© The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com

Tree Physiology 00, 1–16doi:10.1093/treephys/tpu081

Sodium replacement of potassium in physiological processes of olive trees (var. Barnea) as affected by drought

Ran Erel1,2, Alon Ben-Gal1, Arnon Dag3, Amnon Schwartz2,† and Uri Yermiyahu1,4,†

1Institute of Soil, Water and Environmental Sciences, Gilat Research Center, Agricultural Research Organization, Mobile Post Negev 85-280, Israel; 2The Robert H. Smith Institute of Plant Sciences and Genetics in Agriculture, Faculty of Agriculture, Food and Environment, The Hebrew University of Jerusalem, Rehovot, Israel; 3Institute of Plant Sciences, Gilat Research Center, Agricultural Research Organization, Mobile Post Negev 85-280, Israel; 4Corresponding author (Uri4@agri.gov.il)

Received June 23, 2014; accepted August 24, 2014; handling Editor Roberto Tognetti

Potassium (K) is a macro-nutrient understood to play a role in the physiological performance of plants under drought. In some plant species, sodium (Na) can partially substitute K. Although a beneficial role of Na is well established, informa-tion regarding its nutritional role in trees is scant and its function under conditions of drought is not fully understood. The objective of the present study was to evaluate the role of K and its possible replacement by Na in olive's (Olea europaea L.) response to drought. Young and bearing olive trees were grown in soilless culture and exposed to gradual drought. In the presence of Na, trees were tolerant of extremely low K concentrations. Depletion of K and Na resulted in ∼50% reduction in CO2 assimilation rate when compared with sufficiently fertilized control plants. Sodium was able to replace K and recover the assimilation rate to nearly optimum level. The inhibitory effect of K deficiency on photosynthesis was more pronounced under high stomatal conductance. Potassium was not found to facilitate drought tolerance mechanisms in olives. Moreover, stomatal control machinery was not significantly impaired by K deficiency, regardless of water availability. Under drought, leaf water potential was affected by K and Na. High environmental K and Na increased leaf starch content and affected the soluble carbohydrate profile in a similar manner. These results identify olive as a species capable of partly replacing K by Na. The nutritional effect of K and Na was shown to be independent of plant water status. The beneficial effect of Na on photo-synthesis and carbohydrates under insufficient K indicates a positive role of Na in metabolism and photosynthetic reactions.

Keywords: photosynthetic capacity, soluble carbohydrates, stomatal regulation.

Introduction

Potassium (K) is an essential macro-nutrient found in rela-tively high concentrations in most terrestrial plants. Potassium plays a central role in the water economy of plants, especially via osmoregulation and maintenance of cell turgor. Proper K nutrition is acknowledged to raise plant tolerance to water stress (Cakmak 2005, Römheld and Kirkby 2010). Positive effects of K nutrition on drought resistance have been dem-onstrated in species including hibiscus (Hibiscus rosa-sinensis)

(Egilla et al. 2001), maize (Zea mays L.) (Premachandra et al. 1991) and sunflower (Helianthus annuus L.) (Lindhauer 1985). Under drought, adequate K levels are vital for main-taining efficient stomatal control, osmotic adjustment, preven-tion of oxidative damage and high water-use efficiency (WUE, Shabala and Pottosin 2014). In fact, the beneficial effects of high K nutritional levels on plants are often reported exclu-sively under conditions of water stress (Sen Gupta et al. 1989, Marschner 1995, Cakmak 2005). Potassium is the main osmoticum causing stomatal movement (Marschner 1995).

Research paper

†These authors contributed equally.

Tree Physiology Advance Access published October 3, 2014 at IN

RA

Institut National de la R

echerche Agronom

ique on October 10, 2014

http://treephys.oxfordjournals.org/D

ownloaded from

Tree Physiology Volume 00, 2014

Under water stress, K-deficient olive (Olea europaea L.) and sunflower plants failed to regulate stomatal closure (Arquero et al. 2006, Benlloch-Gonzalez et al. 2008). In addition to positively effecting tolerance to abiotic stresses, K was shown to play a crucial role in signaling-stress-causing factors, par-ticularly drought (Anschütz et al. 2014).

Contrary to K, which is acknowledged as an essential macro-nutrient, sodium (Na) is generally regarded in plant nutrition for its adverse toxic effects on plants and its antag-onistic relationship with K. Nonetheless, in some plant spe-cies, Na is apparently capable of at least partially replacing K. Beneficial effects of Na on plant growth and yield have long been acknowledged, specifically under conditions of low K availability. Marschner (1995) classified plants into groups according to their response to Na: species which can substitute some of the K by Na and in which Na can sometimes stimulate enhanced growth even when K is sufficient, and species which enjoy no benefit from Na at all. Since, under insufficient K, high Na-accumulating plant species can better regulate cell osmotic pressure, the presence of Na is expected to enhance drought resistance (Wakeel et al. 2011) by improved stomatal regula-tion, higher turgor and enhanced cell expansion. Unfortunately, there is little solid foundation backing these postulations. In sugar beet, probably the most studied Na utilizing plant spe-cies, Na addition was found to reduce stomatal conductance (gs) under stress and thus enhance the relative water content (Hampe and Marschner 1982). In accordance, the halophyte Aster tripolium was reported to close its stomata in response to high Na concentration and avoid Na accumulation in guard cells (Perera et al. 1997, Véry et al. 1998). In the well-studied glycoside Commelina communis, stomatal response to environ-mental changes of light, abscisic acid (ABA) and CO2 were smaller in the presence of Na compared with equivalent K con-centration (Jarvis and Mansfield 1980). In addition to enhanced drought tolerance, K nutrition has a role in the synthesis, trans-port and accumulation of various metabolites, mainly due to enzyme regulation (Armengaud et al. 2009). However, the question whether Na can partly replace K in carbohydrate (and amino acid) metabolism remains vague. Comprehensive reviews regarding the nutritional interaction of Na and K are available in Subbarao et al. (2003), Wakeel et al. (2011) and, most recently, in Kronzucker et al. (2013).

Although the possible beneficial effect of Na has long been known, reports quantifying the positive effect of Na on the phys-iology of perennial evergreen plants are limited to two species having fundamental physiological and genomic dissimilarity, cacao (Theobroma cacao) and eucalyptus (Eucalyptus grandis). Almeida et al. (2010) demonstrated the positive impact of Na application to eucalyptus trees in K-deficient soil which led to an increased above-ground biomass. Recently, augmented assimi-lation capacity and leaf area index were observed in response to Na application to K-deficient eucalyptus trees (Battie-Laclau

et al. 2013, 2014a). The enhancement in net photosynthesis rate (A) was partly attributed to higher leaf nitrogen and chlo-rophyll concentration. Under water stress, eucalyptus trees sup-plied with K or Na had a higher water demand and thus were more sensitive to drought (Battie-Laclau et al. 2014b). In cacao plants, Gattward et al. (2012) also demonstrated a positive effect of K substitution by Na on assimilation rate and WUE.

Olive trees are grown extensively in their native Mediter-ranean region. In spite of its widespread traditional cultivation, knowledge regarding the role of K nutrition in olive physiology is scarce. Potassium is considered to be particularly impor-tant in olive cultivation (Fernandez-Escobar 2010), partially because of a high correlation between K and yield (Hartmann et al. 1966). However, the assumption of high requirement of K for tree health and productive capability lacks solid scien-tific evidence. Olive is considered to be salt tolerant, depend-ing on the variety (Gucci and Tattini 1997). Salt tolerance in general (Marschner 1995) and the ability to compartmentalize Na in cell vacuoles are crucial if Na will function as a nutri-ent (Wakeel et al. 2011). Hence, plants such as halophytes (Flowers and Lauchli 1983), which exhibit tolerance to salinity, are likely able to use Na as an osmoregulant in vacuoles. The tolerance mechanisms of the olive make it a promising candi-date to utilize Na as a nutrient, however no data are available regarding potential disparate function of Na in olives in gen-eral or specifically when K is deficient. Since K is considered an enhancer of abiotic stress tolerance, the expression of K deficiency may not materialize except under water stress. For example, stomatal control of olives subject to K deficiency was found to differ from K sufficient plants only under water stress (Arquero et al. 2006, Benlloch-Gonzalez et al. 2008). Thus, although olive is an important widespread crop grown in harsh environments, there remains insufficient knowledge regarding the role of K in elevating drought tolerance and its possible replacement by Na. Furthermore, while a general supplemen-tary effect of Na has been established, the literature fails to provide explanations regarding mechanisms for beneficial roles of Na under conditions of water stress.

The effect of K on growth and productive traits was inves-tigated in a preliminary study (Erel et al. 2008, 2013b). Olive trees were grown for 4 years in perlite substrate under six K levels from 0.20 to 5.20 mM. Surprisingly, no considerable effect of K nutrition on growth, fruit production (Erel et al. 2008, 2013b) or oil composition (Erel et al. 2013a) was found. Two possible explanations for the lack of response to K were raised and are the driving hypotheses of the current study: (i) the absence of water stress and (ii) Na replacement of K. The objectives of this work were therefore to investigate the physiological response of olive trees var Barnea to K sta-tus and water stress and to examine the possible role of Na as a substitute for K. This was accomplished by both further examination of the chosen trees from the original experiment

2 Erel et al.

at INR

A Institut N

ational de la Recherche A

gronomique on O

ctober 10, 2014http://treephys.oxfordjournals.org/

Dow

nloaded from

Tree Physiology Online at http://www.treephys.oxfordjournals.org

and through an additional study on young trees. The two trials are complementary and together allow better understanding of the role of K and possible substitution of Na in olive's response to drought.

