Nonselective cation channels (NSCCs) catalyse passive fluxes of cations through plant membranes. NSCCs do not, or only to a small extent, select between monovalent cations, and several are also permeable to divalent cations. Although a number of NSCC genes has been identified in plant genomes, a direct correlation between gene products and in vivo observed currents is still largely absent for most NSCCs. In this review, physiological functions and molecular properties of NSCCs are critically discussed. Recent studies have demonstrated that NSCCs are directly involved in a multitude of stress responses, growth and development, uptake of nutrients and calcium signalling. NSCCs can also function in the perception of external stimuli and as signal transducers for reactive oxygen species, pathogen elicitors, cyclic nucleotides, membrane stretch, amino acids and purines.
I. Introduction 388
II. Constitutive plasma membrane NSCCs 389
III. NSCCs activated by reactive oxygen species (ROS-NSCCs) 392
In plant plasma membranes, the predominant type of ion channel is K+-selective and voltage-dependent (see review by Véry & Sentenac, 2003). However, during many in planta studies, currents were observed in plasma membranes and vacuolar membranes, which lacked voltage dependence and often showed no or little selectivity amongst cations. Initially, these currents were often regarded as artefactual and merely constituting a ‘leak’ without much physiological relevance. However, it became clear that discrete single-channel events underlie these currents, and more recently researchers have come to recognize the importance of nonselective cation channels (NSCCs; see Table 1 for glossary of terms used in this review). NSCCs typically are permeable to a wide range of monovalent cations. For example, most NSCCs demonstrate K+/Na+ selectivity ratios between 0.3 and 3, whereas this ratio is well over 10 in outward rectifying K+-selective channels, and 100 or more in inward rectifying K+-selective channels (Véry & Sentenac, 2003). Similarly, many NSCCs conduct Ca2+ with Ca2+:K+ selectivity ratios between 0.5 and 2, whereas these ratios in animal Ca2+-selective channels range from 10 to over 50.
Table 1. Glossary
Calmodulin-binding domain. A site found on cyclic nucleotide-gated channels that affects binding of ligand and therefore channel activity
Cyclic nucleotide-binding domain. A site found on cyclic nucleotide-gated channels where the binding of ligands (cAMP or cGMP) occurs that leads to channel activation
Cyclic nucleotide-gated channel. A type of nonselective ion channel that is activated by binding of cyclic nucleotides (cAMP or cGMP)
Ion channel activity that can be recorded at all applied membrane voltages and does not require ligand for activation
Depolarization-activated, nonselective cation channel. A nonselective cation channel that shows more activity at positive membrane voltages compared with negative membrane voltages
Decrease in membrane voltage leading to less negative values
Compound released during pathogen attack such as cell wall fragments. Perception of elicitors starts a signalling cascade involving ion channels
‘Fast vacuolar’ channel. A nonselective cation channel characterized in plant tonoplasts, possibly involved in K+ homeostasis
Glutamate receptor-like genes. Gene family of putative glutamate receptors in Arabidopsis
Hyperpolarization-activated, nonselective cation channel. A nonselective cation channel that shows more activity at negative membrane voltages compared with positive membrane voltages
Hyperpolarization-activated, Ca2+ channel. Class of ion channel that has been characterized in planta as conducting Ca2+ influx. Further properties of this channel suggest HACaCs are a subclass of HA-NSCCs
Increase in membrane voltage leading to more negative values
K+ outward rectifier. Class of K+-selective, voltage-dependent channels that is activated only at positive membrane voltages, leading to strong outwardly rectifying current
Compound that binds to the protein. In the context of ion channels, many ligands are known that bind to the channel protein to instigate a transition from the closed to the open state
Mechanosensitive channel. Ion channel whose gating is modulated by physical forces on the membrane
Mechanosensitive-like channel genes. Gene family in Arabidopsis of putative mechanosensitive channels
Nonselective cation channel. Ion channels that conduct cations but not anions and show no or little discrimination between different cations
Phenomenon where transmembrane currents in one direction are different from those in the opposite direction
Reactive oxygen species. Compounds that contain oxygen in a reactive state and cause oxidation of other compounds. Physiologically, most important ROS include singlet oxygen, hydrogen peroxide, superoxide anion radicals and hydroxyl radicals. Note that hydrogen peroxide and superoxide anion radical are not particularly reactive but are nevertheless considered as ROS
‘Slow vacuolar’ channel. A nonselective cation channel characterized in plant tonoplasts, possibly involved in Ca2+ signalling
Voltage-independent, nonselective cation channel. A nonselective cation channel, whose activity is not modulated by the membrane voltage
There are several areas of plant research where interest in NSCCs is substantial. These include Ca2+ nutrition, Ca2+ signalling and salinity stress. Although Ca2+ currents can be recorded from plant membranes, these invariably seem to be mediated by ion channels that are not, or are only weakly, selective for Ca2+. Such Ca2+-permeable NSCCs play a number of physiological roles. In Arabidopsis guard cells, Ca2+-permeable NSCCs are involved in stomatal regulation. The same type of NSCC may contribute to relaying Ca2+ signals derived from plant pathogens and Ca2+ signalling across the vacuolar membrane (Johannes & Sanders, 1995; Peiter et al., 2005), whereas in root epidermis, NSCCs are thought to play a role in Ca2+ acquisition (Demidchik et al., 2002a).
Another area in which NSCCs appear to play a critical role is plant salinity stress. A major question concerns the mechanism of Na+ entry into plant roots and, similar to the situation for Ca2+, no Na+-selective ion channels have been found in plants. Based on the similarity of the Ca2+-dependent block of NSCC current and of Na+ influx in intact tissue, NSCCs were proposed to form a major pathway for Na+ entry into plants (Tyerman et al., 1997) and many later studies appear to confirm this notion (Maathuis & Sanders, 2001; Demidchik & Tester, 2002; Essah et al., 2003).
Here, we will update our current knowledge regarding plant NSCCs. Although NSCCs have been found in several types of plant membrane, this review will mainly focus on plasma mambrane and vacuolar membrane NSCCs and will discuss data that have recently appeared, in particular regarding the physiological roles of NSCCs in plants.
