PKC isoform-specific targeting of the phosphorylation site Serine¹⁹¹ in the GIRK4 subunit induces receptor-dependent modulation of GIRK channel activity

G Protein activated inwardly rectifying K⁺ (GIRK) channels are inhibited during stimulation of G_q Protein-coupled receptors (G_qPCRs) by depletion of phosphatidyl-4,5-bisphosphate (PIP₂) and/or channel phosphorylation by proteinkinase C (PKC). Receptor-specific effects of G_q signaling pathways on GIRK channel activity comprise the activation of different PKC isoforms that target distinct phosphorylation sites within the GIRK4 subunit. The serine residue S¹⁹¹ in the GIRK4 subunit is an important phosphorylation site for PKC which contributes to GIRK channel inhibition. Until now, the specific PKC isoform that phosphorylates this residue is unknown. Furthermore, the functional impact of S¹⁹¹ for G_q-receptor-specific GIRK channel modulation has not been investigated. To evaluate the contribution of this PKC phosphorylation site to receptor-specific GIRK channel modulation, we monitored the activity of phosphorylation-deficient GIRK4 (S191A) channel mutants during stimulation of adrenergic α_1B-receptors (α_1B-ARs) or Angiotensin AT₁-receptors. G_qPCR-dependent modulation of GIRK currents was analyzed in voltage-clamp experiments in rat atrial myocytes and in CHO cells expressing phosphorylation-deficient GIRK channels of various subunit compositions. As a novel finding, we demonstrate that the inhibition of phosphorylation-deficient GIRK4 (S191A) channels is impaired during stimulation of AT₁-Rs, indicating that this phosphorylation site is targeted at least by the Angiotensin II-activated PKCε isoform. In contrast, ablation of the S¹⁹¹ phosphorylation site does not attenuate phenylephrine-induced GIRK reduction, suggesting that S¹⁹¹ is not a primary target for the α_1B-AR activated PKCα. Our data indicate that the different G_qPCR-activated PKC isoforms target distinct phosphorylation sites within the GIRK4 subunit, resulting in receptor-specific regulation of the phosphorylation-deficient GIRK4 (S191A) channel mutant.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00424-026-03157-0.

Highlights

• Phosphorylation-deficient GIRK4 (S191A) channels display reduced PMA-mediated inhibition

• Mutation of S¹⁹¹ alleviates the inhibitory effect of Ang II on GIRK channel activity

• Residue S¹⁹¹ is targeted by the AT₁-R-activated PKCε

• GIRK4 (S191A) inhibition was not abrogated during stimulation of α1B-ARs

• α1B-AR-activated PKCα does not target S¹⁹¹

Supplementary Information

The online version contains supplementary material available at 10.1007/s00424-026-03157-0.

Introduction

Cardiac G protein-activated inwardly rectifying K⁺ (GIRK) channels are heteromeric ion channels composed of GIRK1 and GIRK4 subunits which are directly activated by interaction of their subunits with Gβγ subunits upon stimulation of G_i/o-coupled receptors [6, 14, 22]. The activation of supraventricular GIRK channels during stimulation of muscarinergic M₂-receptors by acetylcholine contributes to the parasympathetic control of heart frequency and cardiac excitability.

GIRK channel activity is inhibited by signaling pathways following activation of G_q Protein-coupled receptors [15, 21, 27, 31, 33]. Signaling pathways downstream of G_q activation result in depletion of phosphatidyl-4,5-bisphosphate (PIP₂) and subsequent protein kinase C (PKC) activation by diacylglycerol (DAG). The phospholipase C (PLC)-catalyzed depletion of PIP₂ induces the rapid component of GIRK channel inhibition [7, 8, 10] whereas the subsequent inhibition of GIRK currents is attributed to PKC-induced phosphorylation of GIRK channels [1, 25, 32].

Despite the fact that the general signaling pathway of activated Gα_q subunits comprises hydrolysis of PIP₂ and subsequent activation of PKC, it was demonstrated that endogenous G_qPCRs inhibit atrial GIRK channels with receptor-specific efficacies and kinetics [8]. These receptor-dependent effects of G_q activation on GIRK channel activity were assessed to distinct spatiotemporal dynamics of PIP₂-depletionand to the different interaction of PIP₂ with its molecular target [7, 8, 10]. Other studies demonstrated species-dependent or receptor-dependent differences in the activation of specific PKC isoforms during stimulation of G_qPCRs [3, 20, 24, 29, 30]. In stably GIRK1/2-transfected HEK293 cells, activation of muscarinic M₁ or M₃ receptors specifically activated the Ca²⁺-independent (novel) nPKC isoform PKCδ [3, 20] whereas α-adrenergic stimulation in neonatal rat atrial myocytes recruited (Ca²⁺-dependent) cPKC isoforms to the cell membrane, resulting in pronounced inhibition of acetylcholine (ACh)-activated currents [30]. The cPKC isoform PKCα induced inhibition of dog atrial GIRK currents under control conditions, whereas the nPKC isoform PKCε induced an increase of constitutive GIRK currents after atrial tachycardia-induced remodelling [24]. More recently, it was demonstrated that receptor-specific activation of the cPKC isoforms PKCα (during stimulation of α₁-adrenergic receptors (α₁-ARs)) and PKCβ (by Endothelin (ET)-receptors) accounts for receptor-dependent differences in atrial GIRK channel inhibition [29].

Although all these studies demonstrated that the specific recruitment of different PKC isoforms has divergent effects on GIRK channel activity, only few studies identified phosphorylation sites within the GIRK channels that are targeted by a specific PKC isoform. In the GIRK1 subunit, the serine residue S¹⁸⁵ has been identified as a PKC phosphorylation site. Site-directed mutagenesis of S¹⁸⁵ to alanine resulted in phosphorylation-deficient GIRK1 (S185A) channels that were insensitive to the PKC activator phorbol 12-myristate 13 acetate (PMA) and displayed weak channel inhibition [25].

Several phosphorylation sites in GIRK4 subunits have been suggested as targets for PKC-induced phosphorylation. Zhang and coworkers mutated 5 putative phosphorylation sites in the GIRK4 subunit to alanine (T37A, T67A, T70A, S209A, S233A) and investigated the inhibition of these phosphorylation-deficient channel mutants by the PKC activator PMA [37]. However, none of these channel mutants displayed reduced PMA-inhibition, indicating that PKC isoforms probably target distinct phosphorylation sites in the GIRK4 subunit.

Additional serine residues in the GIRK4 subunit were identified as putative PKC phosphorylation sites and mutated to alanine [12, 25]. Except the phosphorylation-deficient (S191A) channel mutant, all other GIRK4 channel mutants (e.g. S241A, S320A, S350A) exhibited pronounced channel inhibition by PMA. The phosphorylation-deficient (S191A) channel mutant was further insensitive to the catalytic subunit of PKC or to the neurotransmitter substance P, demonstrating that S¹⁹¹ is an important residue for PKC-induced GIRK channel inhibition. However, since PMA is a non-specific activator of several PKC isoforms, the identity of the specific PKC isoform that targets the serine residue S¹⁹¹ is currently unknown.

