WNK1 is a chloride-stimulated scaffold that regulates mTORC2 activity and ion transport

Mammalian (or mechanistic) target of rapamycin (mTOR) is well recognized as a central controller of cell growth and metabolism (Liu and Sabatini, 2020). It is found in two structurally and functionally distinct multiprotein complexes, mTORC1 and mTORC2 (Fu and Hall, 2020; Loewith et al., 2002; Sarbassov et al., 2004). mTORC1 includes the defining core component, Raptor, and plays central roles in growth, tissue remodeling and immune function by regulating ribosome activity through phosphorylation of p70-S6 kinase and 4-EBP (also known as RPS6KB1 and EIF4EBP1, respectively), among others, in a rapamycin-sensitive manner (Fu and Hall, 2020). mTORC2 is defined by two core subunits, rapamycin-insensitive companion of TOR (Rictor) and mammalian stress-activated protein kinase-interacting protein 1 (SIN1; also known as mSin1 and MAPKAP1), which together control mTORC2 assembly and substrate specificity, in particular targeting the AGC kinases, Akt family proteins (hereafter Akt), SGK1 and protein kinase C (PKC) (Fu and Hall, 2020). It was originally identified as a rapamycin-resistant kinase complex involved in growth control and actin cytoskeleton regulation through phosphorylation of Akt and PKC at a conserved site termed the ‘hydrophobic motif’ (HM) (Jacinto et al., 2004; Sarbassov et al., 2005a,b). HM sequence homology led to studies demonstrating similar mTORC2-dependent phosphorylation of SGK1 at Ser-422, and that it has central roles in electrolyte homeostasis (Garcia-Martinez and Alessi, 2008; Gleason et al., 2015; Grahammer et al., 2016; Lu et al., 2010).

The roles of mTORC2 in responding to hormones, growth factors and cytokines to regulate cell growth, ion transport and metabolism have been extensively investigated; however, recent studies also point to roles in rapid cell autonomous responses to physiological stresses, such as ambient temperature and blood K⁺ levels (Allu et al., 2021; Fu and Hall, 2020; Sørensen et al., 2019). One recent report supported the idea that local extracellular K⁺ concentration rapidly stimulates mTORC2-dependent phosphorylation of SGK1, with downstream effects on the epithelial Na⁺ channel ENaC (Sørensen et al., 2019). This signaling module was further shown to be central to rapid stimulation of K⁺ secretion, consistent with the well-established role of ENaC in controlling the electrical gradient driving apical K⁺ transport, particularly in the context of rapid responses to K⁺ (Frindt and Palmer, 2009). Interestingly, the serine-threonine kinase WNK1 has been shown to potently stimulate mTORC2 activity through a non-catalytic mechanism; however, the functional consequences and underlying mechanism had not been examined (Sørensen et al., 2019). Although most studies have focused on kinase-dependent roles of WNKs in regulating SPAK (also known as STK39) and OSR1 and hence cation-chloride cotransporters such as NCC (SLC12A3) and NKCC1 (SLC12A2) and NKCC2 (SLC12A1) (Murillo-de-Ozores et al., 2020; Terker et al., 2016), a previous report identified non-catalytic stimulation of SGK1 and ENaC (Xu et al., 2005b). Similarly, WNK1-mediated activation of BK channels through non-catalytic mechanisms has been described (Webb et al., 2016). However, the role of mTORC2 was unknown, and the underlying signaling mechanisms for non-canonical signaling have yet to be investigated.

To further explore the mechanism and functional implications of WNK1-stimulated mTORC2 activity, we examined SGK1 phosphorylation and its downstream effects on ENaC activity in mTORC2- and WNK1-deficient renal epithelial cells. Additionally, based on observations that Cl^– can bind WNK1 (Piala et al., 2014) and inhibit its catalytic activity in a K⁺-regulated fashion (Penton et al., 2016; Terker et al., 2015) we also examined the role of Cl^– in controlling non-catalytic effects of WNK1. Our data support the idea that WNK1 is a Cl^–-regulated scaffold for mTORC2-mediated SGK1 phosphorylation, and thereby stimulates ENaC-dependent Na⁺ transport. This mechanism represents a new mode of mTORC2 kinase signaling through which local extracellular K⁺ concentration selectively activates one of the substrates of mTORC2 (SGK1), but not others (Akt and PKC) to regulate ion transport and electrolyte homeostasis.

