Leaf 13C and 15N composition shedding light on easing drought stress through partial K substitution by Na in eucalyptus species

This work aimed to investigate the partial K-replacement by Na supply to alleviate drought-induced stress in Eucalyptus species. Plant growth, leaf gas exchange parameters, water relations, oxidative stress (H2O2 and MDA content), chlorophyll concentration, carbon (C) and nitrogen (N) isotopic leaf composition (δ13C and δ15N) were analyzed. Drought tolerant E. urophylla and E. camaldulensis showed positive responses to the partial K substitution by Na, with similar dry mass yields, stomatal density and total stomatal pore area relative to the well K-supplied plants under both water conditions, suggesting that 50% of the K requirements is pressing for physiological functions that is poorly substituted by Na. Furthermore, E. urophylla and E. camaldulensis up-regulated leaf gas exchanges, leading to enhanced long-term water use efficiency (WUEL). Moreover, the partial K substitution by Na had no effects on plants H2O2, MDA, δ13C and δ15N, confirming that Na, to a certain extent, can effectively replace K in plants metabolism. Otherwise, the drought-sensitive E. saligna species was negatively affected by partial K replacement by Na, decreasing plants dry mass, even with up-regulated leaf gas exchange parameters. The exclusive Na-supplied plants showed K-deficient symptoms and lower growth, WUEL, and δ13C, besides higher Na accumulation, δ15N, H2O2 and MDA content.

. By contrast the N isotope composition (δ 15 N) is a good predictor of the plant growth status and N metabolism 21 . Plants with higher photosynthetic capacity and transpiration efficiency accumulate more N, the most growthlimiting nutrient element for plants. In this context, the δ 13 C and δ 15 N have been used to gain insights into the effect of growing conditions on C and N metabolism, as also the mechanisms responsible for plants responses to nutritional status and abiotic stress throughout its whole cycle, since WUE, N acquisition and C uptake may be correlated [22][23][24][25] . However, studies regarding the interactive effect of partial K substitution by Na on the δ 13 C and δ 15 N of the Eucalyptus seedlings grown under drought condition are scarce.
Despite the similarities between the Na + and K + hydrated ionic radii, the extent of K replacement by Na in plant nutrition not only varies among species, but can also be harmful. Most of agriculturally crops are intolerant to salt, being inhibited by high Na concentrations 15 . Under salinity conditions, which affect more than 20% of the global agriculture lands, the Na exceeds the cells' ability to compartmentalize Na + in the vacuole, increase the osmotic potential and disrupt cell membrane integrity 21 . As a result, Na accumulates in the cytoplasm, leading to enzyme activity inhibition 26 , and sudden death of entire shoots 27 . Currently, studies regarding Na in plants mainly focus on salinity conditions, i.e., > 100 mM of NaCl, and despite recent advances in understanding drought stress and K supply, little attention has been paid to the combined use of K and Na as an agricultural practice to optimize productivity and efficiency of resources in Eucalyptus under drought conditions. Therefore, the effect of partial K-replacement by Na on leaf 13 C, 15 N compositions and photosynthesis-related parameters remains unclear, especially in plants with different drought tolerances when subjected to soil-water availabilities. We hypothesized that (i) the partial K substitution by Na has the potential to alleviate the drought-induced stress in Eucalyptus by up-regulating growth, physiological and biochemical parameters and (ii) confirm the use of natural isotopic abundance (δ 13 C and δ 15 N) as a reliable indicator of the positive effects of partial K substitution by Na on C and N metabolism throughout plants metabolism cycle. Thus, we applied a multitiered approach at a range of leaf and plant levels to (i) characterize leaf C and N isotope compositions, leaf water potentials, oxidative stress; (ii) quantify the long-term plants water use efficiency; and (iii) evaluate K-and Na-uptake as well as eucalyptus species growth.