Up to now, only one phosphorylation site in the C-terminus of the GIRK4 subunit has been identified as a target for a specific PKC isoform. The serine residue S⁴¹⁸ is specifically phosphorylated by the novel PKC isoform PKCε [12]. The authors showed that coexpression of the catalytic subunit of PKCε had a facilitative effect on both basal and ACh-induced GIRK activity which was abolished by mutating serine⁴¹⁸ to alanine. Accordingly, expression of these phosphorylation-deficient GIRK4 (S418A) subunits in Xenopus oocytes reversed the PKCε-induced enhancement of basal and agonist-activated GIRK currents [12]. The functional impact of S⁴¹⁸ as a target for PKCε was further confirmed by experiments on rat atrial myocytes and in the heterologous expression system [16]. GIRK1/4 channel activity was reduced during stimulation of AT₁-receptors via activation of the novel PKC isoform PKCε in native atrial myocytes and in HEK 293 cells. The extent of AT₁-R-induced GIRK current reduction was attributed to the superposition of inhibitory and facilitative effects, depending on distinct phosphorylation sites within the GIRK4 subunit. Whereas the angiotensin-induced PKCε targets S⁴¹⁸ to facilitate GIRK currents [16], the specific phosphorylation site that contributes to the inhibitory effect of PKCε on GIRK channel activity was not identified. This study also demonstrated that GIRK current reduction during stimulation of wildtype α_1B-AR depends on the activation of the cPKC isoform PKCα but does not comprise activation of PKCε. The residue S⁴¹⁸ was excluded to be a primary target for the α_1B-AR-activated PKCα, but it is conceivable that phosphorylation of S¹⁹¹ contributes to α_1B-AR-induced GIRK inhibition.

Since the relevance of the phosphorylation site S¹⁹¹ for α_1B-AR- or AT₁-R-induced GIRK inhibition has not been elucidated up to now, the present study investigated the functional impact of S¹⁹¹ for receptor-specific GIRK inhibition. In electrophysiological whole-cell experiments on atrial myocytes and CHO cells, we monitored the activity of phosphorylation-deficient GIRK4 (S191A) channel mutants and quantified the receptor-specific extent of current inhibition.

As a novel finding, we demonstrate that stimulation of α_1B-ARs induced pronounced GIRK4 (S191A) channel inhibition, suggesting that S¹⁹¹ within the GIRK4 subunit is not subject to PKCα-induced phosphorylation. In contrast, GIRK4 (S191A) channel inhibition was significantly reduced during AT₁-R stimulation, indicating that the inhibitory effect of angiotensin comprises at least phosphorylation of S¹⁹¹.

We further investigated the receptor-dependent modulation of GIRK4 channels with a mutation of isoleucine I²²⁹ to leucine (GIRK4 (I229L)). This GIRK4 channel mutant shows a strengthened interaction of PIP₂ with the GIRK4 subunit [36] and is less sensitive to PKC-induced inhibition as compared to wildtype GIRK4 channels [38].

Since previous studies did not investigate G_qPCR-induced GIRK4 (I229L) inhibition, we aimed to clarify if enhanced PIP₂-channel interaction impedes the PKC-induced channel inhibition depending on the activated receptor species. Given that G_qPCRs receptor-specifically activate distinct PKC isoforms [3, 20, 29, 30], these experiments should figure out potential isoform-specific effects on the channel-PIP₂ interaction.

Materials and methods

Ethical approval

Rats were killed following protocols in accordance with the guidelines of the European Community (86/609/EEC) and approved by the animal welfare officer of the Ruhr-University Bochum.

Isolation and culture of atrial myocytes

Details of the isolation procedure of rat atrial myocytes have been described previously [16]. Briefly, Wistar Kyoto rats were anesthetized and killed by cervical dislocation. The chest was opened, the heart was removed and mounted on the cannula of a Langendorff perfusion system for coronary perfusion at constant flow. Isolated hearts were perfused with a nominally Ca²⁺-free solution (mmol/L) (NaCl 140, KCl 5.4, MgCl₂ 1, HEPES 10, N-acetyl-cysteine 3, pyruvic acid 5, EGTA 1.25, pH 7.4) for 10 min, then with enzyme solution (mmol/L) (NaCl 140, KCl 5.4, MgCl₂ 1, HEPES 10, collagenase II (Worthington Biochemical Corporation, NJ, USA) (1 mg/mL), protease type XIV (Sigma Aldrich) (0.25 mg/mL), DNAse I (Sigma) (0.1 mg/mL), taurine (Sigma) (0.625 mg/mL), 2,3-Butanedione monoxime (Sigma) (0.5 mg/mL)) for 15 min. Hearts were removed from the cannula and atria were mechanically dissected with fine forceps and chopped into 1 mm³ chunks. To isolate single cells, atrial tissue chunks were agitated using a Pasteur pipette in enzyme solution for 30 min. Enzymatic activity was disrupted by adding nominally Ca²⁺-free solution supplemented with fatty acid free albumin (Sigma) (1 mg/mL) and taurine (1.25 mg/mL). [Ca²⁺]_o in the supernatant solution was stepwise increased to 2 mmol/L within 2 h by exchanging the supernatant solution with a solution containing (in mmol/L) NaCl 140, KCl 5.4, MgCl₂ 1, HEPES 10 (pH 7.4) supplemented with 0.5 mmol/L, 1 mmol/L, 1.5 mmol/L and 2 mmol/L CaCl₂.

Cells were plated at a low density (several hundred cells per 35 mm dish) and cultured in FBS-free medium (HEPES-buffered M199, PAA Laboratories) supplemented with 1% ITS (insulin/transferrin/selenium, Gibco), Gentamycin (1.4 µg/mL, Sigma Aldrich) and Kanamycin (0.7 µg/mL, Sigma Aldrich). Cardiac cells were infected with an adenovirus (Ad) encoding for Ad-GFP or Ad-GIRK4 (rat GIRK4) or Ad-GIRK4-S191A (termed GIRK4 (S191A)) one day after isolation and analyzed 48 h after infection. Adenoviral shuttle vectors contain an independent cassette that expresses a green fluorescent protein (GFP) marker. Non-infected myocytes were used experimentally from day 1 until day 6 after isolation.

Cell culture

CHO cells were grown in IMDM medium containing fetal bovine serum (10%), 2% HT supplement, 1% non-essential amino acids (MEM-NEAAs 100x), 400 µg/mL G418 and 100 µg/mL hygromycin B. All cell culture media and supplements were purchased from Gibco. CHO cells were transfected using Lipofectamine (Invitrogen) according to the manufacturer’s instructions.

