Pflügers Archiv - European Journal of Physiology

, Volume 465, Issue 2, pp 233–245

Downregulation of the renal outer medullary K+ channel ROMK by the AMP-activated protein kinase

Authors

  • Balasaheb Siraskar
    • Department of PhysiologyUniversity of Tübingen
  • Dan Yang Huang
    • Department of Pharmacology and Experimental TherapyInstitute of Pharmacology and Toxicology, University of Tübingen
  • Tatsiana Pakladok
    • Department of PhysiologyUniversity of Tübingen
  • Gulab Siraskar
    • Department of PhysiologyUniversity of Tübingen
  • Mentor Sopjani
    • Department of PhysiologyUniversity of Tübingen
    • Faculty of MedicineUniversity of Prishtina
  • Ioana Alesutan
    • Department of PhysiologyUniversity of Tübingen
  • Yulia Kucherenko
    • Department of PhysiologyUniversity of Tübingen
  • Ahmad Almilaji
    • Department of PhysiologyUniversity of Tübingen
  • Vasudharani Devanathan
    • Department of Pharmacology and Experimental TherapyInstitute of Pharmacology and Toxicology, University of Tübingen
  • Ekaterina Shumilina
    • Department of PhysiologyUniversity of Tübingen
  • Michael Föller
    • Department of PhysiologyUniversity of Tübingen
    • Campbell Family Institute for Breast Cancer Research, Ontario Cancer InstituteUniversity Health Network (UHN)
  • Carlos Munoz
    • Department of PhysiologyUniversity of Tübingen
    • Department of PhysiologyUniversity of Tübingen
Ion Channels, Receptors and Transporters

DOI: 10.1007/s00424-012-1180-1

Cite this article as:
Siraskar, B., Huang, D.Y., Pakladok, T. et al. Pflugers Arch - Eur J Physiol (2013) 465: 233. doi:10.1007/s00424-012-1180-1

Abstract

The 5′-adenosine monophosphate-activated serine/threonine protein kinase (AMPK) is stimulated by energy depletion, increase in cytosolic Ca2+ activity, oxidative stress, and nitric oxide. AMPK participates in the regulation of the epithelial Na+ channel ENaC and the voltage-gated K+ channel KCNE1/KCNQ1. It is partially effective by decreasing PIP2 formation through the PI3K pathway. The present study explored whether AMPK regulates the renal outer medullary K+ channel ROMK. To this end, cRNA encoding ROMK was injected into Xenopus oocytes with and without additional injection of constitutively active AMPKγR70Q (AMPKα1-HA+AMPKβ1-Flag+AMPKγ1R70Q), or of inactive AMPKαK45R (AMPKα1K45R+AMPKβ1-Flag+AMPKγ1-HA), and the current determined utilizing two-electrode voltage-clamp and single channel patch clamp. ROMK protein abundance was measured utilizing chemiluminescence in Xenopus oocytes and western blot in whole kidney tissue. Moreover, renal Na+ and K+ excretion were determined in AMPKα1-deficient mice (ampk−/−) and wild-type mice (ampk+/+) prior to and following an acute K+ load (111 mM KCl, 30 mM NaHCO3, 4.7 mM NaCl, and 2.25 g/dl BSA) at a rate of 500 μl/h. As a result, coexpression of AMPKγR70Q but not of AMPKαK45R significantly decreased the current in ROMK1-expressing Xenopus oocytes. Injection of phosphatidylinositol PI(4,5)P2 significantly increased the current in ROMK1-expressing Xenopus oocytes, an effect reversed in the presence of AMPKγR70Q. Under control conditions, no significant differences between ampk−/− and ampk+/+ mice were observed in glomerular filtration rate (GFR), urinary flow rate, serum aldosterone, plasma Na+, and K+ concentrations as well as absolute and fractional Na+ and K+ excretion. Following an acute K+ load, GFR, urinary flow rate, serum aldosterone, plasma Na+, and K+ concentration were again similar in both genotypes, but renal absolute and fractional Na+ and K+ excretion were higher in ampk−/− than in ampk+/+ mice. According to micropuncture following a K+ load, delivery of Na+ to the early distal tubule but not delivery of K+ to late proximal and early distal tubules was increased in ampk−/− mice. The upregulation of renal ROMK1 protein expression by acute K+ load was more pronounced in ampk−/− than in ampk+/+ mice. In conclusion, AMPK downregulates ROMK, an effect compromising the ability of the kidney to excrete K+ following an acute K+ load.

Keywords

Energy depletionK+ channelsROMKKaliuresisNatriuresis

Introduction

The 5′-adenosine monophosphate (AMP)-activated protein kinase (AMPK) is stimulated upon cellular energy depletion [24, 47, 49] and triggers a variety of cellular functions replenishing cellular ATP levels [37], such as stimulation of glucose uptake, glycolysis, fatty acid oxidation, protein degradation, and the activity of enzymes required for ATP production [2, 24, 29, 38, 44, 49]. Moreover, AMPK decreases energy consumption by modifying protein synthesis, gluconeogenesis, and lipogenesis [37, 49]. AMPK [20, 28] and the related salt-inducible kinase [46] decrease Na+/K+ ATPase activity. Most AMPK-regulated functions protect cells during energy depletion [18, 37]. AMPK is, however, not only activated by energy depletion but in addition by hypoxia [15], nitric oxide [32], and increase in cytosolic Ca2+ activity [47]. In the kidney, AMPK participates in ion transport regulation, renal hypertrophy, ischemic injury, and diabetic nephropathy [22].

