Cl⁻ as a bona fide signaling ion

Abstract

Cl⁻ is the major extracellular (Cl⁻_out) and intracellular (Cl⁻_in) anion whose concentration is actively regulated by multiple transporters. These transporters generate Cl⁻ gradients across the plasma membrane and between the cytoplasm and intracellular organelles. [Cl⁻]_in changes rapidly in response to cell stimulation and influences many physiological functions, as well as cellular and systemic homeostasis. However, less appreciated is the signaling function of Cl⁻. Cl⁻ interacts with multiple proteins to directly modify their activity. This review highlights the signaling function of Cl⁻ and argues that Cl⁻ is a bona fide signaling ion, a function deserving extensive exploration.

INTRODUCTION

As the principal anion in all vertebrate cells, Cl⁻ is mainly studied as a homeostatic, rather than as a signaling ion. Plasma and extracellular Cl⁻ concentration ([Cl⁻]_out) maintains body fluid ionic content, volume homeostasis, and blood pressure. [Cl⁻]_out is, therefore, maintained fairly constant, except for disease state associated with metabolic acidosis and alkalosis (26). However, luminal [Cl⁻]_out can vary considerably during epithelial Cl⁻ secretion and absorption. Nevertheless, the activity of several basolateral and luminal proteins and physiological functions are influenced by changes in [Cl⁻]_out. Plasma Cl⁻ homeostasis is maintained mainly by the kidney (26) and the intestine (47) in close association with Na⁺, K⁺, and ${\text{HCO}}_3^{-}$ absorption and secretion (23).

Unlike [Cl⁻]_out, intracellular Cl⁻ concentration ([Cl⁻]_in) varies considerably between cell types and in response to cell stimulation. Resting [Cl⁻]_in in epithelial cells is as high as 60 mM (30, 44, 61, 121), while in neurons and skeletal muscle, it ranges from 10 mM to 30 mM (33, 114). In most cells the major Cl⁻ influx mechanism is the Na⁺/K⁺/2 Cl⁻ cotransporter NKCC1 with contribution of the Cl⁻/${\text{HCO}}_3^{-}$ exchanger AE2 when pH_in is above 7.2, while Cl⁻ efflux is largely mediated by the K⁺/Cl⁻ cotransporters KCC (50). These transporters keep [Cl⁻]_in above electrochemical potential in the face of several types of Cl⁻ channels that are expressed in many cells that are likely minimally active in the basal state. [Cl⁻]_in changes in response to many stimuli that involves activation of Cl⁻ channels that drives [Cl⁻]_in toward its electrochemical potential. Prominent Cl⁻ channels are anoctamin 1 (ANO1, TMEM16A), the Ca²⁺-activated Cl⁻ channel and cystic fibrosis transmembrane conductance regulator (CFTR), a cAMP/PKA activated Cl⁻ channel (57). For example, in the pancreatic and salivary gland acinar cells activation of G_q-coupled receptors evokes Ca²⁺ oscillations that results in oscillatory activation of ANO1 (36, 73), which reduces [Cl⁻]_in due to luminal Cl⁻ secretion (56, 86). CFTR is the major Cl⁻ channel in secretory gland ducts that are activated in response to stimulation of G_s-coupled receptor and increased cellular cAMP to initial luminal Cl⁻ absorption and ${\text{HCO}}_3^{-}$ secretion (56). Duct [Cl⁻]_in is reduced from ~40 mM to 4 mM by basolateral membrane Cl⁻ extrusion, to allow the high ${\text{HCO}}_3^{-}$ concentration (up to 140 mM) in secreted fluids (44, 56, 115). On the other hand, [Cl⁻]_in increases during regulatory volume increase (30, 56) and in lectin-stimulated T cells (55), largely by activation of NKCC1 (50). The inwardly directed Cl⁻ gradient is coupled to the that of Na⁺, K⁺, and ${\text{HCO}}_3^{-}$. The Na⁺ gradient mediates most Cl⁻ influx to keep [Cl⁻]_in above electrochemical equilibrium. K⁺ mediates most Cl⁻ efflux, while the coupling of Cl⁻ to ${\text{HCO}}_3^{-}$ transport has a major role in determining cytoplasmic pH (pH_in).

Although the homeostatic functions of [Cl⁻]_in are extensively studied, signaling by [Cl⁻]_in and its regulatory functions are poorly understood and not completely appreciated. [Cl⁻]_in has all the features of a signaling ion. The concentration of a signaling ion should be actively regulated. There should be gradients of the ion across cellular membranes, the gradients should be rapidly dissipated in response to cell stimulation, and the change in the signaling ion concentration should be translated to an altered cell response. [Cl⁻]_in is regulated by secondary active transporters that generate Cl⁻ gradients across the plasma membrane (56) and intracellular organelles (47). Cell stimulation changes [Cl⁻]_in in multiple cell types rapidly and substantially. For example, carbachol stimulation of salivary gland cells reduces [Cl⁻]_in from 60 to 70 to 30 to 35 mM within a few seconds (30), cAMP stimulation reduces [Cl⁻]_in from 75 to 30 mM in MDCK cells (112), and by 10 mM in collecting duct cells (3). Additionally, lectin stimulation of T cells causes persistent [Cl⁻]_in oscillations (55). Oscillations in [Cl⁻]_in are likely a result of apical receptor-stimulated Ca²⁺ oscillations that activate the Ca²⁺-activated Cl⁻ channels in acinar cells (73). Changes in [Cl⁻]_in affect cellular processes that function in time scales of milliseconds, such as the membrane potential (9, 39) and protein folding (18, 76), to hours and days, such as cell growth and differentiation (13, 53). In this review, we will discuss what we know about Cl⁻ regulation, what we do not know, and what needs to be elucidated as we seek to understand better signaling by Cl⁻. We will discuss first the current evidence for signaling and regulatory functions by [Cl⁻]_out, which are only suggestive at this stage. We will then discuss the much better-established evidence for signaling by [Cl⁻]_in. We will attempt to show that the known regulatory functions of [Cl⁻]_in and dynamic changes in [Cl⁻]_in in response to cell stimulation justify considering [Cl⁻]_in as a prominent regulator of cell function and as a bona fide signaling ion that senses and transmits signals.