Results
The relative effects (increase or decrease) of partial K replacement by Na (50/50% of K/Na) and exclusive Nasupply (0/100% of K/Na) compared to the well K-supply (100/0% of K/Na) in E. saligna, E. urophylla and E. camaldulensis grown under W+ and W− condition can be observed in Supplementary Table 1S. camaldulensis seedlings under three soil levels of K/Na in well-watered (W+) and water-stressed (W−) condition. Different uppercase letters indicate difference between water conditions and different lowercase letters indicate differences between the treatments (% of K/Na). Vertical bars indicate standard errors between blocks (n = 4).  . 3). Regardless water condition, the exclusive Na-supply dramatically decreased the WUE I , WUE T and WUE L of E. saligna, E. urophylla and E. camaldulensis, while the partial K replacement by Na also decreased the WUE L of E. saligna (W−: 25%), relative to well K-supplied plants.

Growth parameters. A significant decrease was observed in the growth parameters of plants grown under
The drought effects on WUE varied among the measurement methods (leaf level or plant scale). Under W− condition, the WUE I was higher in E. saligna (65%) under exclusive Na-supply, whereas the WUE T decreased in E. saligna (45%) under the three levels of K/Na rates and in E. urophylla (15%) at exclusive Na-supply, relative to W+ condition. Moreover, the WUE L was also higher in E. saligna, E. urophylla and E. camaldulensis, under well K-supply, partial K replacement by Na and exclusive Na-supply ( Fig. 3 and Supplementary Table 1S).
Stable isotope natural abundances (δ 13 C, δ 15 N and C/N ratio). Regardless of the water condition, exclusive Na-supplied plants showed the lowest results to δ 13 C (except E. saligna under W −) and C/N ratio, as Figure 2. Photosynthetic rate-A (a-c), leaf transpiration rate-E (d-f) and stomatal conductance-g s (g-i), in leaves of E. saligna, E. urophylla and E. camaldulensis seedlings under three soil levels of K/Na in well-watered (W+) and water-stressed (W−) condition. Different uppercase letters indicate difference between water conditions and different lowercase letters indicate differences between % of K replacement by Na. Vertical bars indicate standard errors between blocks (n = 4).  15 N values (Fig. 4d-f).
Under W+ condition, the exclusive Na-supply decreased the C/N ratio of E. saligna (60%), E. urophylla (20%) and E. camaldulensis (15%), while the partial K replacement by Na decreased its values for E. saligna (27%), and increased for E. urophylla and E. camaldulensis by 20%, relative to well K-supplied plants. Moreover, under W− condition, the exclusive Na-supply decreased the C/N ratio of E. saligna (30%), E. urophylla (20%) and E. camaldulensis (20%), relative to well K-supplied plants at the same water condition. Drought stress decreased the C/N ratio of E. saligna (40%) at the treatment of well K-supplied plants and partial K replacement by Na, as . Intrinsic (a-c), instantaneous (d-f) and long-term (g-i) water use efficiencies in leaves of E. saligna, E. urophylla and E. camaldulensis seedlings under three soil levels of K/Na in well-watered (W+) and waterstressed (W−) condition. Different uppercase letters indicate difference between water conditions and different lowercase letters indicate differences between % of K replacement by Na. Vertical bars indicate standard errors between blocks (n = 4).  . Leaf carbon δ 13 C (a-c), and nitrogen (δ 15 N) isotope composition (d-f) and C/N ratio (g-i) of E. saligna, E. urophylla and E. camaldulensis seedlings under three soil levels of K/Na in well-watered (W+) and water-stressed (W−) condition. Different uppercase letters indicate difference between water conditions and different lowercase letters indicate differences between % of K replacement by Na. Vertical bars indicate standard errors between blocks (n = 4). Stomatal density (Std), stomatal pore area and total stomatal pore area (TSPA). Overall, the partial K replacement by Na increased the Std, while the exclusive Na-supplied plants reduced the Std for adaxial and abaxial sides, as also for TSPA under both water conditions, relative to either well K-supplied plants or the partial K replacement by Na (Table 1). Besides, no stomata were found in the adaxial side of E. saligna and E. urophylla (hypostomatic leaves, i.e. the Std was lower than 5 mm −2 ) 28 , while the leaves of E. camaldulensis were amphistomatic (occurred on both abaxial and adaxial sides, Supplementary Fig. 2S).