AMPK has previously been shown to regulate the epithelial Na+ channel ENaC [1, 5, 8, 13, 21, 22, 35, 40, 50]. ENaC has been shown to be inhibited by AMPK in both Xenopus oocytes and mouse collecting duct cell lines [13, 20, 40]. The inhibitory effect of AMPK on ENaC is presumably due to direct phosphorylation of the ubiquitin protein ligase Nedd4-2, thus leading to enhanced Nedd4-2-dependent retrieval of ENaC from the plasma membrane [20, 40]. Moreover, AMPK-dependent regulation of ENaC may involve phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2) [35]. The in vivo relevance of AMPK-dependent ENaC regulation was documented by increased ENaC expression in AMPKα1 knockout mice [20]. However, no significant decrease of urinary sodium excretion was found in AMPKα1 knockout animals. We have demonstrated before that a moderate high NaCl diet (4 % NaCl) reduced renal expression of activated AMPKα in rats by about three times compared with normal NaCl diet, suggesting that AMPKα subunit expression and activity are regulated by dietary sodium intake [27]. Moreover, treatment with the AMPK activator AICAR significantly increased fluid and sodium load at the end of the proximal tubule in animals on high-salt diet implying that proximal tubular sodium and fluid transport are regulated by AMPK activation.

In addition to regulation of ENaC, AMPK has also been proposed to regulate several potassium channels in myocytes and different epithelia [4, 6, 45]. In the heart, the ATP-sensitive K+ (KATP) channels were observed to be regulated by the AMPKα2 subunit in a model of cardiac ischemic preconditioning [45]. In lung epithelial cells, the calcium-activated K+ channel KCa3.1, which is expressed at the basolateral membrane in a variety of epithelia, interacts with the AMPKγ1 subunit and is inhibited by AMPK [30]. Finally, KCNQ1 potassium channels expressed in collecting duct principal cells are inhibited by AMPK [6, 45]. AMPK regulates those channels by Nedd4-2-dependent mechanisms [6]. Presumably due to altered K+ channel activity, urinary potassium excretion was significantly enhanced in spontaneously collected urine specimens from AMPKα1 knockout mice [5].

The present study explored whether AMPK contributes to the regulation of the renal outer medullary K+ channel ROMK. To determine AMPK sensitivity of ROMK, the K+ current was determined in Xenopus oocytes expressing ROMK with or without additional expression of constitutively active AMPKγR70Q [23] and of inactive AMPKαK45R. Moreover, the impact of AMPK on renal K+ output was tested by means of clearance and micropuncture studies in AMPKα1-deficient (ampk−/−) [48] and wild-type mice (ampk+/+).

Methods

Constructs

For generation of cRNA [25], constructs were used encoding wild-type ROMK1 [51], constitutively active AMPKγR70Q-HA (AMPKα1-HA+AMPKβ1-Flag+AMPKγ1R70Q) [23], kinase-dead mutant AMPKαK45R-HA (AMPKα1K45R+AMPKβ1-Flag+AMPKγ1-HA) [21], and Nedd4-2 [3]. For surface expression of ROMK by chemiluminescence, a tagged ROMK1-HA was included. All constructs used for generation of cRNA were described previously [39].

Voltage clamping in Xenopus oocytes

Xenopus oocytes were prepared as previously described [9]. cRNA encoding ROMK1 (5 ng) was injected with or without 4.6 ng of cRNA encoding either AMPKα1-HA+AMPKβ1-Flag+AMPKγR70Q (constitutively active AMPKγR70Q) or AMPKα1K45R-HA+AMPKβ1-Flag+AMPKγ1-HA (kinase-dead mutant AMPKαK45R) on the day of preparation of the Xenopus laevis oocytes. All experiments were performed at room temperature 3–4 days after injection. In one set of experiment, oocytes were incubated with the AMPK stimulator AICAR (1 mM; Tocris Bioscience) for 24 h or treated with the AMPK stimulator phenformin (1 mM; Sigma Aldrich) for 2 h, with or without AMPK inhibitor compound C (20 μM; Calbiochem) [36]. In another set of experiments, cRNA encoding Nedd4-2 (5 ng) were injected into the oocytes on the day of preparation. Phosphatidylinositol analogues, namely PI(4,5)P2-diC8, PI(3,5)P2-diC8, and PI(3,4)P2-diC8 as derivatives of PI(4,5)P2, PI(3,5)P2, and PI(3,4)P2 (Echelon Biosciences inc., Salt Lake City, UT) were injected (4.6 nl) at a concentration of 1 mg/ml into the oocytes, respectively. The ROMK1 current recordings started 45 min after injection. In order to confirm specificity of the currents, the ROMK1 inhibitor tertiapin LQ (100nM; Tocris Bioscience) was applied under the same condition. The potassium currents were measured 2 h after adding Tertiapin LQ with ND96-A into the bath and measured again 2 h after washing out Tertiapin LQ.

Oocytes were maintained at 17 °C in ND96-A solution containing 88.5 mM NaCl, 2 mM KCl, 1.8 mM CaC12, 1 mM MgC12, 5 mM HEPES, 0.11 mM tretracycline, 4 μM ciprofloxacin, 0.22 mM gentamycin (Refobacin ©), 0.5 mM theophylline (Euphylong ©), as well as 5 mM sodium pyruvate. The pH was adjusted to 7.4 by adding NaOH. Oocytes were superfused continuously with ND-96 buffer containing the following: 96 mM NaCl, 2 mM KCl, 1.8 mM CaCl2, 1 mM MgCl2, and 5 mM HEPES (pH = 7.4). Pipettes were filled with 3 M KCl and had resistances of 0.5–1.0 MΩ. In two-electrode voltage-clamp experiments, ROMK channel currents were elicited every 20 s with 3-s pulses from −160 to +60 mV applied from a holding potential of −80 mV. Pulses were applied in 20 mV increments. The data were filtered at 1 kHz and recorded with a Digidata 1322A A/D-D/A converter and Clampex V.4.2 software for data acquisition and analysis (Axon Instruments) [14]. The data were analyzed with Clampfit 9.2 (Axon Instruments) software [25].

