Effects of temperature on the cuticular transpiration barrier of two desert plants with water-spender and water-saver strategies

The efficacy of the cuticular transpiration barrier and its resistance to elevated temperatures are significantly higher in a typical water-saver than in a water-spender plant growing in hot desert.

Abstract

Water-saver and water-spender strategies are successful adaptations allowing plants to cope with the limitations of hot desert habitats. We investigated whether the efficacy of the cuticular transpiration barrier and its susceptibility to high temperatures are ecophysiological traits differentially developed in the water-spender Citrullus colocynthis and the water-saver Phoenix dactylifera. Minimum leaf conductance (g_min) at 25 °C was six times lower in P. dactylifera (1.1×10^–5 m s^–1) than in C. colocynthis (6.9×10^–5 m s^–1). Additionally, g_min in the range 25–50 °C did not change in P. dactylifera but increased by a factor of 3.2 in C. colocynthis. Arrhenius formalism applied to the C. colocynthis g_min led to a biphasic graph with a steep increase at temperatures ≥35 °C, whereas for P. dactylifera the graph was linear over all temperatures. Leaf cuticular wax coverage amounted to 4.2±0.4 µg cm^–2 for C. colocynthis and 29.4±4.2 µg cm^–2 for P. dactylifera. In both species, waxes were mainly composed of very-long-chain aliphatics. Midpoints of the wax melting ranges of P. dactylifera and C. colocynthis were 80 °C and 73 °C, respectively. We conclude that in P. dactylifera a particular wax and cutin chemistry prevents the rise of g_min at elevated temperatures.

Introduction

Desert plants have evolved a wide range of adaptations that allow them to cope with water shortage and intense solar radiation. There are numerous strategies for dealing with drought stress, and plants can be classified along an ‘adaptation continuum’ (of resistance) from drought tolerance to drought avoidance (Smith et al., 1997). Baquedano and Castillo (2006) showed that drought-avoiding plants utilize two contrasting strategies in response to the scarcity of water. (i) water-saver plants close their stomata early after the onset of drought and thereby prevent damage before the leaf water potential changes substantially; (ii) water-spender plants, by contrast, keep their stomata open for prolonged periods of drought and thus decrease the leaf water potential, allowing them to extract water from the soil to compensate for the high water loss by stomatal transpiration (Guehl and Aussenac, 1987; Lo Gullo and Salleo, 1988).

The stomata and cuticle are the primary components responsible for maintaining a favourable plant water status. At maximum stomatal closure, the water permeability of the cuticle determines the minimum and inevitable water loss and, therefore, is one of the principal features ensuring the successful establishment and survival of plants. The cuticle, a cutin matrix with cuticular waxes embedded within it (intracuticular waxes) and deposited on its surface (epicuticular waxes), covers all primary aerial parts of terrestrial higher plants. The cutin matrix primarily consists of cross-linked C₁₆ and C₁₈ hydroxy and hydroxy epoxy alkanoic acids, whereas the characteristic cuticular waxes comprise very-long-chain (VLC) aliphatic and cyclic components (Jetter et al., 2006). Published values of leaf cuticular permeability reviewed by Schuster et al. (2017) vary among plant species from 2.6×10^–7 m s^–1 (Zamioculcas zamiifolia) to 2.5×10^–4 m s^–1 (Helianthus anomalus). For comparison, representative values of stomatal conductances of mesophytes lie between 4×10^–3 m s^–1 and 20×10^–3 m s^–1, while for xerophytes and trees they range from 1×10^–3 m s^–1 to 4×10^–3 m s^–1 (Nobel, 2009) and, consequently, are one to four orders of magnitude higher than cuticular water permeabilities. The extraction of the cuticular waxes increases the cuticular water permeability by up to several orders of magnitude (Schönherr, 1976; Schönherr and Lendzian, 1981), showing that the waxes make up the main barrier to transpiration. However, so far, there is no evidence that the cuticle thickness or the amount of cuticular waxes per unit area reduces the cuticular water permeability (Schreiber and Riederer, 1996; Riederer and Schreiber, 2001).

In hot deserts, drought often co-occurs with high temperatures. Over the course of the day, relative air humidity can reach low values as the temperature increases, leading to a concomitant rise of the water vapour pressure deficit. Therefore, the efficient control of water loss at high temperatures is of fundamental importance for ensuring the survival, viability, and reproductive fitness of desert plants.

Consequently, the effect of temperature on the cuticular water permeability is a decisive parameter determining the physiological and ecological role the cuticle plays, especially in conditions of high evaporative demand and elevated temperature. The cuticular permeability of all non-desert plants studied so far increased slightly at temperatures from 15 °C to 35 °C, while temperatures above 35 °C caused a drastic increase in permeability (Schreiber, 2001; Riederer, 2006). In contrast to the non-desert plants studied so far, the cuticular water permeability of the desert plant Rhazya stricta increased only slightly between 15 °C and 50 °C, without an abrupt change in slope at higher temperatures (Schuster et al., 2016).

In this study, we investigated the leaf cuticular transpiration barrier and its susceptibility to thermal effects as potential adaptations to aridity and high evaporative demand under elevated temperature in two plant species from the Arabian desert. Citrullus colocynthis (L.) Schrad. (Cucurbitaceae) and Phoenix dactylifera L. (Arecaceae) were chosen as models as both species often grow near each other under identical climatic conditions characterized by frequent periods of intense irradiation, air temperatures of up to 50 °C, and high evaporative demand. C. colocynthis is a broad-leaved summergreen vine, which spreads its trailing stems along the soil surface. Its leaves stand only a few centimetres above the ground (Lange, 1959). Stem growth begins with buds on the taproots of established plants, and after fruit production, the aboveground parts die off (Cunningham et al., 2011). P. dactylifera, by contrast, is a long-lived evergreen woody plant that grows up to 30 m height. The straight leaves are obliquely vertically oriented and form a crown at the end of the trunk (Barrow, 1998).

