WNK Kinases, Renal Ion Transport and Hypertension

Two members of a recently discovered family of protein kinases are the cause of an inherited disease known as pseudohypoaldosteronism type II (PHAII). These patients exhibit arterial hypertension together with hyperkalemia and metabolic acidosis. This is a mirror image of Gitelman disease that is due to inactivating mutations of the SLC12A3 gene that encodes the thiazide-sensitive Na+:Cl– cotransporter. The uncovered genes causing PHAII encode for serine/threonine kinases known as WNK1 and WNK4. Physiological and biochemical studies have revealed that WNK1 and WNK4 modulate the activity of several transport pathways of the aldosterone-sensitive distal nephron, thus increasing our understanding of how diverse renal ion transport proteins are coordinated to regulate normal blood pressure levels. Observations discussed in the present work place WNK1 and WNK4 as genes involved in the genesis of essential hypertension and as potential targets for the development of antihypertensive drugs.

is switched to the right (guytonian hypothesis) [1] . That is, increased arterial blood pressure is required in order to achieve normal salt urinary output.
Because arterial hypertension is polygenic, in order to understand the mechanisms of the disease, it is important to identify genes involved in blood pressure regulation. A very successful approach used over the last decade has been unmasking genes causing monogenic diseases that exhibit high or low blood pressure levels under the assumption that if altered function of a single gene is enough to produce an abnormal change in blood pressure, it is highly likely that the gene could be involved in the genesis of essential hypertension. More than 15 genes have been identified following this strategy [2] . Supporting the guytonian hypothesis, all recognized genes are involved in the maintenance of renal salt balance. Good examples of genes discovered by this strategy that have increased our understanding of hypertension mechanisms are the serine/threonine kinases of the WNK family.

Identification of WNK Kinases as Proteins Involved in the Regulation of Arterial Blood Pressure
Two monogenic diseases resulting in abnormal variations in blood pressure exhibit mirror images in their clinical features. On the one hand, Gitelman disease (OMIM No. 263800) is an autosomal recessive disorder associated with arterial hypotension, together with hypokalemic metabolic alkalosis and hypocalciuria. This monogenic disease is due to mutations in the SLC12A3 gene located in the human chromosome 16 that encodes for the thiazide-sensitive Na + :Clcotransporter (NCC) that is expressed in the apical membrane of the distal convoluted tubule (DCT) [3][4][5] ( table 1 ). The mutations are of the 'loss of function' type resulting in inactivation of the cotransporter [9][10][11] . The absence of NCC activity decreases salt reabsorption in the DCT, resulting in dehydration and arterial hypotension. The consequential increase in salt delivery to the connecting tubule (CNT) and collecting duct (CD) induces increased potassium and hydrogen secretion producing the hypokalemic metabolic alkalosis. On the other hand, a disease known by several names such as Gordon's disease, familial hyperkalemic hypertension or pseudohypoaldosteronism type II (PHAII; OMIM No. 145260) exhibits a mode of inheritance that is compatible with autosomal dominant transmission and features arterial hypertension that is accom-panied by hyperkalemic metabolic acidosis and hypercalciuria. This phenotype is Cldependent and is corrected with low doses of thiazide-type diuretics [12] . Thus, PHAII is a mirror image of Gitelman disease, strongly suggesting that increased activity of the NCC must be implicated. However, initial genomic analysis demonstrated no significant linkage of kindred with PHAII to the SLC12A3 gene on chromosome 16 in patients with PHAII [13] , making it unlikely that the cause of PHAII is activating mutations of the NCC. Later, locus heterogeneity of PHAII was identified by Mansfield et al. [14] when positive linkage analysis was observed in two loci: one located in chromosome 1q31-q42 and another in chromosome 17p11-q21 ( table 1 ). The evidence of genetic heterogeneity of PHAII was increased by Disse-Nicodeme et al. [15,16] , who observed in French kindreds a positive linkage to chromosome 12p13.3 and in another family no linkage to the SLC12A3 gene or to chromosomes 1, 12 and 17. Thus, at least 4 genes are capable of producing the same disease independently ( table 1 ).
