Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Genetic differentiation in cauline-leaf-specific wettability of a rosette-forming perennial Arabidopsis from two contrasting montane habitats

Genetic differentiation in cauline-leaf-specific wettability of a rosette-forming perennial... Abstract Background and Aims An altitudinal gradient of leaf wettability is often observed between and within species. To understand its functional significance, positional variation of leaf surfaces within plants should be taken into account. In rosette-forming plants, rosette leaves are near the ground and their adaxial surfaces are exposed, whereas cauline leaves are lifted from the ground throughout the reproductive season, and their abaxial surfaces are more exposed. Here, we investigated leaf wettability of cauline and rosette leaves of Arabidopsis halleri subsp. gemmifera growing in contrasting montane habitats along an altitudinal gradient at Mt Ibuki, Japan. Methods We conducted field investigations and a growth chamber experiment to determine whether field-observed variation in leaf wettability was caused by genetic differentiation. We further performed gene expression analysis of a wax-related gene, i.e. AhgCER1, a homologue of A. thaliana ECERIFERUM1 (CER1) that may be involved in differentiation of leaf wettability. Key Results We found cauline-leaf specific genetic differentiation in leaf wettability between contrasting montane habitats. Cauline leaves of semi-alpine plants, especially on abaxial surfaces, were non-wettable. Cauline leaves of low-altitudinal understorey plants were wettable, and rosette leaves were also wettable in both habitats. AhgCER1 expression corresponded to observed leaf wettability patterns. Conclusions Low wettability of cauline leaves is hypothesized to keep exposed surfaces dry when they are wrapping flowering buds in early spring, and presumably protects flowering buds from frost damage. The genetic system that controls wax content, specifically for cauline leaves, should be involved in the observed genetic differentiation, and AhgCER1 control is a strong candidate for the underlying genetic mechanism. Altitudinal gradient, Arabidopsis halleri subsp. gemmifera, cauline leaf, CER1, cuticular wax, genetic differentiation, leaf wettability, local adaptation, rosette leaf INTRODUCTION Local adaptation, i.e. genetic differentiation in fitness-related traits between populations, is one of the major interests of evolutionary biologists (Savolainen et al., 2013). Adaptive differentiation between populations occurring along altitudinal gradients has been frequently studied because it provides contrasting habitats within a short geographic distance (Gonzalo-Turpin and Hazard, 2009). Along an altitudinal gradient, plants are exposed to deposition of surface water by various forms of precipitation such as rain, ground fog, cloud mist and dew. The consequence of water deposition may vary depending on ambient temperature and the degree of exposure to surrounding environments. The wettability of leaves, i.e. the amount of water captured and retained on leaf surfaces, differs significantly between plant species (Fogg, 1947; Brewer et al., 1991; Aryal and Neuner, 2010, 2012). Different plant species show a broad range of leaf wettability from being covered by a film of water to being completely water-repellent (Smith and McClean, 1989; Brewer et al., 1991; Brewer and Smith, 1997; Brewer and Nuñez, 2007; Aryal and Neuner, 2010, 2012). Along altitudinal gradients, associations with lower wettability in alpine habitats have been reported both between and within species (Aryal and Neuner, 2010, 2012). The functional significance of low or high leaf wettability appears to be highly varied (Brewer et al., 1991), i.e. it might relate to water supply, pathogen infection, extrinsic ice nucleation and stomatal occlusion (Barthlott and Neinhuis, 1997). The rosette is one of the major forms of herbaceous plants that experience harsh, cold and drought during vegetative growth. Rosette-forming plants are especially interesting in studying leaf wettability, because they produce leaves that have distinctive modes of exposure to atmospheric environments. Rosette leaves (leaves that form a rosette) are successively produced during vegetative growth. Prior to flowering seasons, cauline leaves (leaves on flowering stems) are formed from the shoot apical meristem. Cauline leaves are very compact compared with rosette leaves, but they serve as protection for flowering buds in the initial stages of elongation of flowering stems. Rosette leaves remain near the ground and their adaxial surfaces are more exposed, whereas cauline leaves are lifted from the ground throughout the reproductive season and their abaxial surface is more exposed. Therefore, these positional differences may require distinctive levels of leaf wettability for adaxial and abaxial surfaces of rosette and cauline leaves. The main components of epicuticular wax have hydrophobic properties (Holloway, 1969), suggesting that the physiochemistry of the cuticular wax (e.g. wax content, composition and microstructure) influences leaf surface wettability (Neinhuis and Barthlott, 1997; Bhushan and Jung, 2008; Holder, 2011). The genetic systems controlling wax content have been extensively studied in Arabidopsis thaliana (Aarts et al., 1995; Kosma et al., 2009; Bourdenx et al., 2011). In A. thaliana, alkanes represented 70 % of the total wax in leaves (Jetter and Kunst, 2008). The high proportion of alkanes in the epicuticular wax results in hydrophobic plant surfaces (Holloway, 1969). ECERIFERUM1 (CER1) encodes proteins involved in the biosynthesis of the alkane-forming enzyme in A. thaliana (Aarts et al., 1995), and was reported to be more highly expressed in cauline leaves than rosette leaves in A. thaliana (Bourdenx et al., 2011). We do not know, however, how the genetic control of epicuticular wax differentiates along altitudinal gradients in natural systems. In the present study, we compared leaf wettability of Arabidopsis halleri subsp. gemmifera, a rosette-forming perennial, at contrasting altitudinal habitats within a mountain system. The species is a perennial relative of A. thaliana that has been used in studies on the adaptive significance of gene functions under natural conditions (e.g. Aikawa et al., 2010; Kubota et al., 2015). The study species occurs in two distinctive habitats at Mt Ibuki, Japan; i.e. the understorey of Cryptomeria japonica forests near the base of the mountain (referred to as understorey hereafter) and the semi-alpine open habitat near the peak (referred to as semi-alpine hereafter). Although the peak is not very high (1377 m), the mountain is characterized by a combination of exposure to strong cold winds and snow near the top and relatively warm and mild winters near the base. In this paper we pose the following questions regarding the variation in leaf wettability among plants that occur along a steep environmental gradient. Leaf wettability was assessed in both the field and a laboratory growth experiment to quantify leaf wettability in natural habitats and the degree of genetic differentiation. (1) Is there detectable genetic differentiation in leaf wettability between populations occurring in contrasting habitats in the field? (2) Do the patterns correspond to the degree of exposure of the surface of rosette and cauline leaves? (3) Do the differences in leaf wettability correspond with gene expression of a homologue of CER1 (hereafter referred to as AhgCER1)? By answering these three questions, we will be able to determine how genetics, leaf position and field environments interact to determine leaf wettability. MATERIALS AND METHODS Study species and sites The study plant, Arabidopsis halleri subsp. gemmifera, is distributed in East Asia, including north-eastern China, Japan, Korea and Taiwan, and in the Russian Far East. The study was conducted using natural populations of A. halleri subsp. gemmifera occurring in the understorey and semi-alpine habitats of Mt Ibuki, Japan (35°25′04″ N, 136°24′22″ E at the highest peak; Fig. 1). The mountain has a hiking route to the top that is divided into ten pitches. Along the route, we selected five populations, two from the understorey [430 m and 580 m above sea level (a.s.l.)] and three from the semi-alpine area [1220 m, 1300 m and 1370 m a.s.l.] habitats (Fig. 1). Plants in the semi-alpine habitat of Mt Ibuki are distinct from typical A. halleri plants in that they have dense trichomes on both rosette and cauline leaves, and are sometimes treated as a variety of A. halleri subsp. gemmifera (originally described as Arabis gemmifera var. alpicola). The reported level of genome-wide genetic differentiation between the low- and high-altitude populations on Mt Ibuki was low [FST = 0.017, amplified fragment length polymorphism (AFLP) (Ikeda et al., 2010); GST = 0.043–0.048, genome-wide single-nucleotide polymorphism (SNP) (Kubota et al., 2015)]. Fig. 1. View largeDownload slide Maps showing locations of five sampling populations in understorey (430 and 580 m) and semi-alpine (1220, 1300 and 1370 m) habitats at Mt Ibuki (left) and the geographical location of Mt Ibuki in Japan (upper right). Closed circles indicate the location of the five sites along a hiking trail (thick line). Contour lines in the map indicate altitudes above sea level (m). Corresponding latitudes (N) and longitudes (E) are shown on the left and bottom margins Fig. 1. View largeDownload slide Maps showing locations of five sampling populations in understorey (430 and 580 m) and semi-alpine (1220, 1300 and 1370 m) habitats at Mt Ibuki (left) and the geographical location of Mt Ibuki in Japan (upper right). Closed circles indicate the location of the five sites along a hiking trail (thick line). Contour lines in the map indicate altitudes above sea level (m). Corresponding latitudes (N) and longitudes (E) are shown on the left and bottom margins Sampling of field materials At the five study populations in the understorey (430 and 580 m) and semi-alpine (1220, 1300 and 1370 m) habitats (Fig. 1), we collected leaves of A. halleri subsp. gemmifera in autumn (20 November 2010) and spring (24 June 2011). During each spring and autumn sampling period, we randomly selected 40, 10, 30, 30 and 30 plants from the 430, 580, 1220, 1300 and 1370 m sites, respectively, depending on the availability of plants. In autumn, rosettes are formed by leaves that developed during summer and autumn, and there were no cauline leaves. In spring, plants possessed overwintered leaves on rosettes and newly developed cauline leaves on elongated flowering stalks. From each plant, we collected one rosette leaf in autumn and one rosette and one cauline (6th- to 8th-position leaves, counted basipetally from the top) leaves in spring. All collected leaves were fresh and had little natural damage. Leaves were placed in plastic bags, kept in a cool box and taken to the laboratory for further measurements. For gene expression analysis, leaves were collected in spring (19 June 2017). One cauline leaf (6th- to 8th-position leaves counted basipetally from the top) and one rosette leaf were collected from the same individuals. In total, 12, 6, 6 and 6 sets of cauline and rosette leaves were collected from the 430, 1220, 1300 and 1370 m sites, respectively. Samples were washed in 0.1 % Triton X (0.5 mL) for 10 s and preserved in RNAlater (0.5 mL, Invitrogen, Thermo Fisher Scientific, MA, USA), kept in a cool box and taken to the laboratory. Samples were kept at 4 °C for 24 h and then transferred to −20 °C. Growth experiments During the fruiting season of 2012, seeds were collected from the study populations. Seeds were sown in Petri dishes with moistened quartz sand, placed in an incubator (day/night temperature 22 Cm/15 C5, 14 h light; Koitotron HNM-S, Koito, Osaka, Japan). Seedlings were transferred into plastics pots (75 mm in diameter and 65 mm in depth) containing vermiculite. A total of 72 plants, 36 from the understorey habitats (originated from six and four mother plants from the 430 and 580 m sites, respectively) and 36 from the semi-alpine habitats (originated from two and two mother plants from the 1220 and 1370 m sites, respectively), were prepared. The number of plants prepared per parental plant varied; therefore, results for the understorey and semi-alpine habitats were compared. We placed the pots in the incubator at 20 °C with 12 h light for 6 weeks prior to the first sampling of leaves (hereafter the pre-cold sampling). The light intensity of photosynthetically active radiation (PAR) averaged 65.4 ± 3.6 μm m−2 s−1 at the pot surface. We collected one rosette leaf per plant for all pots. Then, plants were transferred to a 5 °C (12 h light) condition for 6 weeks as a vernalization treatment to induce flowering. After vernalization, plants were again placed in the 20 °C condition (12 h light) for 4 weeks until plants possessed 15–20 flowers. Then, we conducted the second leaf sampling (hereafter post-cold sampling). One rosette and one cauline leaf were sampled from each plant. During the experiments, pots were watered regularly and fertilized once a week (20 times in total) with a solution of 0.1 % Hyponex (a liquid fertilizer, N:P:K = 6:10:5; Hyponex Japan, Osaka). In terms of whether the plants were in vegetative or reproductive phase, pre-cold and post-cold samplings corresponded with autumn and spring samplings in the natural habitats. Measurements For adaxial and abaxial surfaces of all sampled leaves from the field and the growth experiments, we measured leaf wettability, droplet retention, trichome density and density and size of stomata. The measurements were conducted using the leaf surface area where leaves were widest. Care was taken to avoid veins. Leaf wettability was measured as the contact angle (θ, Fig. 2) of a water droplet placed on the leaf surface following the procedure of Brewer et al. (1991). Leaf samples were mounted horizontally on an observation stage using double-sided tape. A 5-µL droplet (recommended standard volume in Brewer et al. 1991) of distilled water was then placed on the leaf surface using a micropipette. A digital photographic image of the horizontal view of the droplet on the leaf surface was taken with a CCD camera connected to a digital microscope (Keyence Japan, Osaka). The digital photographs were processed with ImageJ software (National Institutes of Health, MD, USA). The contact angle (θ) of a line tangent to the water droplet through the point of contact between the droplet and leaf surface was measured by using the ‘tangent 1 default method’ according to Brewer et al. (1991) (Fig. 2). The larger the value of θ, the more repellent the leaf surface; thus, the criteria for judging surface wettability depended on θ. Leaves were classified as highly wettable if θ was <90°, based on previous reports (Fogg, 1948; Challen, 1960; Adam, 1963; Crisp, 1963; Warburton, 1963; Holloway, 1970; Schönherr and Bukovac, 1972) and wettable if θ was <110° (Crisp, 1963). If θ exceeded 110° the leaf was considered non-wettable. Highly non-wettable leaves had θ values >130° (Crisp, 1963). To determine wax deposition on the leaf surfaces, we examined leaves under a scanning electron microscope (Philips XL 30, Eindhoven, The Netherlands). Fig. 2. View largeDownload slide Contact angle (θ) is the angle between the surface of the leaf and the line tangent to the droplet at the point of contact between air, water and the leaf surface. Leaf wettability is categorized into four classes based on θ values (see text for details of categorization). Fig. 2. View largeDownload slide Contact angle (θ) is the angle between the surface of the leaf and the line tangent to the droplet at the point of contact between air, water and the leaf surface. Leaf wettability is categorized into four classes based on θ values (see text for details of categorization). Droplet retention is a measure of the ‘stickiness’ of a leaf surface. We determined water retention for adaxial and abaxial surfaces of sampled leaves using the method of Brewer and Smith (1997). Leaves were mounted on a metal plate fixed to a