1. Introduction

Stomata on the leaves of vascular plants dynamically control transpirational water loss (Meinzer, 1993). Under sufficient soil moisture and saturating light, stomatal conductance to water vapour (g_s) is mainly regulated by vapor pressure difference between leaves and the atmosphere (VPD). Stomata open to maximal extent at a low VPD in the morning, achieving the highest rate of carbon assimilation at this time. Once VPD begins to increase as the day progresses stomata tend to close, reducing water potential decline to avoid a critical water tension that would induce xylem embolism (Sperry, 2000; Choat et al., 2018; Durand et al., 2019). In angiosperms, after a sudden increase in VPD, stomata often open initially because of a passive contraction of the epidermal cells, termed "wrong-way response" (WWR); showing a period of wrong-way opening stomata, and then closing gradually and reaching a new final steady-state (Buckley, 2016). This stomatal closure after a wrong-way opening is believed to be driven by the hormone abscisic acid (ABA), which is synthesized in leaves when mesophyll cells approach turgor loss point (McAdam and Brodribb, 2016). Unlike in species of lycophyte, fern and conifer, in which the kinetics of stomatal responses to VPD can be readily predicted by a passive hydraulic model of stomatal regulation (McAdam and Brodribb, 2015), in angiosperms stomatal closure at high VPD is believed to be largely driven by an active metabolic signal such as ABA (Brodribb and McAdam, 2011). Stomatal kinetics from initial steady-state to final steady-state in response to environmental fluctuation influences the balance between carbon assimilation (A) and water loss, and for this reason differences in kinetics can be adaptively relevant (Buckley, 2016; Meinzer et al., 2017; Lawson and Vialet-Chabrand, 2019; Durand et al., 2019). A fast stomatal kinetics, by closely tracking environmental perturbation, could enable stomata to operate optimally (Drake et al., 2013; Meinzer et al., 2017). Moreover, fast stomatal closure restricts transpirational water consumption and prevents hydraulic failure during rapid changes to high atmospheric water deficit (Martin-StPaul et al., 2017; Rodriguez-Dominguez and Brodribb, 2020).

Firstly, the rapidity of the g_s response to an increase in VPD in angiosperms might arise mechanistically from stomatal morphology (e.g., stomatal size and density). Smaller stomata have faster kinetics and greater responses to change in light, owing to a greater guard cell membrane surface area to volume ratio, which facilitates more rapid ion exchange (Drake et al., 2013). This hypothesis has been supported by some studies (Drake et al., 2013; Raven, 2014; Durand et al., 2019), but has not yet reached consensus (Elliott-Kingston et al., 2016; McAusland et al., 2016; Lawson and Vialet-Chabrand, 2019). Moreover, it has been found that stomatal morphologies, such as stomatal size and density, are correlated with genome size in angiosperms (Beaulieu et al., 2008; Simonin and Roddy 2018). If stomatal morphology mechanistically influences stomatal movement, a key unanswered question is whether the genome size in seed plants influences stomatal responses to fluctuating environmental conditions through its influence on stomatal morphology?

Additionally, stomatal responses to temporal rise in VPD might be affected by the hydraulic feedbacks that depends on leaf hydraulic resistance and water status (Buckley, 2005, 2019). For a species being highly vulnerable to dehydration, leaf hydraulic conductance often declines rapidly between full turgor and turgor loss point, which can further drive stomatal closure as a consequence of decrease in leaf water potential (Ψ_leaf) during leaf dehydration (Brodribb et al., 2014; Skelton et al., 2018). By contrast, in some species the leaf hydraulic system has strong resistance to water stress, with leaf hydraulic conductance declining slowly between full turgor and turgor loss point, and thus delaying stomatal closure during leaf dehydration (Scoffoni et al., 2012, 2017; Skelton et al., 2017). Stomatal regulation is influenced also by leaf water content (WC_leaf) with changing in cell volume, and hence the cell turgor pressure (Trueba et al., 2019; Fu et al., 2019; Xiong and Nadal, 2020). Leaf capacitance (C_leaf) is considered a key trait of water storage capacity in leaves, alleviating against rapid fluctuation in water potential and being involved in desiccation avoidance (Xiong and Nadal, 2020). Stomatal closure may thus be delayed in plants with a high C_leaf. Except in ferns and lycophytes (Martins et al., 2016), the relationship between the speed of stomatal closure and C_leaf in seed plants has been rarely examined explicitly.

