Foliar water uptake in Pinus species depends on needle age and stomatal wax structures

Abstract

Background and aims

Foliar water uptake (FWU) has been documented in many species and is increasingly recognized as a non-trivial factor in plant–water relationships. However, it remains unknown whether FWU is a widespread phenomenon in Pinus species, and how it may relate to needle traits such as the form and structure of stomatal wax plugs. In this contribution, these questions were addressed by studying FWU in current-year and 1-year-old needles of seven Pinus species.

Methods

We monitored FWU gravimetrically and analysed the needle surface via cryo-scanning electron microscopy. Additionally, we considered the effect of artificial wax erosion by application of the surfactant Triton X-100, which is able to alter wax crystals.

Key results

The results show for all species that (1) FWU occurred, (2) FWU is higher in old needles compared to young needles and (3) there is substantial erosion of stomatal wax plugs in old needles. FWU was highest in Pinus canariensis, which has a thin stomatal wax plug. Surfactant treatment enhanced FWU.

Conclusions

The results of this study provide evidence for (1) widespread FWU in Pinus, (2) the influence of stomatal wax plugs on FWU and (3) age-related needle surface erosion.

INTRODUCTION

Foliar water uptake (FWU) is widely acknowledged as an ecologically relevant source of water for many vascular plant taxa (Dawson and Goldsmith, 2018; Berry et al., 2019; Schreel and Steppe, 2020). From an applied perspective, foliar uptake of fertilizers, as well as other agrochemicals, is an established agricultural practice (Knoche, 1994; Fernández et al., 2021). FWU has been demonstrated for angiosperm trees (Eller et al., 2013; Fernández et al., 2014; Carmichael et al., 2020; Schreel et al., 2020; Guzmán-Delgado et al., 2021), conifers (Leyton and Armitage, 1968; Boucher et al., 1995; Burgess and Dawson, 2004; Breshears et al., 2008; Limm et al., 2009; Laur and Hacke, 2014; Cassana et al., 2016), shrubs and herbs (Emery, 2016; Cavallaro et al., 2020) as well as ferns (Limm and Dawson, 2010; Schwerbrock and Leuschner, 2017).

There is evidence that Pinus species are also able to utilize atmospheric water (rain, fog, mist and dew) by FWU (Leyton and Armitage, 1968; Boucher et al., 1995; Baguskas et al., 2016; Miranda et al., 2021; Tianshi and Chau, 2022). Pinus is a species-rich conifer genus with a wide distribution area over the northern hemisphere, spanning a global latitudinal range of more than 70° (Nobis et al., 2012), and many species are of high economic value. As is typical for conifers, stomata are sunken within epistomatal chambers in Pinus, and are occluded by densely meshed wax tubules (Yoshie and Sara, 1985; Kim et al., 2011) also known as ‘stomatal plugs’. There is evidence that these wax tubules decrease maximum stomatal conductance (Jeffree et al., 1971; Brodribb and Hill, 1997). Stomatal plugs are therefore mainly regarded as an anti-transpirant (Brodribb and Hill, 1997), but other effects have also been discussed. In particular, stomatal wax plugs are suggested to prevent pathogens from penetrating the stomatal pore (Deckert et al., 2001, and citations therein).

Given the low water permeability of the cuticle, stomata, hydathodes and trichomes were suggested as entrance sites for water (Berry et al., 2019; Schreel et al., 2020). In the case of coniferous leaves, stomata are the most likely candidate for water entrance, given the absence of other alternative entrance sites. Although there is much evidence of substantial FWU in conifers, it is to be expected that stomatal plugs decrease the ability for FWU. Information on the effect of stomatal plugs on FWU is, however, still largely missing. Also, the mechanisms and pathways of FWU are not completely understood (Berry et al., 2019; Fernández et al., 2021; Guzmán-Delgado et al., 2021). Water may enter a leaf as liquid or – as ‘reverse transpiration’ – in the vapour phase (Vesala et al., 2017; Berry et al., 2019; Guzmán-Delgado et al., 2021). For both phases, it is reasonable to assume that stomatal plugs hamper FWU.

In the case of the typically small and needle- or scale-like leaves of conifers, FWU experiments are often conducted with leafy branches. However, because there is evidence for bark as an alternative entrance site for water in conifers, as reported for Sequoia sempervirens, Picea abies and Taxodium distichum (Katz et al., 1989; Mayr et al., 2014; Mason Earles et al., 2016; Losso et al., 2021), FWU data obtained by using twigs may possibly be confounded by bark water uptake.

In older leaves, the cuticle usually becomes eroded by abrasion (Van Gardingen et al., 1991) and by degradation of wax crystals, which become aggregated into an amorphous mass or vanish completely, including the stomatal wax plug material (Van Gardingen et al., 1991; Burkhardt and Pariyar, 2014). The process is enhanced by aerosols, which can accumulate on the cuticle and in the epistomatal chamber (Burkhardt and Pariyar, 2014). It is therefore conceivable that the wax plug becomes more permeable in older needles. Also, the accumulation of aerosols and organisms settling on the leaf can increase wettability (Burkhardt et al., 2012, and citations therein).

In fact, Burgess and Dawson (2004) found evidence that FWU particularly occurred via older stomata in S. sempervirens, and also suggested that fungal hyphae may be involved in transferring water to the leaf interior. The effect of needle age and surface erosion on FWU was, however, rarely considered. Besides the study of Burgess and Dawson (2004), another investigation on FWU was made in Pinus torreyana in which slightly higher water uptake was reported for older leaves compared to current-year leaves (Tianshi and Chau, 2022).

