Kidney-specific WNK1 regulates sodium reabsorption and potassium secretion in mouse cortical collecting duct

Chih-Jen Cheng, Michel Baum, Chou-Long Huang


Kidney-specific with-no-lysine kinase 1 (KS-WNK1) is a kinase-deficient variant of WNK1 that is expressed exclusively in the kidney. It is abundantly expressed in the distal convoluted tubule (DCT) and to a lesser extent in the cortical thick ascending limb (cTAL), connecting tubule, and cortical collecting duct (CCD). KS-WNK1 inhibits Na+-K+-2Cl- and sodium chloride cotransporter-mediated Na+ reabsorption in cTAL and DCT, respectively. Here, we investigated the role of KS-WNK1 in regulating Na+ and K+ transport in CCD using in vitro microperfusion of tubules isolated from KS-WNK1 knockout mice and control wild-type littermates. Because baseline K+ secretion and Na+ reabsorption were negligible in mouse CCD, we studied tubules isolated from mice fed a high-K+ diet for 2 wk. Compared with that in wild-type tubules, K+ secretion was reduced in KS-WNK1 knockout CCD perfused at a low luminal fluid rate of ∼1.5 nl/min. Na+ reabsorption and the lumen-negative transepithelial potential difference were also lower in the KS-WNK1 knockout CCD compared with control CCD. Increasing the perfusion rate to ∼5.5 nl/min stimulated K+ secretion in the wild-type as well as knockout CCD. The magnitudes of flow-stimulated increase in K+ secretion were similar in wild-type and knockout CCD. Maxi-K+ channel inhibitor iberiotoxin had no effect on K+ secretion when tubules were perfused at ∼1.5 nl/min, but completely abrogated the flow-dependent increase in K+ secretion at ∼5.5 nl/min. These findings support the notion that KS-WNK1 stimulates ROMK-mediated K+ secretion, but not flow-dependent K+ secretion mediated by maxi-K+ channels in CCD. In addition, KS-WNK1 plays a role in regulating Na+ transport in the CCD.

  • WNK kinase
  • cortical collecting duct
  • kidney-specific WNK1
  • in vitro microperfusion

with-no-lysine (WNK) kinases are a family of serine/threonine protein kinases characterized by atypical placement of catalytic lysine residue in subdomain I, rather than in subdomain II as in other protein kinases. In mammals, four members of WNK kinases (WNK1–4) encoded by different genes have been identified and they are widely expressed in tissues (7). Besides the full-length transcript, several variants of WNK1 derived from either alternative splicing and/or alternative translation initiation increase the diversity of function of WNK1. One of these is the kidney-specific WNK1 (KS-WNK1) that is exclusively expressed in the kidney and arises from translation initiated at a unique alternate exon 4A localized between exons 4 and 5 (25). Missing the first four exons of the WNK1 gene that encodes the core of kinase domain and preceding regions, the translated KS-WNK1 protein is shorter than the full-length WNK1 and lacks the kinase activity.

The functions of WNK kinases were first revealed in 2001 by the discovery that mutations of WNK1 and WNK4 cause a human hereditary hypertensive and hyperkalemic disease, pseudohypoaldosteronism type II (23). This discovery has led to many subsequent studies and current understanding that WNK1 and 4 kinases play important roles in renal Na+ and K+ transport through regulation of Na+ and K+ transporters/channels (6, 8). KS-WNK1 is thought to participate in renal Na+ and K+ handling through antagonism of full-length WNK1 in the distal nephron (5, 10, 17, 21). In studies using Xenopus laevis oocytes and mammalian cultured cells, KS-WNK1 reverses the ability of full-length WNK1 to enhance endocytosis of ROMK K+ channels and to activate sodium chloride cotransporter (NCC) and epithelial Na+ channel (ENaC). These results suggest that KS-WNK1 may play a role in regulating K+ secretion and Na+ reabsorption in connecting tubule (CNT) and cortical collecting duct (CCD) where ROMK and ENaC are expressed. Using quantitative RT-PCR of individually isolated tubules, we recently found that KS-WNK1 is also expressed in CNT and CCD (1).