As a first step toward characterizing the relationship between mTORC2 and WNK1 in K⁺-induced ENaC activation, we used mpkCCD cells derived from mouse kidney cortical collecting duct (Bens et al., 1999). We generated mTORC2-deficient cells using CRISPR/Cas9 to delete Rictor (Fig. 1A). As shown previously (Sørensen et al., 2019), wild-type (WT) mpkCCD cells demonstrated robust K⁺-stimulated ENaC current and SGK1 Ser-422 phosphorylation (Fig. 1B,C, respectively). Consistent with prior studies in which mTOR was pharmaceutically inhibited or knocked down through RNAi (Lu et al., 2010; Sørensen et al., 2019), RICTOR^−/− mpkCCD cells showed a marked decrease in both ENaC currents and SGK1 Ser-422 phosphorylation (pSGK1), and showed no response to a change in the medium [K⁺] (Fig. 1B,C). Based on the phenotype of familial hyperkalemia and hypertension (FHHt) (Wilson et al., 2001), and a report that WNK1 could stimulate SGK1 and ENaC (Xu et al., 2005a), we next generated and characterized WNK1^−/− mpkCCD cells, also using the CRISPR/Cas9 system (Fig. 2A). Baseline ENaC currents were reduced in WNK1^−/− cells (Fig. S1A,B), and K⁺-induced activation was abolished (Fig. 2B); pSGK1 and SGK1-dependent phosphorylation of Nedd4-2 (also known as NEDD4L) were also reduced and K⁺-stimulated phosphorylation was lost (Fig. 2C–E; Fig. S2). Expression of active processed form of αENaC (encoded by SCNN1A) (cleaved αENaC) was increased in K⁺-stimulated WT mpkCCD cells but not in WNK1^−/− cells [Fig. 2C, ENaC (Cl)]. It is notable that phosphorylated Akt was also detected (Fig. 2C, pAKT) in WT mpkCCD cells; however, it was not induced by increased K⁺. Interestingly, Akt HM phosphorylation was unaffected by loss of WNK1 (Fig. 2C, pAKT, lanes 4–6), consistent with previous results in HEK-293 cells (Sørensen et al., 2019). Finally, WNK-dependent SPAK phosphorylation was inhibited by elevated [K⁺] in WT mpkCCD cells, and this activity was attenuated in WNK1^−/− cells, and K⁺ regulation was lost (Fig. 2F), as is well established in renal distal convoluted tubule (DCT) cells (Sorensen et al., 2013). Thus, WNK1-deficient mpkCCD cells have severe defects in both mTORC2–SGK1 and canonical WNK–SPAK signaling.

We next examined the role of WNK1 kinase activity on the mTORC2–SGK1 signaling module. First, we demonstrated that transfection of WNK1^−/− mpkCCD cells with WT WNK1 (both long WNK1 and a truncated form known to phosphorylate SPAK) restored K⁺-regulated ENaC activation (Fig. 3A) and SGK1 HM phosphorylation (Fig. 3B, pSGK1; Fig. 3C, upper panel). However, in striking contrast to SPAK regulation (Fig. 3B, pSPAK, pOSR1; Fig. 3C, lower panel), ENaC activation and SGK1 phosphorylation were restored as well by kinase-dead WNK1 (K233M) as it was by WT (Fig. 3A,B). Similar effects were seen in WNK1^−/− HEK-293 cells transfected with ENaC, SGK1 and Nedd4-2 (Fig. S3); ENaC activity (as measured by patch clamp) was nearly undetectable in these cells (Fig. 4A, left panel; Fig. S4) and was fully restored by either WT or kinase-dead WNK1. It is notable that the effects of K⁺ were abrogated in HEK-293 cells transfected with kinase-dead SGK1 (K127M), which had a dominant-negative effect on ENaC, as has been described previously (Fig. 4A, right panel) (Loffing et al., 2001), supporting a central role of SGK1 in these effects. Likewise, extracellular K⁺ stimulation of SGK1 phosphorylation was restored by WT and kinase-dead L-WNK1 and their truncated variants (Fig. 4B).

As an alternative approach to assessing whether WNK kinase activity was dispensable, we examined the effect of a selective WNK inhibitor, WNK463 (Yamada et al., 2016), on mTORC2-dependent SGK1 phosphorylation and ENaC activation. Consistent with the results in Figs 3 and 4, WNK463 had no significant effect on K⁺-induced ENaC current (Fig. 5A) or pSGK1 (Fig. 5B,C) at 100 or 500 nM, concentrations that markedly inhibit pSPAK (Fig. 5D). At 1 µM, WNK463 did have a significant effect on ENaC and pSGK1, possibly due to off-target inhibition of mTORC2 or inhibition of a less-sensitive WNK kinase isoform.