Single-channel patch-clamp recordings

Xenopus oocytes expressing ROMK with or without AMPK as indicated were used within 3–4 days after RNA injection. Before patch clamping, the oocyte vitelline membrane was removed enzymatically by incubating the cells with Pronase E (100 μg/ml, Sigma) dissolved in oocyte Ringer bath solution that contained the following: 105 mM NaCl, 5 mM KCl, 2 mM CaCl2, and 5 mM HEPES/NaOH (pH = 7.5).

Single-channel patch-clamp recordings were performed in inside–out patches at room temperature. The patch pipettes were made of borosilicate glass capillaries (150 TF-10, Clark Medical Instruments) and had a resistance of 6–9 MΩ. Pipettes were connected via an Ag–AgCl wire to the headstage of an EPC 9 patch-clamp amplifier (HEKA). The offset potentials between both electrodes were zeroed before measuring. Data acquisition and data analysis were controlled by a computer equipped with an ITC 16 interface (Instrutech) and by using Pulse software (HEKA). For current measurements, Xenopus oocytes were held at −70 mV holding potential (Vh). The pulses of 200 ms duration from −100 to 0 mV were applied with +20 mV increments. The original current traces are depicted without filtering (acquisition frequency of 3 kHz). Open channel probability (Po) was calculated using the following equation: Po = I/iN, where I is the mean current, i is the unitary current amplitude, and N is the number of functional channels in the patch. Single channel activity was measured in inside-out patches from Xenopus oocytes continuously perfused with fluoride vanadate pyrophosphate (FVPP phosphatase inhibitors) bath solution that prevents spontaneous ROMK channel rundown upon excision [16]. The FVPP bath solution contained the following: 100 mM KCl, 5 mM HEPES, 5 mM EDTA, 3 mM Na3VO4, 4 mM NaF, and 10 mM Na4P2O7 (pH 7.4). The pipette solution contained the following: 100 mM KCl, 2 mM CaCl2, and 5 mM HEPES (pH 7.4).

Detection of ROMK-HA surface expression by chemiluminescence

To determine ROMK-HA cell surface expression by chemiluminescence, defolliculated oocytes were first incubated with primary monoclonal anti-hemaglutinin-horseradish peroxidase antibody (1:1000, Miltenyi Biotec GmbH, Germany). Individual oocytes were then placed in 96-well plates with 20 μl of SuperSignal ELISA Femto Maximum Sensitivity Substrate (Pierce, Rockford, IL) and chemiluminescence of single oocytes was quantified in a luminometer (Walter Wallac 2 Plate Reader, Perkin Elmer, Juegesheim, Germany) by integrating the signal over a period of 1 s. Results display normalized relative light units. Integrity of the measured oocytes was assessed by visual control after the measurement to avoid unspecific light signals from the cytosol.

Animal experiments

Animal experiments were performed in adult AMPKα1-deficient (ampk−/−) and wild-type mice (ampk+/+). The ampk−/− mice have been described previously [48]. All animal experiments were conducted according to the German law for the welfare of animals and were approved by local authorities. The mice were fed a control diet (Sniff, Soest, Germany) and had free access to tap drinking water. Briefly, mice were anesthetized with pentobarbital sodium (80 mg/kg, i.p.). Body temperature was maintained at 37.5 °C by placing the animals on an operating table with a servo-controlled heating plate. The trachea was cannulated for free air breathing throughout the experiment. The femoral artery was cannulated for blood pressure measurements and blood sample withdrawal. The jugular vein was cannulated for continuous infusion (111 mM NaCl, 30 mM NaHCO3, 4.7 mM KCl, and 2.25 g/dl BSA) at a rate of 500 μl/h. For assessment of glomerular filtration rate (GFR), [3H]-inulin was added to deliver 20 μCi/h. Urine sample was collected via a bladder catheter. After surgery, mice were allowed to stabilize for 60 min. Subsequently, to determine basal kidney function, urine was collected during the first 30 min. Blood samples (20 μl) were collected immediately before and after urine collections. After completion of the first collection period, K+ loading was initiated by replacing 250 μl/h of the maintenance infusion with a high K+ solution (111 mM KCl, 30 mM NaHCO3, 4.7 mM NaCl, and 2.25 g/dl BSA). Thirty minutes after starting K+ loading, kidney function was reassessed by performing another 30-min clearance. Additional blood was withdrawn after this final period to measure the serum aldosterone concentration. The plasma and urine concentrations of Na+ and K+ were determined with a flame-emission photometer (ELEX 6361; Eppendorf AG, Hamburg, Germany). The concentrations of [3H]-inulin in plasma and urine were measured by liquid-phase scintillation counting. GFR was calculated according to standard formulas. Serum aldosterone concentration was measured by ELISA (Alpha Diagnostic Intl. Inc. USA).