There were no significant differences between well K-supplied plants and the partial K replacement by Na in E. saligna and E. camaldulensis under both water conditions, and E. urophylla under W+ condition. However, the exclusive Na-supply decreased the Std of E. saligna, E. urophylla and E. camaldulensis compared to well www.nature.com/scientificreports/ K-supplied plants at the same water condition ( Table 1). The adaxial Std in E. camaldulensis also had no significant difference among the three soil levels of K/Na. Drought stress decreased the Std of E. saligna, E. urophylla and E. camaldulensis in partial K replacement by Na, and of E. saligna in well K-supplied plants. Otherwise, an increase in Std was observed in E. saligna and E. urophylla with exclusive Na-supply, when compared to W+ condition. The partial K replacement by Na increased the stomatal pore area of E. urophylla under W+ condition and E. camaldulensis in abaxial and adaxial side relative to well K-supplied plants (Table 1 and Supplementary Table 1S). However, the exclusive Na-supply decreased the stomatal pore area of E. urophylla and E. camaldulensis under W− condition in abaxial and adaxial side, compared to well K-supplied plants. Drought stress decreased the stomatal pore area of E. saligna grown under the three soil levels of K/Na and E. camaldulensis in abaxial side at exclusive Na-supply. Otherwise, the drought significantly increased the stomatal pore area in abaxial of E. urophylla and adaxial side of E. camaldulensis grown under well K-supply.
The exclusive Na-supply decreased the TSPA of E. saligna, E. urophylla and E. camaldulensis relative to well K-supplied plants, meanwhile the partial K replacement by Na reduced its values for E. urophylla under W− condition. Otherwise, an increase was observed for E. urophylla and E. camaldulensis at partial K replacement by Na under W+ condition (Table 1). Drought stress decreased the TSPA of E. saligna at well K-supply, partial K replacement by Na and exclusive Na-supply, and increased its values by 60% in E. urophylla at well K-supply and exclusive Na-supply and in E. camaldulensis at well K-supply, relative to W+ condition.
Chlorophyll and flavonoids content, Na and K accumulation in plants and leaf water potential. A significant decrease was observed in chlorophyll content (Chl) and leaf water potential (Ψw) of plants under drought stress, relative to W+ condition, as also of exclusive Na-supplied plants and those under partial K replacement by Na, when compared to well K-supplied plants. The Na accumulation in plants increased by augmenting Na rates, while the K accumulation decreased by reducing K supply, according to the K/Na combination rates (Table 2 and Supplementary Table 1S).
The partial K replacement by Na and exclusive Na-supply treatments decreased the Chl of E. saligna, E. urophylla, and E. camaldulensis relative to well K-supply ( Table 2). The drought stress decreased the Chl of E. saligna at well K-supply and exclusive Na-supply, of E. urophylla at the three soil levels of K/Na, compared to W+ condition, and E. camaldulensis at partial K replacement by Na and exclusive Na-supply. The flavonoids content (Flav) decreased at partial K replacement by Na in E. saligna, whereas increased in E. urophylla, and at exclusive Nasupply in E. saligna and E. urophylla under W− condition ( Table 2). Flavonoids content on E. saligna subjected to drought-induced stress increased by 22% at exclusive Na-supply, relative to plants under W+ condition.
The Na accumulation in plants augmented at partial K replacement by Na and at exclusive Na-supply relative to well K-supply. The drought stress increased the Na accumulation in plants of E. saligna at exclusive Na-supply, and of E. urophylla at partial K replacement by Na, compared to W+ condition ( Table 2).