Our group and another group led by Hadchouel et al. (4, 13) independently generated mouse models of KS-WNK1 gene knockout (KO) by deleting the initiating exon 4A for KS-WNK1. KS-WNK1-KO mice manifest mild extracellular fluid volume expansion consistent with renal Na+ retention as evidenced by low urinary aldosterone excretion, upregulation of Na+-K+-2Cl (NKCC2) and NCC transporters, and elevated blood pressure when fed a high-salt diet. Hadchouel et al. found downregulation of ENaC protein abundance in the distal nephron in WNK1-KO mice (4). While these studies were indirect, they suggested that downregulation of ENaC is a compensatory response to the increased expression of NCC in the upstream distal convoluted tubule (DCT). In studies by our laboratory and Hadchouel et al., the effect of KS-WNK1-KO on renal K+ transport was less clear. It is also unclear whether KS-WNK1 regulates ROMK and/or the maxi-K+ channel, which are both present in the CCD. The purpose of the present study is to directly examine the potential role of KS-WNK1 in regulating Na+ and K+ transport in CCD using in vitro microperfusion of isolated tubules.



KS-WNK1-KO mice were generated by deleting exon 4A from KS-WNK1 in mice of a pure 129/sv background (13). These experiments were performed on KS-WNK1-KO mice at 8–10 wk of age, and age- and gender-matched wild-type littermates (129/sv). Mice were raised in a 12:12-h day-night cycle and fed a control-K+ (1% KCl) or a high-K+ (10% KCl, Harlan Teklad) diet and tap water ad libitum for 2 wk before experiments. All of the experimental procedures involving these animals were carried out in accordance with relevant laws and institutional guidelines approved by the University of Texas Southwestern Medical Center at Dallas Institutional Animal Care and Use Committee.

In vitro microperfusion of CCD.

After the mouse was killed, the kidney was removed quickly, sliced in thin coronal sections, and placed in Hanks' solution containing (in mM) 137 NaCl, 5 KCl, 0.8 MgSO4, 0.33 Na2HPO4, 0.44 KH2PO4, 1 MgCl2, 10 tris (hydroxymethyl) amino methane hydrochloride, 0.25 CaCl2, 2 glutamine, and 2 l-lactate at 4°C. CCD segments were then dissected under free hand with sharpened Dumont #5 forceps without treatment of collagenase and then transferred to a 1-ml temperature-controlled bathing chamber. Tubules were perfused in vitro as previously described (18).

Isolated CCDs were perfused at either a slow rate (1–2 nl/min) or a fast rate (4∼6 nl/min). The perfusate contained (in mM) 115 NaCl, 25 NaHCO3, 2.3 Na2HPO4, 10 Na acetate, 1.8 CaCl2, 1 MgSO4, 5 KCl, 8.3 glucose, and 5 alanine and had an osmolality equal to that of the bathing solution which contained 6 g/dl of albumin. There were at least three measurements of the perfusion and the collected tubular fluid in each experimental condition. Tubular fluid samples were collected under water-saturated light mineral oil by timed filling of a precalibrated 25-nl volumetric constriction pipette at slow and fast flow rates. The fluid collection rate (fluid volume divided by time to fill the constant volume pipette) was varied by adjusting the height of the perfusate container. The perfusion rate is equal to the collection rate as water transport is zero in the absence of transepithelial osmotic gradients and vasopressin. Na+ and K+ fluxes (JNa, JK) were calculated using the equation: JNa (pmol·min−1·mm−1) = ([Na]perfusate − [Na]collected)(VL/L); JK (pmol·min−1·mm−1) = ([K]perfusate − [K]collected)(VL/L), where VL is the collection rate and L indicates the tubular length (0.4–0.8 mm). The mean of calculated ion fluxes for each collection in the same experiment was used to calculate the flux for that period. The transepithelial potential difference (PDTE) was determined using the perfusion pipette as a bridge into the tubular lumen and referenced to the bathing solution using a Keithley 6517A programmable electrometer (Cleveland, OH). Values for each period were averaged. Amiloride and barium were used to inhibit ENaC and ROMK, respectively (Sigma, St. Louis, MO). Iberiotoxin was used to inhibit maxi-K+ channel (Tocris Bioscience, Bristol, UK). The Na+ and K+ concentrations in perfusate and collected fluid were measured using Na+- and K+-selective electrodes, respectively (Sodium Ionophore II-Cocktail A, Potassium Ionophore I-Cocktail B, Fluka) (1, 15).