At about the time heterogeneity of PHAII was observed, in their effort to clone novel members of the mitogen-activated protein/extracellular signal-regulated protein kinases from brain tissue, the Cobb group in Texas [17] identified a new type of serine/threonine kinase that lacks the canonical lysine observed in all serine/threonine kinases in the subdomain II of the kinase domain. Due to the absence of this lysine, the new kinase was named WNK1 for 'with no lysine (K)'. This report would probably have passed unnoticed by the renal community if not for a study months later in which the Lifton group from Yale, following the positional cloning strategy [8] , identified that the cause of PHAII in families with positive linkage to chromosome 12 was due to deletion of a fragment of intron 1 in the PRKWNK1 gene, encoding WNK1.  [8] uncovered 2 genes causing PHAII. They first observed that 2 kindreds with significant linkage to chromosome 12p13 cosegregated with a 41-or 21kb deletion, respectively, within the first large intron of PRKWNK1 . No mutations or deletions within the WNK1coding sequence were detected in these families. The expression of WNK1 mRNA in leukocytes in affected individuals was observed to be 5-fold higher than in nonaffected members of the family. Thus, genomic deletions of the first intron of WNK1 increase the expression of an otherwise normal WNK1 kinase. Supporting this hypothesis, Delaloy et al. [18] later revealed that the region that goes from -2,500 to -1,200 of the human PRKWNK1 promoter represses transcription of the gene. To date, these are the only 2 PHAII kindreds in which deletion of PRKWNK1 genes have been reported.
The other form of PHAII that is due to mutations in WNK kinases is the one in which positive linkage was observed with human chromosome 17. Four independent missense mutations in the PRKWNK4 gene were observed by Wilson et al. [8] in these families. Interestingly, 3 out of 4 missense mutations occur within a highly conserved acidic region that is located within the carboxylterminal domain, just after the first coiled-coil domain ( fig. 1 ). This negatively charged region contains 10 amino acid residues that are 100% identical between WNK1, WNK2 and WNK4, and 70% with WNK3. Six out of 10 amino acid residues in this region of WNK4 are charged.
The 3 missense mutations that occur in this region change or eliminate the charge of a residue (E562K, D564A and Q565E mutations). Golbang et al. [19] later reported another PHAII kindred in which asparagine 564 is mutated to histidine (D564H). In addition, Wilson et al. [8] also observed that mutation R1185C cosegregates with PHAII in another family. This mutation, although outside the acidic domain, also changes the charge of an amino acid residue. Thus, the PHAII that cosegregates with the PRKWNK4 gene is due to point mutations that change 1 charged amino acid residue. It is believed that this alteration probably affects the way in which WNK4 interacts with other proteins or charged molecules, resulting in modification of the effect of WNK4 upon transport systems.

The WNK Family Is Composed of Four Members
After the cloning and identification of WNK1 from the rat kidney [17] , it was soon observed that the WNK family is composed of 4 members that were named WNK1, WNK2, WNK3 and WNK4 ( fig. 1 ). The corresponding genes PRKWNK1 , PRKWNK2 , PRKWNK3 and PRKWNK4 are located in human chromosomes 12, 9, X and 17, respectively. As shown in figure 1 , WNK kinases are composed of 3 domains. A serine/threonine kinase domain of 274 residues is flanked by a short aminoterminal domain of 146-220 amino acid residues and a large carboxyl-terminal domain of 786-1,888 residues containing an autoinhibitory domain and 2 coiled-coil domains that could be important for protein-protein in- teraction. The degree of identity among the 4 WNKs at the amino-and carboxyl-terminal domains is below 20%, whereas at the serine/threonine kinase it is above 80%. The WNK1 kinase domain crystal structure was resolved at 1.8 Å demonstrating that the catalytic lysine is located in the ␤ -strand 2, rather than strand 3 as in other serine/ threonine kinases [20] . Several in vitro studies have shown that WNK kinases can interact amongst each other at the protein-protein level [21][22][23] or by phosphorylation processes [24] , as well as with many other kinase proteins including mitogen-activated protein kinases [6,25] , transforming growth factor ␤ signaling pathway kinases [26] , synaptotagmin [27] and STE-20-related kinases like serine-proline-alanine-rich kinase (SPAK) or oxidative stress response kinase 1 (OSR1) [7,[28][29][30][31] . WNK1 is a ubiquitously expressed kinase with predominant expression in the kidney, heart, muscle and testis in the rat [17] , mouse [32] and human [33] . In most tissues, it is predominantly expressed in polarized epithelial cells. In the kidney it is expressed along the entire nephron [32,34] . By RT-PCR analysis it has been observed that WNK3 transcripts are present in all tissues [35] , and by immunohistochemistry WNK3 protein was shown to be expressed in all nephron segments, as well as in epithelial cells of several organs such as the pancreas, biliary duct, stomach and intestine [36] . In contrast to WNK1 and WNK4, WNK3 protein is highly expressed in the central nervous system in which it is present in cell bodies of neurons expressing ionotropic ␥ -aminobutyric acid A receptors [36] . WNK4 is mainly expressed in the aldosterone-sensitive distal nephron. At the protein level it has been shown to be present in the DCT and CD [8,34] , and at the mRNA level, by single-nephron RT-PCR, WNK4 transcripts have been shown to be expressed also in the thick ascending limb of Henle's loop [34] . In addition, Kahle et al. [37] observed that WNK4 transcript and protein are present in several epithelial tissues, in which expression is more prominent in tight junctions. Interestingly, all epithelia expressing WNKs are heavily involved in Cltransport.