Next, we addressed the electrolyte determinants of the K⁺ effect on the WNK1–mTORC2–SGK1 signaling module. Cl⁻ has been shown to bind WNK1 in vitro and strongly inhibit its kinase activity toward SPAK (Piala et al., 2014). Prior work further also supports the idea that, in the DCT, this WNK inhibitory effect of Cl⁻ plays a key role in K⁺ inhibition of NCC (Terker et al., 2015). Based on these observations, we speculated that Cl⁻ might stimulate WNK1–mTORC2-dependent SGK1 phosphorylation, which would provide a parsimonious mechanism for integrating K⁺-dependent NCC inhibition and ENaC stimulation. To test this hypothesis, we first examined the Cl⁻-dependence of the K⁺ effect by altering the medium [Cl⁻] from 3 mM to 110 mM. SGK1 HM phosphorylation (Fig. 6A, pSGK1, Fig. 6B, upper panel; Fig. S5) was negligible when cells were incubated in low [Cl^–] medium ([Cl^–]<40 mM), and was not stimulated by raising [K⁺]. At 40 mM [Cl^–], pSGK1 was modestly stimulated, and with [Cl^–] ≥80 mM, K⁺-stimulated SGK1 phosphorylation was markedly increased. In keeping with previous reports (Terker et al., 2015), pSPAK was high and not K⁺-inhibited in low Cl^–, but became K⁺-inhibitable at 40 mM [Cl^–]. The K⁺ sensitivity further increased at 80 and 110 mM [Cl^–] (physiological concentrations). Interestingly, Cl^– inhibitory effect on SPAK phosphorylation (Fig. 6A, pSPAK, pOSR1; Fig. 6B, lower panel) paralleled the stimulation of pSGK1. Also of note, there was no effect of [K⁺] or [Cl^–] on Akt phosphorylation (Fig. 6A, pAKT).

We next demonstrated that increased extracellular [K⁺] does indeed induce an increase in intracellular [Cl^–]. We compared the intracellular Cl⁻ concentration in HEK-293 cells incubated with either 1 mM or 5 mM medium [K⁺] using a Cl^–-sensitive fluorescent protein, mCl-YFP (Zhong et al., 2014). The cells were transfected with Cl-YFP plasmid and fluorescence signals were measured and analyzed in cells incubated with 1 and 5 mM K⁺ medium, as previously described (Kim et al., 2015; Zhong et al., 2014). Based on ionophore-clamped standard curves, the Cl^– concentration in HEK-293 cells treated with 5 mM K⁺ medium was measured as 65.1 mM, whereas in the presence of 1 mM extracellular K⁺ the intracellular [Cl^–] was 46.3 mM (Fig. 6C). We next examined whether altering intracellular [Cl^–] while maintaining [K⁺] constant would have a similar effect on pSGK1. We used the ionophores tributyltin and nigericin to clamp intracellular [K⁺] at 140 mM (a typical intracellular concentration) while clamping [Cl^–] at either 40 or 80 mM (spanning the intracellular [Cl^–] induced by low and high extracellular [K⁺], respectively). SGK1 HM phosphorylation was greater in the presence of high [Cl^–] as compared to low [Cl^–] (Fig. 6D, pSGK1; Fig. 6E, upper panel) conditions. Consistent with in vitro studies (Piala et al., 2014; Terker et al., 2016), WNK-dependent SPAK phosphorylation was inhibited by high [Cl⁻] (Fig. 6D, pSPAK, pOSR1; Fig. 6E, lower panel). Similar effects were observed when a wider range of Cl⁻ concentrations were employed in WNK1^−/−, SIN1^−/− double knockout (KO) HEK-293 cells which had been transfected with WT WNK1 and SIN1 (Fig. S6A,B). Together, these data support the notion that intracellular [Cl⁻] stimulates mTORC2-dependent SGK1 phosphorylation. Of further note, in WNK1^−/− cells, pSGK1 was low and insensitive to changes in extracellular [K⁺] and [Cl⁻] (Fig. S7). Thus, we conclude that extracellular K⁺ induces Cl^– entry into cells, which activates mTORC2 kinase activity selectively toward SGK1 in a WNK1-dependent fashion.