Desert soil surfaces may heat up to extreme temperatures owing to the intense solar radiation. Consequently, trailing plants and other plants with leaves close to the soil surface can be exposed to intense heat stress. A pioneer in desert plant ecology, Lange (1959), pointed out the contrasting ecophysiological strategies of the two species studied here regarding water relation and thermal tolerance. He characterized C. colocynthis as a typical perennial water-spender plant, which is heat avoidant and possesses a deep root system and short-lived leaves (Si et al., 2009). Even under intense solar radiation, an actively transpiring leaf of C. colocynthis maintained its temperature 10–15 °C below that of the surrounding air and far from the limit of heat tolerance of this species, thereby avoiding death by overheating (Lange, 1959; Althawadi and Grace, 1986). The temperature of a cut leaf separated from the water supply, after a short time, exceeded the air temperature and even reached temperatures higher than 60 °C (Lange, 1959). The author associated this strong cooling effect in C. colocynthis with high transpiration rates and consequently high water consumption. In contrast to C. colocynthis, P. dactylifera is a typical woody water-saver plant (Lange, 1959) with evergreen leaflets that are able to thrive in hot and dry conditions as long as there is sufficient moisture around its roots (Barrow, 1998). Leaflets of P. dactylifera exposed to high solar irradiation reached temperatures above that of the air temperature. Despite this leaflet overheating, their temperature did not exceed the limit of leaflet heat tolerance (59 °C), ensuring their survival (Lange, 1959). Surprisingly, it has not yet been investigated whether the efficacy of the cuticular transpiration barrier against water loss and its susceptibility to thermal damage are also traits particular to the water-saver and water-spender strategies.

We hypothesize that water-saver, as compared with water-spender, plants in hot arid environments should have (i) a cuticular transpiration barrier with a lower water permeability and consequently a higher efficacy in reducing transpiration when stomata are closed, and (ii) a cuticular transpiration barrier that is less susceptible to adverse thermal effects. To test these hypotheses, we re-investigated two of the plant species Lange (1959) had used in his pioneering work. The minimum leaf water conductance (g_min), as a proxy for cuticular permeability, of C. colocynthis and P. dactylifera was measured within the range of ecologically relevant temperatures (Almazroui et al., 2014). In the past, it has been shown that g_min determined by the method used in the present study is a good proxy for cuticular permeability with only very small or no contributions of residual transpiration through potentially leaky stomata (Burghardt and Riederer, 2003; Burghardt et al., 2008; Schuster et al., 2017).

Additionally, we documented the fine structure of the leaf surface using scanning electron microscopy. The qualitative and quantitative compositions of the cuticular waxes and the cutin matrix of both species were evaluated chromatographically. The melting behaviour of the cuticular waxes of both species was investigated by Fourier transform infrared spectroscopy (FTIR). Combining the data on the functional and chemical properties of the cuticle of both plant species, we propose a novel hypothesis relating the chemical composition and phase behaviour of the cuticular waxes of both species to the thermal resistance of their cuticular transpiration barrier.

Materials and methods

Plant material and growing conditions

Seeds of Citrullus colocynthis (L.) Schrad. were harvested from a natural population growing near Riyadh, Saudi Arabia. Phoenix dactylifera L. plants approximately 5 years old and 2 m high were purchased from a commercial supplier (‘Der Palmenmann’, Bottrop, Germany). Plants were cultivated in a greenhouse at 24 °C/18 °C and 50±5% relative humidity day/night under a 14 h light regime at 250 µmol m^–2 s^–1 white light (Philips Master Agro, 400 W). Healthy, fully expanded current-year leaves (laminae) of C. colocynthis and leaflets (pinnae) from 1-year-old leaves of P. dactylifera were harvested and used for the experiments.

Morphological leaf traits

Before measurement, leaves were rehydrated in a glass desiccator overnight by immersing their petioles in water. The saturation leaf weights were determined using an analytical balance (MC-1 AC210S, Sartorius; precision 0.1 mg). Dry weight was measured after oven-drying the leaves at 90 °C for 24 h. The actual fresh weights during leaf drying experiments were used to calculate the relative water deficit (RWD) according to:

$$RWD=1-\frac{actualfreshweight-dryweight}{saturatedweight-dryweight}$$ (1)

For determination of leaf area, leaves were scanned at high resolution (600 dpi) using a flatbed scanner. The projected leaf area was measured from the scanned leaf image using Adobe Photoshop image analysis software. Leaf mass per area (LMA) was calculated by dividing the dry weight by the leaf area. The leaf water content was calculated by subtracting the dry weight from the saturation leaf weight and subsequently dividing the value obtained by the saturation weight.

Scanning electron microscopy

The structure of the leaf cuticular surface was characterized using scanning electron microscopy. Small air-dried pieces (2–3 mm) of fully developed leaves of C. colocynthis and leaflets of P. dactylifera were mounted on aluminium stubs using double-sided adhesive tape and sputter-coated with 10–15 nm gold-palladium (150 s, 25 mA, partial argon pressure 0.05 mbar, SCD005 sputter coater, Bal-Tec). The samples were investigated with a field-emission scanning electron microscope (JEOL JSM-7500F) using a 5 kV acceleration voltage and a 10 mm working distance. Pictures were taken from both adaxial and abaxial leaf surfaces.

Minimum leaf water conductance

Leaf minimum water conductance (g_min) represents the lowest water loss a leaf reaches when its stomata are maximally closed due to desiccation stress (Körner, 1995). g_min was determined gravimetrically from the consecutive weight loss of desiccating leaves in an incubator in the dark and with controlled very low air humidity (Burghardt and Riederer, 2003). High-melting-point (68 °C) paraffin wax (Fluka) was used to seal the wounds of cut petioles of water-saturated leaves and leaflets. Subsequently, the sealed leaves were placed in an incubator (IPP 110, Memmert) with controlled temperature. For evaluating the effect of temperature, g_min was measured at 25 °C, 30 °C, 35 °C, 40 °C, 45 °C, and 50 °C. The air temperature and humidity in the incubator were monitored with a digital thermo-hygrometer (Testoterm 6010, Testo). Silica gel (Applichem) was used to control the air humidity in the incubator. The actual weight of desiccating leaves was determined as a function of desiccation time using a high-precision balance (MC-1 AC210S, Sartorius). The transpiration rate (J) was calculated from the change in fresh weight (ΔFW) with time (t) divided by the total projected leaf area (sum of the adaxial and abaxial projected areas, A):

$$J=\frac{\Delta FW}{\Delta t\times A}$$ (2)