REGULATION BY [Cl⁻]_out

[Cl⁻]_out Regulates the Function of Signaling Proteins

There is no clear evidence for signaling by [Cl⁻]_out, although [Cl⁻]_out regulates the function of several proteins, including signaling proteins. Changes in [Cl⁻]_out are observed in select environments, such as the luminal space in epithelia and the synaptic cleft (56, 101). In this manner, changes in [Cl⁻]_out indirectly affect cell signaling to modulate cell function. In addition, [Cl⁻]_out is a cofactor in several enzymes. An extensively studied enzyme regulated by [Cl⁻]_out is the ANG I-converting enzyme. [Cl⁻]_out is an essential cofactor for the testis ANG I-converting enzyme (tACE) (62), and somatic ANG-I (sACE) (60). sACE is found on the surface of vascular and renal endothelial cells, proximal tubules, and podocytes. It is also found on the luminal membrane of other epithelial cells (107) and in macrophages (93). sACE is a zinc metallopeptidase that catalyzes the hydrolysis of carboxy-terminal dipeptides of several biologically active peptides (60). The best understood role of sACE is in the renin-angiotensin system. sACE cleaves ANG I to the vasoconstricting ANG II (95). Additionally, sACE has a significant role in innate and adaptive immunity (92), and tACE has a critical role in fertilization (35). The ACE ectodomain has two subdomains with Cl⁻ binding pockets. The N subdomain binds one Cl⁻ ion, and the C subdomain binds two Cl⁻ ions (62). Affinity of the sACE to Cl⁻ is ~30 mM at pH 7.5 and as high as 100 mM at pH 8.6 (24). It is not obvious how sACE senses and responds to Cl⁻_, since plasma [Cl⁻]_out is ~100 mM. On the other hand, [Cl⁻]_out can be as low as 10–20 mM in the epithelial luminal space (56) and seminal fluid (48) that may affect Cl⁻ binding and sensing by the ACEs. Finally, sACE is also present intracellularly, mainly in the nucleus and mitochondria, where it regulates the renin-angiotensin system (1). Cytoplasmic Cl⁻ can be as low as 4 mM (see channels and transporters regulated by [cl⁻]_in), well within the affinity of the ACEs for Cl⁻ (60).

Several guanylyl cyclases and G protein-coupled receptors (GPCRs) sense [Cl⁻]_out, and their signal transduction is controlled or modulated by [Cl⁻]_out. The natriuretic vasodilatory hormone atrial natriuretic peptide (ANP) is secreted by the heart in response to volume expansion to stimulate NaCl and fluid excretion (100). ANP binds to and activates the A-type natriuretic peptide receptor and signals by increasing cellular cGMP (82). The binding of ANP to the ANPR and the stimulation of cGMP is strictly dependent on Cl⁻ with an apparent affinity of ~1 mM (Fig. 1A). The ANP Cl⁻ binding site (Fig. 1, C and D) is next to the ANP binding site in the extracellular domain of the ANPR (106), and it seems to affect the extracellular domain dimerization, which is required for signaling by ANP (65). The very high affinity of the ANPR to [Cl⁻]_out raises the question of at what situation [Cl⁻]_out becomes physiologically relevant. A good example is in the kidney, where the ANPR is expressed at the luminal membrane of intercalated cells (38). Volume depletion and metabolic alkalosis activate the renin-angiotensin system to reduce luminal [Cl⁻]_out to below 10 mM (31).

Fig. 1.

Signaling by [Cl⁻]_out: effect of [Cl⁻]_out on ligand-receptor interaction. A: effect of [Cl⁻]_out on atrial natriuretic peptide (ANP) binding was measured in Chinese hamster ovary (CHO) cells transfected with the full-length ANP receptors. Isolated membranes were incubated in media containing the indicated Cl⁻ concentrations and ¹²⁵I⁻-ANP to measure ANP binding to the full-length ANP receptor (a). Isolated membranes from CHO cells transfected with the full-length ANP receptors were used to measure stimulation of guanylyl cyclase activity by ANP in the presence and absence of 100 mM NaCl (b). Results are taken from Ref. 65. B: group II and III mGlu receptors (mGluRs) were cotransfected with a chimeric G_i/G_q protein. The receptors were labeled with blue and green fluorescent tags for measurement of FRET. Labeled cells incubated in solutions containing [Cl⁻]_out of 154.6 mM (green) to 2 mM (black) and FRET between Group II and III mGluRs in response to glutamate stimulation was measured. Shown are the changes in the indicated mGluRs conformation in response to receptor stimulation. Results are taken from Ref. 102. C and D: crystal structures of the rat ANP receptor (ANPR: gray; PDB_ID: 1DP4) and human mGlu8 receptor (GluR: green; PDB_ID: 6BSZ) (C) were obtained from the protein data bank website. The models were aligned and visualized using UCSF Chimera 1.11 (77). In D, a closeup of the ANPR chloride binding site residues are shown in red and the chloride ion in blue, with the mGlu8 Cl⁻ binding site in green.

