Significance of current weather conditions for foliar traits of old-growth sessile oak (Quercus petraea Liebl) trees

Sessile oak leaves showed a high degree of plasticity to atmospheric and pedospheric conditions. The aim of the present study was to elucidate the significance of current weather conditions for foliar traits of adult sessile oak (Quercus petraea), one of the most valuable forest tree species in Central Europe. For this purpose, structural and functional traits were analysed in fully expanded, sun exposed leaves collected in south-west Germany from five old-growth forest stands, representing the meteorological and pedospheric conditions in the growing region, but differing in aridity during the 12 days before harvest in two consecutive years. Across the forest stands, most foliar traits differed significantly between wet and dry weather conditions before harvest as indicated by partial least square discriminant analysis (PLS-DA). These traits included fresh weight/dry weight ratio, leaf hydration, leaf-C content, leaf-C/N ratio, structural N, soluble protein-N, total amino acid-N, cell wall composition, numerous specific amino acids as well as soluble sugar content. Structural biomass, δ13C signature, total N and total C as well as H2O2 contents were not affected by the weather before harvest. These results indicate a high plasticity of the foliar metabolism of drought-tolerant sessile oak to current weather conditions. They also suggest that sessile oak is characterized by a high potential to cope with the growth conditions expected as a consequence of future climate change.


Introduction
Besides beech (Fagus sylvatica L.), oak species (Quercus spec.) are the most common deciduous forest tree species in Central Europe (Thomas and Gausling 2000;Hanewinkel et al. 2013). The genus Quercus comprises ~ 400 deciduous and evergreen trees and shrubs occupying a wide variety of habitats in temperate, Mediterranean, subtropical and tropical areas (Nixon 2006). Although oaks are considered to be more tolerant to water shortage than most other deciduous trees in Central Europe, low water availability can limit their growth and development (Bruschi 2010;Jensen and Hansen 2010). In addition, different oak species vary considerably in susceptibility to water shortage. In a macrocosm experiment under controlled conditions, Hu et al. (2013aHu et al. ( , b, 2015 investigated stress responses to drought and increasing temperature in leaves of Quercus robur L., Quercus petraea (Matt.) Liebl., and Quercus pubescens Willd., three widely distributed oak species in Central Europe. The authors found Communicated by T. Koike. that saplings of all three oak species contained higher foliar antioxidant and free amino acid levels in response to drought and air warming. However, these effects were species-specific with highest levels for drought-tolerant Q. pubescens. In addition, effects were more pronounced in ecotypes originating from drier habitats across the species studied (Hu et al. 2013a). Elevated levels of free amino acids in response to drought and increasing temperature were achieved in oak leaves at the expense of structural nitrogen (N), indicating high partitioning flexibility of leaf N contents under stress conditions (Hu et al. 2013b(Hu et al. , 2015. Antioxidative defence and osmotic stress adjustment by amino acid accumulation in oak leaves were more pronounced on calcareous soil with low water holding capacity compared to acidic soil with higher water holding capacity (Hu et al. 2013b(Hu et al. , 2015. These results clearly show that functional traits such as ROS (reactive oxygen species) scavenging capacity by antioxidants, and partitioning flexibility of N influence the tolerance of oak leaves to drought and increasing temperature as well as the plasticity of N metabolism. Because osmotic adjustment by amino acid accumulation was achieved at the expense of structural N that is largely determined by lignin contents, it may be assumed that the change in cellular N partitioning goes hand-in-hand with general changes in cell wall structure and composition. Therefore, measuring amino acids, protein and lignin precursors as well as compatible solutes such as carbohydrates involved in stress defence will provide valuable information on responses of oak trees to changing environmental conditions.
Previous studies indicated that plants of the same species grown in different geographical locations with different environmental conditions contain different levels of carbohydrates depending on water availability and photosynthesis (Scossa et al. 2016;Zanetti et al. 2016). However, the effects of aridity on carbohydrate composition and contents are complex, depending on intensity and duration of drought on one side and the tree species, size, age and tissue on the other side (Hartmann et al. 2015). This is due to the functional diversity of carbohydrates, acting as compatible solutes, precursors of structural polymers as well as mobile and dynamic storage compounds. In addition, soil water availability often shows a strong positive relationship to the δ 13 C signature as long-term average indicator of stomatal conductance as previously reported also for oak trees (Q. robur and Q. petrea, Ponton et al. 2002). Ramírez-Valiente et al. (2010) associated differences in δ 13 C signature of Q. robur with rainfall and temperature at the sites of origin, suggesting local adaptation in response to diverging weather conditions.
To prevent oxidative stress upon water shortage, not only reduced ROS generation, but also enhanced ROS scavenging is required for ROS homeostasis. Scavenging of ROS is achieved in plant cells by chemical reactions with antioxidants such as ascorbate (Asc) and glutathione (GSH), but largely depends on the enzymatic regeneration of these antioxidants in the Foyer-Halliwell-Asada pathway that uses these metabolites as co-substrates (Munné-Bosch et al. 2014;Noctor et al. 2012). Hence, the antioxidative capacity of plant cells depends on the activity of antioxidative enzymes that control the flux of electrons through the Foyer-Halliwell-Asada pathway in different cellular compartments, i.e. dehydroascorbate reductase and glutathione reductase activity. In addition, it depends on the levels of the antioxidants GSH and Asc in the cellular compartment(s) of ROS production (Wrzaczek et al. 2013;Gill and Tuteja 2010). In a study with different oak species, Hu et al. (2013b) found that drought-sensitive Q. robur reinforced its oxidative stress defence systems by enhancing leaf Asc and thiol levels. By contrast, the more drought-tolerant species Q. petraea and Q. pubescens were able to sufficiently scavenge the ROS produced upon oxidative stress at constitutive foliar Asc and thiol levels. Apparently, differences in stress tolerance between oak species seem to be a consequence of differences in ROS scavenging rather than ROS production.
The present study was carried out during summer 2014 and 2015 in five old-growth stands of sessile oak (Quercus petraea) in south-west Germany, one of the economically most important distribution area of this species in Central Europe. The stands chosen represent the meteorological and pedospheric conditions in the growing region, but differed in aridity during the 12 days before harvest in the two consecutive years studied. The aim of this study was to identify responses of foliar structural and physiological traits to current weather conditions before harvest using the De Martonne aridity index based on temperature and precipitation (De Martonne 1926). Investigations were performed with special emphasis on cell wall composition, ROS scavenging as well as nitrogen and carbon partitioning. In this context, we tested the hypothesis that current weather conditions determine structural as well as physiological traits in leaves of adult Quercus petraea trees.

