Stomatal encryption by epicuticular waxes as a plastic trait modifying gas exchange in a Mediterranean evergreen species (Quercus coccifera L.)

Authors

  • ANITA ROTH-NEBELSICK,

    1. State Museum Nat Hist Stuttgart, D-70191, Stuttgart, Germany
    Search for more papers by this author
  • VICTORIA FERNÁNDEZ,

    1. Forest Genetics and Ecophysiology Research Group, School of Forest Engineering, Technical University of Madrid, Ciudad Universitaria s/n, 28040 Madrid, Spain
    Search for more papers by this author
  • JOSÉ JAVIER PEGUERO-PINA,

    1. Grup de Recerca en Biologia de les Plantes en Condicions Mediterrànies, Departament de Biologia, Universitat de les Illes Balears, Carretera de Valldemossa, km 7.5, 07071, Palma de Mallorca, Balears, Spain
    Search for more papers by this author
  • DOMINGO SANCHO-KNAPIK,

    1. Unidad de Recursos Forestales, Centro de Investigación y Tecnología Agroalimentaria, Gobierno de Aragón, 50059, Zaragoza, Spain
    Search for more papers by this author
  • EUSTAQUIO GIL-PELEGRÍN

    Corresponding author
    1. Unidad de Recursos Forestales, Centro de Investigación y Tecnología Agroalimentaria, Gobierno de Aragón, 50059, Zaragoza, Spain
    Search for more papers by this author

E. Gil-Pelegrín. Fax: +34 976716335; e-mail: egilp@aragon.es

ABSTRACT

The adaptive benefit of stomatal crypts remains a matter of controversy. This work studies the effect on gas exchange of cuticular rims that overarch the stomatal pore in the Mediterranean species Quercus coccifera L. growing under Mediterranean (lower relative humidities and high summer temperatures) or oceanic conditions (higher daily relative humidities and mild temperatures). After microscopic assessment of the leaf surfaces and stomatal architecture, the impact of the cuticular ‘cup’ on gas exchange was evaluated by employing three-dimensional finite element models. Here, we provide evidence for a high plasticity of the Q. coccifera cuticular cup, with much larger vents under oceanic conditions compared to small vents under Mediterranean conditions. This structure adds a substantial fixed resistance thereby strongly decreasing gas exchange under Mediterranean conditions. The cuticular cup, which also increases leaf internal humidity, might buffer the rapid changes in vapour pressure deficit (VPD) often observed under Mediterranean conditions. Since water loss of guard and adjacent epidermal cells regulates stomatal aperture, we suggest that this structure allows an efficient regulation of stomatal conductance and optimum use of resources under high VPD. This study provides evidence that plasticity of stomatal architecture can be an important structural component of hydraulic adaptation to different climate conditions.

INTRODUCTION

Water stress has been widely proposed as a major environmental constraint to land colonization by plants and the development of a water-impermeable barrier is considered as a first important step to achieve homoiohydry (Edwards, Kerp & Hass 1998). As a consequence, the surfaces of the aerial organs of the earliest land plants were already covered with a cuticle chiefly made of cutin and waxes (Buschhaus, Herz & Jetter 2007). The cuticle prevents, however, not only water loss but also uptake of CO2 (Woodward 1998). The development of stomata combined restriction of water loss with sufficient CO2 uptake by allowing a regulated gas exchange and enabled plants to colonize terrestrial environments with an upright growth habit (Casson & Hetherington 2010). Since stomata are the main interface for gas exchange between the leaf interior and the free atmosphere (Kaiser 2009), it is generally accepted that their morphological and architectural features represent adaptations to environmental factors influencing transpiration and photosynthesis rates (Larcher 2003).

Recent studies showed the relevance of morphological and mechanical stomatal diversity for gas exchange control and performance (Feild et al. 1998; Franks & Farquhar 2007; Roth-Nebelsick 2007). Stomatal pore dimensions (length, width and depth) and stomatal density define the actual stomatal resistance to gas diffusion (Parlange & Waggoner 1970; Franks & Beerling 2009; Kaiser 2009). Franks & Beerling (2009) found that inverse correlation of changes in stomatal density and stomatal size over geologic time can be linked to atmospheric CO2 in order to optimize gas exchange.