Notably, the extracellular ligand-binding domain of the metabotropic glutamate receptors (mGluR) fold is similar to the fold of the ANPR extracellular domain [(52, 106) and Fig. 1, C and D)]. The Cl⁻ binding site of the ANPR is highly conserved in the various mGluRs with a nearly identical fold (65), and [Cl⁻]_out regulates signaling by the mGluRs (101, 102). The use of FRET between mGluRs extracellular domains and the measurement of intracellular Ca²⁺ ([Ca²⁺]_i) has revealed that [Cl⁻]_out markedly increases the apparent affinity of most mGluRs to their agonist (Fig. 1B). The apparent affinity of mGluRs for [Cl⁻]_out is between 60 and 80 mM, within the range of [Cl⁻]_out in synaptic clefts (102). [Cl⁻]_out interaction with mGluRs shows high cooperativity with a Hill coefficient as high as 6, suggesting a response to an exceptionally narrow change in [Cl⁻]_out (101, 102). The inotropic kainate receptors also bind Cl⁻ to stabilize their extracellular ligand binding sites, but this receptor Cl⁻ binding site structure is different than that of the mGluRs. Nevertheless, this is another example of the regulation of a GPCR by [Cl⁻]_out. [Cl⁻]_out regulates kainate receptor desensitization (79, 80).

COUPLED TRANSPORTERS REGULATED BY [Cl⁻]_out

The Cl⁻/${\text{HCO}}_3^{-}$ exchanger AE1 (also called band 3) functions as a dimer and has major roles in erythrocytes and in the kidney. Mutations in the SLC4A1 (encoding AE1) cause renal tubular acidosis and hereditary spherocytosis (2). AE1 is inhibited by stilbene derivatives, such as DBDS (4,4′-dibenzamido-2,2′-stilbenedisulfonate) and DIDS (4,4′-diisothiocyanato-2,2′- stilbenedisulfonate). A critical residue for transport by AE1 is glutamate 681 (E681). Chemical modification of E681 exposes a cryptic [Cl⁻]_out regulatory site in AE1 that binds Cl⁻ with an apparent affinity of 7–10 mM to modulate the conformation of the AE1 dimer (87). These findings were subsequently confirmed by measurements of Cl⁻ and $\text{S}{\text{O}}_4^{2-}$ fluxes in erythrocytes (46).

A second family of anion transporters is the SLC26 transporter family, in which several members function as electroneutral and electrogenic Cl⁻/${\text{HCO}}_3^{-}$ exchangers, and several function as Cl⁻ channels (56, 67). Among them are SLC26A2, which functions as a $\text{S}{\text{O}}_4^{2-}$/Cl⁻/OH⁻ exchanger (66) and has multiple roles in chondrocytes (72), and SLC26A1, which functions as a $\text{S}{\text{O}}_4^{2-}$ Cl⁻ and oxalate/Cl⁻ exchanger (110). Measurement of $\text{S}{\text{O}}_4^{2-}$/2OH⁻ exchange by SLC26A2 reveals that Cl⁻_out was required to activate SLC26A2 with apparent affinity of ~4 mM (Fig. 2A). The site displays poor selectivity for Cl⁻ with Cl⁻, Br⁻, I⁻, $\text{N}{\text{O}}_3^{-}$, and SCN⁻ having similar potency in the activation of SLC26A2 (66). $\text{S}{\text{O}}_4^{2-}$ and oxalate transport by SLC26A1 is activated by [Cl⁻]_out with an EC₅₀ of ~13 mM, with [Cl⁻]_out modifying the inhibition of SLC26A1 by acidic pH_out (110). The relatively high affinity for regulation of SLC26A2 by [Cl⁻]_out suggests that it functions in low [Cl⁻]_out environment, such as the luminal space of epithelia. There is no structural information on the SLC26 transporters extracellular and cytoplasmic domains except for the STAS domain (110), and therefore, there is no structural information on the potential SLC26 transporters [Cl⁻]_out binding site.

Fig. 2.

Signaling by [Cl⁻]_out: effect of [Cl⁻]_out on ion transporters. A: Xenopus oocytes expressing SLC26A2 and loaded with $\text{S}{\text{O}}_4^{2-}$ in Cl⁻-free solution. The oocytes were then incubated in media containing the indicated Cl⁻ to measure ${\text{SO}}_4^{=}{}_{\text{in}}$/OH⁻_out exchange. The relative rate of ${\text{SO}}_4^{=}{}_{\text{in}}$/OH⁻_out exchange is plotted as a function of [Cl⁻]_out. Results are taken from Ref. 66. B: Chinese hamster ovary (CHO) cells transfected with the indicated K⁺ channel isoforms were used to measure the whole cell K⁺ current with pipette solution containing 150 mM Cl⁻ and bath solutions containing between 5 and 150 mM Cl⁻. The figure shows the effect of [Cl⁻]_out on the magnitude of K_v1.5 and K_v2.1 current. Results are taken from Ref. 59. C: ASIC1a channel current rapidly desensitizes when active by external H⁺. Time constants of desensitization of ASIC1a channel currents are shown when recorded in solutions of varying [Cl⁻]_out, which were normalized to the time constant of desensitization of the current recorded in 143 mM [Cl⁻]_out. Results are taken from Ref. 54.