For electrophysiological experiments, CHO cells were transiently transfected with cDNAs encoding for human muscarinic M₂-receptor (1 µg) (cDNA Resource Center, Bloomsburg University, USA GenBank Accession # AF498916), human GIRK1 (1 µg) and GIRK4 (0.5 µg) (pcDNA3.1-GIRK1 and pcDNA3.1-GIRK4 were kindly provided by Dr. Eric Schulze-Bahr, University Muenster, Germany) or the GIRK4 mutants GIRK4 (S191A) (0.75 µg), GIRK4 (I229L) (0.75 µg) and GIRK4 (S143T) (0.75 µg) and cDNAs encoding for the following receptor species: At1R-P2A-mCherry (0.15 µg), encoding for rat AT₁-R (Addgene plasmid #112934, kindly provided by Dr. Dorus Gadella [26] or 0.75 µg pcDNA3.1⁺ [ADRA1B] encoding for human α_1B-AR (amount per 35 mm culture-dish). In case that the transfected vectors did not contain an independent cassette expressing a green fluorescent protein (GFP) marker, cells were cotransfected with a vector encoding for GFP only (pEGFP-N3, Clontech).

For fluorescence images, cells were transfected with pcDNA3-YFP-GIRK1 (0.75 µg) and pcDNA3-GIRK4 (0.75 µg) or pcDNA3-GIRK4 (S191A) (0.75 µg).

Molecular biology

The mutations S191A, S143T and I229L were introduced into rat GIRK4 cDNA using side-directed mutagenesis. PCR was performed using Phusion High-Fidelity (NEB) DNA polymerase and the following primer pairs:

• GIRK4-S191A-fw: 5´-GAT CGC CCA GCC AAA GAA GAG AGC AGA GAC C-3´.

• GIRK4-S191A-rev: 5´-GGC TGG GCG ATC TTT ATA AAC ATG CAA CCC AC-3´.

• GIRK4-S143T-fw : 5´-CCG CTT TCC TGT TCA CCA TTG AGA CCG AAA CAA CC-3´.

• GIRK4-S143T-rev: 5´-GGT TGT TTC GGT CTC AAT GGT GAA CAG GAA AGC GG-3´.

• GIRK4-I229L-fw: 5´ CTC CGC AAC TCC CAC CTC GTG GAG GCC TCC ATC C-3´.

• GIRK4-I229L-rev: 5´ GGA TGG AGG CCT CCA CGA GGT GGG AGT TGC GGA G-3´.

Successful mutagenesis was verified by DNA sequencing for all generated constructs.

The AdEasy technology [23] was used to generate recombinant adenoviruses.

GIRK4-S191A was subcloned into the shuttle vector pAdTrack-CMV using the restriction enzymes KpnI and EcoRV. The obtained plasmid GIRK4-S191A-pAdTrack-CMV was linearized using PmeI and transformed into AdEasier cells harboring the adenoviral backbone vector pAdEasy-1. Recombinant adenoviruses were produced by transfecting HEK293 cells with pAdEasy-1 plasmids using a lipid-based transfection reagent (Lipofectamine-LTX) at 50% cell confluency (25 cm² cell culture flask). Following transfection, the cells generated adenoviral particles that were released upon cell lysis and subsequently infected neighbouring cells. After an incubation time of 14–20 days, virus particles were harvested by repetitive freeze–thaw cycles, followed by centrifugation to remove cellular debris. The resulting supernatant was used to infect HEK293 cells to amplify virus production. This infection and harvest process was repeated 3–4 times to increase viral titers with progressively shorter times of cell lysis. The final concentrated virus stock was stored at − 80 °C. For infection of atrial myocytes, MOIs were adjusted to reach an infection rate of 50% (assessed as GFP-positive cells).

Current measurement

Membrane currents were measured using whole-cell patch clamp. Pipettes were fabricated from borosilicate glass and were filled with the solution listed below (pipette resistance 4–6 MΩ). Currents were measured by means of a patch clamp amplifier (WPC-100, ESF electronic or Axopatch 200 A, Axon Instruments) connected to a data acquisition system LIH8 + 8 (HEKA, HEKA Elektronik, Germany) for voltage control and data acquisition. Signals were filtered (corner frequency 1 kHz), digitally sampled at 200 Hz and stored on a computer equipped with the software “Patchmaster” (HEKA, Germany) for data recording and analysis. Experiments were performed at ambient temperature (23–26 °C). Application of different solutions was performed by means of a custom-made solenoid-operated flow system, permitting a change of solution around an individual cell with a half time of about 100 ms. For measurements cells devoid of contact with neighboring cells were selected. Transfected CHO/HEK cells or adenoviral infected atrial cells were identified by epifluorescence of GFP.

The cells were voltage-clamped at a holding potential of -90 mV, i.e. negative to the equilibrium potential for K⁺ (E_K = -48 mV), resulting in inward K⁺ currents. In some experiments, voltage ramps (duration 500 ms) from − 120 mV to + 60 mV were applied every 10–20 s to obtain current-voltage relations (membrane currents in response to depolarizing voltage ramps are shown as upward deflections). Zero current levels are indicated by dotted lines in the corresponding figures. The “zero GIRK level” was determined as the current that was measured immediately after patch rupture.

Solutions and chemicals

The extracellular solution for whole cell measurements of membrane currents contained (in mmol/L): NaCl 122, KCl 20, CaCl₂ 2, MgCl₂ 1, HEPES/NaOH 10 (pH 7.4). The internal (pipette) solution contained (in mmol/L): K-aspartate 100, KCl 40, NaCl 5, MgCl₂ 2, Na₂ATP 5, EGTA 2, GTP 0.025, and HEPES/KOH 20 (pH 7.4). The K⁺ reversal potential under this condition was calculated as -48 mV.

In some experiments, 0.5 mmol/L of the non-hydrolyzable GTP-analog GTPγ-S was included in the pipette solution. Standard chemicals were from Merck (Darmstadt, Germany). EGTA, HEPES, Guanosine 5′-triphosphate sodium salt hydrate (GTP), Guanosine 5′-(γ-thio)-triphosphate tetralithium salt (GTPγ-S), Na₂ATP, Acetylcholine-chloride (ACh), XAF-1407, Phorbol 12-myristate 13-acetate (PMA) and phenylephrine were from Sigma-Aldrich. Angiotensin II was from Bio-Techne, (Wiesbaden, Germany).

Statistical analysis

Statistical analysis was performed with the software OriginPro2025 (OriginLab Corporation, Northampton, MA, USA, RRID: SCR_014212). All data are presented as individual observations or summarized data (mean ± S.E. of n cells). To compare the means between two groups, the Student’s t-test was used if the data passed the Shapiro-Wilk normality test and the equal variance test (Levene). P-values less than 0.05 were considered statistically significant. If the data were not distributed normally, the Mann-Whitney Rank Sum test was used.