Besides stomatal size and density, other anatomical or morphological modifications affect gas exchange between the leaf and the turbulent atmosphere (Hill 1998). In many species, the stomatal pore is not on the same level with the leaf surface. Stomata can be clustered in depressed epidermal areas in the form of crypts (Metcalfe & Chalk 1979) or longitudinal grooves (Jordan et al. 2008). In other cases, such protection seems to be achieved by the individual encryption of stomata, through the development of very thick cuticles that overarch it, forming a ‘volcano-like’ structure (Jordan et al. 2008), or a wax plugged antechamber in the sunken stomata of most conifer species (Jeffree, Johnson & Jarvis 1971; Brodribb & Hill 1997). Such stomatal protection has been reported in many xeromorphic species; hence, it is often considered as an adaptation to dry climates (Raven, Evert & Eichorn 2005). However, evidence for any association between different stomatal protection mechanisms and aridity is largely anecdotal (Haworth & McElwain 2008). Evolution of stomatal crypts in various species of Proteaceae shows no clear relationship with water availability (Jordan et al. 2008). By using a three-dimensional finite element (FE) model applied to Banksia ilicifolia R. Br. leaves, Roth-Nebelsick, Hassiotou & Veneklaas (2009) found that stomatal crypts only caused small effects on transpiration. In this regard, Hassiotou et al. (2009) proposed that the stomatal crypts in Banksia species facilitate CO2 diffusion from the abaxial surface to adaxial palisade cells. The development of plugged stomata in conifers and other species may have evolved under wet conditions (Brodribb & Hill 1997) as a device for keeping the stomatal pore free of water (Turrell 1947; Feild et al. 1998).

The anatomical or morphological modifications of stomata can thus be controversially discussed under floristic (Hill 1998), functional (Roth-Nebelsick et al. 2009) or evolutionary (Jordan et al. 2008) aspects. Such controversial views make further investigations desirable. In this study, the effect on water and CO2 fluxes associated with the individual encryption of stomata – in this case by epicuticular waxes – is investigated in Quercus coccifera L. This species was selected since it is a sclerophyllous shrub that is considered one of the most xeromorphic species of the genus Quercus that forms woody vegetation in the most xeric areas of the Mediterranean phytoclimate (Ebro basin) and occurs also under more humid temperate oceanic conditions. This species can survive long periods of drought at very high air temperatures and very low air humidity (Vilagrosa et al. 2003).

In this study, microscopic analyses of the stomatal structure, including the sub-stomatal chamber, were carried out together with three-dimensional FE model calculations of gas exchange that allowed for analysing the effect of Q. coccifera individual stomatal encryption on transpiration. Two different environmental scenarios, comparable to those found in natural habitats, were considered.

MATERIALS AND METHODS

Study sites

This study was performed with Q. coccifera leaves collected from two climatically contrasting locations in the Iberian Peninsula, namely: (1) a temperate site (43°13′N, 02°01′W, 70 m a.s.l.), and (2) a Mediterranean site (41°14′N, 00°02′W, 152 m a.s.l.). For each site, the climate was analysed from daily data collected in nearby meteorological stations (Table 1). To quantify the aridity of each site, we calculated Martonne's aridity index (P/(T + 10), where P is the annual precipitation in mm and T is the mean annual temperature in °C) and the Gaussen index (the number of months in which P < 2T, where P is the monthly precipitation in mm and T is the monthly mean temperature in °C). Moreover, the frequency distribution of the relative humidity values (RH, %) from dawn to sunset was compared during summer for the temperate and Mediterranean site (Fig. 1). It should be noted that during the summer, almost 50% of the daily hours with non-limiting light conditions registered RH values below 30% in the Mediterranean site. Furthermore, the mean of the lowest RH values that occurred throughout the climate recordings (RHmin,abs,d) was determined. For this, the lowest RH value was identified for each summer month and the mean value of these data was calculated (Table 1).

Table 1. Characteristics of the study sites
SiteTemperateMediterranean
  1. T and Tmax,w are, respectively, the mean annual temperature and the mean maximum temperature of the warmest month. P and Pd are, respectively, the total annual precipitation and the total precipitation of the driest month. RHmin,d and RHmin,abs,d are, respectively, the mean minimum relative humidity of the driest month and the mean minimum absolute relative humidity of the driest month. RHmin,abs,d is the mean of the lowest RH values that were found for each summer month within the recorded period.