Statistical analysis.

All results are expressed as means ± SE. Statistical comparisons between two groups of data were made using two-tailed unpaired Student's t-test. Repeated-measures one-way ANOVA was used when the same tubule was analyzed at multiple time points. P value <0.05 (*) was considered to be statistically significant. All statistical analyses were performed using GraphPad Prism, version 5.0 (GraphPad Software, La Jolla, CA).


Reduced K+ secretion in CCDs isolated from KS-WNK1-KO mice.

We recently showed that KS-WNK1 is expressed in CCD (1). Here, we examined the role of KS-WNK1 in K+ secretion in CCDs isolated from KS-WNK1-KO mice and control littermates. The transepithelial K+ flux (JK) was measured using in vitro microperfusion. Previous studies using in vitro microperfusion found that, unlike rabbit tubules, rodent CCDs have relatively smaller Na+ and K+ transport under the basal unstimulated conditions, and these transport activities can be enhanced by feeding animals a high-K+ diet or administration of mineralocorticoids (22). Consistent with these previous observations, we found that K+ secretions (net basal-to-lumen fluxes; −JK) in CCDs from both wild-type and KS-WNK1-KO mice under a control-K+ diet were very small and not statistically significant from each other (Fig. 1; Table 1). K+ secretion was markedly increased (>20-fold) in both control and KS-WNK1-KO tubules isolated from mice fed a high-K+ diet. Compared with that in the wild-type tubules, K+ secretion in the KS-WNK1-KO tubules was significantly reduced. High luminal flow stimulates K+ secretion in CCD (9, 24). We found that increasing the luminal fluid perfusion rate from the basal ∼1.5 to ∼5.5 nl/min stimulated K+ secretion in wild-type as well as KO tubules from mice fed a high-K+ diet. Of note, the magnitude of the flow-stimulated increase in K+ secretion (differences between ∼1.5- and ∼5.5-nl/min flow rates) was similar in wild-type and KS-WNK1-KO tubules, indicating that KS-WNK1 does not affect flow-stimulated K+ secretion (see also Fig. 6 below).

Fig. 1.

Effect of deletion of kidney-specific with-no-lysine kinase 1 (KS-WNK1) on K+ secretion in the cortical collecting duct (CCD) under different dietary K+ content and perfusion rates. K+ secretion rate (basal-to-lumen K+ flux; −JK) was measured in the CCDs isolated from wild-type (WT) or KS-WNK1-knockout (KO) mice fed a control-K+ (1% KCl; CK) or high-K+ (10% KCl; HK) diet for 2 wk (n = 8–10 for each group). *P < 0.05 between indicated groups. #P < 0.001 high-K+ vs. control-K+ diet. †P < 0.05 fast vs. slow perfusion rate. NS, not significant.

View this table:
Table 1.

Summary of mean values for JK, JNa, and PDTE

Reduced lumen-negative PDTE and Na+ reabsorption in the KS-WNK1-KO CCD.

We next examined Na+ reabsorption (net lumen-to-basal Na+ fluxes; JNa) and PDTE in KS-WNK1-KO vs. wild-type CCDs. The lumen-negative PDTE, an index of ENaC-mediated electrogenic Na+ reabsorption, was barely detected in the CCDs of mice fed a normal-K+ diet (Fig. 2A, Table 1), indicating a low ENaC-mediated Na+ reabsorption under this condition. A high-K+ diet greatly enhanced the lumen-negative PDTE in both KO and wild-type CCDs. The high-K+ diet-induced augmentation of lumen-negative PDTE was completely abolished by an ENaC inhibitor amiloride (data not shown), supporting that lumen-negative PDTE is mediated by ENaC-dependent Na+ reabsorption. The magnitude of PDTE in KS-WNK1-KO CCD was significantly less than that in wild-type tubule (Fig. 2A). Similarly, Na+ reabsorption was reduced in KS-WNK1-KO CCD relatively to the wild-type tubule (Fig. 2B). Together, these results support the notion that genetic deletion of KS-WNK1 decreases ENaC-mediated Na+ reabsorption in CCD.

