Root and cell hydraulic conductivity, apoplastic barriers and aquaporin gene expression in barley (Hordeum vulgare L.) grown with low supply of potassium

Abstract

Background and Aims

Limited supply of mineral nutrients often reduces plant growth and transpirational water flow while increasing the ratio of water-absorbing root to water-losing shoot surface. This could potentially lead to an imbalance between water uptake (too much) and water loss (too little). The aim of the present study was to test whether, as a countermeasure, the hydraulic properties (hydraulic conductivity, Lp) of roots decrease at organ and cell level and whether any decreases in Lp are accompanied by decreases in the gene expression level of aquaporins (AQPs) or increases in apoplastic barriers to radial water movement.

Methods

Barley plants were grown hydroponically on complete nutrient solution, containing 2 mm K⁺ (100 %), or on low-K solution (0.05 mm K⁺; 2.5 %), and analysed when they were 15–18 d old. Transpiration, fresh weight, surface area, shoot water potential (ψ), K and Ca concentrations, root (exudation) and cortex cell Lp (cell pressure probe), root anatomy (cross-sections) and AQP gene expression (qPCR) were analysed.

Key Results

The surface area ratio of root to shoot increased significantly in response to low K. This was accompanied by a small decrease in the rate of water loss per unit shoot surface area, but a large (~50 %) and significant decrease in Lp at root and cortex cell levels. Aquaporin gene expression in roots did not change significantly, due to some considerable batch-to-batch variation in expression response, though HvPIP2;5 expression decreased on average by almost 50 %. Apoplastic barriers in the endodermis did not increase in response to low K.

Conclusions

Barley plants that are exposed to low K adjust to an increased ratio of root (water uptake) to shoot (water loss) surface primarily through a decrease in root and cell Lp. Reduced gene expression of HvPIP2;5 may contribute to the decrease in Lp.

INTRODUCTION

Potassium is a macronutrient, which is required by plants in comparatively large quantities due to its ubiquitous role as a major cytosolic cationic osmolyte, ionic interaction with charged side groups of macromolecules and plasma-membrane-related transport processes (Leigh and Jones, 1984; Fricke et al., 1994). Limited supply of K affects plant growth and yield, on a global scale, and the response of plants to biotic and abiotic stresses such as N supply, salinity and soil oxygen availability (for review, see Wang et al., 2013). A typical response of plants to the limited supply of a macronutrient such as N or P is an increase in the fresh weight (FW) or surface area ratio between root and shoot (for reviews see Clarkson et al., 2000; Hermans et al., 2006; Koevoets et al., 2016). Such a response can also be observed for K (e.g. Andrews et al., 1999; Gruber et al., 2013; Singh and Reddy, 2017), though to a smaller and more varied extent than for N and P.

An increase in root-to-shoot ratio makes it possible to allocate resources specifically to that part of the plant (root system) that forages the soil for the limiting nutrient, while decreasing that part of the plant (shoot) that is a sink for that nutrient. Such a shift in the ratio between root and shoot also impacts on the water balance of plants. The water-absorbing portion of the plant increases relative to the water-losing portion of the plant. Thus, a plant grown with limited K supply potentially ends up with the unusual situation that it takes up more water through the root system than it can lose through the shoot. This is de facto not possible, as plant cells have a high volumetric elastic modulus, which causes turgor and water potential (ψ) to increase considerably per fractional volume increase (Volkov et al., 2007; Sharipova et al., 2016) and permits only limited transient storage of water. The two basic mechanisms available to address the potential imbalance between water uptake and loss are to (1) increase the rate of water loss per unit shoot area involving stomatal function; and to (2) decrease the hydraulic properties of the root (root hydraulic conductivity, Lp; volumetric water uptake per unit time, surface area and biophysical driving force) or the biophysical force that drives root water uptake. The latter is essentially the difference in water potential between root xylem and root medium. It is possible, therefore, that some of the plant responses to a low supply of K⁺ are not so much linked to the actual nutritional aspect of K⁺ but are more aimed at regulating the water balance.