  2. MAI, Martonne's aridity index.

T/Tmax,w (°C)14.5/24.515.4/32.3
P/Pd (mm)1631/75298/9
RH min,d/RHmin,abs,d (%)63.8/33.522.9/10.5
MAI6712
Gaussen index08
Figure 1.

Frequency distribution of the relative humidity values (RH, %) from dawn to sunset during summer for the temperate and the Mediterranean sites.

Microscopic observations

The size of the orifice of the cuticular cover above the stomata, stomatal pore dimensions and stomatal density of Q. coccifera were measured on 10 young, fully mature leaves collected from 10 trees (one leaf from each tree) per site. To obtain the dimensions of the stomatal pore that is usually occluded by the cuticular cover, it was necessary to remove epicuticular waxes by immersion twice for 30 s in chloroform at room temperature (Jetter et al. 2000). These parameters were measured in micrographs obtained with a scanning electron microscopy (VP-SEM S-3400N, Hitachi, Japan, low vacuum range 6–270 Pa) using the public domain National Institutes of Health (NIH) Image program (developed at the US National Institutes of Health and available at http://rsb.info.nih.gov/nih-image/).

To further analyse architecture of stomatal apparatus, cuticular cover and sub-stomatal chamber, fresh transversal sections of fully expanded leaves of Q. coccifera were observed with a low-temperature scanning electron microscope (LTSEM, DSM 960 Zeiss, Jena, Germany, acceleration potential 15 kV, working distance 10 mm and probe current 5–10 nA). The sections were frozen in liquid N, gold sputtered and subsequently observed by LTSEM. Micrographs were analysed using NIH Image program. The dimensions of 10 stomatal channels and 10 sub-stomatal chambers per site were measured.

Model technique

Diffusion, according to Fick's law, reads for all three spatial directions as:

image(1)

where J is the diffusional flux, D is the diffusion coefficient, C is the concentration and grad is the differential operator in the three spatial directions (= d/dx, d/dy, d/dz). Only simple geometric domains such as circular or elliptic pores or tubes that do not change their cross-sectional area with height allow for solving this differential equation. For studying diffusion in more complex settings in detail, numerical methods are suitable. In the present study, the diffusion module of the commercial FE-based programme FIDAP 8.7 (Fluent Inc., Lebanon, NH, USA) was applied. The finite element method (FEM) is a numerical approach that transforms differential equations into a system of algebraic equations (Zienkiewicz & Taylor 1989). For this approach, the considered object, for example a gas space confined by a complex wall structure, is divided into a number of elements with prescribed simple geometry, usually triangular or quadrilateral elements. These are interconnected at their nodes. The system of algebraic equations is then solved by iterative numeric procedures for the population of elements which compose the considered geometry. The considered process is then simulated for this structure by calculating the physical parameters (gas concentration in this case) at each element. Changes in geometry require construction of a new mesh, whereas changes of physical constants (such as the diffusion coefficient) are simply achieved by setting the input parameters to the new desired values. In the present study, results were calculated for the steady state.

Creation of the FE model

The basic mesh model was based on anatomical information provided by leaf preparations of Q. coccifera. The model was devised to comprise a stoma consisting of pore and channel plus the sub-stomatal chamber including a portion of the underlying intercellular air space (Fig. 2a). For one model variation, the stomatal cuticular cup found in Q. coccifera growing at Mediterranean sites was placed above the stomatal pore (Fig. 2b,c). For mesh creation, the program GAMBIT was used. Sub-stomatal chamber and stomatal channel were axisymmetric. The total length of the diffusional space was 82 µm and represents stomatal channel, sub-stomatal chamber and a more expanded air space below the chamber surrounded by mesophyll cells. Geometric details are shown in Fig. 2. Two basic models were devised, one without and one with cuticular cup (Fig. 2).

Figure 2.

Basic geometry of the three-dimensional model (a) with two variations: (b) a wax roof with a circular hole, similar to the stomatal encryption realized in Quercus coccifera leaves from the Mediterranean site (b) and (c) a fully opened stoma without wax layer.

The cuticular cup model showed a layer with a circular hole above the stomatal pore, similar to the stomatal encryption observed in Q. coccifera grown under Mediterranean conditions (Fig. 3a) (model with circular pore, CP). The diameter of the cup opening was 2.5 µm and, therefore, a total area of about 5 µm2 results. This area was similar to the mean value measured for Q. coccifera growing at Mediterranean sites (Fig. 4).