Fig. 2.

Effect of deletion of KS-WNK1 on the lumen-negative potential difference (A) and Na+ reabsorption rate (B) under different dietary K+ content and perfusion rates. Lumen-negative transepithelial potential difference (PDTE) and Na+ reabsorption rate (lumen-to-basal Na+ flux; JNa) were measured in the CCDs isolated from WT or KS-WNK1-KO mice fed a control-K+ or high-K+ diet for 2 wk (n = 8–10 for each group). *P < 0.05 between indicated groups. †P < 0.05 fast vs. slow perfusion rate.

It is known that an increase in luminal flow and shear stress activates ENaC and Na+ reabsorption in CCD (2, 20). In agreement with these reports, we found that Na+ reabsorptions in wild-type and KO CCDs were both stimulated by increasing perfusion rate (Fig. 2B). The overall Na+ reabsorption in the KO tubule during high luminal flow remained lower than the wild-type. The lumen-negative PDTE, however, was not increased by the high luminal flow (Fig. 2A). The precise reason why PDTE is not increased along with JNa during high luminal flow is unknown.

Identification of ROMK- and maxi-K+ channel-mediated K+ secretion in CCD.

The last step of transepithelial K+ secretion in CCD is passive diffusion of K+ through apical K+ channels driven by the lumen-negative potential generated by Na+ reabsorption via ENaC. Accordingly, luminal application of a nonspecific K+ channel inhibitor barium markedly inhibited K+ secretion in CCD, and inhibition of ENaC by amiloride further reduced the remaining K+ secretion (Fig. 3). The lumen-negative PDTE was also decreased to near zero by luminal application of amiloride (data not shown). ROMK is a constitutively open apical K+ channel responsible for baseline K+ secretion in the setting of low luminal flow. Maxi-K+ is another apical K+ channel that opens in response to flow-stimulated increases in the intracellular Ca2+ level and thus primarily responsible for flow-stimulated K+ secretion (11). In agreement with this notion, luminal application of a maxi-K+ channel inhibitor iberiotoxin completely inhibited flow-stimulated K+ secretion (Fig. 4A), but had no effect on baseline K+ secretion in wild-type CCD under low flow (Fig. 4B). The specificity of iberiotoxin for maxi-K+ channels is further supported by the finding that flow-stimulated K+ secretion (difference between low and high flow rate) is essentially identical to the iberiotoxin-sensitive component. Thus, at low flow rate all K+ secretion in the CCD is via ROMK.

Fig. 3.

Effect of barium and amiloride on K+ secretion in isolated WT CCD on a high-K+ diet. The rate of K+ secretion was measured in WT CCDs microperfused in vitro at a constant slow (∼1.5 nl/min) perfusion rate. Barium chloride (BaCl2; 2 mM) and amiloride (100 μM) were added to the perfusate in sequence as indicated. Connected dots indicate the same tubule.

Fig. 4.

Effect of flow and iberiotoxin (IBX) on K+ secretion in isolated WT CCD on a high-K+ diet. The rate of K+ secretion was measured in WT CCDs microperfused in vitro at sequential flow rates. A: IBX (50 nM) was added to the perfusate when CCDs were perfused at a rate of ∼5.5 nl/min. B: IBX was added to the perfusate when CCDs were perfused at a rate of ∼1.5 nl/min, followed by a higher rate. Connected dots indicate the same tubule.

Deletion of KS-WNK1 decreases ROMK-mediated but not maxi-K+-mediated K+ secretion in CCD.

Having validated our experimental approach to distinguish between ROMK and maxi-K+-mediated K+ secretion, we examined the effect of KS-WNK1 deletion on K+ secretion through these pathways. We measured K+ secretion in CCDs isolated from KS-WNK1-KO mice fed a high-K+ diet under low and high flow rate and under high flow rate in the presence of iberiotoxin. Increasing flow stimulated K+ secretion in KS-WNK1-KO CCD and iberiotoxin completely inhibited flow-stimulated K+ secretion (Fig. 5).

Fig. 5.