Cuticular water conductance was calculated from J divided by the driving force for water loss from the outer epidermal cell wall to the surrounding air. The driving force for vapour-based conductance is defined by the difference between the saturation concentrations of water vapour at the temperatures of the leaf (C_{wv sat leaf}) and the surrounding atmosphere (C_{wv sat air}) multiplied by the water activities in the epidermal apoplast (α_apo) and the air (α_lair):

$$g=\frac{J}{{\alpha }_{apo}\times C_{wvsatleaf}-{\alpha }_{air}\times C_{wvsatair}}$$ (3)

The water activity of the air (α_air) over silica gel is nearly zero (Slavík, 1974) and the water activity in the apoplast adjacent to the inner side of the cuticle (α_apo) can be assumed to be close to 1. Thus, the active driving force for cuticular transpiration in the setup used here is the saturation concentration of water vapour at the actual leaf temperature (C_{wv sat leaf}; Nobel, 2009). Leaf temperatures were measured at 30 min intervals using an infrared (IR) laser thermometer (Harbor Freight Tools, one-point measurements), and the corresponding water vapour saturation concentrations at leaf temperature were derived from tabulated values (Nobel, 2009).

The leaf conductance was plotted against the corresponding RWD. The conductance was high at low RWDs and declined with progressive leaf desiccation until reaching a plateau where the conductance did not decline further with RWD. This constant low conductance corresponds to g_min and characterizes the leaf water loss at maximum stomatal closure caused by cuticular transpiration (Burghardt and Riederer, 2006; Schuster et al., 2016, 2017).

Chemical composition of cuticular waxes

Cuticular waxes of both species were extracted from leaves of the same age as those used in the determination of morphological leaf traits and cuticular transpiration. The extraction conditions appropriate for each species had been established by prior tests. C. colocynthis leaves were dipped (avoiding the wounds of cut petioles) twice into 10 ml chloroform (≥99.8%, Roth) for 1 min at room temperature. Leaflets of P. dactylifera were immersed (avoiding wounds) twice in 25 ml chloroform (≥99.8%, Roth) and kept for 5 min in an ultrasonic bath at room temperature. N-tetracosane (C₂₄; ≥99.5%, Sigma-Aldrich) was added as an internal standard, and the solutions were reduced to dryness under a gentle flow of nitrogen.

Dry wax samples were derivatized for gas chromatography with N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA, Marchery-Nagel) in anhydrous pyridine (≥99.5%, Roth) for 30 min at 70 °C. Quantification of cuticular wax components was performed with a gas chromatograph equipped with a flame ionization detector and an on-column injector (7890A, Agilent Technologies). Separation of compounds was carried out on a fused-silica capillary column (DB1-ms, 30 m length × 0.32 mm internal diameter, 0.1 µm film, Agilent Technologies) with hydrogen as a carrier gas. The temperature programme consisted of injection at 50 °C for 2 min, then raised by 40 °C min^–1 to 200 °C, held at 200 °C for 2 min, and then raised by 3 °C min^–1 to 320 °C, and held at 320 °C for 30 min. Single compounds were quantified against the internal standard. Qualitative analysis was carried out using a gas chromatograph equipped with a mass spectrometric detector (5975 iMSD, Agilent Technologies) following the same gas chromatographic conditions, except that helium was the carrier gas. Identification of all the compound classes for both species was achieved by comparing a query mass spectrum with reference mass spectra in a library (NIST05) via spectrum matching.

Weighted median carbon chain-lengths (MCL) for the different wax samples were calculated to characterize the chain-length distributions of VLC aliphatic compounds. Molar coverages for each component were calculated from the gas chromatographic data and summed up according to carbon chain-lengths. For each chain-length N_i the molar fraction w_i was obtained and used as a weight for further calculating the MCL. For n distinct ordered chain-lengths N₁, N₂, … N_n with weights w₁, w₂, … w_n, the MCL is the chain-length N_k satisfying

$$\sum _{i=1}^{k-1}{w}_i\le \frac{1}{2}and\sum _{i=k+1}^n{w}_i\le \frac{1}{2}$$ (4)

Chemical composition of the cutin matrix

For cutin depolymerization, delipidated leaves of C. colocynthis and leaflets of P. dactylifera were transesterified with boron trifluoride in methanol (Fluka) at 70 °C overnight to release methyl esters of cutin acid monomers. Sodium chloride-saturated aqueous solution (AppliChem), chloroform, and n-dotriacontane (C₃₂; Sigma-Aldrich) as an internal standard were added to all reaction mixtures. From this two-phase system, the de-esterified cutin components were extracted three times with chloroform. The combined organic phases were dried over anhydrous sodium sulfate (AppliChem). All extracts were filtered, and the organic solvent was evaporated under a continuous flow of nitrogen.

Derivatization and subsequent gas chromatographic analysis with mass spectrometric detection were performed according to Leide et al. (2011). Separation of cutin mixtures was carried out at 50 kPa for 60 min, increased by 10 kPa min^–1 to 150 kPa, and at 150 kPa for 30 min using a temperature programme of 50 °C for 1 min, raised by 10 °C min^–1 to 150 °C, held at 150 °C for 2 min, and then raised by 3 °C min^–1 to 320 °C and held at 320 °C for 30 min. The quantitative composition of the mixtures was studied using capillary gas chromatography and flame ionization detection under the same gas chromatographic conditions.