The molecular diversity of the WNK family has been shown to be increased by expression of several alternative splicing variants described in WNK1 and WNK3. Two independent groups [18,38] observed by Northern blot analysis 2 WNK1 transcripts, 9.0 and 10.5 kb in size, that differ in the length of the 3 untranslated region. These transcripts are present in all tested tissues and are known as L-WNK1. In addition, there is an approximately 8.0-kb transcript expressed only in the kidney that is due to alternative splicing of exons 1-4. The transcription of this isoform is under control of an intron 4 alternative promoter and contains sequences from an extra exon located between exons 4 and 5 which has been designed as exon 4a. As shown in figure 1 , at the protein level the consequence of this splicing mechanism is a shorter WNK1 isoform that lacks the first 437 amino acid residues, including almost the entire kinase domain. This shorter, truncated isoform is known as S-WNK1. By in situ hybridization it has been demonstrated that within the kidney, L-WNK1 isoforms are present all along the entire nephron, whereas S-WNK1 is only present in the aldosterone-sensitive distal nephron, particularly abundant in the DCT and CNT. These locations are very important to understand the mechanism by which overexpression of L-WNK1 produces activation of the NCC (see below). For WNK1 another 5 different transcripts have been detected due to splicing of exons 9, 11 and 12, either separately or in combination. The functional consequences of all these isoforms are unknown. Finally, human WNK3 exhibits at least 2 alternative variants. One is due to usage of an alternative splice donor into exon 18 that introduces 47 amino acid residues, and the other is due to splicing of exon 22 [35] . The functional consequences of these splicing isoforms are unknown.

Modulation of Renal Ion Transport Systems by WNKs
Several lines of evidence suggested that WNKs would be critical regulators of renal ion transport systems: (1) WNK1 and WNK4 are predominantly expressed in the aldosterone-sensitive distal nephron; (2) in addition to hypertension, PHAII patients exhibit hyperkalemia, metabolic acidosis and hypercalciuria consistent with impairment of the normal renal excretion of these ions; (3) PHAII is the mirror image of Gitelman disease that is due to inactivating mutations of the NCC; (4) the PHAII clinical picture is easily reverted by very low doses of thiazide-type diuretics [12] .

Effects of WNK4 upon Distal Nephron Transport Systems
Due to the mirror image between Gitelman disease and PHAII, and the exquisite sensitivity of this latter disease to thiazides, the first transport system that was studied was the NCC, which is the major salt transport pathway in the DCT ( fig. 2 ). The effect of WNK4 upon NCC activity was assessed using the heterologous expression system of Xenopus laevis oocytes, which was used to clone these and other transporters [39,40] , and in which most of the functional properties of the NCC have been analyzed [41] . Two independent groups obtained comparable results simultaneously [42,43] . When coexpressed in oocytes, WNK4 reduces the activity of NCC by decreasing the amount of the cotransporter present on the cell surface. This negative effect upon the NCC was dependent on the WNK4 kinase activity and was also eliminated by introducing the PHAII mutation Q562E (mouse sequence; table 2 ). These observations were later corroborated by other groups using culture cells as expression systems [19,44,45] and, also, added the evidence that decreased expression of NCC on the cell surface is not associated with increased dynamin-induced vesicle internalization but is rather due to increased lysosomal degradation of NCC protein.