Micropuncture of renal tubules

For micropuncture experiments, the left kidney was exposed by flank incision, carefully freed of perirenal fat and immobilized in a lucite cup. The kidney was covered with prewarmed paraffin oil. The distal or proximal tubular configuration of nephrons was identified by a small volume of dye injected into a random proximal tubule on the kidney surface using a fine micropipette (1–3 μm tip) filled with colored artificial tubular fluid (ATF; 113 mM NaCl, 25 mM NaHCO3, 4 mM KCl, 1 mM MgSO4, 2 mM CaCl2, 1 mM Na2HPO4, 5 mM glucose, 5 mM urea, and 0.075 % FD&C green (pH 7.4)). Thus, the last surface loop of a proximal tubule and the first surface loop of a distal tubule of a given nephron were recognized. Fluid collections were made with glass capillaries (5–10 μm tip) for 3 to 5 min under free flow. The tubular fluid volumes were determined from the column length in a constant-bore capillary. Concentrations of [3H]-inulin were measured by liquid-phase scintillation counting. The concentrations of Na+ and K+ in the tubular fluid were determined using a microflame-emission photometer [27].

Western blot analysis

For western blot analysis, kidneys from AMPKα1-deficient (ampk−/−) mice and wild-type mice (ampk+/+) before and after potassium loading, were homogenized and the proteins were separated using 12 % SDS-PAGE. The proteins were then transferred to a nitrocellulose membrane. Transferred proteins on the blot were incubated with rabbit polyclonal anti-potassium channel ROMK1 antibody (1:400, Millipore, Schwalbach, Germany). After washing (TBST), the blots were incubated with anti-rabbit HRP-conjugated IgG antibody (diluted 1:1000, Cell Signaling, Danvers, USA) for 1 h at room temperature. For loading control, the blot was stripped in stripping buffer (Thermo Fisher Scientific, Schwerte, Germany) at room temperature for 20 min. After washing with TBST, the blot was blocked with 5 % nonfat milk in TBST for 1 h at room temperature. The blot was then incubated with a rabbit anti-GAPDH antibody or rabbit anti-actin antibody (diluted 1:1000, Cell Signaling Technology, Danvers, MA) at 4 °C overnight. After washing with TBST, the blot was incubated with anti-rabbit HRP antibody (diluted 1:1000, Cell Signaling Technology) for 1 h at room temperature. Antibody binding was detected with the Pierce ECL Plus Western Blotting Substrate (Thermo Fisher Scientific). Bands were quantified with Quantity One Software (Bio-Rad, Munich, Germany).

Statistical analysis

Data are provided as means ± SEM (n represents the number of experiments). All oocyte experiments were repeated with at least three batches of oocytes; in all repetitions, qualitatively similar data were obtained. Data were tested for significance using ANOVA or t test, and results with p < 0.05 were considered statistically significant.

Results

In order to test, whether AMP-activated kinase AMPK regulates the renal outer medullary K+ channel ROMK1, the K+ current (IK) was determined utilizing dual electrode voltage clamp in Xenopus oocytes expressing ROMK without or with additional expression of AMPK. As illustrated in Fig. 1, IK was low in water-injected oocytes but markedly increased by expression of ROMK. The additional coexpression of constitutively active AMP-activated protein kinase AMPKγR70Q resulted in a significant decrease of IK. Additional experiments explored whether the effect of AMPK on ROMK was mimicked by the catalytically inactive mutant AMPKαK45R (AMPKα1K45R+AMPKβ1+AMPKγ1). As illustrated in Fig. 2, AMPKαK45R did not significantly modify IK in ROMK-expressing oocytes. Thus, kinase activity was required for the inhibitory effect of AMPK on IK.
https://static-content.springer.com/image/art%3A10.1007%2Fs00424-012-1180-1/MediaObjects/424_2012_1180_Fig1_HTML.gif
Fig. 1

Coexpression of constitutively active AMPKγR70Q decreased the K+ current in ROMK-expressing Xenopus oocytes. a Representative original tracings showing currents in Xenopus oocytes injected with water (a), expressing ROMK alone (b), or expressing ROMK with additional coexpression of constitutively-active AMPKγR70Q (c). b Arithmetic means ± SEM (n = 21–58) of the K+ current at −20 mV following depolarization from −80 mV holding potential in Xenopus oocytes injected with water (1st bar) or expressing ROMK without (2nd bar) or with (3rd bar) additional coexpression of constitutively active AMPKγR70Q. ***p < 0.001—statistically significant difference from the absence of AMPKγR70Q. c Arithmetic means ± SEM (n = 21–58) of the current (I) as a function of the potential difference across the cell membrane (V) in Xenopus oocytes injected with water or expressing ROMK without or with additional coexpression of constitutively active AMPKγR70Q

https://static-content.springer.com/image/art%3A10.1007%2Fs00424-012-1180-1/MediaObjects/424_2012_1180_Fig2_HTML.gif
Fig. 2

Constitutively active AMPKγR70Q but not inactive AMPKαK45R decreased current in ROMK-expressing Xenopus oocytes. a Arithmetic means ± SEM (n = 27-55) of the K+ current at −20 mV following depolarization from −80 mV holding potential in Xenopus oocytes injected with water (1st bar) or expressing ROMK without (2nd bar) or with (3rd bar) additional coexpression of constitutively active AMPKγR70Q or of catalytically inactive AMPKαK45R (4th bar). ***p < 0.001—statistically significant difference from the absence of AMPKγR70Q. b Arithmetic means ± SEM (n = 27–55) of the current (I) as a function of the potential difference across the cell membrane (V) in Xenopus oocytes injected with water or expressing ROMK without or with additional coexpression of constitutively active AMPKγR70Q or of catalytically inactive AMPKαK45R

Further experiments addressed the question whether pharmacological modification of AMPK activity similarly impacts on K+ current in Xenopus oocytes. As illustrated in Fig. 3a, b, a 24 h- treatment of oocytes expressing ROMK with AMPK activator AICAR or after a 2 h- treatment with phenformin led to a significant decrease of the outward current, an effect, which was reversed in both cases by additional application of AMPK inhibitor compound C.
https://static-content.springer.com/image/art%3A10.1007%2Fs00424-012-1180-1/MediaObjects/424_2012_1180_Fig3_HTML.gif
Fig. 3