Our data establish that WNK1 enhances mTORC2-dependent phosphorylation of SGK1, and increases ENaC activity without activating Akt or PKC. This effect is markedly and selectively stimulated by K⁺ and, in contrast to SPAK–NCC regulation, WNK1 kinase activity is not implicated. Furthermore, our data provide strong support for the idea that the effect of K⁺ is mediated by Cl^–. Particularly in light of prior publications suggesting that WNK1 can stimulate SGK1 activity through a non-catalytic mechanism (Heise et al., 2010; Xu et al., 2005a,b), these findings suggest that WNK1 might serve as a K⁺-stimulated scaffold to enhance mTORC2 interaction with SGK1 thereby promoting phosphorylation. To examine this possibility, we used WNK1 and SIN1 double KO HEK-293T cells and performed a series of co-immunoprecipitation (co-IP) experiments with selective add-back of WNK1 and SIN1 (Fig. 7; Fig. S8). First, we co-transfected the cells with Flag–SGK1, Myc–WNK1 (either WT or a kinase-dead mutant) and SIN1–V5 (Fig. 7A). The cells were adapted to 1 mM [K⁺], and then were shifted to either 5 mM or maintained in 1 mM [K⁺] for 1 h. Lysates were prepared and Flag–SGK1 was immunoprecipitated (Fig. 7A; Fig. S8A). In the absence of WNK1, SIN1 was slightly detectable by co-IP, and unaffected by [K⁺] (Fig. 7A, lanes 1 and 2). In contrast, in the presence of either WT (Fig. 7A, lanes 3 and 4) or kinase-dead WNK1 (Fig. 7A, lanes 5 and 6), SIN1–V5 was detectable in 1 mM [K⁺] and the co-IP signal was further enhanced in 5 mM [K⁺] (Fig. 7A,B). SGK1–WNK1 interaction was also stimulated by K⁺ (Fig. 7A, Myc; Fig. 7B). SGK1 phosphorylation was concomitantly increased (Fig. 7C). In further experiments, Myc–WNK1 was immunoprecipitated, and blots were stained for either SGK1 or Rictor (Fig. 7D; Fig. S8B). Notably, there was a significant WNK1–SGK1 interaction in the absence of SIN1, which was further enhanced by SIN1. Similarly, there was a strong effect of K⁺ on WNK1 co-IP of Rictor (Fig. 7D, compare lanes 3 and 4); however, SIN1 had no effect (Fig. 7D, compare lanes 4 and 6). A similar effect of WNK1 on promoting the SGK1–mTORC2 interaction was seen when SIN1 was targeted for IP (Fig. S8C). Together, these findings suggest that WNK1 directly interacts with both Rictor and SGK1. Together with previous publications demonstrating that SGK1 interacts with mTORC2 through SIN1 (Cameron et al., 2011; Lu et al., 2011), these data support a model in which WNK1 serves as a K⁺-stimulated scaffold to enhance recruitment of SGK1 to mTORC2 and hence increase its phosphorylation. Furthermore, consistent with the observation that SGK1 phosphorylation is facilitated by high intracellular [Cl⁻] (Fig. 6D), the SGK1–SIN1 interaction was markedly increased by raising extracellular [K⁺] when cells were incubated in medium with a physiological [Cl^–], and remained negligible and unresponsive to increased extracellular [K⁺] when cells were incubated in low [Cl^–] medium (Fig. 7E). These data support the model shown in Fig. 8A and suggest that intracellular [Cl^–] stimulates WNK1 interaction with both SGK1 and mTORC2, and hence enhances the recruitment of SGK1 to mTORC2 and facilitates its phosphorylation. Finally, these data support a Cl^–- and WNK1-dependent mechanism for shifting Na⁺ transport from being electroneutral (via NCC) to electrogenic (via ENaC) thereby synergistically enhancing K⁺ secretion (Fig. 8B).

Here, we set out to investigate the signaling mechanism underlying local regulation of K⁺ secretion in kidney tubule cells. We showed previously that increased basolateral [K⁺] stimulates mTORC2-dependent SGK1 phosphorylation and ENaC activation in renal principal cells to depolarize the apical membrane and increase luminal K⁺ secretion, thereby establishing a novel homeostatic mechanism (Sørensen et al., 2019). We also showed that WNK1 markedly enhances mTORC2-dependent SGK1 phosphorylation; however, the molecular basis and functional implications of this effect remained unknown. As a first step toward addressing these questions, we generated (using CRISPR/Cas9) and characterized WNK1-deficient mpkCCD renal principal cells. These cells demonstrated markedly reduced K⁺-stimulated mTORC2-dependent SGK1 phosphorylation and ENaC activation (Fig. 2), which phenocopied mTORC2-deficient cells (Fig. 1). However, in marked contrast to mTORC2-deficient cells, Akt phosphorylation was unaffected in WNK1^−/− cells. Interestingly, mTORC2-dependent SGK1 activation was restored by transfection of either WT or kinase-dead WNK1 (Figs 3 and 4). That this kinase-independent activity of WNK1 is physiologically relevant was further supported by the observation that the WNK kinase inhibitor WNK463 had no effect on ENaC activity or SGK1 phosphorylation in WT mpkCCD cells at concentrations that inhibited OSR1 and SPAK phosphorylation to baseline (Fig. 5). Thus, the effects of K⁺ on both SGK1 and SPAK phosphorylation are WNK dependent; however, the mechanisms are distinct and the effects of K⁺ are opposite.