Mangrove plants live in intertidal coastal zones of the tropics and subtropics (Duke et al., 1998), which are associated with high salt concentration and hypoxia in the soil rhizosphere as well as high atmospheric evaporative demand (Reef and Lovelock, 2015). To adapt to such habitats, mangroves have developed efficient regulation of water consumption, high resistance to xylem embolism, high tolerance to desiccation and high leaf capacitance, and they have hydraulic redistribution from the foliage to the stem for preserving stem hydraulic function (Reef and Lovelock, 2015; Jiang et al., 2017, 2021, 2022; Aritsara et al., 2022; Beckett et al., 2024). Mangroves have longer guard cell length for a given stomatal density than other vascular plants (Agduma et al., 2022) and evolved decreased genome size in adaptation to the intertidal habitat (He et al., 2022). Mangroves are confronted with increasing challenges for survival under global climate change, for example, a massive mangrove mortality in northern Australia occurred due to a severe El Niño event associated drought (Gauthey et al., 2022). Climate change will result in dramatic increase in VPD in tropical and subtropical regions. Yet, relatively little is known about the stomatal dynamics of mangroves in response to high VPD. In this study, we quantified the speed of stomatal closure in response to a step-rise in VPD in situ, and measured stomatal morphological characteristics, pressure–volume curves and hydraulic vulnerability to dehydration of leaves in nine true mangrove tree species. We aimed to answer the following two questions: (1) Does stomatal morphology influence stomatal dynamics in response to a high VPD in mangroves? with a consideration of possible influence of genome size on stomatal morphology; and (2) Do leaf hydraulic traits influence stomatal sensitivity to VPD in mangroves?

2. Materials and methods 2.1. Study site and plant material

The study sites were located at the Dongzhaigang Harbor National Nature Reserve in Hainan (19°57′ N, 110°35′ E), which has the largest mangrove forest area in China, and the Qinglangang Harbor Provincial Nature Reserve in Hainan (19°37′ N, 110°50′ E), which has the highest richness of mangrove species in China, respectively (Fig. 1). The two mangrove reserves are about 55 km apart and have similar climates, classified as a maritime monsoonal tropical climate. Mean annual temperatures are 23.8 ℃ (Jiang, 2021) and 23.9 ℃ (Leng, 2020) for the two sites, respectively; mean annual precipitation is about 1700 mm (Leng, 2020; Jiang, 2021), which is largely concentrated from May to November (Leng, 2020; Jiang, 2021); mean annual evaporation is about 1800 mm (Han, 2011; Leng, 2020); mean annual relative humidity is 85% (Han, 2011) and 87% (Leng, 2020), respectively. The soil types are mainly sandy saline soil and marsh saline soil; the salinity and water potential of soil is 5.71 g kg⁻¹ and -1.88 MPa in Dongzhaigang, respectively, and 4.85 g kg⁻¹ and -1.14 MPa in Qinglangan (Table 1). The salinity and water potential of tide is 23.84 g L⁻¹ and -1.72 MPa in Dongzhaigang, respectively, and 10.06 g L⁻¹ and -0.84 MPa in Qinglangan (Table 1).

Five mangrove species were selected in Dongzhaigang, including Aegiceras corniculatum (Myrsinaceae), Avicennia marina (Verbenaceae), Ceriops tagal (Rhizophoraceae), Kandelia obovata (Rhizophoraceae), and Rhizophora stylosa (Rhizophoraceae). Four mangrove species were selected in Qinglangang, including Bruguiera gymnorhiza (Rhizophoraceae), B. sexangula (Rhizophoraceae), Sonneratia alba (Sonneratiaceae), and Xylocarpus granatum (Meliaceae). These species are from five different families, all true mangrove species and the dominant or common species in the communities. They vary considerably in leaf hypodermal layers and stomatal morphology (Chen, 2020), and thus we would expect to find a diversity of stomatal regulation strategies across these species. Three to six mature, healthy and sun-exposed individuals of each species were selected for the experimental measurements. The information about height, basal diameter (BD), and the number of cell layers of the leaf hypodermis of the sampled plants are shown in Table 2.