The K accumulation in plants decreased at partial K replacement by Na and at exclusive Na-supply relative to well K-supplied plants. K accumulation in E. saligna under drought-induced stress increased at partial K replacement by Na and exclusive Na-supply and of E. camaldulensis at exclusive Na-supply, compared to plants under to W+ condition ( Table 2).
The partial K replacement by Na and exclusive Na-supply treatments decreased the Ψw of all eucalyptus species relative to well K-supply, at both water conditions. In addition, the drought-induced stress decreased the Table 1. Stomatal density, stomatal area and total stomatal pore area (adaxial plus abaxial surface of leaves) of E. saligna, E. urophylla and E. camaldulensis seedlings under three soil levels of K/Na in well-watered (W+) and water-stressed (W−) condition. Data represent mean values and standard errors between blocks (n = 4). Different uppercase letters indicate difference between water conditions and different lowercase letters indicate differences between % of K replacement by Na according to Tukey test (p < 0.05). -Indicates that no stomata were found. Multivariate analysis: linking carbon-nitrogen isotope compositions, physiological and nutritional responses. The PC1 and PC2 explained 41% and 23% of the total variance (Fig. 6a), respectively. The The hierarchically clustered heat map indicated the formation of two main groups among evaluated traits and treatments (Fig. 6b). Traits were grouped into group 1, composed by growth responses (H, D and TDM), Chl, Table 2. Chlorophyll content, flavonoids (Dualex units), Na and K accumulation and leaf water potential (Ψw) of E. saligna, E. urophylla and E. camaldulensis seedlings under three soil levels of K/Na in well-watered (W+) and water-stressed (W −) condition. Data represent mean values and standard errors between blocks (n = 4). Different uppercase letters indicate difference between water conditions and different lowercase letters indicate differences between % of K replacement by Na according to Tukey test (p < 0.05).  www.nature.com/scientificreports/ WUE, TSPA, A, δ 13 C, and K accumulation; and group 2, formed by Flav, H 2 O 2 , MDA, E, g s , Na and δ 15 N. The treatments were grouped into group 1, comprised by plants grown under well K-supply and partial K replacement by Na, irrespective of species and water condition, except E. saligna at partial K replacement by Na under W−; and group 2 was formed by plants under well K-supply, regardless species and water condition, and E. saligna at partial K replacement by Na under W− condition. These findings are supported by the confidence ellipse, which clearly allow the visualization of differential behavior in these two main groups of treatments, as observed by the high correlation between plants under well K-supply and partial K replacement by Na, and the higher variance in the exclusive Na-supply, irrespective of water condition and species (Fig. 6a). The δ 13 C and WUE L was negatively correlated to δ 15 N in a linear relationship (p < 0.01, Supplementary Fig. 1S) under the three species.

Discussion
The data obtained revealed that the partial K replacement by Na (50/50% of K/Na) up-regulated the leaf gas exchanges (A, E and g s ), boosting WUE L to maximum values in both water conditions tested, compared to the well K-supplied plants (100/0% of K/Na). The highest levels of WUE L observed in plants grown under partial K replacement by Na is probably due to the ability of Na accumulation to promote not only osmotic adjustments, such as better stomatal control and lower leaf osmotic potential 9 , but also mesophyll cell enlargement, enhancing water uptake and storage 11 , thus, alleviating drought impacts in Eucalyptus species. Moreover, drought decreased the Std of E. saligna-the drought sensitive species-grown under 100/0% and 50/50% of K/Na, which also had two-fold higher Std values than E. urophylla and E. camaldulensis, suggesting lower Std value as an effective adaptation tool for survival during dry seasons, as also observed by Frank et al. 29 . The stomatal density and its distribution at each leaf side have a clear impact on carbon physiology 30 , as observed by the positively correlation of TSPA with A and g s , being strongly driven by historic drought regime of the species 31 . In E. saligna and E. urophylla, stomata were located on the abaxial side, which is commonly observed in plants grown under mesophytic areas. However, the E. camaldulensis showed even Std in abaxial and adaxial side, which is common for plants grown under drought ( Supplementary Fig. 2S). These findings explain the greater values of A, E and g s observed in E. camaldulensis, the drought tolerant species, compared to E. saligna and E. urophylla.