With these observations in mind, and in light of prior reports showing that extracellular K⁺ increases intracellular [Cl^–], and that Cl^– binds WNK1 and inhibits its kinase activity (Piala et al., 2014; Terker et al., 2015), we explored the possibility that Cl^– might oppositely regulate WNK1 effects on mTORC2 activity, specifically toward SGK1. In support of this possibility, K⁺-stimulated SGK1 phosphorylation required adequate extracellular [Cl^–] (≥80 mM), similar to that required for K⁺ inhibition of OSR1 and SPAK phosphorylation (Fig. 6A). Furthermore, we found that an increase in extracellular [K⁺] did indeed increase intracellular [Cl^–] (Fig. 6C) and, importantly, that increasing intracellular [Cl^–] (using an ionophore clamp) while holding [K⁺] constant increased pSGK1, whereas it decreased pSPAK. Finally, we demonstrated that WNK1 physically interacts with both SGK1 and mTORC2 and stimulates their mutual interaction in a Cl^–-dependent fashion (Fig. 7E), consistent with it having a role as a Cl^– regulated scaffold that selectively enhances SGK1 (but not Akt or PKC) phosphorylation.

Together, these data support the model for K⁺-induced SGK1 regulation shown in Fig. 8A. According to this view, increased extracellular K⁺ induces a rise in intracellular Cl^–, which binds WNK1 and alters its conformation, enhancing its interaction with both SGK1 and mTORC2 to promote SGK1 phosphorylation. SGK1-dependent activation of ENaC increases the driving force for K⁺ transport through renal outer medullary K⁺ channels (ROMK channels) and possibly BK channels (Fig. 8B). Thus, WNK1 acts as a Cl^– regulated scaffold that senses extracellular K⁺ and stimulates K⁺ secretion to complete a cell-autonomous homeostatic loop. These effects do not require WNK1 kinase activity (Figs 3–5), which is in fact inhibited by rising [Cl^–] (Fig. 6D) (Piala et al., 2014); this latter effect plays a central role in mediating effects of K⁺ to inhibit SPAK phosphorylation, and hence cation-chloride cotransporters such as NCC and NKCC1 and 2 (Murillo-de-Ozores et al., 2020; Sorensen et al., 2013; Terker et al., 2015). Together, these observations suggest that Cl^– induces a WNK1 conformation that favors its scaffolding function while concomitantly inhibiting its kinase activity (Fig. 8A,B). It is clear from earlier work (Piala et al., 2014) that Cl^– induces a change in WNK1 conformation that disrupts catalytic activity; however, how this conformation favors interaction with SGK1 and mTORC2 requires further study. Interestingly, the functional and co-IP data demonstrate that the effect is selective for SGK1, whereas Akt phosphorylation is unaffected.

Cells in the ASDN express both WNK1 and WNK4 (Shekarabi et al., 2017), and we cannot rule out a role for WNK4. We have not succeeded in making a selective WNK4^−/− line mpkCCD cells. It is notable that the effect of extracellular K⁺ on SGK1 phosphorylation and ENaC activity is completely abrogated in WNK1^−/− cells (Fig. 2), supporting its central role. However, a previous study found that mice with knock-in of a Cl^–-binding deficient mutant of WNK4 had a defect in ENaC activation (Chen et al., 2019), and it is possible that WNK4 also contributes to this mechanism. Prior studies have shown interactions of WNK1 and WNK4, and it is possible that WNK1, independent of its kinase activity interacts with other WNKs to modulate their activity and induce SGK1 phosphorylation; also it is possible that the kinase activity of WNK4 (or conceivably another WNK, such as WNK3) is implicated. Along these lines, it is further notable that the WNK kinase inhibitor WNK-463 did, at high concentrations, inhibit mTORC2 phosphorylation of SGK1 (Fig. 5). The concentrations required were 10-fold higher than those inhibiting WNK-dependent SPAK phosphorylation, and hence this could be an off-target effect. In any case, our data strongly suggest that WNK1, independent of its kinase activity, serves as a scaffold to bring together SGK1 and mTORC2 and enhance SGK1 phosphorylation. Further studies will be needed to address the role of WNK4, including its kinase activity. It is also important to note that a shorter variant of WNK1 (KS-WNK1), which lacks kinase activity and is highly expressed in aldosterone-sensitive distal nephron (ASDN) principal cells (in an aldosterone-regulated fashion), interacts with both WNK1 and WNK4, and could play a significant role in modulating ENaC and other transporters, such as ROMK, BK and NCC channels (Argaiz et al., 2018; Cheng et al., 2013; Liu et al., 2009; Ray et al., 2021; Shekarabi et al., 2017; Wade et al., 2006). It will be of interest to explore these possibilities in further studies.