goniometer (Technical Support Division, CER, Kyoto University). A 50-µL droplet (recommended standard volume in Brewer and Smith, 1997) was initially placed on a horizontal leaf surface. Then the angle of leaf inclination was successively increased, and when the droplet began to move the angle of inclination was recorded. Higher angles (>60°) indicated a greater tendency to retain droplets, whereas lower angles (<20°) indicated leaf surfaces that readily shed droplets. We counted the number of stomata and trichomes using Suzuki’s universal micro-printing (SUMP) method (Kenis, Osaka, Japan). Micro-printings were digitized using an optical microscope. We recorded the number of stomata and trichomes within a 0.0625-mm−2 area for both adaxial and abaxial leaf surfaces of each leaf sample and calculated the number of stomata or trichome per unit of leaf area (mm2). The stomatal index was calculated according to the method of Salisbury (1927) as the ratio of number of stomata to number of epidermal cells. Guard cell length and pore length were measured using ImageJ software. Gene expression analysis For AhgCER1 quantification, total RNA was extracted from leaves with a Maxwell® 16 LEV Plant RNA Kit according to the manufacturer’s instructions (Promega, Fitchburg, WI, USA). The amount of extracted RNA was measured using a Quantus Fluorometer (Promega). We used 200 ng of RNA for cDNA synthesis using a High-Capacity cDNA Reverse Transcription Kit (Life Technologies). Quantitative PCR analyses were conducted using the cDNA samples. Two technical replicates were prepared for each sample. PCR conditions were 95 °C for 20 s followed by 40 cycles at 95 °C for 1 s and 60 °C for 20 s. The melting curve analyses were performed by gradually increasing the temperature (0.05 °C s−/) from 60 to 95 °C for 15 s. The PCR reactions were performed using a 10-μL volume containing 200 nm primers and Fast SYBR Green Master Mix (Applied Biosystems). We used AhgACT2 and AhgPP2AA3 as reference genes for normalization among samples (Nishio et al., 2016). Sequences of all primers used are shown in Supplementary Data 1, Table S1. We obtained standard curves using a dilution series of a standard A. halleri cDNA at 1, 1/5, 1/25, 1/125 and 1/625 for AhgACT2 and AhgPP2AA3 and 1, 1/6, 1/36, 1/216 and 1/1296 for AhgCER1 in all analyses. Statistical analysis We conducted two-way analysis of variance (ANOVA) tests and one-way nested ANOVAs for measured leaf wettability (contact angle and droplet retention), as well as stomatal density, stomatal index, guard cell length and pore length/unit area as field measurements for cauline and rosette leaves collected in spring and rosette leaves collected in autumn. For each leaf type, effects of leaf surface (abaxial and adaxial) and population (430, 580, 1220, 1300 and 1370 m) were tested using two-way ANOVAs. We divided the five populations into two altitudinal habitat types, understorey (430 and 580 m) and semi-alpine (1220, 1300 and 1370 m). Effects of between-habitat (understorey versus semi-alpine) and within-habitat variation (between populations) were tested using one-way nested ANOVAs (populations were nested within habitats) separately for each surface. In the understorey habitat (430 and 580 m), the leaves had no trichomes on either leaf surface. Therefore, for altitude effects on trichome density the two-way ANOVA tested effects across the three populations in the semi-alpine habitat (1220, 1300 and 1370 m) and there was no between-habitats term in the one-way nested ANOVAs. Similarly, for growth experiments, seeds were derived from two habitat types [understorey (430 and 580 m) and semi-alpine (1220 and 1370 m)]. For each leaf type, effects of leaf surface (abaxial and adaxial) and habitat (understorey and semi-alpine) were tested using two-way ANOVAs for leaf wettability (contact angle and droplet retention), as well as stomatal density, stomatal index, guard cell length and pore length/unit area of cauline and rosette leaves collected in post- and pre-cold conditions. We did not use a two-way ANOVA for trichome density, and surface effects were evaluated using a one-way ANOVA for plants from the semi-alpine habitat. For gene expression analysis, expression of AhgCER1 relative to that of reference genes was converted to a log2 (x + 1) scale. Effects of leaf type (cauline or rosette), population [understorey (430 m) and semi-alpine (1220, 1300 and 1370 m)] and their interaction were tested using two-way ANOVAs. All statistical analyses were conducted using SPSS (SPSS, Chicago, IL, USA). Prior to conducting ANOVA, we applied Kolmogorov–Smirnov tests, and 253 out of 258 data sets (combinations of experiment × leaf type × surface × habitat/population × trait) did not deviate significantly from normal distributions (Supplementary Data 1, Table S2). All data analysed in this study are available in Supplementary Data 2, 3 and 4 for field measurements, growth experiments and gene expression analyses, respectively. RESULTS Field measurements Cauline leaves from the understorey habitat at low altitudes (430 and 580 m) were highly wettable, as indicated by low θ values (average ranging from 56.1° to 67.5°) for both adaxial and abaxial leaf surfaces (Fig. 3A). However, average θ values were nearly 110° or higher for leaves from the semi-alpine habitat at high altitudes (1220, 1300 and 1370 m), i.e. the leaves were classified as non-wettable for both adaxial and abaxial surfaces (Fig. 3A). A highly significant population effect was detected in the two-way ANOVA. Similarly, there was a significant between-habitats effect, whereas no within-habitat effect was detected in the one-way nested ANOVA, suggesting that most of the variation was attributable to habitat differences (Table 1A). We detected no significant differences between adaxial and abaxial surfaces (Table 1A). For the rosette leaves from spring and autumn, the θ values (leaf wettability) were <110° for all populations (Fig. 3B, C). This indicated that all investigated populations exhibited increased leaf wettability of rosette leaves. Although we detected significant population effects in the two-way ANOVA, they were attributable to variation between populations within habitats (Table 1A). No-significant differences were found in θ values of the two leaf surfaces collected from spring and autumn rosette leaves (Fig. 3B, C, Table 1A). Fig. 3. View largeDownload slide Contact angle (A–C) and droplet retention (D–F) of cauline (A, D) and rosette leaves in spring (B, E) and autumn (C, F) collected in different populations [430, 580 m (understorey); 1220, 1300, 1370 m (semi-alpine)]. Cauline leaves exist only in spring on the flowering stalks. White and grey boxes represent adaxial and abaxial surfaces, respectively. Box plots show the median (line inside the box) and upper and lower quartiles (75 %, 25 %) and maximum and minimum values. (G, H) Scanning electron microscopic images of the abaxial surfaces of cauline leaves from understorey (G) and semi-alpine (H) habitats. Scale bar = 100 µm. Fig. 3. View largeDownload slide Contact angle (A–C) and droplet retention (D–F) of cauline (A, D) and rosette leaves in spring (B, E) and autumn (C, F) collected in different populations [430, 580 m (understorey); 1220, 1300, 1370 m (semi-alpine)]. Cauline leaves exist only in spring on the flowering stalks. White and grey boxes represent adaxial and abaxial surfaces, respectively. Box plots show the median (line inside the box) and upper and lower quartiles (75 %, 25 %) and maximum and minimum values. (G, H) Scanning electron microscopic images of the abaxial surfaces of cauline leaves from understorey (G) and semi-alpine (H) habitats. Scale bar = 100 µm. Table 1. Results of two-way ANOVA and one-way nested ANOVA for leaf wettability (contact angle and droplet retention) of cauline and rosette leaves in spring and rosette leaves in autumn of Arabidopsis halleri subsp. gemmifera collected from five populations in two habitat types. For each leaf type, effects of leaf surfaces and populations were tested by two-way ANOVA. Effects of between-habitat and within-habitat-between-population variation [understorey (430 and 580 m) versus semi-alpine (1220, 1300 and 1370 m)] were tested by one-way nested ANOVA (populations were nested within habitats) for each surface separately. Values are F ratios Trait Two-way ANOVA One-way nested ANOVA Population Surface Interaction Surface Between habitats Within habitat (A) Contact angle, θ (°) Cauline leaf 99.2*** 0.59ns 0.85ns AD 111*** 0.90 ns AB 174*** 0.93 ns Spring rosette leaf 13.3*** 0.21ns 0.91ns AD 4.10ns 2.64ns AB 0.56ns 7.27*** Autumn rosette leaf 4.90** 3.30ns 1.71ns AD 1.71ns 2.30ns AB 0.03ns 5.06** (B) Droplet retention (°) Cauline leaf 1.85ns 0.42ns 1.18ns AD 0.11 ns 1.62 ns AB 0.15 ns 1.86 ns Spring rosette leaf 4.06** 1.08ns 0.97ns AD 4.84ns 0.73ns AB 0.03ns 3.45* Autumn rosette leaf 3.18* 2.48ns 0.52ns AD 0.69ns 1.14ns AB 0.23ns 3.24* Trait Two-way ANOVA One-way nested ANOVA Population Surface Interaction Surface Between habitats Within habitat (A) Contact angle, θ (°) Cauline leaf 99.2*** 0.59ns 0.85ns AD 111*** 0.90 ns AB 174*** 0.93 ns Spring rosette leaf 13.3*** 0.21ns 0.91ns AD 4.10ns 2.64ns AB 0.56ns 7.27*** Autumn rosette leaf 4.90** 3.30ns 1.71ns AD 1.71ns 2.30ns AB 0.03ns 5.06** (B) Droplet retention (°) Cauline leaf 1.85ns 0.42ns 1.18ns AD 0.11 ns 1.62 ns AB 0.15 ns 1.86 ns Spring rosette leaf 4.06** 1.08ns 0.97ns AD 4.84ns 0.73ns AB 0.03ns 3.45* Autumn rosette leaf 3.18* 2.48ns 0.52ns AD 0.69ns 1.14ns AB 0.23ns 3.24* AD, adaxial; AB, abaxial. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant at P < 0.05. View Large Table 1. Results of two-way ANOVA and one-way nested ANOVA for leaf wettability (contact angle and droplet retention) of cauline and rosette leaves in spring and rosette leaves in autumn of Arabidopsis halleri subsp. gemmifera collected from five populations in two habitat types. For each leaf type, effects of leaf surfaces and populations were tested by two-way ANOVA. Effects of between-habitat and within-habitat-between-population variation [understorey (430 and 580 m) versus semi-alpine (1220, 1300 and 1370 m)] were tested by one-way nested ANOVA (populations were nested within habitats) for each surface separately. Values are F ratios Trait Two-way ANOVA One-way nested ANOVA Population Surface Interaction Surface Between habitats Within habitat (A) Contact angle, θ (°) Cauline leaf 99.2*** 0.59ns 0.85ns AD 111*** 0.90 ns AB 174*** 0.93 ns Spring rosette leaf 13.3*** 0.21ns 0.91ns AD 4.10ns 2.64ns AB 0.56ns 7.27*** Autumn rosette leaf 4.90** 3.30ns 1.71ns AD 1.71ns 2.30ns AB 0.03ns 5.06** (B) Droplet retention (°) Cauline leaf 1.85ns 0.42ns 1.18ns AD 0.11 ns 1.62 ns AB 0.15 ns 1.86 ns Spring rosette leaf 4.06** 1.08ns 0.97ns AD 4.84ns 0.73ns AB 0.03ns 3.45* Autumn rosette leaf 3.18* 2.48ns 0.52ns AD 0.69ns 1.14ns AB 0.23ns 3.24* Trait Two-way ANOVA One-way nested ANOVA Population Surface Interaction Surface Between habitats Within habitat (A) Contact angle, θ (°) Cauline leaf 99.2*** 0.59ns 0.85ns AD 111*** 0.90 ns AB 174*** 0.93 ns Spring rosette leaf 13.3*** 0.21ns 0.91ns AD 4.10ns 2.64ns AB 0.56ns 7.27*** Autumn rosette leaf 4.90** 3.30ns 1.71ns AD 1.71ns 2.30ns AB 0.03ns 5.06** (B) Droplet retention (°) Cauline leaf 1.85ns 0.42ns 1.18ns AD 0.11 ns 1.62 ns AB 0.15 ns 1.86 ns Spring rosette leaf 4.06** 1.08ns 0.97ns AD 4.84ns 0.73ns AB 0.03ns 3.45* Autumn rosette leaf 3.18* 2.48ns 0.52ns AD 0.69ns 1.14ns AB 0.23ns 3.24* AD, adaxial; AB, abaxial. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant at P < 0.05. View Large The angle of droplet retention, an index of the stickiness of a leaf surface, was typically <20°. This indicated that the majority of cauline leaves (Fig. 3D) and rosette leaves in both spring (Fig. 3E) and autumn (Fig. 3F) had low droplet retention capacity for both leaf surfaces. A cauline-leaf-specific difference between habitats was not observed in droplet retention (Fig. 3D–F). The results of the two-way ANOVA and nested ANOVA indicated that there were no significant effects for all terms in cauline leaves (Table 1B). Significant differences were observed in spring and autumn rosette leaves for the population term in the two-way ANOVA, but habitat differences were not significant in the nested ANOVA (Table 1B). Scanning electron microscopic images showed that there were no marked differences in surface texture on the abaxial surfaces of cauline leaves between understorey and semi-alpine habitats (Fig. 3G, H). Similarly, no marked and consistent differences were detected between understorey and semi-alpine plants for other surfaces of cauline and rosette leaves (Supplementary Data 1, Fig. S1). Plants from the understorey habitat had no trichomes on either surface of all types of leaves collected in spring and autumn (Table 2A). Semi-alpine leaves had trichomes on both surfaces. The trichome density of semi-alpine leaves was generally higher in cauline leaves (population average range 59.7–70.7 and 72.0–94.4 mm−2 for adaxial and abaxial surfaces, respectively) than rosette leaves from both spring and autumn (population average range 23.6–36.6 and 24.1–39.3 mm−2 for adaxial and abaxial surfaces, respectively; Table 2A). Adaxial and abaxial surfaces exhibited significant differences only for trichome density of cauline leaves in the two-way ANOVA (P < 0.05; Table 2A). Table 2. Trichome density and stomatal density of cauline and rosette leaves in spring and rosette leaves in autumn of Arabidopsis halleri subsp. gemmifera collected from five populations. Average ± s.e. (n = 10–40) for adaxial (AD) and abaxial (AB) surfaces are given. For explanations of the ANOVAs used see the caption of Table 1. Because no plants in the understorey habitat (430 and 580 m) had trichomes on either leaf surface, they were not included in the statistical analyses. Therefore, population effects in two-way ANOVA represent effects across three populations in the semi-alpine habitat, and there is no between-habitats term in the one-way nested ANOVAs for trichome density (shown as –) Leaf type Surface Populations Two-way ANOVA One-way nested ANOVA Understorey habitat (m) Semi-alpine habitat (m) 430 580 1220 1300 1370 Population Surface Inter-action Between habitats Within habitat (A) Trichome density (mm−2) Cauline leaf AD 0.0 ± 0 0.0 ± 0 61.9 ± 3.1 59.7 ± 3.4 70.7 ± 3.1 – 3.28* AB 0.0 ± 0 0.0 ± 0 78.0 ± 5.5 72.0 ± 2.8 94.4 ± 3.8 9.07ns 27.0* 1.20ns – 7.71** Spring rosette leaf AD 0.0 ± 0 0.0 ± 0 36.6 ± 2.6 30.5 ± 2.3 29.6 ± 2.2 – 2.64ns AB 0.0 ± 0 0.0 ± 0 38.4 ± 2.2 35.4 ± 1.6 39.3 ± 2.8 2.03ns 8.57ns 1.51ns – 0.79ns Autumn rosette leaf AD 0.0 ± 0 0.0 ± 0 30.0 ± 2.4 34.2 ± 2.2 23.6 ± 1.9 – 5.22** AB 0.0 ± 0 0.0 ± 0 33.6 ± 2.0 36.4 ± 2.6 24.1 ± 1.7 13.03*** 1.45 ns 0.25 ns – 8.13** (B) Stomatal density (mm−2) Cauline leaf AD 31.3 ± 1.3 37.3 ± 2.4 64.7 ± 2.2 65.6 ± 1.9 73.2 ± 2.1 60** 4.78** AB 38.8 ± 1.4 41.0 ± 2.4 94.0 ± 2.9 82.6 ± 2.2 88.1 ± 2.3 223*** 90.1*** 9.42*** 101** 4.48** Spring rosette leaf AD 34.0 ± 1.3 35.2 ± 2.7 65.0 ± 2.3 68.6 ± 2.6 69.5 ± 1.5 218*** 0.94ns AB 49.6 ± 1.9 38.9 ± 6.0 89.2 ± 2.5 88.0 ± 2.2 97.6 ± 2.6 161*** 120*** 4.84*** 75.64** 4.64** Autumn rosette leaf AD 46.0 ± 2.3 51.7 ± 4.9 83.4 ± 3.0 71.5 ± 2.9 62.8 ± 3.3 7.38 ns 8.68*** AB 60.8 ± 2.6 46.9 ± 3.4 113.6 ± 5.0 80.2 ± 3.1 81.8 ± 2.8 66.05*** 33.52*** 5.02* 5.68 ns 21.43*** Leaf type Surface Populations Two-way ANOVA One-way nested ANOVA Understorey habitat (m) Semi-alpine habitat (m) 430 580 1220 1300 1370 Population Surface Inter-action Between habitats Within habitat (A) Trichome density (mm−2) Cauline leaf AD 0.0 ± 0 0.0 ± 0 61.9 ± 3.1 59.7 ± 3.4 70.7 ± 3.1 – 3.28* AB 0.0 ± 0 0.0 ± 0 78.0 ± 5.5 72.0 ± 2.8 94.4 ± 3.8 9.07ns 27.0* 1.20ns – 7.71** Spring rosette leaf AD 0.0 ± 0 0.0 ± 0 36.6 ± 2.6 30.5 ± 2.3 29.6 ± 2.2 – 2.64ns AB 0.0 ± 0 0.0 ± 0 38.4 ± 2.2 35.4 ± 1.6 39.3 ± 2.8 2.03ns 8.57ns 1.51ns – 0.79ns Autumn rosette leaf AD 0.0 ± 0 0.0 ± 0 30.0 ± 2.4 34.2 ± 2.2 23.6 ± 1.9 – 5.22** AB 0.0 ± 0 0.0 ± 0 33.6 ± 2.0 36.4 ± 2.6 24.1 ± 1.7 13.03*** 1.45 ns 0.25 ns – 8.13** (B) Stomatal density (mm−2) Cauline leaf AD 31.3 ± 1.3 37.3 ± 2.4 64.7 ± 2.2 65.6 ± 1.9 73.2 ± 2.1 60** 4.78** AB 38.8 ± 1.4 41.0 ± 2.4 94.0 ± 2.9 82.6 ± 2.2 88.1 ± 2.3 223*** 90.1*** 9.42*** 101** 4.48** Spring rosette leaf AD 34.0 ± 1.3 35.2 ± 2.7 65.0 ± 2.3 68.6 ± 2.6 69.5 ± 1.5 218*** 0.94ns AB 49.6 ± 1.9 38.9 ± 6.0 89.2 ± 2.5 88.0 ± 2.2 97.6 ± 2.6 161*** 120*** 4.84*** 75.64** 4.64** Autumn rosette leaf AD 46.0 ± 2.3 51.7 ± 4.9 83.4 ± 3.0 71.5 ± 2.9 62.8 ± 3.3 7.38 ns 8.68*** AB 60.8 ± 2.6 46.9 ± 3.4 113.6 ± 5.0 80.2 ± 3.1 81.8 ± 2.8 66.05*** 33.52*** 5.02* 5.68 ns 21.43*** AD, adaxial; AB, abaxial. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant at P < 0.05. View Large Table 2. Trichome density and stomatal density of cauline and rosette leaves in spring and rosette leaves in autumn of Arabidopsis halleri subsp. gemmifera collected from five populations. Average ± s.e. (n = 10–40) for adaxial (AD) and abaxial (AB) surfaces are given. For explanations of the ANOVAs used see the caption of Table 1. Because no plants in the understorey habitat (430 and 580 m) had trichomes on either leaf surface, they were not included in the statistical analyses. Therefore, population effects in two-way ANOVA represent effects across three populations in the semi-alpine habitat, and there is no between-habitats term in the one-way nested ANOVAs for trichome density (shown as –) Leaf type Surface Populations Two-way ANOVA One-way nested ANOVA Understorey habitat (m) Semi-alpine habitat (m) 430 580 1220 1300 1370 Population Surface Inter-action Between habitats Within habitat (A) Trichome density (mm−2) Cauline leaf AD 0.0 ± 0 0.0 ± 0 61.9 ± 3.1 59.7 ± 3.4 70.7 ± 3.1 – 3.28* AB 0.0 ± 0 0.0 ± 0 78.0 ± 5.5 72.0 ± 2.8 94.4 ± 3.8 9.07ns 27.0* 1.20ns – 7.71** Spring rosette leaf AD 0.0 ± 0 0.0 ± 0 36.6 ± 2.6 30.5 ± 2.3 29.6 ± 2.2 – 2.64ns AB 0.0 ± 0 0.0 ± 0 38.4 ± 2.2 35.4 ± 1.6 39.3 ± 2.8 2.03ns 8.57ns 1.51ns – 0.79ns Autumn rosette leaf AD 0.0 ± 0 0.0 ± 0 30.0 ± 2.4 34.2 ± 2.2 23.6 ± 1.9 – 5.22** AB 0.0 ± 0 0.0 ± 0 33.6 ± 2.0 36.4 ± 2.6 24.1 ± 1.7 13.03*** 1.45 ns 0.25 ns – 8.13** (B) Stomatal density (mm−2) Cauline leaf AD 31.3 ± 1.3 37.3 ± 2.4 64.7 ± 2.2 65.6 ± 1.9 73.2 ± 2.1 60** 4.78** AB 38.8 ± 1.4 41.0 ± 2.4 94.0 ± 2.9 82.6 ± 2.2 88.1 ± 2.3 223*** 90.1*** 9.42*** 101** 4.48** Spring rosette leaf AD 34.0 ± 1.3 35.2 ± 2.7 65.0 ± 2.3 68.6 ± 2.6 69.5 ± 1.5 218*** 0.94ns AB 49.6 ± 1.9 38.9 ± 6.0 89.2 ± 2.5 88.0 ± 2.2 97.6 ± 2.6 161*** 120*** 4.84*** 75.64** 4.64** Autumn rosette leaf AD 46.0 ± 2.3 51.7 ± 4.9 83.4 ± 3.0 71.5 ± 2.9 62.8 ± 3.3 7.38 ns 8.68*** AB 60.8 ± 2.6 46.9 ± 3.4 113.6 ± 5.0 80.2 ± 3.1 81.8 ± 2.8 66.05*** 33.52*** 5.02* 5.68 ns 21.43*** Leaf type Surface Populations Two-way ANOVA One-way nested ANOVA Understorey habitat (m) Semi-alpine habitat (m) 430 580 1220 1300 1370 Population Surface Inter-action Between habitats Within habitat (A) Trichome density (mm−2) Cauline leaf AD 0.0 ± 0 0.0 ± 0 61.9 ± 3.1 59.7 ± 3.4 70.7 ± 3.1 – 3.28* AB 0.0 ± 0 0.0 ± 0 78.0 ± 5.5 72.0 ± 2.8 94.4 ± 3.8 9.07ns 27.0* 1.20ns – 7.71** Spring rosette leaf AD 0.0 ± 0 0.0 ± 0 36.6 ± 2.6 30.5 ± 2.3 29.6 ± 2.2 – 2.64ns AB 0.0 ± 0 0.0 ± 0 38.4 ± 2.2 35.4 ± 1.6 39.3 ± 2.8 2.03ns 8.57ns 1.51ns – 0.79ns Autumn rosette leaf AD 0.0 ± 0 0.0 ± 0 30.0 ± 2.4 34.2 ± 2.2 23.6 ± 1.9 – 5.22** AB 0.0 ± 0 0.0 ± 0 33.6 ± 2.0 36.4 ± 2.6 24.1 ± 1.7 13.03*** 1.45 ns 0.25 ns – 8.13** (B) Stomatal density (mm−2) Cauline leaf AD 31.3 ± 1.3 37.3 ± 2.4 64.7 ± 2.2 65.6 ± 1.9 73.2 ± 2.1 60** 4.78** AB 38.8 ± 1.4 41.0 ± 2.4 94.0 ± 2.9 82.6 ± 2.2 88.1 ± 2.3 223*** 90.1*** 9.42*** 101** 4.48** Spring rosette leaf AD 34.0 ± 1.3 35.2 ± 2.7 65.0 ± 2.3 68.6 ± 2.6 69.5 ± 1.5 218*** 0.94ns AB 49.6 ± 1.9 38.9 ± 6.0 89.2 ± 2.5 88.0 ± 2.2 97.6 ± 2.6 161*** 120*** 4.84*** 75.64** 4.64** Autumn rosette leaf AD 46.0 ± 2.3 51.7 ± 4.9 83.4 ± 3.0 71.5 ± 2.9 62.8 ± 3.3 7.38 ns 8.68*** AB 60.8 ± 2.6 46.9 ± 3.4 113.6 ± 5.0 80.2 ± 3.1 81.8 ± 2.8 66.05*** 33.52*** 5.02* 5.68 ns 21.43*** AD, adaxial; AB, abaxial. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant at P < 0.05. View Large The number of stomata per unit area (stomatal density) was higher in leaves from the semi-alpine habitat than in those from the understorey habitat for all cauline leaves and spring and autumn rosette leaves (Table 2B). Stomatal density on the abaxial surfaces was significantly higher than on the adaxial surfaces of all three leaf types (Table 2B). These differences were mostly supported by statistically significant differences in the two-way ANOVA and one-way nested ANOVA (Table 2B). For other stomatal characters, i.e. stomatal index, guard cell length and pore length, there were no consistent habitat differences, except for the stomatal index of spring rosette leaves and pore length of cauline leaves (Supplementary Data 1, Table S3). Growth experiments In the growth chamber experiment, cauline leaves from the understorey habitat exhibited average θ values <110° for both adaxial and abaxial surfaces, i.e. the leaves were classified as wettable surfaces. The adaxial surface of cauline leaves from the semi-alpine habitat had average θ values >110°, i.e. the leaves were classified as non-wettable; furthermore, the average θ value of abaxial surfaces was >130°, i.e. the leaves were classified as highly non-wettable (Fig. 4A). Highly significant effects were detected for habitat, surface and their interaction in the two-way ANOVA (Table 3A). For the post-cold rosette leaves in the understorey and semi-alpine habitat, the θ values were <110° for both adaxial and abaxial surfaces (Fig. 4B). This indicated that almost all the post-cold rosette leaves were wettable surfaces. However, we found significant differences between habitats in the two-way ANOVA (Table 3A). No significant effects were detected for surface and interaction terms (Table 3A). For the rosette leaves at the pre-cold sampling, the average θ values were ~110° (the boundary value between wettable and non-wettable) for both surfaces and habitats (Fig. 4C). We detected no significant effects for all terms in the two-way ANOVA (Table 3A). Fig. 4. View largeDownload slide Contact angle (A–C) and droplet retention (D–F) of cauline (A, D) and rosette leaves in post-cold (B, E) and pre-cold (C, F) leaves collected from the growth chamber. White and grey boxes represent adaxial and abaxial surfaces, respectively. Box plots show the median (line inside the box), upper and lower quartiles (75 %, 25 %) and maximum and minimum values. Fig. 4. View largeDownload slide Contact angle (A–C) and droplet retention (D–F) of cauline (A, D) and rosette leaves in post-cold (B, E) and pre-cold (C, F) leaves collected from the growth chamber. White and grey boxes represent adaxial and abaxial surfaces, respectively. Box plots show the median (line inside the box), upper and lower quartiles (75 %, 25 %) and maximum and minimum values. Table 3. Results of two-way ANOVA for leaf wettability (contact angle and droplet retention) of cauline leaves and post- and pre-cold rosette leaves of Arabidopsis halleri subsp. gemmifera plants grown in growth chambers. Seeds were derived from two habitat types [understorey (430 and 580 m) and semi-alpine (1220 and 1370 m)]. For each leaf type, effects of leaf surfaces and habitats were tested by two-way ANOVAs. Values are F ratios Trait Leaf type Two-way ANOVA Habitat Surface Interaction (A) Contact angle, θ (°) Cauline leaf 164.25*** 13.09*** 33.11*** Post-cold rosette leaf 49.20*** 3.63ns 2.66ns Pre-cold rosette leaf 0.001ns 0.16ns 1.52ns (B) Droplet retention (°) Cauline leaf 3.25ns 1.62ns 0.03ns Post-cold rosette leaf 2.11ns 3.29ns 0.002ns Pre-cold rosette leaf 0.18ns 0.39ns 1.35ns Trait Leaf type Two-way ANOVA Habitat Surface Interaction (A) Contact angle, θ (°) Cauline leaf 164.25*** 13.09*** 33.11*** Post-cold rosette leaf 49.20*** 3.63ns 2.66ns Pre-cold rosette leaf 0.001ns 0.16ns 1.52ns (B) Droplet retention (°) Cauline leaf 3.25ns 1.62ns 0.03ns Post-cold rosette leaf 2.11ns 3.29ns 0.002ns Pre-cold rosette leaf 0.18ns 0.39ns 1.35ns *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant at P < 0.05. View Large Table 3. Results of two-way ANOVA for leaf wettability (contact angle and droplet retention) of cauline leaves and post- and pre-cold rosette leaves of Arabidopsis halleri subsp. gemmifera plants grown in growth chambers. Seeds were derived from two habitat types [understorey (430 and 580 m) and semi-alpine (1220 and 1370 m)]. For each leaf type, effects of leaf surfaces and habitats were tested by two-way ANOVAs. Values are F ratios Trait Leaf type Two-way ANOVA Habitat Surface Interaction (A) Contact angle, θ (°) Cauline leaf 164.25*** 13.09*** 33.11*** Post-cold rosette leaf 49.20*** 3.63ns 2.66ns Pre-cold rosette leaf 0.001ns 0.16ns 1.52ns (B) Droplet retention (°) Cauline leaf 3.25ns 1.62ns 0.03ns Post-cold rosette leaf 2.11ns 3.29ns 0.002ns Pre-cold rosette leaf 0.18ns 0.39ns 1.35ns Trait Leaf type Two-way ANOVA Habitat Surface Interaction (A) Contact angle, θ (°) Cauline leaf 164.25*** 13.09*** 33.11*** Post-cold rosette leaf 49.20*** 3.63ns 2.66ns Pre-cold rosette leaf 0.001ns 0.16ns 1.52ns (B) Droplet retention (°) Cauline leaf 3.25ns 1.62ns 0.03ns Post-cold rosette leaf 2.11ns 3.29ns 0.002ns Pre-cold rosette leaf 0.18ns 0.39ns 1.35ns *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant at P < 0.05. View Large Average droplet retention of water was <20° for both surfaces of cauline leaves and post-cold and pre-cold rosette leaves (Fig. 4D, E and F, respectively). This indicated that all three leaf types had low droplet retention capacity for both surfaces. No significant effects were detected in the two-way ANOVAs (Table 3B). Plants from the understorey habitat had no trichomes on either surface for all three leaf types (Table 4A). Semi-alpine leaves had trichomes on both surfaces. The number of trichomes per unit area was significantly higher for abaxial than adaxial surfaces for cauline leaves only (Table 4A). Table 4. Trichome density and stomatal density of cauline leaves and post- and pre-cold rosette leaves of Arabidopsis halleri subsp. gemmifera plants grown in growth chambers. Seeds were derived from two habitat types [understorey (430 and 580 m) and semi-alpine (1220 and 1370 m)]. Average ± s.e. (n = 36) for adaxial (AD) and abaxial (AB) surfaces are given. For explanation of two-way ANOVAs used see the caption of Table 3. For trichome density, surface effects were evaluated by one-way ANOVAs for plants from the semi-alpine habitat Traits Leaf type Surface Habitat Two-way ANOVA Understorey Semi-alpine Habitat Surface Interaction (A) Trichome density (mm−2) Cauline leaf AD 0.0 ± 0 25.6 ± 1.6 AB 0.0 ± 0 33.8 ± 2.3 – 9.03** – Post-cold rosette leaf AD 0.0 ± 0 17.7 ± 0.8 AB 0.0 ± 0 19.9 ± 1.1 – 2.49ns – Pre-cold rosette leaf AD 0.0 ± 0 16.7 ± 0.3 AB 0.0 ± 0 16.8 ± 0.9 – 0.006ns – (B) Stomatal density (mm−2) Cauline leaf AD 89.3 ± 3.5 105.2 ± 3.6 AB 101.6 ± 3.2 103.7 ± 3.0 7.20** 2.62ns 4.25* Post-cold rosette leaf AD 94.2 ± 3.0 87.7 ± 4.2 AB 101.0 ± 3.1 94.5 ± 3.2 3.57ns 3.91ns 0.00ns Pre-cold rosette leaf AD 45.6 ± 2.3 55.6 ± 3.3 AB 74.5 ± 2.8 68.3 ± 2.7 0.44ns 56.43*** 8.49** Traits Leaf type Surface Habitat Two-way ANOVA Understorey Semi-alpine Habitat Surface Interaction (A) Trichome density (mm−2) Cauline leaf AD 0.0 ± 0 25.6 ± 1.6 AB 0.0 ± 0 33.8 ± 2.3 – 9.03** – Post-cold rosette leaf AD 0.0 ± 0 17.7 ± 0.8 AB 0.0 ± 0 19.9 ± 1.1 – 2.49ns – Pre-cold rosette leaf AD 0.0 ± 0 16.7 ± 0.3 AB 0.0 ± 0 16.8 ± 0.9 – 0.006ns – (B) Stomatal density (mm−2) Cauline leaf AD 89.3 ± 3.5 105.2 ± 3.6 AB 101.6 ± 3.2 103.7 ± 3.0 7.20** 2.62ns 4.25* Post-cold rosette leaf AD 94.2 ± 3.0 87.7 ± 4.2 AB 101.0 ± 3.1 94.5 ± 3.2 3.57ns 3.91ns 0.00ns Pre-cold rosette leaf AD 45.6 ± 2.3 55.6 ± 3.3 AB 74.5 ± 2.8 68.3 ± 2.7 0.44ns 56.43*** 8.49** AD, adaxial; AB, abaxial. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant at P < 0.05. View Large Table 4. Trichome density and stomatal density of cauline leaves and post- and pre-cold rosette leaves of Arabidopsis halleri subsp. gemmifera plants grown in growth chambers. Seeds were derived from two habitat types [understorey (430 and 580 m) and semi-alpine (1220 and 1370 m)]. Average ± s.e. (n = 36) for adaxial (AD) and abaxial (AB) surfaces are given. For explanation of two-way ANOVAs used see the caption of Table 3. For trichome density, surface effects were evaluated by one-way ANOVAs for plants from the semi-alpine habitat Traits Leaf type Surface Habitat Two-way ANOVA Understorey Semi-alpine Habitat Surface Interaction (A) Trichome density (mm−2) Cauline leaf AD 0.0 ± 0 25.6 ± 1.6 AB 0.0 ± 0 33.8 ± 2.3 – 9.03** – Post-cold rosette leaf AD 0.0 ± 0 17.7 ± 0.8 AB 0.0 ± 0 19.9 ± 1.1 – 2.49ns – Pre-cold rosette leaf AD 0.0 ± 0 16.7 ± 0.3 AB 0.0 ± 0 16.8 ± 0.9 – 0.006ns – (B) Stomatal density (mm−2) Cauline leaf AD 89.3 ± 3.5 105.2 ± 3.6 AB 101.6 ± 3.2 103.7 ± 3.0 7.20** 2.62ns 4.25* Post-cold rosette leaf AD 94.2 ± 3.0 87.7 ± 4.2 AB 101.0 ± 3.1 94.5 ± 3.2 3.57ns 3.91ns 0.00ns Pre-cold rosette leaf AD 45.6 ± 2.3 55.6 ± 3.3 AB 74.5 ± 2.8 68.3 ± 2.7 0.44ns 56.43*** 8.49** Traits Leaf type Surface Habitat Two-way ANOVA Understorey Semi-alpine Habitat Surface Interaction (A) Trichome density (mm−2) Cauline leaf AD 0.0 ± 0 25.6 ± 1.6 AB 0.0 ± 0 33.8 ± 2.3 – 9.03** – Post-cold rosette leaf AD 0.0 ± 0 17.7 ± 0.8 AB 0.0 ± 0 19.9 ± 1.1 – 2.49ns – Pre-cold rosette leaf AD 0.0 ± 0 16.7 ± 0.3 AB 0.0 ± 0 16.8 ± 0.9 – 0.006ns – (B) Stomatal density (mm−2) Cauline leaf AD 89.3 ± 3.5 105.2 ± 3.6 AB 101.6 ± 3.2 103.7 ± 3.0 7.20** 2.62ns 4.25* Post-cold rosette leaf AD 94.2 ± 3.0 87.7 ± 4.2 AB 101.0 ± 3.1 94.5 ± 3.2 3.57ns 3.91ns 0.00ns Pre-cold rosette leaf AD 45.6 ± 2.3 55.6 ± 3.3 AB 74.5 ± 2.8 68.3 ± 2.7 0.44ns 56.43*** 8.49** AD, adaxial; AB, abaxial. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant at P < 0.05. View Large Stomatal density was significantly higher in leaves from the semi-alpine habitat than in the understorey for cauline leaves, but not rosettes leaves (Table 4B). Although stomatal density on abaxial surfaces was significantly higher than that on adaxial surfaces of pre-cold rosette leaves, there was no habitat effect in the two-way ANOVA (Table 4B). No significant habitat or surface effects were detected for stomatal index for all three types of leaves (Supplementary Data 1, Table S4A). For other stomatal characters, i.e. guard cell length and pore length, significant habitat effects were detected in guard cell length for pre- and post-cold rosette leaves (Supplementary Data 1, Table S4B) and in pore length for cauline and post-cold rosette leaves (Supplementary Data 1, Table S4C). Gene expression analysis Expression levels of AhgCER1 were significantly higher in cauline leaves of plants from semi-alpine habitats, i.e. 1220, 1300 and 1370 m than in understorey leaves (Fig. 5). Expression was low in cauline leaves for plants from the low-altitude habitat and rosette leaves for both habitats (Fig. 5). The results were similar for AhgCER1/AhgACT2 (Fig. 5) and AhgCER1/AhgPP2AA3 (Supplementary Data 1, Fig. S2). The patterns were statistically supported by high significance in all terms in the two-way ANOVAs (Supplementary Data 1, Table S5). Fig. 5. View largeDownload slide AhgCER1 relative expression (AhgCER1/AhgACT2) in cauline and rosette leaves collected from different populations [430 m (understorey); 1220, 1300, 1370 m (semi-alpine)]. Grey and black boxes represent cauline and rosette leaves, respectively. Averages and standard deviations (error bars) are shown. Fig. 5. View largeDownload slide AhgCER1 relative expression (AhgCER1/AhgACT2) in cauline and rosette leaves collected from different populations [430 m (understorey); 1220, 1300, 1370 m (semi-alpine)]. Grey and black boxes represent cauline and rosette leaves, respectively. Averages and standard deviations (error bars) are shown. DISCUSSION In both field observations and growth experiments with A. halleri subsp. gemmifera populations along an altitudinal gradient, wettability of cauline leaves was significantly decreased in semi-alpine habitats compared with the low-altitude understorey. In the semi-alpine habitats, adaxial surfaces of cauline leaves were non-wettable. Furthermore, abaxial surfaces of cauline leaves became highly non-wettable in