The similar plant yields observed in the treatments with K/Na rates of 100/0% and 50/50% confirm the possibility of using Na to promote CO 2 assimilation and WUE 4 , regardless the water status. The notably positive effects of K replacement by Na in plant growth parameters of E. urophylla and E. camaldulensis than E. saligna are linked to the higher K-use efficiency 32 in drought tolerant species, stimulating greater K uptake of plants grown under stress condition 7 . This mechanism may occur due to the higher ability to sequester Na into vacuoles, while sensitive species allocate Na mainly into cell cytosol 33 .
The stable isotope natural abundances (δ 15 N and δ 13 C) have been utilized as powerful tools to evaluate genotypic variation in WUE and integrate physiological responses to stress due to their sensitivity to environmental factors, delivering useful information about leaf gas exchange 19 , plant yield 34 and drought-tolerant crops selection 35 . However, the WUE measurement is methodologically challenging since it can be determined via several methods, providing information on different spatial and temporal scales 36 . On a short time/leaf level scale, the WUE I can be measured through the ratio of A by g s or WUE T , through the ratio of A by E, producing accurate data of a specific time 36 . On the long time/whole plant scale, the WUE L can also be determined by the relationship between plant dry mass yield and water consumption throughout the experimental period 37 . As observed in this study, the large increases in leaf level WUE (WUE I and WUE T ) were not reflected in significant increases in CO 2 assimilation over the life of the plant, concealing the long-term adjustments in leaf level. This must carefully be considered when integrating these values to create more accurate and complex models for predicting WUE 31 . Thus, the WUE at the whole plant level becomes more reliable when studying gas exchanges and changes in the environmental conditions that occur during the cultivation period.
The magnitude of stomatal and/or non-stomatal limitations, that play an important role in photosynthesis, depends on the severity of the imposed stress. Under mild drought, where stomatal factors are dominant, the decline in A occurs as a consequence of the decrease in g s , and the expected trend is the increase in WUE and δ 13 C and a decrease in C i . However, under severe drought, non-stomatal limitations are preponderant, leading to constraints in mesophyll resistance, ribulose phosphate regeneration, rubisco activity and photochemical activity. Thus, the decrease in A does not occur due to a decrease in g s , and the expected consequence is the decrease in WUE and δ 13 C and an increase in C i 38,39 . As observed in our study, a highest A, δ 13 C and WUE was observed in plants grown under well K-supply and partial K replacement by Na, leading to a more conservative water use strategy 40 . In this context, we can assume that the stomata of plants under partial K replacement by Na were partially closed, optimizing stomatal movements to avoid water loss of Eucalyptus during photosynthesis. In contrast, the exclusive Na supply down-regulated stomata movements and decreased plant's TSPA, A, δ 13 C and WUE while increased the E, due to non-stomatal factors, which may have arisen from increased affinity of RuBisCO in capturing the CO 2 delivered by g s and mesophyll conductance and isotope discrimination 39 . Regarding water condition, the drought decreased considerably the g s , while the WUE and δ 13 C increased, indicating that stomatal factors were responsible for the decline in A of Eucalyptus species.