II. Constitutive plasma membrane NSCCs
Gating of many ion channels is strictly dependent on membrane voltage, and unless the permissible voltage range is applied, no channel activity will be observed. Similarly, ligand-gated channels typically remain inactive without the presence of agonist. However, several types of ion channel have been observed that are active, irrespective of the membrane voltages and without the need for addition of any ligand. Such channels therefore appear to be permanently (constitutively) active at all applied membrane voltages. Nevertheless, the extent of channel activity may still be modulated by many cellular factors and by membrane voltage (Fig. 1). Constitutive NSCCs can therefore be subdivided according to their voltage dependence into voltage-insensitive, depolarization-activated and hyperpolarization-activated NSCCs.
Constitutive NSCCs were among the first cation channels identified in plants. In addition, they were initially considered as constitutive plasma membrane leak conductances (Yurin et al., 1991) or as classes of K+ channel with low selectivity for K+ (Tester, 1990; White & Tester, 1992).
Activation kinetics and voltage dependence of some DA-NSCCs are reminiscent of those observed for outwardly rectifying K+-selective channels (K+ outward rectifiers, KORs). However, in contrast to KORs, DA-NSCCs are insensitive to blocking by extracellular or intracellular Na+ (Shabala et al., 2006). DA-NSCCs can be effectively blocked by extracellular Ca2+ (Shabala et al., 2006), but this has also been observed for many other types of cation channels. Other compounds that may affect DA-NSCCs are TEA+, nifedipine, diltiazem and verapamil (reviewed by Demidchik et al., 2002b; Shabala et al., 2006).
The physiological functions of DA-NSCCs appear related to their ability to catalyse influx and efflux of monovalent cations and influx of divalent cations. The electrochemical gradient for most divalent cations is inwardly directed. This means that DA-NSCCs could form a conduit for the influx of divalent cations. Some indications that Ca2+ influx occurs through DA-NSCCs were found in maize root stelar cells (Roberts & Tester, 1997). In addition, Piñeros & Kochian (2003) demonstrated that passive Zn2+ influx in the Zn-hyperaccumulating Thlaspi caerulescens can occur through DA-NSCC-type conductance.
Involvement of DA-NSCCs in monovalent cation fluxes may be widespread. For example, several studies have shown that DA-NSCCs contribute to K+ loading into the xylem (de Boer & Wegner, 1997) and K+ redistribution in bean seeds (Zhang et al., 2002, 2004). The relatively high permeability for Na+ of many DA-NSCCs suggests they are also important contributors to the uptake and translocation of Na+, and thus their function impacts on salt tolerance. During salt stress, plant cells can rapidly accumulate Na+ to a very high concentration that interferes, in particular, with essential roles of K+ in the cytosol. Thus, maintaining high cytosolic K+/Na+ ratios is an important component of salinity tolerance (Maathuis & Amtmann, 1999; Shabala et al., 2006), and minimizing salt-induced K+ efflux and Na+ influx would contribute to increased tolerance.
Na+ entry constitutes an efficient and relatively cheap form to restore the osmotic balance and, as such, plays an important role in plant salinity tolerance. However, excessive Na+ influx causes Na+ toxicity, and relatively high DA-NSCC K+/Na+ selectivities may therefore provide adaptive advantages in this respect. A comparative study on Arabidopsis thaliana and its halophytic relative Thellungiella halophila showed an approximately fourfold higher PK/PNa of DA-NSCC-like channels in roots of Thellungiella compared with Arabidopsis (Volkov et al., 2004; Volkov & Amtmann, 2006; Wang et al., 2006). The PK/PNa of other NSCCs that are involved in Na+ influx was also found to be lower in Thellungiella halophila.
2. Voltage-independent NSCCs (VI-NSCCs)
The open probability of VI-NSCCs is not modulated by membrane voltage, suggesting these proteins lack any domains for sensing changes in membrane voltage. Thus, most VI-NSCCs equally conduct outward and inward current. VI-NSCCs are not inhibited by organic antagonists of Ca2+ channels such as nifedipine and verapamil, or by conventional blockers of K+ channels (TEA+, TTX, Cs+, Li+ and Na+). However, they are sensitive to Gd3+ and La3+. VI-NSCCs can be subdivided into two groups on the basis of blockage by cations such as Ca2+, Ba2+, Mg2+ and Zn2+. One class is partially blocked, whereas another class is both insensitive and permeable to these divalent cations. Inhibition by quinine is another aspect that pertains to some VI-NSCCs (Demidchik & Tester, 2002), but not others (White & Lemtiri-Chlieh, 1995).
Although conclusive evidence is still lacking, it is widely considered that the main physiological function of VI-NSCCs is to catalyse uptake of cations. This may include monovalent cations like Na+ and NH4+ that do not permeate selective K+ channels, auxiliary uptake capacity for K+ when K+-selective channels are inhibited or not expressed, and the uptake of divalent cations such as Ca2+ and Mg2+ for which no selective pathways have been found in plant membranes at resting membrane potentials. This function can be highly beneficial where the acquisition of essential nutrients such as Ca2+, Mg2+, K+, NH4+, Mn2+ and Zn2+ is concerned, but also creates the potential for influx of harmful ions such as Na+, Cs+, Pb2+, Hg+ and Cd2+. Examples of both processes are discussed later.
All VI-NSCCs studied so far have been shown to be permeable to K+ (Demidchik et al., 2002b) and with the exception of VI-NSCCs in pea leaf epidermis (Elzenga & van Volkenburgh, 1994) and maize root cortex (Roberts & Tester, 1997), K+ is typically the most permeant ion. In most conditions, K+-selective inward rectifiers (KIRs) dominate passive K+ influx into root cells (Maathuis & Sanders, 1993), a contention that was given extra credence after Hirsch et al. (1998) showed that a loss of function in the main root KIR, AKT1, significantly reduced K+ uptake. However, a large component of the low-affinity K+ uptake remains in akt1-1 mutants, and although it is not clear what the exact mechanism is for this influx, VI-NSCCs in conjunction with KORs (Maathuis & Sanders, 1997) could mediate a significant part of this. Thus, VI-NSCCs may provide extra K+ uptake capacity, especially in conditions where the main low-affinity uptake pathway is inhibited.