Comparisons between multiple groups were performed using One-Way Anova after testing normal distribution and equal variance. The Holm-Sidak post-hoc test was assessed for statistical significance (p < 0.05). In case that the assumption of homogeneity of variance between the groups was violated in the Anova analysis, Kruskal-Wallis-Anova and Dunn or Conover post hoc test were used. P-values < 0.05 were considered statistically significant. In general, p-values < 0.05 were marked by an asterisk, p-values < 0.01 were marked by two asterisks and p-values < 0.001 were marked by three asterisks.

Results

PMA-induced current inhibition in atrial myocytes expressing the phosphorylation-deficient GIRK4 (S191A) mutant

Previously published studies on Xenopus oocytes described that mutation of GIRK4 S¹⁹¹ significantly impaired current reduction by the PKC activator PMA (Phorbol 12-myristate 13-acetate) [1, 25]. The question arises if similar effects are also evident in rat atrial myocytes.

In a first series of experiments, we determined the concentration-response curves for wildtype GIRK 4 and mutant GIRK4 (S191A) (phosphorylation-deficient) channels by exposing adenovirus(Ad)-infected rat atrial myocytes to different ACh concentrations. The GIRK current amplitudes during applications of various ACh concentrations were normalized to the maximal response (I/I_max) recorded with 10 µmol/L ACh. The EC₅₀ values for ACh-induced activation of GIRK channels were 0.58 µmol/L for wildtype GIRK4 channels (Fig. 1C) and 0.79 µmol/L for the GIRK4 (S191A) channel mutant (Fig. 1D).

Fig. 1.

PMA-induced inhibition of GIRK currents is significantly reduced in atrial myocytes expressing GIRK4 (S191A) channel subunits. A. Schematic illustration of a GIRK4 subunit highlighting PKC phosphorylation site S¹⁹¹. Residue S¹⁹¹ in the GIRK4 subunit has been identified as a phosphorylation site contributing to PKC-mediated GIRK inhibition. Substitution of serine with alanine (S191A) generates a phosphorylation-deficient variant. B. Scheme of the G_qPCR-induced inhibition of atrial GIRK currents via different PKC-Isoforms. The specific PKC isoform that targets S¹⁹¹ within the GIRK4 subunit is currently unknown. C-D. Concentration-response curves for ACh-activated GIRK4- (C) or GIRK4 (S191A)-currents (D). GIRK currents in rat atrial myocytes were activated by repeated applications (duration 20 s) of different ACh-concentrations (GTP-containing pipette solution). EC₅₀ values of GIRK4- and GIRK4 (S191A)-currents were 0.579 µmol/L and 0.791 µmol/L. Number of experiments (n) for each concentration is indicated in parentheses. E-H. Representative recordings of ACh (10 µmol/L)-activated GIRK currents from non-infected atrial myocytes (control, E) or from atrial myocytes infected with adenoviruses encoding for GFP (F), wildtype-GIRK4 subunits (G) or GIRK4 (S191A) subunits (H). Currents were stable activated by dialysis with a GTPγ-S (0.5 mmol/L)-containing pipette solution. GIRK inhibition was induced by external application of PMA (500 nmol/L). I. Summarized data (mean ± S.E. of n cells) of PMA-induced GIRK current inhibition. Individual data points are indicated by diamonds. Multiple group comparisons were performed with a One-Way Anova after testing normal distribution and equal variance (F-value = 21.8) and Holm-Sidak post-hoc test. P-values < 0.05 were marked by one asterisk, a p-value < 0.001 by three asterisks

Next, the extent of current inhibition by the PKC activator PMA was evaluated in atrial myocytes infected with Ad-GFP or wildtype Ad-GIRK4 or Ad-GIRK4 (S191A). To obtain a stable amplitude of GIRK channel current, cells were continuously exposed to a saturating ACh concentration (10 µmol/L) and loaded with GTPγ-S (0.5 mmol/L) via the patch pipette. Repetitive ACh applications accelerate the activation of IK_(ACh) during loading the cells with GTPγ-S. To achieve full current activation, sometimes more than one application of ACh was required (e.g. Figure 1F). In the absence of PMA, the endogenous atrial GIRK1/4 current remained fairly constant for several minutes of recording (Fig. 1E), indicating irreversible activation of GIRK channels. On average, the GTPγ-S-activated current in the absence of PMA (control) decayed spontaneously by about 15% within 5 min (see summarized data in Fig. 1I).

To quantify the extent of GIRK inhibition, we determined the steady-state amplitude of the GTPγ-S-activated GIRK current and related the PMA-induced current inhibition (I_inhib) to this value. Application of PMA (500 nmol/L) for at least 3 min resulted in pronounced inhibition of GTPγ-S-activated GIRK currents in Ad-GFP-expressing cells to an average of 45% (Fig. 1F, I) and to 64% in Ad-GIRK4-infected atrial myocytes (Fig. 1G, I). These data suggest that homomeric GIRK4 channels display a higher sensitivity to PKC-mediated phosphorylation than heteromeric GIRK1/GIRK4 channels.

It is worth mentioning that the high efficiency of adenoviral infection results in a high number of homomeric GIRK4 channels that exhibit different characteristics from endogenous GIRK1/4 channels such as slower activation kinetics and a loss of desensitization (see also [2, 28]). These characteristic features of GIRK4 homomeric channels are also evident in the current recordings shown in Figs. 1G and 2B, demonstrating that whole cell currents are highly dominated by the properties of homomeric GIRK4 channels.

Fig. 2.

Mutation of S¹⁹¹ to A¹⁹¹ impairs the Ang II-induced GIRK current inhibition but not the Phe-induced current reduction. A-C. Recordings of GTPγ-S (0.5 mmol/L)-activated GIRK currents during coapplication of ACh (10 µmol/L) and Phe (100 µmol/L) from native atrial myocytes (A) or myocytes infected with Ad-GIRK4 (B) or Ad-GIRK4 (S191A) (C). D-F. Representative recordings of GTPγ-S (0.5 mmol/L)-activated GIRK currents during coapplication of ACh (10 µmol/L) and Ang II (1 µmol/L) from native (named GIRK1/4 in G and H) atrial myocytes (D) or myocytes infected with Ad-GIRK4 (E) or Ad-GIRK4 (S191A) (F). G. Summarized data of Phe-induced GIRK inhibition (mean ± S.E.). Individual data points are indicated by diamonds. Statistical significance was tested with a Kruskal-Wallis Anova and Dunn´s post-hoc test. P-value < 0.05 was marked by one asterisk. P-value < 0.001 was marked by three asterisks. n.s. = not significant. H. Summarized data of Ang II-induced GIRK inhibition (mean ± S.E.). Statistical significance was tested with a One-Way Anova (F-value = 17.35) and Holm-Sidak post-hoc test. P-values < 0.001 were considered significant and marked by three asterisks. I. Summarized data of receptor-specific effects on GIRK currents in GIRK4 (S191A)-expressing cells. Statistical significance was tested with a t-test (t-value = 4.96). P-value < 0.001 was marked by three asterisks