Forest stand characteristics
In the present study, leaves of old-growth sessile oak trees from Rhineland Palentine in midwest Germany, one of the most extended growing region of this species, were investigated. For this purpose, five forest stands were selected that largely cover old-growth sessil oak origin and age as well as pedospheric and meteorological conditions of sessile oak stands in the region (Table 1; Fig. 1). Stand origins included natural regeneration, seed plantation and coppice (Table 1). Tree age varied between 94 and 172 years; the Table 1 Field site characteristics of the Quercus petraea stands Information on soils, stand origin, and stand age were provided by the Forest Services of the State of Rhineland-Palatinate. Oak share was calculated on a surface area basis. Leave area index (LAI) was estimated from measurements in a regional oak stand (Bréda et al. 1995).Weather conditions were calculated as mean day-aridities for the 12 days before leaf harvest from precipitation and temperatures as indicated in "Materials and methods" Name of stands  oak share between 26 and 63%, all with a predominating oak crown layer, but different admixed species; stand volume of oak trees between 30 and 327 m 3 ha −1 ; and leaf area index of oak trees between 2.1 and 4.7 m 2 m −2 ( Table 1). The soils of all stands investigated constitute Luvisols to Podzoluvisols (FAO classification) derived from pleistocene periglacial solifluction layers with poor to middle nutrient supply (Tables 1, 2). We used the aridity index (De Martonne 1926) as an integrative variable to characterize weather dryness at the sites of growth. This indicator served previously to identify, locate or delimit origins that are prone to suffer from a deficit of water availability (Baltas 2007;Paparrizos 2018). The index was calculated by the following equation: where P is the mean precipitation (mm) and T (°C) the mean air temperature. The De Martonne aridity index decreases with increasing aridity. Current weather conditions were calculated as the 12 day period that preceded leaf harvest for each stand (Table 1) (see supplementary material) (Seegmueller 2012). Foliar traits were compared between the relatively arid weather conditions before harvest in 2014 and the cooler and wetter weather conditions before harvest in 2015 (Table 1).  Leaf harvest and sample preparation Sun exposed leaves were harvested in mid June 2014 and 2015 at the full expansion of the first flush from same sessile oak trees (Quercus petreae Liebl.) evaluated as undamaged and healthy by visual inspection. Bud break was during May and leaves were fully differentiated at leaf harvest. Within each stand, eight predominant or dominant oak trees as determined by Krafft tree classes were harvested (Kramer 1988). Leaves were taken from light-exposed crown branches harvested by the slingshot technique ( Fig S1). Leaves samples were taken from trees of 16-30 m height (Table 1). Ten leaves were collected from every branch. Each of the eight trees was assessed in three replicate branches. All leaf samples were immediately frozen in liquid N 2 and stored at − 80 °C until further analyses.