The present study provides a species-specific comparison of FWU in Pinus that also considers the factor of needle age. We specifically focused on water absorption of needles and excluded water uptake via the bark. The potential impact of needle surface and wax plug structure was studied using cryo-scanning electron microscopy (cryo-SEM). Furthermore, the influence of Triton X-100 on FWU was considered, because this surfactant is known to degrade wax crystals (Noga et al., 1988), and by treating needles with Triton, the effect of an ‘artificial’ erosion on wax crystals was included. Seven species of Pinus were considered: P. canariensis, P. contorta, P. jeffreyi, P. mugo, P. ponderosa, P. sylvestris and P. wallichiana.

The following questions will be addressed:

• (1) Is FWU widespread in Pinus?

• (2) Are there species-specific differences?

• (3) Are there age-related differences (current-year needles vs. older needles)?

• (4) Is there any evidence for the influence of needle surface and stomatal structure, particularly wax plug structure, on FWU?

• (5) How is FWU affected by surfactants?

MATERIAL AND METHODS

Plant material

Needles of P. contorta, P. jeffreyi, P. mugo, P. ponderosa, P. sylvestris and P. wallichiana used for the study were collected from mature trees growing in the Botanical Garden of the University of Hohenheim and in the case of P. canariensis from the Botanical Garden of the University of Tübingen. All sampled trees from the Botanical Garden Hohenheim are growing outside in the Landesarboretum Baden-Württemberg (GPS coordinates 48.712238, 9.201056). Only P. canariensis trees are cultivated in a glasshouse in the Botanical Garden Tübingen, because this species is sensitive to freezing. For most of the species we used current-year needles (‘young needles’) and 1-year-old needles (‘old needles’). After being collected, the needles were stored loosely in plastic bags and quickly transported to the laboratory at the Museum of Natural History Stuttgart, Germany. The needles were kept in fascicles throughout the entirety of the experiments. Depending on the species, fascicles had two to five needles. Sampling and the subsequent investigations were conducted from October to December 2020.

Hydration experiments

Once the needle fascicles arrived in the laboratory, needles were labelled at their basal end with a permanent marker and their fresh weight was measured with a chemical precision balance (Kern ABS 120-4, Kern & Sohn, Balingen, Germany, resolution: 0.1 mg = 0.0001 g, capacity = 120 g). The needles were left on the laboratory bench to dehydrate for 24 h at 21–22 °C and 25–35 % relative humidity (RH) and their mass after that time was measured (Mass_dehy24). Then, one group of needles (n = 9–30 needles) was immersed in glass cylinders filled with deionized water for 24 h to test for potential FWU at room temperature. At the same time, another group of needles (n = 8–30 needles) was immersed in glass cylinders filled with a 0.1 % (v/v) Triton X-100 (CAS no. 9002-93-1, Sigma-Aldrich, Saint Louis, MO, USA) water mix for 24 h at 21–22 °C and 25–35 % RH. For both treatments, needles were left in fascicles, and each fascicle was placed in a separate glass cylinder filled either with deionized water or Triton solution. In total, there were 459 observations [7 species × 8–30 repetitions × 2 ages × 2 (surfactant) treatments].

Prior to inserting needles into water, we sealed the needle base with waterproof glue (‘tesa Alleskleber wasserfest’, tesa SE, Norderstedt, Germany). Our priority was to prevent any water uptake via the needle base. Once the glue had hardened, the needles were positioned so that their basal ends protruded about 10 mm (5–15 mm, depending on the needle length) above the water surface. The fact that the needle base was not immersed in water meant that some small rate of water loss from the non-immersed needle surface area was possible. This was measurable for a few needles where this loss of water exceeded any uptake of water through the immersed portion of the needle. Hence, our water uptake values are conservative in nature, meaning a slight underestimation of the true water uptake. We tried to minimize this water loss by maximizing the extent to which needles were immersed in water while ensuring that the needle base had no contact with water.

After the 24-h rehydration period was complete, needles were thoroughly blotted dry, and left to air-dry for ~5 min. For this, the needles were teased apart at the individual fascicles to remove all water. This was especially important for long and thin needles. The needles were then weighed (Mass_rehy24).

We conducted a control experiment to quantify any residual water that may remain on the needle surface after removing the fascicles from the deionized water or Triton solution. To do this we used needles whose surfaces were completely dry. These needles were weighed, submerged in deionized water for 1 s and treated as described above (blotted dry and left to air-dry for 3 min), and reweighed. No residual water could be detected in these control experiments.

We then measured the turgid mass (Mass_{rehy_full}) after immersing the needles in deionized water for 48 h. To promote rehydration, the open basal ends of the needles were immersed as well, to facilitate water uptake and to allow for complete rehydration. Finally, the dry mass (Mass_dry) was measured after keeping fascicles in a drying oven at 55 °C for at least 24 h, until the mass remained constant. Relative water content for the various hydration states was calculated as (Laur and Hacke, 2014):

$$RW{C}_{dehy24}=\left(Mas{s}_{dehy24}-Mas{s}_{dry}\right)/\left(Mas{s}_{rehy\_full}-Mas{s}_{dry}\right)$$ (1)

$$RW{C}_{rehy24}=\left(Mas{s}_{rehy24}-Mas{s}_{dry}\right)/\left(Mas{s}_{rehy\_full}-Mas{s}_{dry}\right)$$ (2)

Water uptake after the 24-h rehydration period was calculated as a percentage in relation to the water content before (after the 24-h dehydration period) as follows (Limm et al., 2009):

$$RW{C}_{gain}=\left[\left(RW{C}_{rehy24}/RW{C}_{dehy24}\right)-1\right]\cdot 100$$ (3)

The kinetics of FWU uptake is non-linear and converges to zero when the leaf water capacitance is reached (Guzmán-Delgado et al., 2018). This process can be described by a logistic approach: the water uptake rate is fastest at the beginning of rehydration and then decreases with increasing mass of absorbed water. FWU stops when the maximum leaf water content is reached (at leaf water saturation: RWC_sat = 1) (Boanares et al., 2018; Guzmán-Delgado et al., 2018). To account for this influence of leaf water content on FWU, we also expressed FWU as RWC_{gain_cap} (1 corresponds to RWC_sat when the absorbed mass of water = Mass_{rehy_full}):

$$RW{C}_{gain\_cap}=\left[\left(RW{C}_{rehy24}-RW{C}_{dehy24}\right)\right]/\left[1-RW{C}_{dehy24}\right]$$ (4)

RWC  _{gain_cap} normalizes FWU to capacitance and can attain values between 0 (no water is absorbed) and 1 (full dehydration is reached).