2.2. Kinetics of stomatal response to temporal rise in VPD

Three to six, fully expanded and sun-exposed leaves of each species from different individuals in the field were selected for gas exchange measurement using a portable photosynthesis system (Li–Cor 6800; Li–Cor Inc.) on a clear day between 08:00 and 11:00 h (local time) from August to November 2020. Leaves were acclimatized inside the leaf cuvette (Photosynthetic Active Radiation, PAR, 1600 μmol m⁻² s⁻¹; CO₂ concentration, 400 ppm; relative humidity, RH, 70%; flow rate, 500 μmol s⁻¹) until stomatal conductance (g_s) reached an initial steady-state (SS_initial; defined as a variation < 5% over 3 min), during which the leaf chamber temperature were controlled to 0–2 ℃ below the air temperature. At this point, RH was then switched to 20%, resulting in a step-increase in VPD while no change in other environmental parameters, and g_s showed a transient increase for about 3 min, called the "wrong-way response" (WWR), then declined gradually until it reached a new final steady-state (SS_final). The VPD was about 0.95 kPa and 2.65 kPa before and after changing RH, respectively (Fig. S1). The value of g_s was recorded every 20 s during the entire measurement period. After the measurements of gas exchange, the intact branches from which the leaves grew were marked for subsequent measurements of pressure–volume curves and hydraulic vulnerability to dehydration of leaves.

2.3. Leaf predawn and midday water potential

On the day of measuring gas exchange, leaf predawn water potential (Ψ_pd) and midday water potential (Ψ_md) were measured using a pressure chamber (PMS Instruments) between 4:30 to 6:30 h, and 12:30 to 14:30 h, respectively. For measuring Ψ_pd, two leaves from four-five individuals per species were wrapped up with aluminium foil and plastic bags on the previous evening of gas exchange measurements. For measuring Ψ_md, two leaves from the same individual as the leaves used for Ψ_pd measurements were excised, and then immediately enclosed in aluminium foil and plastic bags. All measurements were made in the field within 10 min after excision.

2.4. Pressure–volume curves

Pressure–volume measurements were performed on leaves using the bench drying procedures described by Tyree and Hammel (1972). A branch neighboring those measured for gas exchange was harvested for each species and rehydrated overnight for next-day measuring from August to November 2020. One fully expanded leaf was collected from each rehydrated branch and dehydrated on the bench in a temperature-controlled room with a good-ventilation. The leaf was repetitively weighted and measured for water potential with a pressure chamber (PMS Instruments). The early and late interval time of each curve measurement was 0.5–2 h and 4–8 h, respectively. Determining a complete pressure–volume curve with least 10 points required leaves to dehydrate for 2–4 days. Furthermore, leaf area was analysed using a scanner and dry mass was determined after oven-drying for 72 h at 60 ℃. From three to six pressure–volume curves for each species (Fig. S4), the following parameters were determined as in Sack and Pasquet-Kok (2011): osmotic potential at full turgor (π_o) and at turgor loss point (π_tlp), bulk modulus of elasticity (ε) and leaf area specific capacitance at full turgor (C_leaf) (Table S1).

2.5. Leaf hydraulic vulnerability curves

Branches from three to six individuals, approximately 110 cm in length that at least twice the maximum vessel length of these species (Jiang, 2021), neighboring those measured for gas exchange for each species were collected and recut immediately under water in May to June 2021. Furthermore, the cut ends of the branches were wrapped in damp towels, which were wrapped tightly by tiny plastic bags with some amount of water in it. Then, the branches were quickly placed in black plastic bags and brought back to the laboratory. All branches were recut underwater again, sealed with in black plastic bags, and allowed to rehydrate overnight in the laboratory. The cut ends of branches were wrapped by wax and parafilm (PM996; BEMIS). Leaf vulnerability curves were obtained using the rehydration method described by Brodribb and Holbrook (2003). According to the leaf vulnerability curve with least 60 points for each species (Fig. S5), we determined the maximum hydraulic leaf conductance (K_leaf), the water potentials inducing 12% (P₁₂) and 50% (P₅₀) loss of the maximum conductance (Table S1).

2.6. Modelling g_s responses to a step-rise in VPD

Stomatal temporal dynamics in response to a single step-rise in VPD was fitted empirically using an analytical sigmoidal model (Vialet-Chabrand et al., 2013) to evaluate the specific parameters in the response curve of g_s (Fig. 2). It describes as (Eq. (1)):

(1)

where g_s (mol·m⁻²·s⁻¹) is the g_s at the corresponding time (s), λ the initial time lag of the sigmoidal curve (s), τ the time constant of g_s response (s), g_s-initial and g_s-final (mol·m⁻²·s⁻¹) are the steady-state values of g_s at the initial and final stages of a sigmoidal curve, respectively. e is Euler's number (c. 2.718). Based on these parameters, we obtained the second parameter, the maximum slope (SL_max) (mmol·m⁻²·s⁻¹) (Eq. (2)) as an estimator of combining speed and amplitude of the g_s response to the step-rise in VPD:

(2)

More details on the model and its parameters can be referred to the literature (Vialet-Chabrand et al., 2013; Durand et al., 2019).