Plants need large amounts of Ca2+, and understanding the mechanism of Ca2+ uptake from the soil is therefore of great value (Marschner, 1995; Welch, 1995). No Ca2+-selective channel has been found in plant membranes (Bothwell & Ng, 2005) and plant VI-NSCCs may be important in plant Ca2+ nutrition as a pathway for Ca2+ uptake, a notion that appears to contradict the blocking effect of Ca2+ on NSCCs (Fig. 2). The Ca2+ block derives from its high-affinity binding to the channel pore. However, the affinity of the pore for Ca2+ is low enough still to allow considerable Ca2+ permeation. A key role of VI-NSCCs in nutritional Ca2+ influx was demonstrated in Arabidopsis root epidermal cells (Demidchik et al., 2002a): 45Ca2+ flux measurements showed that Ca2+ uptake was blocked by Gd3+, a nonspecific blocker of VI-NSCCs. In addition, [Ca2+]cyt measured in Arabidopsis roots by Ca2+/aequorin chemiluminometry revealed a linear voltage dependence, consistent with the voltage independence of VI-NSCCs. A further important feature of VI-NSCCs is their enhanced activity in cells of the elongation zone where Ca2+ influx is particularly high. A reason for this requirement may be the Ca2+-dependent exocytosis that sustains cell expansion.
Interestingly, HACaC activity in root epidermal protoplasts appears approx. 40–60 min after formation of the whole-cell configuration (Demidchik et al., 2002a). This behaviour suggests that HACaCs either can be inserted into the plasma membrane or are activated by some unknown factor such as reactive oxygen species (ROS). Plant annexins have been shown to catalyse transmembrane Ca2+ fluxes (Clark & Roux, 1995) and, in the presence of Ca2+ and ROS, animal annexins move to the plasma membrane and form Ca2+-permeable channels (Gerke & Moss, 2002). Animal annexin-type Ca2+ conductances are similar to HACaCs: they are activated by intracellular Ca2+ and hyperpolarization. Thus, annexins could function as HACaC-like Ca2+ channels in plants (Clark & Roux, 1995; White et al., 2002).
Apart from Ca2+ nutrition and signalling, HA-NSCCs may also have specialized tasks in specific membranes. A very special type of HA-NSCCs was found in the peribacteroid membrane that surrounds bacterial symbionts in nitrogen-fixing root nodules (Tyerman et al., 1995). These hyperpolarization-activated channels showed permeability to monovalent and divalent cations, required cytosolic Mg2+ for their inward rectification and were inhibited by polyamines and Ca2+ in the symbiosome lumen (Whitehead et al., 1998, 2001; Roberts & Tyerman, 2002). The peribacteroid membrane controls fluxes of fixed nitrogen (NH3 or NH4+) between bacteria and the host cells (Obermeyer & Tyerman, 2005) and the HA-NSCCs found in this membrane mediate the NH4+ flux, thus playing an important role in the nitrogen metabolism of legumes (Tyerman et al., 1995). Interestingly, the HA-NSCCs present in the peribacteroid membrane have an extremely low unitary conductance of approx. 0.11 pS, which could only be determined using noise spectrum analysis. Such low unitary conductance is atypical for most ion channels but more often associated with carriers, or ion channels with complex gating (Tyerman et al., 1995; Obermeyer & Tyerman, 2005).
III. NSCCs activated by reactive oxygen species (ROS-NSCCs)
Reactive oxygen species integrate signalling pathways involved in plant stress responses, growth and development, gravitropism, hormone action, and many other physiological phenomena (Apel & Hirt, 2004). However, the molecular mechanisms of ROS regulatory action are poorly understood. In animals, ROS-activated cation channels mostly belong to NSCCs and K+ channel types, and have been shown to play a multitude of roles: from regulation of blood pressure to maintaining neuronal networks (reviewed by Lahiri et al., 2006).
Plant ROS-NSCCs were initially detected in the green alga Nitella flexilis by Demidchik et al. (1996, 1997a,b, 2001). Extracellular application of redox-active transition metals, Cu2+ and Fe3+, which leads to the production of hydroxyl radicals (HRs), activated a voltage-independent, instantaneous, nonselective cationic conductance that was sensitive to divalent cations, protons and the organic Ca2+ channel blocker, nifedipine. Analysis of the temperature dependence of the Nitella conductance activated by Cu2+ suggested an ion channel-based mechanism with a typical Q10 of 1.2–1.6. The idea that HRs can activate cation channels was later examined in higher plants (Demidchik et al., 2003b; Foreman et al., 2003; Inoue et al., 2005). In Arabidopsis epidermal root cells, generation of extracellular HRs led to the activation of cation-selective channels with relative permeabilities of K+ (1.00) ≈ NH4+ (0.91) ≈ Na+ (0.71) ≈ Cs+ (0.67) > Ba2+ (0.32) ≈ Ca2+ (0.24) > TEA+ (0.09).
A special aspect of ROS-NSCCs is their key role in elongation/expansion of plant cells (Demidchik et al., 2003b, 2007; Foreman et al., 2003) (Fig. 4). In elongating root hairs and in root elongation zone cells, HR-activated NSCCs showed significantly higher activity than in mature cells. The resulting increased Ca2+ influx and stimulated actin/myosin interaction lead to accelerated exocytosis, polar vesicle embedment and cell elongation (reviewed by Carol & Dolan, 2006). During this process, a plasma membrane localized NADPH oxidase was shown to produce the ROS necessary for the activation of Ca2+-permeable NSCCs (Foreman et al., 2003) (Fig. 4). Mutants lacking the oxidase (rhd2-1) produced far less extracellular ROS, did not form root hairs and exhibited stunted cell expansion in the elongation zone (Foreman et al., 2003). Interestingly, NADPH oxidase itself is stimulated by cytosolic Ca2+, thus forming a positive feedback mechanism to amplify ROS signals (Fig. 4). Other oxidases, such as cell wall peroxidases, can also produce ROS for Ca2+-driven elongation growth (reviewed by Kawano, 2003), and activation of ROS-NSCCs may also participate in pollen tube growth where a similar interaction between ROS and elevated [Ca2+]cyt has recently been found (Malho et al., 2006; McInnis et al., 2006).