All measurement data are listed in Supplementary Data Table S1.

Needle surface

The following anatomical traits were considered in this study: surface structure, particularly of needle surface waxes, and thickness of stomatal plugs. The traits were generally determined by cryo-SEM.

A SUPRA 40VP-31-79 (Carl Zeiss SMT Ltd, Oberkochen, Germany) cryo-scanning electron microscope equipped with an EMITECH K250X cryo-preparation unit (Quorum Technologies Ltd, Ashford, UK) was used. Pieces of needles 5 mm in length were cut with a razor blade and mounted on metal stubs using polyvinyl alcohol (Tissue-Tek, OCT, Sakura Finetek Europe BV, Alphen aan den Rijn, the Netherlands). Subsequently, the samples were shock-frozen in liquid nitrogen in the slushing chamber, transferred to the cryo-preparation chamber at −140 °C, sublimed for 15 min at −70 °C, sputter-coated with platinum (layer thickness ~10 nm), transferred to the SEM device, and then examined in a frozen state at an accelerating voltage of 5 kV and −100 °C. Cryo-SEM micrographs were taken using the software Smart SEM 05.03.05 (Carl Zeiss). For freeze fractures, the needles were perpendicularly fixed on the sample holder, treated as mentioned above and fractured at a frozen state in the preparation chamber before sublimation, sputter-coating and examination.

To evaluate the effects of the surfactant on the cuticle waxes visually, some of the material was treated with a 0.1 % (v/v) solution of the surfactant Triton X-100 by spraying the needles twice from a spray bottle, from a distance of about 15 cm. Afterwards, the samples were sublimated to remove ice artefacts from the surface.

We attempted to determine the thickness of stomatal wax plugs based on cryo-SEM images by identifying categories. After inspecting interspecific characteristics of wax plugs, three classes were identified (see Fig. 1). Class 1 was assigned to thin and partially loosely structured wax plugs which did not fill the epistomatal chamber and which allowed us to recognize the underlying guard cells. Class 2 wax plugs were thick and dense enough to obscure the view of the guard cells, but the wax plugs did not fill the entire epistomatal chamber. The thickest wax plugs, assigned to class 3, filled the epistomatal chamber and partially protruded above its external aperture.

Fig. 1.

Schematic drawings of stomatal wax plug thickness categories (upper images) and corresponding examples of cryo-SEM images showing longitudinal sections through stomata and epistomatal chambers (lower images, white arrows indicate stomatal slit). Class 1 was assigned to thin and often loosely structured wax plugs, which did not completely fill the epistomatal chamber and which allowed a view of the underlying guard cells. Class 2 wax plugs are thick and dense enough to obscure the view of the guard cells but do not fill the entire epistomatal chamber. The thickest wax plugs, assigned to class 3, fill the epistomatal chamber entirely and even partially protrude above its external aperture. Illustrations of wax tubules forming the stomatal wax plug are not drawn to scale. Greyish bottom with line in drawing of plug category 1: surface of stomatal guard cells and pore slit. Scale bars: 10 µm.

Statistical analysis

Data analysis and visualizations were carried out using the R language and environment for statistical computing (R Core Team, 2021), the glmmTMB package (Brooks et al., 2017) as well as the tidyverse packages (Wickham et al., 2019). Multilevel models, or more specifically generalized linear mixed-effects models (Gelman and Hill, 2006; Bolker et al., 2009; Zuur et al., 2013), were used to fit the increase in relative water content (RWC_gain) in needles of seven Pinus species. Since 52 needle measurements resulted in slightly negative values (indicating evaporative water loss through the non-immersed part of the needle basis, see preceding section: ‘Hydration experiments’), we set those values to zero and ran a logistic regression model first in order to statistically separate those needles that showed no water uptake (set to zero) from those needles that showed water uptake (set to one). Another approach would be to keep the zeros in the main analysis. However, this (1) reduced the RWC_gain of those needles that showed water uptake, hence underestimating the RWC_gain values for this group, and (2) resulted in poor regression diagnostics due to excessive zeros, questioning the validity of the calculated P-values. This ‘two-stage’ modelling approach, which is known under the name of hurdle models, is therefore superior given the data and can also easily be calculated in one function call using the glmmTMB package (Brooks et al., 2017). We assumed RWC_gain to be gamma distributed (positive continuous) with the distribution parameters μ and ν (Faraway, 2016, p. 175). On the predictor side of the modelling equation RWC_dehy24 was used as a continuous covariate to adjust RWC_gain for each needle's dehydration status. Furthermore, Needle Age, Triton and Species were used as fixed categorical predictor variables. All combinations of species and needle age (7 species × 2 needle age groups = 14 clusters) were modelled as a random effect allowing intercepts (α) to vary randomly between clusters in order to account for multiple needle measurements on the same tree and within the same age class. Random effects were assumed to be normally distributed with the distribution parameters μ and σ². The statistical model looks as follows:

$${\text{RWCgain}}_{\text{i}}~\text{Gamma }\left(\mu \text{,}u\right)\text{,} \text{for} \text{observation} \text{i}=1,\dots ,458$$

$$\begin{array}{l}\text{log }\left(\mu \right)={\alpha }_{\text{j }[\text{ i]}}+{\beta }_1\left(\text{RWCdehy24}\right)+{\beta }_2\left({\text{Triton}}_{\text{Yes}}\right)+{\beta }_3\left({\text{NeedleAge}}_{\text{Young}}\right)\hfill \\ \text{+ }{\beta }_4\left({\text{Species}}_{\text{P.\hspace{0.17em}contorta}}\right)\text{ + }{\beta }_5\left({\text{Species}}_{\text{P\hspace{0.17em}.jeffreyi}}\right)\text{ + }{\beta }_6\left({\text{Species}}_{\text{P\hspace{0.17em}.mugo}}\right)\hfill \\ \begin{array}{l}\text{+ }{\beta }_7\left({\text{Species}}_{\text{P\hspace{0.17em}.ponderosa}}\right)\text{ + }{\beta }_8\left({\text{Species}}_{\text{P\hspace{0.17em}.sylvestris}}\right)\hfill \\ +{\beta }_9\left({\text{Species}}_{\text{P\hspace{0.17em}.wallichiana}}\right)\text{ + }{\beta }_{10}\left({\text{Triton}}_{\text{Yes}}{\text{* NeedleAge}}_{\text{Young}}\right)\hfill \end{array}\hfill \end{array}$$

$$\underset{j}{\alpha }~N\left({\mu }_{\underset{j}{\alpha }},{\sigma }^2{}_{\underset{j}{\alpha }}\right),\text{for cluster j }\left(\text{Species: NeedleAge}\right)\text{ j }=1,\dots ,14$$

Since this model describes a linear regression model of categorical data, dummy coding was used (0,1) to designate group membership, with P. canariensis representing the reference group (or intercept), and all other species were coded as 0.

To test the effect of the wax plug thickness category on RWC_gain for those needles that showed water uptake, a second model was fitted. The difference to the model above is that it only included RWC_dehy24 and Plug as predictor variables and was fitted on young needle data only (i.e. excluding the Triton factor and old needle measurements), which excluded P. wallichiana from the analysis since it showed no water uptake for this treatment combination. Since the model was fitted on young needles only, the random effect intercept simplified to Species:

$$RW{\text{C}}_{\text{gaini}}~\text{Gamma }\left(\mu \text{ , }u\right)\text{,} \text{for} \text{observation} \text{i}=1,\dots ,71$$

$$\text{log }\left(\mu \right)={\alpha }_{\text{j }[\text{ i }]}+{\beta }_1\left(\text{RWCdehy24}\right)+{\beta }_2\left({\text{Plug}}_2\right)+{\beta }_3\left({\text{Plug}}_3\right)$$

$$\underset{j}{\alpha }~N\left({\mu }_{\underset{j}{\alpha }},{\sigma }^2{}_{\underset{j}{\alpha }}\right),\text{for cluster j }\left(\text{Species}\right)\text{ j }=1,\dots ,6$$

Model diagnostics were performed on both models using the DHARMa package (Hartig, 2021) and no abnormal residual patterns were detected (Supplementary Data Statistics S1). Wald chi-square tests were conducted to test the fixed effects in generalized linear mixed-effects models for their significance via the Anova() function of the car package (Fox and Weisberg, 2019). Estimated marginal means were calculated using the emmeans package (Lenth, 2020) based on the fitted models. P-values for multiple mean comparisons were adjusted using the Tukey method.

RESULTS

Dependence of FWU on species, needle age and treatment with the surfactant Triton

With some exceptions (untreated current-year needles of P. jeffreyi, untreated current-year needles of P. ponderosa and untreated current-year needles of P. wallichiana, see Table 1), the needles of the considered species exhibited significant FWU. The term ‘untreated’ refers to immersion in pure water (i.e. no Triton). Besides interspecific differences, FWU also depends on the dehydration level RWC_dehy24 (and therefore on the initial capacitance for water uptake). RWC_gain decreases with increasing RWC_dehy24, as expected (Fig. 2A) while the graphs of RWC_{gain_cap} are more complex (Fig. 2B). This is because RWC_gain increases strongly with decreasing RWC_dehy24. Here, the initial dehydration level potentially confounds differences between data groups. RWC_{gain_cap}, being normalized to maximum leaf capacitance, allows the influence of factors other than the initial state of dehydration on FWU, such as needle age and treatment with Triton, to become more apparent (Fig. 2B) (see Discussion).

Table 1.

Mean probability estimates (%) of the logistic regression model. The estimates represent the probability of no needle water uptake by species, Triton treatment (Yes, No) and needle age (Old, Young). In cases when zero probability is estimated, which is the case when all needles in the respective group showed water uptake, a standard error as well as a significant difference cannot be estimated by the logistic regression model since a probability of 1 (or 100%) represents +Infinity on the logit scale

Triton	Needle age	Probability of no water uptake (%)	Standard error	Significant difference
Pinus canariensis
Yes	Old	0	0	n/a
Yes	Young	0.35	0.42	a
No	Old	1.18	1.28	a
No	Young	8.39	7.99	b
Pinus contorta
Yes	Old	0	0	n/a
Yes	Young	5.61	3.52	a
No	Old	16.8	5.54	a
No	Young	60.8	9.69	b
Pinus jeffreyi
Yes	Old	0	0	n/a
Yes	Young	3.56	2.45	a
No	Old	11.1	5.35	a
No	Young	49	11.2	b
Pinus mugo
Yes	Old	0	0	n/a
Yes	Young	1.51	1.06	a
No	Old	4.96	2.3	a
No	Young	28.6	7.61	b
Pinus ponderosa
Yes	Old	0	0	n/a
Yes	Young	2.37	1.72	a
No	Old	7.62	3.99	a
No	Young	38.7	11.2	b
Pinus sylvestris
Yes	Old	0	0	n/a
Yes	Young	0.231	0.275	a
No	Old	0.781	0.84	a
No	Young	5.69	5.51	b
Pinus wallichiana
Yes	Old	0	0	n/a
Yes	Young	0.0654	0.0408	a
No	Old	0.192	0.0715	a
No	Young	0.646	0.103	b

Fig. 2.