Figure 3.

SEM micrographs of stomatal pores with cuticular wax cups from the temperate (a) and the Mediterranean (b) sites, without epicuticular waxes (c and d) and cryo-SEM micrograph of the stomatal channel and the sub-stomatal chamber (e) of Quercus coccifera leaves.

Figure 4.

Left columns: Area of the vent of the cuticular cups for the temperate (black column) and the Mediterranean (grey) growing sites. Right columns: Stomatal pore area (µm2) of Quercus coccifera leaves from the temperate and the Mediterranean sites. Data are mean ± SE. Different letters indicate significant differences at P < 0.05 between the temperate and the Mediterranean sites.

In the second model, no cup was present and the aperture corresponds to a fully opened stomatal pore without a wax cup covering it (Fig. 2c). In this model variation, the stomatal pore matched the elliptic shape of an opened stoma (Fig. 3c,d) (model with elliptical pore, EP). Length of the stomatal pore was 8.8 µm and the width was 4.5 µm. The total area of the elliptic pore amounted to 32.6 µm2, which was similar to the mean value measured for Q. coccifera leaves without waxes for both temperate and Mediterranean sites (Fig. 4).

For the evaporative sites, two cases were realized in the models. In the first case, evaporation was allowed to occur unrestricted everywhere at the boundary of the model with the exception of the guard cell walls, since these were usually distinctly cutinized. However, there was evidence that evaporation was also restricted for larger areas below the stomatal channel by the presence of external cuticles that covered the lower sides of epidermis cells and possibly also walls of mesophyll cells in the vicinity of sub-stomatal chambers, and more deeply within the mesophyll (Boyer 1985; Wullschleger & Oosterhuis 1989). Studies of cross-sections through leaves of Q. coccifera indicated the presence of an internal cuticle (data not shown), but it was not possible to determine how deeply it extended into the mesophyll. Furthermore, an internal cuticle was found in Quercus suber by Molinas (1991), another Mediterranean species of Quercus. In the second model variation, evaporation was restricted to the bottom of the model thereby locating the source of water vapour deeply into the mesophyll. This model setting represented therefore the extreme case of sub-stomatal evaporative restriction by an internal cuticle.

The final mesh densities result from test runs that ensured that minimum density of elements was achieved. This minimum density was obtained by subsequently increasing the number of elements until the calculated results did not change significantly.

Physical parameters – environmental conditions

The basic environmental conditions that were necessary for model calculations were temperature, humidity and CO2 concentration. The selection of the two different environmental settings was based on the prevailing climatic conditions in the two experimental Q. coccifera locations selected (Table 1, Fig. 1). The first setting, the ‘temperate case’ (TM), showed a temperature of 20 °C and an atmospheric RH of 50%. The diffusional constants for water vapour (2.42 10−5 m2 s−1) and CO2 (1.5 10−5 m2 s−1) at 20 °C were applied. The model did not take into account temperature gradients which altered the absolute concentrations of water vapour for a certain RH. Results obtained by model calculations performed by Yianoulis & Tyree (1984) indicated that evaporation at the leaf interior leads to cooling effects which decreased the temperature of the evaporating sites by approximately 0.1–0.4 °C. These altered the concentration gradients slightly but did not change the overall diffusion pattern. In the second setting, the ‘Mediterranean case’ (MD), the temperature was set to 35 °C with 10% RH. This corresponded to extreme conditions that leaves of Q. coccifera were often exposed to (Fig. 1). Diffusional constants were higher, 2.64 10−5 m2 s−1 for water vapour and 1.6 10−5 m2 s−1CO2 (1.5 10−5 m2 s−1) for CO2. The external CO2 concentration was defined to amount to 380 ppm. Since the CO2 influx into leaves was caused by assimilating cells extending deeply into the mesophyll, a CO2 sink was created by assigning a concentration of 70% (= 250 ppm) of the external value to the basis of the model (Franks & Farquhar 2007). A value of 99% RH was assigned to the evaporative sites (Farquhar & Raschke 1978). In total, eight model variations resulted from the combination of presence/absence of cuticular cup, climate and presence/absence of an internal cuticle below the stomatal pore (see Table 2). During a model run, the three-dimensional concentration gradient of water vapour and CO2 was calculated. From these concentration patterns of CO2 and water vapour along the diffusion pathway, the mass fluxes out of and into the stoma could be calculated. This in turn allowed for calculation of stomatal conductance by applying the stomatal density values of Q. coccifera.

