2. K ⁺ homeostasis at the whole plant level

Nutrient allocation within the plant is mediated by the two major vascular systems; the xylem provides a pathway for water and nutrient transport from the roots to the shoot, and the main role of phloem is in translocating sugars from photosynthetically active source tissues, mainly the leaves, to the roots and other below-ground organs as well as to seeds and fruits (Rogiers et al., Reference Rogiers, Coetzee, Walker, Deloire and Tyerman2017; van Bel & Hafke, Reference van Bel and Hafke2005). The xylem is a microfluidic system formed by dead, interconnected tubes, whereas the phloem consists of living cells with flow occurring in the sieve elements. Via xylem and phloem, a circular transport of K⁺ is established (Figure 1).

Figure 1.

A schematic diagram showing K⁺ circulation via xylem and phloem in higher plants; suc., sucrose. For more details, see the text.

Long-distance K⁺ transport, J_K+.i, in the compartment i (either xylem or phloem) can be simply described by the following equation:

(1) $$\begin{align}{J}_{K+,i}={v}_i\ast {A}_i\ast {c}_{K+,i}\end{align}$$

with v_i, A_i and c_K+,i being the flow velocity, the cross-section of all xylem or phloem elements displaying flow and the xylem/phloem K⁺ concentration, respectively. An elegant way of assessing ${v}_i$ and A_i non-invasively in intact plants is by the magnetic resonance imaging (MRI) technique. Using this technique, Windt et al. (Reference Windt, Vergeldt, De Jager and Van As2006) monitored xylem and phloem flow in poplar, tomato, tobacco and Ricinus communis. Phloem flow velocity ranged from 0.25 to 0.4 mm/s in all species during daytime; this corresponded to a volume flow rate (v_i^* A_i) of 0.1 to 0.2 mm³/s in the latter three species and somewhat higher values (~1 mm³/s) in poplar. The xylem flow velocity was higher by a factor of at least 10, and the volume flow rate ranged from 1 mm³/s in tomato to 17 mm³/s in poplar. Similar volume flow rates in the xylem and phloem during nighttime indicated the cycling of water between both compartments. For intact transpiring plants, the xylem K⁺ concentration is directly accessible with the multifunctional xylem probe (Wegner & Zimmermann, Reference Wegner and Zimmermann2002), combining a pressure probe with a K⁺ selective microelectrode. The location of the probe tip in a vessel is verified by recording pressure values below atmospheric or even vacuum level. In transpiring maize seedlings, xylem K⁺ concentrations of about 1 mM were measured, varying slightly, among other things, with K⁺ in the media. Root-to-shoot K⁺ export can also be quantified by combining xylem K⁺ recordings with gravimetrical measurements of water uptake (Wegner & Zimmermann, Reference Wegner and Zimmermann2009). Measuring phloem K⁺ is also a technical challenge. By analysing the exudate from stylets of phloem-feeding aphids, Lohaus et al. (Reference Lohaus, Hussmann, Pennewiss, Schneider, Zhu and Sattelmacher2000) obtained values of about 50 mM K⁺ for maize leaves, indicating that the phloem sap is ~50 times more concentrated than the xylem sap with respect to K⁺. If the xylem volume flow was exceeding that in the phloem by a factor of 50, this would imply a 100% K⁺ circulation between compartments. Unfortunately, we lack MRI data for maize, but this factor was 70 in tomatoes and lower in other species. Apparently, a very large part of xylem K⁺ ends up in the phloem and returns to the root.

Current estimates of the extent to which K⁺ is circulating are still based on less sophisticated techniques which rely on isolating xylem and phloem sap with invasive methods and measuring leaf and root K⁺ content (Jeschke & Pate, Reference Jeschke and Pate1991; for possible technical limitations see Wegner, Reference Wegner2015). Values vary considerably depending on experimental conditions. For maize, Engels and Kirkby (Reference Engels and Kirkby2001), applying different temperature regimes to manipulate root and shoot growth, reported that between 34 and 93% of xylem K⁺ is recirculated via the phloem. Jeschke and Pate (Reference Jeschke and Pate1991) found values between 58.9 and 103 % for Lupin, Ricinus communis and barley at mild salinity. Touraine et al. (Reference Touraine, Grignon and Grignon1988) investigated nitrate-fed soybean and observed that 68% of K⁺ moving upward in the xylem returned back to the root via the phloem. In plants reducing N in the shoot, nitrate is the main counterion for K⁺ in the xylem sap, whereas in the phloem it is frequently malate (so-called Ben Zioni-Lips-model; Ben Zioni et al., Reference Ben Zioni, Vaadia and Lips1971). The extent to which K⁺ is recirculated depends on various factors including K⁺ shoot demand and the K⁺ nutritional status of the plant, suggesting that it serves as a shoot-to-root signal particularly under conditions of K⁺ deficiency.

The most compelling evidence in favour of this hypothesis comes from split-root experiments (Drew & Saker, Reference Drew and Saker1984). K⁺ export from a single seminal barley root supplied with full nutrient solution was monitored using ⁴²K⁺ or ⁸⁶Rb⁺ as a tracer for K⁺, while the rest of the root system that was placed in a separate container was either fed with the same solution or deprived of K⁺. The latter treatment leading to K⁺ deficiency in the shoot strongly stimulated K⁺ export from the single, well-fed root and significantly reduced phloem K⁺ concentration in that root (Drew et al., Reference Drew, Webb and Saker1990) compared to a situation of even K⁺ supply to all roots. Evidence for K⁺ re-circulation via the phloem also comes from blocking phloem transport, for example, in Arabidopsis mutants not expressing the NaKR1 (sodium potassium root defective 1) protein, leading to a K⁺ accumulation in leaves (Tian et al., Reference Tian, Baxter, Lahner, Reinders, Salt and Ward2010).