2.7. Leaf morphology and stomatal anatomy

Leaves used for gas exchange measurements or in the neighboring were collected, enclosed in plastic bags, and immediately refrigerated at 4 ℃. Each species had five leaves from different trees in total. Transverse cross-sections avoiding major veins were made with hands. All sections were stained with 1% Safranin-Alcian blue, washed with distilled water, and then mounted on microscope slides. They were observed under a light microscope (Leica DM 3000 LED Wetzlar Germany), and imaged for measurement of tissue thickness. Paradermal sections were sliced from each leaf with a square puncher (~1 cm²), avoiding major veins, and soaked overnight in a bleach solution composed of hydrogen peroxide and acetic acid (H₂O₂: CH₂COOH = 1:1) at 70 ℃ in an oven. For the measurement of leaf vein density, transparent paradermal sections were stained with 0.5% safranin, washed with distilled water, and then mounted on microscope slides. They were observed under the light microscope, capturing five images per leaf at magnifications of 5× for vein length measurements. For the measurement of stomatal traits, transparent epidermal samples were isolated from the mesophyll and washed with distilled water. These samples were stained with 1% Safranin-Alcian blue, washed with distilled water, and then mounted on microscope slides. They were observed under the light microscope, capturing five images per leaf at magnifications of 10× for stomatal density and 40× for stomatal size measurements. ImageJ software was used to measure leaf thickness (LT), hypodermis thickness (HT), vein length, stomatal number and guard cell length (GCL). The fraction of hypodermis to lamina thickness (HT_fraction) was calculated as HT_fraction = HT/LT. Stomatal density (SD) was the number of stomata per mm². Leaf vein density (VD) was calculated as the total length of veins per mm². Anatomical maximum stomatal conductance (G_s-max) was calculated according to the method described by de Boer et al. (2016b). Stomatal opening ratio at the initial steady-state (G_s-ratio) was calculated as G_s-ratio = g_s-initial/G_s-max.

2.8. Statistics and graphics

Genome sizes of nine mangrove species were collected from the two literatures, He et al., (2022); Hu et al., (2020), and were shown in Table S1. Remarkably, mangroves have a smaller genome assembly compared to other angiosperms (Jiang et al., 2023).

An independent-samples t-test was used to compare the differences in salinity and water potential of soil and tide between the two sites. Stomatal dynamics was fitted using a sigmoidal model in SigmaPlot 12.5, excluding stomatal transient WWR. The leaf vulnerability curves were fitted in SigmaPlot 12.5. Linear regression analyses were performed using species' average values to examine the interspecific relationship between leaf water transport efficiency, gas exchange and water relations in nine mangrove species. Correlations were computed using mean values by performing linear regression analyses with stomatal dynamics parameters as dependent variables and stomatal morphology, total assembled genome size, leaf osmotic pressure, leaf dehydration tolerance, leaf predawn water potential or leaf capacitance as continuous independent variables across the nine mangrove species. The correlations were considered statistically significant if p < 0.05. All regression analyses, figures and t-tests were made using the R-project v.4.1.1 (https://www.r-project.org/).

Partial Least Squares Path Modeling (PLS-PM) was used to build the network of relationships between the measured variables through latent variables for exploring the cause-and-effect link between environmental factors of the sites, leaf water status, leaf morphology, hydraulic traits, stomatal anatomy and stomatal dynamics. The conceptual framework among latent variables was established based on known relationships and previous understandings of environmental factors of the sites, leaf water status, leaf hydraulic traits, stomatal anatomy and stomatal dynamics. In this framework, the environmental factors of the sites were presented by Ψ_soil, Ψ_tide SAL_soil and SAL_tide. Leaf water status was described by Ψ_pd and Ψ_md. Leaf morphology was described by LMA, VD, LT, HT, HT_fraction. Stomatal anatomy was presented by SD, GCL, G_s-max and G_s-ratio. Hydraulic trait was presented by C_leaf, K_leaf, π₀, P₁₂ and P₅₀. Stomatal dynamics was determined by g_s-initial, g_s-final, Δg_s, λ, τ and SL_max. PLS-PM were run in the R-project v.4.1.1 (https://www.r-project.org/) using packages "plspm", and performed with 1000 bootstraps to obtain path coefficients and R² values for each latent variable and its statistical significance and standardized loading values of each measured variable to its latent variable and Goodness-of-Fit index of the path models.

