Molecular physiology of cation-coupled Cl− cotransport: the SLC12 family
The electroneutral cation-chloride-coupled cotransporter gene family (SLC12) was identified initially at the molecular level in fish and then in mammals. This nine-member gene family encompasses two major branches, one including two bumetanide-sensitive Na+-K+-2Cl− cotransporters and the thiazide-sensitive Na+:Cl− cotransporter. Two of the genes in this branch (SLC12A1 and SLC12A3), exhibit kidney-specific expression and function in renal salt reabsorption, whereas the third gene (SLC12A2) is expressed ubiquitously and plays a key role in epithelial salt secretion and cell volume regulation. The functional characterization of both alternatively-spliced mammalian Na+-K+-2Cl− cotransporter isoforms and orthologs from distantly related species has generated important structure-function data. The second branch includes four genes (SLC12A4–7) encoding electroneutral K+-Cl− cotransporters. The relative expression level of the neuron-specific SLC12A5 and the Na+-K+-2Cl− cotransporter SLC12A2 appears to determine whether neurons respond to GABA with a depolarizing, excitatory response or with a hyperpolarizing, inhibitory response. The four K+-Cl− cotransporter genes are co-expressed to varying degrees in most tissues, with further roles in cell volume regulation, transepithelial salt transport, hearing, and function of the peripheral nervous system. The transported substrates of the remaining two SLC12 family members, SLC12A8 and SLC12A9, are as yet unknown. Inactivating mutations in three members of the SLC12 gene family result in Mendelian disease; Bartter syndrome type I in the case of SLC12A1, Gitelman syndrome for SLC12A3, and peripheral neuropathy in the case of SLC12A6. In addition, knockout mice for many members of this family have generated important new information regarding their respective physiological roles.
KeywordsThiazideBumetanideBartter's syndromeGitelman's syndromeCorpus callosumPeripheral neuropathyEpilepsyGABA (γ-aminobutyric acid)
Introduction: from fish to cotransporters
SLC12—the electroneutral cation-Cl coupled cotransporter family (TAL thick ascending limb of the loop of Henle, ACCPN peripheral neuropathy with or without agenesis of the corpus callosum
Human gene name
Transport type/coupling ions*
Tissue distribution, cellular/subcellular expression
Link to disease#
Human gene locus
Sequence accession ID
Splice variants and their specific features
Na+, K+, Cl−
Kidney-specific: apical membrane of the thick ascending limb
G: Bartter's syndrome type I
1) Isoforms A, B, F. Differ in ion affinities and exhibit axial expression along TAL. 2) A shorter C-terminal isoform encodes a Na+:Cl− cotransporter.
Na+, K+, Cl−
Ubiquitous: basolateral membrane of epithelial cells. Also in non-epithelial cells
A shorter C-terminal isoforms due to absence of exon 21
Kidney-specific: apical membrane of the distal convoluted tubule; bone?
G: Gitelman's syndrome
Two N-terminal variants, generated by different first coding exons
Widespread, minimal in brain
The Na+-K+-2Cl− cotransporters
Two genes encoding Na+-K+-2Cl− cotransporter isoforms have been identified. The kidney-specific SLC12A1 [34, 102] and the ubiquitously expressed SLC12A2 [24, 147]. Both isoforms are inhibited by loop diuretics such as bumetanide and furosemide.