In order to understand how K⁺ allocation in the plant is organized, potential sites of control need to be identified. These are (i) xylem loading, that is, K⁺ release from stellar cells into adjacent dead xylem vessels in the root, (ii) K⁺ unloading from the xylem in leaf tissues and (iii) uptake of K⁺ by sieve tubes in assimilating leaf tissues. All three sites are associated with membrane transport steps that have been studied in some depth.

3. K ⁺  homeostasis at the cellular level

Modelling approaches have been proven already as rather helpful in filling gaps that are hard to address in wet-lab experiments. The above-mentioned concept of the K⁺ battery in the phloem, for instance, was initially developed in computational simulations. Based on the experience from these dry-laboratory experiments pointed experiments with transgenic plants under energy-limiting conditions could be designed, in that case with low ambient O₂ to reduce ATP formation by respiration. The successful wet-laboratory experiment was thus the proof of concept of the model (Gajdanowicz et al., Reference Gajdanowicz, Michard, Sandmann, Rocha, Corrêa, Ramírez-Aguilar, Gomez-Porras, González, Thibaud, van Dongen and Dreyer2011; Sandmann et al., Reference Sandmann, Skłodowski, Gajdanowicz, Michard, Rocha, Gomez-Porras, González, Corrêa, Ramírez-Aguilar, Cuin, van Dongen, Thibaud and Dreyer2011). Recent advances in modelling have also provided new insights into K⁺ homeostasis at the cellular level. The modelling of cellular steady-state conditions, initially thought to be ‘trivial’, revealed that an adjustable steady-state of the transmembrane K gradient can only be achieved if at least two differently energized K⁺ transporter types are present in the membrane (Dreyer, Reference Dreyer2021). If this condition is met, the transmembrane K⁺ gradient can be adjusted by regulating the transporter activities. However, if it is not fulfilled, the regulation of transporter activity has no influence on the steady-state gradient (Dreyer, Reference Dreyer2021; Dreyer et al., Reference Dreyer, Hernández-Rojas, Bolua-Hernández, Tapia-Castillo, Astola-Mariscal, Díaz-Pico, Mérida-Quesada, Vergara-Valladares, Arrey-Salas and Rubio-Meléndez2024). Remarkably, the gain of control over the transmembrane homeostatic condition has a metabolic cost. The steady state is inevitably accompanied by transmembrane H⁺ and K⁺ cycles driven by ATP-consuming proton pumping (Figure 2). Such nutrient cycling was reported experimentally, for instance, under salt stress (Munns et al., Reference Munns, Day, Fricke, Watt, Arsova, Barkla, Bose, Byrt, Chen, Foster, Gilliham, Henderson, Jenkins, Kronzucker, Miklavcic, Plett, Roy, Shabala, Shelden and Tyerman2020; Shabala et al., Reference Shabala, Chen, Chen and Pottosin2020) and considered “futile” cycles (Britto & Kronzucker, Reference Britto and Kronzucker2006). The model analysis indicated, however, that these cycles are not as futile as they initially appear, but an essential part of the homeostatic process (Dreyer, Reference Dreyer2021).

Figure 2.

Homeostatic (steady state) condition of a transmembrane K⁺ gradient. In this example, the membrane voltage in a steady state is V _ss = −130 mV. At this voltage, there is an efflux of H⁺ released by the proton pump. The charge transport is compensated by a K⁺ influx of the same magnitude via the K⁺ channel (KC). Without further compensation, the system would not be stable in steady state but mediate an H⁺/K⁺ antiport, which would at least change the transmembrane potassium gradient. E _K would become increasingly negative. The system would only be stable if the red and green dots overlap with the light blue dot at −200 mV. Stability at V _ss = −130 mV can be obtained with an additional H⁺/K⁺ antiporter. This transporter releases each accumulated K⁺ ion and in the same process retrieves the released H⁺ in an electroneutral manner. The system is in steady state, and V _ss or E _K do not change. However, there are ATP-consuming H⁺ and K⁺ cycles.

If the system is brought out of steady state by changing the potassium concentration on one side of the membrane, the transporter network (called homeostat) mediates a largely electroneutral H⁺/K⁺ antiport to restore the altered gradient to its original value. Such an H⁺/K⁺ exchange also takes place if there are no H⁺/K⁺ antiporters in the membrane via symporters and channels (Contador-Álvarez et al., Reference Contador-Álvarez, Rojas-Rocco, Rodríguez-Gómez, Rubio-Meléndez, Riedelsberger, Michard and Dreyer2025). The exchange of H⁺ for K⁺ upon a change in the K⁺ gradient has been known since long (Blatt & Slayman, Reference Blatt and Slayman1987; Conway & O’Malley, Reference Conway and O'Malley1944; Maathuis & Sanders, Reference Maathuis and Sanders1994; Rodriguez-Navarro et al., Reference Rodriguez-Navarro, Blatt and Slayman1986) and was also recently observed in guard cells and mesophyll cells from Nicotiana benthamiana or Nicotiana tabacum (Li et al., Reference Li, Grauschopf, Hedrich, Dreyer and Konrad2024). The homeostat concept now provides a detailed mechanistic explanation of these observations. Interestingly, H⁺/K⁺ antiport occurred in a coordinated manner across both the plasma membrane and the vacuolar membrane supporting further the model prediction that there is an electrochemical coupling to sequential membranes (Dreyer et al., Reference Dreyer, Vergara-Valladares, Mérida-Quesada, Rubio-Meléndez, Hernández-Rojas, Riedelsberger, Astola-Mariscal, Heitmüller, Yanez-Chávez, Arrey-Salas, San Martín-Davison, Navarro-Retamal and Michard2023). Remarkably, besides organizing transmembrane K⁺ transport the homeostat also generates a cytosolic pH signal that additionally depends on various cellular parameters (Contador-Álvarez et al., Reference Contador-Álvarez, Rojas-Rocco, Rodríguez-Gómez, Rubio-Meléndez, Riedelsberger, Michard and Dreyer2025). Thus, in addition to its role as an information carrier at the whole plant level, K⁺ can also transmit messages at the cellular level.