Materials and methods

The first hypothesis, that the role of K nutrition is contingent on conditions of water stress, was tested using trees from Erel et al.’s (2013b) preliminary experiment. Olive (O. europaea L.) trees that had grown outdoors for 4 years under various K levels were subjected to imposed drought stress and physi-ological parameters were monitored (EXP-I). Since EXP-I was originally designed to study solely the role of K+, Na+ was abundant in the irrigation solution (i.e., 1.8–4.0 mM). The source of variability in Na concentration was the use of K or Na as a counter-cation of NO3, therefore high K concentration was accompanied by low Na and vice versa. In addition, due to the high water consumption in EXP-I, the water used to prepare the irrigation solutions was from a local water supplier which contained ∼0.13 mM K. In order to eliminate the inherited limi-tations of EXP-I and investigate the second hypothesis that Na can at least partially replace K in physiological processes in olive, an additional experiment was run (EXP-II) in which both K and Na were studied independently in 1-year-old olive plants in a greenhouse, allowing accurate evaluation of the role of Na in the presence or absence of K. In EXP-II the source of water was a reverse osmosis plant providing practically zero back-ground K and Na concentration. Both trials were conducted at the Gilat Research Center, located in southern Israel (latitude 31°20′N, longitude 34°39′E).

Experimental design and treatments

EXP-I Olive plants var. Barnea were grown in 500 l contain-ers filled with perlite substrate. The experimental setup was random block designed with three repetitions per treatment.

For 4 years prior to the current study, trees were irrigated with a solution containing K concentration ranging from 0.20 to 5.20 mM K (Table 1). The main anion in the solution was NO3

− (62 mg l−1 for all treatments) which was accompanied with either K or Na according to the K level in the solution. With the exception of K, Na and SO4, the remaining miner-als were provided at identical concentrations to all treatments. During the experiment, developmental and productive traits were monitored. For detailed experiment description, please refer to Erel et al. (2013b). Trees were irrigated daily in excess from September 2006 to July 2010 using nutrient solutions as described above. Thereafter, eight trees, two from each treat-ment: T1, T2, T3 and T5 were placed on large scales connected to a data-logger (CR-1000, Campbell Scientific, Logen, UT, USA). Each container (tree) was equipped with a 40-l draining tank which was automatically emptied each morning at 05:00. Irrigation was scheduled to start at 24:00 to avoid interference with water uptake measurements. Diurnal water uptake was monitored for a month. Average measured evapotranspiration (ET) in July was 98, 92, 126 and 99 l per tree for treatments T1, T2, T3 and T5, respectively. Meteorological conditions were monitored at a station located 300 m from the experimental site. Diurnal solar radiation, air temperature and vapor pressure deficit (VPD) are plotted in Figure 1. During the experiment period (July–August 2010) the conditions were fairly stable, average class A pan evaporation was 8.8 ± 0.92 mm. Water stress was realized by gradually decreasing irrigation volume to 75, 50 and 25% of the initial average water uptake of each individual tree. Midday gs and stem water potential (ψs) were measured every 2–3 days in order to decide when to employ each of the drying steps, i.e., only when ψs and gs stabilized. Each stress step lasted for ∼10 days.

EXP-II One-year-old olive plants var. Barnea were planted in 60 l containers filled with perlite. Plants were grown in a controlled greenhouse equipped with wet pad cooling and a

Sodium replacement of potassium in the olive 3

Table 1. Average ± standard deviation of mineral concentration in the irrigation solution during the 4 years of study in EXP-I and the corresponding dry biomass, total K and Na uptake and leaf K and Na concentrations at the end of the experiment. Shoot was defined as the total above-ground biomass (leaves, branches, limbs and stem) and root is the below-ground biomass (root neck, big roots and small roots). The type (linear, quadric, logarithmic [Log] or exponential [Exp]) and significance of best-fit trend lines of plant biomass and mineral partitioning as a function of K level in irrigation water are noted. n = 3. *P < 0.05, **<0.001, ***<0.0001. N.S., not significant.

Treatment Mineral concentration (mM) Cumulative dry weight (kg tree−1)

Cumulative K (g tree−1)

Cumulative Na (g tree−1)

Leaf concentration (%DW)

K Na Shoot Root Fruit Shoot Root Shoot Root K Na

T1 0.20 ± 0.10 4.02 ± 0.33 81.2 42.9 21.2 147 47 181 263 0.28 0.37T2 0.47 ± 0.09 3.85 ± 0.17 75.9 29.1 21.2 223 82 150 182 0.58 0.19T3 0.73 ± 0.09 3.72 ± 0.13 91.4 37.6 21.0 442 107 117 287 0.93 0.09T4 2.06 ± 0.15 3.25 ± 0.11 80.2 37.4 20.6 571 245 93 265 1.42 0.06T5 2.80 ± 0.19 2.71 ± 0.14 89.8 37.7 18.8 858 550 89 191 1.71 0.05T6 5.20 ± 0.12 1.78 ± 0.22 73.7 41.1 18.8 857 548 48 157 2.20 0.06

N.S. N.S. N.S. **Log *Log *Exp N.S. ***Log **Exp

at INR

A Institut N

ational de la Recherche A

gronomique on O

ctober 10, 2014http://treephys.oxfordjournals.org/

Dow

nloaded from

Tree Physiology Volume 00, 2014

removable thermal screen cover. Diurnal solar radiation, air temperature and VPD are plotted in Figure 1. After 3 months of acclimation, trees were pruned to 65-cm single shoots with 20 leaves per plant. Differential nutrient application treatments were initiated immediately after pruning. Treatment K and Na concentrations and nomenclature are given in Table 2. Three treatments contained no added K and increasing levels of Na with target concentrations of 0 mg l−1 Na (K0-Na0), 10 mg l−1 Na (K0-Na10) and 58 mg l−1 Na (K0-Na58). Three treatments contained no added Na and increasing levels of K with target concentrations of 0 mg l−1 K (K0-Na0), 15 mg l−1 K (K15-Na0) and 100 mg l−1 K (K100-Na0). A control treatment contained

high levels of K and Na with target concentrations of 100 mg l−1 K and 58 mg l−1 Na (K100-Na58). Divergence from target con-centrations was mostly due to the impurity of the salts and traces of minerals in the desalinated irrigation water. The experi-ment used a ‘Latin square’ design (balanced rows and columns) with six replicates. During the growth period trees were pruned three times to comparable size and water uptake. Target con-centrations were achieved by proportionally dissolving commer-cial grade salts of Ca(NO3)2, NH4H2PO4, KCl, MgSO4, Mg(Cl)2 and NaCl. Differences between treatments were exclusively as K, Na and Cl concentrations. Micronutrients (Fe, Mn, Zn, Cu, Mo and B) were applied at a rate of 0.1 ml l−1 of commercial solution

4 Erel et al.

Figure 1. Solar radiation, air temperature and vapor pressure deficit (VPD) on dates when physiological parameters were measured. Open air: Measurements taken from a nearby meteorological station before (solid lines, 27 July 2010) and during (dashed lines, 28 August 2010) imposed drought. Greenhouse: Measurements taken from the center of the greenhouse before (solid lines, 22 May 2013) and during (dashed lines, 17 June 2013) imposed drought.

Table 2. Average ± standard deviation of mineral concentrations in irrigation solutions during EXP-II and the corresponding leaf K and Na con-centrations, dry biomass and stem circumference. Shoot was defined as the total above-ground biomass (leaves, branches, limbs and stem) and root is the below-ground biomass (root neck, big roots and small roots). Treatment effect was tested by one-way ANOVA (Tukey–Kramer multiple comparisons test), different letters indicate statistical significance (n = 6).

Treatment Mineral in irrigation solutions (mM)

Leaf concentration (%DW)

Cumulative DW (g tree−1) Stem circumference (mm)

K Na K Na Shoot Root Total

K0-Na0 0.09 ± 0.03 0.13 ± 0.04 0.19c 0.03c 2150c 320b 2470d 96cK0-Na10 0.07 ± 0.04 0.57 ± 0.07 0.21c 0.10b 2304c 296b 2599cd 98bcK0-Na58 0.08 ± 0.03 2.42 ± 0.15 0.20c 0.19a 2533bc 512a 3045bc 105bK15-Na0 0.39 ± 0.04 0.15 ± 0.04 0.47b 0.02c 2789ab 332b 3130b 116aK100-Na0 2.41 ± 0.17 0.27 ± 0.04 1.05a 0.01c 2970a 362ab 3332ab 123aK100-Na58 2.48 ± 0.20 2.84 ± 0.31 1.01a 0.01c 3159a 496a 3655a 121a

at INR

A Institut N

ational de la Recherche A

gronomique on O

ctober 10, 2014http://treephys.oxfordjournals.org/

Dow

nloaded from

Tree Physiology Online at http://www.treephys.oxfordjournals.org

(‘Koratin + B’, Fertilizers and Chemicals Ltd, Haifa, Israel). For each treatment, salts were dissolved in desalinated water with electrical conductivity <50 µS m−1. Average measured mineral concentrations of specific elements were: Ca: 2.0 ± 0.1 mM, Mg: 1.2 ± 0.2 mM, NH4: 0.27 ± 0.04 mM, NO3: 4.6 ± 0.1 mM and P: 0.30 ± 0.02 mM. The pH was in the range of 5.7–5.8.