PHAII is associated with hyperkalemia and potassium excretion in the nephron that predominantly occurs in the aldosterone-sensitive distal nephron. Thus, another obvious target for WNK4 regulation was the renal outer medullary potassium channel (ROMK) that is the major renal potassium-secretory pathway in the apical membrane of the distal nephron ( fig. 2 ). Similar to NCC, WNK4 is a negative regulator of the ROMK [46] . However, there are important differences between WNK4 ef-fects upon the NCC than upon the ROMK ( table 2 ). First, WNK4 catalytic activity is not required for ROMK inhibition. Second, the mechanism is due to a WNK4-induced increased internalization of the ROMK via clathrin-coated pits. These observations performed originally in Xenopus oocytes have been confirmed in mammalian cells [47] , in which it has been observed that WNK4 interacts with a protein known as intersectin that is in-  Salt and potassium transport systems in the DCT and CNT. The DCT begins few cells after the macula densa; it is divided into 2 portions, the DCT1 in which the salt reabsorption pathway is the NCC and the DCT2 in which salt reabsorption takes place through both, the NCC and ENaC. Then, the ENaC is the major salt reabsorption pathway in the CNT and downstream in the CD. In all these segments, the ROMK is the major potassiumsecretory pathway. As discussed in the text, L-WNK1, S-WNK1 and WNK4 are heavily expressed in these nephron segments. volved in the clathrin-induced endocytosis. Additionally, treatment of HEK-293 cells with siRNA against WNK1 and WNK4 resulted in increased activity of the ROMK, indicating that, indeed, basal activity of the channel is modulated by both kinases [48] . Third, while for the NCC PHAII type mutations in WNK4 prevented the dominant negative effect, for ROMK they resulted in increased inhibition. This observation can explain the hyperkalemia in PHAII patients since inhibition of the ROMK is increased by mutant WNK4.
A third candidate target for WNK4 is the apical epithelial sodium channel (ENaC) that in the distal nephron is an important pathway for sodium reabsorption ( fig. 2 ). In addition, activity of the ENaC is required for ROMK potassium excretion since the luminal negative potential that is generated by sodium translocation is required for potassium to be secreted. Ring et al. [49] demonstrated that WNK4 reduces the activity of the ENaC. This effect does not require WNK4 catalytic activity and is lost by elimination of the PPXY motif of the ENaC ␤ -or ␥ -subunit, indicating that, similar to ROMK, it is due to clathrin-mediated endocytosis. Interestingly, PHAII type mutations in WNK4 prevent the inhibitory effect of WNK4 upon the ENaC ( table 2 ).
The fourth identified targets for WNK4 are the proteins mediating parecellular Clflux in the distal nephron known as claudins. Two independent groups [50,51] observed simultaneously in MDCKII cells that WNK4 increases the paracellular transport of Clbut not that of Na + . This effect requires WNK4 catalytic activity and is associated with phosphorylation of claudin 4 in its carboxyl-terminal domain. In addition, PHAII-type mutations in WNK4 are associated with a further increase in Clpermeability and phosphorylation of claudin 4 ( table 2 ). Finally, WNK4 also inhibits the activity of the K-Cl cotransporters KCC1, KCC3 and KCC4 [52] , which in the distal nephron are critical for potassium secretion in the DCT [53] and acid secretion in the CD [54] . KCC4 knockout mice feature metabolic acidosis. The WNK4 inhibitory effect is kinase dependent but was not affected by PHAII-type mutations ( table 2 ).