The homeostat concept, which considers transporter networks instead of individual transporters, has so far explained well several observations, especially on K⁺ homeostasis. However, it also provided new insights into the homeostasis of other ions and even hormones (Dreyer et al., Reference Dreyer, Li, Riedelsberger, Hedrich, Konrad and Michard2022; Dreyer et al., Reference Dreyer, Hernández-Rojas, Bolua-Hernández, Tapia-Castillo, Astola-Mariscal, Díaz-Pico, Mérida-Quesada, Vergara-Valladares, Arrey-Salas and Rubio-Meléndez2024; Geisler & Dreyer, Reference Geisler and Dreyer2024; Li et al., Reference Li, Grauschopf, Hedrich, Dreyer and Konrad2024). As just mentioned before, homeostats appear also to constitute important parts of signalling processes in plants (Contador-Álvarez et al., Reference Contador-Álvarez, Rojas-Rocco, Rodríguez-Gómez, Rubio-Meléndez, Riedelsberger, Michard and Dreyer2025). Furthermore, the homeostat concept applies not only to the plasma membrane but also to endomembrane (Li et al., Reference Li, Grauschopf, Hedrich, Dreyer and Konrad2024). It could therefore contribute in the future to a better understanding of the role of K⁺ at the subcellular level.

4. From the cellular to the subcellular level: stress-induced modulation of cytosolic K⁺ as a signal in plant adaptive responses to hostile environment

In recent years, a significant bulk of evidence has been accumulated demonstrating that, in addition to a well-known role of K⁺ as a nutrient, changes in intracellular K⁺ concentrations may act as important developmental and adaptive signals (reviewed in Anschütz et al., Reference Anschütz, Becker and Shabala2014; Shabala, Reference Shabala2017; Wu et al., Reference Wu, Zhang, Giraldo and Shabala2018), affecting key processes such as cell fate determination and autophagy. Stress-induced K⁺ loss has been demonstrated in response to a broad range of abiotic and biotic stresses (reviewed in Shabala & Pottosin, Reference Shabala and Pottosin2014). These responses are often very fast (<1 min) and comparable with reported changes in cytosolic Ca²⁺ signals that are traditionally considered being upstream of stress signalling (Dong et al., Reference Dong, Wallrad, Almutairi and Kudla2022). It was also shown that such stress-induced K⁺ signalling is highly tissue-specific (Ahmed et al., Reference Ahmed, Shabala and Shabala2021; Chakraborty et al., Reference Chakraborty, Bose, Shabala and Shabala2016; Shabala et al., Reference Shabala, Zhang, Pottosin, Bose, Zhu, Fuglsang, Velarde-Buendia, Massart, Hill, Roessner, Bacic, Wu, Azzarello, Pandolfi, Zhou, Poschenrieder, Mancuso and Shabala2016a), and that kinetics of stress-induced K⁺ loss differ drastically between plant species, leading to a new concept of the stress-induced K⁺ flux ‘signatures’ (Rubio et al., Reference Rubio, Nieves-Cordones, Horie and Shabala2020), similar to those reported for cytosolic Ca²⁺. The physiological role of stress-induced modulation of cytosolic K⁺ and the mechanisms by which these K⁺ “signatures” are decoded are understood much less and require more discussion.

4.5. Decoding K ⁺  “signatures”

In the Arabidopsis genome, 26 CIPK and 10 CBL are present, providing a plethora of possibilities for decoding stress specificity. Such decoding machinery for K⁺ “signatures” remains mostly unexplored. Here, two possible mechanisms are put forward and discussed.

4.5.1. Decoding by regulation of purine metabolism

One of the enzymes with a strong K⁺ dependency is the adenosine kinase (ADK) that catalyses the phosphorylation of adenosine to 5^’-AMP using ATP as a source of phosphate. It was shown that binding of K⁺ to Asp³¹⁰ is essential for a transition to a catalytically productive structure and removing auto-inhibition of ADK (de Oliveira et al., Reference de Oliveira, Morales-Neto, Rocco, Sforca, Polo, Tonoli, Mercaldi, Cordeiro, Murakami and Franchini2018). Thus, a drop in [K⁺]_cyt will reduce the amount of cAMP (3’,5’-cyclic adenosine monophosphate). The latter, in turn, is a second messenger for a wide range of stress responses (Blanco et al., Reference Blanco, Fortunato, Viggiano and de Pinto2020), with demonstrated ability to regulate plant ionic homeostasis by controlling activity of various ion channels and transporters (Kaplan et al., Reference Kaplan, Sherman and Fromm2007). One of them is SOS1 H⁺/Na⁺ exchanger that is considered as a critical component of the salt exclusion mechanism (Zhao et al., Reference Zhao, Zhang, Song, Zhu and Shabala2020). It was shown recently that SOS1 contains a cyclic nucleotide-binding domain (CNBD; Zhang et al., Reference Zhang, Hu, Han, Chen, Lai and Wang2023) between residues 732 and 883. Moreover, mutations at G777 and G784 sides affected folding and stability of CNBD thus significantly decreasing transport activity of SOS1 (Wang et al., Reference Wang, Pan, Chen, Xie, Gao, He, Li, Dong, Jiang and Zhao2023). These findings suggest a possible causal link between salt-induced reduction in cytosolic K⁺ content, ADK-mediated cAMP production and SOS1 activation. Also, in response to stress, small amounts of ATP (nanomolar concentrations; Xiao et al., Reference Xiao, Zhou, Xie, Li, Su, He, Jiang, Zhu and Qu2024) are released to apoplastic space to be sensed by purinoreceptors that activate secondary signalling cascades (Kim et al., Reference Kim, Yanders and Stacey2023).