Trees were irrigated daily in excess using the described solutions for 15 months. In a preliminary trial, irrigation was discontinued in 10 similar olive pots. The pots were weighed daily until visual stress symptoms (wilting) appeared. The amount of water lost during the drought period averaged 12.0 l per pot. This amount of water, between pot capacity and wilting, was determined to be the ‘plant available water’ for subsequent trials. In May 2013, drought was imposed by gradually decreasing the plant available water by 4 l pot−1 in three steps (8, 4 and 0 l pot−1). In order to achieve accurate and identical water content for each plant, pots were weighted daily early in the morning and the volume of available water was determined by comparing with the pot’s weight at capac-ity accordingly: AW = 12 − (Wpc − Ws) where AW is the avail-able water, Wpc is the pot weight after excesses irrigation and drainage (weighed at dawn) and Ws is the pot weight early in the morning during stress. Immediately following daily weigh-ing, water was manually supplied to reach the target weight for each pot if necessary. Each stress step was carried out over a week. During the periods of imposed stress, morning stomatal conductance and leaf water potential were deter-mined twice a week in order to monitor plant water status. The average daily ET before initiation of drought (26 May 2013) was 4.8 ± 0.42 l per pot.

Mineral concentration, partition and biomass

Leaf analysis was performed according to the protocol formerly used for olive as described in Erel et al. (2013b). Upon con-cluding the trials, plants were removed and divided into leaves, branches, limbs, stems, root necks, thick roots (>10 mm diam-eter) and fine roots (<10 mm diameter). Each plant fraction was sampled, weighed, dried and ground. Mineral concen-trations were determined as described for leaves. For sim-plicity, in Tables 1 and 2 the dry weight (DW) and mineral uptake rates and concentrations of combined total shoot and root (total above- and below-ground biomass) are presented. In EXP-I, fruits were harvested and weighed in October each year. Fruit DW content was determined by oven drying of fruit paste. Stem circumference was measured ∼10 cm above the substrate surface.

Gas exchange, chlorophyll fluorescence and water potential

Net CO2 assimilation rate (A) and stomatal conductance of H2O vapor (gs) were measured simultaneously by a portable gas exchange measuring system (LI-6400–40, LI-COR, Inc., Lincoln,

NE, USA). In general, chamber conditions were set to an envi-ronmental CO2 concentration of 400 µmol CO2 mol−1, a photo-synthetic photon flux density (PPFD) of 1000 µmol mol m−2 s−1, a humidity of 10 mmol H2O mol−1 and a block temperature of 25 °C. Measurement periods predawn and after sunset were taken in accordance with the environment natural light with chamber light turned off (PPFD = 0). Electron transport rate (ETR) was measured using a portable chlorophyll fluorometer, Mini-PAM (Walz, Effeltrich, Germany) in EXP-I and OS5p (Opti-Sciences, Hudson, NH, USA) in EXP-II. For both gas exchange and chlorophyll a fluorescence (F), measurements were taken from fully illuminated leaves (except early in the morning and after sunset). During the day, ETR was measured only if PPFD ≥500. Firstly, F was logged and immediately after saturation a pulse was illuminated in order to log maximal fluorescence (Fm′). Electron transport rate was calculated accordingly: (Fm′ − F)/ Fm′ × 0.5 × PPFD × 0.875. Stem and leaf water potential (ψs and ψl) were measured in a pressure chamber (Arimad-3000, MRC, Holon, Israel). The chamber flow rate was set to 0.03 MPa s−1.

EXP-I Diurnal measurements were taken twice during the experiment: a day before imposing drought stress on 27 July 2010 and during the stress on 29 August 2010. Measurements were taken from sunrise until after sunset at ∼2-h intervals. Gas exchange measurements were made on leaves on the tree’s periphery, five leaves per tree and two trees per treatment. ψs was measured on two shoots per tree as described by Naor et al. (2006) on shaded shoot endings containing six to seven leaves which were covered by aluminum bags at least 2 h in advance.

EXP-II Two diurnal measurements were performed before and during drought stress (22 May and 17 June 2013, respec-tively) on five replicate trees of treatments K0-Na0, K0-Na10, K0-Na58 and K15-Na0, and the control. K100-Na0 was essen-tially identical to the control in regards to leaf mineral concen-trations and was therefore not measured. First measurements were taken before sunrise and the last set was carried out after sunset. Gas exchange was measured on three to four leaves per tree and ψl on one uncovered shoot per tree, each at ∼2-h intervals.

Starch and soluble carbohydrate determination (EXP-II)

Starch and soluble carbohydrates (SCH) were determined a day before drought was initiated. Treatment K100-Na0 was excluded. Leaves were collected and immediately frozen in liq-uid nitrogen, lyophilized, milled and stored at −20 °C. Soluble carbohydrates were extracted as follows: 20–25 mg leaves were extracted three times with 80% ethanol at 80 °C. The extracts were evaporated to dryness and re-dissolved in 4 ml of double distilled water from which aliquots of 500 µl were lyophilized. The dried material was re-dissolved and derivatized

Sodium replacement of potassium in the olive 5

at INR

A Institut N

ational de la Recherche A

gronomique on O

ctober 10, 2014http://treephys.oxfordjournals.org/

Dow

nloaded from

Tree Physiology Volume 00, 2014

for 90 min at 37 °C (in 40 µl of 20 mg ml−1 methoxyamine hydrochloride in pyridine), followed by a 30-min treatment with 70 µl of N-methyl-N-(trimethylsilyl) trifluoroacetamide at 37 °C and centrifugation. After derivatization, the SCH was analyzed via Agilent 6850 GC/5795C MS (Agilent Technology, Palo Alto, CA, USA). One microliter was injected in splitless mode at 230 °C. Helium carrier gas was used at a flow rate of 1 ml min−1. Chromatography was performed using a HP-5MS capillary column (30 m × 0.250 mm × 0.25 µm). Mass spec-tra were collected at 2.4 scans s−1 with an m/z 50–550 scan. Mass spectrometry (MS) temperature was set to 230 °C. Both ion chromatograms and mass spectra were evaluated using MSD ChemStation E.02.00.493 software. Carbohydrates were identified by comparison of retention time and mass spectra with those of authentic standards (sigma) and each SCH was quantified using a calibration curve.

Starch content was determined by digesting ethanol insoluble residue with amyloglucosidase as described in Vishnevetsky et al. (2000). Glucose in solution was determined by colorimet-ric reaction with dinitrosalicylic acid according to Miller (1959).

Data analysis

EXP-I Relationships between plant biomass and mineral par-titioning to the K level in irrigation water were tested using Sigmaplot 10.0 software (Systat Software, San Jose, CA, USA). Linear, quadric, logarithmic or exponential trend lines were fit-ted to data. The type of best fit regression and significance are presented in Table 1.

EXP-II Effect of treatments on plant biomass, stem circum-ference, Fv/Fm, SCH and starch were analyzed by one-way analysis of variance (ANOVA, Tukey–Kramer multiple compari-sons test) using JMP 7 statistical software (SAS Institute, Cary, NC, USA). The diurnal measurements were integrated over the day for each individual tree, assuming continuity between the measured points. Instead of gs, transpiration rate was used for evaluation. The sum of daily A, transpiration and e− transport for each tree was analyzed using ANOVA as described above. The regression of gs and A was plotted using SigmaPlot (Systat Software, Inc.) along with 95% confidence levels. The relation-ship between gs and ψl was plotted using SigmaPlot using data measured during the stress development stages.

Results

EXP-I

Cumulative biomass, yield and minerals Olive trees were grown in an inert substrate for 4 years under a wide range of K in irrigation solutions. Nonetheless, accumulated growth of all treatments was similar: i.e., 124 and 115 kg DW per tree for T1 and T6, respectively. Likewise, no effect of K on biomass partitioning was noted (Table 1). Cumulative dry fruit yield was

also independent of treatments. Potassium in the shoot and root increased in a logarithmic fashion reaching maximum at the two highest K levels: T5 and T6. The K content in shoots of the high K treatment was more than fivefold greater than that of T1. In the roots, the high K treatments accumulated more than 11 times greater K. Sodium accumulated in the shoots at relatively high levels in the lowest K treatment and decreased exponentially as K in the irrigation solution increased. Sodium accumulation in roots was variable and was not significantly affected by K level. As described for shoot K, leaf K concentra-tion increased in a logarithmic fashion with K in the irrigation solution rising from 0.28% up to 2.20%. Treatment T1 had K concentration below the commonly accepted K deficiency level of 0.4% (Fernandez-Escobar 2010). Potassium in T2 leaves fell between 0.4% and the correspondingly accepted sufficiency threshold of 0.8%. Sodium leaf concentration was inversely correlated to leaf K, decreasing exponentially from 0.37% in T1 down to ∼0.06% for T4–T6. Thus, the T1 treatment had significantly higher Na content compared with K in leaves. Leaf Na concentration was almost double the accepted threshold for toxicity of 0.2% (Fernandez-Escobar 2010). Nonetheless, this did not reduce growth, even after 4 years.