PMA inhibition was significantly reduced for GIRK4 (S191A)-containing channels (Fig. 1H) compared to wildtype GIRK4 channels (Fig. 1G) and to endogenous GIRK1/4 channels in GFP-infected cells (Fig. 1F) and amounted to about 26% (Fig. 1H and summarized data in Fig. 1I). The reduction of PMA-induced inhibition of atrial GIRK4 (S191A)-channels described above is in line with previous findings on Xenopus oocytes [1, 25]. Furthermore, the significant difference in the GIRK current reduction of GFP- and GIRK4 (S191A)-infected cells indicates a sufficient expression of GIRK4 (S191A) subunits that associate with endogenous GIRK subunits to form heteromeric phosphorylation-deficient GIRK4 (S191A)-containing GIRK channels.

Receptor-specific differences in the inhibition of GIRK4 (S191A)-containing channels in atrial cells

As mentioned above, several studies described G_qPCR-specific activation of distinct PKC isoforms [3, 16, 20, 24, 29, 30], resulting in receptor-dependent modulation of GIRK channel activity. However, it is not known if the phosphorylation site S¹⁹¹ within the GIRK4 subunit is targeted by different PKC isoforms. We therefore aimed to investigate if S¹⁹¹ represents a specific target for the Ang II-activated PKCε isoform and/or for the α_1B-AR activated PKCα (see cartoon in Fig. 1B).

To evaluate the inhibitory potential of different G_qPCR agonists on mutant GIRK channel activity, we stable activated GIRK currents by GTPγ-S via the patch pipette in the continuous presence of ACh and then coapplied phenylephrine (100 µmol/L) or angiotensin (1 µmol/L) to achieve full activation of α_1B-ARs and AT₁-Rs. Figure 2A-F shows representative current recordings from native atrial myocytes or from myocytes infected with Ad-GIRK4 or Ad-GIRK4 (S191A) during exposure to ACh and Phe.

In untreated (native) atrial myocytes (named GIRK1/4 in Fig. 2G), activation of α_1B-ARs resulted in a gradual GIRK current decline by about 50% (Fig. 2A and summarized data in Fig. 2G). The upward reflections represent current responses to depolarizing voltage ramps (-120 mV to + 60 mV, duration 500 ms, applied every 10 s). These voltage ramps were applied to obtain the current-voltage relation of ACh-activated GIRK currents (see Online Resource Supplemental Fig. 1).

In atrial myocytes, overexpressing wildtype GIRK4 subunits, Phe-induced current inhibition was significantly increased compared to control cells and amounted to about 80% (Fig. 2B and summarized data in Fig. 2G).

Adenoviral-induced expression of phosphorylation-deficient GIRK4 (S191A) subunits should give rise to the formation of heteromeric GIRK4 (S191A)-containing GIRK channels. Activation of α_1B-ARs results in an inhibition of GIRK4 (S191A) currents that amounted to an average of about 73% (Fig. 2C) which is almost identical to the Phe-induced wildtype GIRK4 inhibition (see summarized data in Fig. 2G). Since the exact subunit composition of GIRK channels containing the mutated GIRK4 (S191A) subunits cannot be predicted, the reducing effect of GIRK4 (S191A) subunits on current inhibition might be blunted by endogenous atrial GIRK channels. If this is the case, the α_1B-AR-induced current inhibition should be in the same order of magnitude in both native GIRK1/4-expressing cells and in GIRK4(S191A)-infected atrial cells. However, the Phe-induced reduction of GIRK4 (S191A) currents was significantly larger than that of endogenous atrial GIRK1/4 channels (see Fig. 2G). Thus, despite an unknown stoichiometry of GIRK4 (S191A)-containing GIRK channels, the different extent of Phe-induced current reduction in untreated atrial cells and GIRK4 (S191A)-expressing cells suggests that the whole cell current is dominated by the properties of heteromeric GIRK4 (S191A) channels.

Analogous experiments on GTPγ-S-dialyzed atrial myocytes expressing either GIRK4 or GIRK4 (S191A) subunits were performed during coapplication of Ang II (1 µmol/L) and ACh (10 µmol/L). As shown in the representative trace in Fig. 2D and in the summarized data in Fig. 2H, the extent of Ang II-induced GIRK inhibition in untreated atrial myocytes was determined to about 40% (see also [16]). In Ad-GIRK4-infected cells, Ang II caused an IK_(ACh) inhibition of about 75% (Fig. 2E) which was significantly reduced to about 46% in atrial cells coexpressing GIRK4 (S191A) subunits (Fig. 2F, see also summarized data in Fig. 2H). We directly compared the change in IK_(ACh) inhibition caused by α_1B-ARs and AT₁-Rs for the GIRK4 (S191A) mutant (Fig. 2I) and found significant receptor-dependent differences in the extent of G_qPCR-induced GIRK inhibition. More precisely, inhibition of GIRK (S191A)-containing channels was impeded during stimulation of AT₁-Rs, but not during activation of α_1B-ARs.

Coexpression of GIRK1 and GIRK4 (S191A) subunits in CHO cells results in functional heteromeric GIRK channels

The data presented in Figs. 1 and 2 suggest that the interaction of overexpressed GIRK4 (S191A) subunits with endogenous GIRK1 or GIRK4 subunits determines the properties of the whole cell IK_(ACh). However, the exact subunit stoichiometry of GIRK4 (S191A)-containing channels is unknown. It is conceivable that GIRK4 (S191A) subunits replace the endogenous GIRK4 subunits or that a “non-conventional” stoichiometry of GIRK1/GIRK4/GIRK4 (S191A)-containing channels occurs. We therefore repeated these experiments in CHO cells transfected with muscarinic M₂-Rs, the G_qPCRs α_1B-AR or AT₁-R and with various GIRK subunits to ensure a more defined expression of mutant GIRK channels. We first tested if coexpression of M₂-receptors and GIRK4 (S191A) subunits in CHO cells give rise to functional homomeric GIRK4 (S191A) channels. As shown in Fig. 3A, single expression of the GIRK4 channel mutant S191A did not result in detectable ACh-activated inward currents. However, GIRK4 (S191A) subunits are capable to form functional heteromeric channels with coexpressed GIRK1 subunits (Fig. 3B) that display a characteristic inwardly rectifying current-voltage relation (Fig. 3F).

Fig. 3.

