Disease-causing mutations in KLHL3 impair its effect on WNK4 degradation

Abstract

Mutations in with-no-lysine (K) kinase 4 (WNK4) and a ubiquitin E3 ligase complex component kelch-like 3 (KLHL3) both cause pseudohypoaldosteronism II (PHAII), a hereditary form of hypertension. We determined whether WNK4 or its effector is regulated by KLHL3 in Xenopus oocytes. KLHL3 inhibited the positive effect of WNK4 on Na⁺-Cl⁻ cotransporter (NCC) by decreasing WNK4 protein abundance without decreasing that of NCC and the downstream kinase OSR1 directly. Ubiquitination and degradation of WNK4 were induced by KLHL3. The effect of KLHL3 on WNK4 degradation was blocked by a dominant negative form of cullin 3. All five PHAII mutations of KLHL3 tested disrupted the regulation on WNK4. We conclude that KLHL3 is a substrate adaptor for WNK4 in a ubiquitin E3 ligase complex.

1. Introduction

Pseudohypoaldosteronism II (PHAII), also known as familial hyperkalemia and hypertension or Gordon’s syndrome, is a Mendelian form of hypertension with hyperkalemia, mild metabolic acidosis, and low plasma renin [1;2]. In 2001, mutations in WNK1 and WNK4, two genes of with-no-lysine (K) (WNK) serine/threonine kinase family, were identified to be associated with PHAII [3]. Among the two genes, WNK4 harbors missense mutations that may alter the function/regulation of WNK4 protein [3]. Indeed, PHAII mutations in the acidic motif alter Ca²⁺-sensitivity of WNK4 kinase [4]. The R1185C mutation, on the other hand, resides inside a calmodulin binding domain in the vicinity of three phosphorylation sites for the serum- and glucocorticoid-induced protein kinase 1 (SGK1) [5–7]. The calmodulin binding and SGK1 phosphorylation sites are parts of a regulatory domain that inhibits WNK4 activity, and the R1185C mutation alleviates the inhibitory effect [5]. WNK4 is an integrative regulator of renal electrolyte transport with the thiazide-sensitive Na⁺-Cl⁻ cotransporter (NCC) as the main target [8–10]. The effects of WNK4 on NCC are complex: on the one hand, WNK4 inhibits NCC and increases its lysosomal degradation [11–14]; on the other, NCC is stimulated by WNK4 via proline-alanine-rich STE20-related kinase (SPAK) and oxidative stress-responsive 1 (OSR1) in a phosphorylation cascade [10;15;16]. Furthermore, in the presence of angiotensin II, the positive effect overtakes the inhibitory effect of WNK4 in a SPAK-dependent manner [17]. The PHAII-causing WNK4 mutants up-regulate NCC as if WNK4 is activated by angiotensin II [17]. Thus, PHAII mutations likely disrupt the Ca²⁺-dependent activation mechanism of WNK4 and lock WNK4 at a state equivalent to that stimulated by the renin-angiotensin-aldosterone system [4;5].

A decade after the identification of the association of WNK1 and WNK4 mutations with PHAII, mutations in two genes of a likely cullin-RING ubiquitin E3 ligase complex, kelch-like 3 (KLHL3) and cullin 3 (CUL3) were also identified to cause PHAII [18;19]. CUL3 is a scaffold protein that connects the substrate adaptor protein and RING-box 1 ubiquitin ligase E3 in the cullin-RING ubiquitin E3 ligase complexes [20]. Members of the kelch-like family were shown to serve as substrate adaptors in these complexes [21;22]. Thus, it is likely that KLHL3 and CUL3 are part of a cullin-RING ubiquitin E3 ligase complex that specifically targets renal electrolyte transporters or their regulators involved in Na⁺ and/or K⁺ transport. While CUL3 is ubiquitously expressed in the kidney, KLHL3 is highly expressed in distal convoluted tubule (DCT) where WNK4 and NCC are expressed [18;19]. KLHL3 decreases NCC plasma membrane abundance when they were co-transfected into cells [19]. It is unclear, however, which step of the regulation cascade is the target of the ubiquitin E3 ligase complex containing KLHL3.