Table 2. The different model variations that result from combination of structural parameters and environmental parameters
Model identifierModel characteristics
Wax roofClimateInternal cuticle
EPTMAbsentTemperateAbsent
EPTMICAbsentTemperatePresent
EPMDAbsentMediterraneanAbsent
EPMDICAbsentMediterraneanPresent
CPTMPresentTemperateAbsent
CPTMICPresentTemperatePresent
CPMDPresentMediterraneanAbsent
CPMDICPresentMediterraneanPresent

Statistical analysis

Student's t-tests were used to compare stomatal pore areas of leaves grown under temperate and Mediterranean conditions. All statistical analyses were carried out using SAS version 8.0 (SAS, Cary, NC, USA).

RESULTS

Stomatal anatomy

Figure 4 shows that the aperture of the cuticular cup of Q. coccifera leaves from the Mediterranean site (Fig. 3a) was significantly lower (P < 0.05) than that from the temperate site (Fig. 3b). However, no significant differences (P < 0.05) were found for the stomatal dimensions between both growing sites when epicuticular waxes were removed (Fig. 4). These results indicated that both sites only differed with respect to the degree of encryption, being identical the dimensions of stomatal apparatuses. Moreover, no differences were found in the dimensions of the stomatal channel and the sub-stomatal chamber between both sites (data not shown). Both structures are shown in Fig. 3e.

Simulation of gas exchange

Encryption of stomata, that is, the presence of a cuticular cup as developed under Mediterranean conditions, decreased gas fluxes substantially (Fig. 5). The cuticular cup added a considerable fixed resistance to the system thereby decreasing the stomatal conductance (gs) (Fig. 5a). As a result, transpiration rate was markedly decreased in the models with cuticular cup as compared to models without cup. The presence of an internal cuticle that shifted the evaporating site deeply into the mesophyll further reduced transpiration rate by decreasing the intercellular conductance (gias). Inspection of the results presented in Fig. 5a showed that the presence of a cuticular cup decreased transpiration rate at about 80%, the internal cuticle as arranged in the present study decreased transpiration rate at about 55–60% and with both structures present at about 90% (provided that the stomata were open for all cases).

Figure 5.

(a) Transpiration rate (E, mmol m−2 s−1) (black columns) and stomatal conductance (gs, mm s−1) (white columns). Whereas E differs between all models, gs is only different between the models with and without a cuticular cup. (b) CO2 influx (mmol m−2 s−1). Since the internal cuticle affects water vapour diffusion only, there is no difference between model pairs with and without an internal cuticle and therefore sufficient to present the results of the models with an internal cuticle. (c) Water-use efficiency (WUE, CO2 influx/E, µmol CO2 mmol−1 H2O) for all the model variations studied.

Accordingly, transpiration rate was highest under Mediterranean conditions without internal cuticle and cuticular cup, reaching a value of more than 35 mmol m−2 s−1. Presence of cuticular cup and internal cuticle reduced the transpiration rate down to about 3 mmol m−2 s−1. Since CO2 influx was assumed to reach deeply into the mesophyll to the assimilating sites, it was not affected by the internal cuticle. CO2 influx, however, was affected by the presence of a cup in the same way as water vapour diffusion, and about 80% lower than in the models without a cup (Fig. 5b). Values of water-use efficiency (WUE) were much higher for leaves grown in temperate versus Mediterranean conditions (Fig. 5c). However, the presence of the cuticular cup did not change WUE since restriction of gas conductance affected both gases likewise (Fig. 5).

The influence of the cuticular cup on the gas concentration gradient along and above the stomatal pore was affected by the internal cuticle. The presence of an internal cuticle led to quite low RH beneath the stomatal channel (Fig. 6a). For the models with an internal cuticle, RH was increased along the whole sub-stomatal space and particularly along the stomatal channel if a cup was present (Fig. 6a). Without internal cuticle, the humidity below the stomatal channel rose drastically (Fig. 6b). Additionally, the RH gradient became steeper towards the stomatal pore in all cases when a cuticular cup was present. For the models without an internal cuticle, the difference in RH along the channel was, however, much lower (Fig. 6b).