Hydrogen peroxide (H2O2) did not activate whole-cell currents in protoplasts isolated from the Arabidopsis mature root epidermis. However, H2O2 does induce inward Ca2+ currents in protoplasts from the epidermal elongation zone (Demidchik et al., 2007). Several mechanisms could be responsible for these contrasting results: (i) ROS-NSCCs in growing tissues may have a different structure and/or regulatory properties; (ii) young elongating cells may contain a higher density of H2O2-permeable aquaporins (Eisenbarth & Weig, 2005; Bienert et al., 2006), allowing H2O2 to access potential regulatory sites in the cytosol; (iii) elongating cells have a greater capacity to generate H2O2. Localization of H2O2 activation sites at the cytoplasmic side has recently been confirmed by Demidchik et al. (2007) (Fig. 4). These authors applied H2O2 extracellularly and intracellularly to mature epidermal protoplasts in different patch-clamp configurations. H2O2-induced activation was observed only when H2O2 was applied to the cytosolic side of the membrane. The same study showed that the recorded channels had no selectivity amongst divalent cations and a unitary Ca2+ conductance that was similar to that of HACaCs in the same membranes. The latter finding suggests that ROS-NSCCs and HACaCs are the same channel. The data also suggest that there are at least two different types of ROS-NSCC: HR-activated, which were found in all tested root cell types; and H2O2-activated, in the elongation zone.
Reactive oxygen species modulation of NSCCs tightly links activity of these channels to signalling processes that involve ROS (reviewed by Pitzschke et al., 2006). Of particular interest in this respect is abscisic acid (ABA) signalling (reviewed by Pei & Kuchitsu, 2005). This hormone is involved in the regulation of stomatal closure, seed dormancy, flowering, activation of antioxidants and defence reactions, stress responses and other phenomena, which are also known to be accompanied by ROS accumulation. ABA can stimulate NADPH oxidase-mediated generation of ROS, leading to the activation of Ca2+-permeable NSCCs, Ca2+ influx and stomatal closure (Pei et al., 2000; Kwak et al., 2003). Guard cell H2O2-activated NSCCs showed selectivity and voltage dependence similar to H2O2-activated channels in root epidermis. However, they were not studied at the single-channel level and further examination is required to reveal whether root and leaf ROS-NSCCs are the same.
IV. Cyclic nucleotide-gated channels (CNGCs)
The second messengers 3′,5′-cyclic adenosine monophosphate (cAMP) and 3′,5′-cyclic guanosine monophosphate (cGMP) participate in many aspects of growth and development of higher plants. Examples include phytochrome signalling, gibberellic acid-induced signalling, pollen tube tip growth, plant cell cycle progression and salt stress tolerance (Maathuis & Sanders, 2001; Rubio et al., 2003; Talke et al., 2003; Newton & Smith, 2004).
In animal cells, many cyclic nucleotide-based signalling cascades are relayed via the cyclic nucleotide-dependent kinases PKA and PKG. However, no orthologues of PKA and PKG have been found in plants. The number of plant gene products with putative cyclic nucleotide binding sites is limited to c. 40 in Arabidopsis (Maathuis, 2006b) and mainly consists of membrane transporters. Within this category, the cyclic nucleotide-gated ion channels (CNGCs) constitute the largest group.
In animals, CNGCs function in the transduction of sensory input and in Ca2+ signalling (Biel et al., 1998) and are gated via binding of cAMP or cGMP to a cyclic nucleotide-binding domain (CNBD) near the C-terminus (Fig. 5). Additional control over gating is provided by a calmodulin-binding domain (CaMBD) at the N-terminus which interacts with the cyclic nucleotide-binding domain, thereby lowering the affinity for cAMP or cGMP. Similarly, plant CNGCs contain a C-terminal CNBD but differ from their animal counterparts regarding the CaMBD, with the plant CaMBD located at the C-terminus and partly overlapping the CNBD (Fig. 5). Nevertheless, CaM binding appears to have a comparable function in plant CNGCs, lowering their activity by preventing cyclic nucleotide binding. CNGCs have a generalized predicted structure of six transmembrane domains, S1–S6, with a pore domain (P loop) between S5 and S6, although the number of transmembrane spans varies considerably depending on the hydrophobicity algorithm that is used. The region that endows ion selectivity, the P loop selectivity sequence, is significantly different in plant CNGCs compared with animal CNGCs.
It has proved frustrating to achieve routine functional expression of plant CNGCs and this has greatly hampered the evaluation of their physiological roles. In a number of cases, heterologous expression in oocytes was achieved (Leng et al., 2002; Balague et al., 2003), in particular with Arabidopsis CNGC isoforms. The data showed that activation is cGMP- and/or cAMP-dependent, that most CNGCs do not discriminate amongst monovalent cations, have a limited Ca2+ permeability and are blocked by Cs+ and external Mg2+. A notable exception is AtCNGC2 which exhibited a high degree of K+ selectivity with regard to Na+ (Hua et al., 2003), a feature that is unknown in animal CNGCs. Interestingly, AtCNGC1, AtCNGC2 and NtCBP4, but not AtCNGC4, all show considerable voltage dependence, resulting in a strong inward rectification. This phenomenon is unknown in animal CNGCs and it remains to be revealed where this rectification originates.
Information on the physiological roles of plant CNGCs mainly derives from forward and reverse genetics and appears to cover two broad categories: plant pathogen interactions and plant cation nutrition. Null mutants in AtCNGC2 (Clough et al., 2000) and AtCNGC4 (Balague et al., 2003) showed that both these isoforms may participate in signalling events during pathogen attack. More recently, data from a deletion mutant leading to a CNGC11-CNGC12 chimera also suggested a role in the pathogen response for these CNGCs (Yoshioka et al., 2006). It is well documented that Ca2+ as well as K+ and Cl− fluxes occur during the early phase of plant defence responses and, in addition, cyclic nucleotides have been implicated in plant defence-related signalling. For example, ROS production requires cAMP and Ca2+ in French bean cells (Bindschedler et al., 2001), whereas exposure of tobacco suspension cells to nitric oxide causes a transient increase in cGMP concentrations (Durner et al., 1998). Thus, both Ca2+ and cyclic nucleotide-based signalling form part of defence responses where CNGCs may be involved.