Sensitivity of ROMK1 to AMPK activators AICAR and phenformin, and to inhibitor compound C. a Arithmetic means ± SEM (n = 10–13) of the normalized K+ current at −20 mV following depolarization from −80 mV holding potential in Xenopus oocytes injected with water (left bars) or expressing ROMK1 and incubated without (3rd bar) or with (4th bar) AMPK activator AICAR alone (1 mM) or with (5th bar) the AMPK inhibitor compound C (20 μM). ***p < 0.001—statistically significant difference from the absence of AICAR. b Arithmetic means ± SEM (n = 8–17) of the normalized K+ current at −20 mV following depolarization from −80 mV holding potential in Xenopus oocytes injected with water (1st bar) or expressing ROMK1 and incubated without (2nd bar) or with (3rd bar) AMPK activator phenformin alone (1 mM) or with (4th bar) the AMPK inhibitor compound C (20 μM). *p < 0.05—statistically significant difference from the absence of phenformin

As AMPK is known to phosphorylate and thus to activate the ubiquitin ligase Nedd4-2, additional experiments were performed to test whether Nedd4-2 is involved in the AMPK-dependent regulation of ROMK. As a result the current was 1,867 ± 27 nA (n = 26) in oocytes expressing ROMK alone and 1,707 ± 13 nA (n = 25) in oocytes expressing both, ROMK and Nedd4-2. Thus, as shown in Fig. 4a, ROMK is apparently not a target of Nedd4-2 and the ubiquitin ligase does not mediate the downregulation of ROMK by AMPK. As AMPK is effective in part through the PIP2 pathway in connection with PI3K, we further analyzed the effect of three PIP2 analogues on the ROMK-dependent K+ current in the presence and absence of AMPK. As a result, the ROMK-dependent K+ current was significantly increased by PI(4,5)P2 but not by PIP(3,4)P2 or PIP(3,5)P2 (Fig. 4b). To confirm specificity of the ROMK currents, the specific ROMK1 inhibitor Tertiapin LQ (100 nM) was applied. As shown in Fig. 5, Tertiapin LQ abolished the difference of the K+ current in oocytes expressing ROMK alone and coexpressing the constitutively active AMPKγR70Q, an effect that could be reversed by washing out of tertiapin LQ.
https://static-content.springer.com/image/art%3A10.1007%2Fs00424-012-1180-1/MediaObjects/424_2012_1180_Fig4_HTML.gif
Fig. 4

Effects of Nedd4-2 and phosphatidylinositol analogues on AMPK-induced ROMK1 currents. a Arithmetic means ± SEM (n = 14–29) of the K+ current at −20 mV following depolarization from −80 mV holding potential in Xenopus oocytes expressing water (1st bar) or ROMK1 without (2nd and 3rd bars) or with (4th and 5th bars) coexpresion of Nedd4-2 without (white bars) or with (black bars) coexpresion of the constitutively active AMPKγR70Q. **p < 0.01; ***p < 0.001—statistically significant difference from absence of AMPK. b Arithmetic means ± SEM (n = 12–18) of the K+ current at −20 mV following depolarization from −80 mV holding potential in Xenopus oocytes expressing water (1st bar) or ROMK1 without (2nd and 3rd bars) and with additional injection of PI(3.4)P2 (4th and 5th bars), PI(4.5)P2 (6th and 7th bars), or PI(3.5)P2 (8th and 9th bars) with (black bars) and without (white bars) coexpresion of the constitutively active AMPKγR70Q. **p < 0.01; ***p < 0.001—statistically significant difference from absence of AMPK; ###p < 0.001—statistically significant difference from absence of the respective PIP2

https://static-content.springer.com/image/art%3A10.1007%2Fs00424-012-1180-1/MediaObjects/424_2012_1180_Fig5_HTML.gif
Fig. 5

Effects of ROMK inhibitor Tertiapin LQ on AMPK-induced ROMK1 currents. Arithmetic means ± SEM (n = 12–15) of the normalized K+ current at −20 mV following depolarization from −80 mV holding potential in Xenopus oocytes injected with water (1st bar) or expressing ROMK1 without (white bars) or with coexpression of the constitutive active AMPKR70Q (black bars) incubated for 2 h with the ROMK1 channel inhibitor tertiapin LQ (4th and 5th bars) and washed out for 2 h (6th and 7th bars). ***p < 0.001—indicates statistically significant difference from the absence of the constitutively active AMPKγR70Q. ###p < 0.001 indicates statistically significant difference from the presence of Tertiapin LQ

The effect of AMPK on ROMK was further characterized by single channel patch-clamping. Figure 6a shows typical K+ channel current traces from excised, inside–out patches of Xenopus oocytes expressing ROMK. The injection of cRNA encoding AMPKγR70Q into ROMK-expressing oocytes decreased single-channel conductance (middle). Open-channel probability measured at −100 mV was 0.86 ± 0.04 (n = 4) in the ROMK-expressing cells and not significantly modified by AMPKγR70Q (0.88 ± 0.03, n = 4). Water-injected Xenopus oocytes showed no detectable ROMK channel activity (Fig. 6b). As shown in Fig. 6b, the associated IV curves display slight inward rectification typical of ROMK channels [33].
https://static-content.springer.com/image/art%3A10.1007%2Fs00424-012-1180-1/MediaObjects/424_2012_1180_Fig6_HTML.gif
Fig. 6