Root water uptake is particularly limited through the radial transport properties of the cellular path from the epidermis to the xylem across the root cylinder (Frensch and Steudle, 1989; Steudle, 2000). In plants such as barley, the vast majority of water moves radially from cell to cell along a path that involves the crossing of membranes, as opposed to an apoplastic path (Steudle and Jeschke, 1983; Knipfer and Fricke, 2010). This does not rule out the possibility that apoplastic barriers, such as the suberin lamella in the endodermis, which are located perpendicular to the radial flow direction and external to the plasma membrane, affect the rate of radial water transport (for review see Geldner 2013). Transport of water across membranes involves aquaporin (AQP) function, in particular those AQPs that are localized at the plasma membrane (plasma membrane intrinsic proteins, PIPs). Aquaporins belong to the family of major intrinsic proteins. They are best known for their ability to facilitate the diffusion of water but can also transport a range of small-molecular-weight solutes such as boron, ammonia, urea, silicon, glycerol, hydrogen peroxide and gases such as carbon dioxide. Aquaporins have been shown to be regulated at the gene transcriptional, protein and trafficking levels and are key to the regulation of root Lp in many plant species (for reviews see Chaumont and Tyerman, 2014; Maurel et al., 2015; Gambetta et al., 2017; Tyerman et al., 2017), including barley (Knipfer et al., 2011). Mineral nutrition affects AQP gene expression and root Lp (for reviews see Aroca et al., 2012; Wang et al., 2016), but detailed studies linking the two and including Lp analyses at cellular level combined with root anatomical studies in plants grown with a low K⁺ supply are missing.

The aim of the present study was to carry out a detailed analysis of changes in gene expression level (qPCR analyses) of PIP aquaporins in barley grown with a low K supply and relate these changes to parallel changes in root Lp (exudation analyses), root cortex cell Lp (cell pressure probe analyses), root anatomy (cross-sections stained for Casparian bands and suberin lamella) and any changes in root-to-shoot ratio. Transpirational water loss per unit shoot surface area was determined to investigate any stomatal regulation of shoot water loss. Cell pressure probe analyses (turgor) were accompanied by picolitre osmometry of extracted cell sap (osmotic pressure) to calculate cortex cell water potential. Together with analyses of leaf water potential, this made it possible to estimate the biophysical force driving water uptake across the root cylinder of intact, transpiring plants.

MATERIALS AND METHODS

Plant material and growth conditions

Barley (Hordeum vulgare ‘Quench’) plants were grown on modified half-strength Hoagland solution in a growth chamber (Microclima, MC1000HE, CEC Technology, Glasgow, UK) as described previously (Knipfer and Fricke, 2011). The control nutrient solution contained 2 mm K⁺ (100 %). The low-K solution contained 0.05 mm K⁺ (2.5 %). As some of the nitrate in the control solution was added as KNO₃, and to make sure that low-K plants received the control level of N, low-K plants received an additional 1 mm Ca(NO₃)₂ (control, 2 mm Ca(NO₃)₂; low K, 3 mm Ca(NO₃)₂).

The root medium was aerated, and plants grew at a day/night length of 16/8 h and temperature of 21/15 °C. Relative humidity was 75 % and photosynthetically active radiation at plant level was 300–350 μmol m⁻² s⁻¹. The vapour pressure difference was 0.427 kPa during the night and 0.622 kPa during the day. All plants were germinated for 6 d on CaSO₄ solution (0.5 mm), and then transferred to control or low-K nutrient solution. Plants were analysed when they were 15–18 d old. At that developmental stage, leaf 3 was expanding and leaf 4 was just emerging. All analyses were carried out 4–8 h into the photoperiod.

The water potential of the medium at the time of plant harvest was determined with a Vapro (Wescor Inc., South Logan, UT, USA) osmometer for four or five replicate samples.

Transpiration measurements

Transpirational water loss of plants in the growth chamber was determined gravimetrically (Knipfer and Fricke, 2011) using a balance (Model CP323P, Sartorius, Göttingen, Germany) as described previously (Meng et al., 2016). To obtain a measure of stomatal conductance (m s⁻¹), transpirational water loss rate (m³ s⁻¹) was related to shoot surface area (m²).

Fresh weight and surface area

Following transpiration analyses, the FW of the shoot and root system of each plant was determined with an analytical balance. The shoot and root system was scanned (Canon 9900F) for subsequent determination of surface area. Scanned images were analysed with the freely available software ImageJ (www.imagej.nih.gov/ij/). To increase the contrast of root images, roots were stained in 0.25 % Coomassie Brilliant Blue for 2 d prior to scanning (Kano-Nakata et al., 2012). Roots were treated as cylinders for surface area calculations (Suku et al., 2014).

Hydraulic analyses

Hydraulic analyses were carried out in a normal laboratory environment, without (root exudation) or with (cell pressure probe analyses on intact plants) supplementary lighting to keep plant transpiration rates comparable with those in the growth chamber. Ambient air temperature and temperature of root media was between 18 and 22 °C.