Soil sampling
Soil samples were taken in the outer areas of the crown projection from the assessed oak trees summing up to eight replicates per stand in 2015. Samples were collected with a soil auger from 0 to 30 cm soil depths (A: 0-10 cm and B:10-30 cm), representing the main fine root zone under local conditions (Kern et al. 1961). The soil samples were frozen at − 80 °C until further preparation. For soil preparation, samples were sieved (< 2 mm) to remove roots and coarse material and dried at 60 °C for 48 h to determine total C and N contents.

Dry mass and leaf hydration determination
Leaf samples were homogenized in liquid N 2 to a fine powder and aliquots of 100 mg were dried at 60 °C for at least 48 h to weight constancy for dry mass determination. Leaf hydration (H; g H 2 O g -1 DW) was determined as (FW-DW)/DW, where FW is the fresh mass and DW is the dry mass (Contin et al. 2014).

Biochemical analyses
Determination of δ 13 C signature, total C and N contents in leaf and soil material δ 13 C signature as well as total C and N contents in leaf material dried at 60 °C for at least 48 h to weight constancy were determined in aliquots (0.8-1 mg) weighed into tin capsules (IVA Analysentechnik, Meerbusch, Germany). Samples were analyzed using a C/N elemental analyzer (NA 2500, CE Instruments, Milan, Italy) coupled via a Conflo II interface (Finnigan MAT GmbH, Bremen, Germany) to an isotope ratio mass spectrometer (Delta plus, Thermo Finnigan MAT GmbH, Bremen, Germany) as previously described in Geßler et al. (2005). For analysis of soil samples according to Liu et al. (2015a), 4 mg soil from the A (0-10 cm), and 20 mg from the B (10-30 cm) horizon were weighed into tin capsules and analysed using a Costech elemental analyzer (Costech International S.p.A., Milan, Italy) fitted with a Zero Blank autosampler, and coupled via a ConFlo-III interface to a Thermo Finnigan Delta V Plus isotope ratio mass spectrometer (Thermo Scientific, Waltham, MA, USA). Glutamic acid was used as the working standard and calibrated against the international standard USGS-40 (glutamic acid, δ 13 C PDB = − 26.39‰) for δ 13 C analyses. Working standards were analyzed after every tenth sample to detect any potential instrument drift over time.

Determination of cell wall composition
Structural biomass was determined after removal of soluble cellular constituents (Arab et al. 2019). Lignin contents were analyzed by the thioglycolate method (Brinkmann et al. 2002), cellulose contents with the anthrone reagent (Updegraff 1969) using a UV-DU650 spectrophotometer (Beckman Coulter Inc., Fullerton, CA, USA) as previously described (Arab et al. 2019).

Determination of glutathione (GSH) and ascorbate (Asc), hydrogen peroxide contents, glutathione reductase (GR) and dehydroascorbate reductase (DHAR) activities
GSH and Asc were extracted from 50 mg frozen and powdered leaf material using 0.1 N HCl in the presence of 50 mg of polyvinylpolypyrrolidone and 10 µL 1 mM EDTA (Schwanz et al. 1996). From this extract, ascorbate was separated isocratically on a HP 1090 series II HPLC (Agilent, Waldbronn, Germany) at 40 °C with 10% methanol (MeOH), pH 2.0 and determined at λ = 248 nm (Schwanz et al. 1996). GSH was determined in 200 µL of the above extract after reduction of GSSG (oxidized glutathione) and derivatisation with monobromobimane (mBBr) (Schupp andRennenberg 1988, Strohm et al. 1995 (Schupp et al. 1991). Hydrogen peroxide contents were quantified with the potassium iodide assay using a UV-DU650 spectrophotometer at λ = 390 nm (Beckman Coulter Inc., Fullerton, CA, USA) as previously reported (Loreto and Velikova 2001). GR activity was measured kinetically from the NADPH consumption for GSSG reduction with a spectrophotometer at λ = 340 nm (Polle et al. 1990). DHAR activity was assayed directly by following the increase in absorbance at 1 3 265 nm, resulting from GSH-dependent production of Asc from DHA (dehydroascorbate) (Polle et al. 1990).