CHANNELS REGULATED BY [Cl⁻]_out

Several anion and cation channels are regulated by [Cl⁻]_out, which affects channel conductance and/or selectivity. The ionotropic P2X₇ receptor (P2X₇R) functions as a Cl⁻-permeable, nonselective cation channel (58). The P2X₇R channel pore expands upon continued stimulation to accommodate large molecules the size of ATP and NMDG⁺ (64). [Cl⁻]_out is required for cation current by P2X₇R. Moreover, although [Cl⁻]_out regulates only P2X₇R outward current of large molecules, P2X₇R pore expansion is also regulated by [Cl⁻]_out (58).

[Cl⁻]_out activates the Cl⁻ channel cystic fibrosis transmembrane conductance regulator (CFTR). Increasing [Cl⁻]_out increases CFTR current by 50–70% with a complex [Cl⁻]_out dependence (109), suggesting complicated modes of [Cl⁻]_out interaction with CFTR. Increasing [Ca²⁺]_i appears to increase the apparent affinity for [Cl⁻]_out without affecting maximal current density. Single-channel analysis showed that [Cl⁻]_out increases CFTR open probability without affecting channel conductance (109). Reduction in [Cl⁻]_out has been reported to increase CFTR Cl⁻/${\text{HCO}}_3^{-}$ permeability in a time-dependent manner (89), but this may relate to the effect of [Cl⁻]_in on CFTR (see Ref. 71 and below). More recent analysis has shown that the effect of [Cl⁻]_out requires Arg-899 in the fourth extracellular loop of CFTR and may affect interaction of ATP with the nucleotide binding domain 1 of CFTR (11). This would argue for the presence of a specific extracellular Cl⁻ binding site in CFTR, although such a site has not been observed in the structure of CFTR (120). Regulation of CFTR by [Cl⁻]_out is significant for epithelial fluid and ${\text{HCO}}_3^{-}$ secretion during which duct luminal Cl⁻ is reduced from about 120 to below 20 mM (56). At 20 mM, [Cl⁻]_out CFTR activity should be substantially reduced.

The 4-Ap-sensitive outward rectifying K⁺ channel (K_v) is one of the two major K⁺ channels in the outer hair cell (OHC) and has several kinetic components when activated by a depolarizing pulse (59). The channel is activated by [Cl⁻]_out with a shallow linear dependence on [Cl⁻]_out between 5 and 150 mM (Fig. 2B), with [Cl⁻]_out differentially affecting each of its kinetic components (59). Topical expression of various outward rectifying channels has shown that K_v2.1 and K_v1.5 are activated by [Cl⁻]_out with the properties of K_v2.1 most resembling the K⁺ channel of OHC and indicating that K_v2.1 accounts for the [Cl⁻]_out sensitivity of OHC (59). The identity and nature of the site mediating the effect of [Cl⁻]_out in K_v channels is unknown.

The acid-sensing ion channels (ASICs) and the epithelial Na⁺ channel ENaC comprise a group of homologous channels regulated by [Cl⁻]_out. The crystal structure of ASIC1a reveals Cl⁻ binding sites reside at the subunit-subunit interface at low pH. These sites are coordinated by Arg and Glu residues on the thumb domain and by a Lys residue on the palm domain of a neighboring subunit (45). The residues between the thumb and palm domain involved in Cl⁻ bindings are conserved among all ASICs and in ENaC channels. Cl⁻ interacts with these sites to regulate proton-dependent gating, desensitization kinetics, and the tachyphylaxis of ASIC1a with an apparent affinity of ~50 mM [Fig. 2C and (54)]. [Cl⁻]_out inhibits ENaC, in part, by enhancing Na⁺-self inhibition of the channel (16). Interestingly, the effects of the three [Cl⁻]_out binding sites in ENaC on channel activity are not equivalent and are reduced or eliminated at alkaline pH (16, 17). Indeed, the structure of ASIC shows destabilization of the Cl⁻ binding sites between the thumb and palm domain at high pH, but also reveals Cl⁻ interaction at the mouth of the extracellular fenestrations, just above the transmembrane domain (117). Hence, Cl⁻ binding may control access of the conducted Na⁺ to the pore.

The effects of [Cl⁻]_out discussed above demonstrate the breath and range of allosteric regulation by [Cl⁻]_out. These regulatory functions require interaction of [Cl⁻]_out with sites of vastly different affinity, some with apparent affinity in the low mM range and others with an affinity close to physiological concentration of 100 mM. At present, there is limited structural information on the nature of the extracellular Cl⁻ binding sites, and therefore, no consensus motifs or even essential residues have emerged to mediate Cl⁻ binding to extracellular sites to allow searching for similar sites in other proteins. Moreover, the physiological significance of regulation by [Cl⁻]_out is rarely demonstrated, even when mutations affecting [Cl⁻]_out sensing are known. Hopefully the availability of a knockout/knockin by CRISPR/Cas will change this to increase understanding and appreciation of the sensing and regulation by [Cl⁻]_out.