Excised root systems were analysed through exudation experiments as described previously (Suku et al., 2014). Exudate was collected for 30–60 min, and the osmotic pressures of exudates and root media were analysed using a Vapro (Wescor, South Logan, UT 84321, USA) osmometer. Root hydraulic conductivity (m s⁻¹ MPa⁻¹) was calculated by relating the exudate flow rate (m³ s⁻¹) to root surface area (m²) and the osmotic force driving water uptake (difference in osmotic pressure between exudate and root medium; unit MPa; for calculations see Knipfer and Fricke 2010, 2011).

Cell turgor, half-time (t_1/2) of water exchange, elastic modulus (ε) and cell Lp were determined with the cell pressure probe as described previously (e.g. Knipfer et al., 2011; Sharipova et al., 2016). Cell osmotic pressure (π), which is required for calculation of cell Lp, was determined through picolitre osmometry of extracted cell sap (Fricke and Peters, 2002). The volume and surface area of cells, which are required for calculation of cell Lp, were determined using free-hand cross-sections assuming that cells were shaped like cylinders. Root cortex cells were analysed in the root hair region. The cells were located 0.9–1.1 cm from the tip of seminal roots, in the first three cortical cell layers beneath the root epidermis.

Molecular analyses

Gene expression analyses [quantitative PCR (qPCR)] were carried out as detailed previously (Meng et al., 2016). The harvested (liquid N) root systems of two or three plants were pooled together into one biological replicate and ground into a fine powder in liquid N. The powder was used to extract RNA (RNeasy Mini Kit, Qiagen), and the extract was treated with DNase I (Invitrogen) and used to synthesize cDNA (SuperScript II Reverse Transcriptase, Invitrogen, San Diego, CA, USA). Real-time qPCR was carried out using a Stratagene rapid cycler and SYBR Green as reagent (Takara Bio, Otsu, Shiga, Japan) on 96-well plates. Eleven or 12 biological replicates, which were derived from four different batches of plants (two or three replicates per batch), were analysed for each treatment. The control and low-K treatments were always grown in parallel.

The calculation of qPCR data was based on the ΔCt method (Pfaffl, 2001); details of calculations are provided in Meng et al. (2016). Originally, we wanted to use three housekeeping genes as references of expression, as in previous studies (Besse et al., 2011; Knipfer et al., 2011; Meng et al., 2016; Meng and Fricke, 2017). We tested five reference genes (ubiquitin, tubulin, plasma membrane H⁺-ATPase, HSP70, cyclophylin); however, all reference genes except ubiquitin gave occasionally very high or low expression values (Ct) for low-K plants. For this reason, we decided to use only ubiquitin as reference gene and base all calculations of ∆Ct values on this reference gene. Nine AQPs were analysed. This included eight PIP genes, which had previously shown significant expression in roots of barley (Besse et al., 2011; Knipfer et al., 2011), and a highly expressed tonoplast intrinsic aquaporin protein (TIP), TIP1;1 (Knipfer et al., 2011). The latter was included to see whether any expression response of PIPs to low K may be PIP-specific. Sequences of primers are given in Besse et al. (2011).

Root anatomical analyses

Root anatomy was studied on free-hand cross-sections, which were made from plant material stored in 50 % ethanol. Sections were made at two positions along the root axis, at 9–11 mm (within the root hair region) and at 65–70 mm from the tip (within the mature, lateral root region). Sections were observed with a Leica microscope (DM IL; Leica, Wetzlar, Germany) and captured with a digital camera (DFC300 FX; Leica, Wetzlar, Germany). For the detection of Casparian bands, sections were stained for 50–60 min with 0.1 % (w/v) berberine hemisulphate, counterstained for 20–30 min with 0.5 % (w/v) aniline blue and mounted in 0.1 % (w/v) FeCl₃ in 50 % (v/v) glycerol (Brundrett et al., 1988). Sections were viewed under fluorescent light using a UV/violet filter with an excitation wavelength of 390–420 nm (Knipfer and Fricke, 2011). Suberin and lipid deposits were visualized by staining sections with Sudan Red 7B for 1.5 h (Brundrett et al., 1991). The same sections were also used to detect autofluorescence of lignified tissues, using the filter setup (UV illumination) as employed for berberine hemisulphate-stained sections.

Shoot water potential

Shoot water potential was determined using a plant pressure chamber and following the manufacturer’s instructions (SKPM 1400 SKYE, UK).