Melting of cuticular waxes

The melting of the leaf cuticular waxes of C. colocynthis and P. dactylifera was analysed by FTIR according to the method developed by Merk et al. (1997). A Bruker Tensor 27 FTIR spectrometer with a BIO-ATR II® unit (Bruker, Ettlingen, Germany) was used for recording the spectra in horizontal attenuated total reflection mode (h-ATR). The BIO-ATR II® unit consists of a silicon crystal in a stainless steel holder. The wax solution in chloroform (dry weight of a single sample ~200 µg) was deposited on the crystal at 35 °C, and the solvent was quantitatively evaporated. Subsequently, the crystal was cooled to 24 °C for IR measurement. IR spectra were recorded in the wavenumber range from 4000 to 670 cm^–1 at temperatures from 24 °C to 94 °C. The temperature of the holder was controlled by connecting it to a thermostat (Thermo Scientific Haake DC30-K20, Karlsruhe, Germany). The BIO-ATR II® unit was purged continuously with dry CO₂-free air (K-MT-LAB 3, Parker Hannifin, Kaarst, Germany). The initial temperature was set to 24 °C and was increased to 48 °C in intervals of 4 °C . From 48 °C to 94 °C, IR spectra were recorded at temperature intervals of 1 °C . The resolution was set to 2 cm^–1 with an acquisition time of 120 scans. OPUS 7 software (Bruker, Ettlingen, Germany) was used to control the spectrometer and the thermostat, as well as to analyse the spectra. The maxima of the CH₂ asymmetric stretching bands were determined in the range from 2195 to 2922 cm^–1 and used for determination of the wax melting ranges (Merk et al., 1997).

Statistical analysis

Statistical analyses were performed using SPSS Statistics software version 23.0 (IBM Corporation). Normal distribution of data was tested using the Shapiro–Wilk test. Subsequently, a t-test was used to compare the minimum conductances of both plant species at 25 °C. Spearman rank correlation was performed to investigate the impact of temperature increase on g_min. Quantitative differences between the cuticular wax coverage of the species were investigated using t-tests or Mann–Whitney U-tests depending on the dataset distribution.

Results

Morphological leaf traits

The morphological traits of C. colocynthis leaves and P. dactylifera leaflets are shown in Table 1.

Table 1.

Morphological traits of Citrullus colocynthis leaves and Phoenix dactylifera leaflets

Trait	Citrullus colocynthis	Phoenix dactylifera
Total projected leaf area a (×10–2 m2)	0.45±0.13	0.74±0.25
Saturation leaf weight (g)	0.82±0.34	1.27±0.42
Dry weight (g)	0.14±0.06	0.59±0.21
Leaf water content (g g–1)	0.83±0.02	0.54±0.03
Leaf mass per area (g m–2)	60.11±15.95	161.83±22.41

^a Sum of projected adaxial and abaxial leaf surfaces.

Each value represents the mean ±SD (n≥61).

Leaf surface properties

Leaves of C. colocynthis and leaflets of P. dactylifera were analysed by scanning electron microscopy to examine the morphology of the leaf surface. The presence of stomata, trichomes, and epicuticular wax structures were the primary features investigated. Both plant species had stomata on both the adaxial and abaxial leaf surfaces (Figs 1 and 2). Stomata of C. colocynthis were distributed without any distinct pattern over the whole leaf surface. The cuticular surface was smooth without any explicit crystalline epicuticular deposits (Fig. 1B, D) and the density of trichomes varied among the adaxial (Fig. 1A) and abaxial (Fig. 1C) surfaces. In contrast, the stomata of P. dactylifera were arranged in rows along both the adaxial (Fig. 2A) and abaxial (Fig. 2C) leaflet surfaces. Chimney-like cuticular wax surrounded the stomata (Fig. 2B, D), leading to a locally increased boundary layer resistance affecting only stomatal transpiration. Trichomes could not be detected.

Fig. 1.

Scanning electron micrographs of the adaxial (A, B) and abaxial (C, D) surface of Citrullus colocynthis leaves.

Fig. 2.

Scanning electron micrographs of the adaxial (A, B) and abaxial (C, D) surface of Phoenix dactylifera leaflets.

Leaf drying curves and minimum leaf water conductance

The actual leaf temperature at a given temperature of the surrounding air was measured to calculate the driving force for cuticular transpiration. Leaf temperatures of both C. colocynthis and P. dactylifera were slightly lower than the corresponding air temperatures at 25 °C and 30 °C. With rising air temperatures, the leaf–air temperature differences (ΔT_leaf–air) of C. colocynthis increased continuously until reaching a value of approximately –4 °C at an air temperature of 50 °C. The ΔT_leaf–air of P. dactylifera remained at approximately –1 °C at all temperatures above 30 °C (see Supplementary Table S1 at JXB online). Minimum cuticular transpiration rates (J_min) increased with air temperature. J_min rose from 1.3±0.4×10^–3 g m^–2 s^–1 at 25 °C to 13.2±2.8×10^–3 g m^–2 s^–1 at 50 °C in C. colocynthis, and from 0.2±0.0×10^–3 g m^–2 s^–1 to 1.0±0.2×10^–3 g m^–2 s^–1 in P. dactylifera (Supplementary Table S1).

The kinetics of water loss from detached leaves of C. colocynthis and leaflets of P. dactylifera exposed to dry air were measured over time. Conductances were calculated from the water loss rates for each pair of consecutive data points along the leaf drying curve. For both plant species, the conductance was high in the initial phase of leaf drying. During leaf dehydration, the conductance declined to reach a plateau yielding the minimum conductance (g_min) at maximum stomatal closure (Fig. 3). Under such conditions, water loss became linear over time (Supplementary Fig. S1). Leaf drying curves at 50 °C showed essentially the same pattern as those at 25 °C for P. dactylifera, but not for C. colocynthis. The conductance of C. colocynthis at 50 °C, after the initial phase, progressively decreased until a late stomatal closure, when the RWD reached a value of ~0.6 (Table 2).

Fig. 3.

Leaf conductance to water vapour as a function of the relative water deficit (RWD). The drying curves shown are representative for a leaf of Citrullus colocynthis (A) and a leaflet of Phoenix dactylifera (B) at 25 °C. The RWD at maximum stomatal closure is indicated by the transition between the declining phase and the plateau phase of leaf conductance. After maximum stomatal closure, leaf conductance remains constant at a minimum, defined as the minimum leaf conductance (g_min).

Table 2.