(A) Gamma generalized linear model fits of relative water content increase (RWC_gain) modelled as a function of relative water content at the dehydrated state (RWC_dehy24) for all combinations of Triton treatment and needle age. Due to the nature of the RWC_gain data (truncated at zero and positive continuous), the Gamma distribution was used to model the data. (B) Beta generalized linear model fits of relative water content increase normalized to capacitance (RWC_{gain_cap}) modelled as a function of relative water content at the dehydrated state (RWC_dehy24) for all combinations of Triton treatment and needle age. Due to the nature of the RWC_{gain_cap} data (falling in the interval of 0 and 1), the beta distribution was used to model the data.

The results of both approaches to calculate FWU, RWC_gain as well as RWC_{gain_cap} were analysed statistically. The regression diagnostics for RWC_{gain_cap} were somewhat weaker than for RWC_gain. However, both approaches resulted in identical statistical group differences (see Supplementary Data). In the following, the results for RWC_gain will be presented. The results of RWC_{gain_cap} are summarized in the Supplementary Data. To account for the observed effect of RWC_dehy24 on RWC_gain, RWC_dehy24 has been added as a covariate in all statistical analyses.

Across all species RWC_gain was lower in young needles than in old needles, and RWC_gain was promoted by Triton. These effects are statistically significant (Fig. 3 and Table 2).

Fig. 3.

Increase in relative water content (RWC_gain) for all Triton treatment and needle age combinations, pooled for all species. Error bars represent 95% confidence intervals. Significant differences (α = 0.05) are indicated by different letters. P-value adjustment was conducted using the Tukey method for comparing a family of four estimates. Note: error bars may be asymmetric since the Gamma distribution can model data which are asymmetric in shape (such as right skewed data).

Table 2.

Analysis of deviance table with Wald chi-square tests for the gamma generalized mixed-effects models 1 and 2

Model 1
Response: RWC  gain
	χ  2	d.f	Pr(>χ  2  )
RWC  dehy24	134.4	1	<0.0001
Triton	110.9	1	<2.2e-16
NeedleAge	12.7	1	0.0004
Species	16.1	6	0.013
Triton:NeedleAge	3.1	1	0.079
Model 2
Response: RWC  gain
	χ  2	d.f.	Pr(>χ  2  )
RWC  dehy24	24.7	1	<0.0001
Plug	10.58	1	0.005

Analysing the results separately by species reveals species-specific FWU (Fig. 4). Both effects, needle age and treatment with Triton, are, however, significant for each species: in all species, FWU is higher in older needles than in younger needles and enhanced by treatment with Triton. In younger needles treated with Triton, FWU increases to a similar level to that of older untreated needles (Fig. 4).

Fig. 4.

Increase in relative water content (RWC_gain) for all Triton treatment and needle age combinations split by species. Error bars represent 95% confidence intervals. Significant differences (α = 0.05) are indicated by different letters. P-value adjustment was conducted by species using the Tukey method for comparing a family of four estimates.

Dependence of FWU on stomatal wax plug thickness

Statistical analysis of the effect of stomatal wax plug thickness category on FWU was conducted for young needles without treatment with Triton, because in this group the plug is still largely unaltered by erosion (see subsequent section). Therefore, the influence of the original stomatal wax plug thickness should be particularly distinct in this group. The data indicate that the thickness of the stomatal wax plug affects FWU, with RWC_gain being significantly promoted by a thin plug (Fig. 5).

Fig. 5.

Increase in relative water content (RWC_gain) for the wax plug categories 1, 2 and 3. Category 1 consists of Pinus canariensis; category 2 consists of P. contorta, P. jeffreyi, P. mugo and P. sylvestris; category 3 consists of P. ponderosa. Statistical analysis was conducted for young needles without treatment with Triton, because in this group the plug is still largely unaltered by erosion. Pinus wallichiana was excluded from this analysis, because – under these conditions – no net RWC gain occurred for this species. Error bars represent 95% confidence intervals. Significant differences (α = 0.05) are indicated by different letters. P-value adjustment was conducted using the Tukey method for comparing a family of four estimates. Note: error bars may be asymmetric since the Gamma distribution can model data which are asymmetric in shape (such as right skewed data).

Needle surface

In the following, the terminology for epicuticular wax structures of Barthlott et al. (1998) is applied. The epicuticle of the considered species consists mainly of a crust often showing longitudinal grooves (Figs 6A and 7A; Supplementary Data Figs S1–S4). This crust is partially covered with wax crystals, mainly tubules, occurring preferentially or exclusively in relation with stomata (Figs 7A, B and 8A, C; Figs S2–S4, S7 and S8). Stomatal rows are partially or completely covered by wax tubules; the latter also form the stomatal wax plugs (Figs 6–8; Figs S5, S6 and S8). The wax structures are subject to erosion and therefore differ in older leaves compared to current-year needles, as will be considered further below. The following descriptions refer to the state of current-year needles.

Fig. 6.