It is also interesting to consider the potential implications of this mechanism for other processes, such as thermogenesis, inflammation and immune modulation. Recent studies suggest that Cl⁻-regulated inflammatory responses might be implicated in a variety of diseases, including inflammatory bowel disease, cystic fibrosis, atherosclerosis and hypertension (Lamkanfi and Dixit, 2014; Madhur et al., 2021; Scambler et al., 2019). One recent publication implicated WNK1 in mediating Cl⁻ modulation of NLRP3 inflammasome activity (Mayes-Hopfinger et al., 2021) and a separate study similarly identified SGK1 as an important inflammasome modulator (Clauzure et al., 2021); however, neither study addressed the potential connection between SGK1 and WNK1. mTORC2 modulation has also been implicated in responses to acute cold (Allu et al., 2021; Sanchez-Gurmaches et al., 2018) – one of these reports, showed that mTORC2 plays a central role in rapid direct responses to acute cold, stimulating catabolic and inhibiting anabolic activity in the absence of sympathetic nervous system activation (Allu et al., 2021). However, the cell-autonomous sensing mechanism to control mTORC2 activity was not elucidated.

Finally, in regard to the implications of our present findings for K⁺ homeostasis, it is worth emphasizing that concomitantly inhibiting WNK1 kinase activity through K⁺ while stimulating its interactions with SGK1 and mTORC2 provides a parsimonious mechanism for rapidly shifting Na⁺ transport from electroneutral NaCl reabsorption (which defends extracellular fluid volume) to electrogenic reabsorption (which drives K⁺ secretion to bring down extracellular [K⁺]) (Pearce et al., 2022; Sørensen et al., 2019) (Fig. 8B). Modulators of this aldosterone-independent mechanism for K⁺ secretion might provide new avenues for treating hyperkalemia, particularly when induced by inhibitors of the renin–angiotensin–aldosterone system.

Flag–SGK1 (WT), SGK1 (K127M) and SIN1–V5, were generated as described previously (Lu et al., 2011). Myc-tagged WT WNK1 (1–491) and Myc–WNK1 (1-491) (K233M) expression constructs were generous gifts from Drs Melanie H. Cobb (UT Southwestern, Dallas, TX, USA) and Chou-long Huang (University of IOWA, IOWA City, IA, USA; Xu et al., 2005b). Myc–L-WNK1 (WT) and HA–L-WNK1 (WT and K233M) expression constructs were generous gifts from Dr David H. Ellison (Oregon Health and Science University, Portland, OR, USA), and Drs Thomas R. Kleyman and Marcelo D. Carattino, University of Pittsburgh, PA, USA, respectively.

mpkCCDc14 (mpkCCD) (a kind gift from Alain Vandewalle, Institut National de la Santé et de la Recherche Médicale), mpkCCD WNK1^−/− and mpkCCD RICTOR^−/− cells were maintained in modified DMEM/Ham's F12 (1:1) medium (‘regular medium’) as described previously (Bhalla et al., 2005). mpkCCD WNK1^−/− cells were transfected using Lipofectamine 3000 (Thermo Fisher Scientific).

HEK-293T (ATCC), HEK-293T WNK1^−/− (Roy et al., 2015), HEK-293T SIN1^−/− (Gleason et al., 2019) and HEK-293T WNK1^−/− SIN1^−/− (double knockout) cells were grown in DMEM (DME H-21, CCFAA005, UCSF Cell culture facility) supplemented with 10% FBS and 1% penicillin and streptomycin antibiotics. These cell lines were routinely tested for mycoplasma contamination. All cell lines were acquired from reputable sources, but have not recently been authenticated. Cells were transfected using polyethylenimine (PEI). For all experiments, cells were serum-starved overnight. For experiments with different K⁺ concentrations, DMEM without KCl was used. K⁺-free DMEM was supplemented with KCl, NaCl or choline chloride to get the desired concentrations of K⁺ in the medium as indicated in the figure legends.

For experiments with different extracellular Cl⁻ concentrations, DMEM without KCl and NaCl was used (UCSF cell culture facility). The medium was supplemented with sodium gluconate and NaCl to vary Cl⁻ concentrations as indicated in the figures. Different K⁺ concentrations were achieved by adding potassium gluconate and sodium gluconate or KCl and NaCl to the media to maintain same osmolarity in all cultures.