growth experiments. Our growth experiments were designed to detect genetically based phenotypic differences by growing plants from seeds in a common set of environments. The results of the experiments indicated that the difference in leaf wettability between the two distinct altitudinal habitats had a genetic basis, at least in part. At the level of phenotypic variation, intraspecific altitudinal variation in leaf wettability has been demonstrated in previous studies, and leaf wettability generally decreased as altitude increased (Aryal and Neuner, 2010, 2012). In the field site, environmental factors are likely to be critical determinants of leaf wettability (Fogg, 1947; Weiss, 1988). We found that wettability of leaves was generally higher in the growth experiments than in the field measurements, probably because the surface of leaves had been exposed to various conditions, including sun, rain and snow in the natural habitats. We should note that the altitudinal differentiation of leaf wettability was specific to cauline leaves. Cauline leaves of the semi-alpine plants had the genetically fixed non-wettable character. This led us to assume the existence of a specific function of non-wettable cauline leaves that only occurs in the semi-alpine habitat. Because cauline leaves cover a flowering bud at the centre of rosettes in the early stage of bolting, we suspect that low wettability is required for young cauline leaves to serve for protection of flowering buds in very early spring. At the summit of Mt Ibuki, monthly minimum temperature (± s.d.) averaged −10.7 ± 1.1 °C, −6.6 ± 1.6 °C and 0.3 ± 1.9 °C in March, April and May, respectively (1971–2000; Japan Meteorological Agency). Monthly maximum wind velocity averaged 28.5 ± 5.2, 29.3 ± 6.3 and 27.5 ± 5.6 m s−1 in March, April and May, respectively. We suspect that highly non-wettable surfaces reduce frost damage in cauline leaves and the flower buds inside. The lower wettability of the abaxial surface of cauline leaves relative to that of the adaxial surface supports this idea, because the abaxial surface is exposed during the early stage of bolting when cauline leaves cover flower buds. In temperate zones, there is a risk of spring frost, which can cause critical damage to flowers or buds (Rodrigo, 2000). Previous studies also suggested that spring damage of flowers occurs more often at the cold margins of plant distributions (Charrier et al., 2015). Additionally, it has been demonstrated that low leaf wettability creates a frost-protective effect by experimentally spraying tomato plants with a hydrophobic kaolin particle film (Wisniewski et al., 2002). We suspect that the cause of cauline-leaf-specific low wettability of the semi-alpine plants is a difference in the wax characteristics, such as content, composition and microstructure, of leaf surfaces because we found higher expression of AhgCER1, a CER1 homologue in the study species, in the corresponding leaves. CER1 was first verified to be responsible for the biosynthesis of highly hydrophobic molecules in the epidermal wax of A. thaliana (Aarts et al., 1995; Jenks et al., 1995). More recently, upregulation of CER1 was reported to increase the alkane content of A. thaliana leaves, and to improve the cuticle barrier properties against biotic and abiotic stress (Kosma et al., 2009; Bourdenx et al., 2011). Kubota et al. (2015) conducted genome-wide scans for populations of A. halleri subsp. gemmifera along the altitudinal gradient of Mt Ibuki and identified genes that contained SNPs with altitudinal differentiation in allele frequencies. AhgCER1 and other wax-related genes were not listed in their analyses. Further analyses are required to identify the responsible nucleotide substitutions underlying the altitudinal differentiation of AhgCER1 expression in cauline leaves. The angle of droplet retention of leaves in both the field and the growth experiment was low (<20°), indicating that all investigated leaf types had less sticky surface on which water droplets could easily move. Run-off of water droplets depends on the stickiness of the surface and the leaf inclination. In our case, rosette leaves were horizontally arranged, whereas cauline leaves were vertically arranged. It is likely that droplets would run off the leaf surface immediately after contact on vertically angled cauline leaves from semi-alpine habitats, which have been found to be non-wettable. A low retention angle may also be important in providing dry surfaces for cauline leaves in the semi-alpine habitat. Cauline and rosette leaves of understorey plants had no trichomes, but the trichome density was high on both surfaces of leaves from the semi-alpine habitat. For rosette leaves, the substantial difference in trichome density, i.e. presence or absence of trichomes, between understorey and semi-alpine plants did not correspond to the similarity in leaf wettability between the habitats. It has been reported that trichomes on leaf surfaces enhance water repellence by the formation of spherical droplets (Brewer et al., 1991; Brewer and Smith, 1997; Pandey and Nagar, 2003; Aryal and Neuner, 2010), whereas another study has reported that a high density of trichomes holds water droplets more effectively than does a medium or low density (Wang et al., 2015). High trichome formation in cauline leaves of semi-alpine plants may contribute to low leaf wettability of these leaves if the trichome surface is hydrophobic (Neinhuis and Barthlott, 1997). The stomatal density of cauline and rosette leaves increased progressively from understorey to semi-alpine habitats in field-collected samples. Similar altitudinal patterns have been reported in other species (Aryal and Neuner, 2012). Our growth chamber experiments indicated that the difference was not the result of genetic differentiation. In A. halleri, the altitudinal changes in stomata density represent a phenotypic response. Because the stomatal index was constant between habitats in the growth experiment, the plasticity of leaf size may result in the altitudinal gradients of stomatal density per unit leaf area. In conclusion, we identified genetic differentiation of leaf wettability between distinctive montane habitats in the rosette-forming perennial Arabidopsis. To our knowledge, this is the first report of genetic differentiation in leaf wettability between natural plant populations within a single species. The lower wettability specific to cauline leaves of the semi-alpine plants is supportive of the postulate that the dry surface of cauline leaves is necessary for the protection of floral buds from frost under exposure to prevailing cold wind. The theory needs to be tested in future field and experimental studies. We also identified a candidate gene that could explain the differential leaf wettability between cauline and rosette leaves and between cauline leaves from the two habitats. A genome-wide study on segregated populations derived from crosses between plants from the two habitats will likely reveal the identity of the responsible SNP(s) underlying the discovery in this study. SUPPLEMENTARY DATA Supplementary data are available online at https://academic.oup.com/aob and consist of the following. Supplementary Data 1 consists of Figure S1: scanning electron microscopic images of the adaxial surfaces of cauline leaves and adaxial and abaxial surfaces of rosette leaves. Figure S2: relative expression of AhgCER1 in cauline and rosette leaves. Table S1: list of primer sequences used in this study. Table S2: results of Kolmogorov–Smirnov tests for data normality. Table S3: stomata-related leaf characteristics of cauline and rosette leaves in spring and rosette leaves in autumn of Arabidopsis halleri subsp. gemmifera collected from five different populations. Table S4: stomata-related leaf characteristics of cauline leaves and post- and pre-cold rosette leaves of Arabidopsis halleri subsp. gemmifera plants grown in growth chambers. Table S5: results of two-way ANOVAs for relative gene expression of AhgCER1 in cauline and rosette leaves of A. halleri subsp. gemmifera collected from the field. Supplementary Data 2, 3 and 4 are the original data sets of this study for field measurements, growth experiments and gene expression analyses, respectively. ACKNOWLEDGEMENTS We thank Professor K. Agata, who provided the CCD camera. This work was supported by the Japan Society for Promotion of Science (grant KAKENHI JP26221106) and the Japan Society and Technology Agency (grant CREST JPMJCR15O1) to H.K. LITERATURE CITED Aarts MG , Keijzer CJ , Stiekema WJ , Pereira A . 1995 . Molecular characterization of the CER7 gene of Arabidopsis involved in epicuticular wax biosynthesis and pollen fertility . Plant Cell 7 : 2115 – 2127 . Google Scholar CrossRef Search ADS PubMed Adam NK . 1963 . Principles of water-repellency . In: Moilliet JL , ed. Waterproofing and water-repellency . Amsterdam : Elsevier , 1 – 23 . Aikawa S , Kobayashi MJ , Satake A , Shimizu KK , Kudoh H . 2010 . Robust control of the seasonal expression of the Arabidopsis FLC gene in a fluctuating environment . Proceedings of the National Academy of Sciences of the USA 107 : 11632 – 11637 . Google Scholar CrossRef Search ADS PubMed Aryal B , Neuner G . 2010 . Leaf wettability decreases along an extreme altitudinal gradient . Oecologia 126 : 1 – 9 . Google Scholar CrossRef Search ADS Aryal B , Neuner G . 2012 . Leaf wettability in bilberry Vaccinium myrtillus L. as affected by altitude and openness of the growing site . Phyton 52 : 245 – 262 . Barthlott W , Neinhuis C . 1997 . Purity of sacred lotus, or escape from contamination in biological surfaces . Planta 202 : 1 – 8 . Google Scholar CrossRef Search ADS Bhushan B , Jung YC . 2008 . Wetting, adhesion and friction of superhydrophobic and hydrophilic leaves and fabricated micro/nanopatterned surfaces . Journal of Physics: Condensed Matter 20 : 1 – 24 . Bourdenx B , Bernard A , Domergue F et al. 2011 . Overexpression of Arabidopsis ECERIFERUM1 promotes wax very-long-chain alkane biosynthesis and influences plant response to biotic and abiotic stresses . Plant Physiology 156 : 29 – 45 . Google Scholar CrossRef Search ADS PubMed Brewer CA , Nuñez CI . 2007 . Patterns of leaf wettability along an extreme moisture gradient in western Patagonia, Argentina . International Journal of Plant Sciences 168 : 555 – 562 . Google Scholar CrossRef Search ADS Brewer CA , Smith K . 1997 . Patterns of leaf surface wetness for montane and subalpine plants . Plant, Cell & Environment 20 : 1 – 11 . Google Scholar CrossRef Search ADS Brewer CA , Smith WK , Vogelmann TC . 1991 . Functional interaction between leaf trichomes, leaf wettability and the optical properties of water droplets . Plant, Cell & Environment 14 : 955 – 962 . Google Scholar CrossRef Search ADS Challen SB . 1960 . The contribution of surface characteristics to the wettability of leaves . Journal of Pharmacy and Pharmacology 12 : 307 – 311 . Google Scholar CrossRef Search ADS PubMed Charrier G , Ngao J , Saudreau M , Améglio T . 2015 . Effects of environmental factors and management practices on microclimate, winter physiology, and frost resistance in trees . Frontiers in Plant Science 6 : 1 – 18 . Google Scholar PubMed Crisp DJ . 1963 . Waterproofing in animals and plants . In: Moilliet JL , ed. Waterproofing and water-repellency . Amsterdam : Elsevier , 416 – 481 . Fogg GE . 1947 . Quantitative studies on the wetting of leaves by water . Proceedings of the Royal Society B 134 : 503 – 522 . Google Scholar CrossRef Search ADS Fogg GE . 1948 . Adhesion of water to the external surfaces of leaves . Faraday Discussions of the Chemical Society 3 : 162 – 166 . Google Scholar CrossRef Search ADS Gonzalo-Turpin H , Hazard L . 2009 . Local adaptation occurs along altitudinal gradient despite the existence of gene flow in the alpine plant species . Journal of Ecology 97 : 742 – 751 . Google Scholar CrossRef Search ADS Holder CD . 2011 . The relationship between leaf water repellency and leaf traits in three distinct biogeographical regions . Plant Ecology 212 : 1913 – 1926 . Google Scholar CrossRef Search ADS Holloway PJ . 1969 . Chemistry of leaf waxes in relation to wetting . Journal of the Science of Food and Agriculture 20 : 124 – 128 . Google Scholar CrossRef Search ADS Holloway PJ . 1970 . Surface factors affecting the wetting of leaves . Pesticide Science 1 : 156 – 163 . Google Scholar CrossRef Search ADS Ikeda H , Setoguchi H , Morinaga S . 2010 . Genomic structure of lowland and highland ecotypes of Arabidopsis halleri subsp. gemmifera (Brassicaceae) on Mt Ibuki . Acta Phytotaxonomica et Geobotanica 61 : 21 – 26 . Jenks MA , Tuttle HA , Eigenbrode SD , Fieldmann KA . 1995 . Leaf epicuticular waxes of the eceriferum mutants in Arabidopsis . Plant Physiology 108 : 369 – 377 . Google Scholar CrossRef Search ADS PubMed Jetter R , Kunst L . 2008 . Plant surface lipid biosynthetic pathways and their utility for metabolic engineering of waxes and hydrocarbon biofuels . Plant Journal 54 : 670 – 683 . Google Scholar CrossRef Search ADS PubMed Kosma DK , Bourdenx B , Bernard A et al. 2009 . The impact of water deficiency on leaf cuticle lipids of Arabidopsis . Plant Physiology 151 : 1918 – 1929 . Google Scholar CrossRef Search ADS PubMed Kubota S , Iwasaki T , Hanada K et al. 2015 . A genome scan for genes underlying microgeographic-scale local adaptation in a wild Arabidopsis species . PLoS Genetics 11 : 1 – 26 . Neinhuis C , Barthlott W . 1997 . Characterization and distribution of water repellent, self cleaning plant surfaces . Annuals of Botany 79 : 667 – 677 . Google Scholar CrossRef Search ADS Nishio H , Buzas DM , Nagano AJ et al. 2016 . From the laboratory to the field: assaying histone methylation at FLOWERING LOCUS C in naturally growing Arabidopsis halleri . Genes & Genetic Systems 91 : 15 – 26 . Google Scholar CrossRef Search ADS PubMed Pandey S , Nagar PK . 2003 . Pattern of leaf surface wetness in some important medicinal and aromatic plants of western Himalaya . Flora 198 : 349 – 357 . Google Scholar CrossRef Search ADS Rodrigo J . 2000 . Spring frosts in deciduous fruit trees—morphological damage and flower hardiness . Scientia Horticulturae 85 : 155 – 173 . Google Scholar CrossRef Search ADS Salisbury EJ . 1927 . On the causes and ecological significance of stomatal frequency with special reference to the woodland flora . Philosophical Transactions of the Royal Society of London B 216 : 1 – 65 . Google Scholar CrossRef Search ADS Savolainen O , Lascoux M , Merliä J . 2013 . Ecological genomics of local adaptation . Nature Reviews Genetics 14 : 807 – 820 . Google Scholar CrossRef Search ADS PubMed Schönherr J , Bukovac MJ . 1972 . Penetration of stomata by liquids . Plant Physiology 49 : 813 – 819 . Google Scholar CrossRef Search ADS PubMed Smith WK , McClean TM . 1989 . Adaptive relationship between leaf water repellency, stomatal distribution, and gas exchange . American Journal of Botany 76 : 465 – 469 . Google Scholar CrossRef Search ADS Wang H , Shi H , Wang Y . 2015 . The wetting of leaf surfaces and its ecological significances . In: Aliofkhazraei M , ed. Wetting and wettability . Rijeka, Croatia : InTech , 295 – 321 . Google Scholar CrossRef Search ADS Warburton FL . 1963 . The effect of structure on waterproofing . In: Moilliet JL , ed. Waterproofing and water-repellency . Amsterdam : Elsevier , 24 – 51 . Weiss A . 1988 . Contact angle of water droplets in relation to leaf water potential . Agricultural and Forest Meteorology 43 : 251 – 259 . Google Scholar CrossRef Search ADS Wisniewski M , Glenn DM , Fuller MP . 2002 . Use of a hydrophobic particle film as a barrier to extrinsic ice nucleation in tomato plants . Journal of American Society of Horticulture Science 127 : 358 – 364 . © The Author(s) 2018. Published by Oxford University Press on behalf of the Annals of Botany Company. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Annals of Botany Oxford University Press

Genetic differentiation in cauline-leaf-specific wettability of a rosette-forming perennial Arabidopsis from two contrasting montane habitats

Annals of Botany , Volume Advance Article (7) – Mar 20, 2018

Loading next page...