In this study, we provide evidence that WNK4, rather than the downstream targets of WNK4, is a substrate of the E3 ligase complex containing KLHL3. PHAII mutations in KLHL3 may lead to elevated WNK4 protein abundance and therefore elevated WNK4 activity, leading to disease.

2. Materials and methods

2.1. cDNA constructs

The human WNK4 cDNA was a generous gift from Dr. Xavier Jeunemaitre and was used previously [4;5;23]. The cDNA for mouse NCC (BC038612) [4] was purchased from Open Biosystems (Huntsville, AL). The human OSR1 and KLHL3 cDNAs were cloned from HEK-293 cells and human CUL3 cDNA was cloned from SW480 cells using reverse transcription PCR approach. These cDNAs were subcloned into Xenopus laevis (X. laevis) oocytes expression vector pIN and their sequences were verified by sequencing. FLAG, HA, or GST tag was added to NH₂-termini of the cDNAs for detecting protein expression or for GST pull-down experiments. The PHAII mutations in KLHL3 were generated using QuikChange II site-directed mutagenesis kit (Stratagene, La Jolla, CA).

2.2. Western blot analysis

Western blot analyses were performed as described previously [23]. Lysates from ten oocytes were extracted in each group. Monoclonal anti-HA antibody (product # H9658, 1:5000 dilution) and anti-FLAG antibody (F7425, 1:2,000 dilution) were purchased from Sigma-Aldrich (St. Louis, MO). Anti-GST (27-4577-01, 1:2000) was purchased from GE Healthcare Life Sciences (Piscataway, NJ). Antibody specific for phosphorylated serine 325 of OSR1 was purchased from EMD Millipore (product # 07-2273, Billerica, MA). The anti-β-actin antibody (sc-47778, 1:5,000) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Appropriate horseradish peroxidase (HRP)-conjugated secondary antibodies (1:5,000 dilution) were purchased from Santa Cruz Biotechnology. Chemiluminescence signals were detected using SuperSignal West Femto Maximum Sensitivity Substrate kit or SuperSignal West Pico Chemiluminescent Substrate kit (Pierce Biotechnology, Rockford, IL).

2.3. GST pull-down assay

X. laevis oocytes were injected with cRNAs encoding GST (as control), GST-WNK4, GST-CUL3, or HA-KLHL3 and were cultured in 0.5 × L-15 medium at 18 °C. Two days after injection, 50 oocytes were lysed with lysis buffer (NaCl 100 mM, Tris·Cl 20 mM, Triton X100 1%, plus protease inhibitor cocktail, pH 7.6) at 20 µl/oocyte. After vigorous vortex, the oocytes were centrifuged at 3,500 g for 10 min at 4 °C to remove the cellular debris and yolk proteins. The supernatant of protein extract from GST or GST-WNK4 group was incubated with 50 µl slurry glutathione Sepharose (GE Healthcare, Piscataway, NJ) for 2 hrs at room temperature. The Sepharose beads were washed 3 times with 500 µl lysis buffer, and then were incubated with 600 µl lysate from oocytes expressing HA-KLHL3. After rocking at 4 °C overnight, the Sepharose beads were washed 3 times again with 500 µl lysis buffer supplemented with protease inhibitor cocktail. Then GST, GST-CUL3, or GST-WNK4 and proteins associated with them were eluted from Sepharose beads by incubation with 10 mM L-glutathione reduced for 1 hr at 4 °C. Both the supernatants and the proteins bound to the Sepharose beads were respectively subjected to SDS-PAGE. The proteins were transferred from the SDS-PAGE to PVDF membrane and immunoblotting experiments were carried out with anti-GST (27-4577-01, GE healthcare, 1:2,000 dilution) and anti-HA (H9658, Sigma–Aldrich, 1:5,000 dilution) antibodies to determine whether HA-KLHL3 was associated with GST-fusion proteins.