Minimum conductance (g_min) and relative water deficit at maximum stomatal closure (RWD_SC) of Citrullus colocynthis leaves and Phoenix dactylifera leaflets obtained from leaf drying curves as functions of air temperature (T_air)

T air (°C)	g min×105 (m s–1)		RWDSC
	C. colocynthis	P. dactylifera	C. colocynthis	P. dactylifera
25	6.94±2.00	1.11±0.24	0.14±0.05	0.014±0.004
30	7.83±1.23	0.92±0.16	0.15±0.06	0.010±0.006
35	7.74±1.46	1.00±0.12	0.16±0.03	0.014±0.001
40	8.79±2.41	1.19±0.47	0.20±0.08	0.011±0.006
45	13.70±5.33	1.37±0.32	0.30±0.08	0.015±0.003
50	21.98±5.19	1.30±0.26	0.60±0.04	0.016±0.005

Each value represents the mean ±SD (n≥8).

The transition point representing the RWD at the onset of maximum stomatal closure (RWD_SC) was calculated from the intersection of the steep declining phase and the plateau phase of the leaf drying curves. The two plant species showed different responses in terms of stomatal closure. In the temperature range from 25 °C to 40 °C, C. colocynthis maximally closed the stomata at RWDs between 0.1 and 0.2. At higher temperatures, maximum stomatal closure shifted to higher RWDs of up to 0.6 at an air temperature of 50 °C. In comparison, P. dactylifera maximally closed the stomata at RWD values 10 times lower than in C. colocynthis, regardless of the air temperature (Table 2).

The g_min of C. colocynthis at 25 °C, 6.9±2.0×10^–5 m s^–1, was more than six times higher [t (14.6)=–10.8, P<0.001] compared with the g_min of P. dactylifera at the same temperature (1.1±0.2×10^–5 m s^–1). The g_min of C. colocynthis increased with temperature by a factor of 3.2 between 25 °C and 50 °C, from 6.9±2.0×10^–5 m s^–1 to 22.0±5.2×10^–5 m s^–1. In comparison, the g_min of P. dactylifera remained constant over the whole temperature range. The C. colocynthis g_min did not change significantly between 25 °C and 35 °C but then steeply increased at temperatures above 35 °C (Fig. 4A).

Fig. 4.

Minimum conductance (g_min; A) and Arrhenius plots of the g_min (B) of Citrullus colocynthis leaves and Phoenix dactylifera leaflets obtained from drying curves as a function of air temperature (n≥8). Actual leaf temperatures deviated from the corresponding air temperature by a maximum of 4 °C. For calculating g_min the driving force derived from the actual leaf temperature was used. The Arrhenius formalism applied to the g_min of C. colocynthis led to a biphasic graph with no statistically significant (P=0.383) change of conductance from 25 °C to 35 °C and a sharp significant increase (P=0.017) at higher temperatures (R²=0.99, ln g_min=20.3–9.3×10³ 1/T). In P. dactylifera, there is no statistically significant change (P=0.087) of the conductance over the whole range of temperatures, and no transition.

Chemical composition of cuticular waxes

The cuticular waxes of C. colocynthis leaves and P. dactylifera leaflets were analysed to investigate differences between the plant species and detect potential relationships between cuticular transpiration and the cuticular wax quantity and composition. The leaf cuticular wax coverage was 4.2±0.4 µg cm^–2 for C. colocynthis and 29.4±4.2 µg cm^–2 for P. dactylifera. The cuticular waxes of both plant species were mainly composed of VLC aliphatic compounds (Fig. 5). VLC aliphatic compounds with carbon chain-lengths ranging from C₂₀ to C₄₆ amounted to 90% of total cuticular waxes in C. colocynthis (Fig. 6A), and the main constituents were octacosan-1-ol (C_28; 13% of total cuticular waxes), triacontan-1-ol (C_30; 26%), and dotriacontan-1-ol (C₃₂; 13%). In P. dactylifera, the VLC aliphatic compounds with carbon chain-lengths ranging from C₂₀ to C₆₂ amounted to 88% of total cuticular waxes (Fig. 6B), with dotriacontanoic acid (C₃₂; 15% of total cuticular waxes) and alkyl esters with carbon chain-length C₄₈ (10%), C₅₆ (6%), and C₅₈ (6%) as the dominant constituents (Supplementary Table S2). Only a minor fraction of cuticular waxes contained cyclic compounds in both plant species (<1% of total cuticular waxes in C. colocynthis and 4% in P. dactylifera).

Fig. 5.

Cuticular wax coverage of Citrullus colocynthis leaves and Phoenix dactylifera leaflets arranged by compound class. Each value represents the mean ±SD (n≥3). Asterisks indicate significant differences in cuticular wax coverage between the species (P<0.05).

Fig. 6.

Chain-length distributions of the very-long-chain (VLC) aliphatic compounds of cuticular waxes from leaves of Citrullus colocynthis (A) and leaflets of Phoenix dactylifera (B). Bars represent the mole-based contribution of a single chain-length to the total VLC fraction. Asterisks denote the weighted median carbon chain-lengths (MCL; the 50% weighted percentile of the chain-lengths) of the VLC aliphatic fraction with chain-lengths <40 (black asterisk) and ≥40 (grey asterisk) carbon atoms.

Chemical composition of the cutin matrix

The coverage of the cutin monomers of C. colocynthis leaves was 7.8±1.3 µg cm^–2 and that of P. dactylifera leaflets 57.4±4.2 µg cm^–2. For both plant species, the cutin matrix was composed of ~88% aliphatic and ~12% aromatic cutin monomers. For the leaf cutin of C. colocynthis, the carbon chain-lengths of aliphatic compounds ranged from C₁₆ to C₃₂, but for P. dactylifera leaflets they ranged up to C₃₄. The ratio of the cutin acids with a carbon chain-length of C₁₆ and C₁₈ was 6:1 for C. colocynthis leaves and 1:11 for P. dactylifera leaflets. The predominant cutin acid was 9/10,16-dihydroxy hexadecanoic acid (51% of total cutin monomers) in C. colocynthis leaves and 18-hydroxy 9,10-epoxy octadecanoic acid (60%) in P. dactylifera leaflets. The 9,10-epoxy 18-hydroxy octadecanoic acid was absent from the C. colocynthis leaf cutin. The aromatic cutin constituent was mainly trans-4-hydroxycinnamic acid, representing 7% of total cutin monomers in both C. colocynthis and P. dactylifera (Supplementary Table S3).