1
 
/lp/ou_press/genetic-differentiation-in-cauline-leaf-specific-wettability-of-a-tW3ClV5wg3

References (44)

Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the Annals of Botany Company. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
0305-7364
eISSN
1095-8290
DOI
10.1093/aob/mcy033
Publisher site
See Article on Publisher Site

Abstract

Abstract Background and Aims An altitudinal gradient of leaf wettability is often observed between and within species. To understand its functional significance, positional variation of leaf surfaces within plants should be taken into account. In rosette-forming plants, rosette leaves are near the ground and their adaxial surfaces are exposed, whereas cauline leaves are lifted from the ground throughout the reproductive season, and their abaxial surfaces are more exposed. Here, we investigated leaf wettability of cauline and rosette leaves of Arabidopsis halleri subsp. gemmifera growing in contrasting montane habitats along an altitudinal gradient at Mt Ibuki, Japan. Methods We conducted field investigations and a growth chamber experiment to determine whether field-observed variation in leaf wettability was caused by genetic differentiation. We further performed gene expression analysis of a wax-related gene, i.e. AhgCER1, a homologue of A. thaliana ECERIFERUM1 (CER1) that may be involved in differentiation of leaf wettability. Key Results We found cauline-leaf specific genetic differentiation in leaf wettability between contrasting montane habitats. Cauline leaves of semi-alpine plants, especially on abaxial surfaces, were non-wettable. Cauline leaves of low-altitudinal understorey plants were wettable, and rosette leaves were also wettable in both habitats. AhgCER1 expression corresponded to observed leaf wettability patterns. Conclusions Low wettability of cauline leaves is hypothesized to keep exposed surfaces dry when they are wrapping flowering buds in early spring, and presumably protects flowering buds from frost damage. The genetic system that controls wax content, specifically for cauline leaves, should be involved in the observed genetic differentiation, and AhgCER1 control is a strong candidate for the underlying genetic mechanism. Altitudinal gradient, Arabidopsis halleri subsp. gemmifera, cauline leaf, CER1, cuticular wax, genetic differentiation, leaf wettability, local adaptation, rosette leaf INTRODUCTION Local adaptation, i.e. genetic differentiation in fitness-related traits between populations, is one of the major interests of evolutionary biologists (Savolainen et al., 2013). Adaptive differentiation between populations occurring along altitudinal gradients has been frequently studied because it provides contrasting habitats within a short geographic distance (Gonzalo-Turpin and Hazard, 2009). Along an altitudinal gradient, plants are exposed to deposition of surface water by various forms of precipitation such as rain, ground fog, cloud mist and dew. The consequence of water deposition may vary depending on ambient temperature and the degree of exposure to surrounding environments. The wettability of leaves, i.e. the amount of water captured and retained on leaf surfaces, differs significantly between plant species (Fogg, 1947; Brewer et al., 1991; Aryal and Neuner, 2010, 2012). Different plant species show a broad range of leaf wettability from being covered by a film of water to being completely water-repellent (Smith and McClean, 1989; Brewer et al., 1991; Brewer and Smith, 1997; Brewer and Nuñez, 2007; Aryal and Neuner, 2010, 2012). Along altitudinal gradients, associations with lower wettability in alpine habitats have been reported both between and within species (Aryal and Neuner, 2010, 2012). The functional significance of low or high leaf wettability appears to be highly varied (Brewer et al., 1991), i.e. it might relate to water supply, pathogen infection, extrinsic ice nucleation and stomatal occlusion (Barthlott and Neinhuis, 1997). The rosette is one of the major forms of herbaceous plants that experience harsh, cold and drought during vegetative growth. Rosette-forming plants are especially interesting in studying leaf wettability, because they produce leaves that have distinctive modes of exposure to atmospheric environments. Rosette leaves (leaves that form a rosette) are successively produced during vegetative growth. Prior to flowering seasons, cauline leaves (leaves on flowering stems) are formed from the shoot apical meristem. Cauline leaves are very compact compared with rosette leaves, but they serve as protection for flowering buds in the initial stages of elongation of flowering stems. Rosette leaves remain near the ground and their adaxial surfaces are more exposed, whereas cauline leaves are lifted from the ground throughout the reproductive season and their abaxial surface is more exposed. Therefore, these positional differences may require distinctive levels of leaf wettability for adaxial and abaxial surfaces of rosette and cauline leaves. The main components of epicuticular wax have hydrophobic properties (Holloway, 1969), suggesting that the physiochemistry of the cuticular wax (e.g. wax content, composition and microstructure) influences leaf surface wettability (Neinhuis and Barthlott, 1997; Bhushan and Jung, 2008; Holder, 2011). The genetic systems controlling wax content have been extensively studied in Arabidopsis thaliana (Aarts et al., 1995; Kosma et al., 2009; Bourdenx et al., 2011). In A. thaliana, alkanes represented 70 % of the total wax in leaves (Jetter and Kunst, 2008). The high proportion of alkanes in the epicuticular wax results in hydrophobic plant surfaces (Holloway, 1969). ECERIFERUM1 (CER1) encodes proteins involved in the biosynthesis of the alkane-forming enzyme in A. thaliana (Aarts et al., 1995), and was reported to be more highly expressed in cauline leaves than rosette leaves in A. thaliana (Bourdenx et al., 2011). We do not know, however, how the genetic control of epicuticular wax differentiates along altitudinal gradients in natural systems. In the present study, we compared leaf wettability of Arabidopsis halleri subsp. gemmifera, a rosette-forming perennial, at contrasting altitudinal habitats within a mountain system. The species is a perennial relative of A. thaliana that has been used in studies on the adaptive significance of gene functions under natural conditions (e.g. Aikawa et al., 2010; Kubota et al., 2015). The study species occurs in two distinctive habitats at Mt Ibuki, Japan; i.e. the understorey of Cryptomeria japonica forests near the base of the mountain (referred to as understorey hereafter) and the semi-alpine open habitat near the peak (referred to as semi-alpine hereafter). Although the peak is not very high (1377 m), the mountain is characterized by a combination of exposure to strong cold winds and snow near the top and relatively warm and mild winters near the base. In this paper we pose the following questions regarding the variation in leaf wettability among plants that occur along a steep environmental gradient. Leaf wettability was assessed in both the field and a laboratory growth experiment to quantify leaf wettability in natural habitats and the degree of genetic differentiation. (1) Is there detectable genetic differentiation in leaf wettability between populations occurring in contrasting habitats in the field? (2) Do the patterns correspond to the degree of exposure of the surface of rosette and cauline leaves? (3) Do the differences in leaf wettability correspond with gene expression of a homologue of CER1 (hereafter referred to as AhgCER1)? By answering these three questions, we will be able to determine how genetics, leaf position and field environments interact to determine leaf wettability. MATERIALS AND METHODS Study species and sites The study plant, Arabidopsis halleri subsp. gemmifera, is distributed in East Asia, including north-eastern China, Japan, Korea and Taiwan, and in the Russian Far East. The study was conducted using natural populations of A. halleri subsp. gemmifera occurring in the understorey and semi-alpine habitats of Mt Ibuki, Japan (35°25′04″ N, 136°24′22″ E at the highest peak; Fig. 1). The mountain has a hiking route to the top that is divided into ten pitches. Along the route, we selected five populations, two from the understorey [430 m and 580 m above sea level (a.s.l.)] and three from the semi-alpine area [1220 m, 1300 m and 1370 m a.s.l.] habitats (Fig. 1). Plants in the semi-alpine habitat of Mt Ibuki are distinct from typical A. halleri plants in that they have dense trichomes on both rosette and cauline leaves, and are sometimes treated as a variety of A. halleri subsp. gemmifera (originally described as Arabis gemmifera var. alpicola). The reported level of genome-wide genetic differentiation between the low- and high-altitude populations on Mt Ibuki was low [FST = 0.017, amplified fragment length polymorphism (AFLP) (Ikeda et al., 2010); GST = 0.043–0.048, genome-wide single-nucleotide polymorphism (SNP) (Kubota et al., 2015)]. Fig. 1. View largeDownload slide Maps showing locations of five sampling populations in understorey (430 and 580 m) and semi-alpine (1220, 1300 and 1370 m) habitats at Mt Ibuki (left) and the geographical location of Mt Ibuki in Japan (upper right). Closed circles indicate the location of the five sites along a hiking trail (thick line). Contour lines in the map indicate altitudes above sea level (m). Corresponding latitudes (N) and longitudes (E) are shown on the left and bottom margins Fig. 1. View largeDownload slide Maps showing locations of five sampling populations in understorey (430 and 580 m) and semi-alpine (1220, 1300 and 1370 m) habitats at Mt Ibuki (left) and the geographical location of Mt Ibuki in Japan (upper right). Closed circles indicate the location of the five sites along a hiking trail (thick line). Contour lines in the map indicate altitudes above sea level (m). Corresponding latitudes (N) and longitudes (E) are shown on the left and bottom margins Sampling of field materials At the five study populations in the understorey (430 and 580 m) and semi-alpine (1220, 1300 and 1370 m) habitats (Fig. 1), we collected leaves of A. halleri subsp. gemmifera in autumn (20 November 2010) and spring (24 June 2011). During each spring and autumn sampling period, we randomly selected 40, 10, 30, 30 and 30 plants from the 430, 580, 1220, 1300 and 1370 m sites, respectively, depending on the availability of plants. In autumn, rosettes are formed by leaves that developed during summer and autumn, and there were no cauline leaves. In spring, plants possessed overwintered leaves on rosettes and newly developed cauline leaves on elongated flowering stalks. From each plant, we collected one rosette leaf in autumn and one rosette and one cauline (6th- to 8th-position leaves, counted basipetally from the top) leaves in spring. All collected leaves were fresh and had little natural damage. Leaves were placed in plastic bags, kept in a cool box and taken to the laboratory for further measurements. For gene expression analysis, leaves were collected in spring (19 June 2017). One cauline leaf (6th- to 8th-position leaves counted basipetally from the top) and one rosette leaf were collected from the same individuals. In total, 12, 6, 6 and 6 sets of cauline and rosette leaves were collected from the 430, 1220, 1300 and 1370 m sites, respectively. Samples were washed in 0.1 % Triton X (0.5 mL) for 10 s and preserved in RNAlater (0.5 mL, Invitrogen, Thermo Fisher Scientific, MA, USA), kept in a cool box and taken to the laboratory. Samples were kept at 4 °C for 24 h and then transferred to −20 °C. Growth experiments During the fruiting season of 2012, seeds were collected from the study populations. Seeds were sown in Petri dishes with moistened quartz sand, placed in an incubator (day/night temperature 22 Cm/15 C5, 14 h light; Koitotron HNM-S, Koito, Osaka, Japan). Seedlings were transferred into plastics pots (75 mm in diameter and 65 mm in depth) containing vermiculite. A total of 72 plants, 36 from the understorey habitats (originated from six and four mother plants from the 430 and 580 m sites, respectively) and 36 from the semi-alpine habitats (originated from two and two mother plants from the 1220 and 1370 m sites, respectively), were prepared. The number of plants prepared per parental plant varied; therefore, results for the understorey and semi-alpine habitats were compared. We placed the pots in the incubator at 20 °C with 12 h light for 6 weeks prior to the first sampling of leaves (hereafter the pre-cold sampling). The light intensity of photosynthetically active radiation (PAR) averaged 65.4 ± 3.6 μm m−2 s−1 at the pot surface. We collected one rosette leaf per plant for all pots. Then, plants were transferred to a 5 °C (12 h light) condition for 6 weeks as a vernalization treatment to induce flowering. After vernalization, plants were again placed in the 20 °C condition (12 h light) for 4 weeks until plants possessed 15–20 flowers. Then, we conducted the second leaf sampling (hereafter post-cold sampling). One rosette and one cauline leaf were sampled from each plant. During the experiments, pots were watered regularly and fertilized once a week (20 times in total) with a solution of 0.1 % Hyponex (a liquid fertilizer, N:P:K = 6:10:5; Hyponex Japan, Osaka). In terms of whether the plants were in vegetative or reproductive phase, pre-cold and post-cold samplings corresponded with autumn and spring samplings in the natural habitats. Measurements For adaxial and abaxial surfaces of all sampled leaves from the field and the growth experiments, we measured leaf wettability, droplet retention, trichome density and density and size of stomata. The measurements were conducted using the leaf surface area where leaves were widest. Care was taken to avoid veins. Leaf wettability was measured as the contact angle (θ, Fig. 2) of a water droplet placed on the leaf surface following the procedure of Brewer et al. (1991). Leaf samples were mounted horizontally on an observation stage using double-sided tape. A 5-µL droplet (recommended standard volume in Brewer et al. 1991) of distilled water was then placed on the leaf surface using a micropipette. A digital photographic image of the horizontal view of the droplet on the leaf surface was taken with a CCD camera connected to a digital microscope (Keyence Japan, Osaka). The digital photographs were processed with ImageJ software (National Institutes of Health, MD, USA). The contact angle (θ) of a line tangent to the water droplet through the point of contact between the droplet and leaf surface was measured by using the ‘tangent 1 default method’ according to Brewer et al. (1991) (Fig. 2). The larger the value of θ, the more repellent the leaf surface; thus, the criteria for judging surface wettability depended on θ. Leaves were classified as highly wettable if θ was <90°, based on previous reports (Fogg, 1948; Challen, 1960; Adam, 1963; Crisp, 1963; Warburton, 1963; Holloway, 1970; Schönherr and Bukovac, 1972) and wettable if θ was <110° (Crisp, 1963). If θ exceeded 110° the leaf was considered non-wettable. Highly non-wettable leaves had θ values >130° (Crisp, 1963). To determine wax deposition on the leaf surfaces, we examined leaves under a scanning electron microscope (Philips XL 30, Eindhoven, The Netherlands). Fig. 2. View largeDownload slide Contact angle (θ) is the angle between the surface of the leaf and the line tangent to the droplet at the point of contact between air, water and the leaf surface. Leaf wettability is categorized into four classes based on θ values (see text for details of categorization). Fig. 2. View largeDownload slide Contact angle (θ) is the angle between the surface of the leaf and the line tangent to the droplet at the point of contact between air, water and the leaf surface. Leaf wettability is categorized into four classes based on θ values (see text for details of categorization). Droplet retention is a measure of the ‘stickiness’ of a leaf surface. We determined water retention for adaxial and abaxial surfaces of sampled leaves using the