SLC12A1 (NKCC2; BSC1)
SLC12A2 (NKCC1; BSC2)
The NKCC1 protein primary structure was elucidated by cloning the corresponding cDNAs from shark rectal gland , mouse inner medullary collecting duct cells , and human colonic epithelial cells . As shown in Fig. 2 for NKCC2, NKCC1 is a 1212-amino acid residue protein exhibiting 12 TM that are flanked by hydrophilic NH2 and COOH-terminal domains. This topology has been confirmed experimentally . An elegant series of comparative studies using human and shark NKCC1 have indicated that TM2 plays a major role in cation affinity, whereas TM4 and TM7 affect anion affinity [33, 57, 58, 59]. The NH2-terminal domain contains a RVXFXD motif that binds protein phosphatase-1  in addition to a (R/K)FX(V/I) motif required for binding stress-related protein kinases . NKCC1 appears to form homodimers, although the functional role of cotransporter dimerization is as yet unclear . One putative splicing isoform has been reported, exhibiting a shorter COOH-terminal domain due to the absence of exon 21 . In the nervous system, NKCC1 is present in the apical membrane of choroids plexus, in neurons, oligodendrocytes and in dorsal root ganglion neurons [110, 112], where it is critical for the appropriate response to GABA in these cells, and thus for normal sensory perception . To date no human disease has been liked to mutations in SLC12A2. However, the targeted disruption of this gene  results in several interesting phenotypes: (i) an inner-ear dysfunction characterized by deafness and shaker/waltzer phenotype indicates the major role of this transporter in the endolymph secretion; (ii) complete infertility due to a defect in spermatocyte production ; (iii) reduction in the production of saliva , and (iv) sensory perception abnormalities . Hypotension in SLC12A2 null mice is not observed consistently [31, 99].
The Na+-Cl− cotransporter
One gene, SLC12A3, encoding an Na+-Cl− cotransporter has been identified. The Na+-Cl− cotransporter represents the major target site for clinically useful benzothiadiazine (or thiazide)-type diuretics like chlorothiazide (Diuril) used in the treatment of hypertension, states of extracellular fluid volume overload, and renal stone disease.
SLC12A3 (NCC; TSC)
Thiazide diuretics were introduced into clinical medicine in 1957 , representing a major breakthrough in the treatment of hypertension. The nephron segment and mechanism of action of thiazide diuretics, however, remained unknown for a number of years until Kunau and coworkers showed in 1975 clear evidence that thiazides acted in the distal nephron giving rise to an increase in urinary Cl− excretion . Subsequently, Stokes and coworkers showed that thiazides inhibited an electrically silent Na+-Cl− cotransport in winter flounder urinary bladder , and, later, Ellison and coworkers presented evidence for a luminal Na+-Cl− cotransport mechanism in the rat distal convoluted tubule that was specifically inhibited by thiazides . Beamount and coworkers demonstrated that tracer [3H]metolazone binds to high-affinity receptor—it was proposed to be the cotransporter—exclusively in the renal superficial cortex . Taken together, these studies allowed the conclusion that the major salt reabsorption pathway in the apical membrane of the distal convoluted tubule is the thiazide-sensitive Na+-Cl− cotransporter (Figs. 3 and 4).
Molecular identification of the SLC12 family of cotransporters began with the cloning of an thiazide-sensitive, electroneutral Na+-Cl− cotransporter from the urinary bladder of the winter flounder, Pseudopleuronectes americanus . Subsequently, SLC12A3 homologues were cloned from rat , mouse , rabbit  and human [88, 124]. The predominant expression of Na+-Cl− cotransporter mRNA is in kidney [49, 88], although transcripts have also been identified in an osteoblast-like cell line . Transcript and protein expression in the mammalian kidney is predominantly in the initial distal convoluted tubule (DCT) but variably extends into the connecting segment in all species studied [4, 10, 81, 98, 111] except rabbit . The NCC cotransporter is thus expressed in kidney segments known to transport thiazide-sensitive Na+ and Cl− in a coupled and electroneutral manner [19, 29, 139]. SLC12A3 expression in X. laevis oocytes gives rise to thiazide-sensitive Cl−-dependent Na+ uptake with a 1:1 stoichiometry [35, 92]. A transport model has been proposed with a random order of ion binding where each ion facilitates the binding of the counter-ion .