Elemental analyses

Concentrations of K and Ca were determined through flame atomic absorption spectrometry in root and shoot tissue and in exudate of excised root systems. Plants, which were 16 d old, had their shoot removed. The shoot was used for extraction of K and Ca, and the remaining root system was used for exudation measurements, in the same way as for determination of root hydraulic conductivity. At the end of exudation measurements, the exudate was collected, as was the root system, and both were used for determination of concentrations of K and Ca. Root and shoot tissues were transferred into centrifuge tubes that contained a mesh insert at the bottom, and tissues were frozen, thawed and then spun for 5 min at 10 000 g to collect undiluted sap at the bottom of the centrifuge tube. This sap was used, together with root exudates, for determination of K and Ca. Samples were diluted between 250- and 10 000-fold, and dilution factors were used to convert the p.p.m. outputs of the flame atomic absorption spectrometer into millimolar concentrations using known molecular weights of K and Ca.

Statistical analyses

Data were subjected to Student’s t-test (Figs 1–4) and one-/two-way analysis of variance (ANOVA) (Supplementary Data Fig. S1) using functions in Minitab.

Fig. 1.

Transpiration, fresh weight and root and shoot surface area of 15- to 18-d-old barley plants grown on nutrient-sufficient (control, 2 mm K⁺) or low-K medium (0.05 mm K⁺). (A, B) Transpirational water loss (A) per plant and (B) per root surface area (RSA) and shoot surface area (SSA). (C) Fresh weight of plants, shoot and root system. (D) Surface area of shoot and root system. (E) Root-to-shoot ratios of fresh weight (FW) and surface area (SA). Results are averages and s.e. (error bars) of 62 plant analyses, which were derived from 11 independent batches of plants (control and low-K plants always grown in parallel). *P < 0.05; **P < 0.01; ***P < 0.001.

Fig. 4.

Gene expression of eight plasma membrane intrinsic proteins (PIPs) and one tonoplast intrinsic aquaporin protein (TIP) in roots of 15- to 18-d-old barley plants grown on nutrient-sufficient (control 2 mm K⁺) or low-K medium (0.05 mm K⁺). Expression of aquaporin (AQP) genes was related to that of the reference gene ubiquitin (∆Ct values); the 2^−(∆Ct) values shown represent the fold-time expression of the AQP gene relative to that of the reference gene (value set to 1.0). Results are averages and s.e. (error bars) of the analysis of 11 or 12 plant samples derived from four independent batches of plants (control and low-K plants always grown in parallel) and each containing the pooled root system of two or three plants. All differences in gene expression between control and low-K plants were statistically non-significant (Student’s t-test).

RESULTS

Transpiration, fresh weight, surface area, root-to-shoot ratios and concentrations of K and Ca

The rate of daytime transpirational water loss decreased significantly and by almost 40 % in response to low-K treatment, from 4.91 × 10⁻¹¹ m³ s⁻¹ in control to 2.99 × 10⁻¹¹ m³ s⁻¹ in low-K plants (Fig. 1A). Transpirational water loss (reflecting root water uptake), expressed per unit shoot and root surface area, also decreased significantly in response to low K, particularly in relation to root surface area (Fig. 1B). Plant and shoot FW decreased highly significantly in response to low K. At the root level, FW decreased by only 12 % (P < 0.05; Fig. 1C). Shoot surface area decreased, whereas root surface area increased significantly in response to low K (Fig. 1D). As a result, the root-to-shoot ratio of FW and surface area increased significantly by 35 and 55 %, respectively, in response to low K (Fig. 1E). Shoot water potential averaged −0.59 MPa in control and −0.64 MPa in low-K plants; this difference was statistically non-significant (eight to ten plant analyses; not shown).

The concentration of K in roots and shoots was significantly lower in low-K compared with control plants (Supplementary Data Fig. S1). Calcium was significantly higher in the shoot, but not root, of low-K compared with control plants (Supplementary Data Fig. S1).

Root exudation analyses

The exudation rate of excised root systems decreased non-significantly, by 18 %, in response to low K (Fig. 2A). The osmotic force, which drove root water uptake during exudation analyses, decreased by 7 % (Fig. 2B). Due to an increased surface area of the low-K compared with control roots used for exudation analyses (not shown), the rate of exudate flow per unit driving force and root surface area (exudation Lp) decreased significantly by 42 % in response to the low-K treatment (Fig. 2C).

Fig. 2.

Root exudation data of 15- to 18-d-old barley plants grown on nutrient-sufficient (control, 2 mm K⁺) or low-K medium (0.05 mm K⁺). (A) Exudation rate, (B) osmotic force driving radial root water uptake, and (C) exudation root hydraulic conductivity (Lp). Results are averages and s.e. (error bars) of 8–17 plant analyses, which were derived from three independent batches of plants (control and low-K plants always grown in parallel). *P < 0.05; ns, not significant.