3. Results 3.1. Correlations between stomatal dynamics, stomatal morphology, vein network and genome size

The nine mangrove species varied considerably in stomatal size and stomatal density (SD) (Fig. 3). Guard cell length (GCL) varied from 23.68 to 45.84 μm, and SD from 123.62 to 386.64 no·mm⁻². By the bivariate trait-relationship analysis, these stomatal morphological properties were not related to the speed of stomatal responses to a step-rise in VPD across all species from five families (Fig. 3). But for the five mangrove species of Rhizophoraceae, the stomatal response time constant (τ) was positively correlated with GCL (R² = 0.773; Fig. 3a) and negatively with SD (R² = 0.78; Fig. 3b); and the maximum speed of stomatal closing (SL_max) also showed a positive relationship with SD (R² = 0.797; Fig. 3c). However, in our path modeling, the variation in stomatal dynamics was affected directly by stomatal anatomy with the total effect of -0.64 (Fig. 7). These results indicate that stomatal behavior in response to a step-rise in VPD are regulated directly by stomatal anatomy, with smaller and more numerous stomata have faster kinetics than leaves with larger stomata, while across diverse mangrove genera there is no general relationship between stomatal anatomy and responsiveness.

Furthermore, genome size was significantly correlated negatively with GCL and positively with leaf vein density (VD) among the mangrove species excluding an outlier (Fig. 3d, f). However, genome size was not associated statistically with the speed of stomatal responses to a step-rise in VPD across all species, or even within Rhizophoraceae (Fig. S2). We also observed that SD was positively correlated with VD and the leaf osmotic potential at full turgor (π_o) (Figs. S3a–b), and the initial steady-state g_s (g_s-initial), the final steady-state g_s (g_s-final) and the amplitude of variation in g_s (Δg_s) in response to a step-increase in VPD were significantly and positively correlated with π_o among the mangrove species excluding an outlier (Fig. 4a–c). However, π_o was not statistically related to the parameters of stomatal dynamics in responses to a temporal rise in VPD across the nine mangrove species, even if excluding an outlier (Fig. 4d–f). g_s-initial was significantly and positively correlated with the maximum leaf hydraulic conductance (K_leaf) among the mangrove species (Fig. S3c). These results indicate that operating steady-state g_s is related to leaf osmotic pressure and hydraulic conductance, which is closely linked to stomatal morphology and vein construction, controlled by genome size.

3.2. Stomatal dynamics in relation to leaf hydraulic traits

Firstly, the species with less negative P₁₂ had a higher maximum stomatal response speed (SL_max) and faster the time constant (τ) and the lag time (λ) of g_s response during stomatal closing response to a step rising VPD (Fig. 5a–c), but the water potential inducing 50% loss of the leaf maximum conductance (P₅₀) was not statistically related to SL_max, τ, and λ (Fig. 5d–f), indicating that P₁₂ and not P₅₀, might be critical for stomatal regulation and that a lower vulnerability of leaves to dehydration could lengthen stomatal closing time in response to a step-rising of VPD. Moreover, leaf capacitance (C_leaf) was positively and significantly correlated (R² = 0.775) with τ and was negatively correlated (R² = 0.714) with SL_max (Fig. 6a–b) among the mangrove species excluding a single outlier species (Ceriops tagal), indicating that under a high transpiration demand, high C_leaf could delay the stomatal closure. Moreover, the results of the path analysis reveal that the variations in stomatal dynamics were explained directly by leaf hydraulic traits with the total effect of 0.88 (Fig. 7), which reinforced our observation based on bivariate relationship stated above.

Additionally, the species with lower bulk modulus of elasticity (ε) had higher leaf capacitance (C_leaf) (Fig. S3d), which was related significantly to lamina hypodermis thickness (HT) and hypodermis fraction (HT_fraction) (Figs. S3e–f). These results indicate that leaf structural traits that are associated with water storage, especially in the hypodermis, have indirect effects on stomatal dynamics through the influence on hydraulic traits (Fig. 7). Leaf water status, which was proportional to site salinity, had also indirect impacts on stomatal dynamics by changing in hydraulic traits (Fig. 7), as that species with more negative P₁₂ and P₅₀ had a more negative leaf predawn water potential (Ψ_pd), and the latter was negatively correlated with λ (Fig. 5g–i).