The human NCC cotransporter is a 1021-amino acid residue glycoprotein protein  that has a general topology similar to NKCC1 and NKCC2 (see Fig. 2) with a central core of 12 TM. The two N-linked glycosylation sites in the large extracellular loop between TM 7 and 8 [49, 88] are essential for efficient function and surface expression of the cotransporter and elimination of glycosylation allows much greater access of thiazide diuretics to their binding site . Recent studies have established that regulation of NCC cotransporter abundance in the DCT plays important roles in water and salt homeostasis: (i) Na+-Cl− cotransporter abundance is increased with deamino-Cys1-d-Arg8-vasopressin (dDAVP) and associated with vasopressin escape occurring in the syndrome of inappropriate antidiuresis [26, 27]; (ii) the cotransporter co-localizes with 11-β-hydroxysteroid dehydrogenase-2 (11-β-HSD2) and mineralocorticoid receptors [13, 140]; (iii) its abundance is up-regulated by aldosterone  and with a low-sodium diet ; (iv) cotransporter abundance is increased by estradiol , loop diuretics  and in the obese Zucker rat model ; (v) Na+-Cl− cotransporter abundance is increased with modulation of kidney size associated with prenatal programming using a low-protein diet ; (vi) cotransporter abundance is decreased in chronic hypokalemia , following ureteral obstruction  and by spironolactone.
Pseudohypoaldosteronism type II (PHAII; Fig. 4; OMIM 145260) is an autosomal dominant disease caused by mutations in one of two serine-threonine kinases (WNK1 and WNK4; ). Individuals with PHAII exhibit hypertension that is Cl−-dependent and corrected by thiazide diuretics, suggesting a role for the Na+-Cl− cotransporter in this salt-dependent form of hypertension. Recently, WNK4 has been shown to negatively regulate the surface expression, and therefore transport function, of the Na+-Cl− cotransporter expressed in X. laevis oocytes [145, 148]. Loss of this negative regulation by mutations in WNK4 protein is suggested to be the molecular pathogenesis of this form of hypertension . Because modulation of Na+-Cl− cotransporter function in Gitelman's syndrome and PHAII gives rise to either hypotension or hypertension, respectively, SLC12A3 is one of the genes involved in defining the blood pressure .
The K+-Cl− cotransporters
The K+-Cl− cotransporters (KCCs) were first described as a swelling-activated K+ efflux pathway in red cells [25, 74]; the cellular physiology of erythroid K+-Cl− cotransport merits a brief review here, given the central importance of two decades of work [72, 73] on red cells in guiding characterization of the recombinant KCCs. Red cell K+-Cl− cotransport is strongly activated by cell swelling , and appears to function in regulatory volume decrease. The cotransport of K+ and Cl− in red cells is interdependent, with a 1:1 stoichiometry and low affinity constants for both ions ; under most physiological conditions erythroid KCCs mediate K+-Cl− efflux, whereas neuronal KCCs may mediate both efflux and influx (see KCC2 section). Increased activity of erythroid K+-Cl− cotransport has been implicated in dehydration and hence sickling of red cells in sickle cell anemia and other hemoglobinopathies [72, 73]. The main physiological activators of erythroid K+-Cl− cotransport include cell swelling, low pH, high PO2, and urea [72, 73]. Pharmacological activation of red cell K+-Cl− cotransport can also be achieved with the thiol-alkylating agent N-ethyl maleimide (NEM) and by oxidizing agents such as hydroxylamine, H2O2, and diamide; these reagents are considered to act on thiol groups present in upstream regulatory proteins (kinases, phosphatases, etc.; ). The activation of K+-Cl− cotransport by either cell swelling, NEM, or urea is generally blocked by protein phosphatase inhibitors such as vanadate, calyculin-A, and/or okadaic acid [18, 72]. Recent data suggest the involvement of membrane-bound protein phosphatase-1 (PP1) and PP2A .
KCC1 was cloned by the identification of expressed sequence tags (ESTs) homologous to the other cation-chloride cotransporters; full-length cDNAs have been reported for rat , rabbit , human , pig , and mouse . Northern blot analysis has shown KCC1 to be expressed ubiquitously, suggestive of a housekeeping role in cell volume regulation. Very little tissue localization data has been published, and in the absence of a knockout mouse nothing is known about the global physiological role of this cotransporter. In adult rat brain there is widespread, low-level expression of KCC1 , with more abundant transcript in choroid plexus, olfactory bulb, and cerebellum . Embryonic expression of KCC1 is almost exclusive to the choroid plexus .