Figure 6.

The gradients of relative humidity (RH, %) along a line transect from the bottom of the sub-stomatal chamber to the atmosphere for Quercus coccifera stomata under Mediterranean conditions. (a) Models with internal cuticle, with cuticular cup (dashed line) and without cuticular cup (solid line). (b) Models without internal cuticle, with cuticular cup (dashed line) and without cuticular cup (solid line).

DISCUSSION

The individual encryption of stomata by epicuticular waxes in Q. coccifera leaves can be considered a xeromorphic trait, which was largely influenced by the prevailing environmental conditions during growth. This idea was supported by the higher degree of encryption found in the Mediterranean site (Fig. 3a) when compared to that found in the temperate site (Fig. 3b). As a consequence, the area of the vent of the cuticular cup at the Mediterranean site was much lower than at the temperate site (Fig. 4).To the best of our knowledge, this is the first report which provides evidence for the plasticity of stomatal protection means in relation to contrasting climatic conditions for plant growth (Table 1, Fig. 1).

Several studies focusing on the role of anatomical or morphological modifications of stomata could not confirm a significant reduction of transpiration by them (Feild et al. 1998; Hassiotou et al. 2009; Roth-Nebelsick et al. 2009). In contrast, our results in Q. coccifera showed that epicuticular wax cups caused a strong decrease in stomatal conductance (Fig. 5a). The decreased conductance led to a particularly high decrease in transpirational net water loss under Mediterranean conditions (Fig. 5a). Therefore, our results indicated that the degree of individual encryption of stomata similar to the cuticular cups in Q. coccifera had immediate implications in the context of the soil–plant–atmosphere continuum. This was in accordance with the results of Jordan et al. (2008), who found a relation between aridity and occurrence of stomatal encryption only in those cases in which the encryption was very deep, both for single stomata (individual crypts) or multiple stomatal arrangement (tightly revolute margins or longitudinal grooves).

In this regard, in a natural Q. coccifera macchia, Tenhunen et al. (1985) observed that midday water potential (Ψmd) did not decrease below −2.0 MPa in the early summer, whereas E was around 3.5 mmol m−2 s−1 and gs was c. 150 mmol m−2 s−1. These values were close to those obtained in the simulations for the Mediterranean case with cuticular cup and internal cuticle (Fig. 5). Similar Ψmd and gs values for fully irrigated Q. coccifera plants grown in containers under Mediterranean summer conditions have been reported by other authors (Vilagrosa et al. 2003; Castro-Díez & Navarro 2007; Peguero-Pina et al. 2009). Such a value of Ψmd was far from the onset of cavitation events (c. −5 MPa) and also far from the Ψ value inducing the 50% stem conductivity loss (PLC50) in this species (Vilagrosa et al. 2003). Therefore, restriction of evaporation by individual stomatal encryption in Q. coccifera appeared to be crucial for maintaining the leaf water potential above values that would cause xylem cavitation and loss of hydraulic conductance in the stem and, as a consequence, to prevent irreversible hydraulic failure (Oren et al. 1999). The safety margin in Q. coccifera– in terms of a possible loss of conductivity due to drought-induced stem cavitation – was very high even for the Mediterranean case, assuming measured values of E and values of E derived from the model with cuticular cup. Stomatal aperture during midday might not be conditioned by hydraulic factors when soil water availability was not a limiting factor as was also indicated by constant gs over a day (see later). However, if we consider the evaporative water loss that would occur with open stomata in the absence of encryption under Mediterranean conditions (c. five times higher, Fig. 5a), then stomatal closure under these conditions should become mandatory for the plant to avoid a drop in water potential to values close to the stem xylem cavitation threshold (Vilagrosa et al. 2003).

This begs, however, the question why this regulation could not simply be achieved by a decrease in stomatal aperture, without any encryption (Lambers, Chapin & Pons 2008). What is the benefit of the fixed resistance in the shape of a cuticular cup with a narrow vent? In the lack of stomatal encryption, much higher CO2 influx rates were possible that could be realized under favourable conditions (Fig. 5). The definitive regulation of E could then be achieved by a more dynamic control of stomatal aperture. At this point, an alternative explanation for the functional benefit of the cuticular cups under Mediterranean conditions is worth considering.