The concentration of K in the exudate of excised root systems averaged close to 10 mm in control plants but only ~1 mm in low-K plants (Supplementary Data Fig. S1). The Ca concentration in exudate was around 2 mm in both treatments. As a result, the concentration ratio of K:Ca in exudate was significantly higher in control compared with low-K plants; it did not differ significantly from the K:Ca concentration ratio in bulk root tissue (Supplementary Data Fig. S1).

Cell pressure probe and osmotic analyses

The t_1/2 of water exchange of root cortex cells increased significantly from 1.71 s in control to 2.92 s in low-K plants (Fig. 3A). Cell elastic modulus, equalling the change in turgor pressure per fractional change in cell volume (∆V/V), did not differ significantly between treatments (Fig. 3B). The same applied to the volume of cells, which averaged 146 pL in control and 135 pL in low-K plants (average of 50–60 microscopic cell analyses using cross-sections and longitudinal root sections; not shown). Cell turgor decreased highly significantly in response to low K, from 0.64 MPa in control to 0.51 MPa in low-K plants (Fig. 3C). The same applied to cell osmotic pressure, which decreased from 0.81 to 0.60 MPa (Fig. 3D). As a result, cell water potential, being calculated as the difference between average turgor and osmotic pressure, was −0.179 MPa in control and −0.086 MPa in low-K plants. In comparison, the water potential of nutrient media at the time of plant harvest averaged −0.056 MPa (not shown). Cell Lp decreased significantly by 55 % in response to the low-K treatment (Fig. 3F).

Fig. 3.

Cell pressure probe and picolitre osmometry analyses of root cortex cells of 15- to 18-d-old barley plants. Plants were grown on nutrient-sufficient (control, 2 mm K⁺) or low-K medium (0.05 mm K⁺). Cortex cells were analysed in the root hair region of intact, transpiring plants. (A–C) A cell pressure probe was used to determine the cell (A) half-time of water exchange, (B) elastic modulus and (C) turgor. (D) Cell osmotic pressure was analysed by picolitre osmometry of extracted cell sap, where each sample consisted of the pooled sap of four or five cells sampled in quick (seconds) succession. (E) Data on average turgor and osmotic pressure were used to calculate cell water potential. (F) Cell pressure probe data, osmotic pressure and data on cell dimensions (not shown) were used to calculate cell hydraulic conductivity. Results are averages and s.e. (error bars) of 10–15 cell analyses (cell pressure probe) or analyses of five samples (osmotic pressure). *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant.

Aquaporin gene expression

A total of eight PIPs and one TIP AQP were analysed for gene expression. There was no general trend in gene expression response to low K, and none of the changes observed was statistically significant (Fig. 4). Five of the AQPs (HvPIP1;1, HvPIP1;2, HvPIP2;4, HvPIP2;5, HvTIP1;1) showed a decrease, two AQPs (HvPIP1;3, HvPIP2;2) showed an increase and two AQPs (HvPIP1;4, HvPIP2;3) showed no change in gene expression in response to low K. The largest decrease (49 %) in gene expression, and lowest P-value (0.159), in response to low K was observed for HvPIP2;5.

Root anatomy

Sections in the root hair region showed mostly no (Fig. 5A, B) or occasionally very few Casparian bands (not shown), irrespective of the growth conditions. In the mature, lateral root region, where endodermis development has proceeded to stages II and III (Enstone et al., 2003; Knipfer and Fricke, 2011), sections of control and low-K plants showed Casparian bands and suberin lamella. Heavily thickened secondary walls in endodermal cells could also be observed (Fig. 5C, D). There was no apparent difference in the expression of these anatomical features between control plants and plants grown with low K. None of the sections studied showed any signs of an exodermis (not shown).

Fig. 5.