Determination of soluble protein contents and total amino acid
Total soluble protein in leaf samples was quantified as reported by Dannenmann et al. (2009). For determination of total amino acids, aliquots of 50 mg frozen and powdered leaf material were extracted according to Winter et al. (1992) in 200 µL 20 mM HEPES buffer (pH 7.0), containing 5 mM ethylene glycol tetraacetic acid (EGTA), 10 mM NaF and 1 mL methanol/chloroform (3.5:1.5 v/v). Total amino acids in the extracts were quantified after derivatization with ninhydrin using a UV-DU650 spectrophotometer (Beckman Coulter Inc., Fullerton, CA, USA) at λ = 570 nm. Specific amino compounds were analyzed using an ultra-performance liquid chromatography (UPLC) system (Waters Corp., Milford, MA, USA) as described by Luo et al. (2009).

Structural N calculation
Structural N content in plant material was calculated by subtracting the N fractions in total amino acids and soluble protein from total N. Since N assimilation takes place in the roots of sessile oak trees and, therefore, foliar nitrate contents are minute, nitrate was not considered in this calculation (Hu et al. 2013a).

Carbohydrate analysis
Carbohydrates were extracted from homogenized leaves samples and derivatized by a modification of the method by Kreuzwieser et al. (2009). For each sample, approximately 50 mg frozen leaf powder was weighed into a pre-frozen 2 mL round-bottom Eppendorf tube and 500 µL cold 95% (v/v) methanol (Sigma-Aldrich GmbH) were added as extraction medium. 30 µL of a solution of 2 mg mL −1 ribitol was used as internal standard. Tubes were rapidly heated to 70 °C and shaken at 1400g for 15 min. After brief vortexing, 200 µL chloroform and 400 µL distilled water were added, the samples were shaken at 1400g for 5 min, and centrifuged at 14,000g for 5 min. 10 µL aliquots of the supernatant were freezedried, methoxinated by adding 20 µL of a 20 mg mL −1 solution of methoxyamine hydrochloride in anhydrous pyridine (Sigma-Aldrich GmbH), and incubated at 30 °C for 90 min under shaking at 1400g. For trimethylsilylation, 35 µL N-methyl-N-(trimethylsilyl) trifluoroacetamide (MSTFA; Sigma-Aldrich GmbH) were added to each tube. Tubes were incubated at 37 °C for 30 min with 1400g shaking. Subsequently, 5 µL of an alkane standard solution (C8-C20, Sigma-Aldrich GmbH) were added to each sample and 50 µL supernatant were transferred to GC-MS vials (Agilent Technologies, Palo Alto, CA, USA) for GC-MS analysis. The derivatized metabolite samples were analyzed on an Agilent GC/MSD system consisting of an Agilent GC 6890 N gas chromatograph (Agilent Technologies, Wilmington, DE, USA) equipped with an autosampler (GC-PAL, CTC Analytics, Zwingen, Switzerland) and connected to a 5975C Inert XL EI MSD quadrupole MS detector (Agilent Technologies). The separation was achieved on a capillary column (HP-5MS 5% Phenyl Methyl Silox, length: 30 m, diameter: 0.25 mm, film thickness: 0.25 µm; Agilent Technologies) at a helium flow of 1 mL min −1 ; GC-MS run conditions were set according to Kreuzwieser et al. (2009). The raw data files were processed with the Mass Hunter (version B.07.00/ Build 7.0.457.0) workstation software for GC-MS (Agilent Technologies). Mass spectra were searched against the Golm Metabolome Database (Kopka et al. 2005) and carbohydrates were identified based on internal standards, spectrum similarity match, and retention time.