2.4. Na⁺ uptake assay

X. laevis oocytes were used for Na⁺ uptake assay. The animal protocol used in this study was approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Alabama at Birmingham. X. laevis oocytes were collected and cultured in 0.5 × L-15 medium as previously described [23;24]. Oocytes were microinjected with in vitro transcribed capped complementary RNAs (cRNAs) of untagged, HA or FLAG tagged NCC, WNK4, KLHL3, CUL3 at 6.25–12.5 ng/oocyte. When two or three cRNAs were co-injected, the total amount of cRNA injected was kept under 50 ng/oocyte so that the oocyte translation system was not overwhelmed. Unless stated otherwise, the oocytes will be assayed two days (36–48 hrs) after injection by radiotracer ²²Na⁺ uptake or by extracting proteins for Western blot analysis. NCC activity was assessed by ²²Na⁺ tracer uptake in isotonic uptake solution as reported [11;12;17], with or without overnight Cl⁻ free treatment. After uptake, oocytes were washed six times and then lysed in 10% SDS solution. Radioactivity of each oocyte was determined using a scintillation counter. Statistical significance is defined as P < 0.05 by Student’s t-test or Mann-Whitney Rank Sum Test using SigmaPlot 10 (Systat Software, Inc., Chicago, IL).

3. Results and discussion

3.1. KLHL3 abolishes the effect of WNK4 on NCC by decreasing WNK4 protein abundance

KLHL3 decreases the plasma membrane level of NCC in HEK293 cells [19]. However, it was unclear whether NCC or a regulator of NCC is a substrate of the ubiquitin E3 ligase complex with KLHL3 as the substrate adaptor. We examined this further using X. laevis oocyte system that is conventionally utilized to evaluate NCC activity. Without Cl⁻ depletion treatment, WNK4 exerted a positive effect on NCC-mediated Na⁺ uptake in X. laevis oocytes (Fig. S1A). This effect was abolished when oocytes were treated in a Cl⁻ free medium overnight before uptake experiments (Fig. S1A). Consistent with previous report [15], the level of phosphorylated serine 325 (p-S325) in OSR1 was increased by WNK4 without Cl⁻ depletion; it was also elevated by Cl⁻ depletion in the absence of exogenous WNK4 (Fig. S1B). Thus, the WNK4-OSR1/SPAK-NCC pathway might have been largely masked by Cl⁻ depletion. However, the level of p-S325 did not entirely correlate with the activity of NCC (Fig.S1), suggesting the pathway activated by WNK4 and that activated by Cl⁻ depletion may overlap but may not be identical. We chose to examine the effect of KLHL3 on the WNK4-OSR1-NCC pathway without Cl⁻ depletion treatment so that the positive effect of WNK4 on NCC could be observed.

When KLHL3 was co-expressed with NCC or with both NCC and WNK4, the activity of NCC and the positive effect of WNK4 on NCC were both inhibited (Fig. 1A). Because KLHL proteins serve as substrate adaptors in CUL3-RING ubiquitin E3 ligase complexes [21;22], we reasoned that it is most likely KLHL3 is an adaptor for NCC or its regulator. We thus tested the effects of KLHL3 on the protein abundance of NCC and its regulators WNK4 and OSR1. An HA- or FLAG-tag was inserted in the NH₂-terminus of each of the proteins for assessing their abundance with Western blot analysis. Neither the abundance of NCC nor that of OSR1 was reduced by KLHL3 (Fig. 1B). In contrast, WNK4 was not detectable and p-S325 in OSR1 was lowered in the presence of KLHL3 (Fig. 1B). Similarly, KLHL3 also inhibited the positive regulation of WNK4 on NKCC2 without reducing the protein abundance of NKCC2 (data not shown).

Fig. 1.

KLHL3 inhibited the effect of WNK4 on NCC-mediated Na⁺ uptake by decreasing WNK4 protein abundance. (A) KLHL3 inhibited NCC-mediated Na⁺ uptake and blocked the effect of WNK4 on NCC when expressed in X. laevis oocytes. Data from 33–45 oocytes/group from 4 independent experiments are shown as means ± S.E. * indicates P < 0.05; NS, not significant (P > 0.05). (B) The effects of KLHL3 on the protein abundance of NCC, WNK4, and OSR1. HA-tagged NCC, WNK4, or OSR1 was expressed with HA- or FLAG-tagged KLHL3 in X. laevis oocytes and their protein abundance and that of phosphorylated serine 325 (p-S325) of OSR1 and β-actin (loading control) were assessed by Western blot analysis two days later. The levels of WNK4 and p-S325, but not those of NCC and total OSR1, were reduced by KLHL3.