Needle surface and stomatal region of Pinus canariensis (cryo-SEM). (A) Surface of a young needle with longitudinal grooves and a stomatal band. (B) An epistomatal chamber of a young needle, showing the thin wax plug and allowing the stomatal pore slit to be recognized. (A) A longitudinal section through a stoma and the epistomatal chamber of a young needle. The chamber is almost empty, due to the sparse amount of wax tubules. White arrow: guard cell; white circle: aperture of epistomatal chamber. (D) An epistomatal chamber of an old needle. The wax tubules forming the stomatal wax plug are altered by erosion, showing fusion and agglutination. Large pores are present in the tubule mass.

Fig. 7.

Needle surface and stomatal region of Pinus jeffreyi (cryo-SEM). (A) Surface of a young needle with longitudinal grooves and a stomatal band. The surface region along the stomatal band is densely covered with wax tubules. (B) An epistomatal chamber of a young needle. The stoma is occluded by the stomatal wax plug, which does not completely fill the chamber. (C) A longitudinal section through an epistomatal chamber of a young needle. The walls of the chamber are covered by wax tubules. The stomatal wax plug fills approximately the lower half of the chamber. Black circle: aperture of epistomatal chamber. (D) An epistomatal chamber of an old needle. The chamber aperture is covered with debris whereas the wax tubules which covered the chamber walls have been eroded away. There is also debris in the epistomatal chamber. The wax tubules forming the stomatal wax plug have been altered by erosion.

Fig. 8.

Needle surface and stomatal region of Pinus sylvestris (cryo-SEM). (A) Surface of a young needle with stomatal bands. The surface is densely covered with wax tubules. (B) A longitudinal section through an epistomatal chamber of a young needle. The walls of the chamber are covered by wax tubules. The stomatal wax plug fills approximately the lower half of the chamber and is more dense in its lower part. The upper part of the stomatal wax plug consists of loosely packed long wax tubules. Black circle: aperture of the epistomatal chamber. (C) An epistomatal chamber of a young needle. The stoma is occluded by the stomatal wax plug which does not completely fill the chamber. (D) Needle surface with stomatal bands of an old needle. The wax tubule cover has eroded. (E) An epistomatal chamber of an old needle with eroded stomatal wax plug.

Extension of the tubule cover is species-specific. For instance, stomatal rows in P. jeffreyi are covered completely with tubules (Fig. 7A), whereas a tubule cover is sparse or even absent in P. canariensis (Fig. 6A) and P. contorta (Supplementary Data Fig. S1). The stomatal wax plugs consist of a lattice of wax tubules, and their size and density are species-specific. Pinus canariensis shows the most weakly pronounced plug (Fig. 6B, C). It consists of a thin layer of irregularly arranged, perpendicularly oriented or sloped tubules barely covering the guard cells which are partially visible in various cases. The wax plug in P. canariensis was therefore assigned to ‘plug class 1’ (see stomatal wax plug categories summarized in Fig. 1), as the only one of the considered species showing this trait. In the other considered species, the wax tubule bodies forming the plug either fill the epistomatal chamber partially (‘plug class 2’, Fig. 1), such as in P. sylvestris (Fig. 8B, C; Figs S5 and S8) or completely (‘plug class 3’, Fig. 1) such as in P. ponderosa (Fig. 1; Fig. S6). Florin rings are distinct in most of the considered species (such as P. jeffreyi, Fig. 7A, P. sylvestris, Fig. 8A, D and P. wallichiana, Fig. S4).

When compared to current-year needles, erosion of wax structures is obvious in older needles. If present, the tubule structures along the stomatal bands are eroded or appear to be smeared on the surface of older needle leaves (Fig. 8D). The surface of older needles can be covered with deposited particles (Supplementary Data Fig. S10), including spores (Fig. S12) and fungal hyphae (Figs S9 and S11). Stomatal wax plugs are also affected by erosion in older needles (Figs 6D, 7D and 8E; Fig. S9). Fusing and agglutination of tubules occur (Fig. 6D) and can lead to larger pores within the eroded plug matrix (Fig. 8E), or the wax plug becomes lost completely (Fig. S9). Also, particles can be deposited in the epistomatal chamber (Fig. 7D; Fig. S9).

Application of Triton leads to substantial alteration and reorganization of wax crystals. The surface of the cuticle crust becomes amorphous, with Triton residues appearing like a highly porous cover (Supplementary Data Figs S13 and S14). Superficial wax tubules often smear (Fig. S14). Also, fusing and agglutination of wax tubules is promoted after Triton treatment. These alterations also occur with the wax tubules forming the stomatal plugs; these can become completely covered with Triton and amorphous with distinct holes in the eroded plug body (Figs S14–S16).

DISCUSSION

Comparison with previous studies

In immersion experiments with conifers, an increase in relative water content of about 3–10 % was observed (Limm et al., 2009; Berry and Smith, 2014; Laur and Hacke, 2014; Carmichael et al., 2020). However, direct comparison of these data with the data obtained in this study is difficult due to various methodological differences. For instance, entire shoots were often immersed instead of single needles or the immersion times differ among studies. Also, dehydration level was usually not considered as well as leaf age.

Dependence of FWU on dehydration and capacitance

Besides conductance of the leaf surface for water, FWU depends on the water potential gradient between the external water source and the leaf interior and is limited by the maximum capacitance for water uptake (Goldsmith et al., 2017; Boanares et al., 2018). Leaf rehydration can therefore be described as being analogous to the charging of a capacitor (Guzmán-Delgado et al., 2018) and the water uptake rate follows an exponential kinetics which can be expressed as ΔC = ΔC_max (1 − e^−kt) (where ΔC = difference between actual and initial leaf water content, ΔC_max = difference between maximum and initial leaf water content, k = absorption coefficient which can be understood as conductance divided by capacitance, and t = time which is 24 h in our case) (Boanares et al., 2018, and citations therein).