GIRK4 (S191A) subunits do not form functional homomeric channels but generate functional heteromeric channels with coexpressed GIRK1 subunits. A. Representative recording of a membrane current during application of ACh (10 µmol/L) in a M₂-R/GIRK4 (S191A)-transfected CHO cell (out of 6 cells). B. ACh-activated GIRK current in a M₂-R/GIRK1/GIRK4 (S191A)-expressing cell (out of 12 cells). Experiments were performed with a GTP-containing pipette solution. Membrane currents in response to depolarizing voltage ramps (from − 120 mV to + 60 mV, duration 500 ms applied every 10 s) are shown as upward deflections. C-E. Fluorescence images taken from CHO cells transfected with a cDNA encoding for N-terminally-tagged YFP-GIRK1 subunits (C) or from CHO cells coexpressing untagged GIRK4 subunits (D) or untagged GIRK4 (S191A) subunits (E). F. current-voltage relation (I/V curve) of an ACh-induced current in a M₂-R/GIRK1/GIRK4 (S191A)-cotransfected CHO cell. The I/V curve was obtained by subtracting the basal current response to a depolarizing voltage ramp (Ba) from an ACh-activated current (Bb). G. Summarized amplitudes of ACh (10 µmol/L)-activated GIRK currents in CHO cells expressing GIRK4(S191A) subunits alone or coexpressing GIRK4 and/or GIRK1 subunits. Individual data points are indicated by diamonds. Statistical significance was tested with a Kruskal-Wallis Anova and Conover post hoc test. P-values < 0.001 were marked by three asterisks

The moderate activation of GIRK1/GIRK4 (S191A) channels and the resulting small amplitude of IK_(ACh) let the question arise if the GIRK4 (S191A) subunit is properly folded or adequately transported to the plasma membrane.

We therefore investigated if GIRK4 (S191A) subunits were able to translocate GIRK1 subunits to the plasma membrane of CHO cells. As depicted in the fluorescence image in Fig. 3C, CHO cells, transiently transfected with N-terminally YFP (Yellow fluorescent protein)-fused GIRK1 subunits, showed a predominant cytosolic expression of YFP-GIRK1 subunits. Coexpression of untagged GIRK4 subunits resulted in pronounced targeting of YFP-GIRK1 subunits to the cell surface (Fig. 3D). These data are in line with previous results from Kennedy and coworkers [17] who evaluated the function of GIRK4 subunits in promoting the cell surface localization of heteromeric GIRK1/4 channels. In that study, it was demonstrated that GIRK1/GIRK4 chimeras containing the C-terminal tail of GIRK1 were retained inside COS-7 cells, while chimeras containing the C terminus of GIRK4 were localized at the cell surface.

In the present study, we observed fluorescence of the cell membrane upon coexpression of YFP-GIRK1 and untagged GIRK4 (S191A) subunits, indicating that GIRK1 and GIRK4 (S191A) subunits colocalize at the cell surface and form heteromeric GIRK1/GIRK4 (S191A) channels (Fig. 3E). These data refute the objection that the small current amplitude of heteromeric GIRK4 (S191A)-containing channels results from a partial rescue of nonfunctional GIRK4 (S191A) proteins by GIRK1 subunits. Upon coexpression of wildtype GIRK4, GIRK1 and GIRK4 (S191A) subunits, the IK_(ACh) amplitudes significantly increased (Fig. 3G), probably by increasing the formation of heteromeric GIRK1/4 channels.

Phe- and Ang II-induced inhibition of mutant GIRK4 (S191A)-containing channels in CHO cells

Although the coexpression of GIRK1 and GIRK4 (S91A) subunits results in functional heteromeric channels in CHO cells, the small amplitudes of ACh-activated inward currents are inadequate to analyze receptor-specific differences in GIRK channel regulation.

Before testing other subunit compositions, we aimed to repeat the experiments that were performed on atrial myocytes (Fig. 2) by coexpressing M₂-receptors and the subunits GIRK1, GIRK4 and GIRK4 (S191A) subunits in CHO cells. The purpose of this experimental design is to evaluate possible cell-type specific differences in G_q-signaling pathways that might affect receptor-dependent inhibition of GIRK4 (S191A) currents.

First, we determined the extent of G_qPCR-induced GIRK1/4 current inhibition without coexpression of GIRK4 (S191A) subunits. Data obtained under these experimental conditions served as control and are entitled “-GIRK4 (S191A)” in Fig. 4E and F. According to different expression levels of α_1B-ARs and AT₁-R in CHO cells and in native atrial myocytes, we reduced the concentration of Phe to 10 µmol/l and the Ang II concentration to 100 nmol/L to activate a sizeable number of receptors but to avoid maximal current inhibition (see also [16]). The representative recordings of GTPγ-S-activated GIRK currents in GIRK1/4-transfected cells show pronounced inhibition by about 80% during stimulation of α_1B-ARs (Fig. 4A) and during activation of AT₁-Rs (Fig. 4C, see also summarized data in E and F).

Fig. 4.

Receptor-species dependent inhibition of mutated GIRK4 (S191A) channels in CHO cells. A. Recordings of GTPγ-S (0.5 mmol/L)-activated GIRK currents during coapplication of ACh (10 µmol/L) and Phe (10 µmol/L) from CHO cells transfected with M₂-R, GIRK1, GIRK4 and α_1B-AR (control condition, termed -GIRK4 (S191A)). B. GIRK currents from CHO cells additionally expressing GIRK4 (S191A) subunits. C-D. Representative GTPγ-S-activated GIRK currents during application of ACh (10 µmol/L) and Ang II (100 nmol/L) from CHO cells transfected with M₂-R, GIRK1, GIRK4 and AT₁-R (C) or from cells additionally expressing GIRK4 (S191A) subunits (D). E. Summarized data of Phe-induced GIRK inhibition (mean ± S.E.). Single data points are indicated by diamonds. Significance was evaluated with a Mann-Whitney test. Number of experiments as indicated. n.s. = not significant. F. Summarized data of Ang II-induced GIRK inhibition (mean ± S.E.). Significance was evaluated with a Student´s t-test (t-value = 3.35). P-value < 0.01 was marked by two asterisks. G. Summarized data of receptor-specific effects on GIRK currents in M₂-R/GIRK1/4/G_qPCR-transfected CHO cells coexpressing GIRK4 (S191A) subunits. Statistical significance was tested with a Student´s t-test (t-value = 2.58). P-value < 0.05 was marked by one asterisk

To evaluate the receptor-specific inhibition of the GIRK4 channel mutant S191A in CHO cells, we next compared the rate of current reduction during coapplication of ACh and Phe or Ang II in cells transfected with M₂-R, GIRK1, GIRK4, GIRK4 (S191A) and the corresponding G_qPCRs. Analogous to the experiments on atrial myocytes (Fig. 2), application of Phe (10 µmol/L) resulted in pronounced inhibition of GTPγ-S-activated GIRK currents (Fig. 4B) by about 74%, indicating that the Phe-induced inhibition of mutant GIRK (S191A)-containing channels was not impeded (see summarized data in Fig. 4E). In contrast, Ang II-mediated current inhibition was significantly reduced from 80% under control conditions (Fig. 4C) to 45% (Fig. 4D) for GIRK4 (S191A)-containing channels (summarized data in Fig. 4F). As shown by the summarized data in Fig. 4G, significant receptor-specific differences in the G_qPCR-induced modulation of GIRK4 (S191A) channel activity can also be confirmed in the heterologous expression system. Furthermore, the qualitative differences in Phe- and Ang II-effects in S191A-transfected cells indicate that, apart from wildtype GIRK1/4 channels, a substantial amount of S191A-containing channels is expressed in the heterologous expression system.