The predicted KCC1 protein is 1085 amino acids long, with the highest identity (75%) to KCC3. Expression in Xenopus oocytes [89, 95, 129] and HEK293 [39, 40, 52] cells reveals the functional characteristics expected of a K+-Cl− cotransporter. In particular, transport activity is not detected under isotonic conditions and swelling-induced activation is blocked by phosphatase inhibition [40, 89, 129]. Of interest, baseline Na+-K+-2Cl− cotransport activity is elevated and intracellular [Cl−] reduced in KCC1-HEK293 cells, compared with the parental cell line ; NKCC1 is thought to be activated by decreases in intracellular [Cl−], via phosphorylation by an as yet unidentified Cl−-sensitive kinase . KCC1 and KCC3 share several residues within TM2, a region implicated in cation affinity of Na+-K+-2Cl− cotransport [58, 108]. The functional correlate of this similarity is that the cation affinities of these two KCCs are much lower than those of KCC2 and KCC4 [89, 101, 125]. The Km values of KCC1 for K+ and Cl− are 25.5 and 17.5 mM respectively, with an anion preference of Cl−>SCN−=Br−>>PO43− . As discussed elsewhere , these characteristics differ from red cell K+-Cl− cotransport, which has a much lower K+ affinity and seems to prefer Br− to Cl−. Therefore, although KCC1 is evidently expressed in red cells from several species [75, 129], it seems unlikely that this is the sole erythroid KCC.
The KCC2 protein is expressed at the cell membrane of both neuronal somata and dendrites [44, 143], and is colocalized at inhibitory synapses with subunits of the GABAA receptors  and associated proteins . Robust expression of KCC2 is already detectable at embryonic day (E)12.5 in the ventral horns of the developing spinal cord [55, 79]. Spinal motoneurons at E18.5 respond to GABA and glycine with a depolarizing and excitatory response, which is absent in KCC2-null mice . The phenotypic correlates in KCC2-null mice include both a spastic posture and immediate postnatal death from apnea, the latter attributed to a defect in the regulation of respiratory motoneurons . The developmental pattern of KCC2 expression appears to correlate with neuronal maturation in several regions of the CNS [79, 90]. GABA was shown recently to induce expression of the KCC2 protein, thus limiting its brief window of neurotrophic effect . In contrast, brain-derived neurotrophic factor (BDNF) and neurotrophin-4 dramatically decrease KCC2 expression, thus amplifying neuronal excitability in conditions associated with up-regulation of BDNF, such as epilepsy .
A feature unique to KCC2 is an inserted protein domain of ~100 amino acids, rich in proline residues and negatively charged residues, near the end of the cytoplasmic COOH-terminus [103, 125]. Functional expression in both HEK293 cells and Xenopus oocytes reveals that KCC2 is unique in mediating isotonic K+-Cl− cotransport [125, 128], and it is tempting to speculate that this property is conferred by this cytoplasmic KCC2-specific domain. The modest increase in KCC2 activity induced by cell swelling is blocked by phosphatase inhibition; isotonic activity is not affected, suggesting a distinct mechanism for constitutive activation of KCC2 . Mutagenesis studies suggest that a COOH-terminal tyrosine predicted to be a substrate for tyrosine kinases is required for activity of rat KCC2 in Xenopus oocytes . Of note, although tyrosine kinase inhibitors do not affect KCC2 in oocytes, the same reagents inhibit K+-Cl− cotransport in primary neuronal culture .