Statistical analysis
All analytical measurements were performed with leaves of eight replicate trees of five forest sites each. This includes determination of water content, FW/DW ratios, δ 13 C, total C, total N, SBM, lignin, cellulose, hydrogen peroxide, glutathione and ascorbate contents, total protein-N, total amino acid-N, specific amino acids, carbohydrate compounds as well as enzyme activities of glutathione reductase (GR) and dehydroascorbate reductase (DHAR). Data were positively tested for normal distribution and homoscedasticity. Significant differences between pooled data from all stands of each year were assessed using the unpaired Student's t test. Moreover, to characterize differences between the years, datasets were subjected to partial least square discriminant analysis (PLS-DA) using Meta-boAnalyst 3.0 (Xia et al. 2015).

Results
Weather conditions before harvest were similar for all stands and relatively dry in 2014, whereas in 2015, weather conditions before harvest were cooler and wetter compared to 2014, as indicated by a higher aridity index (Table 1). Thus, water availability before harvest was higher in 2015 compared to 2014 across all stands. Weather conditions before harvest strongly affected the foliar traits analysed in the present study. PLS-DA analysis across stands and parameters measured revealed a clear separation between the 2014 and 2015 ( Fig. 2A). To elucidate the foliar traits that determine this separation and, hence, the response to different weather conditions before harvest, PLS-DA analyses were performed with subsets of the foliar parameters studied across the five oak stands.
Full separation was observed for structural parameters investigated (Fig. 2C), although foliar SBM was similar in leaves harvested after arid and wet weather conditions before harvest across the sessile oak stands. However, total foliar C content was significantly reduced at the more arid weather conditions. This reduction was reflected by reduced lignin, sucrose and fructose contents of the leaves.
Since cellulose and glucose contents increased, also the cellulose/lignin and glucose/sucrose ratios were increased at more arid conditions before harvest (Table 3). In addition, the foliar contents of various other soluble carbohydrates including lyxose, sorbose, myo-inositol, and glucose-6-phosphate were enhanced at more arid conditions before harvest (Table 4). These increases resulted in a full separation of the foliar carbohydrate composition between arid and wet weather conditions before harvest across the five oak stands (Fig. 2D). Surprisingly, such a separation was not observed for the parameters of the antioxidative metabolism studied (Fig. 3A). The H 2 O 2 content of the leaves as an indicator of ROS accumulation was not significantly affected by more arid weather conditions before harvest (Table S1). The antioxidants ascorbate and glutathione as well as the activity of the antioxidative enzymes GR and DHAR as rate limiting components of the Foyer-Haliwell-Asada cycle of ROS scavenging were even reduced at more arid weather conditions (Table 3). These results indicate that enhanced ROS production at the arid weather conditions before harvest could be scavenged by the constitutive antioxidative metabolism of the leaves.
Similar to the antioxidatve metabolism of oak leaves, also N metabolism did not show a clear separation between wet and more arid weather conditions before harvest, neither of total N and its partitioning, nor for the specific amino acid composition (Fig. 3B, C). Whereas the foliar total N content was not affected by the weather conditions before harvest across the stands studied, soluble protein-N declined, structural N and amino acid-N increased at more arid weather conditions before harvest. The increase in amino acid-N was achieved by increased amounts of 3-phospho-glycerate derived, pyruvate derived and oxalo acetate derived amino acids (Table 3, S1). At the level of individual amino acids, this increase was achieved by significantly enhanced foliar glycine, isoleucine, glutamine, proline, aspartate, threonine, phenylalanine and tyrosine contents, whereas other amino acids were not significantly affected by the weather conditions before harvest (Table 5).

Impacts of current weather conditions on water relations and structural traits of sessile oak leaves
In the present study, foliar FW/DW ratio and leaf hydration were higher across stands in 2015 at wet conditions before harvest compared to 2014 at dry conditions before harvest. This result indicates modulation of foliar water relations by current water availability, as also observed in previous studies (Contin et al. 2014;Baslam et al. 2014;Suguiyama et al. 2016). The carbon isotope composition (i.e. δ 13 C signature) is widely considered a sensitive indicator of long-term water availability to plants, provided photosynthetic processes are not impaired (Warren et al. 2001). In the present study, leaf δ 13 C signatures did not differ between 2014 and 2015, indicating a limited responsiveness of this parameter to short-term aridity. Aridity did not affect the share of foliar SBM on dry weight in all stands in 2014, but reduced the lignin content of SBM. Apparently, oak stands respond to short-term aridity by changing the foliar cell wall composition, as also found in other species subjected to drought (Joly and Zaerr 1987;Donaldson 2002;Zwiazek 1991).