Heteromeric GIRK4 (S143T)/GIRK4 (S191A) channels display reduced Ang II-induced inhibition

Although the reducing effect of the S191A mutation on Ang II-induced IK_(ACh) inhibition was observed in both atrial myocytes and GIRK1/GIRK4/GIRK4 (S191A)-transfected CHO cells, strong inhibition of heteromeric GIRK1/4 channels might still mitigate the reducing effect of GIRK4 (S191A) subunits. To avoid coexpression of GIRK1 subunits, but to obtain GIRK4 (S191A)-containing channels with sizeable current amplitudes, we introduced the S191A mutation in a GIRK4 mutant with a threonine substitution of serine¹⁴³ (GIRK4 (S143T)). This S143T mutant was initially described by Vivaudou and coworkers who demonstrated that expression of GIRK4 (S143T) subunits in Xenopus oocytes resulted in highly active homomeric GIRK4 (S143T) channels with large GIRK current amplitudes [34]. The high activity of GIRK4 (S143T) channels was also confirmed upon expression of this mutant in HEK 293 cells [5].

A previous study of Adney and coworkers [1] showed that the expression of double mutated GIRK4 (S143T-S191A) subunits in Xenopus oocytes resulted in inward currents through homomeric GIRK4 (S143T-S191A) channels with similar current levels compared to GIRK4 (S143T) channels. In contrast to the experiments in Xenopus oocytes, we were unable to elicit sizeable ACh-activated currents in CHO cells expressing muscarinic M₂-receptors and GIRK4 (S143T-S191A) subunits alone. Furthermore, coexpression of GIRK1 and GIRK4 (S143T-S191A) subunits (see Online Resource Supplemental Fig. 2) did not result in highly active heteromeric GIRK4 (S191A)-containing channels.

As we demonstrated that GIRK4 (S191A) subunits are able to translocate GIRK1 subunits to the cell membrane (see Fig. 3), the question arises if the mutation at position S¹⁴³ affects the translocation of GIRK4 (S143T-S191A) subunits to the plasma membrane. In a previous study, Vivaudou and coworkers demonstrated a strong fluorescence signal of GFP-fused GIRK4 (S143F) channel mutants at the plasma membrane, indicating that a point mutation at the position 143 of GIRK4 did not alter the subcellular localization of GIRK4 subunits [34]. Thus, the reduced channel activity of homomeric GIRK4 (S143T-S191A) channels is unlikely to result from altered membrane localization. Previously, a region in the carboxyl tail of GIRK4 (amino acids 209–245) has been shown to be crucial for Gβγ binding and IK_(ACh) activation [19]. Point mutations at the positions GIRK4²²⁶, GIRK4²³³ GIRK4²³⁷, GIRK4²⁴³, and GIRK4²⁴⁵ completely abolished the activity of GIRK1/mutant GIRK4 channels. Moreover, a single point mutation at position 216 in the GIRK4 subunit from cysteine to threonine (GIRK4 (C216T)) significantly reduced the sensitivity of heteromeric GIRK1/mutant GIRK4 (C216T) channels to Gβγ more than 50 fold as compared to wildtype GIRK1/GIRK4 channels. Although we did not investigate the Gβγ sensitivity of the GIRK4 (S191A) mutant, it is tempting to speculate that a mutation in close proximity to the Gβγ binding region of GIRK4 affects the channel activation.

Since the amplitudes of GIRK4 (S143T-S191A) currents are inadequate to monitor receptor-specific GIRK regulation, we cotransfected CHO cells with single cDNAs encoding for GIRK4 (S143T) and GIRK4 (S191A) subunits to obtain heteromeric GIRK4 (S143T)/GIRK4 (S191A) channels.

In a first series of experiments, we investigated the properties of ACh-activated homomeric GIRK4 (S143T) channels in CHO cells coexpressing muscarinic M₂ receptors (Fig. 5A). Application of ACh (10 µmol/L) induced rapidly activating inward currents that slightly declined in the continuous presence of ACh. In line with previous experimental findings of Vivaudou and coworkers [34], the current-voltage relation of ACh-activated GIRK4 (S143T) channels in CHO cells (Fig. 5B) displayed a weaker inward rectification than native atrial GIRK1/GIRK4 channels (see Online Resource Supplemental Fig. 1). As soon as the IK_(ACh) amplitude remained stable, 1 µmol/L of the selective GIRK1/4 inhibitor XAF-1407 was coapplied which induced an inhibition of IK_(ACh) by about 80% (Fig. 5A). Upon washout of XAF, the amplitude of IK_(ACh) partially recovered during repetitive applications of acetylcholine. The current-voltage relation of the XAF-inhibited current (Fig. 5C) displayed a weak inward rectification, indicating that XAF-1407 efficiently blocks homomeric GIRK4 (S143T) channels.

Fig. 5.

ACh-activated GIRK currents in CHO cells transfected with mutant GIRK4 (S143T) subunits (± GIRK4 (S191A)) and muscarinic M₂ receptors (M₂-R). A. Membrane recording of ACh-evoked GIRK4 (S143T) currents in CHO cells (GTP-containing pipette solution). The dotted line indicates the zero current level. The dashed lines represent the steady-state current and the GIRK current at the end of the XAF-application. Application of ACh (10 µmol/L) and XAF-1407 (1 µmol/L) is indicated by bars. Membrane currents in response to depolarizing voltage ramps (from − 120 mV to + 60 mV, duration 500 ms, applied every 20 s) are shown as upward deflections B. representative current-voltage (I/V) curve, obtained by subtraction of membrane currents in the absence (Aa) or presence (Ab) of ACh (10 µmol/L) of a CHO cell expressing homomeric GIRK4 (S143T) channels. C. I/V curve of the XAF-inhibited current, obtained by subtraction of GIRK4 (S143T) currents in the absence (Ab) or presence (Ac) of XAF-1407. D. current recording of an ACh-activated current of a CHO cell coexpressing GIRK4 (S143T), GIRK4 (S191A) and M₂-R in the presence or absence of XAF. E. representative I/V curve of IK_(ACh) (ACh 10 µmol/L) of a CHO cell expressing M₂-R and GIRK4 (S143T) and GIRK4 (S191A) subunits. F. Summarized data of the amplitudes of basal and ACh-induced membrane currents of homomeric GIRK4 (S143T) channels and heteromeric GIRK4 (S143T)/GIRK4 (S191A) channels. Individual data points are indicated by diamonds. Pairwise comparison of basal and ACh-activated currents was performed by a Student´s t-test (t-values were 0.47 (basal currents) and 3.21 (ACh-activated currents)). p-value < 0.01 was marked by two asterisks. n.s. not significant