4. Discussion

We found evidence that the morphology of stomata and hydraulic traits of leaves influences stomatal response to a step-increase in VPD in mangrove trees. Rhizophoraceae species (but not all mangrove species) with smaller and denser stomata had faster stomatal responses to the step-rise of VPD. Across all species those with higher vulnerability to initial loss of hydraulic conductance during dehydration had faster stomatal responses and the opposite was true for the species with higher leaf capacitance. In addition, partial least squares path modeling showed that the water potential and salinity of the sites determined leaf water status of mangroves. They in turn influenced leaf morphology, which indirectly affected on stomatal dynamics through the variations in both stomatal anatomy and hydraulic traits in mangroves. These regulatory behaviors of stomata could be of vital importance for the long-term survival of mangrove trees in a stressful environment.

4.1. A key role of stomatal morphology in stomatal dynamics

Smaller stomata have often considered to have faster kinetics (Drake et al., 2013). In the present study, we observed a trend of increasing speed of stomatal closure with increasing stomatal density (SD) and decreasing guard cell length (GCL) in the five mangrove species of Rhizophoraceae, but this pattern was not found across all nine mangrove species from different families. Elliott-Kingston et al. (2016) found no relationship between stomatal closure rapidity and size or density of stomata in an evolutionary diverse series of species (including fern, cycad, conifers and angiosperms). Likewise, McAusland et al. (2016) examined rapidity of stomatal closure from species over a range of crops with differences in stomatal morphology (kidney- or elliptical-shaped) and found that stomatal kinetics was also not explained by the size of stomata. Overall, our study is generally consistent with previous studies and suggests that the relationship between stomatal closing speed in response to a step-rise in VPD and stomata size is conserved in closely related species, but not across lineages. There may be mechanisms that affect stomatal movement speed in response to ambient environmental fluctuation independent of stomatal size across species with diverse stomata morphology. As shown by the present study that stomatal behavior was strongly regulated by hydraulics (e.g., the water potential inducing 12% loss of leaf hydraulic conductance, P₁₂) and water-relations traits of leaves (discussed below).

We also observed that operating steady-state g_s in responses to a step-increase in VPD were determined by leaf osmotic adjustment, and that species with a higher SD had a less negative bulk leaf osmotic potential at full turgor (π_o). The positively relationship between SD and π_o would likely constrain stomatal behavior in response to high atmospheric water deficit. These findings are consistent with the recent proposal that species with less negative π_o and osmotic potential at turgor loss point (π_tlp) had smaller, denser stomata and higher maximum operating steady-state g_s and greater sensitivity in stomatal closure during leaf dehydration over a wide range of species (Henry et al., 2019). Our work on the nine mangrove species, which have more negative π_o and π_tlp than most of terrestrial plants, provides an important support for the idea of stomatal safety-efficiency trade-off (Henry et al., 2019). However, leaf osmotic pressure was not correlated with stomatal dynamics in responses to a temporal rise in VPD across the nine mangrove species in this study. The stomatal response to a high VPD might involve in the dynamics of synthesis and breakdown of some biochemical materials (e.g., abscisic acid and protein kinases) (McAdam and Brodribb, 2015; Jalakas et al., 2021).

Our results also showed that genome size was correlated significantly and negatively with guard cell length (GCL), and positively with vein density (VD) among mangrove species excluding an outlier. It must be noted that the inverse relationship of genome size and stomatal size across the mangrove species is not in conflict with previous studies (Beaulieu et al., 2008). The universal relationship between genome size and GCL is found at large-scale comparative analyses over a wide range of species (Beaulieu et al., 2008), while the negative relationship in mangroves of our study was shown by a little variation in genome size at small-scale from nine species. This finding is consistent with recent work on genome size, leaf cell size and cell packing density relationships in a range of mangroves (Jiang et al., 2023). Nevertheless, genome size showed nonsignificant association with the rapidity of stomatal closure response to an increase in VPD among the nine mangrove species. Compared to other angiosperms, mangroves had a smaller genome (Jiang et al., 2023), a faster VPD response (Fig. S6), and a larger GCL and higher vein density at a given SD (Agduma et al., 2022). Therefore, mangroves adapted to the stressful intertidal habitat may change the genome size-stomatal dynamics relationship for optimal water use.