Impacts of current weather conditions on oxidative stress and its compensation in leaves
Arid weather conditions can stimulate the formation of reactive oxygen species (ROS), particularly at high irradiation (Munné-Bosch and Peñuelas 2004). Non-enzymatic antioxidants such as Asc and GSH and their regeneration in the Foyer-Haliwell-Asada cycle play a key role in the detoxification of ROS (Hossain et al. 2013). In the present study, the contents of both antioxidants were lower in 2014 at drier weather conditions before harvest compared to 2015 across stands. Since this was also observed for GR and DHAR activity that are involved in the regeneration of Asc and GSH, it appears that part of the antioxidative metabolism of sessile oak leaves is even down-regulated at dry weather conditions. This result is surprising, because also ROS levels such as H 2 O 2 contents that are thought to be kept under control mainly by the components of the Foyer-Haliwell-Asada cycle (Noctor et al. 2012;Schnaubelt et al. 2015) remained unchanged at drier weather conditions before harvest in 2014. Therefore, it may be concluded that in sessile oak leaves other antioxidative compounds such as secondary metabolites get enhanced significance for ROS scavenging upon dry weather conditions. This view is supported by a recent report from a free-air cross-exchange experiment with sessile oak saplings indicating a high significance of secondary metabolites in sessile oak leaves for ROS scavenging (Arab et al. 2020).

Impacts of current weather conditions on N and C contents and partitioning in leaves
Drought stress was found to increase the foliar N concentration in Malus domestica (Jie et al. 2010) and in Q. Robur (Picon et al. 1997). However, Dichio et al. (2007) reported that leaf N concentration decreased in drought-treated plants of Prunus persica. In the present study, neither differences in soil N contents nor in aridity before harvest changed the foliar total N contents significantly, supporting Fig. 3 A PLS-DA analysis of ROS and parameters of the antioxidative metabolism including H 2 O 2 (µmol g-1 DW), ascorbate (µmol g −1 DW), glutathione (nmol g −1 DW), GR activity (µmol g −1 DW), DHAR activity (µmol g −1 DW), B PLS-DA analysis of N content and its partitioning including total N (mg g −1 DW), soluble protein-N (mg g −1 DW), structural N (mg g −1 DW), total amino acid-N (mg g −1 DW), 3-phospho-glycerate derived amino acids (mg g −1 DW), pyruvate derived amino acids (mg g −1 DW), 2-ketoglutarat drived amino acids (mg g −1 DW), Oxalo acetate drived amino acids (mg g −1 DW), aromatic amino acids (mg g −1 DW), other amino compounds (mg g-1 DW). C PLS-DA analysis of specific amino acids concentrations (mg −1 g DW) including histidine, cysteine, serine, glycine, alanine, valine, leucine, isoleucine, glutamate, glutamine, proline, argenine, aspartate, asparagine, methionine, threonine, tryptophan, phenylalanine, tyrosine, ornithine, GABA, AABA ▸ the view that foliar N content is not subjected to short-term changes. In line with our results, Li et al. (2013) found that air warming and drought did not affect leaf N contents in different Quercus species⁄ecotypes. However, differences to short-term aridity before harvest modulated N partitioning in sessile oak leaves in the present study. We observed higher protein contents at lower N availability in the soil and decreasing soluble protein-N at the benefit of both total amino acid-N and structural N at dry weather conditions before harvest. Similar results were previously obtained for Quercus robur and Quercus petraea, where ecotypes from drier origins showed lower soluble protein but higher structural N compared with ecotypes from wetter areas (Hu et al. 2013a). Moreover, Kimura et al. (1998) and Uemura et al. (2000) showed that leaves of Fagus japonica tree were well adjusted to the current-year light environment both anatomically and physiologically as mass-based N and Chl concentrations and N rearrangements within a leaf substantially adjusted to the current-year light environment. The significance of winter bud traits on foliar traits of fully expanded leaves of mature trees also depends on the overall storage capacity of the trees that usually overroles bud traits at the end of leaf development. In addition, among the traits studied mostly soluble sugars and total nitrogen contents may be affected by bud traits. Since soluble sugars were largely higher in a wet year after a dry year and total N contents were not different, a significant influence of bud traits developed in the dry year on leaf traits in the subsequent wet year is unlikely. Still, we cannot exclude that bud traits developed in the previous year may have contributed to the differences in foliar traits observed. The authors assumed that this N partitioning pattern might be related to differences in leaf structural characteristics, but also indicates a tight control of N acquisition by the roots at sufficient N supply to the trees. Amino acids, amines, organic acids, sugars and sugar alcohols constitute major solutes that are responsive to drought stress. These primary metabolites perform important functions in plant adaptation to drought, because they can serve as energy reserves, osmolytes to maintain cell turgor, antioxidants, and/or as signal transduction molecules (Bartels and Sunkars 2005;Seki et al. 2007;Shulaev et al. 2008;Silvente et al. 2012). In the present study, drier weather conditions before harvest caused an obvious increase of foliar amino acid-N across stands in 2014 compared to 2015, mostly at the expense of soluble protein-N. In addition, a coordinated increase in the foliar concentration of amino acids derived from 3-phospho-glycerate, pyruvate and oxaloacetate was observed across all stands at drier weather conditions before harvest in 2014. Apparently, drought strengthens the use of N for osmotic adjustment, as also observed in a recent study in dwarf bamboo under shortterm drought stress (Liu et al. 2015b). These results support the view that accumulation of amino acids can aid stress tolerance in plants through osmotic adjustment, detoxification of reactive oxygen species, and by intracellular pH regulation (Silvente et al. 2012;Muttucumaru, et al. 2015). Also for other species, weather conditions were reported to cause a high variation in the concentrations of individual amino acids from year to year, although their relative proportions on total free amino acids remained constant (Ortega-Heras et al. 2014;Bouzas-Cid et al. 2018).
In the present study, the significant accumulation of foliar lyxose, glucose, sorbose, glucose-6-phosphate, myo-inositol, cellobiose, cellobitol, raffinose and mannitol contents across stands in 2014 under dry weather conditions before harvest suggest a close association between these metabolites and drought tolerance of sessile oak. Lu et al. (2009) also found a significant increase of total soluble sugar and sucrose in hybrid Bermudagrass under drought stress. Apparently, accumulation of soluble sugars in water-stressed tissues as osmoprotectants is a more general phenomenon in herbaceous and woody plants (Skirycz et al. 2010;Munnik and Vermeer 2010;Perera et al. 2008). Yoshimura et al. (2016) reported that the stored carbohydrate is quickly reformed between soluble sugar and starch within sapwood during several days, due to water stress. However, our results show that foliar carbohydrate contents were mostly lower in the dry compared to the wet year indicating that a compensation of reduced foliar carbohydrate accumulation under dry conditions by the mobilisation of carbohydrates from sapwood was not effective. Furtheremore, Viklund et al (2008) showed that the contents of free amino acids and sugars in potatoes can vary significantly between years and that these variations can be explained by differences in precipitation and temperature. Therefore, it can be assumed that the accumulation of carbohydrates in the leaves could contribute to an increased osmotic potential at reduced cell hydration, and may facilitate the maintenance of cell turgor, thereby mitigating injurious effects of shortterm drought to sessile oak leaves.