Stomata are sensitive to low RH values. This is explained by direct water loss of guard cells across their external wall and cuticle (Appleby & Davies 1983; Shope, Peak & Mott 2008), which induces a decline in guard cell turgor (Eamus et al. 2008) in relation to epidermal cells (Buckley 2005). The mechanisms underlying stomatal response to low air humidity remain, however, a matter of controversy (Kaiser & Legner 2007), but there is ample evidence that two phases can be distinguished during a sudden rise of atmospheric vapour pressure deficit (VPD), that is: a transient, ‘wrong-way’ response, inducing a hydropassive opening followed by a steady-state, ‘right-way’ response inducing a hydroactive stomatal closure (Buckley 2005). The hydroactive stomatal closure controlling the response to changes in VPD has been attributed to a metabolic response within guard cells after a change in the water status of the ‘epidermis–guard cells’ complex (Franks, Cowan & Farquhar 1997; Buckley 2005; Buckley, Sack & Gilbert 2011). Thereby, the passive regulation of water losses – even at high VPD values– achieved by the wax cup might constitute a more effective mechanism in terms of metabolic cost.

As shown by the simulations, stomatal encryption also had considerable influence on leaf internal air humidity (Fig. 6), particularly if an internal cuticle was present. The arrangement in the models equipped with an internal cuticle represented the extreme case of a cuticular lining reaching deeply into the mesophyll. However, there was evidence that such a structure can extend quite below the internal stomatal pore (Boyer 1985; Wullschleger & Oosterhuis 1989). The exact locations of the evaporative sites within a leaf are still unknown, but it is conceivable that they may reach quite far into the mesophyll (Boyer 1985). For Vicia faba, Kaiser (2009) found an additional resistance to water vapour diffusion apparently seated in the mesophyll which could be attributed to an internal cuticle. This additional resistance decreased leaf conductance to about 25%. It is quite difficult to identify internal cuticles reliably because cuticular layers as thin as 1 µm are sufficient to restrict evaporation (Kerstiens 1996).

Without an internal cuticle, evaporation will preferentially occur close to the stomatal pore (Meidner 1975; Tyree & Yianoulis 1980; Roth-Nebelsick 2007). For Q. coccifera, this would mean a rapid water loss from guard cells and adjacent epidermal cells under Mediterranean conditions with stomata open. It is therefore to be expected that an internal cuticle is present to some extent. But even then, a quite considerable load of evaporative water loss may be caused by high VPD for the cells close to the pore. The model simulations indicated that humidity below the stomatal pore may reach levels that were lower than usually assumed for the leaf internal air space, particularly for the situations without a cuticular cup Thereby, the need for an active regulation of water loss at Mediterranean sites (Table 1, Fig. 1) should imply an early regulation of stomatal aperture in the morning, that also decreases the net carbon gain during most of the late spring and summer (Tenhunen et al. 1985). Contrary to this expectation, however, absence of any midday stomatal closure was observed by Vilagrosa (2002) and Peguero-Pina et al. (2008, 2009) in Q. coccifera, which showed low variation in gs, net CO2 assimilation and WUE throughout the day under Mediterranean summer conditions.

There are numerous reports that disturbances in VPD lead to ‘wrong-way’ responses due to sudden turgor loss of epidermal cells close to the guard cells (Buckley 2005). Furthermore, also stomatal oscillations have been observed under these conditions and are strongly promoted by high VPD (Jarvis et al. 1999; Kaiser & Kappen 2001; Steppe et al. 2006). Patchy stomatal aperture may also be at least partly related to this phenomenon. Transient ‘wrong-way’ response and stomatal oscillations are contrary to an efficient stomatal regulation and indicate that the regulation system can be ‘off balance’ and overstrained by high VPD. It is possible that a cuticular cup serves as an upstream ‘humidity buffer’ that is able to damp strong VPD jumps and prevents the regulation system from overshooting or other non-efficient stomatal responses.