Root anatomy of barley plants grown under control conditions (with 2 mm K) or in the presence of low K (0.05 mm K). Plants were 16–18 d old at the time of harvest. Sections were made at two positions along the main axis of seminal roots, at 9–11 mm (root hair region) and 65–70 mm from the tip (mature, lateral root region). Casparian bands (bright yellow–blueish signal) were stained with berberine hemisulphate, and suberin and lipid deposits (intense red signal) were visualized by staining with Sudan Red 7B. The Sudan Red-stained sections were also used to detect autofluorescence (blue) of lignified tissues (UV light illumination). (A–D) Subsection of the root cross-section, focusing on the endodermis, pericycle, xylem parenchyma and early metaxylem vessels. (A, B) Sections taken 9–11 mm from the root tip of (A) control and (B) low-K plants, and (A1, B1) viewed under bright light, (A2, B2) stained and viewed for Casparian bands, (A3, B3) stained and viewed for suberin lamella, and (A4, B4) showing autofluorescence of lignified tissues. (C, D), as for (A, B), except that sections were taken 65–70 mm from the root tip. The same sections were viewed under bright light and for Casparian bands, and for suberin lamella and autofluorescence. (E) Explanation of tissues and anatomical features, showing a berberin hemisulphate-stained/aniline blue-counterstained section viewed under UV illumination. The central part of root section, containing the stele, is shown together with several layers of cortex cells; an exodermis near the epidermis could not be observed in any of the sections studied (not shown). The yellow/red arrows in C2 and D2 point to Casparian bands, and point to suberin lamella in C3 and D3. CC-L, cortex cell layers; PC-L, pericycle layer; EMX, early metaxylem vessel; LMX, late metaxylem vessel; END-L, endodermis layer; CBs, Casparian bands; LSL, location of suberin lamella in endodermis; PH, phloem; TSW, thickened secondary wall in endodermis; XP, xylem parenchyma.

DISCUSSION

Involvement of aquaporins in the Lp response to low K

Most, if not all, radial water uptake in barley roots occurs along a cell-to-cell path (Steudle and Jeschke, 1983; Knipfer and Fricke, 2010) and involves AQP function (Knipfer et al., 2011). Application of the AQP inhibitor HgCl₂ reduced Lp at root level by 52 % and at cortex cell level (root hair zone) by 96 % in barley (Knipfer et al., 2011). Suberin lamellas, which are located externally to the tangential planes of the plasma membrane of endodermal cells (Geldner, 2013) and slow down the entry and exit of water into and out of the endodermal protoplast, could have also contributed to the decrease in root Lp. However, contrary to a recent study on Arabidopsis (Barberon et al., 2016), we could not detect any increased formation of suberin lamella in response to low K. We also analysed root sections at locations between the root hair and mature lateral root region to see whether low K promoted the developmental onset of suberin formation as reported by Barberon et al. (2016), but could not see any differences between treatments (not shown). Therefore, it appears justified to conclude that the present reductions in root and cell Lp in response to low K involved primarily a downregulation of AQP activity at the plasma membrane.

Reductions in the gene expression level of PIP AQPs in response to a low supply of mineral nutrients have been reported previously (for review, see Clarkson et al., 2000; Wang et al., 2016), and this included low K supply in Arabidopsis (Maathuis et al., 2003; Armengaud et al., 2004). In contrast, Liu et al. (2006), studying rice, observed an increase in AQP gene expression level in response to low K. In the present study, we observed both increases and decreases in gene transcript abundance, with the latter prevailing. The isoform-specific response may explain some of the differences in results between the present and above-cited studies.

HvPIP2;5 showed the largest reduction in gene expression level of all PIP AQPs tested, and the reduction was in the same range (49 % decrease) as reductions in root and cell Lp, yet was statistically non-significant. It is possible that analyses of an even larger set of plants (here, n = 9–10 biological replicates) or random selection of plants showing less intrinsic plant-to-plant variation in expression may have led to a significant decrease in HvPIP2;5 expression. We observed, similar to the considerable batch-to-batch variation in the response of root-to-shoot ratio to low K, a considerable variation in AQP gene expression response between plant batches – there was no obvious relation between the two, e.g. individual plants showing a particularly large decrease in AQP gene expression did not necessarily show at the same time a particularly large increase in the root-to-shoot ratio in response to low K, an observation (variation in AQP gene expression) previously noted also for nitrate-deprived Lotus japonicus plants (Clarkson et al., 2000). We do not have any explanation for this variable response, but one possible explanation could be that some plant-intrinsic factor (such as ratio of hormones) that differs to some extent randomly between individual plants causes a switch between a gene expression response and a response operating at the level of protein.

It has been shown repeatedly that the abundances of transcript and protein of a particular PIP isoform do not have to change in concert in response to stress, and that additional means of regulation of PIP activity in the plasma membrane, such as phosphorylation and relocalization, have to be considered (Boursiac et al., 2005, 2008; Lee and Zwiazek, 2015; Wudick et al., 2015; for review see Maurel et al., 2015). Muries et al. (2011; and authors cited therein) concluded from studies of salt-stressed broccoli (Brassica napus) plants that inverse changes in transcript and protein levels of PIPs may point to a negative feedback of accumulating protein on the transcription of gene. Aroca et al. (2005) concluded for maize (Zea mays) that the hydraulic response to chilling involved an AQP-mediated mechanism, but was not restricted to such a mechanism. The authors suggested that oxidative stress and associated membrane damage should also be considered. Potassium starvation has been shown to cause oxidative stress in plants (Hernandez et al., 2012), which in turn could affect the gating (Ye and Steudle, 2006) and plasma membrane localization (Wudick et al., 2015) of PIPs through altering the cellular level of hydrogen peroxide. This could provide an additional, or alternative, explanation of the observed reduction in root and cell Lp in response to low K in the absence of significant decreases in PIP gene transcript levels.