Effect of flow and IBX on K+ secretion in isolated KS-WNK1-KO CCD on a high-K+ diet. The rate of K+ secretion was measured in KS-WNK1-KO CCDs microperfused in vitro at sequential flow rates. IBX (50 nM) was added to the perfusate when CCDs were perfused at ∼5.5 nl/min. Connected dots indicate the same tubule.

Figure 6 summarizes results of ROMK-mediated and maxi-K+-mediated K+ secretion in wild-type and KS-WNK1-KO CCDs. We attributed K+ secretion under low perfusion rate as ROMK-mediated and the iberiotoxin-sensitive K+ secretion (difference before and after drug) under high flow rate as maxi-K+ mediated. As shown in Fig. 6, ROMK-mediated K+ secretion in KS-WNK1-KO CCD was decreased by >50% compared with wild-type tubule, whereas maxi-K+-mediated K+ secretion was not different between KO and wild-type tubules. Thus, these data are consistent with KS-WNK1 regulating ROMK but not the maxi-K+ channel in the CCD.

Fig. 6.

Effect of KS-WNK1 on ROMK- and maxi-K+-mediated K+ secretion. ROMK (left) and maxi-K+ (right)-mediated K+ secretion in the CCDs isolated from WT or KS-WNK1-KO mice were analyzed as described in the text (n = 8–10 for each group). *P < 0.05 between indicated groups.


In this study, we showed that deletion of KS-WNK1 in mice decreases Na+ reabsorption and K+ secretion in CCD of mice fed a high-K+ diet. The effect of KS-WNK1 on K+ secretion is primarily from ROMK-mediated K+ secretion, and KS-WNK1 deletion has no detectable effect on flow-stimulated K+ secretion mediated by maxi-K+.

Several previous studies provide information for potential mechanism by which KS-WNK1 might regulate Na+ reabsorption in CCD. Hadchouel et al. (4) reported that the expression of ENaC is downregulated in KS-WNK1-KO mice. The authors found that NCC is upregulated in KS-WNK1-KO mice but the systemic blood pressure is only mildly elevated, and they suggested that the decrease in ENaC expression is a compensatory response to the increase in NCC expression in the upstream DCT. In contrast, cell-based studies using X. laevis oocytes and mammalian cultured cells suggested that KS-WNK1 inhibits ENaC by antagonizing full-length WNK1 activation of ENaC via serum- and glucocorticoid-induced kinase (5). It is possible that KS-WNK1 could inhibit ENaC in CCD by antagonism of full-length WNK1. However, the finding that ENaC-mediated Na+ reabsorption is reduced in KS-WNK1-KO CCD is not consistent with this being the predominant mechanism for regulation of ENaC by KS-WNK1. Instead, the present results support the notion that compensatory downregulation of ENaC is the primary cause for the decrease in Na+ reabsorption in KS-WNK1-KO mice. Reciprocal compensatory changes of ENaC expression in response to changes in Na+ reabsorption in the upstream tubular segments have been well-described. For example, genetic inactivation of NCC in mice increases the expression of ENaC (14). Alterations of aldosterone levels probably underlie the compensatory changes of ENaC expression.

With respect to the mechanism of regulation of K+ secretion by KS-WNK1, cell-based studies have shown that full-length WNK1 inhibits ROMK by enhancing its endocytosis and KS-WNK1 stimulates the channel by reversing the inhibition (10). In support of these in vitro results, mice with transgenic overexpression of KS-WNK1 in the kidney have increased expression of ROMK in the distal nephron (12). Our finding of deletion of KS-WNK1 decreases ROMK-mediated K+ secretion further supports the notion that KS-WNK1 is an important physiological stimulator of ROMK expression and K+ secretion in CCD. Reduced Na+ reabsorption may also contribute to the decrease in K+ secretion in KS-WNK1-KO CCD, but it is probably not the predominant mechanism. Flow-stimulated K+ secretion via maxi-K+ is at least partially dependent on Na+ reabsorption (16), yet we found that maxi-K+ is not significantly affected in KS-WNK1-KO mice.