Several studies have reported on the role of CNGCs in plant nutritional aspects. Loss of function in CNGC1 led to mutants that showed lower shoot Ca2+ contents and altered gravitropic root response, and the authors concluded that CNGC1 may participate in plant Ca2+ uptake from the external medium (Ma et al., 2006). AtCNGC3 is predominantly expressed in epidermal and cortical root tissues and a null mutation in this gene altered both short-term Na+ influx and K+ uptake in high external K+ conditions, indicating it may play a role in nonselective monovalent cation uptake (Gobert et al., 2006). Apart from having a role in pathogen response which presumably is limited to shoot tissue, AtCNGC2 in roots influences the homeostasis of cations such as Ca2+. Mutant plants became hypersensitive to elevated Ca2+ without increasing tissue Ca2+ contents (Chan et al., 2003). Another isoform, CNGC10, that is also relatively highly expressed in root tissue was able to complement the K+ uptake deficient phenotype of the akt1-1 loss of function mutant, showing it can form a root K+ uptake pathway and considerably augment K+ uptake when overexpressed in the akt1-1 genotype (Li et al., 2005). In addition, the application of membrane permeable analogues of cAMP and cGMP can affect Na+ influx, Na+ efflux and K+ influx (Maathuis & Sanders, 2001; Essah et al., 2003; Rubio et al., 2003; Maathuis, 2006b), suggesting that here, too, CNGCs may be involved. CNGCs may also impact on the uptake of toxic monovalent cations such as Cs+ (Hampton et al., 2005).
Thus, it appears that plant CNGCs may function in roots as important contributors to nonselective uptake of cations, whereas in shoots their function may be mainly in early signalling events that form part of the pathogen response. In the context of signalling cascades, CNGCs may form important elements in crosstalk between cyclic nucleotides and Ca2+ (Talke et al., 2003), since CNGC activation is sensitive to both types of second messenger and CNGCs also conduct Ca2+. For example, the observation that externally applied membrane-permeable analogues of cyclic nucleotides causes transient elevations in [Ca2+]cyt (Volotovski et al., 1998) may indicate that Ca2+ signals are directly downstream of cyclic nucleotides, and are generated through the activity of CNGCs.
V. Amino acid-gated NSCCs
Ionic conductances activated by amino acids, specifically by glutamate and glycine, are critical for synaptic transmission and other complex physiological phenomena in animals (Dingledine et al., 1999). Amino acid-activated conductances are mediated by ionotropic glutamate receptors. After binding glutamate or glycine, ionotropic glutamate receptors (iGluRs) form cation channels with variable selectivity, conductance, kinetics and pharmacology (Dingledine et al., 1999). Genes with similarities to those encoding animal ionotropic glutamate receptors have been found in plants (Lam et al., 1998). Twenty glutamate receptor-like genes (termed GLRs; Fig. 5) are found in the genome of Arabidopsis thaliana and are divided into three subgroups based on sequence similarity (Lacombe et al., 2001; Chiu et al., 2002). The family of GLR genes in plants shows a great divergence from its animal counterpart, particularly in the pore region (Davenport, 2002). Since functional analysis is lacking, it is still unknown what type of channel is encoded by GLRs, but on the basis of overall homology to their animal counterparts, GLRs are believed to be a subclass of ligand-gated NSCC (Davenport, 2002).
In animals, many iGluRs are NSCCs. Some of them form Ca2+-permeable NSCCs with properties similar to constitutive VI-NSCCs of higher plants. Dennison & Spalding (2000) reported that extracellular application of glutamate generated transient increases in [Ca2+]cyt in Arabidopsis seedlings. Lanthanides inhibited this effect, as did removal of bath Ca2+, suggesting the Ca2+ entered the cytosol through glutamate-activated cation channels. Using the same experimental system, Dubos et al. (2003) have found that glycine addition could also elevate [Ca2+]cyt. They also showed that glycine has synergistic effects when added simultaneously with glutamate, probably because it increases the affinity of glutamate binding to the receptor. Demidchik et al. (2004) obtained data on protoplasts isolated from different root tissues and demonstrated that similar amounts of glutamate induced larger increases in [Ca2+]cyt in mature epidermal and cortical cells compared with cells from deeper tissues. These authors have also carried out the first electrophysiological characterization of glutamate-activated currents in plants and found that the probability of observing glutamate-activated conductances is low and increases with increasing glutamate concentrations. The recorded glutamate-activated conductances were nonselective for monovalent cations, Ca2+-permeable, voltage-independent, and revealed instantaneous activation kinetics. Channel activity was sensitive to quinine and lanthanides, resembling the pharmacology of constitutive VI-NSCCs.
The functionality of plant ionotropic glutamate receptors has been questioned, particularly with respect to the availability of ligand in the extracellular compartment. However, several studies have shown that apoplastic concentrations of glutamate and glycine can range from 0.01 to 1 mm (Lohaus et al., 1995, 2001; Lohaus & Heldt, 1997). Such concentrations are high enough to activate ionotropic glutamate receptors, and apoplastic glutamate and glycine could therefore function in plant Ca2+ uptake and/or signalling.
Despite the lack of electrophysiological characterization, effects of exogenous amino acids on plant cell physiology and phenotypic properties of glr knockout mutants have been a subject of intensive investigations during the last 5 years. A number of studies has shown that exposure to extracellular glutamate modifies Ca2+-dependent physiological processes, such as depolymerization of microtubules, cell elongation and responses to aluminium (Sivaguru et al., 2003), root branching (Walch-Liu et al., 2006) and sugar sensing (Dubos et al., 2005). Loss of function mutations in Oryza sativa GLR3.1 (Li et al., 2006) and AtGLR3.2 (Kim et al., 2001; Turano et al., 2002) showed their involvement in Ca2+ accumulation and Ca2+-mediated reactions, for example stress responses, programmed cell death, cell division and differentiation. AtGLR 1.1 may be involved in seed germination and ABA-mediated processes (Kang & Turano, 2003; Kang et al., 2004). Overexpression of a radish GLR (Kang et al., 2006) in Arabidopsis enhanced glutamate-activated transient [Ca2+]cyt elevation and altered Ca2+-mediated mechanisms such as necrosis, growth and development. Additionally, it resulted in enhanced resistance to a fungal pathogen, possibly because of the up-regulation of jasmonic acid-responsive defensin genes. Strong evidence that GLR3.3 forms Ca2+-permeable channels has recently been obtained by Qi et al. (2006). Depolarization and elevation of [Ca2+]cyt induced by glutamate were prevented in glr3.3 knockout mutants. Interestingly, in addition to glutamate, five other amino acids (glycine, alanine, serine, asparagine, and cysteine) and glutathione (g-glutamyl-cysteinyl-Gly) were demonstrated to be agonists of the GLR3.3-induced responses. Overall, these data strongly suggest that glutamate causes physiological effects through an increase in [Ca2+]cyt catalysed by iGluRs. Some of these effects, for example altered ABA signalling, can also be mediated by ROS, because an increase in [Ca2+]cyt activates membrane-bound NADPH oxidases (Sagi & Fluhr, 2001).