The alternative performance under low RH values – the reduction in maximum gs by the cuticular cup – may provide Q. coccifera with the possibility to realize a more stable and robust stomatal regulation and therefore to obtain longer effective time period for CO2 assimilation both on a daily and a yearly basis. The environmental conditions in Mediterranean-type climates where this species thrives limit plant growth during a long winter period due to low temperatures (Corcuera et al. 2005) and during summer when soil water availability is severely restricted and the VPD is extremely high (Vilagrosa et al. 2010). It should be noted that the time course of soil drying in Mediterranean areas is quite independent of the rising of atmospheric VPD, due to the response of soil water content to precipitation events (Errea et al. 2001; Ramos 2006). A strategy based on a conservative, more inert stomatal functioning would have a double advantage under this scenario: (1) to increase the span of effective carbon gain during the vegetative period and (2) to avoid early soil water consumption, which should induce a chronic decrease in gs and net CO2 assimilation rate (Vilagrosa et al. 2003, 2010). Goulden (1996) found that gas exchange in Quercus agrifolia responded to shallow soil water availability, which increased rapidly after a rain event, as also reported by Quevedo & Francés (2008) for Q. coccifera. The ability for responding to soil water changes more than to VPD might be considered a challenge in these evergreen plants, due to the apparent independence between these two parameters. The application of the model showed that WUE, which is a parameter of outstanding importance in plant performance under Mediterranean-type climates (Medrano et al. 2007; Medrano, Flexas & Galmés 2009), was not negatively affected by the presence of individually encrypted stomata in Q. coccifera (Fig. 5). However, the absolute carbon gain should be higher in the absence of such extra resistance in stomata under a less stressful environment, such as that found in the temperate site.

Individually encrypted stomata in Q. coccifera were generated by the accumulation of epicuticular waxes in the protruding surface of guard cells, which drastically reduced the pore dimensions (Figs 3 & 4). In fact, stomata of Q. coccifera leaves cannot be considered a typical example of sunken stoma as described by Metcalfe & Chalk (1979), Jordan et al. (2008) or as considered by Roth-Nebelsick (2007). In our case, the existence of cuticular ridges overarching the stomatal pore created itself a slight individual encryption that allowed the extension of the wax cup, as described by Molinas (1991) for Q. suber. We therefore suggest that the differences found between both experimental sites in the degree of encryption of Q. coccifera stomata were caused by the different amount of wax deposition in the leaves. Different stress factors such as air RH, temperature or irradiation have been reported to affect epicuticular wax morphology, composition and quantity (Shepherd & Wynne Griffiths 2006). We may suggest that air humidity, temperature and irradiations were the principal factors that may trigger the increased production of waxes and the formation of individually encrypted stomata in Q. coccifera plants grown in more arid environments. However, further investigations will be required to assess the major environmental factors leading to the formation of individual wax crypts in Q. coccifera and to analyse epicuticular waxes qualitatively and quantitatively in relation to differential growing conditions. The environmental conditions under the Mediterranean case, and specially the low air humidity (Koch et al. 2006; Shepherd & Wynne Griffiths 2006), may favour a higher wax production than in temperate conditions.

In conclusion, with the potential to form individually encrypted stomata, Q. coccifera may improve its physiological performance by means of a plastic response in gs that allows it to cope with high VPD in the most xeric areas of its geographical range. The formation of stomatal wax roof implied a reduction in the effective pore dimensions while the stomatal density (data not shown) and the stomatal size once the wax had been removed remained constant irrespective of the growth environmental conditions (Fig. 4). Villar-Salvador et al. (1997) found that Q. coccifera responded to variation in rainfall along an Atlantic-Mediterranean climatic gradient by a reduction in hydraulic conductivity, which is interpreted as a mechanism that could regulate the soil water consumption. This apparent plasticity in the water transport properties seems to be consistent with our results, where the development of an ‘extra’ mechanism affecting gs may also be understood as a strategy to allow for optimally using the effective vegetative period of Q. coccifera under Mediterranean-type climates. Therefore, we suggest that the individual epicuticular wax encryption of stomata in Q. coccifera allows this species to open stomata even at the high VPD values registered during the Mediterranean summer.

ACKNOWLEDGMENTS

This study was partially supported by the projects SUM2008-00004-C03-03 and AGL2010-21153-C02-02 (Ministerio de Economía y Competitividad). Financial support from Gobierno de Aragón (A54 research group) is also acknowledged. The work of José Javier Peguero-Pina is supported by a ‘Juan de la Cierva’ post-doctoral contract (Ministerio de Economía y Competitividad, Spain). Victoria Fernández is supported by a ‘Ramón y Cajal’ contract (MICINN, Spain), co-financed by the European Social Fund.

Ancillary