Therefore, the dehydration level greatly affects FWU. However, bench drying of the pine needles in the present study resulted in different levels of dehydration as shown by the different RWC_dehy24 values, probably promoted by variation of environmental conditions during the period of sampling. For instance, needles of P. canariensis were collected from plants growing in a glasshouse, and these needles showed a relatively uniform dehydration level after bench drying. As illustrated by the results of the present study, the dependence of FWU on the dehydration level (and probably also differences in C_max) can potentially obscure intra- and interspecific differences in FWU as well as effects of treatments (here with Triton), making capacitance a substantial confounding factor. Showing FWU as RWC_gain plotted against RWC_dehy24 accounts for this influence of the dehydration level on leaf water uptake (Fig. 2A). Here, the declining ability for water uptake with decreasing initial dehydration level (higher values of RWC_dehy24) is reflected by the converging shape of the graph showing RWCgain plotted against RWCdehy24 for all groups (Fig. 2A), It is therefore important to include the dehydration level as a covariate into statistical analyses of RWC_gain.

In the case of RWC_{gain_cap}, however, RWC_dehy24 should have no effect unless the absorption coefficient k changes with the hydration state. In fact, RWC_dehy24 has apparently no effect on RWC_{gain_cap} in the case of young needles (Fig. 2B). In contrast, the data for RWC_{gain_cap} of old needles show a weak positive trend with RWC_dehy24, for untreated needles as well as for needles treated with Triton (Fig. 2B). This may indicate that the absorption coefficient is negatively affected by rehydration for old needles. Currently, we cannot explain this phenomenon. It has been suggested that drying may lead to contractions of cuticle material (Boanares et al., 2018), but this is speculative, given that young needles should be also affected in that case. A possible effect of hydration on water uptake kinetics may merit further studies.

Needle-age related differences in FWU

Stomata are the most probable entrance site for external water in conifers (Berry et al., 2019). There are reports of higher FWU in older needles compared to younger needles (S. sempervirens, Burgess & Dawson, 2004; P. torreyana, Tianshi & Chau, 2022), and it was suggested that particularly stomata of older leaves are preferential sites of entrance for water due to degradation and erosion of wax (Burgess and Dawson, 2004). Principally, higher FWU in older needles as compared to current-year needles might be attributed to: (1) higher stomatal aperture (leading to higher stomatal conductance), (2) higher cuticle permeability or (3) possibly higher physiological ability to absorb and symplastically transport external water in mesophyll cells. For the third process, aquaporins are important. Laur and Hacke (2014) found enhanced expression of aquaporins in various needle tissues of Picea glauca after exposure to high-humidity air, indicating facilitated symplastic water transport. To our knowledge, however, there are no data on age-related aquaporin expression in coniferous needles.

With respect to cuticle conductance, g_min, Anfodillo et al. (2002) found for P. cembra that older needles show somewhat higher g_min than current-year needles. Cuticle conductance is, however, much lower than stomatal conductance g_s (Hubbard et al., 1999; Han et al., 2008; Wyka et al., 2016). Age-related g_s data for Pinus are sparse. However, the available data do not support a distinct age-related difference of g_s (Han et al., 2008; Wyka et al., 2016).

Wax erosion and water uptake

It has been discussed that foliar water uptake via stomata should be possible by water vapour absorption (‘reverse transpiration’) (Vesala et al., 2017; Binks et al., 2020) and not by absorption of liquid water. Guzmán-Delgado et al. (2021) found evidence for stomatal water vapour uptake in the angiosperm species Pyrus communis and Prunus dulcis. In addition, the rehydration of P. glauca in air saturated with humidity observed by Laur and Hacke (2014) supports the concept of FWU by absorption of water vapour.

Eichert et al. (2008) confirmed, however, stomatal uptake of liquid water by observing suspended nanoparticles penetrating stomata. There is also evidence that leaf surface wettability affects FWU (Berry et al., 2019) which should be not relevant for water vapour absorption. Available contact angle (CA) data do not indicate that Pinus needles are distinctly hydrophobic (meaning a CA > 90°) (Leyton and Armitage, 1968; Pogorzelski et al., 2014; Tianshi and Chau, 2022). Both Pogorzelski et al. (2014) and Tianshi and Chau (2022) found that CA tends to decrease with increasing needle age, caused by aerosol deposition, colonization with microorganisms and wax erosion (Turunen and Huttunen, 1990; Van Gardingen et al., 1991; Burkhardt and Pariyar, 2014; Burkhardt et al., 2018). In addition, in the material used in the present study, the surfaces of current-year needles differ greatly from those of older needles, particularly at the region of the stomatal bands (Fig. 8E). Accumulation of aerosols cannot be distinguished from eroded wax crystals (Burkhardt and Pariyar, 2014).

Erosion also occurs in the stomatal wax plug material of older needles (Figs 6D, 7D and 8D, E; Supplementary Data Figs S9 and S11). It is therefore conceivable that eroded stomatal wax plugs are more permeable for water. Also, the enhancing effect of Triton on RWC_gain found in this study can be ascribed to wax plug erosion, because surface tension depression by this surfactant is not sufficient to allow for stomatal infiltration (Schönherr and Bukovac, 1972). Triton was reported to alter and dissolve the wax crystal layer when applied to leaves (Noga et al., 1988). In fact, according to the cryo-SEM images, Triton also affected wax tubules, including those forming the stomatal wax plugs.