These results suggest that KLHL3 is the substrate adaptor for WNK4, not for NCC or OSR1, in a ubiquitin E3 ligase complex. This does not exclude the possibility that KLHL3 regulates NCC in a way unrelated to its degradation or KLHL3 regulates other regulators of NCC because KLHL3 inhibited NCC activity without exogenous WNK4 (Fig. 1A). However, if the basal activity of NCC depends on endogenous WNK4, which has been shown to be expressed in X. laevis oocytes [25], the effect of KLHL3 on NCC could be due to the elimination of endogenous WNK4 by KLHL3. The reduction of phosphorylated and total NCC proteins in WNK4 knockout mice indicates that WNK4 is essential to NCC phosphorylation and stability [10]. Phosphorylation of NCC via the WNK4-SPAK/OSR1 pathway may increase cell surface level of NCC and prevent NCC degradation [26]. We didn’t observe a significant effect of KLHL3 on NCC protein abundance likely due to the absence of significant level of the ubiquitin E3 ligase for NCC degradation in oocytes.

3.2. Cullin 3 is involved in the regulation of WNK4 stability

CUL3 connects substrate adaptor and RING-box 1 ubiquitin ligase E3 in ubiquitin E3 ligase complexes [20]. Unlike the KLHL3, the effect of CUL3 on NCC-mediated Na⁺ uptake was modest (Fig. 2A). Based on the Expressed Sequence Tag (EST) database, cullin 3 is expressed in X. laevis oocytes. Therefore, it is likely that KLHL3 is the rate-limiting factor. As shown in Figure 2B, HA-WNK4 protein was at a steady state without HA-KLHL3 twelve hrs after injection. However, only four hrs after injection of HA-KLHL3, HA-WNK4 protein was largely degraded. The NH₂-terminal region of CUL3 interacts with substrate adaptor and thus has a dominant-negative effect on CUL3 [27]. When the NH₂-terminal 400 amino-acid segment of CUL3 (CUL3 1–400) was expressed, the effect of KLHL3 was partially blocked. In some (but not all) batches of oocytes, injection of HA-CUL 1–400 also increased the level of HA-WNK4 in a dose-dependent manner (Fig. 2C). This effect may depend on the expression of endogenous KLHL3, which is likely low and variable. These results with dominant-negative CUL3 support a role of CUL3 in regulating WNK4 stability. GST-CUL3 and GST-WNK4 but not GST alone (control) pulled down HA-KLHL3 expressed in oocytes (Fig. 2D). This indicates that KLHL3 has the ability to form a protein complex with both CUL3 and WNK4

Fig. 2.

The role of CUL3 in WNK4 stability. (A) HA-CUL3 moderately inhibited Na⁺ uptake of HA-NCC without reducing its protein abundance. Data from 33–35 oocytes/group from 4 independent experiments are shown as means ± S.E. * indicates P < 0.05. (B) Left panel, oocytes were injected with HA-WNK4 cRNA and 12 hrs later, HA-KLHL3 cRNA (or water in the control group) was injected. After 4 hrs, oocytes were treated with protein synthesis blocker cycloheximide (CHX) at 100 µg/ml. Proteins of oocytes were extracted at 12, 16, and 20 hrs after injection of HA-WNK4 cRNA. Upper panel shows the time points of experiment, and lower panel shows the level of HA-WNK4 and HA-KLHL3 at these points. Right panel, co-injection of HA-CUL3 1–400 partially blocked HA-KLHL3 induced reduction in HA-WNK4 abundance. (C) HA-CUL3 1–400 dose-dependently increases the level of HA-WNK4 protein in some batch of oocytes. (D) When the lysate of oocytes expressing GST-CUL3, GST-WNK4, or GST alone was mixed with lysate of oocytes expressing HA-KLHL3, HA-KLHL3 was pulled down by either GST-CUL3 or GST-WNK4 but not by GST alone.