Diurnal physiological parameters of stressed and non-stressed trees Diurnal measurements of non-stressed plants were collected on 27 July 2012 and drought stressed measurements on 29 August following gradual drought. Both days were clear and hot, typical to the arid climate of the research center (Figure 1). Both diurnal sets, stressed and non-stressed, are presented in Figure 2. For non-stressed trees, regardless of the K level, A rose from nearly null early in the morning to a maximum of ∼15 µmol m−2 s−1 at 10:30 (Figure 2). During the day, A was quite stable until 17:00 to sunset when it decreased. Stressed trees had consider-ably lower A during the day. Maximum A was measured at 09:10 followed by a gradual decrease towards the afternoon and a minor peak around 18:30. Under both stressed and non-stressed conditions, no significant difference in A due to K was found. In non-stressed trees, gs increased from early in the morning until 10:30 and remained fairly stable (0.26 mol H2O m−2 s−1) until 17:10, followed by a sharp decline towards sunset (Figure 2). Stressed trees exhib-ited very low gs which peaked at around 9:00 to a value of ∼0.1 mol H2O m−2 s−1 and later dropped to extremely low values. Again, K level had no effect on gs in both cases. Diurnal ETR of well-irrigated trees of T1 and T2 started from almost zero in the morning to ∼120 µmol e− m2 s−1 com-pared with ∼140 for T3 and T5. This difference was not significant. Electron transport rate remained stable most of the day (Figure 2). Electron transport rate of the stressed trees peaked at 10:30 (∼76 µmol e− m2 s−1) and gradually decreased until the end of the day. The ψs was substantially

6 Erel et al.

at INR

A Institut N

ational de la Recherche A

gronomique on O

ctober 10, 2014http://treephys.oxfordjournals.org/

Dow

nloaded from

Tree Physiology Online at http://www.treephys.oxfordjournals.org

affected by irrigation level (Figure 2). In the non-stressed trees, ψs was −0.7 MPa early in the morning, decreased to −1.8 by midday and recovered to −1.0 around sunset. Potassium level did not affect ψs. In the stressed trees, ψs decreased continuously from −1.7 MPa at dawn to −4.5 in the afternoon and evening. T2 had the lowest ψs for most of the day, however, differences were insignificant.

EXP-II

Biomass and mineral content In addition to the difference in tree size and age, EXP-II differed from Exp-I, in part, due to the manipulations made on K levels with or without Na. Leaf K increased as a function of K in irrigation water from 0.19 to 0.20%, when no K was added, to slightly >1% in response to 2.5 mM K in solution (Table 2). When no K was added, Na concentration in leaves increased as a function of Na in solution from 0.03 to 0.19%. In the presence of K, Na was almost completely excluded from leaves. Treatments K0-Na58 and K100-Na58 (control) received similar Na levels

in their irrigation solution but had entirely different effects on leaf Na concentration, 0.19% if no K was added compared with 0.01% at the highest K level. Shoot biomass production (including pruned branches) tended to increase with Na level with no statistical difference (Table 2). The three no-K treat-ments produced significantly lower shoot biomass compared with K100-Na0 and the control. Root biomass was highest in Na58-K0 and the control, significantly higher than Na0-K0, Na15-K0 and K15-Na0. When no K was added, total produced biomass significantly increased with Na, but did not reach the levels of the trees which received K. Maximum biomass was found in the control. Stem circumference was comparable to plant biomass, showing a significant increase in response to Na in K-depleted plants and a maximum value for the highest K levels (Table 2).

Starch and SCHs in leaves Starch and SCHs were deter-mined in May before imposing drought. Of the eight inden-tified SCHs, glucose was the most dominant followed by

Sodium replacement of potassium in the olive 7

Figure 2. Average (symbols) and standard deviation (error bars) of daily measurements of assimilation rate (A), stomatal conductance (gs), electron transport rate (ETR) and stem water potential (ψ s) of fully irrigated (continuous lines, empty symbols) or drought stressed trees (dashed lines, full symbols) of olive trees grown for 4 years under different K levels: 0.2 mM (circles), 0.5 mM (diamonds), 0.7 mM (triangles) and 2.8 mM (squares) in EXP-I.

at INR

A Institut N

ational de la Recherche A

gronomique on O

ctober 10, 2014http://treephys.oxfordjournals.org/

Dow

nloaded from

Tree Physiology Volume 00, 2014

mannitol and sucrose; the three composed ∼90% of the total SCH. Potassium and Na nutrition significantly affected six of the eight SCHs. The main trend of elevated K was found to be an increase in mannitol at the expense of sucrose. Galactose, myo-inositol and galactinol significantly decreased with increased K while raffinose was increased. The total SCH level, however, was not modified by treatment and fluc-tuated ∼120 mg g−1. Leaf starch concentration was signifi-cantly higher in high Na and high K treatments: 86 mg g−1, compared with a mere 66 mg g−1 for Na0-K0 (Table 3). The intermediate K and Na treatments had intermediate starch concentrations.

Diurnal physiological parameters of stressed and non-stressed trees Diurnal measurements of non-stressed plants for EXP-II were collected on 22 May 2013 and stressed measurements on 17 June. Both days were relatively warm (Figure 1). The diurnal measurements are presented in Figure 3. In order to improve visualization, results from treat-ment K0-Na10 are presented in the summary table (Table 4) but not in Figure 3. Under conditions of high water availability, A was considerably affected by Na and K levels (Figure 3). Throughout the day, treatment K0-Na0 had the lowest A, roughly half of the control. The addition of Na to zero-K trees (K0-Na58) recovered most of the A, although it remained lower than the control. The A of treatment K15-Na0 was quite similar to the control. Stomatal conductance and ETR diurnal curves generally followed those of A (Figure 3). Predawn gs was notably high for all treatments even though measurements were taken in the absence of significant PPFD radiation and with the chamber internal light turned off. The zero-K treat-ments, K0-Na0 and K0-Na58, exhibited the lowest gs between 09:00 and 14:00 but not at dawn or later in the afternoon. After sunset, maximal gs was measured in treatment Na0-K0. Potassium and Na nutrition had a pronounced effect on ETR. During daylight hours (09:00–17:00) ETR was mostly stable, averaging 141, 143, 116 and 79 µmol e− m−2 s−1 for control, K15-Na0, K0-Na58 and K0-Na0, respectively. All treatments exhibited identical and typical ψl diurnal curves, starting at −0.7 MPa at dawn, decreasing to a minimum of −1.5 MPa at

around 14:00 and partly recovering in the evening, post dusk, hours.

Subsequent to prolonged gradual drought stress, A, gs, ETR and ψl each declined considerably (Figure 3). A and gs in stressed trees were highest at 09:00, gradually depleted as the day advanced and demonstrated occasional minor peaks in the afternoon between 16:00 and 18:00 (Figure 3). Except for ψl, the nature of the response to nutrition (but not the actual values) was similar in irrigated and drought condi-tions. Assimilation rate, gs and ETR were higher in K15-Na0 and the control in the morning and through most of the day. A, gs and ETR were generally lowest in K0-Na0 while K0-Na58 had either lower or similar values compared with the control. ψl was notably affected by K and Na levels under drought stress (Figure 3). Throughout the day, the lowest ψl was measured for the control, followed by K15-Na0, K0-Na58 and K0-Na0. Regardless of the nutritional level, ψl was highest in the morn-ing, lowest around 14:00 and subsequently partially recovered towards the night.

Cumulative daily A, transpiration, ETR and WUE In order to interpret the diurnal A, transpiration and ETR, each curve was integrated to provide accumulated daily values of assimi-lation, transpiration and e− transport of each individual tree. Additionally, cumulative A was divided by transpiration to yield intrinsic WUE (Table 4). Under irrigated conditions, total daily assimilated CO2 was significantly lower in K0-Na0 and K0-Na10: 0.40–0.47 mol m−2 day−1, compared with 0.75–0.78 in the control and K15-Na0. Total daily assimilated CO2 in treatment K0-Na58 lay roughly in between, signifi-cantly higher than K0-Na0 and K0-Na10 but lower than the K treatments. Accumulated CO2 assimilation decreased more than three times under drought. However, the relative effect of treatments remained more or less the same. The diur-nal transpired H2O of non-stressed trees was the lowest in K0-Na0, significantly less than K15-Na0 and the control but not than K0-Na58 and K0-Na10. Under drought, transpiration dropped by a magnitude of ∼6 and K0-Na0 was the lowest, although not significantly different. The effect of treatments on diurnal e− transport strongly resembled that reported for

8 Erel et al.

Table 3. Average SCHs and starch concentrations in leaves 11 months after treatment initiation in EXP-II. Treatment effect was tested by one-way ANOVA (Tukey–Kramer multiple comparisons test), different letters indicate statistical significance (n = 5).