4.5.2. Decoding by regulating H⁺-ATPase activity

Stress-induced depolarization of the membrane potential (MP) is arguably one of the fastest events detected at the cellular level in response to a broad range of abiotic stresses (reviewed in Shabala et al., Reference Shabala, Zhang, Pottosin, Bose, Zhu, Fuglsang, Velarde-Buendia, Massart, Hill, Roessner, Bacic, Wu, Azzarello, Pandolfi, Zhou, Poschenrieder, Mancuso and Shabala2016). This is also true in response to K⁺ deficiency, where the plasma membrane (PM) is hyperpolarized within a few minutes (Maathuis & Sanders, Reference Maathuis and Sanders1993; Nieves-Cordones et al., Reference Nieves-Cordones, Miller, Aleman, Martinez and Rubio2008), while other signals such as ROS or ethylene start to operate after at least several hours (Behera et al., Reference Behera, Long, Schmitz-Thom, Wang, Zhang, Li, Steinhorst, Manishankar, Ren, Offenborn, Wu, Kudla and Wang2017). Equally fast are changes in MP caused by hypoxia or salinity (Bose et al., Reference Bose, Shabala, Pottosin, Zeng, Velarde-Buendía, Massart, Poschenrieder, Hariadi and Shabala2014, Reference Bose, Rodrigo-Moreno, Lai, Xie, Shen and Shabala2015) (Figure 3). As many K⁺-permeable channels are voltage gated (Demidchik & Maathuis, Reference Demidchik and Maathuis2007; Jegla et al., Reference Jegla, Busey and Assmann2018), this comes with important implications for intracellular ion homeostasis. Recently, Maierhofer et al. (Reference Maierhofer, Scherzer, Carpaneto, Müller, Pardo, Hänelt, Geiger and Hedrich2024) have shown that AtHAK5 operates as a proton-driven K⁺ transporter that is activated by membrane hyperpolarization, emphasising the essentiality of voltage gating in regulation of K⁺ homeostasis in plant cells.

Figure 3.

Difference in NaCl-induced K⁺ flux ‘signatures’ explain the difference in the strategy of adapting to salt stress between pea and barley. (a) Transient changes in MP of epidermal root cells in wheat and pea in response to acute NaCl treatment (based on Cuin et al., Reference Cuin, Betts, Chalmandrier and Shabala2008 and Bose et al., Reference Bose, Shabala, Pottosin, Zeng, Velarde-Buendía, Massart, Poschenrieder, Hariadi and Shabala2014, respectively). (b) Net K⁺ fluxes measured from plant roots by non-invasive ion flux measuring technique. (c) Predicted changes in cytosolic K⁺ concentration in pea and wheat root cytosol assuming no buffering from the vacuole. (d) Effect of vacuolar K⁺ buffering on cytosolic K⁺ kinetics in pea roots exposed to acute salinity treatment. Percentage values are the size of the vacuolar K⁺ available to be used to compensate cytosolic K⁺ loss.

In plants, the critical contributor to MP maintenance is the H⁺-ATPase pump. When the plasma membrane is depolarized under stress conditions (e.g., salinity stress), H⁺-ATPase activity is rapidly upregulated (Chen et al., Reference Chen, Pottosin, Cuin, Fuglsang, Tester, Jha, Zepeda-Jazo, Zhou, Palmgren, Newman and Shabala2007; Shabala et al., Reference Shabala, Zhang, Pottosin, Bose, Zhu, Fuglsang, Velarde-Buendia, Massart, Hill, Roessner, Bacic, Wu, Azzarello, Pandolfi, Zhou, Poschenrieder, Mancuso and Shabala2016), to restore MP. Such activation also provides an additional driving force for secondary active (H⁺-coupled) uptake of essential nutrients (i.e. PO₄⁻, NO₃⁻ or K⁺) and exclusion of toxins (such as Na⁺). K⁺ can bind to the proton pump at a site involving Asp⁶¹⁷ in the cytoplasmic phosphorylation domain (Buch-Pedersen et al., Reference Buch-Pedersen, Rudashevskaya, Berner, Venema and Palmgren2006). Such binding can induce dephosphorylation of the E₁P reaction cycle intermediate and, therefore, control the H⁺/ATP coupling ratio and uncouple ATP hydrolysis from transport. In other words, K⁺-induced dephosphorylation serves as a negative regulator of the transmembrane electrochemical gradient. We believe that this mechanism is ideally suited for decoding stress-induced K⁺ flux “signatures”, as illustrated in an example later.