Effects of WNK1 upon Distal Nephron Transport Systems
The effect of WNK1 upon distal transport systems is probably more complex than WNK4 since it has been demonstrated that WNK1 is a regulator of other WNKs by phosphorylation processes [55] or by WNK-to-WNK interactions at the protein level [21,22] . For instance, Yang et al. [43] observed that WNK1 has no direct effect upon the NCC, but prevents the WNK4-induced inhibition of the NCC. That is, when the NCC and WNK4 were coexpressed together in X. laevis oocytes, WNK4 induced inhibition of NCC activity. However, when WNK1 was added to the coexpression cocktail, the WNK4 inhibitory effect of the NCC was lost. Later on, Subramanya et al. [56] analyzed the effect of the L-WNK1 and S-WNK1 variants of WNK1 upon the NCC and its WNK4 regulation. Their observations are very important to understand the mechanism by which increased expression of a normal L-WNK1 in PHAII patients produces increased activity of the NCC. These authors demonstrated that while L-WNK1 prevented the WNK4-induced inhibition of the NCC, S-WNK1 by interacting with L-WNK1 in a dominant negative fashion eliminated the L-WNK1induced inhibition of WNK4. As was discussed above, while L-WNK1 is present in several epithelial cells and along the entire nephron, S-WNK1 is mainly expressed in the DCT and CNT [34,38] . Thus, it is proposed that in normal subjects the ratio of S-WNK1/L-WNK1 expression in the DCT and CNT is in favor of S-WNK1, and thus, the shorter isoform by highjacking L-WNK1 prevents its inhibitory effect upon WNK4, allowing WNK4 to keep the NCC inhibited. In PHAII patients due to intronic deletions of the PRKWNK1 gene that increases expression of L-WNK1, the ratio S-WNK1/L-WNK1 is reduced. As a consequence, L-WNK1 is able to inhibit WNK4 and thus, NCC activity is increased, augmenting salt reabsorption in the DCT and thus arterial pressure.
Similarly, it has been demonstrated that S-WNK1 and L-WNK1 interact with each other regulating ROMK activity. Lazrak et al. [48] observed in HEK cells transfected with ROMK that L-WNK1 decreased the activity of the channel. This effect of L-WNK1 requires the L-WNK1 catalytic activity and can be prevented by coexpression of S-WNK1. In the same study it was observed that exposing rats to a K + -deficient diet was associated with increased expression of L-WNK1 and decreased expression of S-WNK1, consistent with a situation in which ROMK activity is decreased since K + secretion is maximally reduced. Consistent with these findings, O'Reilly et al. [34] also observed that the ratio S-WNK1/L-WNK1 and the expression of WNK4 at the mRNA level are modulated by low-and high-potassium diets.
WNK1 is also a modulator of ENaC activity. In a very interesting study, Xu et al. [57] demonstrated that ENaC can be regulated by WNK1, via another kinase known as serum glucocorticoid kinase (SGK). WNK1 induces a phosphorylation of SGK which in turn phosphorylates and inhibits a protein named Need4 that is known to re-duce ENaC activity by promoting its endocytosis via a clathrin-dependent mechanism. Therefore, ENaC activity is enhanced. Interestingly, SGK is known to be a key kinase through which aldosterone achieves its effects upon distal nephron transport mechanisms. Finally, similar to WNK4, WNK1 also increases paracellular Clfluxes, an effect that was associated with increased phosphorylation of claudin 4 ( table 2 ).

WNKs Lie Upstream of Kinases Involved in the Regulation of Ion Transporters
Biochemical and functional analysis revealed that WNK kinases and the cation-coupled chloride cotransporter (SLC12 family) interact with each other and with another family of serine/threonine kinases known as SPAK and OSR1 that belong to the STE-20 family. On the one hand, Piechota et al. [28,58] observed that some members of the SLC12 family, such as the K + -Clcotransporter KCC3 and the basolateral isoform of the Na + : K + :2Clcotransporter, NKCC1, interact with SPAK in a 2-hybrid yeast system. They observed that these interactions occurred through an 'SPAK-binding domain' [RFx(V/I)] present in the cotransporter sequence. Simultaneously, it was demonstrated by Dowd and Forbush [59] that modulation of NKCC1 activity by intracellular chloride concentration involved the activity of this kinase. On the other hand, Vitari et al. [29] observed that immunoprecipitation of WNK1 or WNK4 brought SPAK as an interacting protein. In this study they observed in an in vitro phosphorylation assay using the amino-terminal domain of NKCC1 that the presence of WNK1, WNK4, SPAK or OSR1 alone was not enough to achieve NKCC1 phosphorylation. Only when WNK1 or WNK4 was added together with SPAK or OSR1, did NKCC1 become phosphorylated. This effect disappeared when catalytically inactive kinases were used. The conclusion of the study was that WNK1 or WNK4 is able to induce phosphorylation of NKCC1 only when SPAK or OSR1 is present. Similar observations were obtained by Moriguchi et al. [31] for NKCC1, and also for the apical isoform of the Na + :K + :2Clcotransporter, NKCC2, and for the NCC. These observations are supported by the observations of Gagnon et al. [7] at the functional level. These authors observed in X. laevis oocytes that injecting WNK4 cRNA had no effect upon the NKCC1 cotransporter, unless SPAK cRNA was added to the injecting cocktail. They observed that injecting cRNA of both kinases resulted in a significant activation of NKCC1. Interestingly, this interaction does not seem to be required for WNK4-induced inhibition of the K + -Clcotransport-ers [52] . Finally, detailed definitions of the interacting domains between WNK4 and SPAK were informed [30,60] . Thus, it is currently believed that at least for some cotransporters, WNK1 or WNK4 lie upstream of SPAK/ OSR1 kinases.