3.3. KLHL3 increases the ubiquitination and degradation of WNK4

To determine the effect of KLHL3 on the ubiquitination of WNK4, we detected the ubiquitination of WNK4 using the approach we used previously [24]. We co-injected HA-ubiquitin and GST-WNK4 or GST (as control), and one day later, FLAG-KLHL3 was injected. GST-WNK4 proteins were undetectable in the samples pulled down by glutathione beads 3 and 6 hrs after FLAG-KLHL3 injection. Meanwhile, high molecular weight smears became detectable by HA antibody (Fig. 3A). Such smears were not detected in the control group that expressed GST instead of GST-WNK4 (Fig. 3B). The disappearance of GST-WNK4 band and the concomitant appearance of the smear higher than GST-WNK4 indicate that the smear represents poly-HA-ubiquitinated GST-WNK4. Because the heterogeneously polyubiquinated WNK4 molecules contained multiple HA-tags but only one GST-tag, they were detectable by HA antibody but not by GST antibody. This result indicates that KLHL3 promotes WNK4 polyubiquitination, which marks WNK4 for proteolysis.

Fig. 3.

KLHL3 increased WNK4 ubiquitination. cRNAs for HA-ubiquitin and GST-WNK4 (or GST as control) were injected into oocytes. After 12 hrs, oocytes were injected with cRNA for FLAG-KLHL3 or water as control. GST-WNK4 (A) and GST (B) proteins were pulled down at 0, 3, and 6 hrs after injection of FLAG-KLHL3 or water and the ubiquitinated proteins were detected by HA-antibody (upper panel). The levels of GST-WNK4 (or GST) in pulled down samples and FLAG-KLHL3 in oocytes lysate were detected with GST antibody and FLAG antibody, respectively.

3.4. PHAII mutations in KLHL3 decrease its ability to reduce WNK4 protein abundance

If WNK4 is the true target of KLHL3 and the regulation of WNK4 stability by KLHL3 is essential to the pathogenesis of PHAII, one would expect that PHAII mutations in KLHL3 will result in alterations in WNK4 protein abundance. KLHL3 contains BTB (Bric-a-brac, Tramtrack, Broad-complex) and BACK (BTB and COOH-terminal Kelch) domains in the NH₂-terminal region and six kelch repeats in the COOH-terminal region (Fig. 4A). The BTB and BACK domains of KLHL7 are involved in the interaction with CUL3 [22]. The kelch repeats form β-propellers for substrate recognition [28]. We tested five representative PHAII mutations, including A77E in the BTB domain, C164F in the BACK domain, and Q309R, L387P, and R528C in kelch repeats (Fig. 4A). After HA-WNK4 was co-expressed with the wild-type or individual KLHL3 mutant for 36 hrs, HA-WNK4 was undetectable in the presence of wild-type KLHL3; however, it was detectable in the presence of each of the KLHL3 mutants tested (Fig. 4B). The A77E mutant in the BTB domain completely abolished the negative effect of KLHL3 on HA-WNK4. The C164F mutation in the BACK domain also exhibited a strong effect. The mutations in the kelch repeats disrupted the suppression of HA-WNK4 to lesser extents. These results suggest that mutations in different functional domains of KLHL3 impair its function differently. Wild-type KLHL3 eliminated WNK4 protein in 3–6 hrs (Fig. 3), whereas the least effective R528C mutant didn’t completely eliminate WNK4 in 36 hrs (Fig. 4B). Thus, these PHAII mutations drastically decreased the ability of KLHL3 in reducing WNK4 protein abundance.

Fig. 4.