The δ 13 C at 100/0% and 50/50% of K/Na increased up to 3%, 1.7% and 2% due to drought in E. saligna (drought sensitive), E. urophylla (moderate drought tolerance) and E. camadulensis (drought tolerance), reflecting an increase in WUE L of 50%, 20% and 25%, respectively. It may be expected that even more significant increases in δ 13 C and WUE L correlation would be attained for longer drought periods 41 . Thus, the δ 13 C provided a reliable estimation of the K substitution by Na effect on water status of leaves, acting as a long-term indicator of plant metabolism and integrating CO 2 assimilation and WUE throughout the plant life. The exclusive Na supply led to the appearance of leaf chlorosis and necrosis, visual symptoms of K deficiency (Supplementary Fig. 3S) www.nature.com/scientificreports/ consequently to the lower plant yield, since K is an essential element that cannot be fully replaced 42 . Meanwhile, the K-deficiency also affected plants natural profusion of 15 N, which higher δ 15 N values indicate lower N uptake and accumulation in plant tissues, affecting plants carbon-nitrogen balance (leaf C/N ratio) 43 . As observed by the negative correlation of growth parameters and leaf gas exchange (Fig. 6a), δ 13 C and WUE L (Fig. 1S) with δ 15 N, our results corroborate the utilization of δ 15 N for crop screening on N metabolism and growth conditions 44 . The elevated ROS concentrations (H 2 O 2 ), harm the cellular membranes and other cellular components, resulting in oxidative stress and cell death 45 . Potassium is responsible for stomatal movement and its deficiency may lead to stomatal closure, which causes lower light energy absorption during CO 2 fixation and consequently intensify the electron transport chain, resulting in ROS accumulation and cell membrane oxidation 46 . As also observed for well K-supplied plants, the partial K replacement by Na maintained high values of soluble proteins, indicating higher detoxifying enzyme activities to reduce oxidative damage (lowest values of H 2 O 2 ) 18 . Furthermore, the partial Na supply also relieved MDA concentration 11 probably by entering into vacuole of stomatal cells and enabling stomatal aperture and CO 2 fixation, decreasing therefore H 2 O 2 synthesis 47 . Thus, partial Na supply appears to avoid oxidative stress (indicated by H 2 O 2 ) and, consequently, cell membrane disruption (indicated by MDA) in plants grown under limited K supply. The Chl levels reduced with Na supply under both water conditions in all species (Table 2). These findings provide an interesting insight: partial Na supply significantly improved A, g s , and WUE L , besides decreasing the Chl pigment content, possibly due an efficient regulation of the available amount of light and lower ROS formation by enhanced soluble proteins.
Otherwise, the exclusive Na supply may have exceeded the capacity of H 2 O 2 scavenging enzymes, as indicated by the low soluble protein values, depressing the photosystem activity and electron transport rate 48 and therefore resulting in elevated concentration of H 2 O 2 and MDA (final product of lipid peroxidation). Consequently, as typical symptom of oxidative stress, the K deficiency led to photo-oxidation of chlorophyll pigments of all Eucalyptus species, whose low values indicate hampered photosynthetic capacity 49 . Thus, Eucalyptus plants receiving less than 50% of their critical K supply could not be fully compensated by Na supply, suggesting that at least half the K requirements are necessary for functions that is poorly substituted by Na.
As mentioned, the K-deficient plants were characterized by high Na accumulation, MDA and H 2 O 2 values. Nevertheless, the novelty was the similarity of these values with high δ 15 N and low δ 13 C and WUE L (Fig. 6a). Thus, the data presented in this study confirms the potential of partial K replacement by Na in Eucalyptus plants with genotypes adapted to drought, such as E.urophylla e E.camaldulensis. Nonetheless, according to our hypothesis, we validate the use of natural isotopic abundance (δ 13 C and δ 15 N) as a useful sensitive indicator of WUE, stress responses and growth parameters, representing the long-term metabolism of plant life. Experimental design. The experiment was performed in randomized blocks with four replicates per treatment, in a 3 × 2 factorial design: three rates of K replacement by Na and two water conditions. Based on soil earlier trials 15 , the K replacement by Na was applied (as KCl and NaCl) in three soil levels: 100/0 (well K-supplied plants), 50/50 (partial K replacement by Na) and 0/100 (exclusive Na-supplied plants) % of K / % of Na, containing, respectively, 0/0.90, 0.44/0.44 and 0.90/0 mmol c dm −3 of Na/mmol c dm −3 of K, reaching the K level required for forest species development (< 1.20 mmol c dm −3 of K) 7 .