HEK-293 cells were transfected with a modified Cl⁻ sensor YFP plasmid, monomeric Cl-YFP (Zhong et al., 2014) in 10 cm² plates. At 24 h post transfection, cells were re-plated in 96-well plates coated with poly-L-lysine. After 24 h, cells were incubated with 1 and 5 mM [K⁺] solution (composed of 110 mM NaCl, 2 mM CaCl₂, 1.2 mM MgCl₂, 10 mM HEPES, 25 mM glucose, and 30 mM Na-gluconate, pH 7.2, with 1 and 5 mM [K⁺] adjusted with equimolar potassium gluconate and sodium gluconate) for 30 min–1 h prior to measurement of fluorescent signals. To produce a standard curve, cells were incubated with high K⁺ solutions composed of 140 mM K⁺ (KCl or potassium gluconate), 2 mM CaCl₂, 1.2 mM MgCl₂, 10 mM HEPES and 25 mM glucose, containing different chloride concentrations (adjusted with equimolar KCl and potassium gluconate), and the ionophores nigericin (5 µM; cat. #B7644, APExBIO) and tributyltin (10 µM; cat. #T50202, Sigma Aldrich). Fluorescence signals were measured on a Synergy H4 Hybrid Multi-Mode Microplate Reader with excitation at 485 nm and emission at 520 nm. Fluorescence intensity values measured in each solution were normalized to that of 2 mM [Cl⁻] standard solution in each well.

For experiments with modulation of intracellular [Cl^–], cells were incubated for 1 h in high K⁺ solutions with different [Cl^–] in presence of nigericin (5 µM) and tributyltin (10 µM) to clamp intracellular and extracellular [Cl^–].

For CRISPR knockouts of the WNK1 and RICTOR gene in mpkCCD cells, WNK1 sgRNA CRISPR Lentiviral vector (pLenti-U6-sgRNA-PGK-Neo; #504391140195, Applied Biological Materials) or RICTOR sgRNA CRISPR Lentiviral vector (#395791140195, Applied Biological Materials) and Cas9-expressing lentiviral vector (pLenti-EF1a-Cas9-Puro; #K002, Applied Biological Materials) were used. The lentiviral plasmid DNA were packaged into lentivirus by co-transfection with Virapower (Invitrogen) in HEK-293FT cells. Supernatant containing lentivirus was used to infect mpkCCD cells and then infected cells were selected in puromycin (5 μg/ml) and neomycin (for WNK1^−/−) or hygromycin (for RICTOR^−/−). Single colonies were selected either by fluorescence-activated cell sorting (FACS) into a 96-well plate, or using cloning rings and expanded and tested for WNK1 and Rictor expression by western blotting.

WNK1 SIN1 double-knockout (WNK1^−/− SIN1^−/−) HEK-293T cells were generated by knocking out the SIN1 gene in HEK-293T WNK1^−/− cells (Roy et al., 2015) cells using the CRISPR/Cas9 Lentiviral system. SIN1 sgRNA CRISPR/Cas9 (All in one) lentiviral vector containing a sgRNA targeting the 3′ end of exon 3 of the human MAPKAP1 genomic locus, and a Cas9 (Gleason et al., 2019) expressing locus was used for lentivirus production. HEK-293 WNK1^−/− cells were then infected with the lentivirus and subsequently single clones of the WNK1 SIN1 double-knockout cell lines (WNK1^−/− SIN1^−/−) were generated as described above. The knockout cells were tested for SIN1 expression by western blotting.

For electrophysiological measurements, mpkCCDc14 cells were seeded on type I collagen-coated filters (Transwell, pore size 0.4 µm, Corning Costar) until the cell monolayers reached a transepithelial resistance of >1000 Ω.cm². They were then maintained in serum-free plain DMEM for at least 16 h before treatment with aldosterone (1 µM for 3 h) to induce endogenous SGK1 expression. For extracellular [K⁺] stimulation–recovery experiments, cells were adapted for at least 2 h to 1 mM or 3 mM [K⁺] on the basolateral side. At t=0, the medium [K⁺] was increased by addition of potassium gluconate or KCl or equimolar sodium gluconate or sodium chloride and incubated for 1 h prior to measurement of amiloride-sensitive current. Transepithelial resistance and potential difference across the cell monolayer were measured using Millepore ERS voltohmmeter (MilliCell ERS; Millipore) (Lu et al., 2010). The equivalent short-circuit current was calculated using Ohm's law. Amiloride (10 µM; cat. #A7410, Sigma-Aldrich) was added to the medium on the apical side at the end of the experiments to derive amiloride-sensitive component of the current. Amiloride almost completely (>98%) inhibited the total current, indicating its ENaC dependence. In some experiments, cells were adapted to 5 mM [K⁺] and then transferred to media with different [K⁺] as specified.