PHAII-causing mutants of KLHL3 were less effective in reducing WNK4 protein abundance in X. laevis oocytes. (A) Representative PHAII-causing mutations examined in this study. (B) HA-WNK4 protein abundance assessed by Western blot analysis in the presence or absence of wild-type (WT) or mutants of HA-KLHL3. HA antibody detected both HA-WNK4 and HA-KLHL3 at different molecular weights. A summary of the band intensity of HA-WNK4 from three experiments is shown in the upper panel. (C) NCC-mediated Na⁺ uptake in the presence of WNK4 and WT or mutant HA-KLHL3. Similar to the Western blot experiments in (B), uptake experiments were performed at 36 hrs after cRNA injection. Data from 18 oocytes/group derived from 2 frogs are shown as means ± S.E. * indicates P < 0.05 vs. WT HA-KLHL3 group; # indicates P < 0.05 vs. the group without HA-KLHL3.

In order to examine to what extent the remaining WNK4 level under the influence of KLHL3 mutants promotes NCC activation, we evaluated NCC-mediated Na⁺ uptake in the presence of WNK4 and individual KLHL3 mutants in oocytes. The level of NCC-mediated Na⁺ uptake in the presence of each of the five KLHL3 mutants was significantly increased compared to that in the presence of wild-type KLHL3 (Fig. 4C). Intriguingly, the NCC-mediated Na⁺ uptake was relatively low even though the WNK4 abundance was relatively high in the presence of A77E or C164F mutant (Figs. 4B & 4C). The reason for this discrepancy is unclear. Unlike PHAII mutations in the kelch repeats that likely disrupt the interaction with WNK4, A77E and C164F mutations in the BTB-BACK domain may perturb the interaction with CUL3 [22]. Thus, these two mutations may interfere with all substrates of KLHL3-CUL3 ubiquitin E3 ligase complex. If one of the substrates is an unidentified “NCC inhibitor” in oocytes, the scenario of lowered NCC activity in the presence of higher level of WNK4 could occur due to the enhanced inhibition of NCC by the inhibitor.

The results of this study indicate that WNK4 is a substrate of the E3 ligase complex with KLHL3 as a substrate adaptor. This ubiquitin E3 ligase complex specifically decreases the protein abundance of WNK4 without decreasing that of OSR1 and NCC. KLHL3 is abundantly expressed in the DCT and to a much lesser extent in the connecting tubule and collecting duct [18;19]. Thus, PHAII mutations in KLHL3 are expected to increase WNK4 abundance mostly in the DCT, where NCC is expressed. Thus, PHAII mutation in KLHL3 is not equivalent to the overall overexpression of WNK4 in the WNK4 transgenic mouse model reported previously [8].

Point mutations in either WNK4 or KLHL3 may result in PHAII [3;18;19]. PHAII mutations in WNK4 disrupt the regulation of WNK4 kinase activity and lock WNK4 at an activated state equivalent to that stimulated by angiotensin II and aldosterone [4;5;17]. The total activity of WNK4 depends not only on its specific activity, but also on the total amount of WNK4 protein. The results of this study suggest that protein abundance of WNK4 is controlled by KLHL3. PHAII mutations in KLHL3 lead to elevated WNK4 protein abundance and in turn, elevated kinase activity. Thus, PHAII mutations in both WNK4 and KLHL3 affect the same pathway. Similar conclusions were reached in two independent studies published during the revision of this manuscript [29;30]. While we emphasized the effect of KLHL3 on WNK4 in the context of NCC function and the impacts of PHAII mutations of KLHL3 on WNK4 protein abundance, these studies indicate that KLHL3 is also a substrate adaptor for WNK1 [29], and PHAII mutations in the acidic motif of WNK4 affect the regulation of WNK4 by KLHL3 [29;30].

Highlights.

• ➢KLHL3 decreases protein abundance of WNK4 but not that of NCC.
• ➢KLHL3 reduces the level of phosphorylated OSR1 but not that of total OSR1 protein.
• ➢Both cullin 3 and KLHL3 are involved the degradation of WNK4.
• ➢KLHL3 increases the ubiquitination of WNK4 protein.
• ➢Mutations in KLHL3 reduce the inhibitory effect of KLHL3 on WNK4 protein abundance.

Abbreviations

CUL3

cullin 3

KLHL3

kelch-like 3

NCC

Na⁺-Cl⁻ cotransporter

OSR1

oxidative stress-responsive 1

PHAII

pseudohypoaldosteronism type II

WNK4

with-no-lysine (K) kinase 4