SCHs and starch concentration (mg g−1)

Glucose Mannitol Sucrose Fructose Galactose myo-Inositol Raffinose Galactinol Total SCH Starch

K0-Na0 54.0a 17.4b 23.6a 3.7a 2.9a 2.8a 2.1b 12.1a 119a 66bK0-Na10 57.1a 18.7b 22.0ab 3.8a 3.2a 2.2ab 2.3ab 10.2ab 119a 69abK0-Na58 55.3a 22.6ab 23.4a 3.2a 3.0a 2.4ab 2.8a 10.9a 124a 86aK15-Na0 60.7a 26.2a 15.9ab 3.4a 2.5ab 1.8ab 2.5ab 7.6ab 121a 78abK100-Na58 51.3a 26.1a 15.2b 3.0a 1.3b 1.6b 2.5ab 5.3b 106a 86aTreatment effect: 0.8492 0.0005 0.0097 0.3048 0.0009 0.0099 0.0165 0.0055 0.6833 0.0011

at INR

A Institut N

ational de la Recherche A

gronomique on O

ctober 10, 2014http://treephys.oxfordjournals.org/

Dow

nloaded from

Tree Physiology Online at http://www.treephys.oxfordjournals.org

assimilation, i.e., lowest in K0-Na0, higher in K0-Na58 and highest in K15-Na0 and control. Under drought, daily accu-mulated e− transport was much lower but the response to the treatments was similar. Water-use efficiency was much higher under drought. For example: 5.0 compared with 7.8 mmol CO2 mol−1 H2O in the control. Under irrigation, the

low K and Na treatments, K0-Na0 and K0-Na10, had signifi-cantly lower WUE. Under drought, all treatments had roughly similar WUE, an average of 8.0 mmol mol−1.

Maximum quantum yield of PSII (Fv/Fm) Under fully irrigated conditions, predawn Fv/Fm was high (0.82–0.83)

Sodium replacement of potassium in the olive 9

Figure 3. Average (symbols) and standard deviation (error bars) of daily measurements of assimilation rate (A), stomatal conductance (gs), electron transport rate (ETR) and leaf water potential (ψl) of fully irrigated (left column) or drought stressed (right column) trees with different K and Na nutrition. No added K and Na (circles); No K and 58 mg l−1 Na (triangles); 15 mg l−1 K and no Na (diamonds); 100 mg l−1 K 58 mg l−1 Na (squares).

at INR

A Institut N

ational de la Recherche A

gronomique on O

ctober 10, 2014http://treephys.oxfordjournals.org/

Dow

nloaded from

Tree Physiology Volume 00, 2014

for all treatments except K0-Na0 and K0-Na10 (0.77–0.78) (Figure 4). Under drought, predawn Fv/Fm was impaired in all treatments (Fv/Fm < 0.80), indicating photodamage. Yet, Fv/Fm was lowest for K0-Na0 and was significantly higher in the high Na treatment.

Visual symptoms For the first 3 years in EXP-I, trees could not be visually distinguished according to their K level. In the fourth year, T1 exhibited willow-like branch growth patterns. Subsequently, by the end of the experiment, visual examination revealed occasional leaf tip necrosis. The appearance of ‘tip burn’ was likely due to Na toxicity rather than K deficiency as leaf Na (Table 1) reached approximately double the accepted threshold values for olives (Fernandez-Escobar 2010). In EXP-II, visual evidence of K deficiency, chlorotic leaves in the K0-Na0 treatment, was first noticed ∼4 months after treatment initiation. Thereafter, symptoms were accelerated, trees showed

retarded growth, ‘willow-like’ branching and, eventually, typical necrosis spreading from leaf tips to edges (Figure 5). All symp-toms were remarkably reduced by the addition of Na to zero-K plants.

Discussion

General response of olive to limited K and its substitution by Na

Potassium is understood to play many vital roles in plant physi-ology, including osmotic regulation, enzyme activation, elec-trochemical balance, phloem and xylem transport and stress signaling (Wakeel et al. 2011, Shabala and Pottosin 2014). Typical K deficiency symptoms are characterized by retarded cell growth, accumulation of simple carbohydrates, lower chlorophyll content, photosynthesis inhibition, and eventually, reduced growth and productivity (Römheld and Kirkby 2010). In the current study, the physiological response of olive to K levels and drought was investigated in the presence of ample Na on bearing trees (EXP-I) or under manipulated levels of Na on young trees (EXP-II). The data obtained indicates rel-atively low demand of the olive tree for K, especially in the presence of Na. In most plants, K is needed in concentrations comparable with those of Na, typically in the range of 2–5% of plant DW. In contrast, in species that are capable of replacing K by Na, adequate leaf K requirement is much lower, 0.5–1.0% (Marschner 1995). In the olive, in the presence of Na, leaf K dropped to 0.3% with no appearance of visual symptoms and with no negative effect on growth and yield as demonstrated in EXP-I. Under the conditions of EXP-I, leaf Na was more than double compared with leaf K. Olive is considered to be salt tolerant, depending on the variety (Gucci and Tattini 1997) and its tolerance to salinity is accredited mainly to the ability to exclude and retain salt ions in roots (Chartzoulakis 2005) and therefore might be considered as an ‘excluder’. However, when K is limited Na is accumulated in the leaf and augments

10 Erel et al.

Figure 4. Predawn maximum PSII quantum yield (Fv/Fm) before (black bars) and during (gray bars) imposed drought stress. Treatment effect was tested by one-way ANOVA (Tukey–Kramer multiple comparisons test), different letters indicate statistical significance (n = 6).

Table 4. Integrated daily assimilation, transpiration, ETR and intrinsic WUE under irrigation and drought regime. The diurnal measurements were integrated by summing the additive values over the day for each individual tree (n = 5) assuming continuity between the measured points. Treatment effect was tested by one-way ANOVA (Tukey–Kramer multiple comparisons test), different letters indicate statistical significance.

Treatment Daily assimilation (mol CO2 m−2)

Daily transpiration (mol H2O m−2)

Daily ETR (mol e− m−2)

WUE (mmol CO2 mol−1 H2O)

Irrigation K0-Na0 0.40c 123b 3.2c 3.3b K0-Na15 0.47c 137ab 3.7cb 3.4b K0-Na58 0.62b 134ab 4.7b 4.6a K15-Na0 0.78a 159a 5.8a 4.9a K100-Na58 0.75a 152a 5.8a 5.0aDrought K0-Na0 0.13B 16.8A 0.59B 7.6A K0-Na15 0.12B 17.4A 0.63AB 6.6A K0-Na58 0.17AB 19.8A 0.79AB 8.8A K15-Na0 0.26A 27.7A 1.01A 9.3A K100-Na58 0.22AB 28.3A 0.91AB 7.8A

at INR

A Institut N

ational de la Recherche A

gronomique on O

ctober 10, 2014http://treephys.oxfordjournals.org/

Dow

nloaded from

Tree Physiology Online at http://www.treephys.oxfordjournals.org

physiological performance, as seen in both EXP-I and EXP-II. Hence, the olive may also be considered an ‘includer’, i.e., hav-ing the ability to accumulate high amounts of Na ions in shoots (Marschner 1995). The choice of salt translocation mecha-nism, ‘includer’ or ‘excluder’, is largely dependent on the olive K level. In the presence of adequate K, Na is excluded from olive shoots. Only when K is limited, Na is transported to the shoots and substitutes for K. The ability of the olive to flourish under conditions of extremely low environmental K levels indicates an exceptional ability to utilize K or replace it with Na, probably as a result of physiological adaptation to shallow, rocky infertile soils common in its natural habitat (Fernández 2014). It is of little surprise, therefore, that the olive tree is considered to have a low nutrient demand. The combination of low soil fertil-ity and close vicinity of the olive habitat to the Mediterranean sea likely drove adaptation mechanisms enabling the use of Na when the supply of K is short, as found previously for eucalyp-tus (Almeida et al. 2010).