In inhibitor study experiments, cells were treated for 15 min with inhibitors – WNK463 (0.1–1 µM, apical and basolateral sides; cat. #A16205, Adooq Bioscience), AZD8055 (1 µM, apical and basolateral sides; cat. #S1555, SelleckChem), or equal volumes of vehicle as the control prior to raising basolateral K⁺ concentration from 1 to 5 mM.

In multichannel cell attached mode net current (in pA) is measured in individual patches at specific holding potentials in cell attached mode. The measured pA signal in this case reflects the sum of all the channel activities in one patch.

To determine protein expression levels, western blot analysis was performed as previously described (Lu et al., 2010; Sørensen et al., 2019). Briefly, cells were lysed in 1% Triton X-100 buffer (40 mM HEPES pH 7.5, 1 mM EDTA pH 8, 10 mM sodium pyrophosphate, 10 mM glycerophosphate, 50 mM sodium fluoride, 120 mM sodium chloride and 1% Triton X-100) containing complete protease inhibitor cocktail and PhoSTOP phosphatase inhibitor. Total protein (40 μg) from each cell extract supernatant was electrophoresed on 8% polyacrylamide gels. The blots were then probed with primary antibodies (Table S1) followed with horseradish peroxidase (HRP)-conjugated secondary antibodies. Each experiment was repeated more than three times. The bands corresponding to the proteins of interest were quantified on scanned films using NIH Image J software, as previously described.

Cell lysates were prepared using 1% CHAPS buffer (40 mM HEPES pH 7.5, 1 mM EDTA pH 8, 10 mM sodium pyrophosphate, 10 mM glycerophosphate, 50 mM sodium fluoride, 120 mM sodium chloride and 1% CHAPS) containing complete protease inhibitor cocktail and PhoSTOP phosphatase inhibitor (cat. #04906837001, Roche), and centrifuged at 14,000 g for 10 min. Supernatant (cell extract) was removed and protein content was estimated by performing a Bradford assay. Immunoprecipitation was performed using anti-FLAG M2 affinity gel (#A2220, Sigma-Aldrich), anti-Myc agarose gel (#20168, Thermo Fisher Scientific), and anti-V5-affinity gel (#A7345, Sigma). To immunoprecipitate Flag–SGK1, SIN1–V5 and Myc–WNK1, 100–200 µg of cell extract protein was rotated overnight at 4°C with 20 µl of 50% slurry anti-FLAG or anti-V5 affinity gel or 25% slurry anti-Myc affinity gel respectively. The agarose beads were collected by centrifugation (5000 g for 1 min), washed three times with 1% CHAPS cell lysis buffer. Protein was eluted from the beads by adding Laemmli sample buffer and subsequently boiled and denatured, and then separated by SDS-PAGE and transferred to PVDF membrane as described under immunoblotting. Each experiment was repeated at least three times independently with similar results.

GraphPad Prism (GraphPad Software, USA) and Microsoft Excel were used for data analysis. Comparison between two groups was performed by unpaired two-tailed t-test. Comparison between more than two groups was analyzed by one-way ANOVA with a Bonferroni's multiple comparison post-test. Statistical test used are stated in each figure legend. P<0.05 is considered significant and P-values are either shown directly in the figures or symbols for P-values are explained in figure legends.

We thank Dr Arohan R. Subramanya, (University of Pittsburgh) for providing WNK1-knockout HEK-293 cells, Drs Melanie Cobb (UTSW), Chou-long Huang (UIOWA), Chao-Ling Yang and David Ellison (OHSU) for WNK constructs, Dr Lily Jan (UCSF) for Cl-YFP plasmid, Drs Johannes Loffing and Dominique Loffing-Cueni for anti-αENaC antibody and helpful discussions, Dr Jacob Erikson for help with intracellular Cl^– measurements, and Dr Aylin Rodan (University of Utah) for helpful suggestions. We are grateful to Prof. Jens Leipziger and Prof. Alicia McDonough (USC), for helpful discussions. Wen-Hui Wang is gratefully acknowledged for advice and instruction in patch-clamp techniques. We also thank the Center for Advanced Technology and Fabrication and Design Center of UCSF for their technical support.

The peer review history is available online at https://journals.biologists.com/jcs/lookup/doi/10.1242/jcs.260313.reviewer-comments.pdf.