Kinetic characterization of rat and human KCC2 indicates a much higher cation affinity than the other three KCCs, with Km values of ~5.2 and 9.2 mM, respectively [101, 125]. There is much greater discrepancy in the reported Cl− affinity of human KCC2 and rat KCC2, with Km values of >50 mM and 6.8 mM, respectively. The lower Km is closer to the intracellular [Cl−] activity of mature neurons that express KCC2 . The high ion affinities of KCC2 befit a buffer of external K+ and internal Cl−, as was first noted by Payne . As extracellular K+ increases to 10–12 mM during neuronal activity, the range wherein KCC2 is highly active, the driving force for net K+-Cl− cotransport will switch from efflux to influx. In this regard, reversibility of K+-Cl− cotransport mediated by neuronal KCCs has been verified experimentally [21, 60, 61]. Bi-directional transport via neuronal KCCs may explain activity-dependent disinhibition , whereby repetitive activation of GABA receptors results in increased extracellular [K+], an increase in neuronal [Cl−] due to K+-Cl− influx, and a reduction in the inhibitory GABA effect .
The role of KCC2 in GABA-mediated neuronal inhibition and activity-dependent disinhibition suggest its involvement in human epilepsy. Indeed, KCC2 knockout mice with a >90% reduction in KCC2 expression (versus a complete deficiency; ) manifest a severe seizure disorder and early postnatal mortality . It is thus conceivable that genetic variation in human SLC12A5 might affect epilepsy and the neuronal response to injury, particularly since GABA has been shown to have excitatory effects after neuronal trauma . In addition, genetic variation in the negative transcriptional response of KCC2 to neurotrophins may have a role in sustained epileptic activity .
The KCC3 transcript is abundant in muscle, brain, spinal cord, kidney, heart, pancreas, and placenta [51, 95, 114]. A differentiating feature of KCC3 is that the SLC12A6 gene encoding this transporter has two separate first coding exons . This results in the expression of two separate KCC3 isoforms, with different NH2-terminal ends; we have designated these isoforms KCC3a and KCC3b. KCC3a expression is more widespread than that of KCC3b, which is particularly abundant in kidney . The predicted KCC3a protein is longer by 50 amino acids , and contains several potential phosphorylation sites for protein kinase C, that are not present in the KCC3b-unique NH2-terminus . The functional consequences of this variation are as yet unclear.
Expression of KCC3 in the brain and spinal cord was of particular interest, in light of the genetic linkage between the region encompassing the SLC12A6 gene on chromosome 15q14 and several neurological syndromes . Genomic characterization of SLC12A6 was thus pursued, revealing a 26-exon gene spanning ~160 kb. Whereas patients with family juvenile myoclonic epilepsy and rolandic epilepsy linked to 15q14 were not found to carry coding sequence mutations in SLC12A6 , four loss-of-function mutations were recently characterized in kindreds with peripheral neuropathy with or without agenesis of the corpus callosum (ACCPN, OMIM 21800; ). This syndrome is particularly common in the Saguenay-St. Lawrence region of Quebec, Canada, due to a founder effect. The complete syndrome encompasses severe peripheral neuropathy, agenesis of the corpus callosum, mental retardation, psychosis, and a progressive course suggestive of a neurodegenerative process. The Quebec founder effect mutation is a single base deletion at the end of coding exon 18, resulting in an mRNA splicing error and a premature stop codon. The predicted protein is missing the COOH-terminal 338 residues and is non-functional when expressed in Xenopus oocytes . Of note, unlike loss-of-function mutations in NCC  and other transport proteins, deletion of the COOH-terminus of KCC3 does not affect glycosylation, homo-dimerization, or expression at the surface of Xenopus oocytes .
Despite the clear genetic and functional evidence implicating KCC3 in ACCPN, the pathophysiology of this neurological deficit is not yet clear. Although there is heavy oligodendrocytic expression of the KCC3 protein in CNS white matter tracts, including the corpus callosum , KCC3 knockout mice have histologically normal brains with intact corpora callosa . However, agenesis/dysgenesis of the corpus callosum is a partially penetrant phenotype in ACCPN, even within affected families. Expression of KCC3 in large neurons  and the human NT2-N neuronal cell line  has also been reported, compatible perhaps with the neuropsychiatric manifestations of ACCPN. At the cellular level, KCC3 has been implicated in the response of cells to both vascular endothelial growth factor  and insulin-like growth factor-1 ; loss of KCC3 may thus affect the proliferation and/or survival of neurons or oligodendrocytes. A major paradox is the severe peripheral neuropathy associated with loss of KCC3 in both mice and man , despite the minimal expression in peripheral nervous tissue .