SIGNALING BY [Cl⁻]_in

The evidence for signaling by [Cl⁻]_in is quite strong. [Cl⁻]_in rapidly changes in response to cell stimulation to modulate the activity of various cellular proteins, including transcription factors, ion transporters and channels, and protein kinases. Regulation of proteins and physiological functions by [Cl⁻]_in has been described for many years but was not, and for the most part, is not viewed as a signaling by [Cl⁻]_in but rather as a form of regulation. The recognition of [Cl⁻]_in as a signaling ion is in its infancy, but should gain acceptance once signaling by [Cl⁻]_in to various pathways is understood to the extent of [Cl⁻]_in signaling in the WNK [With No lysine (K)] kinases pathway, which was established with the seminal discovery of a Cl⁻ binding site in WNK1 (78). This discovery, more than any other, is drawing attention to [Cl⁻]_in as an important signaling ion, especially in the transport field. We will discuss below regulation by [Cl⁻]_in from genes to proteins to physiological processes, emphasizing the effects of [Cl⁻]_in binding and that are independent of effect of [Cl⁻]_in on the transport of other ions. We close with the most complete studies of the role of [Cl⁻]_in in the regulation of protein kinases, including the WNK pathway, which has been established from the identification of the WNK Cl⁻ binding site to its role in mouse physiological.

GENES REGULATED BY [Cl⁻]_in

Cells respond to changes in [Cl⁻]_in by affecting expression of various Cl⁻ transporters (49, 69, 83). This likely involves sensing of [Cl⁻]_in changes by transcription factors that regulate expression of multiple Cl⁻ transporter genes. A well-demonstrated example is the change in GABA receptor subunits during neuronal maturations in vitro (96) and in vivo (97). A change in GABA receptor subunit mRNA expression during neuronal maturation causes a change in GABAergic responses (25). Embryonic and young neurons express high levels of NKCC1 mRNA and NKCC1 protein, which maintains high [Cl⁻]_in. Neuronal maturation involves upregulation of KCC2 mRNA and KCC2 protein and reduction in [Cl⁻]_in. Reduced [Cl⁻]_in signals a change in the relative expression of α3-1 and δ-containing GABA_A receptor subunits to tune the strength of GABAergic responses (96). Although strongly suggestive of gene regulation by [Cl⁻]_in, it should be noted that [Cl⁻]_in regulates the concentrations of other ions like Ca²⁺, Na⁺, K⁺, and H⁺, as well as cell volume; the contribution of these factors to gene regulation in this case needs a more careful evaluation, to separate them from direct gene regulation by [Cl⁻]_in.

Limited evidence for sensing of and direct regulation of genes by [Cl⁻]_in is available. The gram-positive, halophilic bacterium Halobacillus halophilus lives in an environment of changing salt and is dependent on Cl⁻ for growth. In this bacterium, Cl⁻ directly induces expression of the fliC gene that determines synthesis of the major subunit of the flagellum FliC, and of genes that control five stress-related proteins (85). In ciliated human nasal epithelial cells, [Cl⁻]_in regulates ciliary beating frequency and amplitude (43), although the protein affected by Cl⁻_in that transmit the change in [Cl⁻]_in to the cilia is not known at present. [Cl⁻]_in directly regulates the activity of the epithelial Na⁺ channel ENaC (see below), but also has a long-term effect by regulating the expression of αENaC subunit mRNA and protein levels (63). Evidence for gene regulation by [Cl⁻]_in was reported in a mammalian bronchial cell line in association with changes in reduced activity of CFTR that resulted in increased [Cl⁻]_in (105). Cl⁻_in appeared to differentially regulate several genes, with inhibition of the RPS27 gene expression at [Cl⁻]_in between 25 and 75 mM, and reverse the inhibition as [Cl⁻]_in was increased between 75 and 125 mM. By contrast, the GLRX5 gene, which encodes glutaredoxin-related protein 5, was activated by Cl⁻_in with an EC₅₀ of 35 mM (104). These [Cl⁻]_in concentrations are well within the physiological changes in [Cl⁻]_in observed in many cell types (56). The nature of the Cl⁻_in interacting protein(s) and the way Cl⁻_in binding is translated to gene regulation are unknown.

CHANNELS AND TRANSPORTERS REGULATED BY [Cl⁻]_in

Several studies have reported direct regulation of ENaC by [Cl⁻]_in. Early work with isolated salivary glands intralobular duct cells has shown that [Cl⁻]_in inhibits ENaC current with half maximal inhibition at 50 mM. The Cl⁻ site mediating this inhibition displays poor anion selectivity with similar inhibition observed with Cl⁻, Br⁻, and $\text{N}{\text{O}}_3^{-}$ (20). A subsequent study used excised inside-out patches to show a direct effect of [Cl⁻]_in on ENaC current, again with poor anionic selectivity (6). However, in both studies only high anion concentrations were used, and potential differences in potency among the anions in inhibition of ENaC were not determined. Single channel analysis revealed that [Cl⁻]_in reduced both ENaC open probability and channel conductance with most of the effects observed at [Cl⁻]_in between 4 and 60 mM (34). Importantly, truncation analysis localized the effect of [Cl⁻]_in on ENaC to the cytoplasmic COOH terminus of the α and β ENaC subunits (6).

Another group of channels regulated by [Cl⁻]_in are K⁺ channels. The K⁺ current in astrocytes mediated by K_v channels is activated when [Cl⁻]_in is increased from 40 to 80 mM, which affects channel activation kinetics (8). Similar regulation by [Cl⁻]_in was observed with the rapidly inactivating K⁺ channel isoforms K_v1.4 and K_v4.2 (7). The mechanism by which [Cl⁻]_in affects K_v channels function is not known. However, regulation of K_v1.4 and K_v4.2 by [Cl⁻]_in could be observed only after disruption of the cytoskeleton, suggesting that a cytoplasmic channel domain that is shielded by the cytoskeleton senses and mediates channel gating by [Cl⁻]_in (7). The Na⁺- and Ca²⁺-activated SLO channels display a unique form of regulation by [Cl⁻]_in. Regulation by [Cl⁻]_in was discovered when the properties of the Caenorhabditis elegans SLO-2 channel were examined. SLO-2, but not SLO-1 and SLO-3, is regulated by [Cl⁻]_in and regulation by [Cl⁻]_in and Ca²⁺ is synergistic (118). This is illustrated in Fig. 3, showing that [Cl⁻]_in left-shifts the voltage dependence of SLO-2, that [Cl⁻]_in-dependent activation is modulated by [Ca²⁺]_i and that Ca²⁺-dependent activation is modulated by [Cl⁻]_in. The mammalian SLO channels Slick and SLAK (KCNT1 and KCNT2) are also regulated by Na⁺ and [Cl⁻]_in with half maximal activation at [Cl⁻]_in of 35 mM (10, 75), which is of physiological relevance.