HvPIP2;5 shows high water channel activity when tested in heterologous expression systems (Besse et al., 2011) and accounts for 64 % of the combined gene expression of all PIP2s in seminal roots of barley (Knipfer et al., 2011). Higher gene expression of HvPIP2;5 in adventitious compared with seminal roots in barley has been shown to coincide with a 3- to 4-fold higher Lp in the former type of root [Knipfer et al., 2011; adventitious roots were little developed at the plant developmental state studied here (Suku et al., 2014)]. In addition, we observed in a recent study on barley plants exposed to high Zn the largest and most significant decrease in gene expression of all PIPs tested for HvPIP2;5, parallel to a significant decrease in root and cell Lp (Gitto and Fricke, 2018). The rice homologue of HvPIP2;5, OsPIP2;5, has been proposed in several studies as a key PIP AQP that regulates changes in root Lp in response to stress or shoot removal (Sakurai-Ishikawa et al., 2011; Ahamed et al., 2012; Meng et al., 2016). Based on the above data, we consider HvPIP2;5 to be a prime candidate for future studies into the molecular regulation of water balance of barley plants exposed to low K.

Regulation of plant water balance at low K

The surface area ratio of root to shoot increased on average by 59 % in response to low K (Fig. 1). This is a typical response of plants to the limited supply of mineral nutrients such as N and P, with a more varied response reported for K (e.g. Andrews et al., 1999; Clarkson et al., 2000; Koevoets et al., 2016). We observed some considerable variation in the response of root-to-shoot ratio (of either surface area or FW) to the low-K treatment between plant batches, with some batches showing very little change in the ratio yet others showing large increases; note that the rather small error bars in Fig. 1, which shows average root-to-shoot ratios, represent standard errors for a large sample size (n = 62). We do not know the cause of this variation, as plants were grown under identical conditions and from the same batch of seeds. Also, plants of all batches showed consistent responses to low K, such as a decrease in transpiration rate and shoot surface area. Therefore, we decided to pool data from all batches.

The average increase in root-to-shoot ratio was due as much to a 23 % reduction in shoot surface area as to a 17 % increase in root surface area. At the same time, shoot water loss per unit shoot surface area, being indicative of stomatal conductance, decreased in response to low K, by 21 %. Decreases in stomatal conductance in response to the limited supply of mineral nutrients, including K⁺, have been reported previously (N, P; e.g. Carvajal et al., 1996; K, Rhee et al., 2011). The decrease in stomatal conductance, being due to either a reduced number or opening of stomatal pores, was the opposite of what one would have predicted if regulation of shoot water loss was a means to counter the potential imbalance between the rate of root water uptake (too high) and shoot water loss (too low) under conditions of low K. Rather, the rate of water uptake per unit root surface area decreased by 45 % in response to low K; this rate equals the numeric product of root hydraulic conductivity (Lp) and the biophysical driving force for radial root water uptake. By regulating or having changed either of the two (Lp, driving force), plants grown on low K adjusted root water uptake to shoot water loss. One could argue that the decrease in stomatal conductance may have rendered shoot water potential, and through this root xylem water potential, less negative, thereby decreasing the driving force. However, shoot water potential in low-K plants was almost identical to, and even slightly more negative than, shoot water potential in control plants. However, Zwieniecki et al. (2001) concluded that increasing concentrations of KCl in xylem reduced the axial hydraulic resistance of xylem through a hydrogel mechanism (see also Van Doorn et al., 2011). In low-K plants, xylem K concentrations will have been lowered – as also observed here for exudate of excised root systems, where K concentration decreased almost 10-fold (Supplementary Data Fig. S1) – and any associated increase in axial hydraulic resistance will have rendered the water potential in the root xylem less negative than suggested by the shoot water potential. We do not know how large this effect was, yet it would have reduced the driving force for radial root water uptake and contributed to a decreased rate of root water uptake in low-K plants.