4.2. Leaf hydraulic vulnerability and capacitance affect stomatal regulation behavior

Our results showed that the water potential inducing 12% loss of the maximum conductance of leaves (P₁₂) is a strong regulator of the kinetics of stomatal closure in response to a temporal rise in VPD in the mangroves. Firstly, P₁₂, which is the initial inflection point of the vulnerability of leaf hydraulic conductance to dehydration, is considered as a potential trigger for stomatal closure as explained by simple hydraulic feedback (Buckley and Mott, 2002; Buckley, 2005). Secondly, P₁₂ reflects the primary hydraulic resistance to bulk flow of the liquid-phase moving from xylem into the bundle sheath to epidermal cells and guard cells within leaves during dehydration (Sack and Holbrook, 2006). It may drive an initial increase in the opposing effect of the epidermal pavement cells on guard cells (Darwin, 1898; Cowan, 1977), triggering hydraulic feedback for stomatal closure (Buckley, 2019). Lastly, the whole-leaf hydraulic resistance also includes the vapour-phase diffusion from the evaporating sites to stomata (Buckley et al., 2017), which is linked with the liquid transport in transpiring leaves and closely related to water potential (Rockwell et al., 2014). It, therefore, is likely that species with more negative P₁₂ have faster vapour diffusion through stomata (Peak and Mott, 2011), which would delay stomatal closure in response to atmospheric drought. However, the water potential corresponding to 50% loss in leaf hydraulic conductance (P₅₀) was not significantly correlated with the kinetics of stomatal closure in response to a step-rise in VPD across the mangrove species. The adaptation of P₁₂, not P₅₀, as a strong linker of kinetics of stomatal closure in response to a step-rise in VPD may contribute to the trigger of hydraulic feedbacks and translate this message into next effector that participates in the dynamic behavior of stomata (Tardieu, 2016; Choat et al., 2018). This initial hydraulic P₁₂ message drives fast stomatal closure in response to a step rise VPD to limit transpiration and timely prevent a dramatic decline in leaf water potential (Buckley, 2019), and thus is of crucial importance for the long-term survival of mangrove trees.

Additionally, we observed that mangroves developed thick or even upper and lower hypodermis, contributing a large proportion of lamina thickness and storing a considerable amount of water inside leaves, thus had high leaf capacitance (C_leaf). Leaves with a high C_leaf can buffer fluctuation in leaf water potentials when facing atmospheric drought, by compensating transpirational water loss from storage water. Therefore, we observed a significant trend of increasing time of stomatal closure in response to a step-rise in VPD in the mangrove species with higher C_leaf. This is consistent with previous studies (Martins et al., 2016; Fu et al., 2019; Xiong and Nadal, 2020). Although the high C_leaf of the mangroves may delay stomatal closure at high VPD, the mangrove species with a high C_leaf had a faster stomatal VPD response than the six glycophytic woody species measured by Fu et al. (2019) (Fig. S6). Thin leaves of these glycophytic species could systematically alleviate the instantaneous decline of leaf water potential under an atmospheric drought due to a shorter hydraulic pathway from vascular bundle to stomata so that generate a smaller water potential gradient across mesophyll tissues (Brodribb et al., 2007; de Boer et al., 2016a; Buckley et al., 2017) and thus slower stomatal VPD response, because water potential would not drop close to π_tlp when abscisic acid (ABA) would be synthesized in leaves (McAdam and Brodribb, 2016). By contrast, the thick leaves of mangrove trees with longer outside-xylem pathways of water transport are difficult to maintain water balance between mesophyll cells when facing rapid changes in atmospheric moisture (Brodribb et al., 2007; de Boer et al., 2016a; Scoffoni et al., 2017), resulting in a larger change in turgor pressure of guard cells. This might accelerate the transmission of hormonal signal for ABA biosynthesis inside leaves for faster stomatal closure to reduce water loss (Tardieu, 2016; McAdam and Brodribb, 2016; Buckley, 2019) and thus faster stomatal VPD response in these mangrove species. Moreover, epidermal cells of mangroves are smaller than other angiosperms and also smaller than their own stomata (Jiang et al., 2023), meaning that stomata of mangroves have a higher density of peripheral epidermal cells. This design feature might contribute to the stomatal movement as suggested by Gérardin et al. (2018), and thus faster stomatal closure for mangroves. Furthermore, mangroves for the adaptation to super-saline intertidal habitat have evolved a series of water-protection properties, such as highly embolism-resistant xylem (Jiang et al., 2017, 2021, 2022), high hydraulic-capacitance in leaf and root tissues (Aritsara et al., 2022), and strong osmotic regulation (Reef and Lovelock, 2015). They therefore may have enhanced stomatal regulation with a highly sensitive stomatal response to VPD suited to this adaptation to water-deficit prone habitats.