VI. Purine signalling
In animals, extracellular purines such ATP and ADP function as signalling molecules, activating specific ionotropic (P2X) and G-protein-coupled (P2Y) receptors. Animal P2X receptors form NSCCs after interaction with ATP, ADP and sometimes AMP (Ralevic & Burnstock, 1998). Permeability to Ca2+ is an important characteristic of several P2X receptors (North, 2002), and P2X receptors are involved in physiological activities ranging from neurotransmission to cell death and cell proliferation (reviewed by Ralevic & Burnstock, 1998).
Physiological effects of extracellular purines in plants have received little attention. In 1960s and 1970s, the effect of purines on plants was interpreted in terms of energy supplementation or Ca2+ chelation, but not in the context of signalling. The extracellular presence of purines was found to affect plant movement (Jaffe, 1973), K+ transport (Luttge et al., 1974), and activity of cell-degrading enzymes (Udvardy & Farkas, 1973). Although such effects were often observed in the presence of micromolar concentrations of ATP, for example when studying the Venus fly trap closure, they were not considered to be the result of receptor activity (Jaffe, 1973).
Recent interest into the notion that extracellular purines can function as signalling agents in plants was triggered by two reports, showing: (i) depolarization of the plasma membrane by extracellular ATP and ADP in Arabidopsis root hairs (Lew & Dearnaley, 2000); and (ii) transient elevations in [Ca2+]cyt induced by extracellular purines (Demidchik et al., 2003a). In these studies, low micromolar concentrations of ADP and nonhydrolysable ATP analogues were found to be effective, indicating that signalling rather than energetics underlies purine activity. The effect of purines on [Ca2+]cyt was blocked by lanthanides and by nonspecific (suramin) and specific (pyridoxalphosphate-6-azophenyl-2′,4′-disulfonic acid) antagonists of animal P2X receptors, and absent when external Ca2+ was removed.
Although conclusive evidence is still lacking, the accumulative data strongly suggest that purinergic receptors exist in plants and that these are involved in signalling events during stress and wounding.
VII. Mechanosensitive ion channels
Many environmental cues lead to alterations in physical forces on membranes. In plants, such cues are in particular associated with changes in turgor but also mechanical perturbation for example by wind. Such external signals can be relayed into electrical and/or Ca2+ signals through the action of mechanosensitive channels (MCs) whose gating depends on changes in tension forces on the membrane. MC open probability is often positively correlated to membrane stretch, hence the terms MC and stretch activated channel are often used to denote the same type of channel.
Several MCs have been characterized in plant plasma membranes using patch-clamp analyses showing both anion and cation permeability in MCs (Cosgrove & Hedrich, 1991; Pickard & Ding, 1993; Qi et al., 2004). In guard cells, MCs are believed to mediate anion and cation movement across the guard cell plasma membrane and thus contribute to volume and turgor regulation of stomata (Cosgrove & Hedrich, 1991). MCs are blocked by trivalent cations such as Gd3+ and Al3+, and although most MCs are nonselective, occasionally an MC conductance with considerable selectivity is described, such as the small-conductance (3 pS) Ca2+-selective MC in Vicia faba guard cells (Cosgrove & Hedrich, 1991). The nonselective nature of MCs causes a large Ca2+ influx upon activation and several authors have suggested that the resulting Ca2+ signal forms a crucial intermediate in the transduction of mechanical stimuli. For example, Sato et al. (2001) studied mechanical stimulation of the chloroplast avoidance response in ferns: mechanical pressure on protonemal cells resulted in chloroplasts moving away from the site within an hour. This response only occurred when extracellular Ca2+ was present and it was sensitive to Gd3+, suggestive of a Ca2+-permeable MC-based process. Pollen tubes were shown to contain several types of MC (Dutta & Robinson, 2004), one of which may be involved in maintaining the polar Ca2+ gradient necessary for directional growth of the tube.
On the basis of homology to bacterial MC genes, a family of 10 MC-like genes was discovered in Arabidopsis (MSL1–10, Haswell & Meyerowitz, 2006). MSL3 can rescue the osmotic shock sensitivity of bacterial mutants, whereas null mutations in MSL2 and MSL3 led to abnormalities in shape and size of Arabidopsis plastids. Both MSL2 and MSL3 localized to the plastid envelope and the authors hypothesize that these proteins mediate plastidic ion release in response to increased osmotic pressure within the plastid. Whether an intracellular differential in osmotic pressure that is large enough to activate MCs ever occurs remains a question.
VIII. Vacuolar NSCCs
Plant vacuoles can constitute 90% of the cellular volume and play essential roles in turgor provision, chlorophyll breakdown, programmed cell death and development, mineral storage, and as a depository for xenobiotics and toxic compounds. Vacuoles also constitute a major intracellular Ca2+ store and are therefore important in cellular signalling (Allen & Sanders, 1997).
The amenability of vacuoles to patch-clamp methodology ensured that tonoplast channels were amongst the first and best characterized plant ion channels. The accumulative data show the presence of two NSCCs (Allen & Sanders, 1997): the ubiquitous slow vacuolar (SV) channel has K+/Na+ and K+/Ca2+selectivity ratios of around 1 and 4, respectively, is activated by tonoplast depolarization, has slow kinetics and requires elevated cytoplasmic Ca2+ concentrations. The fast vacuolar (FV) channel has a similar K+/Na+ selectivity ratio to the SV channel and is inhibited by elevated [Ca2+]cyt. The SV channel is now well established as a voltage-dependent NSCC which is sensitive to both cytoplasmic and luminal Ca2+ concentrations and further regulated by a host of mechanisms including phosphorylation (Bethke & Jones, 1997), 14-3-3 proteins (van den Wijngaard et al., 2001), organic cations and redox potential (Scholtz-Starke et al., 2005).