Methods
In addition, the seedlings were fertilized with basal nutrients (nitrogen, phosphorus, calcium, magnesium, sulfur, copper, zinc, iron, boron, manganese and molybdenum), as defined by Novais et al. 51 . Sixty days after the onset of the treatments, the seedlings were watered with two different regimes: well-watered (W+) and water stressed (W −) condition. The pots were daily weighed and watered with deionized water until their respective soil relative water content reached W+ : 80% and W−: 35% respectively, accordingly to the gravimetric method 52 to replace the evaporated and transpired water.
To analyze the influence of Na supply in extractable cations levels (K, Ca, Mg and Al), another supplementary soil incubation experiment was performed. Pots with same treatments and basal nutrients fertilization were watered daily to W+ soil relative water content and the soil was mixed monthly for 120 days and then analyzed. There were not any significant effects of Na supply (50/50 and 0/100% of K/Na) on exchangeable K, Ca, Mg and Al levels compared to the well K-supplied soil (100/0% of K/Na, data not shown).
Leaf gas exchange. Pior to harvesting, gas exchange measures on fully expanded leaves were performed (between 9 to 11 am), with a Li-Cor 6400XT (Li-Cor, Inc., Lincoln, NE, USA). Cuvette temperature was set to 25 °C and relative humidity to 65% maintaining vapor pressure deficit in the cuvette at around 1.1 kPa. The www.nature.com/scientificreports/ ration (E) were measured and averaged over 5-10 min after sample stabilization. Intrinsic water use efficiency (WUE I ) and instantaneous water use efficiency (WUE T ) of the leaf were determined by dividing the values of A by g s and A by E, respectively 36 . Long-term water use efficiency (WUE L ) was calculated by dividing the total dry mass value (belowground plus aboveground) by water use throughout the experiment (g dry mass/kg H 2 O) 37 . The leaf water potential (Ψw) was measured in the same leaves at noon (12 p.m) with a Scholander Pressure Chamber.
Chlorophyll and flavonoids content, stomatal density and total stomatal pore area. Relative leaf chlorophyll content and flavonoids content were expressed in Dualex units, and estimated using a portable Dualex-DX4 (FORCE-A, Orsay, France). Stomatal density (Std; stomates mm −2 ) was calculated using two fully expanded leaves per plant counting both abaxial and adaxial surfaces 9 . The number of stomata was counted using ImageJ program (https:// imagej. nih. gov/ ij/). The stomatal pore area (µm 2 ) was considered as an ellipse (π × 0.25 × stomatal length × stomatal width) 28 , which observations were made at 1000 × magnification by scanning electron microscopy (JEOL JSM-IT300 LV, Tokyo-Japan) at 20 kV. The length and width of stomates of four randomly selected fields of view of each leaf surface were measured per plant. The total stomatal pore area (TSPA; 10 −2 mm mm −2 ) was calculated as the product of stomatal pore area and Std.
Growth measurements. A graduated ruler and a digital pachymeter were used to measure the growth in height (cm/plant) and collar diameter (mm/plant), respectively. Plants were harvested on the 120th day of the experiment and their leaves, stems, and roots were separated. The dry mass production of each part was calculated by drying the samples in a forced air ventilation oven at 60 °C for 72 h and then weighing them.