method of Brewer and Smith (1997). Leaves were mounted on a metal plate fixed to a goniometer (Technical Support Division, CER, Kyoto University). A 50-µL droplet (recommended standard volume in Brewer and Smith, 1997) was initially placed on a horizontal leaf surface. Then the angle of leaf inclination was successively increased, and when the droplet began to move the angle of inclination was recorded. Higher angles (>60°) indicated a greater tendency to retain droplets, whereas lower angles (<20°) indicated leaf surfaces that readily shed droplets. We counted the number of stomata and trichomes using Suzuki’s universal micro-printing (SUMP) method (Kenis, Osaka, Japan). Micro-printings were digitized using an optical microscope. We recorded the number of stomata and trichomes within a 0.0625-mm−2 area for both adaxial and abaxial leaf surfaces of each leaf sample and calculated the number of stomata or trichome per unit of leaf area (mm2). The stomatal index was calculated according to the method of Salisbury (1927) as the ratio of number of stomata to number of epidermal cells. Guard cell length and pore length were measured using ImageJ software. Gene expression analysis For AhgCER1 quantification, total RNA was extracted from leaves with a Maxwell® 16 LEV Plant RNA Kit according to the manufacturer’s instructions (Promega, Fitchburg, WI, USA). The amount of extracted RNA was measured using a Quantus Fluorometer (Promega). We used 200 ng of RNA for cDNA synthesis using a High-Capacity cDNA Reverse Transcription Kit (Life Technologies). Quantitative PCR analyses were conducted using the cDNA samples. Two technical replicates were prepared for each sample. PCR conditions were 95 °C for 20 s followed by 40 cycles at 95 °C for 1 s and 60 °C for 20 s. The melting curve analyses were performed by gradually increasing the temperature (0.05 °C s−/) from 60 to 95 °C for 15 s. The PCR reactions were performed using a 10-μL volume containing 200 nm primers and Fast SYBR Green Master Mix (Applied Biosystems). We used AhgACT2 and AhgPP2AA3 as reference genes for normalization among samples (Nishio et al., 2016). Sequences of all primers used are shown in Supplementary Data 1, Table S1. We obtained standard curves using a dilution series of a standard A. halleri cDNA at 1, 1/5, 1/25, 1/125 and 1/625 for AhgACT2 and AhgPP2AA3 and 1, 1/6, 1/36, 1/216 and 1/1296 for AhgCER1 in all analyses. Statistical analysis We conducted two-way analysis of variance (ANOVA) tests and one-way nested ANOVAs for measured leaf wettability (contact angle and droplet retention), as well as stomatal density, stomatal index, guard cell length and pore length/unit area as field measurements for cauline and rosette leaves collected in spring and rosette leaves collected in autumn. For each leaf type, effects of leaf surface (abaxial and adaxial) and population (430, 580, 1220, 1300 and 1370 m) were tested using two-way ANOVAs. We divided the five populations into two altitudinal habitat types, understorey (430 and 580 m) and semi-alpine (1220, 1300 and 1370 m). Effects of between-habitat (understorey versus semi-alpine) and within-habitat variation (between populations) were tested using one-way nested ANOVAs (populations were nested within habitats) separately for each surface. In the understorey habitat (430 and 580 m), the leaves had no trichomes on either leaf surface. Therefore, for altitude effects on trichome density the two-way ANOVA tested effects across the three populations in the semi-alpine habitat (1220, 1300 and 1370 m) and there was no between-habitats term in the one-way nested ANOVAs. Similarly, for growth experiments, seeds were derived from two habitat types [understorey (430 and 580 m) and semi-alpine (1220 and 1370 m)]. For each leaf type, effects of leaf surface (abaxial and adaxial) and habitat (understorey and semi-alpine) were tested using two-way ANOVAs for leaf wettability (contact angle and droplet retention), as well as stomatal density, stomatal index, guard cell length and pore length/unit area of cauline and rosette leaves collected in post- and pre-cold conditions. We did not use a two-way ANOVA for trichome density, and surface effects were evaluated using a one-way ANOVA for plants from the semi-alpine habitat. For gene expression analysis, expression of AhgCER1 relative to that of reference genes was converted to a log2 (x + 1) scale. Effects of leaf type (cauline or rosette), population [understorey (430 m) and semi-alpine (1220, 1300 and 1370 m)] and their interaction were tested using two-way ANOVAs. All statistical analyses were conducted using SPSS (SPSS, Chicago, IL, USA). Prior to conducting ANOVA, we applied Kolmogorov–Smirnov tests, and 253 out of 258 data sets (combinations of experiment × leaf type × surface × habitat/population × trait) did not deviate significantly from normal distributions (Supplementary Data 1, Table S2). All data analysed in this study are available in Supplementary Data 2, 3 and 4 for field measurements, growth experiments and gene expression analyses, respectively. RESULTS Field measurements Cauline leaves from the understorey habitat at low altitudes (430 and 580 m) were highly wettable, as indicated by low θ values (average ranging from 56.1° to 67.5°) for both adaxial and abaxial leaf surfaces (Fig. 3A). However, average θ values were nearly 110° or higher for leaves from the semi-alpine habitat at high altitudes (1220, 1300 and 1370 m), i.e. the leaves were classified as non-wettable for both adaxial and abaxial surfaces (Fig. 3A). A highly significant population effect was detected in the two-way ANOVA. Similarly, there was a significant between-habitats effect, whereas no within-habitat effect was detected in the one-way nested ANOVA, suggesting that most of the variation was attributable to habitat differences (Table 1A). We detected no significant differences between adaxial and abaxial surfaces (Table 1A). For the rosette leaves from spring and autumn, the θ values (leaf wettability) were <110° for all populations (Fig. 3B, C). This indicated that all investigated populations exhibited increased leaf wettability of rosette leaves. Although we detected significant population effects in the two-way ANOVA, they were attributable to variation between populations within habitats (Table 1A). No-significant differences were found in θ values of the two leaf surfaces collected from spring and autumn rosette leaves (Fig. 3B, C, Table 1A). Fig. 3. View largeDownload slide Contact angle (A–C) and droplet retention (D–F) of cauline (A, D) and rosette leaves in spring (B, E) and autumn (C, F) collected in different populations [430, 580 m (understorey); 1220, 1300, 1370 m (semi-alpine)]. Cauline leaves exist only in spring on the flowering stalks. White and grey boxes represent adaxial and abaxial surfaces, respectively. Box plots show the median (line inside the box) and upper and lower quartiles (75 %, 25 %) and maximum and minimum values. (G, H) Scanning electron microscopic images of the abaxial surfaces of cauline leaves from understorey (G) and semi-alpine (H) habitats. Scale bar = 100 µm. Fig. 3. View largeDownload slide Contact angle (A–C) and droplet retention (D–F) of cauline (A, D) and rosette leaves in spring (B, E) and autumn (C, F) collected in different populations [430, 580 m (understorey); 1220, 1300, 1370 m (semi-alpine)]. Cauline leaves exist only in spring on the flowering stalks. White and grey boxes represent adaxial and abaxial surfaces, respectively. Box plots show the median (line inside the box) and upper and lower quartiles (75 %, 25 %) and maximum and minimum values. (G, H) Scanning electron microscopic images of the abaxial surfaces of cauline leaves from understorey (G) and semi-alpine (H) habitats. Scale bar = 100 µm. Table 1. Results of two-way ANOVA and one-way nested ANOVA for leaf wettability (contact angle and droplet retention) of cauline and rosette leaves in spring and rosette leaves in autumn of Arabidopsis halleri subsp. gemmifera collected from five populations in two habitat types. For each leaf type, effects of leaf surfaces and populations were tested by two-way ANOVA. Effects of between-habitat and within-habitat-between-population variation [understorey (430 and 580 m) versus semi-alpine (1220, 1300 and 1370 m)] were tested by one-way nested ANOVA (populations were nested within habitats) for each surface separately. Values are F ratios Trait Two-way ANOVA One-way nested ANOVA Population Surface Interaction Surface Between habitats Within habitat (A) Contact angle, θ (°) Cauline leaf 99.2*** 0.59ns 0.85ns AD 111*** 0.90 ns AB 174*** 0.93 ns Spring rosette leaf 13.3*** 0.21ns 0.91ns AD 4.10ns 2.64ns AB 0.56ns 7.27*** Autumn rosette leaf 4.90** 3.30ns 1.71ns AD 1.71ns 2.30ns AB 0.03ns 5.06** (B) Droplet retention (°) Cauline leaf 1.85ns 0.42ns 1.18ns AD 0.11 ns 1.62 ns AB 0.15 ns 1.86 ns Spring rosette leaf 4.06** 1.08ns 0.97ns AD 4.84ns 0.73ns AB 0.03ns 3.45* Autumn rosette leaf 3.18* 2.48ns 0.52ns AD 0.69ns 1.14ns AB 0.23ns 3.24* Trait Two-way ANOVA One-way nested ANOVA Population Surface Interaction Surface Between habitats Within habitat (A) Contact angle, θ (°) Cauline leaf 99.2*** 0.59ns 0.85ns AD 111*** 0.90 ns AB 174*** 0.93 ns Spring rosette leaf 13.3*** 0.21ns 0.91ns AD 4.10ns 2.64ns AB 0.56ns 7.27*** Autumn rosette leaf 4.90** 3.30ns 1.71ns AD 1.71ns 2.30ns AB 0.03ns 5.06** (B) Droplet retention (°) Cauline leaf 1.85ns 0.42ns 1.18ns AD 0.11 ns 1.62 ns AB 0.15 ns 1.86 ns Spring rosette leaf 4.06** 1.08ns 0.97ns AD 4.84ns 0.73ns AB 0.03ns 3.45* Autumn rosette leaf 3.18* 2.48ns 0.52ns AD 0.69ns 1.14ns AB 0.23ns 3.24* AD, adaxial; AB, abaxial. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant at P < 0.05. View Large Table 1. Results of two-way ANOVA and one-way nested ANOVA for leaf wettability (contact angle and droplet retention) of cauline and rosette leaves in spring and rosette leaves in autumn of Arabidopsis halleri subsp. gemmifera collected from five populations in two habitat types. For each leaf type, effects of leaf surfaces and populations were tested by two-way ANOVA. Effects of between-habitat and within-habitat-between-population variation [understorey (430 and 580 m) versus semi-alpine (1220, 1300 and 1370 m)] were tested by one-way nested ANOVA (populations were nested within habitats) for each surface separately. Values are F ratios Trait Two-way ANOVA One-way nested ANOVA Population Surface Interaction Surface Between habitats Within habitat (A) Contact angle, θ (°) Cauline leaf 99.2*** 0.59ns 0.85ns AD 111*** 0.90 ns AB 174*** 0.93 ns Spring rosette leaf 13.3*** 0.21ns 0.91ns AD 4.10ns 2.64ns AB 0.56ns 7.27*** Autumn rosette leaf 4.90** 3.30ns 1.71ns AD 1.71ns 2.30ns AB 0.03ns 5.06** (B) Droplet retention (°) Cauline leaf 1.85ns 0.42ns 1.18ns AD 0.11 ns 1.62 ns AB 0.15 ns 1.86 ns Spring rosette leaf 4.06** 1.08ns 0.97ns AD 4.84ns 0.73ns AB 0.03ns 3.45* Autumn rosette leaf 3.18* 2.48ns 0.52ns AD 0.69ns 1.14ns AB 0.23ns 3.24* Trait Two-way ANOVA One-way nested ANOVA Population Surface Interaction Surface Between habitats Within habitat (A) Contact angle, θ (°) Cauline leaf 99.2*** 0.59ns 0.85ns AD 111*** 0.90 ns AB 174*** 0.93 ns Spring rosette leaf 13.3*** 0.21ns 0.91ns AD 4.10ns 2.64ns AB 0.56ns 7.27*** Autumn rosette leaf 4.90** 3.30ns 1.71ns AD 1.71ns 2.30ns AB 0.03ns 5.06** (B) Droplet retention (°) Cauline leaf 1.85ns 0.42ns 1.18ns AD 0.11 ns 1.62 ns AB 0.15 ns 1.86 ns Spring rosette leaf 4.06** 1.08ns 0.97ns AD 4.84ns 0.73ns AB 0.03ns 3.45* Autumn rosette leaf 3.18* 2.48ns 0.52ns AD 0.69ns 1.14ns AB 0.23ns 3.24* AD, adaxial; AB, abaxial. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant at P < 0.05. View Large The angle of droplet retention, an index of the stickiness of a leaf surface, was typically <20°. This indicated that the majority of cauline leaves (Fig. 3D) and rosette leaves in both spring (Fig. 3E) and autumn (Fig. 3F) had low droplet retention capacity for both leaf surfaces. A cauline-leaf-specific difference between habitats was not observed in droplet retention (Fig. 3D–F). The results of the two-way ANOVA and nested ANOVA indicated that there were no significant effects for all terms in cauline leaves (Table 1B). Significant differences were observed in spring and autumn rosette leaves for the population term in the two-way ANOVA, but habitat differences were not significant in the nested ANOVA (Table 1B). Scanning electron microscopic images showed that there were no marked differences in surface texture on the abaxial surfaces of cauline leaves between understorey and semi-alpine habitats (Fig. 3G, H). Similarly, no marked and consistent differences were detected between understorey and semi-alpine plants for other surfaces of cauline and rosette leaves (Supplementary Data 1, Fig. S1). Plants from the understorey habitat had no trichomes on either surface of all types of leaves collected in spring and autumn (Table 2A). Semi-alpine leaves had trichomes on both surfaces. The trichome density of semi-alpine leaves was generally higher in cauline leaves (population average range 59.7–70.7 and 72.0–94.4 mm−2 for adaxial and abaxial surfaces, respectively) than rosette leaves from both spring and autumn (population average range 23.6–36.6 and 24.1–39.3 mm−2 for adaxial and abaxial surfaces, respectively; Table 2A). Adaxial and abaxial surfaces exhibited significant differences only for trichome density of cauline leaves in the two-way ANOVA (P < 0.05; Table 2A). Table 2. Trichome density and stomatal density of cauline and rosette leaves in spring and rosette leaves in autumn of Arabidopsis halleri subsp. gemmifera collected from five populations. Average ± s.e. (n = 10–40) for adaxial (AD) and abaxial (AB) surfaces are given. For explanations of the ANOVAs used see the caption of Table 1. Because no plants in the understorey habitat (430 and 580 m) had trichomes on either leaf surface, they were not included in the statistical analyses. Therefore, population effects in two-way ANOVA represent effects across three populations in the semi-alpine habitat, and there is no between-habitats term in the one-way nested ANOVAs for trichome density (shown as –) Leaf type Surface Populations Two-way ANOVA One-way nested ANOVA Understorey habitat (m) Semi-alpine habitat (m) 430 580 1220 1300 1370 Population Surface Inter-action Between habitats Within habitat (A) Trichome density (mm−2) Cauline leaf AD 0.0 ± 0 0.0 ± 0 61.9 ± 3.1 59.7 ± 3.4 70.7 ± 3.1 – 3.28* AB 0.0 ± 0 0.0 ± 0 78.0 ± 5.5 72.0 ± 2.8 94.4 ± 3.8 9.07ns 27.0* 1.20ns – 7.71** Spring rosette leaf AD 0.0 ± 0 0.0 ± 0 36.6 ± 2.6 30.5 ± 2.3 29.6 ± 2.2 – 2.64ns AB 0.0 ± 0 0.0 ± 0 38.4 ± 2.2 35.4 ± 1.6 39.3 ± 2.8 2.03ns 8.57ns 1.51ns – 0.79ns Autumn rosette leaf AD 0.0 ± 0 0.0 ± 0 30.0 ± 2.4 34.2 ± 2.2 23.6 ± 1.9 – 5.22** AB 0.0 ± 0 0.0 ± 0 33.6 ± 2.0 36.4 ± 2.6 24.1 ± 1.7 13.03*** 1.45 ns 0.25 ns – 8.13** (B) Stomatal density (mm−2) Cauline leaf AD 31.3 ± 1.3 37.3 ± 2.4 64.7 ± 2.2 65.6 ± 1.9 73.2 ± 2.1 60** 4.78** AB 38.8 ± 1.4 41.0 ± 2.4 94.0 ± 2.9 82.6 ± 2.2 88.1 ± 2.3 223*** 90.1*** 9.42*** 101** 4.48** Spring rosette leaf AD 34.0 ± 1.3 35.2 ± 2.7 65.0 ± 2.3 68.6 ± 2.6 69.5 ± 1.5 218*** 0.94ns AB 49.6 ± 1.9 38.9 ± 6.0 89.2 ± 2.5 88.0 ± 2.2 97.6 ± 2.6 161*** 120*** 4.84*** 75.64** 4.64** Autumn rosette leaf AD 46.0 ± 2.3 51.7 ± 4.9 83.4 ± 3.0 71.5 ± 2.9 62.8 ± 3.3 7.38 ns 8.68*** AB 60.8 ± 2.6 46.9 ± 3.4 113.6 ± 5.0 80.2 ± 3.1 81.8 ± 2.8 66.05*** 33.52*** 5.02* 5.68 ns 21.43*** Leaf type Surface Populations Two-way ANOVA One-way nested ANOVA Understorey habitat (m) Semi-alpine habitat (m) 430 580 1220 1300 1370 Population Surface Inter-action Between habitats Within habitat (A) Trichome density (mm−2) Cauline leaf AD 0.0 ± 0 0.0 ± 0 61.9 ± 3.1 59.7 ± 3.4 70.7 ± 3.1 – 3.28* AB 0.0 ± 0 0.0 ± 0 78.0 ± 5.5 72.0 ± 2.8 94.4 ± 3.8 9.07ns 27.0* 1.20ns – 7.71** Spring rosette leaf AD 0.0 ± 0 0.0 ± 0 36.6 ± 2.6 30.5 ± 2.3 29.6 ± 2.2 – 2.64ns AB 0.0 ± 0 0.0 ± 0 38.4 ± 2.2 35.4 ± 1.6 39.3 ± 2.8 2.03ns 8.57ns 1.51ns – 0.79ns Autumn rosette leaf AD 0.0 ± 0 0.0 ± 0 30.0 ± 2.4 34.2 ± 2.2 23.6 ± 1.9 – 5.22** AB 0.0 ± 0 0.0 ± 0 33.6 ± 2.0 36.4 ± 2.6 24.1 ± 1.7 13.03*** 1.45 ns 0.25 ns – 8.13** (B) Stomatal density (mm−2) Cauline leaf AD 31.3 ± 1.3 37.3 ± 2.4 64.7 ± 2.2 65.6 ± 1.9 73.2 ± 2.1 60** 4.78** AB 38.8 ± 1.4 41.0 ± 2.4 94.0 ± 2.9 82.6 ± 2.2 88.1 ± 2.3 223*** 90.1*** 9.42*** 101** 4.48** Spring rosette leaf AD 34.0 ± 1.3 35.2 ± 2.7 65.0 ± 2.3 68.6 ± 2.6 69.5 ± 1.5 218*** 0.94ns AB 49.6 ± 1.9 38.9 ± 6.0 89.2 ± 2.5 88.0 ± 2.2 97.6 ± 2.6 161*** 120*** 4.84*** 75.64** 4.64** Autumn rosette leaf AD 46.0 ± 2.3 51.7 ± 4.9 83.4 ± 3.0 71.5 ± 2.9 62.8 ± 3.3 7.38 ns 8.68*** AB 60.8 ± 2.6 46.9 ± 3.4 113.6 ± 5.0 80.2 ± 3.1 81.8 ± 2.8 66.05*** 33.52*** 5.02* 5.68 ns 21.43*** AD, adaxial; AB, abaxial. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant at P < 0.05. View Large Table 2. Trichome density and stomatal density of cauline and rosette leaves in spring and rosette leaves in autumn of Arabidopsis halleri subsp. gemmifera collected from five populations. Average ± s.e. (n = 10–40) for adaxial (AD) and abaxial (AB) surfaces are given. For explanations of the ANOVAs used see the caption of Table 1. Because no plants in the understorey habitat (430 and 580 m) had trichomes on either leaf surface, they were not included in the statistical analyses. Therefore, population effects in two-way ANOVA represent effects across three populations in the semi-alpine habitat, and there is no between-habitats term in the one-way nested ANOVAs for trichome density (shown as –) Leaf type Surface Populations Two-way ANOVA One-way nested ANOVA Understorey habitat (m) Semi-alpine habitat (m) 430 580 1220 1300 1370 Population Surface Inter-action Between habitats Within habitat (A) Trichome density (mm−2) Cauline leaf AD 0.0 ± 0 0.0 ± 0 61.9 ± 3.1 59.7 ± 3.4 70.7 ± 3.1 – 3.28* AB 0.0 ± 0 0.0 ± 0 78.0 ± 5.5 72.0 ± 2.8 94.4 ± 3.8 9.07ns 27.0* 1.20ns – 7.71** Spring rosette leaf AD 0.0 ± 0 0.0 ± 0 36.6 ± 2.6 30.5 ± 2.3 29.6 ± 2.2 – 2.64ns AB 0.0 ± 0 0.0 ± 0 38.4 ± 2.2 35.4 ± 1.6 39.3 ± 2.8 2.03ns 8.57ns 1.51ns – 0.79ns Autumn rosette leaf AD 0.0 ± 0 0.0 ± 0 30.0 ± 2.4 34.2 ± 2.2 23.6 ± 1.9 – 5.22** AB 0.0 ± 0 0.0 ± 0 33.6 ± 2.0 36.4 ± 2.6 24.1 ± 1.7 13.03*** 1.45 ns 0.25 ns – 8.13** (B) Stomatal density (mm−2) Cauline leaf AD 31.3 ± 1.3 37.3 ± 2.4 64.7 ± 2.2 65.6 ± 1.9 73.2 ± 2.1 60** 4.78** AB 38.8 ± 1.4 41.0 ± 2.4 94.0 ± 2.9 82.6 ± 2.2 88.1 ± 2.3 223*** 90.1*** 9.42*** 101** 4.48** Spring rosette leaf AD 34.0 ± 1.3 35.2 ± 2.7 65.0 ± 2.3 68.6 ± 2.6 69.5 ± 1.5 218*** 0.94ns AB 49.6 ± 1.9 38.9 ± 6.0 89.2 ± 2.5 88.0 ± 2.2 97.6 ± 2.6 161*** 120*** 4.84*** 75.64** 4.64** Autumn rosette leaf AD 46.0 ± 2.3 51.7 ± 4.9 83.4 ± 3.0 71.5 ± 2.9 62.8 ± 3.3 7.38 ns 8.68*** AB 60.8 ± 2.6 46.9 ± 3.4 113.6 ± 5.0 80.2 ± 3.1 81.8 ± 2.8 66.05*** 33.52*** 5.02* 5.68 ns 21.43*** Leaf type Surface Populations Two-way ANOVA One-way nested ANOVA Understorey habitat (m) Semi-alpine habitat (m) 430 580 1220 1300 1370 Population Surface Inter-action Between habitats Within habitat (A) Trichome density (mm−2) Cauline leaf AD 0.0 ± 0 0.0 ± 0 61.9 ± 3.1 59.7 ± 3.4 70.7 ± 3.1 – 3.28* AB 0.0 ± 0 0.0 ± 0 78.0 ± 5.5 72.0 ± 2.8 94.4 ± 3.8 9.07ns 27.0* 1.20ns – 7.71** Spring rosette leaf AD 0.0 ± 0 0.0 ± 0 36.6 ± 2.6 30.5 ± 2.3 29.6 ± 2.2 – 2.64ns AB 0.0 ± 0 0.0 ± 0 38.4 ± 2.2 35.4 ± 1.6 39.3 ± 2.8 2.03ns 8.57ns 1.51ns – 0.79ns Autumn rosette leaf AD 0.0 ± 0 0.0 ± 0 30.0 ± 2.4 34.2 ± 2.2 23.6 ± 1.9 – 5.22** AB 0.0 ± 0 0.0 ± 0 33.6 ± 2.0 36.4 ± 2.6 24.1 ± 1.7 13.03*** 1.45 ns 0.25 ns – 8.13** (B) Stomatal density (mm−2) Cauline leaf AD 31.3 ± 1.3 37.3 ± 2.4 64.7 ± 2.2 65.6 ± 1.9 73.2 ± 2.1 60** 4.78** AB 38.8 ± 1.4 41.0 ± 2.4 94.0 ± 2.9 82.6 ± 2.2 88.1 ± 2.3 223*** 90.1*** 9.42*** 101** 4.48** Spring rosette leaf AD 34.0 ± 1.3 35.2 ± 2.7 65.0 ± 2.3 68.6 ± 2.6 69.5 ± 1.5 218*** 0.94ns AB 49.6 ± 1.9 38.9 ± 6.0 89.2 ± 2.5 88.0 ± 2.2 97.6 ± 2.6 161*** 120*** 4.84*** 75.64** 4.64** Autumn rosette leaf AD 46.0 ± 2.3 51.7 ± 4.9 83.4 ± 3.0 71.5 ± 2.9 62.8 ± 3.3 7.38 ns 8.68*** AB 60.8 ± 2.6 46.9 ± 3.4 113.6 ± 5.0 80.2 ± 3.1 81.8 ± 2.8 66.05*** 33.52*** 5.02* 5.68 ns 21.43*** AD, adaxial; AB, abaxial. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant at P < 0.05. View Large The number of stomata per unit area (stomatal density) was higher in leaves from the semi-alpine habitat than in those from the understorey habitat for all cauline leaves and spring and autumn rosette leaves (Table 2B). Stomatal density on the abaxial surfaces was significantly higher than on the adaxial surfaces of all three leaf types (Table 2B). These differences were mostly supported by statistically significant differences in the two-way ANOVA and one-way nested ANOVA (Table 2B). For other stomatal characters, i.e. stomatal index, guard cell length and pore length, there were no consistent habitat differences, except for the stomatal index of spring rosette leaves and pore length of cauline leaves (Supplementary Data 1, Table S3). Growth experiments In the growth chamber experiment, cauline leaves from the understorey habitat exhibited average θ values <110° for both adaxial and abaxial surfaces, i.e. the leaves were classified as wettable surfaces. The adaxial surface of cauline leaves from the semi-alpine habitat had average θ values >110°, i.e. the leaves were classified as non-wettable; furthermore, the average θ value of abaxial surfaces was >130°, i.e. the leaves were classified as highly non-wettable (Fig. 4A). Highly significant effects were detected for habitat, surface and their interaction in the two-way ANOVA (Table 3A). For the post-cold rosette leaves in the understorey and semi-alpine habitat, the θ values were <110° for both adaxial and abaxial surfaces (Fig. 4B). This indicated that almost all the post-cold rosette leaves were wettable surfaces. However, we found significant differences between habitats in the two-way ANOVA (Table 3A). No significant effects were detected for surface and interaction terms (Table 3A). For the rosette leaves at the pre-cold sampling, the average θ values were ~110° (the boundary value between wettable and non-wettable) for both surfaces and habitats (Fig. 4C). We detected no significant effects for all terms in the two-way ANOVA (Table 3A). Fig. 4. View largeDownload slide Contact angle (A–C) and droplet retention (D–F) of cauline (A, D) and rosette leaves in post-cold (B, E) and pre-cold (C, F) leaves collected from the growth chamber. White and grey boxes represent adaxial and abaxial surfaces, respectively. Box plots show the median (line inside the box), upper and lower quartiles (75 %, 25 %) and maximum and minimum values. Fig. 4. View largeDownload slide Contact angle (A–C) and droplet retention (D–F) of cauline (A, D) and rosette leaves in post-cold (B, E) and pre-cold (C, F) leaves collected from the growth chamber. White and grey boxes represent adaxial and abaxial surfaces, respectively. Box plots show the median (line inside the box), upper and lower quartiles (75 %, 25 %) and maximum and minimum values. Table 3. Results of two-way ANOVA for leaf wettability (contact angle and droplet retention) of cauline leaves and post- and pre-cold rosette leaves of Arabidopsis halleri subsp. gemmifera plants grown in growth chambers. Seeds were derived from two habitat types [understorey (430 and 580 m) and semi-alpine (1220 and 1370 m)]. For each leaf type, effects of leaf surfaces and habitats were tested by two-way ANOVAs. Values are F ratios Trait Leaf type Two-way ANOVA Habitat Surface Interaction (A) Contact angle, θ (°) Cauline leaf 164.25*** 13.09*** 33.11*** Post-cold rosette leaf 49.20*** 3.63ns 2.66ns Pre-cold rosette leaf 0.001ns 0.16ns 1.52ns (B) Droplet retention (°) Cauline leaf 3.25ns 1.62ns 0.03ns Post-cold rosette leaf 2.11ns 3.29ns 0.002ns Pre-cold rosette leaf 0.18ns 0.39ns 1.35ns Trait Leaf type Two-way ANOVA Habitat Surface Interaction (A) Contact angle, θ (°) Cauline leaf 164.25*** 13.09*** 33.11*** Post-cold rosette leaf 49.20*** 3.63ns 2.66ns Pre-cold rosette leaf 0.001ns 0.16ns 1.52ns (B) Droplet retention (°) Cauline leaf 3.25ns 1.62ns 0.03ns Post-cold rosette leaf 2.11ns 3.29ns 0.002ns Pre-cold rosette leaf 0.18ns 0.39ns 1.35ns *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant at P < 0.05. View Large Table 3. Results of two-way ANOVA for leaf wettability (contact angle and droplet retention) of cauline leaves and post- and pre-cold rosette leaves of Arabidopsis halleri subsp. gemmifera plants grown in growth chambers. Seeds were derived from two habitat types [understorey (430 and 580 m) and semi-alpine (1220 and 1370 m)]. For each leaf type, effects of leaf surfaces and habitats were tested by two-way ANOVAs. Values are F ratios Trait Leaf type Two-way ANOVA Habitat Surface Interaction (A) Contact angle, θ (°) Cauline leaf 164.25*** 13.09*** 33.11*** Post-cold rosette leaf 49.20*** 3.63ns 2.66ns Pre-cold rosette leaf 0.001ns 0.16ns 1.52ns (B) Droplet retention (°) Cauline leaf 3.25ns 1.62ns 0.03ns Post-cold rosette leaf 2.11ns 3.29ns 0.002ns Pre-cold rosette leaf 0.18ns 0.39ns 1.35ns Trait Leaf type Two-way ANOVA Habitat Surface Interaction (A) Contact angle, θ (°) Cauline leaf 164.25*** 13.09*** 33.11*** Post-cold rosette leaf 49.20*** 3.63ns 2.66ns Pre-cold rosette leaf 0.001ns 0.16ns 1.52ns (B) Droplet retention (°) Cauline leaf 3.25ns 1.62ns 0.03ns Post-cold rosette leaf 2.11ns 3.29ns 0.002ns Pre-cold rosette leaf 0.18ns 0.39ns 1.35ns *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant at P < 0.05. View Large Average droplet retention of water was <20° for both surfaces of cauline leaves and post-cold and pre-cold rosette leaves (Fig. 4D, E and F, respectively). This indicated that all three leaf types had low droplet retention capacity for both surfaces. No significant effects were detected in the two-way ANOVAs (Table 3B). Plants from the understorey habitat had no trichomes on either surface for all three leaf types (Table 4A). Semi-alpine leaves had trichomes on both surfaces. The number of trichomes per unit area was significantly higher for abaxial than adaxial surfaces for cauline leaves only (Table 4A). Table 4. Trichome density and stomatal density of cauline leaves and post- and pre-cold rosette leaves of Arabidopsis halleri subsp. gemmifera plants grown in growth chambers. Seeds were derived from two habitat types [understorey (430 and 580 m) and semi-alpine (1220 and 1370 m)]. Average ± s.e. (n = 36) for adaxial (AD) and abaxial (AB) surfaces are given. For explanation of two-way ANOVAs used see the caption of Table 3. For trichome density, surface effects were evaluated by one-way ANOVAs for plants from the semi-alpine habitat Traits Leaf type Surface Habitat Two-way ANOVA Understorey Semi-alpine Habitat Surface Interaction (A) Trichome density (mm−2) Cauline leaf AD 0.0 ± 0 25.6 ± 1.6 AB 0.0 ± 0 33.8 ± 2.3 – 9.03** – Post-cold rosette leaf AD 0.0 ± 0 17.7 ± 0.8 AB 0.0 ± 0 19.9 ± 1.1 – 2.49ns – Pre-cold rosette leaf AD 0.0 ± 0 16.7 ± 0.3 AB 0.0 ± 0 16.8 ± 0.9 – 0.006ns – (B) Stomatal density (mm−2) Cauline leaf AD 89.3 ± 3.5 105.2 ± 3.6 AB 101.6 ± 3.2 103.7 ± 3.0 7.20** 2.62ns 4.25* Post-cold rosette leaf AD 94.2 ± 3.0 87.7 ± 4.2 AB 101.0 ± 3.1 94.5 ± 3.2 3.57ns 3.91ns 0.00ns Pre-cold rosette leaf AD 45.6 ± 2.3 55.6 ± 3.3 AB 74.5 ± 2.8 68.3 ± 2.7 0.44ns 56.43*** 8.49** Traits Leaf type Surface Habitat Two-way ANOVA Understorey Semi-alpine Habitat Surface Interaction (A) Trichome density (mm−2) Cauline leaf AD 0.0 ± 0 25.6 ± 1.6 AB 0.0 ± 0 33.8 ± 2.3 – 9.03** – Post-cold rosette leaf AD 0.0 ± 0 17.7 ± 0.8 AB 0.0 ± 0 19.9 ± 1.1 – 2.49ns – Pre-cold rosette leaf AD 0.0 ± 0 16.7 ± 0.3 AB 0.0 ± 0 16.8 ± 0.9 – 0.006ns – (B) Stomatal density (mm−2) Cauline leaf AD 89.3 ± 3.5 105.2 ± 3.6 AB 101.6 ± 3.2 103.7 ± 3.0 7.20** 2.62ns 4.25* Post-cold rosette leaf AD 94.2 ± 3.0 87.7 ± 4.2 AB 101.0 ± 3.1 94.5 ± 3.2 3.57ns 3.91ns 0.00ns Pre-cold rosette leaf AD 45.6 ± 2.3 55.6 ± 3.3 AB 74.5 ± 2.8 68.3 ± 2.7 0.44ns 56.43*** 8.49** AD, adaxial; AB, abaxial. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant at P < 0.05. View Large Table 4. Trichome density and stomatal density of cauline leaves and post- and pre-cold rosette leaves of Arabidopsis halleri subsp. gemmifera plants grown in growth chambers. Seeds were derived from two habitat types [understorey (430 and 580 m) and semi-alpine (1220 and 1370 m)]. Average ± s.e. (n = 36) for adaxial (AD) and abaxial (AB) surfaces are given. For explanation of two-way ANOVAs used see the caption of Table 3. For trichome density, surface effects were evaluated by one-way ANOVAs for plants from the semi-alpine habitat Traits Leaf type Surface Habitat Two-way ANOVA Understorey Semi-alpine Habitat Surface Interaction (A) Trichome density (mm−2) Cauline leaf AD 0.0 ± 0 25.6 ± 1.6 AB 0.0 ± 0 33.8 ± 2.3 – 9.03** – Post-cold rosette leaf AD 0.0 ± 0 17.7 ± 0.8 AB 0.0 ± 0 19.9 ± 1.1 – 2.49ns – Pre-cold rosette leaf AD 0.0 ± 0 16.7 ± 0.3 AB 0.0 ± 0 16.8 ± 0.9 – 0.006ns – (B) Stomatal density (mm−2) Cauline leaf AD 89.3 ± 3.5 105.2 ± 3.6 AB 101.6 ± 3.2 103.7 ± 3.0 7.20** 2.62ns 4.25* Post-cold rosette leaf AD 94.2 ± 3.0 87.7 ± 4.2 AB 101.0 ± 3.1 94.5 ± 3.2 3.57ns 3.91ns 0.00ns Pre-cold rosette leaf AD 45.6 ± 2.3 55.6 ± 3.3 AB 74.5 ± 2.8 68.3 ± 2.7 0.44ns 56.43*** 8.49** Traits Leaf type Surface Habitat Two-way ANOVA Understorey Semi-alpine Habitat Surface Interaction (A) Trichome density (mm−2) Cauline leaf AD 0.0 ± 0 25.6 ± 1.6 AB 0.0 ± 0 33.8 ± 2.3 – 9.03** – Post-cold rosette leaf AD 0.0 ± 0 17.7 ± 0.8 AB 0.0 ± 0 19.9 ± 1.1 – 2.49ns – Pre-cold rosette leaf AD 0.0 ± 0 16.7 ± 0.3 AB 0.0 ± 0 16.8 ± 0.9 – 0.006ns – (B) Stomatal density (mm−2) Cauline leaf AD 89.3 ± 3.5 105.2 ± 3.6 AB 101.6 ± 3.2 103.7 ± 3.0 7.20** 2.62ns 4.25* Post-cold rosette leaf AD 94.2 ± 3.0 87.7 ± 4.2 AB 101.0 ± 3.1 94.5 ± 3.2 3.57ns 3.91ns 0.00ns Pre-cold rosette leaf AD 45.6 ± 2.3 55.6 ± 3.3 AB 74.5 ± 2.8 68.3 ± 2.7 0.44ns 56.43*** 8.49** AD, adaxial; AB, abaxial. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant at P < 0.05. View Large Stomatal density was significantly higher in leaves from the semi-alpine habitat than in the understorey for cauline leaves, but not rosettes leaves (Table 4B). Although stomatal density on abaxial surfaces was significantly higher than that on adaxial surfaces of pre-cold rosette leaves, there was no habitat effect in the two-way ANOVA (Table 4B). No significant habitat or surface effects were detected for stomatal index for all three types of leaves (Supplementary Data 1, Table S4A). For other stomatal characters, i.e. guard cell length and pore length, significant habitat effects were detected in guard cell length for pre- and post-cold rosette leaves (Supplementary Data 1, Table S4B) and in pore length for cauline and post-cold rosette leaves (Supplementary Data 1, Table S4C). Gene expression analysis Expression levels of AhgCER1 were significantly higher in cauline leaves of plants from semi-alpine habitats, i.e. 1220, 1300 and 1370 m than in understorey leaves (Fig. 5). Expression was low in cauline leaves for plants from the low-altitude habitat and rosette leaves for both habitats (Fig. 5). The results were similar for AhgCER1/AhgACT2 (Fig. 5) and AhgCER1/AhgPP2AA3 (Supplementary Data 1, Fig. S2). The patterns were statistically supported by high significance in all terms in the two-way ANOVAs (Supplementary Data 1, Table S5). Fig. 5. View largeDownload slide AhgCER1 relative expression (AhgCER1/AhgACT2) in cauline and rosette leaves collected from different populations [430 m (understorey); 1220, 1300, 1370 m (semi-alpine)]. Grey and black boxes represent cauline and rosette leaves, respectively. Averages and standard deviations (error bars) are shown. Fig. 5. View largeDownload slide AhgCER1 relative expression (AhgCER1/AhgACT2) in cauline and rosette leaves collected from different populations [430 m (understorey); 1220, 1300, 1370 m (semi-alpine)]. Grey and black boxes represent cauline and rosette leaves, respectively. Averages and standard deviations (error bars) are shown. DISCUSSION In both field observations and growth experiments with A. halleri subsp. gemmifera populations along an altitudinal gradient, wettability of cauline leaves was significantly decreased in semi-alpine habitats compared with the low-altitude understorey. In the semi-alpine habitats, adaxial surfaces of cauline leaves were non-wettable. Furthermore, abaxial surfaces of cauline leaves became highly non-wettable in growth experiments. Our growth experiments were designed to detect genetically based phenotypic differences by growing plants from seeds in a common set of environments. The results of the experiments indicated that the difference in leaf wettability between the two distinct altitudinal habitats had a genetic basis, at least in part. At the level of phenotypic variation, intraspecific altitudinal variation in leaf wettability has been demonstrated in previous studies, and leaf wettability generally decreased as altitude increased (Aryal and Neuner, 2010, 2012). In the field site, environmental factors are likely to be critical determinants of leaf wettability (Fogg, 1947; Weiss, 1988). We found that wettability of leaves was generally higher in the growth experiments than in the field measurements, probably because the surface of leaves had been exposed to various conditions, including sun, rain and snow in the natural habitats. We should note that the altitudinal differentiation of leaf wettability was specific to cauline leaves. Cauline leaves of the semi-alpine plants had the genetically fixed non-wettable character. This led us to assume the existence of a specific function of non-wettable cauline leaves that only occurs in the semi-alpine habitat. Because cauline leaves cover a flowering bud at the centre of rosettes in the early stage of bolting, we suspect that low wettability is required for young cauline leaves to serve for protection of flowering buds in very early spring. At the summit of Mt Ibuki, monthly minimum temperature (± s.d.) averaged −10.7 ± 1.1 °C, −6.6 ± 1.6 °C and 0.3 ± 1.9 °C in March, April and May, respectively (1971–2000; Japan Meteorological Agency). Monthly maximum wind velocity averaged 28.5 ± 5.2, 29.3 ± 6.3 and 27.5 ± 5.6 m s−1 in March, April and May, respectively. We suspect that highly non-wettable surfaces reduce frost damage in cauline leaves and the flower buds inside. The lower wettability of the abaxial surface of cauline leaves relative to that of the adaxial surface supports this idea, because the abaxial surface is exposed during the early stage of bolting when cauline leaves cover flower buds. In temperate zones, there is a risk of spring frost, which can cause critical damage to flowers or buds (Rodrigo, 2000). Previous studies also suggested that spring damage of flowers occurs more often at the cold margins of plant distributions (Charrier et al., 2015). Additionally, it has been demonstrated that low leaf wettability creates a frost-protective effect by experimentally spraying tomato plants with a hydrophobic kaolin particle film (Wisniewski et al., 2002). We suspect that the cause of cauline-leaf-specific low wettability of the semi-alpine plants is a difference in the wax characteristics, such as content, composition and microstructure, of leaf surfaces because we found higher expression of AhgCER1, a CER1 homologue in the study species, in the corresponding leaves. CER1 was first verified to be responsible for the biosynthesis of highly hydrophobic molecules in the epidermal wax of A. thaliana (Aarts et al., 1995; Jenks et al., 1995). More recently, upregulation of CER1 was reported to increase the alkane content of A. thaliana leaves, and to improve the cuticle barrier properties against biotic and abiotic stress (Kosma et al., 2009; Bourdenx et al., 2011). Kubota et al. (2015) conducted genome-wide scans for populations of A. halleri subsp. gemmifera along the altitudinal gradient of Mt Ibuki and identified genes that contained SNPs with altitudinal differentiation in allele frequencies. AhgCER1 and other wax-related genes were not listed in their analyses. Further analyses are required to identify the responsible nucleotide substitutions underlying the altitudinal differentiation of AhgCER1 expression in cauline leaves. The angle of droplet retention of leaves in both the field and the growth experiment was low (<20°), indicating that all investigated leaf types had less sticky surface on which water droplets could easily move. Run-off of water droplets depends on the stickiness of the surface and the leaf inclination. In our case, rosette leaves were horizontally arranged, whereas cauline leaves were vertically arranged. It is likely that droplets would run off the leaf surface immediately after contact on vertically angled cauline leaves from semi-alpine habitats, which have been found to be non-wettable. A low retention angle may also be important in providing dry surfaces for cauline leaves in the semi-alpine habitat. Cauline and rosette leaves of understorey plants had no trichomes, but the trichome density was high on both surfaces of leaves from the semi-alpine habitat. For rosette leaves, the substantial difference in trichome density, i.e. presence or absence of trichomes, between understorey and semi-alpine plants did not correspond to the similarity in leaf wettability between the habitats. It has been reported that trichomes on leaf surfaces enhance water repellence by the formation of spherical droplets (Brewer et al., 1991; Brewer and Smith, 1997; Pandey and Nagar, 2003; Aryal and Neuner, 2010), whereas another study has reported that a high density of trichomes holds water droplets more effectively than does a medium or low density (Wang et al., 2015). High trichome formation in cauline leaves of semi-alpine plants may contribute to low leaf wettability of these leaves if the trichome surface is hydrophobic (Neinhuis and Barthlott, 1997). The stomatal density of cauline and rosette leaves increased progressively from understorey to semi-alpine habitats in field-collected samples. Similar altitudinal patterns have been reported in other species (Aryal and Neuner, 2012). Our growth chamber experiments indicated that the difference was not the result of genetic differentiation. In A. halleri, the altitudinal changes in stomata density represent a phenotypic response. Because the stomatal index was constant between habitats in the growth experiment, the plasticity of leaf size may result in the altitudinal gradients of stomatal density per unit leaf area. In conclusion, we identified genetic differentiation of leaf wettability between distinctive montane habitats in the rosette-forming perennial Arabidopsis. To our knowledge, this is the first report of genetic differentiation in leaf wettability between natural plant populations within a single species. The lower wettability specific to cauline leaves of the semi-alpine plants is supportive of the postulate that the dry surface of cauline leaves is necessary for the protection of floral buds from frost under exposure to prevailing cold wind. The theory needs to be tested in future field and experimental studies. We also identified a candidate gene that could explain the differential leaf wettability between cauline and rosette leaves and between cauline leaves from the two habitats. A genome-wide study on segregated populations derived from crosses between plants from the two habitats will likely reveal the identity of the responsible SNP(s) underlying the discovery in this study. SUPPLEMENTARY DATA Supplementary data are available online at https://academic.oup.com/aob and consist of the following. Supplementary Data 1 consists of Figure S1: scanning electron microscopic images of the adaxial surfaces of cauline leaves and adaxial and abaxial surfaces of rosette leaves. Figure S2: relative expression of AhgCER1 in cauline and rosette leaves. Table S1: list of primer sequences used in this study. Table S2: results of Kolmogorov–Smirnov tests for data normality. Table S3: stomata-related leaf characteristics of cauline and rosette leaves in spring and rosette leaves in autumn of Arabidopsis halleri subsp. gemmifera collected from five different populations. Table S4: stomata-related leaf characteristics of cauline leaves and post- and pre-cold rosette leaves of Arabidopsis halleri subsp. gemmifera plants grown in growth chambers. Table S5: results of two-way ANOVAs for relative gene expression of AhgCER1 in cauline and rosette leaves of A. halleri subsp. gemmifera collected from the field. Supplementary Data 2, 3 and 4 are the original data sets of this study for field measurements, growth experiments and gene expression analyses, respectively. ACKNOWLEDGEMENTS We thank Professor K. Agata, who provided the CCD camera. This work was supported by the Japan Society for Promotion of Science (grant KAKENHI JP26221106) and the Japan Society and Technology Agency (grant CREST JPMJCR15O1) to H.K. LITERATURE CITED Aarts MG , Keijzer CJ , Stiekema WJ , Pereira A . 1995 . Molecular characterization of the CER7 gene of Arabidopsis involved in epicuticular wax biosynthesis and pollen fertility . Plant Cell 7 : 2115 – 2127 . Google Scholar CrossRef Search ADS PubMed Adam NK . 1963 . Principles of water-repellency . In: Moilliet JL , ed. Waterproofing and water-repellency . Amsterdam : Elsevier , 1 – 23 . Aikawa S , Kobayashi MJ , Satake A , Shimizu KK , Kudoh H . 2010 . Robust control of the seasonal expression of the Arabidopsis FLC gene in a fluctuating environment . Proceedings of the National Academy of Sciences of the USA 107 : 11632 – 11637 . Google Scholar CrossRef Search ADS PubMed Aryal B , Neuner G . 2010 . Leaf wettability decreases along an extreme altitudinal gradient . Oecologia 126 : 1 – 9 . Google Scholar CrossRef Search ADS Aryal B , Neuner G . 2012 . Leaf wettability in bilberry Vaccinium myrtillus L. as affected by altitude and openness of the growing site . Phyton 52 : 245 – 262 . Barthlott W , Neinhuis C . 1997 . Purity of sacred lotus, or escape from contamination in biological surfaces . Planta 202 : 1 – 8 . Google Scholar CrossRef Search ADS Bhushan B , Jung YC . 2008 . Wetting, adhesion and friction of superhydrophobic and hydrophilic leaves and fabricated micro/nanopatterned surfaces . Journal of Physics: Condensed Matter 20 : 1 – 24 . Bourdenx B , Bernard A , Domergue F et al. 2011 . Overexpression of Arabidopsis ECERIFERUM1 promotes wax very-long-chain alkane biosynthesis and influences plant response to biotic and abiotic stresses . Plant Physiology 156 : 29 – 45 . Google Scholar CrossRef Search ADS PubMed Brewer CA , Nuñez CI . 2007 . Patterns of leaf wettability along an extreme moisture gradient in western Patagonia, Argentina . International Journal of Plant Sciences 168 : 555 – 562 . Google Scholar CrossRef Search ADS Brewer CA , Smith K . 1997 . Patterns of leaf surface wetness for montane and subalpine plants . Plant, Cell & Environment 20 : 1 – 11 . Google Scholar CrossRef Search ADS Brewer CA , Smith WK , Vogelmann TC . 1991 . Functional interaction between leaf trichomes, leaf wettability and the optical properties of water droplets . Plant, Cell & Environment 14 : 955 – 962 . Google Scholar CrossRef Search ADS Challen SB . 1960 . The contribution of surface characteristics to the wettability of leaves . Journal of Pharmacy and Pharmacology 12 : 307 – 311 . Google Scholar CrossRef Search ADS PubMed Charrier G , Ngao J , Saudreau M , Améglio T . 2015 . Effects of environmental factors and management practices on microclimate, winter physiology, and frost resistance in trees . Frontiers in Plant Science 6 : 1 – 18 . Google Scholar PubMed Crisp DJ . 1963 . Waterproofing in animals and plants . In: Moilliet JL , ed. Waterproofing and water-repellency . Amsterdam : Elsevier , 416 – 481 . Fogg GE . 1947 . Quantitative studies on the wetting of leaves by water . Proceedings of the Royal Society B 134 : 503 – 522 . Google Scholar CrossRef Search ADS Fogg GE . 1948 . Adhesion of water to the external surfaces of leaves . Faraday Discussions of the Chemical Society 3 : 162 – 166 . Google Scholar CrossRef Search ADS Gonzalo-Turpin H , Hazard L . 2009 . Local adaptation occurs along altitudinal gradient despite the existence of gene flow in the alpine plant species . Journal of Ecology 97 : 742 – 751 . Google Scholar CrossRef Search ADS Holder CD . 2011 . The relationship between leaf water repellency and leaf traits in three distinct biogeographical regions . Plant Ecology 212 : 1913 – 1926 . Google Scholar CrossRef Search ADS Holloway PJ . 1969 . Chemistry of leaf waxes in relation to wetting . Journal of the Science of Food and Agriculture 20 : 124 – 128 . Google Scholar CrossRef Search ADS Holloway PJ . 1970 . Surface factors affecting the wetting of leaves . Pesticide Science 1 : 156 – 163 . Google Scholar CrossRef Search ADS Ikeda H , Setoguchi H , Morinaga S . 2010 . Genomic structure of lowland and highland ecotypes of Arabidopsis halleri subsp. gemmifera (Brassicaceae) on Mt Ibuki . Acta Phytotaxonomica et Geobotanica 61 : 21 – 26 . Jenks MA , Tuttle HA , Eigenbrode SD , Fieldmann KA . 1995 . Leaf epicuticular waxes of the eceriferum mutants in Arabidopsis . Plant Physiology 108 : 369 – 377 . Google Scholar CrossRef Search ADS PubMed Jetter R , Kunst L . 2008 . Plant surface lipid biosynthetic pathways and their utility for metabolic engineering of waxes and hydrocarbon biofuels . Plant Journal 54 : 670 – 683 . Google Scholar CrossRef Search ADS PubMed Kosma DK , Bourdenx B , Bernard A et al. 2009 . The impact of water deficiency on leaf cuticle lipids of Arabidopsis . Plant Physiology 151 : 1918 – 1929 . Google Scholar CrossRef Search ADS PubMed Kubota S , Iwasaki T , Hanada K et al. 2015 . A genome scan for genes underlying microgeographic-scale local adaptation in a wild Arabidopsis species . PLoS Genetics 11 : 1 – 26 . Neinhuis C , Barthlott W . 1997 . Characterization and distribution of water repellent, self cleaning plant surfaces . Annuals of Botany 79 : 667 – 677 . Google Scholar CrossRef Search ADS Nishio H , Buzas DM , Nagano AJ et al. 2016 . From the laboratory to the field: assaying histone methylation at FLOWERING LOCUS C in naturally growing Arabidopsis halleri . Genes & Genetic Systems 91 : 15 – 26 . Google Scholar CrossRef Search ADS PubMed Pandey S , Nagar PK . 2003 . Pattern of leaf surface wetness in some important medicinal and aromatic plants of western Himalaya . Flora 198 : 349 – 357 . Google Scholar CrossRef Search ADS Rodrigo J . 2000 . Spring frosts in deciduous fruit trees—morphological damage and flower hardiness . Scientia Horticulturae 85 : 155 – 173 . Google Scholar CrossRef Search ADS Salisbury EJ . 1927 . On the causes and ecological significance of stomatal frequency with special reference to the woodland flora . Philosophical Transactions of the Royal Society of London B 216 : 1 – 65 . Google Scholar CrossRef Search ADS Savolainen O , Lascoux M , Merliä J . 2013 . Ecological genomics of local adaptation . Nature Reviews Genetics 14 : 807 – 820 . Google Scholar CrossRef Search ADS PubMed Schönherr J , Bukovac MJ . 1972 . Penetration of stomata by liquids . Plant Physiology 49 : 813 – 819 . Google Scholar CrossRef Search ADS PubMed Smith WK , McClean TM . 1989 . Adaptive relationship between leaf water repellency, stomatal distribution, and gas exchange . American Journal of Botany 76 : 465 – 469 . Google Scholar CrossRef Search ADS Wang H , Shi H , Wang Y . 2015 . The wetting of leaf surfaces and its ecological significances . In: Aliofkhazraei M , ed. Wetting and wettability . Rijeka, Croatia : InTech , 295 – 321 . Google Scholar CrossRef Search ADS Warburton FL . 1963 . The effect of structure on waterproofing . In: Moilliet JL , ed. Waterproofing and water-repellency . Amsterdam : Elsevier , 24 – 51 . Weiss A . 1988 . Contact angle of water droplets in relation to leaf water potential . Agricultural and Forest Meteorology 43 : 251 – 259 . Google Scholar CrossRef Search ADS Wisniewski M , Glenn DM , Fuller MP . 2002 . Use of a hydrophobic particle film as a barrier to extrinsic ice nucleation in tomato plants . Journal of American Society of Horticulture Science 127 : 358 – 364 . © The Author(s) 2018. Published by Oxford University Press on behalf of the Annals of Botany Company. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Annals of BotanyOxford University Press

Published: Mar 20, 2018

There are no references for this article.