The identification of the Arabidopsis SV channel as AtTPC1 (Peiter et al., 2005) led the way to delineating the role of the SV channel through overexpression and loss of function genetics. ABA-induced delay of seed germination was significantly less in a tpc1-1 knockout mutant, whereas it was augmented in overexpressing lines. In guard cells, ABA-dependent closure was not affected by the expression of AtTPC1, but high external Ca2+, another well-documented closing stimulus, largely failed to evoke stomatal closure in the knockout mutant. The stomatal phenotype in tpc1-1 mutants is reminiscent of det1-3 and cas1-1 mutants, both of which are altered in their Ca2+ signature. Thus, the SV channel, at least in guard cells, might mediate Ca2+ release from the vacuole which contributes to cytoplasmic Ca2+ signatures.
There are currently two sources of debate regarding the SV channel. Firstly, in species such as rice and tobacco the SV channel activity is also found in vacuolar membranes but orthologues of AtTPC1 were proposed to be expressed at the plasma membrane and not at the vacuolar membrane. This implies that orthologous proteins may be targeted to different membranes depending on plant species. It also suggests that rice and tobacco SV channels are not coded by the AtTPC1 homologues NtTPC1 and OsTPC1. Manipulation of NtTPC1 and OsTPC1 expression showed that these proteins may be involved in elicitor-induced Ca2+ signalling and in pathogen response (Kurusu et al., 2004).
Secondly, a function of SV channels in Ca2+ signalling has been disputed. Although the SV channel can conduct Ca2+ in certain experimental conditions, it is questionable whether this occurs in vivo. SV channel activity requires high cytoplasmic Ca2+ concentrations and is severely blocked by luminal Ca2+. Thus, several studies have shown that, with physiological transtonoplast Ca2+ gradients and voltages, no SV channel activity is observed (Pottosin et al., 1997).
Far less is known about the other tonoplast NSCC, the FV channel. Originally observed in red beet storage tissue (Hedrich & Neher, 1987), few further studies have been published. FV channels become increasingly inactive whenever the cytoplasmic Ca2+ concentration exceeds c. 200 nm, and FV open probability has been reported to be largely insensitive to tonoplast potential. Subsequent publications reported on the presence of FV channels in other tissues, such as barley mesophyll vacuoles (Tikhonova et al., 1997), showing moderate outward rectification and a biphasic voltage dependence. One of the physiological roles for the FV channel may be in maintaining cellular K+ homeostasis, since both luminal and cytoplasmic K+ concentrations affect FV open probability (Pottosin & Martínez-Estévez, 2003). Other putative roles for the FV channel include providing a shunt conductance for the V-ATPase (Davies & Sanders, 1995), osmoregulation and regulation of the tonoplast potential (Allen & Sanders, 1997).
IX. Cation channels sensitive to elicitors
Elicitors are compounds released during pathogen attack. They can be produced by pathogens themselves or derive from host cell wall fragments and are important cues to plants in their capacity of evoking defence mechanisms, such as the hypersensitive response. Elicitors themselves may have pore-forming ability (Klüsener & Weiler, 1999; Lee et al., 2001) and thus cause electrical signals directly when they are present. However, in most cases, it is assumed that their action is via host receptor proteins that transform perception of these chemical signals into electrical and Ca2+ signals. One of the earliest reports on this mechanism described the activation of Ca2+-permeable channels in tomato protoplasts by fungal elicitors through intermediate steps that included a putative G-protein and channel phosphorylation (Gelli et al., 1997). A similar process has been described in parsley protoplasts treated with Phytopthora-derived cell wall elicitor (Zimmermann et al., 1997). However, in tomato, hyperpolarization-activated channels carried the Ca2+ influx, whereas in parsley channel activation occurred when cells were far more depolarized.
Addition of yeast elicitor and chitosan induced whole-cell hyperpolarization-activated currents in Arabidopsis guard cells (Klüesener et al., 2002). This Ca2+ current required the production of ROS through the action of NADPH oxidases suggestive of a linear pathway, elicitor > ROS > Ca2+, and pointing to a general mechanism where stress leads to the production of ROS and subsequent Ca2+ signals. Nevertheless, elicitor-induced Ca2+ influx has been reported to occur both before (Blume et al., 2000) and after (Kawano & Muto, 2000) ROS production and therefore different classes of Ca2+ channel and Ca2+ store are probably involved during the perception of elicitors.
In tobacco suspension cells, cosuppression of NtTPC1 led to a significant decrease in the cryptogene-induced Ca2+ signal (Kadota et al., 2004). Similarly, overexpression of OsTPC1 generated an increase in Ca2+ signal in response to fungal elicitor and also a larger production of ROS. Both NtTPC1 and OsTPC1 are believed to be targeted to the plasma membrane, and the data suggest TPC1 may be one of the initial pathways that links elicitor perception to cytoplasmic Ca2+ signals and subsequent production of ROS. ROS, in turn, have been shown to directly activate HACaCs (Pei et al., 2000), thus creating a further amplification of the initial signal.
Although an extremely high Ca2+ selectivity was reported for some elicitor-activated cation channels (Gelli et al., 1997), a proper evaluation of channel selectivity is absent for any of them. Such studies are required to show whether elicitor-activated cation channels form a separate group of highly Ca2+-selective cation channels or belong to NSCCs. The latter seems more likely since, as mentioned earlier, plants do not appear to have genes that encode Ca2+-selective channels.
X. NSCCs acting in concert
In spite of belonging to a diverse range of gene families and having various predicted structures, NSCCs catalyse the same process – passive transport of cations. Influx and efflux of cations require careful regulation, since cations are involved in many crucial aspects of cellular physiology. A number of gene families and isoforms that encodes NSCCs may contribute to cell- and tissue-specific fine-tuning of signalling events and ion fluxes. There are many scenarios where the activity of various classes of NSCCs may converge to mediate highly relevant physiological processes. One example would be the short- and long-term response of plants to salt stress (Fig. 6). The early response to the sudden onset of salt stress probably reflects turgor changes and involves both cGMP-based and Ca2+-based signalling (Donaldson et al., 2004). The observed rapid rise in cellular cGMP may derive from a receptor kinase-cyclase relay that senses turgor change and would lead to the activation of root cell CNGCs. As already described, CNGCs may form important nodes in signalling networks where cyclic nucleotide signals are converted into Ca2+ signals (Talke et al., 2003). The cytoplasmic Ca2+ signal could have many downstream targets and may include membrane oxidases leading to the production of ROS. Similar to other stresses, ROS have been recorded during salinity and drought stress, and via activation of ROS-NSCCs could sustain further Ca2+ influx. Sustained Ca2+cyt elevation often depends on Ca2+ release from internal stores, such as the vacuole, and this function could be mediated by SV-type channels, although no salinity-dependent phenotype was observed in either tpc1-1 loss of function or TPC1-1 overexpressing plants (Peiter et al., 2005). Elevated cGMP can directly inactivate root VI-NSCCs and thus reduce the influx of harmful Na+ (Maathuis & Sanders, 2001; Rubio et al., 2003). In addition, both elevated Ca2+ and cGMP are also known to affect gene transcription, including those that encode NSCCs (Maathuis, 2006b). The latter may also affect monovalent cation homeostasis during salt stress and reduce Na+ uptake through NSCCs such as CNGC3 (Gobert et al., 2006).