Conclusions
The present investigation shows that structural and physiological traits in leaves of adult sessile oak trees are largely determined by the weather conditions before harvest.
Thus, the central hypothesis of this study was validated for foliar traits such as lignin, cellulose and some soluble sugar (sucrose, glucose, fructose, and glucose-6-phosphate) contents. Our results also show that only few physiological traits of leaves from adult sessile oak trees are independent of current weather conditions across the five forest stands investigated, indicating a high degree of plasticity of sessile oak leaves to atmospheric and pedospheric conditions. However, a combined analysis of foliar and sapwood traits over an extended period of time with multiple sampling would provide more detailed information of plasticity of sessile oak leaves to claimate change in future.
Author contribution statement LA: performed most of the measurements, analyzed data and wrote the manuscript with input from all authors. SS: conceived and planned the experiments, analyzed the antioxidants and edited the manuscript. JK: helped with GC-MS measurements and edited the manuscript. ME: helped with UPLC measurements. MD: helped with soil measurements and edited the manuscript. HR: designed the project, supervised the study and edited the manuscript. All authors contributed to the final version of the manuscript.
Funding Open Access funding enabled and organized by Projekt DEAL. This study was part of the APEK project (No. 2047441501) funded by the Bundesministerium für Ernährung und Landwirtschaft (BMEL) and the Bundesminesterium für Umwelt, Naturschaft, Bau und Reaktorsicherheit (BMUB) based on the decision of the German Federal Parliament.

Conflict of interest The authors declare no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.