Exudation Lp, measured on excised root systems, and cortex cell Lp, measured on intact, transpiring plants, were reduced by 42 and 56 %, respectively, in response to low K. Together with the above considerations, this would suggest that plants exposed to low K reduced root water uptake primarily through a reduction in root Lp. Similarly, if we take shoot water potential (control, −0.59 MPa; low K, −0.64 MPa; for comparison, root medium water potential −0.056 MPa) as a close approximation of xylem water potential, despite the uncertainty of the effect of any hydrogel mechanism (Zwieniecki et al., 2001) the transpiration and root surface area data shown in Fig. 1 calculate to a root Lp in intact, transpiring plants of 3.59 × 10⁻⁸ m s⁻¹ MPa⁻¹ for control and 1.68 × 10⁻⁸ m s⁻¹ MPa⁻¹ for low-K plants – a reduction by 47 %. There exist several studies in which root Lp was studied in response to low or no supply of a particular mineral nutrient (for reviews, see Clarkson et al., 2000; Aroca et al., 2012; Wang et al., 2016). Most of these studies have been concerned with N and P and less with K. The majority of studies show a decrease in root Lp. Rhee et al. (2011) analysed root Lp also at cell level, using the cell pressure probe, and observed a small (7 %) decrease for fig-leaf gourd plants grown under low supply of N, P and K. Aquaporin inhibition experiments suggested that the decrease in root Lp at low N and low P, but not at low K, was due to a downregulation of AQP activity (Rhee et al., 2011). In contrast, the present data show that a large and significant reduction in root Lp in response to low K is accompanied by a large and significant reduction in root cortex cell Lp. The cells analysed were located in the root hair zone along the main axis of seminal roots, a zone that contributes significantly to the overall water uptake of barley roots at the plant developmental stage analysed (Sanderson, 1983; Knipfer et al., 2011). Triboulot et al. (1997) observed for maritime pine (Pinus pinaster) that K⁺ deficiency reduced cortex cell turgor and osmotic pressure in the mature zone of tap roots by 0.05 MPa each, though non-significantly. In the present study on barley, the decreases in cell turgor and osmotic pressure were larger (0.13 and 0.21 MPa, respectively) and significant. The differences in cellular response to low K between the present study on barley and the studies on fig-leaf gourd (Rhee et al., 2011) and maritime pine (Triboulot et al., 1997) most likely reflect species- and root-type specific responses.

Conclusions

Barley plants grown with a low supply of K adjust water flow to an increase in root-to-shoot ratio mainly through root-associated sizes (Fig. 6). This includes a decrease in root Lp, which is also observed at cell level, and involves AQP function. HvPIP2;5 is a prime candidate to facilitate the decrease in Lp. Root anatomical changes in the endodermis, in particular the formation of Casparian bands and suberin lamella, do not contribute to the Lp response. A reduction in xylem K may lead, through a hydrogel mechanism (Zwieniecki et al., 2001), to an increased axial hydraulic resistance and less negative water potential in root xylem, which in turn reduces the water potential gradient, which drives water movement between root xylem and medium.

Fig. 6.

Schematic of changes at root and shoot level that affect uptake by and flow of water through barley plants grown with low K. The upper panel shows the scenario for control plants, and values are set here to 100 %. The lower panel shows the scenario for low-K plants. Exposure to low K reduces the rate of transpirational water loss per unit shoot surface area (H₂O/S) to 79 %, and shoot surface area (S) to 77 % of the control value; in contrast, root surface area (R) and the root-to-shoot surface area ratio (R:S) increase to 117 and 159 %, respectively, of the control value. These changes could potentially lead to a mismatch between the amount of water taken up through the root system (too much) and the amount of water lost through the shoot system (too little) in plants exposed to low K. One reason why this does not happen is that the rate of water uptake per unit root surface area (H₂O/R) decreases to 55 % of the control value in low-K plants. This is accompanied by, and is probably due to, a reduction in radial root and cortex cell hydraulic conductivity (Lp) to 58 % (root) and 44 % (cell) of the value in control plants. Reduced concentrations of K⁺ in the xylem of low-K plants may lead, through a hydrogel mechanism and increased axial hydraulic resistance of xylem, to a less negative (larger) water potential (ψ) in root xylem. This, in turn, reduces the difference in ψ between root medium and xylem, which drives radial water uptake, and may contribute to the reduced rate of water uptake per unit root surface area in low-K plants. Symbol size represents qualitative changes (increase, decrease) of the size considered. Xyl, xylem; End, endodermis, with Casparian bands (yellow) and suberin lamella (orange), which do not change in intensity in response to low K; CC with AQPs, cortex cells with aquaporins (filled red circles); the blue arrows represent water flow.

SUPPLEMENTARY DATA

Supplementary data are available online at https://academic.oup.com/aob and consist of the following. Figure S1: K and Ca concentrations in plant tissues and exudate