Plant water status and gas exchange parameters The olive is morphologically and physiologically well adapted to long peri-ods of drought (Fernández 2014). The physiological response patterns of olive to drought in the current experiment were comparable to those previously reported (Angelopoulos et al. 1996, Moriana et al. 2002, Bacelar et al. 2007, Diaz-Espejo et al. 2007, Ben-Gal et al. 2010). Subsequent to drought, the peaks of gs and A are early in the morning when WUE is highest (Moriana et al. 2002, Testi et al. 2008) followed by a substantial reduction of all parameters for the remain-der of the day. Similar diurnal patterns of gas exchange and water status parameters were found in the current study’s two experiments. Nonetheless, absolute values were significantly affected by tree nutritional status of extreme K deficiency and,

even more significantly, by limiting Na. Under low supply of 0.20 mM K and sufficient Na (4.0 mM), as demonstrated in EXP-I, the measured physiological response to drought was almost unchanged, only ETR seemed to be impaired by low K. Most importantly, in contrast to our initial hypothesis, neither K nor Na nutrition enhanced drought resistance of the olive. In eucalyptus, K and Na nutrition resulted in more severe drought stress compared with unfertilized control due to rapid decline of soil water (Battie-Laclau et al. 2014b). However, the funda-mental difference between the current and the Battie-Laclau studies is the way drought was controlled. In the Battie-Laclau et al. (2013) study, plants were rainfed whereas in the current study water availability was maintained similar among plants by monitoring and supplying water differentially. In any case, the current study joins the former in challenging the generally accepted understanding that K elevates drought tolerance in plants (Cakmak 2005, Römheld and Kirkby 2010). The main impact of K and Na deficiency found in the current study was pronounced only at the lowest K treatment level, where A, gs and ETR were notably impaired, regardless of the water avail-ability conditions. Unlike the gas exchange parameters, ψl was considerably affected by K and Na nutrition under drought. Leaf or stem water potential is considered to be an accurate indicator for water stress in many fruit trees (Naor et al. 2006). Thus, a possible interaction between ψl and environmental or intrinsic factors is of high interest. Fruit load has been shown to affect ψs under both stressed and non-stressed conditions in grapevines (Naor et al. 1997) and in olives (Ben-Gal et al. 2011, Trentacoste et al. 2011, Naor et al. 2013). The pres-ent study indicates that K and Na nutritional levels also influ-ence the absolute values of ψl, but only under drought. Leaf water potential was highly dependent on the K and Na nutri-tional status (Figure 3). In general, a higher concentration of

Sodium replacement of potassium in the olive 11

Figure 5. Typical appearance of trees with no K and no Na (left), no K and 2.5 mM Na (center) and 2.5 mM K and Na (right) in the irrigation solu-tion. Picture was taken a year after treatment initiation.

at INR

A Institut N

ational de la Recherche A

gronomique on O

ctober 10, 2014http://treephys.oxfordjournals.org/

Dow

nloaded from

Tree Physiology Volume 00, 2014

monovalent cations was correlated with lower ψl. The lower ψl may have been a result of the higher gs and higher tran-spiration leading to tissue dehydration. Similar to our findings, K and Na were found to reduce ψl in eucalyptus trees dur-ing the dry season (Battie-Laclau et al. 2013, 2014b). In the rainy season, predawn leaf water potential was similar for low K, high K or high Na trees. As mentioned, the lower ψl may be a result of enhanced water uptake of K sufficient trees. Additionally, the differences in ψl may be a result of leaf mor-phology and hydraulic properties and due to the effect of K and Na on leaf area. In eucalyptus, parenchyma thickness and total intercellular space increased in response to K and Na nutrition (Battie-Laclau et al. 2014a). Furthermore, bulk elas-tic modulus (ε) has a role in plant water relations (Saito et al. 2006). In eucalyptus leaves, ε was significantly affected by K and Na nutrition (Battie-Laclau et al. 2013). The difference in ψl between trees with comparable water availability is impor-tant for broader understanding of ψl as a stress indicator. As stated above, fruit load has been shown to affect ψl by gen-eral decrease in ψl with elevated fruit load. As a result, olive trees with high fruit load had higher gs for a given ψl (Naor et al. 2013). In a similar manner, K and Na nutritional levels were shown in the present study to impact the gs–ψl interac-tion (Figure 6). Throughout the range of ψl, gs was lower in the deficient treatment, higher in response to Na application and highest in the control. Hence, K and Na nutrition levels should be taken into account when interpreting ψl.

Discussed widely by Kronzucker et al. (2013), the possible role of Na in stomatal movement is of high interest, but as

yet undetermined. The similar response to drought of olive trees with various K and Na levels indicates that their ability to regulate transpiration according to water availability was not impaired, even under severe K deficiency. The lack of response of stomatal movement (not conductance) to either K or Na is surprising as K is considered the main ion accumu-lating in guard cells (Marschner 1995) and since the K level has previously been shown to effect stomatal regulation in olives (Arquero et al. 2006, Benlloch-Gonzalez et al. 2008). Simple carbohydrates such as sucrose also play a substan-tial role in stomatal movement (Talbott and Zeiger 1998) and stomatal regulation through feedback inhibition (Kelly et al. 2013). Sodium was also shown to enable stomatal opening in detached epidermis of C. communis (Willmer and Mansfield 1969, Jarvis and Mansfield 1980). Thus, in K-deficient olive trees, either Na or SCHs may be responsible for the adjust-ment of guard cell osmotic pressure allowing appropriate stomatal opening. Sodium, however, is expected to be less efficient in stomatal response to environmental changes, such as light, CO2 and ABA (Jarvis and Mansfield 1980) since the presence of Na causes impaired stomatal closing. In the halo-phyte A. tripolium, even under high leaf Na concentration, Na was excluded from guard cells and not found to replace the role of K (Perera et al. 1997). The results presented here for olives suggest that stomatal response to drought remains normal even at extremely low K levels with or without Na. Although stomatal regulation was not impaired in the current study, the absolute values of gs tended to be lower under K deficiency, whether under drought or not. Likewise, Battie-Laclau et al. (2014a) found greater gs in eucalyptus trees exposed to high K and high Na. The negative effect of K defi-ciency on gs may be regulated by the ‘stress hormone’ ABA, a major regulator of stomatal movement (Blatt 2000). Abscisic acid is produced in roots in response to drought and increases exponentially as soil water content declines (Dodd 2007). In a split root experiment, Dodd demonstrated that total water availability and heterogeneous water distribution affected ABA transport and consequential root-to-shoot signaling and plant gas exchange parameters (Dodd 2007, Dodd et al. 2008). Hence, variation in root or water distribution within the pots in the current study might have affected the behavior of gs as presented in Figure 3, i.e., nutritional level may have affected root distribution which, in turn, influenced ABA production and stomatal behavior. Such an indirect effect between phytohor-mones and K nutrition was reported in sunflower, which failed to regulate stomatal closure under drought probably due to the involvement of K in ethylene synthesis (Benlloch-González et al. 2010). In a similar way, K and Na nutrition may have had an indirect effect on gs by triggering, encouraging or inhibiting phytohormone synthesis.

One remarkable effect of nutrition was the non-stressed pre-dawn and post-sunset gs seen in EXP-II. All treatments had

12 Erel et al.

Figure 6. Relationship between leaf water potential (ψl) and stoma-tal conductance (gs) for the four extreme treatments: K0-Na0 (filled circles), K0-Na58 (triangles), K15-Na0 (diamonds) and K100-Na58 (open circles). Data were obtained on five selected days during the period of imposed water stress with the addition of the two diurnal measurements. All data were collected between 08:30 and 12:00. Each point represents an average ± standard error of five trees per treatment. Exponential regression was fitted to each treatment.

at INR

A Institut N

ational de la Recherche A

gronomique on O

ctober 10, 2014http://treephys.oxfordjournals.org/

Dow

nloaded from

Tree Physiology Online at http://www.treephys.oxfordjournals.org

approximately double gs at dawn (0.11–0.15 mol H2O m−2 s−1) compared with 0.07 mol H2O m−2 s−1 for the control. Night-time stomatal opening has been reported for other plant species, including those well adapted to dry environments (Caird et al. 2007). Although night transpiration is well documented, the phenomenon is not yet fully understood (Auchincloss et al. 2014). The slow response of stomata to darkness in the pres-ence of Na, as found in C. communis (Jarvis and Mansfield 1980), may partly explain these results. The literature is not consistent in regard to a nutritional effect on stomatal move-ment (Caird et al. 2007). We speculate that night-time transpi-ration, as encountered in the current study, may be an adaptive mechanism to enhance mineral uptake and/or redistribution and utilization under deficit conditions as suggested by Daley and Phillips (2006).

Nutritional effect on photosynthesis, WUE and photoinhibition

EXP-II clearly showed inhibition of assimilation caused by K deficiency as occasionally has been reported in the literature (Cakmak 2005, Römheld and Kirkby 2010, and references within). Our results show that, similar to eucalyptus, Na partially restored A of K-deficient plants (Battie-Laclau et al. 2014a). Cacao K-deficient plants were also shown to have enhanced A under the presence of Na (Gattward et al. 2012). Two major factors may affect net assimilation: (i) stomatal conductance and (ii) assimilation efficiency. The higher A rate in high K trees as displayed in Figure 2 is likely to be partly a consequence of lower stomatal limitation. However, this cannot explain the sig-nificant difference between K0-Na58 and K0-Na0 treatments, which had comparable daytime gs but significantly different A. The assimilation efficiency of the trees was therefore likely affected by K and Na nutrition. In order to better evaluate inter-actions between A, gs and nutrition, we have plotted A response to gs (Figure 7) revealing an asymptotic curve similar to that found by Tognetti et al. (2004). Both K levels and the high Na level had fairly identical response curves, i.e., identical A effi-ciencies. However, at almost any given gs, treatment K0-Na0 had lower A. Thus, in comparison with K0-Na0, the superior A of the high K treatments was derived from both greater gs (Figure 2) and greater assimilation efficiency (Figure 7). The higher A of treatment K0-Na58 compared with K0-Na0 can also be credited to better assimilation efficiency. Therefore, at low environmental K, Na is capable of maintaining proper A by preserving normal assimilation capacity. The reduced assimila-tion efficiency of K- and Na-depleted plants is probably the cause of the lower WUE found in this treatment. This reduction in WUE was completely recovered in the presence of 2.5 mM Na in the irrigation solution. Nutritional level affected WUE only for the irrigated regime and not under drought. The significant effect of water availability on interaction between WUE and K nutrition is likely also to be due to the nature of the A response

to gs. At gs <0.07 mol m−2 s−1, which is typical for drought con-ditions, A was nearly identical among treatments (Figure 7). The difference between K0-Na0 and the remaining treatments was augmented as gs increased. Thus, when water was highly available, gs was generally high and the assimilation rate of K- and Na-deficient plants was substantially impaired.

