Guard cell protoplast (GCP) isolation was carried out as described by Leonhardt et al. (2004). Healthy and fully expanded leaves from three individuals of each species (V. faba, M. struthiopteris, and S. moellendorffii; as three replicates) from well-watered plants were excised before the onset of the light period in the growth chamber and then immediately immersed in cold tap water. The abaxial epidermis was peeled from the leaves, cut into squares of about 5 mm × 5 mm. The GCP of V. faba was isolated using digesting solution of 0.8% (w/v) cellulose RS (Yakult Honsha Co., Tokyo, Japan), 0.08% (w/v) Pectolyase Y-23 (Seishin Pharmaceutical Co., Tokyo, Japan), 0.1% (w/v) Macerozyme R-10 (Yakult Honsha Co., Tokyo, Japan), 0.5% (w/v) polyvinyl pyrrolidone (Sigma, USA), 1-mM CaCl₂, 10-mM Mes/Tris, pH 6.1, and 550-mM D-sorbitol for 2–4 h. For M. struthiopteris and S. moellendorffii, the epidermal strips were pretreated with 3% (w/v) Driselase (Sigma, USA) for 30 min, then the GCP of M. struthiopteris were isolated using digesting solution of 2.4% (w/v) cellulose RS (Yakult Honsha Co., Tokyo, Japan), 0.24% (w/v) Pectolyase Y-23 (Seishin Pharmaceutical Co., Tokyo, Japan), 0.3% (w/v) Macerozyme R-10 (Yakult Honsha Co., Tokyo, Japan), 0.5% (w/v) polyvinyl pyrrolidone (Sigma, USA), 1-mM CaCl₂, 10-mM Mes/Tris, pH 6.1, and 550 mM D-sorbitol for 3 h, and GCP of S. moellendorffii were isolated using digesting solution of 4.0% (w/v) cellulose RS (Yakult Honsha Co., Tokyo, Japan), 0.4% (w/v) Pectolyase Y-23 (Seishin Pharmaceutical Co., Tokyo, Japan), 0.5% (w/v) Macerozyme R-10 (Yakult Honsha Co., Tokyo, Japan), 0.5% (w/v) polyvinyl pyrrolidone (Sigma, USA), 1-mM CaCl₂, 10-mM Mes/Tris, pH 6.1, and 550-mM D-sorbitol solution for 5 h. Isolated protoplasts of all three species were re-suspended in 4 mL of 550-mM D-sorbitol solution (isotonic solution) and kept in the dark before treatment with 50-μM ABA for 60 min. Swelling measurements were performed at 20°C. P_f (in) of guard cell protoplasts was determined from the initial rate of volume increase by transfer of individual protoplasts from a 550-mM D-sorbitol isotonic solution to a 300-mM D-sorbitol solution (hypotonic solution) under a microscope. P_f (out) were obtained when transferred from a 550-mM D-sorbitol isotonic solution to a 950-mM D-sorbitol (hypertonic solution). Around 14–19 guard cell protoplasts were measured for each independent replicate, and mean values and standard errors of P_f (in) and P_f (out) were calculated on the basis of three independent replicates per treatment per species.

Statistical analysis

We used two-way ANOVA to test differences in stomatal morphology and physiology traits using the SPSS 15.0 program (SPSS software, SPSS Inc.). Within our statistical models we treated species and treatments (ABA, Ca²⁺, and D-sorbitol) as fixed factors, and post hoc pairwise mean comparisons on all significant effects of treatments within species were carried out using Fisher’s least significant difference test with SPSS syntax editor. We also used nested ANOVA to test the difference in stomatal aperture induced by compounds on ABA-induced stomatal closure within species using lmer function in R 3.6.2 (https://www.R-project.org/), and used t test to test the difference between P_f (in) and P_f (out) under ABA treatment within each species using the SPSS 15.0 program (SPSS software, SPSS Inc). Different letters or asterisks above columns in the figures indicate statistical differences (P < 0.05) across treatments or between the treatment and control groups.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. Stomatal responses to exogenous ABA applied to three lineages of vascular plants.

Supplemental Figure S2. ABA stimulated generation of H₂O₂ in guard cells of the angiosperm S. japonica and gymnosperm G. biloba.

Supplemental Figure S3. ABA stimulated NO synthesis in guard cells of the angiosperm S. japonica and gymnosperm G. biloba.

Supplemental Figure S4. ABA stimulated an increase of the content of [Ca²⁺]_cyt in guard cells of the angiosperm S. japonica and gymnosperm G. biloba.

Supplemental Figure S5. Changes in the vacuoles in guard cells of the fern M. struthiopteris and lycophyte S. moellendorffii in response to D-sorbitol treatment.

Supplemental Figure S6. Changes in actin filaments in guard cells of the fern M. struthiopteris and lycophyte S. moellendorffii in response to D-sorbitol treatment.

Supplemental Figure S7. Microscope images of guard cell protoplasts exposed to hypertonic, isotonic and hypotonic D-sorbitol solutions.

Supplemental Table S1. F, P, and least significant difference (LSD)_{P  = 0.05} values for two-way ANOVA analysis of the effect of species, ABA, Ca²⁺, and D-sorbitol treatments, and their interactions on stomatal aperture in Figure 2.

Supplemental Table S2. F, P, and LSD_{P  = 0.05} values for two-way ANOVA analysis of the effect of species and ABA treatment, and their interaction on H₂O₂ fluorescence, NO fluorescence, Ca²⁺ fluorescence in Figure 4.

Supplemental Table S3. F, P, and LSD_{P  = 0.05} values for two-way ANOVA analysis of the effect of species and ABA treatment, and their interaction on net K⁺ and net Ca²⁺ flux rate in Figure 5.

Supplemental Table S4. F, P, and LSD_{P  = 0.05} values for two-way ANOVA analysis of the effect of species and ABA treatment, and their interaction on vacuole diameter, vacuole number, and stomatal aperture in Figure 6.

Supplemental Table S5. F, P, and LSD_{P  = 0.05} values for two-way ANOVA analysis of the effect of species and ABA treatment, and their interaction on filament occupancy, filament skewness, and stomatal aperture in Figure 7.

Supplemental Table S6. F, P, and LSD_{P  = 0.05} values for two-way ANOVA analysis of the effect of species and D-sorbital treatment, and their interaction on the vacuole diameter, vacuole number, and stomatal aperture in Supplemental Figure S5.

Supplemental Table S7. F, P, and LSD_{P  = 0.05} values for two-way ANOVA analysis of the effect of species and D-sorbital treatment, and their interaction on the filament occupancy, filament skewness, and stomatal aperture in Supplemental Figure S6.

Supplemental Table S8. F, P, and LSD_{P  = 0.05} values for two-way ANOVA analysis of the effect of species and ABA treatment, and their interaction on P_f (in) and P_f (out) in Figure 8.

 

X.W.F., L.G., and X.D.L. conceived and designed the experiments; L.G. and X.D.L. performed most of the experiments; Y.Y.Z., X.Q.T., and Y.L.L. analyzed the data and assisted with the experiments; L.G., X.W.F., and N.C.T. wrote the manuscript.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/general-instructions) is: Xian-Wen Fang (fangxw@lzu.edu.cn).