Earlier, we have shown that pea and wheat plants have different strategies of responding to salt. Pea plants were able to quickly restore their MP within minutes upon acute salt treatment (Bose et al., Reference Bose, Shabala, Pottosin, Zeng, Velarde-Buendía, Massart, Poschenrieder, Hariadi and Shabala2014) (Figure 3a), while the MP in wheat root epidermal cells remained depolarized for much longer period (Cuin et al., Reference Cuin, Betts, Chalmandrier and Shabala2008) (Figure 3a). This difference was accompanied by a striking difference in the kinetics of the NaCl-induced K⁺ loss from their roots (Figure 3b). In pea, this K⁺ efflux was massive but transient while in wheat K⁺ efflux was an order of magnitude lower and remained more or less constant over the first 10 min. Can these differences in K⁺ flux “signatures” affect the operation of H⁺-ATPase pumps and explain the differences in MP kinetics?

To answer this question, some model calculations were done. Assuming the diameter of the root cell being 50 μm, then the cell volume is 6.54×10⁻¹⁴ m³ (65.4 pL) and the cell surface area is 7.85×10⁻⁹ m². Accepting a basal cytosolic K⁺ concentration ([K⁺]_cyt) as 100 mM, then the total amount of K⁺ in one cell is 6.54 pmol under control (unstressed conditions). In pea plants, the total K⁺ loss from root cells ranged from ∼ 520 nmol m⁻²s⁻¹ 1.5 min after stress onset to ∼ -40 nmol m⁻²s⁻¹ 20 min after (Figure 3b). The total K⁺ loss (in nmol) from the cell can be calculated as

(2) $$\begin{align}&\left[\mathrm{K}\;\mathrm{loss}\right]=\mathrm{Flux}\;\left(\mathrm{nmol}\;\mathrm{m}-2\mathrm{s}-1\right)\times \mathrm{surface}\kern0.17em \mathrm{area}\;\left(\mathrm{m}2\right)\nonumber\\&\qquad\qquad\times \mathrm{time}\;\left(\sec \right)\end{align}$$

Assuming that the efflux measured by non-invasive MIFE technique originates from the topmost (epidermal) cell layer; the cytosol representing 10% of the cell volume; no vacuolar buffering (e.g., no ability to replace lost cytosolic K⁺ from the vacuolar K⁺ pool); and feeding the flux data from Figure 3b into (Eq. 2), one can see that [K⁺]_cyt in pea will drop by about 2-fold within 2 min after stress onset and will be reduced to zero within 6 min (Figure 3c). At 5% vacuolar buffering capacity, a 50% decline in [K⁺]_cyt will occur by 6 min, and at 10% buffering – at 18 min (Figure 3d). Such rapid and drastic changes in [K⁺]_cyt may explain rapid activation of H⁺-pumping (Bose et al., Reference Bose, Shabala, Pottosin, Zeng, Velarde-Buendía, Massart, Poschenrieder, Hariadi and Shabala2014) and repolarization of the MP (Figure 3a) following the removal of K⁺-induced dephosphorylation of H⁺-ATPase. In wheat, however, NaCl-induced reduction in [K⁺]_cyt was much slower, and even without vacuolar buffering within 15 min [K⁺]_cyt will drop only by 15 mM (Figure 1c). As a result, the H⁺-ATPase remained dephosphorylated for much longer and failed to rapidly restore the MP.

As shown in the above-mentioned examples, the extent of H⁺-ATPase dephosphorylation will be critically dependent not only on the magnitude of stress-induced K⁺ loss but also on the plant’s ability to replenish [K⁺]_cyt loss from the vacuolar pool which may be highly specific for different cell types/tissues, as well as plant species. Thus, modulation of the coupling between H⁺-ATPase hydrolytic and transport activity by [K⁺]_cyt “signatures” may represent a decoding mechanism and explain the difference in species’ strategies of handling salt stress.

The above modelling has been done without considering the fact that cells are connected by plasmodesmata, so cortical cells may also provide some additional buffering. However, giving the lack of the quantitative data on how rapid K⁺ from the neighbouring cells may be transported to root epidermis, modelling of this process is nearly impossible. It should be also kept in mind that these (cortical) cells will be also exposed to NaCl stress and will undergo the same membrane depolarisation and, hence, K⁺ loss. Thus, such additional “buffering capacity” is likely to play a relatively minor role in determining cytosolic K⁺ levels in root epidermal cells.

5. Vacuolar K⁺-permeable channels sense vacuolar K⁺ content

The vacuole is the main intracellular K⁺ store in plant cells and, except extreme K⁺-starving conditions, the electrochemical K⁺ gradient across the tonoplast is cytosol-directed. Thus, K⁺ accumulation in the vacuole is mediated by an active K⁺/H⁺ antiport, whereas passive vacuolar K⁺ release is channel-mediated (Li et al., Reference Li, Grauschopf, Hedrich, Dreyer and Konrad2024). The vacuolar K⁺ concentration tends to be more variable as compared to cytosolic K⁺. During stomata closure there is a massive decrease of intracellular (presumably, mainly vacuolar) K⁺ content, from 450 to 95 mM in Commelina communis (Leigh, Reference Leigh1997). For several plant species, collectively, upon stomatal closure the cytosolic K⁺ concentration decreases from 150–247 mM to 55–93 mM, whereas the vacuolar one from 181–454 mM to 38–92 mM, (Jezek & Blatt, Reference Jezek and Blatt2017). In open stomata after the midday, K⁺ is substituted by 50% with sugars (Talbott & Zeiger, Reference Talbott and Zeiger1996). In salinized barley leaves K⁺ dropped (substituted by Na⁺) both in the vacuole and the cytosol; these changes were dramatic (about 5-fold) in epidermal cells, and less prominent in mesophyll cells, where cytosolic K⁺ tends to be maintained at the expense of the vacuolar K⁺ pool (Cuin et al., Reference Cuin, Miller, Laurie and Leigh2003), thus improving cytosolic K⁺/Na⁺ ratio, underlying a higher photosynthetic activity and salt tolerance (James et al., Reference James, Munns, Von Caemmerer, Trejo, Miller and Condon2006). The refilling of the cytosol by vacuolar K⁺ is most pronounced in the case of K⁺ deficiency: the vacuolar K⁺ activity can drop 10-fold, whereas the cytosolic K⁺ activity remains almost constant (Walker et al., Reference Walker, Leigh and Miller1996).