The maximum photosystem II (PSII) quantum yield, Fv/Fm, is considered a reliable indicator of proper activity of PSII of the photosynthetic system (Maxwell and Johnson 2000) and has been found relevant for olive (Angelopoulos et al. 1996). In the current study, K-deficient plants with no Na suffered damage to their photosynthetic apparatus regardless of the imposed water stress. Sodium addition diminished the apparent pho-toinhibition. Under drought, predawn Fv/Fm indicated photo-damage to all plants, in accordance with Angelopoulos et al. (1996). In spite of this, K0-Na0 had significantly lower Fv/Fm compared with K0-Na58, implying that photoinhibition, similar to A, was positively affected by the presence of Na regardless of plant water status.

Starch and SCHs

The SCH profile of the olive leaf is complex and composed of several non-structural SCHs. A number of SCHs such as mannitol are thought to play a role in responses to abiotic stress (Tattini et al. 1996, Dichio et al. 2009). In the pres-ent study, both starch content and SCHs were significantly affected by K and Na nutrition. Starch concentration was low-est in the absence of K and Na as reported for eucalyptus trees (Battie-Laclau et al. 2014a), probably a result of impaired net assimilation (Table 4) combined with inhibition of enzyme

Sodium replacement of potassium in the olive 13

Figure 7. The relationship between stomatal conductance and assimi-lation rate of the four extreme treatments: K0-Na0 (circles), K0-Na58 (triangles), K15-Na0 (diamonds) and K100-Na58 (squares). For all treatments exponential rise regression was fitted. The raw data from the two diurnal measurements were used, excluding measure-ments taken before 08:30 and after 16:30. Thin lines represent 95% confidence levels.

at INR

A Institut N

ational de la Recherche A

gronomique on O

ctober 10, 2014http://treephys.oxfordjournals.org/

Dow

nloaded from

Tree Physiology Volume 00, 2014

starch synthase (Nitsos and Evans 1969). In contrast to starch, total SCH level was not modified. In eucalyptus, Na-, and to a greater extent, K-fertilized trees had a lower SCH compared with an unfertilized control (Battie-Laclau et al. 2013, 2014a). The SCH concentrations found here are comparable to those reported in the literature for olive (Bustan et al. 2011) but the relative partition of SCHs was strongly influenced by nutrition level. Mannitol content increased with K and Na. Increase in mannitol content was reported previously in olive in response to drought (Dichio et al. 2009), salinity (Tattini et al. 1996) and even K (Chartzoulakis et al. 2006). The opposite (i.e., a decrease) was found for galactinol concentration, which was highest in the absence of K and Na. Galactinol was found to provide protection from abiotic stresses and particularly from oxidative damage (Nishizawa et al. 2008). The higher galac-tinol concentration may be an indicator for oxidative stress in the low K and Na plants as also reflected in the maximum PSII quantum yield.

Conclusions

The physiological response to K and Na nutrition and drought was studied in two experiments using bearing and young olive trees. Olive was found to be highly efficient in K utilization, allowing leaf K to drop to extremely low levels prior to the appearance of deficiency symptoms. Potassium is consid-ered an enhancer of drought resistance; however, we found no evidence of such in olive exposed to gradual water stress. Moreover, stomatal control was not significantly impaired by K deficiency whether under well watered or drought condi-tions. The most pronounced effect of K and Na deficiency was lower stomatal conductance and inferior assimilation capac-ity. Consequentially, the general assimilation rate if no K and Na were applied was roughly half that of sufficient fertilized control. The presence of Na in irrigation solution recovered the assimilation rate to a nearly optimum level by enhancing both conductivity and photosynthetic efficiency. The inhibitory effect of K deficiency on assimilation was more pronounced when stomatal conductance was high and therefore, under drought the effect of K declined. Under drought, leaf water potential decreased with K and Na concentration. High K and Na nutritional levels caused increased leaf starch concentra-tion. Total SCH concentrations were not modified by nutrition, however, the relative partitioning of individual sugars was strongly related to K and Na nutrition. To the best of our knowl-edge, this is the first documentation that the olive is capable of partially replacing K with Na. Under conditions of deficit K, the addition of Na resulted in enhanced growth, higher stomatal conductivity, superior assimilation capacity, higher leaf starch concentration and reduced deficiency symptoms. The positive effect of Na on root growth, assimilation performance, photo-damage and carbohydrate profile indicates that Na may play a

metabolic role in addition to its acknowledged osmotic role. The nutritional effect of K and Na was shown to be independent of plant water status.

Conflict of interest

None declared.

References

Almeida JCR, Laclau J-P, Gonçalves JLdM, Ranger J, Saint-André L (2010) A positive growth response to NaCl applications in euca-lyptus plantations established on K-deficient soils. For Ecol Manag 259:1786–1795.

Angelopoulos K, Dichio B, Xiloyannis C (1996) Inhibition of photo-synthesis in olive trees (Olea europaea L.) during water stress and rewatering. J Exp Bot 47:1093–1100.

Anschütz U, Becker D, Shabala S (2014) Going beyond nutrition: regulation of potassium homoeostasis as a common denominator of plant adaptive responses to environment. J Plant Physiol 171: 670–687.

Armengaud P, Sulpice R, Miller AJ, Stitt M, Amtmann A, Gibon Y (2009) Multilevel analysis of primary metabolism provides new insights into the role of potassium nutrition for glycolysis and nitrogen assimila-tion in Arabidopsis roots. Plant Physiol 150:772–785.

Arquero O, Barranco D, Benlloch M (2006) Potassium starva-tion increases stomatal conductance in olive trees. HortScience 41:433–436.

Auchincloss L, Easlon HM, Levine D, Donovan L, Richards JH (2014) Pre-dawn stomatal opening does not substantially enhance early-morning photosynthesis in Helianthus annuus. Plant Cell Environ 37:1364–1370.

Bacelar E, Santos D, Moutinho-Pereira J, Lopes J, Gonçalves B, Ferreira T, Correia C (2007) Physiological behaviour, oxidative damage and antioxidative protection of olive trees grown under different irriga-tion regimes. Plant Soil 292:1–12.

Battie-Laclau P, Laclau J-P, Piccolo M et al. (2013) Influence of potas-sium and sodium nutrition on leaf area components in Eucalyptus grandis trees. Plant Soil 371:19–35.

Battie-Laclau P, Laclau J-P, Beri C et al. (2014a) Photosynthetic and anatomical responses of Eucalyptus grandis leaves to potassium and sodium supply in a field experiment. Plant Cell Environ 37:70–81.

Battie-Laclau P, Laclau J-P, Domec J-C et al. (2014b) Effects of potas-sium and sodium supply on drought-adaptive mechanisms in Eucalyptus grandis plantations. New Phytol 203:401–413.

Ben-Gal A, Kool D, Agam N et al. (2010) Whole-tree water balance and indicators for short-term drought stress in non-bearing ‘Barnea’ olives. Agric Water Manag 98:124–133.

Ben-Gal A, Yermiyahu U, Zipori I, Presnov E, Hanoch E, Dag A (2011) The influence of bearing cycles on olive oil production response to irrigation. Irrig Sci 29:253–263.

Benlloch-Gonzalez M, Arquero O, Fournier JM, Barranco D, Benlloch M (2008) K+ starvation inhibits water-stress-induced stomatal closure. J Plant Physiol 165:623–630.

Benlloch-González M, Romera J, Cristescu S, Harren F, Fournier JM, Benlloch M (2010) K+ starvation inhibits water-stress-induced sto-matal closure via ethylene synthesis in sunflower plants. J Exp Bot 61:1139–1145.

Blatt MR (2000) Cellular signaling and volume control in stomatal movement in plants. Annu Rev Cell Dev Biol 16:221–241.

Bustan A, Avni A, Lavee S, Zipori I, Yeselson Y, Schaffer AA, Riov J, Dag A (2011) Role of carbohydrate reserves in yield production of intensively cultivated oil olive (Olea europaea L.) trees. Tree Physiol 31:519–530.

14 Erel et al.

at INR

A Institut N

ational de la Recherche A

gronomique on O

ctober 10, 2014http://treephys.oxfordjournals.org/

Dow

nloaded from

Tree Physiology Online at http://www.treephys.oxfordjournals.org

Caird MA, Richards JH, Donovan LA (2007) Nighttime stomatal conduc-tance and transpiration in C3 and C4 plants. Plant Physiol 143:4–10.

Cakmak I (2005) The role of potassium in alleviating detrimental effects of abiotic stresses in plants. J Plant Nutr Soil Sci 168:521–530.