In vivo Mouse Models of PHAII All the above observations regarding modulation of renal ion transport systems by WNKs were done in heterologous expression systems of either amphibian or mammalian origin. Thus, it was necessary to demonstrate that the proposed mechanisms by which PAHIItype mutations in WNKs produce the disease actually occur in an in vivo model. Regarding WNK1 the only study in which genetic manipulation of WNK1 has been performed is in the WNK1 knockout mice produced by Zambrowicz et al. [61] . Mice homozygous for WNK1 deletion died before day 13 of gestation, indicating that WNK1 is an important protein for embryogenesis. However, WNK1 heterozygous mice survived to adulthood and displayed a significant reduction in arterial blood pressure, indicating that WNK1 plays a key role in maintaining normal blood pressure.
Two groups have produced transgenic mice overexpressing WNK4 containing a PHAII-type mutation. First, Lalioti et al. [62] generated bacterial artificial chromosome transgenic mice overexpressing either wild-type WNK4 or a WNK4 gene containing one of the PHAIItype mutations (Q562E). Interesting observations were done in this study: when compared with control mice, PHAII-WNK4 transgenic mice were hypertensive, whereas wild-type WNK4 transgenic mice were hypotensive. When exposed to a low-K + diet, PHAII-WNK4 transgenic mice were hyperkalemic (approx. 5.3 m M ) and when exposed to a high-K + diet, these animals developed a dramatic increase in serum K + to values over 8.0 m M . In contrast, wild-type WNK4 transgenic mice were hypokalemic (approx. 3.7 m M ). Furthermore, PHAII-WNK4 transgenic mice exhibited metabolic acidosis and were hypercalciuric. Within the histological analysis, the observations were more impressive: wild-type WNK4 transgenic mice exhibited a significant reduction in the number and size of DCTs. In contrast, PHAII-WNK4 transgenic mice exhibited a remarkable increase in the number and size of distal tubules, that is DCT hypertrophy and hyperplasia. Therefore, overexpression of PHAII-WNK4 recapitulated the PHAII phenotype, and interestingly, overexpression of wild-type WNK4 recapitulated a phenotype that is similar to Gitelman disease, in which NCC activity is absent. Finally, to find out to what extent the effects of WNK4 were due to modulation of NCC activity, the PHAII-WNK4 transgenic mice were crossed with NCC knockout mice. The PHAII phenotype disappeared in the new colony. Renal histology was normal, with no changes in the DCT when compared with normal controls. Arterial blood pressure and all other parameters were corrected. Even the hyperkalemia during the high-K + diet was prevented, implicating NCC as a key protein, not only for the development of hypertension, but also for hyperkalemia. In the same study, treating PHAII-WNK4 transgenic mice with thiazides also corrected the phenotype. A second transgenic model that has been formed is a knockin of a PHAII mutation (D561A) that was introduced into the endogenous WNK4 locus [63] . In this study it was also observed that knockin mice reproduced the phenotype of the disease with hypertension, hyperkalemia and metabolic acidosis. At the molecular level it was shown that NCC, ENaC and BK K + channel expression was increased, with no change in WNK4 and the ROMK. Increased phosphorylation of SPAK/OSR1 was observed in the knockin mice supporting the hypothesis that WNK4 affects the activity of NCC by interacting with, and/or phosphorylation of, STE-20 kinases. In addition, it was also observed that PHAII-WNK4 induced phosphorylation of the NCC at serine 71 where it was previously shown that its phosphorylation was associated with the activity of the NCC [64] . These studies strongly suggest that, as predicted in the heterologous expression systems, wild-type WNK4 is a potent inhibitor of NCC, and PHAII-type mutations eliminate this WNK4 property releasing the NCC from a tonic inhibition by this kinase. However, the consequences of transgenic wildtype WNK4 or PHAII-type WNK4 upon transport systems such as the ROMK, ENaC or paracelin were not reported in detail, and thus, the WNK4 in vitro effects upon this transport system still need to be confirmed in vivo.