Probably due to compensatory responses from other tubular segments, disturbances in the overall K+ homeostasis in KS-WNK1-KO mice appear to be relatively modest. In a recent study, we found that serum K+ levels and daily urinary K+ excretion were not significantly different between KS-WNK1-KO and wild-type mice on a control-K+ diet (1). We also found that KS-WNK1 inhibits NKCC2 in cortical thick ascending limb (cTAL) measured using isolated tubular microperfusion and that NKCC2-mediated Na+ reabsorption is increased in cTAL isolated from KS-WNK1-KO mice. Inhibition of NKCC2 is believed to play an important role in causing diuresis and natriuresis, and thus flow-stimulated kaliuresis during a high-K+ diet. The increase in Na+ reabsorption in cTAL in KS-WNK1-KO mice would be expected to impair kaliuresis during a high-K+ diet. The impairment in kaliuresis in KS-WNK1-KO mice on a high-K+ diet, however, is only transient, occurring during the second and third day of high-K+ intake. We suggested that high-K+ diet-induced inhibition of Na+ reabsorption in the proximal tubule (where KS-WNK1 is not normally expressed) augments flow-stimulated K+ secretion and counterbalances the impairment of K+ excretion resulting from increased NKCC2 function in KS-WNK1-KO mice. Our current finding that KS-WNK1-KO does not affect flow-stimulated K+ secretion by maxi-K+ in CCD is consistent with this notion. Other possible mechanisms may explain the relatively mild disturbances in the overall K+ homeostasis in KS-WNK1-KO mice.

In principal cells of the aldosterone-sensitive distal nephron, K+ secretion via apical K+ channels causes hyperpolarization of the apical membrane limiting further K+ efflux. ENaC-mediated Na+ influx depolarizes the apical membrane potential to maintain the driving force for K+ efflux through K+ channels. Thus, K+ secretion is dependent on Na+ reabsorption in the aldosterone-sensitive distal nephron. Recently, Frindt and Palmer (3) reported that while K+ secretion under a control-K+ diet is amiloride-sensitive, a large fraction of K+ secretion in rats after being fed a high-K+ diet is amiloride-insensitive, thus independent of ENaC activity. This interesting finding makes sense from the physiological perspective. Assuming that ENaC in the distal nephron mediates reabsorption of 2–3% of glomerular-filtered Na+ load, ∼5–10% of the total ENaC-mediated Na+ reabsorption would be devoted to the exchange for K+ secretion under a control-K+ diet. In the setting of high-K+ intake, particularly in experimental animals where a 5 to 10 time higher K+ load is given, Na+-dependent K+ secretion would amount to ∼50% of Na+ reabsorption via ENaC. A mechanism for Na+-independent K+ secretion, such as via parallel K+ and Cl secretion (19), would allow animals to uncouple K+ secretion from Na+ reabsorption and volume homeostasis. On the surface, our finding that amiloride completely inhibits K+ secretion in CCD from high K+-treated mice seems to contradict to the notion of ENaC-independent K+ secretion. However, our study used conventional isotonic solutions in which both luminal and bath fluids contained 155 mM Na+ and 125 mM Cl. This perfusion fluid NaCl concentration favors Na+ reabsorption rather than Cl secretion in exchange for K+ secretion. The normal physiological concentration of NaCl in fluid entering CCD in vivo is ∼10–20 mM, which would instead favor parallel K+ and Cl secretion if pathway(s) for Cl secretion is indeed upregulated under a high-K+ diet. Future studies will investigate K+ secretion in CCD under a high-K+ diet using perfusates resembling in vivo physiological fluid composition.


This work was supported by National Institutes of Health Grants [UTSW O'Brien Center Grant DK79328, DK41612 (to M. Baum), and DK59530 (to C. L. Huang)]. C. J. Cheng is supported by a scholarship grant from the Ministry of Defense, Taiwan.


No conflicts of interest, financial or otherwise, are declared by the author(s).


Author contributions: C.-J.C., M.B., and C.-L.H. conception and design of research; C.-J.C. performed experiments; C.-J.C., M.B., and C.-L.H. analyzed data; C.-J.C., M.B., and C.-L.H. interpreted results of experiments; C.-J.C. and C.-L.H. prepared figures; C.-J.C. drafted manuscript; C.-J.C., M.B., and C.-L.H. edited and revised manuscript; C.-J.C., M.B., and C.-L.H. approved final version of manuscript.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
View Abstract