A second example where different types of NSCC interact is in the uptake of nutrients such as Ca2+. At resting membrane potentials, constitutive VI-NSCCs catalyse steady-state Ca2+ uptake for nutritional needs. HACaC activity will dominate at hyperpolarized voltages and/or when [Ca2+]cyt is elevated, and provides an additional Ca2+ loading that is required for the stimulation of exocytosis and growth (Fig. 6). A small increase in the activity of constitutive VI-NSCCs in growing tissues is probably sufficient to induce an initial increase in basal [Ca2+]cyt, leading to activation of HACaCs and further Ca2+ influx (Demidchik et al., 2002a). Similar amplification mechanisms could exist for the generation of Ca2+ signals in response to stress, signalling agents and other stimuli, such as gravity and hormones. Glutamate, glycine, purines, ROS, elicitors and membrane stretch all activate specific Ca2+-permeable NSCCs and NSCC-like conductances that elevate [Ca2+]cyt, which in turn stimulates HACaCs and NADPH oxidase (Fig. 6). As described for salt stress, the production of extracellular ROS may also form part of this positive feedback system.
XI. Conclusions and perspectives
Many classes of NSCC can be distinguished on the basis of electrophysiological, biochemical and genomics data (see Table 2 for an inventory of various types of plant nonselective ion channels). An increasingly recurrent theme is the important role this type of channel plays in signalling events. Many NSCCs show Ca2+ permeability and have been shown to mediate changes in [Ca2+]cyt that can be linked to important physiological processes, such as elicitor perception or ROS perception. Plant NSSCs therefore appear to substitute the Ca2+-selective channels in animal systems that carry out similar functions (Bothwell & Ng, 2005). The other main function of plant NSCCs is mainly confined to root tissues where they mediate uptake of important ions such as NH4+, Na+ and K+.
Table 2. An inventory of various types of plant nonselective ion channels and their physiological functions
For an explanation of the category definitions, see the text. The physiological functions ascribed to the different channel types are based on published studies but specific evidence is lacking in many cases.
The lack of selectivity implies less control over transmembrane fluxes and it remains unclear why these functions in plants are not mediated by selective ion channels, as is the case in animals. One main difference is the potential of large fluctuations in ionic conditions that plants may be exposed to, particularly root cells. Such fluctuations may require rapid adjustment of membrane voltage and osmotic potentials, functions that do not necessarily involve specific ions. The occurrence of salinity stress exemplifies this: a rapid reduction of the external water potential obliges cells to take up inorganic ions rapidly to increase cellular osmolarity. During K+ starvation, similar mechanisms may prevent excessive turgor loss. These processes are potentially beneficial but their nonselective nature would necessitate tight control over NSCC activity.
In principle, plant Ca2+ signalling does not appear to be different from that observed in animals. Yet in animals, Ca2+ signalling proceeds through Ca2+-selective channels, whereas in plants NSCCs seem to provide the rise in [Ca2+]cyt. Opening of NSCCs inevitably leads to transmembrane fluxes of other cations, such as K+ and Na+, which may be more problematic in animal cells where extracellular concentrations of 100–110 mm Na+ prevail. Most terrestrial plants are not exposed to large amounts of Na+ and hence such a disadvantage would not occur. However, it can be surmised that during salinity stress, opening of NSCCs to mediate Ca2+ signalling may well lead to toxic Na+ effects and it would be interesting to see whether halophytes have adaptations to avoid this.
In many cases, different types of NSCC may be manifestations of the same protein. For example, ROS-, cyclic nucleotide- and glutamate-activated NSCCs all show little or no voltage dependence and may very well constitute subclasses of VI-NSCCs. Additionally, ROS-NSCCs and constitutive HA-NSCCs demonstrate very similar unitary conductances (Véry & Davies, 2000; Demidchik et al., 2007). Both DA-NSCCs and HA-NSCCs are involved in monovalent cation uptake but in certain cell types could form elicitor-activated NSCC-like conductances. A proper delineation of NSCC functions requires molecular and electrophysiological tools, but most importantly a direct link between specific NSCC conductances and NSCC-encoding genes. Only in one case has conclusive evidence been reported for such a link, for the vacuolar NSCC, TPC1. In other cases, heterologous expression of cyclic nucleotide-gated channels led to nonselective currents (Leng et al., 2002; Balague et al., 2003), but the associated in vivo currents have not been reported. Several gene families have been identified, such as the CNGCs and GLRs that encode putative NSCCs, but in situ electrophysiological data are missing. Although it is imperative that the genomics and electrophysiological data for NSCCs can be combined, this is fraught with difficulties. Successful cloning and characterization of K+-selective channels (Sentenac et al., 1992; Hirsch et al., 1998) was based on yeast complementation assays and plant loss of function studies, both strategies that are not easily applied to NSCCs. The extensive gene families that encode putative NSCCs hamper loss of function studies, particularly because a large degree of functional redundancy may be present. Further transcriptomics and proteomics studies will aid in this respect, for example by revealing tissue- and membrane-specific expression of NSCC isoforms. The latter will greatly help in assigning potential functions but more importantly in targeting electrophysiological studies to the right cell type and cellular compartment. In vivo currents are often mediated by ion channels consisting of heteromers. Thus, to allow comparison between in vivo and heterologous systems, knowledge regarding channel subunit composition is urgently needed, for example through application of FRET-based studies. In combination, these approaches will provide vital steps in making further progress regarding the physiological functions of NSCCs.