We also found an indirect effect of leaf water status on stomatal dynamics that was mediated mainly by leaf hydraulic traits in our path model. Leaf water status is related to the water absorbed by roots from the soil and water evaporated from stomata into the atmosphere (Cowan, 1965; Buckley, 2019). The water potential in leaf mesophyll and epidermis firstly were reduced after a step increase in VPD, which increased resistance in the leaf water transport (Buckley et al., 2017; Buckley, 2019). As the above results shown in this study that species with more negative leaf predawn water potential (Ψ_pd) had more negative P₁₂ and P₅₀, and thus regulating stomatal dynamics. For mangroves, the water potential and salinity of the soil and tide determine the water status of leaves in the present study. Thus, although the water potential and salinity of the sites alone cannot explain the mechanism of stomatal VPD response, the continuous pattern of soil-plant hydraulics is often important for understanding how stomatal dynamics are regulated by leaf hydraulic traits (Rodriguez-Dominguez and Brodribb, 2020; Cai et al., 2023).

Our results also showed that the leaves of mangroves had a high bulk modulus of elasticity (ε). This means that cell walls are rigid, and could provide mechanical strength and maintain cellular hydration during dehydration. Our results also showed a negative relationship between C_leaf and ε across the mangrove species. This finding was consistent with recent framework on water movements and storage dynamics (Xiong and Nadal, 2020), suggesting that a tight coordination between capacitance and resistance to deformation and shrinkage under tension. Our results also showed that high operating initial steady-state g_s (g_s-initial) was closely linked to high leaf hydraulic conductance owing to high vein density in mangroves. This finding is consistent with previous suggestion that a high operating steady-state g_s, resulting from a large aera of leaf-scale stomatal pore, in angiosperms inevitably accelerates transpiratory water loss (Brodribb et al., 2007), which must be supported by water supply of a high vein density (Sack and Frole, 2006; Brodribb et al., 2007). As the path analysis results of the present study found that leaf morphology had an important direct effect on stomatal anatomy, and then had an indirect effect on stomatal dynamics. This is thought to signify the balance between water demand and supply within leaves for mangroves.

5. Conclusions

We provide evidence that stomatal response to a temporal rise in VPD is related directly to stomatal anatomy and hydraulic traits in nine true mangrove species. We also found that the stomata of mangroves have strong sensitivity in response to a step rise in VPD. Faster stomatal closure would be a strategy for conservation water during atmospheric drought, although with higher capacitance and greater embolism resistance of mangrove leaves than those of glycophytic plants. This regulation behavior is suitable for adaptation to water-deficit prone habitats and may improve the advantage of survival in rising global VPD in the future for mangroves. Besides passive stomatal control through hydraulic feedback, the active regulation appears to play a key role in stomatal closure under water deficit in mangroves. A quantitative analysis of both passive and active stomatal regulation behavior is urgently needed for plants living in stressful habitats and confronting with the changing environment.

Acknowledgements

We acknowledge the collaboration of the Dongzhaigang National Nature Reserve and Qinglan Harbor Provincial Mangrove Nature Reserve, and the Forestry Department of the Hainan Province for permission to conduct the field experiment in the mangrove reserves. The authors thank Prof. Jun-Jie Zhu for the experimental design, Yang Wei for field assistance, Jin-Yan Lei and Pei-Xin Cui for their help in leaf anatomy. This study was financially supported by a grant of the National Natural Science Foundation of China (31670406) and the Bagui Fellow scholarship (C33600992001) of Guangxi Zhuang Autonomous Region to KFC.

CRediT authorship contribution statement

Ya-Dong Qie: Writing – review & editing, Writing – original draft, Software, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Qi-Wei Zhang: Writing – review & editing, Visualization, Software, Methodology, Investigation, Formal analysis, Conceptualization. Scott A.M. McAdam: Writing – review & editing, Visualization, Validation, Methodology, Formal analysis, Conceptualization. Kun-Fang Cao: Writing – review & editing, Visualization, Validation, Supervision, Resources, Project administration, Funding acquisition, Data curation, Conceptualization.

Declaration of competing interest

The authors declare no conflict of interest.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.pld.2024.02.003.