Much like KCC1, KCC4 is widely expressed. A differentiating feature is however that KCC4 expression within specific tissues is more discrete than that of KCC1. Although transcript is undetectable in Northern analyses of human or mouse brain, there is expression of KCC4 within select CNS neurons . Expression is particularly robust in the kidney, where the KCC4 protein has been localized at the basolateral membrane of type-A intercalated cells and proximal tubule . There is a burgeoning literature on renal K+-Cl− cotransport, and swelling-activated Cl− exit mediated by KCC4 and other KCCs is thought to play a role in transepithelial transport of Na+-Cl− by the proximal tubule . Mice deficient in KCC4 manifest a renal tubular acidosis attributed to defects in acid secretion by intercalated cells . Like some forms of human hereditary distal renal tubular acidosis, these KCC4−/− mice are also profoundly deaf. Careful histological analysis reveals degeneration of outer hair cells, presumably due to impaired K+ uptake by the supporting Deiter's cells .
Functional characterization in Xenopus oocytes reveals Km values for K+ and Cl− of 17.5 and 16.2 mM ; the higher K+ affinity is in keeping perhaps with an analogous role to that of KCC2  in mediating both influx and efflux of K+-Cl− . However, unlike KCC2, which is ~75% identical, KCC4 does not mediate K+-Cl− cotransport under isotonic conditions in Xenopus oocytes and mediates higher swelling-activated cotransport .
SLC12A8 and SLC12A9
Two other separate and distinct branches of the cation-chloride cotransporter family have recently emerged, each with only one paralog in the Drosophila, C. elegans, human, and murine genomes. The first was denoted cation-chloride cotransporter interacting protein (CIP) due to its ability to inhibit the functional expression of co-expressed NKCC1 . Arguably, the conservation of this gene during evolution suggests that, rather than performing an accessory role in transport, CIP transports substrates that have simply not yet been defined. The 918-residue CIP protein predicts an extracellular loop between TM5 and TM6, similar to the KCCs. A unique feature of CIP is the presence of a type-I PDZ-interacting motif (T-D-L) at the extreme COOH-terminal end.
Finally, a ninth member of the family was recently identified through the characterization of a psoriasis-susceptibility locus on chromosome 3q21 . This widely expressed family member, provisionally denoted SLC12A8, has a unique predicted membrane topology; the much shorter protein (714 amino acids) does not seem to contain predicted extracellular loops within the central core of hydrophobic domains, and may end with a glycosylated extracellular COOH-terminal tail. The transport function of the SLC12A8 protein is again not yet known, however there are clear orthologs within the Drosophila and C. elegans genomes.
The NKCC2 and NCC proteins were defined in large part by the differential sensitivity to the loop diuretics and thiazides, respectively. Given the extensive clinical experience with these widely prescribed drugs, the incentive to develop newer agents seems minimal. In contrast, there is increasing interest in targeting the K+-Cl− cotransporters for drug development. Inhibition of red cell K+-Cl− cotransport is a particularly attractive goal in sickle cell anemia, in which abnormal regulation of this and other swelling-activated pathways may contribute to red cell dehydration and sickling [72, 73]. Activating agents specific for KCC2 are expected to have an antiepileptic effect (see above), and may be neuron protective . Identification of SLC12A6/KCC3 as the ACCPN gene suggests broader roles for this transporter in peripheral neuropathy, demyelinating disease, and schizophrenia (reviewed in ). Finally, cardiac KCCs (mostly KCC3 and KCC4) may be a major pathway for K+ efflux after myocardial ischemia, a major cause of ischemic arrhythmias (reviewed in ).
This work was supported by research grants 36124-M from the Mexican Council of Science and Technology (CONACYT) and No. 75197-553601 from Howard Hughes Medical Institute to GG, a Career Development Award from the VA to DBM, and from the National Institutes of Health: DK36803 (SCH and GG); DK57708 (DBM).