Constitutively active AMPKγR70Q but not inactive AMPKαK45R reduces ROMK channel activity in inside–out patches from Xenopus oocytes. a Representative single-channel tracings, recorded with K+-containing (100 mM KCl) pipette solution, in inside–out patches from Xenopus oocytes injected with water or with cRNA encoding ROMK1 or with cRNA encoding ROMK1 and AMPKγR70Q in FVPP bath solution. b Arithmetic means (±SEM) of the current as a function of voltage (I/V curve) in inside–out patches from Xenopus oocytes injected with water (closed diamonds, n = 3), with cRNA encoding ROMK without (closed triangles, n = 7) or with additional cRNA encoding AMPKγR70Q (open squares, n = 8), or of AMPKαK45R (open triangles, n = 6) in FVPP bath solution. c Arithmetic means (±SEM) of single-channel conductance (as calculated for the currents at −100/−60 mV by linear regression) in inside–out patches from Xenopus oocytes expressing ROMK without (white bar, n = 7) or with coexpression of AMPKγR70Q (KA, black bar, n = 8) or of AMPKαK45R (KD, striped bar, n = 6) in FVPP bath solution. *p ≤ 0.05—significant difference from ROMK1 alone

The inhibitory effect of the AMPK on ROMK could result from decreased carrier protein abundance in the plasma membrane. To explore this possibility, chemiluminescence experiments were performed. As illustrated in Fig. 7, the coexpression of AMPKγR70Q resulted in a significant decrease in the ROMK-HA protein abundance within the oocyte membrane.
https://static-content.springer.com/image/art%3A10.1007%2Fs00424-012-1180-1/MediaObjects/424_2012_1180_Fig7_HTML.gif
Fig. 7

ROMK surface expression detected by chemiluminescence in oocytes expressing constitutively active AMPKγR70Q. Arithmetic means ± SEM (n = 69–79) of the normalized chemiluminescence signal of HA-tagged ROMK in Xenopus oocytes injected with water (grey bar) or expressing ROMK without (white bar) or with (black bar) additional coexpression of constitutively active AMPK (AMPKγR70Q). ***p < 0.001—statistically significant difference from the absence of AMPKγR70Q coexpresion

In order to determine whether the AMPK-dependent regulation of ROMK is relevant for the renal expression and function of ROMK, clearance experiments were performed in AMPKα1-deficient mice (ampk−/−) and wild-type mice (ampk+/+). In a first step, we explored whether AMPK-dependent regulation of ROMK impacts on the ability of the kidneys from ampk−/− mice to eliminate K+ following an acute K+ load. Prior to an acute K+ load, no differences were observed between ampk−/− and ampk+/+ mice in GFR (8.95 ± 1.54 vs. 7.60 ± 1.36 μl min−1 g−1 BW), urinary flow rate (62.10 ± 15.68 vs. 45.52 ± 10.34 nl min−1 g−1 BW), plasma Na+ (Fig. 8) or K+ (Fig. 9) concentration, urinary Na+ (Fig. 8) or K+ (Fig. 9) excretion, and fractional urinary Na+ (Fig. 8) or K+ (Fig. 9) excretion or serum aldosterone concentration (Fig. 10). Basal mean arterial blood pressure tended to be lower in ampk−/− mice than in ampk+/+ mice (Table 1). Following an acute K+ load (111 mM KCl, 30 mM NaHCO3, 4.7 mM NaCl, and 2.25 g/dl BSA), the plasma Na+ (Fig. 8) and K+ (Fig. 9) concentration was again not significantly different between the genotypes. No significant differences were observed between ampk−/− and ampk+/+ in GFR (5.8 ± 0.7 vs. 6.8 ± 1.1 μl min−1 g−1 BW) and urinary flow rate (110 ± 38 vs. 68 ± 14 nl min−1 g−1 BW) following a K+ load. The reduction in GFR and the concomitant rise in urinary flow rate in response to K+ loading occurred to the same extent in both groups of mice. The serum aldosterone concentration tended to increase slightly in both genotypes after the acute K+ load, an effect, however, not reaching statistical significance (Fig. 10). The urinary excretion and fractional urinary excretion of both, Na+ (Fig. 8) and K+ (Fig. 9) were significantly higher in ampk−/− mice than in ampk+/+ mice following the K+ load. Accordingly, the elimination of K+ following an acute K+ load was more pronounced in ampk−/− mice than in ampk+/+ mice.
https://static-content.springer.com/image/art%3A10.1007%2Fs00424-012-1180-1/MediaObjects/424_2012_1180_Fig8_HTML.gif
Fig. 8

Plasma Na+ concentration, urinary Na+ excretion and fractional urinary Na+ excretion in AMPKα1-deficient and wild-type mice. Arithmetic means ± SEM (n = 8) of the plasma Na+ concentration (left), urinary Na+ excretion (middle), and fractional urinary Na+ excretion (right) in AMPKα1-deficient mice (AMPKα1−/−, black bars) and wild-type mice (AMPKα1+/+, white bars) prior to (left bars, basal) and following (right bars, +K) an acute K+ load (111 mM KCl, 30 mM NaHCO3, 4.7 mM NaCl, and 2.25 g/dl BSA) at a rate 500 μl/h. *p < 0.05—difference from wild-type animals

https://static-content.springer.com/image/art%3A10.1007%2Fs00424-012-1180-1/MediaObjects/424_2012_1180_Fig9_HTML.gif
Fig. 9