Fig. 3.

Signaling by [Cl⁻]_in: effect of [Cl⁻]_in on Caenorhabditis elegans and mouse K⁺ channel activity. [Cl⁻]_in modulates C. elegans SLO-2 K⁺ channel properties. A: shows C. elegans SLO-2 channel conductance measured at constant 100 µM Ca²⁺ and at [Cl⁻]_in of 0 (○) 10 (■), 50 (Δ), and 100 mM (●). B shows C. elegans SLO-2 channel conductance measured at constant Cl⁻ of 50 mM and at [Ca²⁺]_i of 0 (○), 25 (■), 100 (Δ), and 250 μM (●). C and D show that mouse SLO-1 is insensitive to [Cl⁻]_in compared with SLO-2. In C and D, channel conductance was measured at constant [Ca²⁺]_i of 250 µM and at [Cl⁻]_in of 0 (●), 10 (Δ), 50 (■), and 100 mM (○). Results are taken from Ref. 118.

Cl⁻ has been well described as a cosubstrate for several neurotransmitter transporters, and evidence for regulation of the transporters by [Cl⁻]_in is also available (28). More recently, a whole endosomal recording of glutamate-mediated current by the vesicular glutamate transporters revealed allosteric activation of glutamate transport by [Cl⁻]_in that took place mostly between 10 and 50 mM [Cl⁻]_in. An arginine in transmembrane domain 4 was required for both the transport and the allosteric regulation by [Cl⁻]_in (14).

A well-understood example of [Cl⁻]_in-mediated regulation of a transporter is the Na⁺-${\text{HCO}}_3^{-}$ cotransporter family, in particular, the ubiquitous NBCe1-B isoform (90) that has a prominent role in epithelial ${\text{HCO}}_3^{-}$ secretion (56). NBCe1-B (94, 116), as well as several other transporters (37, 51, 74, 116), are regulated by the scaffolding protein IP₃ receptor binding protein released with IP₃ (IRBIT). IRBIT binds to and inhibits the IP₃ receptors (5), and once released from the IP₃Rs by elevation of cytoplasmic IP₃, it mediates the synergistic activation of the transporters by Ca²⁺ and cAMP signaling pathways (74). The basal activity of NBCe1-B is determined by its NH₂ terminus autoinhibitory domain (AID) and is minimally inhibited by [Cl⁻]_in. IRBIT interacts with the AID to markedly activate NBCe1-B and exposes a cryptic Cl⁻_in sensing site with an apparent affinity of ~10 mM that potently inhibits the NBCe1-B activity. Inhibition of the NBCe1-B by [Cl⁻]_in occurs with exquisite selectivity and is not affected by Br⁻, I⁻, $\text{N}{\text{O}}_3^{-}$, or SCN⁻ (90).

Although the structure and nature of the NBCe1-B [Cl⁻]_in sensing site is not known, it is mediated by two conserved GXXXP motifs. The first motif is ³²GXXXP³⁶ in the NBCe1-B AID, and the second motif is ¹⁹⁴GXXXP¹⁹⁸. GXXXP motifs function as Cl⁻ sensing sites (22) and participate in Cl⁻ sensing by SLC26A2 (66), NBCe1-B (SLC4A4), and NBCe2-C (SLC4A5) (90). [Cl⁻]_in sensing by the NBCe1-B GXXXP motifs is regulated by protein kinases and phosphatases acting on specific serine residues (Fig. 4). The ³²GXXXP³⁶ motif is regulated by Ser-65, which is phosphorylated by the SPAK kinase and dephosphorylated by the phosphatase PP1. The ¹⁹⁴GXXXP¹⁹⁸ motif is regulated by Ser-12, which is phosphorylated by CaMKII kinase and dephosphorylated by the phosphatase calcineurin [see Fig. 4 and (103)]. IRBIT recruits the kinases and phosphatases to NBCe1-B to determine their phosphorylation state. When Ser-12 and Ser-65 are phosphorylated, the [Cl⁻]_in sensing sites are cryptic and their dephosphorylation uncovers the [Cl⁻]_in sensing sites to allow regulation of NBCe1-B by [Cl⁻]_in (103).

Fig. 4.