Melting behaviour of cuticular waxes

The melting ranges of the leaf cuticular waxes of C. colocynthis and P. dactylifera were determined by FTIR. The wavenumbers of the maxima of the CH₂ asymmetric stretching band (ν_(as)) were plotted against temperature. In long-chain alkyl crystals the wavenumbers of the maxima of ν_(as) increase with temperature as a consequence of increasing disorder of the C-C chain (Snyder et al., 1982). This disorder reaches its maximum when the long-chain alkyl sample is entirely liquid. Since waxes are multicomponent materials, they do not display a sharp melting point, but more or less extended melting ranges (Merk et al., 1997). C. colocynthis wax melted in the temperature range from 60 °C to 80 °C, while the wax of P. dactylifera started to melt at ~70 °C and wholly melted at 88 °C. The midpoints of the wax melting ranges were identified as the temperatures where 50% of the total frequency shift was reached. The midpoints of the melting ranges of P. dactylifera and C. colocynthis waxes were 80 °C and 73 °C, respectively (Fig. 7).

Fig. 7.

Melting curves of cuticular waxes of Citrullus colocynthis leaves (A) and Phoenix dactylifera leaflets (B) derived from the temperature-dependent shift of the absorption maxima of the asymmetric CH₂ stretching modes. Dashed lines represent the midpoints of wax melting, which correspond to the 50% frequency change.

Discussion

Water-saver and water-spender plants growing in hot arid environments are characterized by common syndromes of traits on the morphological, physiological, biochemical, cellular, and molecular levels allowing them to cope with the unique challenges of such habitats. The intention of our study of C. colocynthis and P. dactylifera was to investigate whether the efficacy of the cuticular transpiration barrier against water loss and its susceptibility to thermal damage are as yet unidentified traits pertaining to the water-saver and water-spender strategies.

Morphological leaf traits are very often directly related to life strategies. The LMA reflects the trade-off between carbon gain and longevity (Díaz et al., 2016), while the LWC roughly indicates leaf density (Garnier and Laurent, 1994). The evergreen leaflets of P. dactylifera have an LMA three times higher than the leaves of C. colocynthis, indicating the much higher investment into the long-lived fronds of P. dactylifera. The comparably high costs for producing long-lived evergreen leaves are the price to be paid for a higher survival rate in the face of abiotic and biotic hazards (Coley et al., 1985; Reich et al., 1997; Wright et al., 2004). Hence, the high LMA and low LWC confer higher stress resistance in P. dactylifera compared with C. colocynthis.

In the leaf drying experiments, the minimum conductance as a proxy for cuticular transpiration of P. dactylifera and C. colocynthis was measured. For C. colocynthis, a g_min of 6.9±2.0×10^–5 m s^–1 at 25 °C was found, which lies at the lower end of the interquartile range of cuticular water permeabilities or minimum conductances measured so far for non-evergreen forbs from temperate climates (Fig. 8). The g_min of C. colocynthis is very close to the g_min of the slightly xerophytic species Hippocrepis comosa, Sanguisorba minor, and Salvia pratensis from temperate climates (Schuster, 2016). The g_min of P. dactylifera leaflets is about six times lower but is still at the lower end of the interquartile range formerly reported for evergreen woody plants from various climates (Fig. 8). Thus, the cuticles of the two hot desert species studied, while being low, do not provide, at mild temperatures, particularly efficient protection against non-stomatal leaf water loss as might be intuitively expected for plants living in such habitats.

Fig. 8.

Comparison of the minimum conductances at 25 °C measured in leaves of P. dactylifera and C. colocynthis with minimum conductances or cuticular permeabilities from the corresponding ecological groups of evergreen woody plants (n=54) and non-evergreen forbs (n=18). Data for these two latter groups were taken from Table S1 in Schuster et al. (2017).

The RWD_SC of C. colocynthis was 0.2 at 25 °C and drastically increased above 40 °C to ~0.6 at 50 °C, indicating that at high temperatures, leaf desiccation proceeds to a greater degree before the stomata close. This agrees well with field studies showing that the cooling effect of stomatal transpiration keeps the leaf temperatures below the thermal tolerance of this plant species (46 °C), avoiding death by overheating (Lange, 1959). Allowing the leaf water potential to decrease maintains the water flow from the soil to compensate for the water loss by stomatal transpiration that is necessary for leaf cooling. Thus, C. colocynthis employs a typical strategy of water-spender plants adapted to heat stress (Lange, 1959; Althawadi and Grace, 1986; Beerling et al., 2001; Schulze et al., 2005). In contrast to this behaviour, Lange (1959) did not find a cooling effect with P. dactylifera. The leaflet temperature was up to 11 °C above that of the air, and the transpiration rates were almost unnoticeable with the method used by this author in the field. The RWD_SC in P. dactylifera is 10 times lower than for C. colocynthis and remains low over the whole temperature range measured (Table 2). Hence, P. dactylifera is very sensitive to leaf water loss and closes its stomata early in the course of desiccation (Müller et al. 2017). P. dactylifera closes its stomata at small changes in the RWD at all experienced temperatures, thus preventing leaf damage, and has a low g_min, minimizing water loss. Both features are typical for water-saver plants, irrespective of their phylogeny (Guehl and Aussenac, 1987; Lo Gullo and Salleo, 1988; Baquedano and Castillo, 2006).