There are three types of ion channels, which mediate passive K⁺ fluxes across the tonoplast: (i) VK (K⁺-selective) channel, mainly encoded by TPK1, (ii) slow vacuolar (SV) channels; non-selective cation channels encoded by TPC1 and (iii) fast vacuolar (FV) channels (also non-selective; with unknown molecular identity). Only VK/TPK channels, but not SV/TPC1 or FV, can discriminate between K⁺ and Na⁺ (Brüggemann et al., Reference Brüggemann, Pottosin and Schönknecht1999a; Gobert et al., Reference Gobert, Isayenkov, Voelker, Czempinski and Maatthuis2007; Pottosin et al., Reference Pottosin, Martínez-Estévez, Dobrovinskaya and Muñiz2003; Pottosin & Dobrovinskaya, Reference Pottosin and Dobrovinskaya2014; Ward & Schroeder, Reference Ward and Schroeder1994). FV and SV channel activity and/or conductance is modulated by a variety of cytosolic and vacuolar factors, including Ca²⁺ (and Mg²⁺), voltage and polyamines. FV activity is suppressed by cytosolic Ca²⁺, Mg²⁺ and polyamines in the micromolar–submillimolar range (Brüggemann et al., Reference Brüggemann, Pottosin and Schönknecht1998, Reference Brüggemann, Pottosin and Schönknecht1999b; Dobrovinskaya et al., Reference Dobrovinskaya, Muñiz and Pottosin1999; Tikhonova et al., Reference Tikhonova, Pottosin, Dietz and Schönknecht1997). In the virtual absence of these cations at both tonoplast sides and symmetric K⁺ in a physiological range, it is activated at an increased voltage difference of either sign (Allen et al., Reference Allen, Amtmann and Sanders1998; Bonales-Alatorre et al., Reference Bonales-Alatorre, Shabala, Chen and Pottosin2013; Pottosin et al., Reference Pottosin, Martínez-Estévez, Dobrovinskaya and Muñiz2003; Tikhonova et al., Reference Tikhonova, Pottosin, Dietz and Schönknecht1997), with a minimum open probability at about resting tonoplast potential of −20 to −30 mV (Miller et al., Reference Miller, Cookson, Smith and Wells2001; Walker et al., Reference Walker, Leigh and Miller1996; Wang et al., Reference Wang, Dindas, Rienmuller, Krebs, Schumacher, Wu, Hedrich and Roelfsema2015). Vacuolar polycations at millimolar concentrations convert FV into an outward rectifier, suppressing the cytosol-directed K⁺ flux (Tikhonova et al., Reference Tikhonova, Pottosin, Dietz and Schönknecht1997). The SV/TPC1 channel is activated at cytosol-positive voltage and its opening requires high (ten micromolar) cytosolic Ca²⁺, whereas vacuolar Ca²⁺ (and Mg²⁺) locks the channel in the resting state, thus increasing the threshold for its voltage activation (reviewed in Pottosin & Dobrovinskaya, Reference Pottosin and Dobrovinskaya2022). VK/TPK channels are mainly expressed in guard cells and to a lesser extent in root and mesophyll, yielding 10 to 200 VK copies per vacuole (Pottosin et al., Reference Pottosin, Martínez-Estévez, Dobrovinskaya and Muñiz2003; Tang et al., Reference Tang, Zhao, Yang, Wang, Li, Kleist, Lemaux and Luan2020; Ward & Schroeder, Reference Ward and Schroeder1994), whereas a typical vacuole contains thousands of active copies of SV/TPC1 and FV channels. Yet, due to the fact that VK/TPK channels are voltage-independent, insensitive to Mg²⁺ and polyamines and only require elevated (>1 μM) cytosolic Ca²⁺, which can activate the channel directly or via the CBL-CIPK pathway (Allen et al., Reference Allen, Amtmann and Sanders1998; Pottosin et al., Reference Pottosin, Martínez-Estévez, Dobrovinskaya and Muñiz2003; Tang et al., Reference Tang, Zhao, Yang, Wang, Li, Kleist, Lemaux and Luan2020; Ward & Schroeder, Reference Ward and Schroeder1994), it is conceivable that they can dominate K⁺ flux from vacuole at moderate cytosolic Ca²⁺ elevations. Strict negative control of SV/TPC1 channels by physiologically relevant factors can question its direct contribution to tonoplast K⁺ and Ca²⁺ fluxes. However, it has been shown that in vivo SV/TPC1 activation by a depolarization is transmitted in some way to the VK/TPK activation, mediating vacuolar K⁺ release by the latter, as if the activity of these two tonoplast channels is functionally coupled (Jaslan et al., Reference Jaslan, Dreyer, Lu, O'Malley, Dindas, Marten and Hedrich2019). It is conceivable that such a transduction occurs in a direct mechanical way so that the movement of the voltage sensor of the SV/TPC1 channels favours the activation of associated VK/TPK ones; for such mechanism full activation (opening) of SV/TPC1 channels, which is only observed at rather high cytosolic Ca²⁺, is not necessary so that the channel protein may operate in a non-canonical ion flux-independent way (Pottosin & Dobrovinskaya, Reference Pottosin and Dobrovinskaya2022).