Signaling by [Cl⁻]_in: regulation of the Na⁺/${\text{HCO}}_3^{-}$ cotransporter NBCe1-B by [Cl⁻]_in is modulated by kinases and phosphatases. Na⁺-${\text{HCO}}_3^{-}$ cotransport by NBCe1-B was measured as the outward current triggered by the addition of ${\text{HCO}}_3^{-}$ to human embryonic kidney (HEK) cells transfected with NBCe1-B and IRBIT (binding protein released with IP₃). IRBIT activates NBCe1-B and exposes its cryptic [Cl⁻]_in sensitive sites ³²GXXXP³⁶ and ¹⁹⁴GXXXP¹⁹⁸ (103). Current was initiated by addition of ${\text{HCO}}_3^{-}$ to the bath and [Cl⁻]_in was set by changes of Cl⁻ concentration in the patch pipette solution. NBCe1-B current due to Na⁺-2${\text{HCO}}_3^{-}$ cotransport in the absence of IRBIT (○) shows low sensitivity to [Cl⁻]_in. IRBIT activates NBCe1-B and exposes regulatory [Cl⁻]_in sites (●). When protein phosphatase 1 (PP1) is activated, regulation of NBCe1-B by [Cl⁻]_in sensing through the ³²GXXXP³⁶ site (●) is eliminated, while active constitutively active calcineurin (CA-CaN) eliminates [Cl⁻]_in sensing by the ¹⁹⁴GXXXP¹⁹⁸ site (●). Activation of both PP1 and CA-CaN eliminates inhibition of NBCe1-B by [Cl^-]_in (●). Results are taken from Ref. 103.

REGULATION OF PROTEIN KINASES BY [Cl⁻]_in

One of the best understood form of regulation by [Cl⁻]_in is that of various protein kinases. An indirect evidence for activation of Jun kinase 2 (JNK2) by [Cl⁻]_in is suggested by the finding that reduction in [Cl⁻]_in prevents JNK2-dependent induction of the Na⁺/K⁺ pump γ subunit in response to hypotonicity (12). In addition, upregulation of the cyclin-dependent kinase inhibitor p21, which is mediated by p38 and JNK kinases, requires an increase in [Cl⁻]_in (68). Regulation of JNK by [Cl⁻]_in may be cell specific since in macrophages low [Cl⁻]_in activates the p38 and the JNK kinases (111). [Cl⁻]_in strongly inhibits the activity of purified apoptosis signal-regulating kinase 1 (ASK1) and mitogen-activated kinase 6 (MEK6) (78), but the physiological function of this inhibition is not clear at this time. Increasing [Cl⁻]_in between 50 and 70 mM, prominently phosphorylates and activates purified glucocorticoid-inducible protein kinase 1 (SGK1) in airway epithelial cells, to mediate the response to LPS-induced reduction in cellular cAMP (119). Although it is not known how [Cl⁻]_in regulates SGK1 and whether the regulation is by direct effect of [Cl⁻]_in on the kinase, the physiological significance of the regulation is evident from the finding that it occurs at the physiological range [Cl⁻]_in and that SGK1 is markedly activated in patients with bronchiectasis, a chronic airway inflammatory disease (119).

One conclusive piece of evidence for the regulation of protein kinases by [Cl⁻]_in with profound physiological consequences is demonstrated with the WNK kinases (41, 91). The WNK kinases were discovered as members of the MAP kinase superfamily (113), with the WNK subfamily consisting of four members with conserved kinase domain and variable NH₂ and COOH termini (41). Interest in the WNKs increased greatly with the discovery that mutations in WNK1 and WNK4 cause hypertension in humans (108). The primary action of the WNKs is determining ion transporters surface expression (91) by regulating their endocytosis (32, 42). The WNK kinases can act by directly phosphorylating several transporters, or indirectly by phosphorylating the oxidative stress-responsive kinase 1 and/or the STE20/SPS1-related proline/alanine-rich kinase (SPAK) (19), which phosphorylate the transporters to control their plasma membrane level (40).

Regulation of the WNK/SPAK pathway by [Cl⁻]_in was established long before direct regulation of the kinases by [Cl⁻]_in and identification of their Cl⁻ binding site was established. For instance, the Na⁺/K⁺/2Cl⁻ cotransporter NKCC1 is phosphorylated and activated by SPAK only at low [Cl⁻]_in (21), suggesting that SPAK is inhibited by [Cl⁻]_in. Similar findings were reported for the NaCl cotransporter NCC (70). The phosphorylation of SPAK by WNK3 to regulate NKCC2 function requires low [Cl⁻]_in, suggesting that WNK3 is inhibited by [Cl⁻]_in (81). Phosphorylation of CFTR by WNK1 changes its selectivity from a Cl⁻-selective to a Cl⁻ and ${\text{HCO}}_3^{-}$ conducting channel, a function observed only at low [Cl⁻]_in (71). Interest in the regulation of the WNKs by [Cl⁻]_in has dramatically changed with the discovery that autophosphorylation, stability (Fig. 5A), and activity (Fig. 5B) of WNK1 are regulated by Cl⁻, and with the identification of the Cl⁻ binding site in the kinase domain (Fig. 5C) (78). This discovery initiated a number of studies demonstrating that WNK3 and WNK4 are also inhibited by [Cl⁻]_in, with WNK4 showing the highest sensitivity to [Cl⁻]_in [(98) and Fig. 5D)], highlighting the importance of [Cl⁻]_in in the regulatory function of the WNKs (91).

Fig. 5.

Signaling by [Cl⁻]_in: effect of [Cl⁻]_in on the activity of the WNK [With No lysine (K)] kinases. A: effect of Cl⁻, Br⁻, and I⁻ concentration on WNK1 kinase stability evaluated from measurement of a change in melting temperature (ΔT_m). B: WNK1 kinase activity as a function of Cl⁻ concentration. C: cartoon of WNK1 halide binding site. α helices are pink, β sheets are cyan, and loops are gray. The Cl⁻ position is shown in green. Inset: 3/10 helix and hydrogen bonding distances are shown. Results in A–C are taken from Ref. 78. D: effects of Cl⁻ concentration on in WNK kinase activity was measured using the WNKs kinase domain (WNKxKD) and STE20/SPS1-related proline/alanine-rich kinase (SPAK) as a substrate. The WNK4KD activity is more sensitive to inhibition by Cl⁻ than WNK1KD WNK3KD. Results are taken from Ref. 98.