Besides altering the wettability of the stomatal region, aerosols deposited on a leaf can also create a hydrophilic layer which extends into the leaf interior and could act as a kind of ‘wick’ with the ability for bidirectional water transport (Burkhardt, 2010). When exposing conifer needles or branches to a particle-laden airstream, Burkhardt et al. (1995) found that particles mostly settled close to stomata and stomatal rows of P. sylvestris, and concluded that the micro-roughness of the epicuticular wax tubules covering the stomatal region enhanced particle deposition. Besides erosion, colonization of the needle surface and particularly the stomatal region, including the stomatal wax plugs themselves (Deckert et al., 2001), may promote stomatal water uptake, as was already suggested by Burgess and Dawson (2004). For instance, the epistomatal chamber in P. strobus can – depending on the growing site – be colonized by fungal hyphae which are sheathed by a polysaccharide matrix, and chambers colonized by fungi showed degraded wax plugs (Deckert et al., 2001).

Interspecific differences

The data obtained in the present study indicate particularly high FWU in P. canariensis. This result is consistent with various evidence for P. canariensis benefitting from fog events. Pinus canariensis is endemic to mountain regions in the Canary Islands where orographic lifting of moist winds leads quite frequently to fog formation on slopes (Fernández-Palacios and de Nicolás, 1995). It was found that P. canariensis is adversely affected by lower fog frequency (Rozas et al., 2013). The high capacity for FWU observed in this study supports the ecophysiological importance of fog harvesting in P. canariensis, which involves direct water uptake by needles. The thin wax plug in P. canariensis may therefore be adaptive for efficient FWU. Interestingly, P. canariensis is able to control transpiration tightly (Rozas et al., 2013, and citations therein), and this appears to be inconsistent with the presumed role of stomatal wax plugs as an anti-transpiration adaptation (Jeffree et al., 1971; Yoshie and Sara, 1985; Brodribb and Hill, 1997). It was found that xeromorphic traits develop in needles of P. canariensis trees growing in drier areas (Grill et al., 2004). Possible changes in stomatal wax plugs were, however, not considered in the study of Grill et al. (2004). Note that P. canariensis was the only species in this study which was cultivated in a glasshouse. This might influence the cuticle because of lower levels of UV radiation which can affect cuticles in Pinus, as was shown for mediterranean species (P. halepensis and P. pinea), which showed increased cuticle thickness under enhanced UV-B radiation, thereby possibly increasing cuticle resistance (Paoletti, 2005). Cuticle resistance, however, is much higher than stomatal resistance (Hubbard et al., 1999; Han et al., 2008; Wyka et al., 2016) and it is therefore unlikely that this putative effect might have influenced FWU in P. canariensis substantially. To our knowledge, no reports on the effects of UV on epistomatal chamber structure are available in the literature.

Whereas a high ability for FWU would therefore fit with the generally acknowledged importance of fog for P. canariensis, the substantial amount of water uptake in needles of P. sylvestris – as found in this study – was unexpected. To our knowledge, no systematic study has considered exploitation of atmospheric water (including dew) by P. sylvestris. As noted in the preceding section, particles accumulated on P. sylvestris needles mostly in the zones of wax tubules, and therefore at the stomatal bands (Burkhardt et al., 1995). We therefore suggest that needles of P. sylvestris are able to absorb water from fog, mist or drizzle events, thereby using all available humidity sources for water supply.

CONCLUSIONS

To briefly summarize, the results obtained in this study indicate that FWU in Pinus is species-specific and tends to be higher in older needles than in current-year needles. The results also illustrate the dependence of FWU on the hydration status of the needles. Thick stomatal wax plugs appear to reduce the ability for FWU and may contribute to the interspecific differences in FWU of Pinus. The tendency for higher FWU in older needles may well be explained by erosion of stomatal wax plugs. However, given the inter- and intraspecific differences in FWU and the confounding effect of the dehydration state, as well as micro-structural plug traits, more data are desirable to elucidate the effect of stomatal plugs on FWU in Pinus. Since stomatal wax plugs are widespread in conifers, their influence on FWU may be of general importance.

SUPPLEMENTARY DATA

Supplementary data are available online at https://academic.oup.com/aob and consist of the following.

Table S1: Measurement data.

Statistics S1: Residuals with respect to Triton and needle age. Residuals with respect to stomatal wax plugs.

Figure S1: Needle surface of Pinus contorta with stomatal band.

Figure S2: Needle surface of Pinus mugo with stomatal band. Figure S3: Needle surface of Pinus ponderosa with stomatal bands.

Figure S4. Needle surface of Pinus wallichiana with stomatal bands.

Figure S5. Epistomatal chamber with stomatal wax plug of Pinus contorta.

Figure S6. Epistomatal chamber with stomatal wax plug of Pinus ponderosa.

Figure S7. Needle surface of Pinus wallichiana, covered with wax tubules.

Figure S8. Needle surface of Pinus mugo, with stoma.

Figure S9. Epistomatal chamber of an old needle of Pinus sylvestris.

Figure S10. Particles on the surface of an old needle of Pinus jeffreyi.

Figure S11. Two epistomatal chambers with Florin rings on the surface of an old needle of Pinus wallichiana.

Figure S12. Pollen on the surface of an old needle of Pinus wallichiana.

Figure S13. Surface of a needle of Pinus jeffreyi treated with Triton.

Figure S14. Surface of a needle of Pinus wallachiana with aperture of an epistomatal chamber, treated with Triton.

Figure S15. Surface of a needle of Pinus mugo with aperture of an epistomatal chamber, treated with Triton.

Figure S16. Aperture of an epistomatal chamber of a needle of Pinus mugo treated with Triton.

FUNDING

U.H. acknowledges funding from a Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery grant.

CONFLICT OF INTEREST

The authors declare they have no conflict of interest.