VK/TPK channels were proven to play key roles in vacuolar K⁺ release upon ABA-induced stomatal closure (Gobert et al., Reference Gobert, Isayenkov, Voelker, Czempinski and Maatthuis2007) and vacuolar K⁺ remobilization upon K⁺ starvation (Gu et al., Reference Gu, Lu, Ren and Lu2024; Tang et al., Reference Tang, Zhao, Yang, Wang, Li, Kleist, Lemaux and Luan2020). VK/TPK activity is postulated to be predominant over the activity of non-selective SV/TPC1 and FV channels under salt stress, not only for the means of refilling the cytosolic K⁺ pool but also for a proper Na⁺ compartmentalization, avoiding a futile energy-consuming vacuolar Na⁺ cycling (Shabala et al., Reference Shabala, Chen, Chen and Pottosin2020; Wu et al., Reference Wu, Zhang, Giraldo and Shabala2018).

Vacuolar K⁺ dynamics appears to be controlled by the vacuolar K⁺ content. Firstly, upon the increase of the K⁺ concentration in the growth medium, K⁺ cellular (principally, vacuolar) content tends to approach the upper limit of 100–250 mM when external K⁺ reaches 24–95 μM (species-dependent) and does not increase upon further increase of external K⁺ to 1 mM (Leigh, Reference Leigh2001). On the other hand, under stomatal closure, induced by low and optimal ABA concentrations, the release of the vacuolar K⁺ analogue ⁸⁶Rb, albeit substantially slower at low ABA dosage, ceased when the same loss of vacuolar tracer resulted as if “different numbers of channels are activated at high and low amounts of ABA, but the behaviour of individual ion channels is sensitive to ion content (volume)” (MacRobbie, Reference MacRobbie1995, Reference MacRobbie1998). MacRobbie opted for a volume/ stretch sensitive K⁺ channel and osmo-/stretch sensitivity was indeed demonstrated for TPK1 channels from diverse plant species (Maathuis, Reference Maathuis2011). In contrast to FV and SV/TPC1 channels, which are equally permeable for Rb⁺ and K⁺ (Brüggemann et al., Reference Brüggemann, Pottosin and Schönknecht1999a; Pottosin & Dobrovinskaya, Reference Pottosin and Dobrovinskaya2014), VK/TPK channels hardly conduct Rb⁺ (Gobert et al., Reference Gobert, Isayenkov, Voelker, Czempinski and Maatthuis2007; Ward & Schroeder, Reference Ward and Schroeder1994). Yet, a significant role of SV/TPC1 channels in ABA-induced stomata closure seems to be ruled out (Islam et al., Reference Islam, Munemasa, Hossain, Nakamura, Mori and Murata2010; Peiter et al., Reference Peiter, Maathuis, Mills, Knight, Pelloux, Hetherington and Sanders2005), but what about FV channels? For the sake of completeness, one needs to take also into account the coupling between the fluxes of K⁺ and small anions (Cl⁻), governing stomatal movements, not only for the same membrane, but also between plasma membrane and tonoplast. Indeed, plasma membrane and tonoplast communicate across the common space, cytosol. Thus, the mutation of the main anion pathway in the plasma membrane, formed by SLAC1 channel, predictably slows down the global transport across both membranes in series (Horaruang et al., Reference Horaruang, Hills and Blatt2020). Voltage dependence of the outwardly rectifying K⁺ channel GORK in the guard cell plasma membrane senses external K⁺, which changes manifold upon stomatal close-open transitions (Jezek & Blatt, Reference Jezek and Blatt2017). This K⁺ modulation ensures that at any external K⁺ GORK channels will mediate K⁺ efflux (Dreyer & Blatt, Reference Dreyer and Blatt2009). However, such a mechanism defines the direction of K⁺ flux but not the condition of its cessation. A more compatible K⁺-sensing mechanism, which is the voltage-independent control of K⁺ channels activity by cytosolic K⁺ was described for SKOR, another outward-rectifying K⁺ channel, with a high degree of homology to GORK (Liu et al., Reference Liu, Li and Luan2006). But to the best of our knowledge, yet such a mechanism has not been confirmed for GORK. Thus, next we consider K⁺ sensing mechanisms of the major tonoplast cation channels.