The physiological in vivo significance was recently established in a notable study, in which a Cl⁻-insensitive mutant WNK4 replaced the native WNK4 in mice (15). In mammals, a change in systemic K⁺ is translated to an increase in [Cl⁻]_in, which, in turn, results in inhibition of the WNK4 activity. The altered WNK4 activity changes regulation of KCC and perhaps NKCC2 and, consequently, K⁺ excretion, resulting in hyperkalemic hypertension (91, 99). Recapitulation in mice expressing the Cl⁻-insensitive WNK4, the human hypertensive and hyperkalemic disease pseudohypoaldosteronism type II, and the observation that the WNK4 mutant mice lost the renal response to maneuvers that change Cl⁻ homeostasis (15) conclusively establish WNK4 as a physiological [Cl⁻]_in sensor and [Cl⁻]_in as a signaling ion.

CONCLUSIONS AND CHALLENGES

The evidence discussed in the examples above and summarized in Fig. 6 provides a compelling case for Cl⁻ as a signaling ion used by cells to regulate key physiologic functions. Cl⁻ gradients are generated across the plasma and intracellular organelles membranes; the Cl⁻ gradients are dissipated rapidly by cell stimulation to change [Cl⁻]_in and even generate [Cl⁻]_in oscillations; and changes in cellular and systemic Cl⁻ are sensed by a variety of transcription factors to regulate many genes, protein kinases, and ion transporters, to alter cellular activity. These properties fulfill all of the key criteria for an ion to be a bona fide signaling ion. Increasing recognition of Cl⁻ as a signaling ion and its physiological consequences, and not only as a transported ion that affects the concentration of other ions and cell homeostasis through changes in cellular ionic contents poses a major challenge.

Fig. 6.

A model summarizing the signaling functions of Cl⁻. Intracellular Cl⁻ concentration tightly controls by Cl⁻ influx and efflux transporters that keep [Cl⁻]_in above electrochemical potential. Cell stimulation often rapidly reduces [Cl⁻]_in or causes [Cl⁻]_in oscillations. The changes in [Cl⁻]_in are translated to cellular responses by direct binding of Cl⁻ to a variety of proteins and changes in their activity. The best understood form of signaling by [Cl⁻]_in is regulation of protein kinases and phosphatases. However, regulation of several transporters by binding of Cl⁻ to them, including transporters that do not transport Cl⁻, is well documented. By regulation of K⁺ channels, [Cl⁻]_in indirectly affects the membrane potential, in addition to its direct effect on the membrane potential as a conductive ion. [Cl⁻]_in also has long-term effects by binding and activating transcription factors to regulate gene expression. Although [Cl⁻]_in binding to the various proteins and regulation of their activity are well documented, the [Cl⁻]_in binding sites are not always known, and the physiological roles of signaling by [Cl⁻]_in in animal studies is mostly not documented. The exception is the role of [Cl⁻]_in in the WNK/SPAK pathway. Regulation of enzymes, ion channels and transporters, and signaling by G protein-coupled receptors by [Cl⁻]_out is also well documented. However, this is the least understood form of signaling by Cl⁻, and its physiological role is rarely documented or understood.

A significant difficulty in studying the signaling function of [Cl⁻]_in is separating between the physiological and regulatory function of [Cl⁻]_in. This is exacerbated by the general lack of structural information on the regulatory Cl⁻ binding sites. Because the affinity for Cl⁻ is at the mM range and the presence of [Cl⁻]_in binding sites in cytoplasmic domains that tend to be poorly ordered, the Cl⁻ electron density may not be resolved very well in structural studies. Yet, the case of the WNK kinases illustrates how valuable such information is. With knowledge of the WNK1 Cl⁻ binding site, it became possible to examine the importance of regulation of WNK1, WNK3, and WNK4 by Cl⁻ in epithelial fluid and electrolyte homeostasis, both in model expression system and in vivo. Availability of the structure of the increasing number of regulatory Cl⁻ binding sites should allow better understanding of signaling by Cl⁻ and its physiological relevance.

Second messengers like Ca²⁺ and cAMP act in cellular microdomains that are generated by compartmentalization of signaling protein complexes in discreet cellular compartments (29, 88). This is also the case for [Cl⁻]_in. Specific Cl⁻ transporters localize to specific membranes, such as the luminal and basolateral membranes of epithelia, and membrane subdomains, such as lipid rafts. In addition, diverse Cl⁻ channels and transporters are present in different organelles. In principal, such compartmentalization allows generation of [Cl⁻]_in signals in specific cellular compartments to exert a selective, rather than a global, effect. Detection and demonstration of localized effects of Ca²⁺ and cAMP are greatly aided by the availability of genetically targeted protein sensors (4). Genetically encoded fluorescent protein sensors are now available for Cl⁻ (84), and their targeting to various cellular compartments should generate important information on local [Cl⁻]_in signals and their function.

GRANTS

This work was funded by Intramural NIH Grant DE000735-9 from the National Institute of Dental and Craniofacial Research.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

E.O. prepared figures; B.P.L., L.V., and S.M. drafted manuscript; B.P.L., L.V., E.O., and S.M. edited and revised manuscript; B.P.L., L.V., E.O., and S.M. approved final version of manuscript.