Being a kinetic phenomenon, the diffusion velocity of water in the cuticular transpiration barrier rises with temperature, which in all cases studied so far also led to an increase in cuticular permeability (Schreiber, 2002; Riederer, 2006). The temperature dependence of a kinetic parameter can be analysed quantitatively and mechanistically using the Arrhenius formalism by plotting the natural logarithm of the rate constant—in this case, the cuticular permeability—versus the inverse absolute temperature. An abrupt increase or decrease in the slope of the Arrhenius plot at a specific temperature indicates a change in the properties of the surroundings, on a molecular scale, of the diffusing water molecule. The Arrhenius formalism applied to the cuticular permeance of C. colocynthis leaves led to two distinct linear plots for the lower and the higher temperature ranges (Fig. 4B): (i) no significant change of permeance at lower temperatures ranging from 25 °C to 35 °C and (ii) a sharp increase at higher temperatures from 40 °C to 50 °C, with a transition temperature of 37.7 °C. This implies that at higher temperatures the water molecules diffuse through pathways that are not accessible at lower temperatures. The activation energy (E_p) of cuticular permeation of C. colocynthis derived from the slope of the Arrhenius plot at higher temperatures (for details of the calculations see Riederer, 2006) is 76.3±33.4 kJ mol^–1. This activation energy coincides with the range of the respective values reported so far in the literature (Riederer, 2006).

In contrast to C. colocynthis, the Arrhenius plot for the minimum conductance of P. dactylifera is linear and parallel to the x-axis over the whole range of temperatures studied. Permeability does not change from 25 °C to 50 °C, and no activation energy can be estimated. This pronounced resistance to a thermally induced deterioration of the cuticular transpiration barrier is comparable to that recently described for the hot-desert shrub R. stricta (Schuster et al., 2016), suggesting that resistance to substantial thermal effects on cuticular permeability is an adaptive trait of desert water-saver plants from hot arid habitats.

What is the basis for the observed difference in the thermal susceptibility of the cuticular transpiration barrier of the water-saver plant P. dactylifera and the water-spender C. colocynthis? The differences can be assumed to be related to the chemical composition and physical structure of the cuticle of the two species, especially differences in the qualitative and quantitative composition of the main cuticular components—cutin and cuticular waxes—and the aggregates they form. The cutin matrix of both plant species mainly consisted of ester-cross-linked C₁₆ and C₁₈ hydroxy alkanoic acids, predominantly 9/10,16-dihydroxy hexadecanoic acids in C. colocynthis leaves and 18-hydroxy-9,10-epoxy octadecanoic acid in P. dactylifera leaflets. The cutin composition of C. colocynthis leaves represents the C₁₆ type, and of P. dactylifera leaflets the C₁₈ type, of cutin, according to the classification proposed by Holloway (1982). The amount of cutin monomers per unit surface area was seven times lower in C. colocynthis leaves compared with P. dactylifera leaflets. Differences in the amount and composition of cutin have been discussed as the cause of differences in the degree of cross-linking of the polymeric network and, consequently, in the matrix structure and the viscoelastic properties of the cutin (Fich et al., 2016; Khanal and Knoche, 2017). The mid-chain epoxy groups of the 18-hydroxy-9,10-epoxy octadecanoic acid dominating the cutin of P. dactylifera can form very stable ether bonds conferring particular robustness on the cuticle of this species (Schmidt and Schönherr, 1982; Riederer and Schönherr, 1988), which in turn may decrease the thermal vulnerability of its cuticular transpiration barrier. Schreiber and Schönherr (1990) hypothesized that discrepancies in the thermoelastic properties of cutin and waxes might cause defects in the cuticular transpiration barrier, leading to a steep increase of its water permeability above a specific transition temperature.

Another reason for the divergence in the thermal susceptibility of the cuticular transpiration barriers of the two species may lie in differences in the cuticular wax composition, specifically in the VLC aliphatic part, which has been shown to be the primary factor determining cuticular water permeability (Vogg et al., 2004; Leide et al., 2007, 2011; Buschhaus and Jetter, 2012; Jetter and Riederer, 2016). A model proposed for the molecular structure of VLC cuticular waxes by Riederer and Schreiber (1995) states that cuticular waxes are multiphase systems composed of highly ordered crystalline lamellae interspersed with amorphous zones. According to this model, the inner parts of the alkyl chains of the VLC aliphatic molecules make up the crystalline domains of the cuticular waxes, while the irregularly protruding chain ends, which do not fit into the crystallites, form the adjacent amorphous zones. The cuticular waxes of both P. dactylifera and C. colocynthis fit this model. Although P. dactylifera and C. colocynthis qualitatively present similar aliphatic compound classes, the quantitative contributions of each class to the total cuticular wax load and the carbon chain-lengths of homologous compounds vary considerably. The cuticular waxes of P. dactylifera contain higher amounts of all compound classes except for primary alcohols, which occur in equal quantities. The total VLC wax coverage of the leaflets of P. dactylifera was seven times higher than that of C. colocynthis leaves.

Although the carbon chain-lengths of the VLC aliphatic compounds range from C₂₀ to C₄₆ in C. colocynthis and from C₂₀ to C₆₂ in P. dactylifera, the overall weighted MCL, which are a proxy for the chain-length distributions, differ by only about three carbon atoms (31.8 in P. dactylifera and 29.4 in C. colocynthis). However, the patterns of the respective chain-length distributions of VLC wax components drastically differ between the two species. The VLC aliphatic compounds of both species can be divided into two groups: (i) components with fewer than 40 carbon atoms and (ii) components with chains of 40 or more carbon atoms. N-alkanes, n-alkenes, n-aldehydes, alkanoic acids, and primary and secondary alcohols make up the fraction with chain-lengths <C₄₀ whereas the cuticular wax fraction with chain-length ≥C₄₀ consists exclusively of VLC alkyl esters from long-chain fatty acids and primary alcohols. The aliphatic components with chain-lengths <C₄₀ and ≥C₄₀ comprise 69 mol % and 31 mol %, respectively, of the total VLC aliphatic fraction in P. dactylifera. In contrast, only a minor fraction (1 mol %) of the VLC aliphatic components was ≥C₄₀ in C. colocynthis, while almost all compounds belonged to the shorter chain-length fraction of the cuticular waxes (Fig. 6). The MCL of the shorter-chain components is 29.4 in C. colocynthis and 31.3 in P. dactylifera, while the MCL of the VLC ester fraction in P. dactylifera is 51.3 (Fig. 6).