Plasma K+ concentration, urinary K+ excretion, and fractional urinary K+ excretion in AMPKα1-deficient and wild-type mice. Arithmetic means ± SEM (n = 8) of the plasma K+ concentration (left), urinary K+ excretion (middle), and fractional urinary K+ excretion (right) in AMPKα1-deficient mice (AMPKα1−/−, black bars) and wild-type mice (AMPKα1+/+, white bars) prior to (left bars, basal) and following (right bars, +K) an acute K+ load (111 mM KCl, 30 mM NaHCO3, 4.7 mM NaCl, and 2.25 g/dl BSA) at a rate 500 μl/h. *p < 0.05—difference from wild-type animals; #p < 0.05—significant difference from respective basal value prior to K+ load

https://static-content.springer.com/image/art%3A10.1007%2Fs00424-012-1180-1/MediaObjects/424_2012_1180_Fig10_HTML.gif
Fig. 10

Serum aldosterone prior to and following an acute K+ load in AMPKα1-deficient and wild-type mice. Arithmetic means ± SEM (n = 6) of the serum aldosterone concentration in AMPKα1-deficient mice (AMPK−/−, black bars) and wild-type mice (AMPK+/+, white bars) prior to (basal) and following an acute K+ load (+K+)

Table 1

Micropuncture parameters prior to (basal) and following an acute K+ load

 

AMPK+/+

AMPK−/−

Basal

+K+

Basal

+K+

Body weight (g)

27.0 ± 1.0

24.8 ± 8.8

Mean arterial blood pressure (mmHg)

106 ± 5

101 ± 8

92 ± 5

87 ± 8

SNGFR up to the LP (nl/min)

36.4 ± 2.0

37.0 ± 1.6

38.7 ± 1.4

36.7 ± 1.8

SNGFR up to the ED (nl/min)

35.3 ± 1.2

35.0 ± 1.1

36.6 ± 1.3

35.9 ± 1.0

Absolute Na+ delivery up to the LP (μmol/min)

3.658 ± 166

3.558 ± 198

3.372 ± 207

3.727 ± 160

Fractional Na+ delivery up to the LP (%)

68 ± 3

72 ± 5

67 ± 3

75 ± 5

Absolute Na+ delivery up to the ED (μmol/min)

147 ± 16

138 ± 15

147 ± 17

218 ± 18*, **

Fractional Na+ delivery up to ED (%)

4 ± 1

4 ± 1

9 ± 5

10 ± 5

Absolute K+ delivery up to the LP (μmol/min)

101 ± 5

98 ± 7

94 ± 6

104 ± 5

Fractional K+ delivery up to the LP (%)

57 ± 3

60 ± 5

57 ± 3

65 ± 5

Absolute K+ delivery up to the ED (μmol/min)

13 ± 1

11 ± 1

12 ± 2

12 ± 1

Fractional K+ delivery up to ED (%)

7 ± 1

6 ± 1

7 ± 1

7 ± 1

Arithmetic means ± SEM (n = 17–19 nephrons each group)

SNGFR single nephron filtration rate, LP late proximal tubule, ED early distal tubule

*p < 0.05 vs. basal; **p < 0.05 vs. AMPK+/+

To gain insight into nephron segments involved in alteration of K+ transport in ampk−/− mice, we performed micropuncture experiments before and following a K+ load. Single-nephron GFR up to the late proximal and early distal tubule was comparable in both genotypes prior and after acute K+ load (Table 1). However, the absolute delivery of Na+ to the early distal tubule was significantly higher in ampk−/− mice following the K+ load (Table 1) implying reduced tubular Na+ transport between the late proximal and early distal segments of ampk−/− mice. Both, absolute and fractional delivery of K+ up to the late proximal and early distal tubule were not significantly altered in both genotypes following a K+ load, indicating an effect further downstream of the early distal tubules, i.e., in the late distal tubules and collecting duct.

Western blotting did not reveal significant differences in renal ROMK1 protein expression between ampk−/− and ampk+/+ prior to an acute K+ load (Fig. 11). An acute K+ load, however, resulted in an increase in renal ROMK1 protein expression, which was twofold in ampk+/+ mice and fourfold in ampk−/− (Fig. 11). Thus, the effect of acute K+ load on renal ROMK1 protein expression was more pronounced in the absence of AMPK.
https://static-content.springer.com/image/art%3A10.1007%2Fs00424-012-1180-1/MediaObjects/424_2012_1180_Fig11_HTML.gif
Fig. 11

ROMK1 protein expression in kidneys prior to and following an acute K+ load in AMPKα1-deficient and wild-type mice. Representative blot (a) and arithmetic means ± SEM (n = 4) of densitometry (b) from kidneys of AMPKα1-deficient mice (AMPK−/−, black bars) and wild-type mice (AMPK+/+, white bars) prior to (basal) and following an acute K+ load (+K). *p < 0.05 vs. basal; #p < 0.05 vs. AMPK+/+

Discussion

The present observations disclose a novel function of the AMP-activated kinase AMPK in the kidney. The kinase decreases the activity of the outer medullary K+ channel ROMK in vitro and compromises the elimination of K+ following an acute K+ load in vivo. ROMK channel activity is not only inhibited by constitutively active AMPK but also by pharmacological activation of AMPK in oocytes.

As AMPK is known to phosphorylate and thus activate the ubiquitin ligase Nedd4-2 [1, 2, 4], AMPK could have been effective through Nedd4-2. However, according to the present observations, Nedd4-2 does not account for the downregulation of ROMK, as the channel is apparently not a target of Nedd4-2. It is noteworthy, that an interaction of CFTR, ROMK, and AMPK has been suggested to participate in the regulation of ROMK channel activity [12].

AMPK could be effective through phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2) [35]. Similar to what has been shown earlier [26, 52], ROMK channels are upregulated by PI(4,5)P2. ROMK current increased up to 65 % in the presence of a PI(4,5)P2 analogue. This observation is in agreement with the previous report that PI(4,5)P2 is the most important phosphoinositide for the stability and regulation of ROMK1 channels [26, 52]. Moreover, our observations support the notion that AMPK inhibits ROMK channel activity probably by reducing ROMK membrane abundance rather than by modifying channel open probability.