Chartzoulakis KS (2005) Salinity and olive: growth, salt tolerance, photosynthesis and yield. Agric Water Manag 78:108–121.

Chartzoulakis K, Psarras G, Vemmos S, Loupassaki M, Bertaki M (2006) Response of two olive cultivars to salt stress and potassium supplement. J Plant Nutr 29:2063–2078.

Daley MJ, Phillips NG (2006) Interspecific variation in nighttime tran-spiration and stomatal conductance in a mixed New England decidu-ous forest. Tree Physiol 26:411–419.

Diaz-Espejo A, Nicolás E, Fernández JE (2007) Seasonal evolution of diffusional limitations and photosynthetic capacity in olive under drought. Plant Cell Environ 30:922–933.

Dichio B, Margiotta G, Xiloyannis C, Bufo S, Sofo A, Cataldi TI (2009) Changes in water status and osmolyte contents in leaves and roots of olive plants (Olea europaea L.) subjected to water deficit. Trees 23:247–256.

Dodd IC (2007) Soil moisture heterogeneity during deficit irrigation alters root-to-shoot signalling of abscisic acid. Funct Plant Biol 34:439–448.

Dodd IC, Egea G, Davies WJ (2008) Abscisic acid signalling when soil moisture is heterogeneous: decreased photoperiod sap flow from drying roots limits abscisic acid export to the shoots. Plant Cell Environ 31:1263–1274.

Egilla J, Davies F Jr, Drew M (2001) Effect of potassium on drought resistance of Hibiscus rosa-sinensis cv. Leprechaun: plant growth, leaf macro- and micronutrient content and root longevity. Plant Soil 229:213–224.

Erel R, Dag A, Ben-Gal A, Schwartz A, Yermiyahu U (2008) Flowering and fruit set of olive trees in response to nitrogen, phosphorus, and potassium. J Am Soc Hort Sci 133:639–647.

Erel R, Kerem Z, Ben-Gal A, Dag A, Schwartz A, Zipori I, Basheer L, Yermiyahu U (2013a) Olive (Olea europaea L.) tree nitrogen status is a key factor for olive oil quality. J Agric Food Chem 61:11261–11272.

Erel R, Yermiyahu U, Van Opstal J, Ben-Gal A, Schwartz A, Dag A (2013b) The importance of olive (Olea europaea L.) tree nutritional status on its productivity. Sci Hortic 159:8–18.

Fernandez-Escobar R (2010) Fertilization. In: Barranco D, Fernandez-Escobar R, Rallo L (eds) Olive growing. RIRDC, Canberra, Australia, pp 267–297.

Fernández J-E (2014) Understanding olive adaptation to abiotic stresses as a tool to increase crop performance. Environ Exp Bot 103:158–179.

Flowers T, Lauchli A (1983) Sodium versus potassium: substitution and compartmentation. In: Läuchli A, Bieleski RL (eds) Inorganic plant nutrition. Springer, Berlin, pp 651–681.

Gattward JN, Almeida A-AF, Souza JO, Gomes FP, Kronzucker HJ (2012) Sodium–potassium synergism in Theobroma cacao: stimula-tion of photosynthesis, water-use efficiency and mineral nutrition. Physiol Plant 146:350–362.

Gucci R, Tattini M (1997) Salinity tolerance in olive. Hortic Rev 21:177–214.Hampe T, Marschner H (1982) Effect of sodium on morphology,

water relations and net photosynthesis of sugar beet leaves. Z Pflanzenphysiol 108:151–162.

Hartmann HT, Uriu K, Lilleland O (1966) Olive nutrition. In: Childers NF (ed) Fruit nutrition. Horticultural Publications, Rutgers University, NJ, pp 252–261.

Jarvis RG, Mansfield TA (1980) Reduced stomatal responses to light, carbon dioxide and abscisic acid in the presence of sodium ions. Plant Cell Environ 3:279–283.

Kelly G, Moshelion M, David-Schwartz R, Halperin O, Wallach R, Attia Z, Belausov E, Granot D (2013) Hexokinase mediates stomatal clo-sure. Plant J 75:977–988.

Kronzucker HJ, Coskun D, Schulze LM, Wong JR, Britto DT (2013) Sodium as nutrient and toxicant. Plant Soil 369:1–23.

Lindhauer MG (1985) Influence of K nutrition and drought on water relations and growth of sunflower (Helianthus annuus L.). Z Pflanzen Bodenkunde 148:654–669.

Marschner H (1995) Mineral nutrition of higher plants, 2nd edn. Academic Press, London.

Maxwell K, Johnson GN (2000) Chlorophyll fluorescence—a practical guide. J Exp Bot 51:659–668.

Miller GL (1959) Use of dinitrosalicylic acid reagent for determination of reducing sugar. Anal Chem 31:426–428.

Moriana A, Villalobos FJ, Fereres E (2002) Stomatal and photosyn-thetic responses of olive (Olea europaea L.) leaves to water deficits. Plant Cell Environ 25:395–405.

Naor A, Gal Y, Bravdo B (1997) Crop load affects assimilation rate, stomatal conductance, stem water potential and water relations of field-grown Sauvignon blanc grapevines. J Exp Bot 48:1675–1680.

Naor A, Gal Y, Peres M (2006) The inherent variability of water stress indicators in apple, nectarine and pear orchards, and the validity of a leaf-selection procedure for water potential measurements. Irrig Sci 24:129–135.

Naor A, Schneider D, Ben-Gal A, Zipori I, Dag A, Kerem Z, Birger R, Peres M, Gal Y (2013) The effects of crop load and irrigation rate in the oil accumulation stage on oil yield and water relations of ‘Koroneiki’ olives. Irrig Sci 31:781–791.

Nishizawa A, Yabuta Y, Shigeoka S (2008) Galactinol and raffinose constitute a novel function to protect plants from oxidative damage. Plant Physiol 147:1251–1263.

Nitsos RE, Evans HJ (1969) Effects of univalent cations on the activity of particulate starch synthetase. Plant Physiol 44:1260–1266.

Perera LKRR, De Silva DLR, Mansfield TA (1997) Avoidance of sodium accumulation by the stomatal guard cells of the halophyte Aster tripolium. J Exp Bot 48:707–711.

Premachandra GS, Saneoka H, Ogata S (1991) Cell membrane stability and leaf water relations as affected by potassium nutrition of water-stressed maize. J Exp Bot 42:739–745.

Römheld V, Kirkby E (2010) Research on potassium in agriculture: needs and prospects. Plant Soil 335:155–180.

Saito T, Soga K, Hoson T, Terashima I (2006) The bulk elastic modulus and the reversible properties of cell walls in developing Quercus leaves. Plant Cell Physiol 47:715–725.

Sen Gupta A, Berkowitz GA, Pier PA (1989) Maintenance of photosyn-thesis at low leaf water potential in wheat: role of potassium status and irrigation history. Plant Physiol 89:1358–1365.

Shabala S, Pottosin I (2014) Regulation of potassium transport in plants under hostile conditions: implications for abiotic and biotic stress tolerance. Physiol Plant 151:257–279.

Subbarao GV, Ito O, Berry WL, Wheeler RM (2003) Sodium—a functional plant nutrient. Crit Rev Plant Sci 22:391–416.

Talbott LD, Zeiger E (1998) The role of sucrose in guard cell osmo-regulation. J Exp Bot 49:329–337.

Tattini M, Gucci R, Romani A, Baldi A, Everard JD (1996) Changes in non-structural carbohydrates in olive (Olea europaea) leaves during root zone salinity stress. Physiol Plant 98:117–124.

Testi L, Orgaz F, Villalobos F (2008) Carbon exchange and water use efficiency of a growing, irrigated olive orchard. Environ Exp Bot 63: 168–177.

Tognetti R, d'Andria R, Morelli G, Calandrelli D, Fragnito F (2004) Irrigation effects on daily and seasonal variations of trunk sap flow and leaf water relations in olive trees. Plant Soil 263:249–264.

Trentacoste ER, Sadras VO, Puertas CM (2011) Effects of the source : sink ratio on the phenotypic plasticity of stem water potential in olive (Olea europaea L.). J Exp Bot 62:3535–3543.

Sodium replacement of potassium in the olive 15

at INR

A Institut N

ational de la Recherche A

gronomique on O

ctober 10, 2014http://treephys.oxfordjournals.org/

Dow

nloaded from

Tree Physiology Volume 00, 2014

Véry A-A, Robinson MF, Mansfield TA, Sanders D (1998) Guard cell cation channels are involved in Na+-induced stomatal closure in a halophyte. Plant J 14:509–521.

Vishnevetsky J, Zamski E, Ziv M (2000) Carbohydrate metabolism in Nerine sarniensis bulbs developing in liquid culture. Physiol Plant 108:361–369.

Wakeel A, Farooq M, Qadir M, Schubert S (2011) Potassium substitu-tion by sodium in plants. Crit Rev Plant Sci 30:401–413.

Willmer CM, Mansfield TA (1969) A critical examination of the use of detached epidermis in studies of stomatal physiology. New Phytol 68:363–375.

16 Erel et al.

at INR

A Institut N

ational de la Recherche A

gronomique on O

ctober 10, 2014http://treephys.oxfordjournals.org/

Dow

nloaded from