Elementary and isotopic analyses. Subsequently, the plant material was ground in a Wiley type mill to quantify the K and Na concentration by nitric-perchloric digestion 53 . The accumulated K and Na (g plant −1 ) were calculated as the product of the concentration of each element in the plant part by the dry mass production of the respective tissue. To determine the leaf isotope composition (δ 13 C and δ 15 N) and the total leaf C and N, used to calculate the leaf C/N ratio, aliquots of different mass were packed into tin capsules and analyzed with an automatic nitrogen carbon analyzer (ANCA-GLS) interfaced to a continuous-flow isotope ratio mass spectrometer (Hydra 20-20, Sercon Ltd, Crewe, UK) 54 with an analytical precision of ± 0.3‰. The isotopes composition were calculated as the following equation 20 : δ = (R (sample) /R (standard) − 1); where R is the isotope ratio of 13 C/ 12 C or 15 N/ 14 N. The δ 13 C results were reported relative to the Vienna Pee Dee Belemnite (VPDB standard) and δ 15 N is relative to the standard atmospheric N 2 . Measurements were carried out at Isotope Ratio Mass Spectrometry Core Facility (CM-IRMS CENA) of the University of São Paulo.
Protein extraction. The samples (0.25 g of frozen leaves) were homogenized with a mortar and pestle to 2.5 mL of 100 mmol L −1 phosphate buffer (pH 7.8) containing 0.04 g of PVPP and 2% Triton X-100. The homogenate compound was centrifuged at 10,000 rpm for 30 min at 4 °C, and the supernatant was stored in 0.2 mL aliquots at − 80 °C. The soluble protein concentration was determined by Bradford 55 method using bovine serum albumin as a standard in a spectrophotometer at 595 nm (Genesys 10S UV-VIS,Thermo Fisher Scientific, Walthan, USA).

Determination of hydrogen peroxide (H 2 O 2 ) and lipid peroxidation.
The leaf concentration of hydrogen peroxide (H 2 O 2 , µmol g −1 fresh mass −1 ) was determined according to the method described by Alexieva et al. 56 . Lipid peroxidation was estimated by the leaf concentration of malondialdehyde (MDA, nmol g −1 fresh mass −1 ), according to the thiobarbituric acid (TBA) test 57 . The samples (0.2 g of frozen fully-expanded leaves) were ground in 2 mL of 0.1% (w/v) trichloroacetic acid (TCA) and 0.04 g of polyvinyl polypyrrolidone (PVPP) and centrifuged at 10,000 rpm for 10 min at 4 °C. For H 2 O 2 , 0.20 mL of supernatant was added to 0.2 mL of 100 mmol L −1 potassium phosphate buffer (pH 7.5) and 0.8 mL of 1 mmol L −1 potassium iodide. The tubes remained at room temperature in the dark for 1 h in continuous dark. Readings at 390 nm were measured using a spectrophotometer (Genesys 10S UV-VIS,Thermo Fisher Scientific, Walthan, USA). Three independent replicates of each plant were used. For MDA determination, the initial procedures were the same for H 2 O 2 measurements as described above. Following centrifugation, 0.25 mL of supernatant was added to 1 mL of 20% (w/v) TCA containing 0.5% TBA and heated in a water bath for 30 min at 95 °C. After cooling for 20 min on ice, the precipitate was removed by centrifugation at 10,000 rpm for 10 min and then, its absorbance read with a spectrophotometer (Genesys 10S UV-Vis, Thermo Fisher Scientific, Walthan, USA) at 532 and 600 nm. Three independent replicates of each plant were used. The leaf concentration of MDA was calculated as the following equation 58 : Statistical procedures. The K replacement by Na and water regimes were statistically analyzed by twoway ANOVA followed by post-hoc Tukey test (p < 0.05) using the software SAS version 9.1 (SAS Institute Inc, 2012). The original data were standardized to be analyzed via principal components analysis (PCA), integrating the measured variables in each treatment, genotype and water condition. For the PCA, we used the treatments with Na supply for the first two principal components (PC1 and PC2) and 95% confidence ellipses to visualize the multivariate trends of Na application under W+ and W− conditions using R package (R Development Core www.nature.com/scientificreports/ Team, 2018). The results are indicated as mean ± standard error of four independent biological replicates and graphically visualized using SigmaPlot 11.0 (Systat Software Inc., San Jose, CA, USA).
Research involving plants. The  www.nature.com/scientificreports/ Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
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:// creat iveco mmons. org/ licen ses/ by/4. 0/.