A focused study demonstrated that FV channels sense both vacuolar and cytosolic K⁺ changes (Pottosin & Martínez-Estévez, Reference Pottosin and Martínez-Estévez2003). An increase in the cytosolic K⁺ resulted in a shift of the whole voltage dependence in parallel with the E_K shift. In contrast, increase of vacuolar K⁺ only affected the voltage-dependent process at cytosol negative potentials, decreasing the respective voltage threshold and increasing channel’s open probability, hence potentially inducing vacuolar K⁺ release. Overall, this analysis demonstrated that at tonoplast attainable potentials the FV activity was almost independent of cytosolic and strongly potentiated by vacuolar K⁺ (Pottosin & Martínez-Estévez, Reference Pottosin and Martínez-Estévez2003). Meanwhile, SV/TPC1 was also sensitive to vacuolar K⁺: in the virtual absence of vacuolar divalent cations an increase of vacuolar K⁺ resulted in the increase of the threshold for voltage activation in parallel with the right-hand E_K shift (Ivashikina & Hedrich, Reference Ivashikina and Hedrich2005; Pottosin et al., Reference Pottosin, Martínez-Estévez, Dobrovinskaya and Muñiz2005). Substitution of K⁺ with NMDG⁺ demonstrated that this shift was caused by shielding of the negative membrane surface charge. Yet, the effect was beyond the totally unspecific cation binding, so that vacuolar Na⁺ caused a larger shift as compared to K⁺ (Ivashikina & Hedrich, Reference Ivashikina and Hedrich2005; Pottosin & Dobrovinskaya, Reference Pottosin and Dobrovinskaya2014). It was speculated then that an increase of the threshold for voltage activation by high vacuolar Na⁺ can be a mechanism for the prevention of a Na⁺ leak from the vacuole at salt stress (Ivashikina & Hedrich, Reference Ivashikina and Hedrich2005). Yet, in the presence of submillimolar Ca²⁺ at the luminal side the direction of the K⁺ (Na⁺)-induced shift of the voltage dependence is reversed, so that the channel tends to open at more cytosol negative potentials. Luminal Ca²⁺ is a principle negative regulator of the SV/TPC1 activity, and shielding the negative surface charge by K⁺, and, more strongly, by Na⁺, decreases the local Ca²⁺ concentration in the vicinity of the membrane surface/luminal Ca²⁺ binding site and alleviates the Ca²⁺ inhibitory effect (Pérez et al., Reference Pérez, Wherrett, Shabala, Muñiz, Dobrovinskaya and Pottosin2008; Pottosin et al., Reference Pottosin, Martínez-Estévez, Dobrovinskaya and Muñiz2005; Pottosin & Dobrovinskaya, Reference Pottosin and Dobrovinskaya2014). So that, at physiological vacuolar Ca²⁺, an increased content of Na⁺ in the lumen would promote vacuolar Na⁺ release via SV/TPC1 channels. The same appears to be true for FV channels (Bonales-Alatorre et al., Reference Bonales-Alatorre, Shabala, Chen and Pottosin2013). To summarize, at non-salinized conditions, when K⁺ is a dominant cellular cation, both FV and SV/TPC1 channels act as vacuolar K⁺ sensors, so that their open probability increases with the increase of vacuolar K⁺ and decreases when vacuolar K⁺ dropped. The latter can be suitable for a condition of extreme K⁺ deficiency, when the electrochemical K⁺ gradient is vacuole-directed (Walker et al., Reference Walker, Leigh and Miller1996), to diminish channel-mediated K⁺ re-absorption by the vacuole (Figure 4). Yet, this sensing mechanism is not selective for K⁺ vs Na⁺, so that at salt stress additional regulatory factors need to be invoked to suppress the non-selective channels´ activity and reduce the vacuolar Na⁺ leak (Pottosin et al., Reference Pottosin, Olivas-Aguirre, Dobrovinskaya, Zepeda-Jazo and Shabala2021; Shabala et al., Reference Shabala, Chen, Chen and Pottosin2020).

Figure 4.

Potassium permeable channels of FV and SV types act as safety valves, preventing K⁺ overload at excessive K⁺; vacuolar K⁺ filling is mediated by K⁺ (Na⁺)/ H⁺ antiporter of NHX type (left). Suppression of the activity of FV and SV channels by low vacuolar K⁺ prevents channel-mediated vacuolar K⁺ re-uptake and allows the maintenance of cytosolic K⁺ at the expense of the vacuolar K⁺ pool. Cytosol refilling by vacuolar K⁺ can be mediated by a hypothetical K⁺: H⁺ symporter (right). Note that in the K⁺ replete conditions a significant ‘futile’ H⁺ and K⁺ cycling results, whereas under K⁺ deficiency (upon the change in the direction of transtonoplast K⁺ gradient) this cycling is reduced. Above transport system acts as K⁺ sensor/effector at non-saline conditions but is unable per se to discriminate between K⁺ and Na⁺ under high salinity (see text).

6. Conclusion and outlook

Potassium is the most abundant and probably also the most versatile cation in plants. Besides its rather static role as an osmotic agent and charge balancer, there is more and more evidence that it also plays significant roles in intra-plant and intra-organelle communication. In addition, potassium gradients can also serve as an important supporting energy source, for example, for phloem loading. Although our knowledge of K in plants seems to be by far the most advanced of all relevant ions, there are still many open questions that need to be answered in future research. Status of current knowledge provides the basis for a comprehensive quantitative modelling approach including properties of membrane transport proteins, volume flow and regulatory features of K⁺. Recent computational modelling revealed the role of apparently futile K⁺ and H⁺ cycling for intracellular K⁺ homeostasis and steady state membrane potential maintenance. These conclusions received an experimental confirmation, but more wet-lab work is needed to support them. The loss of K⁺ is a common denominator of abiotic stress responses and there is an emerging view for its physiological importance as an energy switch from metabolism to defence needs. There is contrasting behaviour of different plant species and tissues in handling of intracellular K⁺ and membrane potential in a response to stress, which raises the question of the specificity of K⁺ sensing, decoding and downstream signalling. Some plausible K⁺ sensing and decoding mechanisms are addressed in this review. Several K⁺ and K⁺-conducting channels and transporters such as SKOR, HAK5 and vacuolar SV/TPC1 and FV may be plausible candidates to act as K⁺ sensors and effectors regulating the activity of target proteins. Therefore, the puzzle of K⁺ homeostasis and signalling in plants is gaining shape. Obviously, to complete this job, a systemic modelling approach with focused experimentation are required.