Under control K+ intake, the urinary elimination of K+ was not significantly different between AMPK-deficient mice (ampk−/−) and wild-type mice (ampk+/+). Thus, lack of AMPK does not lead to K+ wasting. AMPKα1 knockout mice have previously been shown to excrete significantly more K+ in urine [5], an observation possibly due to increased K+ intake of those animals. In the present study, urinary K+ excretion was only enhanced following a K+ load. Thus, either AMPK is not significantly activated under control K+ intake, or ROMK activity is limiting for K+ excretion only following K+ loading. Indeed, renal ROMK1 protein abundance became significantly different between ampk−/− and ampk+/+ only following K+ load but not prior to K+ load, indicating that loss function of AMPK promotes urinary K+ loss possibly by upregulation of ROMK during high K+ intake. We also observed in our previous study that only under high salt diet but not under normal diet, tubular absorption of Na+ and K+ could be altered by pharmacological activation of AMPK [27]. Therefore, AMPK may participate in the regulation of renal electrolyte transport only following a respective challenge.

To determine which nephron segments may account for AMPK-dependent K+ excretion during acute K+ load, we analyzed Na+ and K+ reabsorption up to the late proximal and early distal tubules. Both ampk+/+ and ampk−/− displayed comparable K+ delivery up to the late proximal and to the early distal tubules after K+ load, indicating an effect downstream of the early distal segments, i.e., late distal and/or collecting tubules, where ROMK1 and ENaC are mainly expressed. Indeed, our western blot data confirmed that acute K+ load upregulates ROMK1 expression in both ampk+/+ and ampk−/− mice, implying an important role of ROMK1 in K+ excretion following an acute K+ load. Interestingly, AMPK is also involved in Na+ transport between late proximal and early distal nephrons following an acute K+ load while AMPK regulates proximal Na+ absorption under high salt diet [27]. Thus, AMPK may play different roles in distinct nephron segments upon different metabolic challenges. To which extent KCl infusion influences AMPK activity and/or AMPK-sensitive urinary K+ excretion and Na+ reabsorption cannot be derived from the present observations. It is noteworthy, however, that AMPK inhibits the Na+/K+ ATPase [5] and ENaC [7, 8], effects which would both be expected to compromize renal tubular Na+ absorption and K+ secretion.

Since ROMK is the major K+ channel contributing to K+ secretion in renal distal nephrons especially under a high K diet, upregulation of apical ROMK K+ channels during AMPK depletion is expected to increase urinary K+ excretion. Indeed, our patch clamp data support the notion that activation of AMPK inhibits ROMK activity, while inhibition of AMPK increased ROMK activity. Furthermore, AMPK has been reported to regulate Na+/K+ ATPase at the basolateral cell membrane of other cell types [20]. Thus, the effect of AMPK on K+ secretion could be in part the result of altered Na+/K+ ATPase activity. Moreover, the kidney-specific Na+–K+–2Cl co-transporter (NKCC2) which mediates transport of Na+, K+, and Cl across the luminal membrane of the thick ascending limb of the loop of Henle, has been observed to be phosphorylated by AMPK [19]. Thus, loss of interaction between AMPK and NKCC2 might account for the reduced Na+ transport in the knockout mice during acute K+ load. Inhibition of K+ channels during energy depletion could further curtail cellular K+ loss and thus counteract suicidal cell death [7, 11, 17, 42, 43]. In HCO3-permeable cells, such as the proximal tubule, inhibition of K+ channels with subsequent depolarization decreases electrogenic HCO3 exit leading to cytosolic alkalinization, which decreases the requirement for H+ extrusion by the Na+/H+ exchanger [31]. Decreased Na+ entry via the Na+/H+ exchanger again decreases the requirement for energy-consuming Na+ extrusion by the Na+/K+ ATPase [31]. On the other hand, AMPK may stimulate the Na+/H+ exchanger in order to alkalinize the cell [41]. Cytosolic alkanization may counteract death of apoptotic cells [34]. Moreover, cytosolic alkalinization stimulates glycolysis and may thus facilitate utilization of glucose for energy generation [10]. AMPK enhances cellular glucose uptake through upregulation of the facilitative glucose carriers [53] and stimulates glucose degradation by fostering glycolysis [24]. AMPK supports the generation of ATP further by stimulation of fatty acid oxidation and expression of several enzymes required for ATP production [49]. AMPK-dependent ROMK regulation may not only be relevant for energy depletion but AMPK may be activated and thus downregulate ROMK similarly following an increase in cytosolic Ca2+ activity [47], during hypoxia [15], and following formation of nitric oxide [32].

In conclusion, the present observations reveal an inhibitory effect of the AMP-activated kinase AMPK on the renal outer medullary K+ channel ROMK. Renal AMPK activity thus participates in the regulation of renal tubular electrolyte transport during metabolic challenge.

Acknowledgments

The authors gratefully acknowledge the generosity of Benoit Viollet, Institut Cochin, Université Paris Descartes, Centre National de la Recherche Scientifique (UMR8104), Paris, France; INSERM, U1016, Paris, France, for providing the ampk−/− mice. The authors further acknowledge the technical assistance of E. Faber. The manuscript was meticulously prepared by S. Rübe. This study was supported by the Deutsche Forschungsgemeinschaft (DFG HU 1600/1-2) and the IZKF of the University of Tübingen (Nachwuchsgruppe to M.F.).

Copyright information

© Springer-Verlag Berlin Heidelberg 2012