We hypothesize that the high proportion of VLC aliphatic compounds ≥C₄₀ in P. dactylifera significantly affects the physical properties of the leaf cuticular transpiration barrier, thereby leading to more efficient control of cuticular water loss over the range of temperatures experienced by this plant in its natural habitat. The reason for this particular thermal stability may be the high melting temperature of cuticular waxes, which consist to a considerable degree of VLC alkyl esters. Indeed, the melting temperatures of cuticular waxes as determined by FTIR differ significantly between the two species, with midpoints of the solid–liquid transition occurring at 73 °C in C. colocynthis and 80 °C in P. dactylifera (Fig. 7). This finding shows that waxes of both species differ in their susceptibility to thermally induced changes to their physical properties. Based on evidence obtained from the study of technical waxes, we propose that the presence of appreciable amounts of VLC alkyl esters in the cuticular waxes of P. dactylifera affects the melting range of the cuticular wax in a particular way. It was shown that VLC aliphatic esters intercalate simultaneously into two adjacent crystalline lamellae formed by the shorter-chain VLC aliphatic compounds of a wax (Reynhardt and Riederer, 1994; Dorset, 1999, 2002, 2005) under the condition that the chain-length of the esters is appreciably higher than that of the shorter chain-length fraction. This condition is fulfilled in the case of P. dactylifera wax, with an MCL of 51.3 for the longer-chain fraction and 31.3 for the shorter-chain fraction.

Consequently, a representative ester molecule of 51 C atoms will be integrated simultaneously by up to 25 C atoms each into the two adjacent lamellae composed of molecules with an MCL of 31. These bridging molecules represent covalent anchors holding the crystalline lamellae together (Fig. 9B) and thereby increase the melting point of the cuticular waxes. The tight bridging of the crystalline lamellae also affects the extent of the effect temperature has on the mechanical properties of the wax (Reynhardt and Riederer, 1994). The thermal expansion coefficient of the leaf cuticular wax of the palm Copernicia prunifera (Mill.) H.E. Moore (carnauba wax), which has an overall composition and an alkyl ester content equal to that of P. dactylifera (Basson and Reynhardt, 1988), is much lower than the coefficient for waxes without VLC alkyl esters (Craig et al., 1965).

Fig. 9.

Model proposed for the molecular structure of the cuticular waxes of Citrullus colocynthis leaves (A) and Phoenix dactylifera leaflets (B). The inner parts of the alkyl chains of very-long-chain (VLC) aliphatic molecules (blue zigzag lines) form the highly ordered crystalline domains (blue background), whereas the chain ends make up the adjacent amorphous zones of the cuticular waxes. Dots symbolize functional groups. The absolute chain-lengths and their distribution represent the corresponding experimental values for both species, with weighted median chain-lengths of 29.4 for C. colocynthis and 31.8 for P. dactylifera. The VLC alkyl esters (orange zigzag lines) integrate within the two crystalline lamellae. These bridging molecules form covalent anchors connecting two adjacent crystalline flakes.

Further, paraffin waxes can be mechanically reinforced by the addition of VLC esters. To quantify this effect, the mechanical deformability of paraffin with varying amounts of carnauba wax admixed was measured as a function of temperature by Craig et al. (1966). The addition of 25% of carnauba wax, which leads to a VLC ester content of only 8% in the final mixture, increased the temperature at which the mixed wax could be deformed by a given load by ~10 °C (Craig et al., 1966).

We hypothesize that the absence of an effect of temperature on the cuticular water permeability of P. dactylifera, in contrast to that seen with C. colocynthis, is due to a higher and less temperature-sensitive mechanical stability of the alkyl ester-containing waxes of the former species, even under elevated temperatures. The presence of alkyl ester-containing ‘bridged’ waxes can be imagined to mechanically reinforce the cuticular transpiration barrier of P. dactylifera and, thus, prevent the occurrence of thermally induced structural defects that are thought to lead to the abrupt increases of water permeability at higher temperatures observed in C. colocynthis and many other species (Schreiber, 2001; Riederer, 2006). We further hypothesize that the production of alkyl ester-reinforced cuticular waxes is particularly beneficial to plant species living in hot environments, and that this is comparable to the advantage the desert plant R. stricta is assumed to derive from the incorporation of appreciable amounts of pentacyclic triterpenoids into its cutin matrix (Schuster et al., 2016). The hypotheses we propose here are testable. Therefore, investigations into the phase behaviour and the effect of temperature on the mechanical and rheological properties of cuticular waxes of varying compositions and chain-length distributions are ongoing.

In conclusion, our findings show for the first time that a plant exhibiting the water-saver strategy such as P. dactylifera differs from a water-spender plant like C. colocynthis not only in terms of the response of stomatal regulation to water shortage but also by a constitutive difference in cuticular barrier properties. The cuticle of the water-saver P. dactylifera (i) more efficiently reduces cuticular transpirational water loss and (ii) does not experience thermal damage even under the impact of high temperatures up to 50 °C. The cuticular waxes play an essential role in the efficacy and stability of the transpirational barrier to water loss. We hypothesize that the cuticular transpiration barrier of P. dactylifera leaves is thermally resistant because an appreciable fraction of the wax consists of VLC alkyl esters, which bridge the crystalline lamellae of the cuticular waxes and, thus, reinforce the transport-limiting barrier.

Supplementary data

Supplementary data are available at JXB online.

Table S1. Minimum transpiration (J_min) and leaf to air temperature difference (ΔT_leaf–air) of Citrullus colocynthis leaves and Phoenix dactylifera leaflets obtained from drying curves as functions of air temperatures (T_air).

Table S2. Cuticular wax coverage and components of Citrullus colocynthis leaves and Phoenix dactylifera leaflets.

Table S3. Cutin monomeric coverage of Citrullus colocynthis leaves and Phoenix dactylifera leaflets.

Figure S1. Representative transpiration kinetic curves of Citrullus colocynthis leaves and Phoenix dactylifera leaflets.

Glossary

Abbreviations:

FTIR

Fourier transform infrared spectroscopy

g_min

minimum leaf conductance

MCL

weighted median carbon chain-lengths

RWD

relative water deficit

RWD_SC

relative water deficit at maximum stomatal closure

VLC

very-long-chain

In memory of Otto Ludwig Lange (1927–2017).