Implications of WNK4 for Renal Physiology
The effects of WNK4 upon renal transport systems shown in table 2 reveal that WNK4 exists in at least 3 different states of function. One is the wild-type WNK4 that inhibits NCC, ROMK and ENaC, with slight stimulation of paracellular Clflux. One can think of this as an equilibrium state. A second clearly identified state is the one resulting from PHAII mutations. In this state, WNK4 no longer inhibits the NCC and ENaC resulting in increased activity of these pathways, and increased inhibition of the ROMK. Simultaneously, paracellular Clflux is also enhanced. A third state involves the effect of WNK4 phos-phorylation by SGK kinase. WNK4 contains a canonical SGK phosphorylation site at serine 1169. Ring et al. [65] first showed that WNK4 is phosphorylated at this site by SGK. Then they observed that mutation of S1169 for alanine, which eliminates the possibility that WNK4 becomes phosphorylated at this site, had no effect upon WNK4-induced inhibition of the ROMK or ENaC. In contrast, mutation of serine 1169 for aspartic acid (S1169D) that mimics the phosphorylation state of this serine resulted in a WNK4 kinase that lost the inhibitory effect upon the ENaC or ROMK, resulting in increased activity of these channels [65] . However, this mutation S1169D has no effect upon the WNK4-induced inhibition of the NCC [Gamba G., unpubl. observations]. The differential functional properties of WNK4 revealed by these different states, together with observations in PHAII patients and transgenic mice, place WNK4 as the potential kinase and the key molecular switch that explains how the kidney handles the effect of aldosterone as a salt-retaining and K + -losing hormone. In physiological conditions, losing fluid volume increases aldosterone levels, which in turn activates salt-retaining mechanisms, without affecting the normal urinary potassium excretion. This is achieved by activating the NCC, ENaC and paracellular Clflux, without changing or even reducing the ROMK. As discussed above, WNK4 harboring a PHAII-type mutation mimics this situation. In contrast, another stimulus for aldosterone secretion is hyperkalemia. In this situation, aldosterone release is associated with increased K + excretion, without affecting salt retention. This is achieved by increasing ENaC and ROMK activity, without affecting the NCC. Thus, salt reabsorption is not increased in the DCT1 allowing enough sodium to be delivered to the DCT2 and CNT, in which the increased activity of the ENaC and ROMK promotes potassium secretion into the lumen ( fig. 2 ).

Future Research for WNKs, Renal Ion Transport and Human Hypertension
Observations from PHAII patients, transgenic animals and in vitro transfection models together strongly suggest that WNKs and several ion transport systems in the kidney work together to maintain salt and potassium homeostasis and, as a consequence, arterial blood pressure homeostasis. Alterations in the activity of WNKs and/or transport systems are clearly associated with human diseases featuring arterial hypertension or hypotension [2] . One important area that is currently being stud-ied is defining the role that WNKs can have upon the most common forms of hypertension, particularly since a considerable proportion of patients with essential hypertension exhibit low plasma renin activity, suggesting that increased blood pressure in these patients is due to salt retention. In this regard, the WNK4 gene in chromosome 17 is just 1 Mb from the locus D17S1299 that has shown the strongest linkage with hypertension in the Framingham Heart Study population [66] , and recent studies suggest the association of SNPs or variants of PRKWNK1 or PRKWNK4 with arterial hypertension [67,68] . It is equally important to define drug development against WNK kinases. The facts that mice heterozygous for WNK1 deletion exhibit a significantly lower blood pressure than normal mice and that PHAII-WNK4 transgenic or knockin mice are hypertensive suggest that drugs affecting the functional properties of WNK1 or WNK4 could play a potential role in the treatment of arterial hypertension.