In CHO cells, coexpressing M₂-Rs and the subunits GIRK4 (S143T) and GIRK4 (S191A), ACh-activated GIRK currents were effectively blocked by 1 µmol/L XAF-1407 (Fig. 5D). The current-voltage relation of IK_(ACh) in these cells was characterized by a stronger inward rectification (Fig. 5E) compared to homomeric GIRK4 (S143T) currents. By evaluating the amplitudes of agonist-independent (basal) and ACh-activated currents of homomeric GIRK4 (S143T) channels and GIRK4 (S191A)-containing channels, we measured almost identical mean basal current amplitudes but a significant reduction of ACh-activated currents upon coexpression of GIRK4 (S191A) subunits (Fig. 5F) from 1272 ± 254 pA to 383 ± 136 pA (mean ± S.E.). Both the reduction of ACh-activated current amplitudes and the strong inward rectification of IK_(ACh) indicate a significant expression of functional heteromeric GIRK4 (143T)/GIRK4 (S191A) channels.

Next, we investigated the α_1B-AR and AT₁-R-induced inhibition of mutated homomeric GIRK4 (S143T) and heteromeric GIRK4 (S143T)/GIRK4 (S191A) channels to evaluate potential receptor-specific differences in GIRK channel modulation. GIRK currents in M₂-R/G_qPCR/GIRK4 (S143T)-expressing CHO cells were stable activated by internal dialysis with GTPγ-S (0.5 mmol/L). We then coapplied acetylcholine (10 µmol/L) and phenylephrine (10 µmol/L) or angiotensin to measure the receptor-induced inhibition of homomeric GIRK4 (S143T) channels. Interestingly, a concentration of 100 nmol/L Ang II reduced GTPγ-S activated currents of GIRK1/4 channels by about 80%, but only by 40% for GIRK4 (S143T) currents (see Online Resource Supplemental Fig. 3), indicating a lower Ang II sensitivity of homomeric GIRK4 (S143T) channels compared to wildtype GIRK1/4 channels. According to the lower sensitivity of GIRK4 (S143T) channels to angiotensin II, we increased the Ang II concentration to 1 µmol/L to achieve a sizeable current inhibition.

Application of Phe induced a pronounced reduction of GTPγ-S-activated GIRK4 (S143T) currents by about 80% (Fig. 6A and E). The Ang II-induced inhibition of GIRK4 (S143T) currents amounted to about 70% (Fig. 6C and F). During stimulation of α_1B-receptors in GIRK4 (S143T)/GIRK4 (S191A) expressing cells, IK_(ACh) was reduced by 85% (Fig. 6B and E), indicating that the Phe-induced inhibition of mutant GIRK (S191A)-containing channels was not alleviated. In contrast, Ang II-induced current inhibition was significantly reduced to 52% for heteromeric GIRK4 (S143T)/GIRK4 (S191A) channels (Fig. 6D and F), demonstrating a reducing effect of GIRK4 (S191A) subunit expression on Ang II-mediated GIRK channel inhibition.

Fig. 6.

Currents through heteromeric GIRK4 (S143T)/GIRK4 (S191A) channels show impaired Ang II-induced inhibition. A. Current recordings of GTPγ-S (0.5 mmol/L)-activated GIRK currents during coapplication of ACh (10 µmol/L) and Phe (10 µmol/L) from a CHO cell transfected with M₂-R, GIRK (S143T) and α_1B-AR. B. GIRK currents from a CHO cell additionally expressing GIRK4 (S191A) subunits. C-D. Representative GIRK currents during coapplication of ACh (10 µmol/L) and Ang II (1 µmol/L) from CHO cells expressing M₂-R, GIRK4 (S143T) and AT₁-R (C) or from cells additionally transfected with a cDNA encoding for GIRK4 (S191A) subunits (D). E. Summarized data of Phe-induced GIRK inhibition (mean ± S.E.). Individual data points are indicated by diamonds. Significance was evaluated with a Mann-Whitney test. Number of experiments as indicated. n.s. = not significant. F. Summarized data of Ang II-induced GIRK inhibition (mean ± S.E.). Significance was evaluated with a Student´s t-test (t-value = 2.99). P-value < 0.01 was marked by two asterisks. G. Summarized data of receptor-specific effects on GIRK currents in M₂-R-, GIRK4 (S143T)/GIRK4 (S191A)- and G_qPCR-transfected CHO cells. Statistical significance was tested with a Student´s t-test (t-value = 4.75). P-value < 0.001 was marked by three asterisks

We compared the extent of GIRK inhibition induced by Phe and Ang II for GIRK4 (S143T)/GIRK4 (S191A) channels (see Fig. 6G) and found significant differences between the current reduction caused by α_1B- and AT₁-R-receptor stimulation. These data further support the notion that the phosphorylation site S¹⁹¹ is differentially addressed by α_1B-ARs and AT₁-Rs.

PKC-sensitivity of mutant GIRK4 (I229L) channels

As mentioned above, increased PIP₂-channel interaction should modify the PKC-induced channel inhibition [38]. A point mutation of isoleucine²²⁹ in the GIRK4 subunit to leucine resulted in GIRK channels that strongly interact with PIP₂ and display minimal PKC sensitivity during application of PMA [36, 38]. However, PMA directly activates several PKC isoforms without prior stimulation of G_q signaling pathways. Up to now, the PKC sensitivity of mutant GIRK4 (I229L) channels during stimulation of different G_qPCRs has not been investigated. We therefore quantified the inhibition of homomeric GIRK4 (I229L) channels during stimulation of α_1B-ARs or AT₁-Rs to evaluate potential receptor-specific differences in the PKC-sensitivity of mutant channels.

In the original study of Zhang et al. the point mutation I229L was introduced into the highly active GIRK4 (S143T) subunit to achieve large current amplitudes [36].

We coexpressed M₂ receptors with GIRK4 (I229L) subunits in CHO cells to investigate if GIRK4 (I229L) subunits form homomeric channels with sizeable current amplitudes. With GTP in the pipette solution, application of ACh (10 µmol/L) frequently activated GIRK currents through homomeric GIRK4 (I229L) channels (see representative trace in Fig. 7A). Internal dialysis of GIRK4 (I229L)-expressing cells with GTPγ-S (0.5 mmol/L) induced slowly increasing GIRK currents (Fig. 7C). As soon as the IK_(ACh) amplitude remained stable for at least 20 s, we applied XAF-1407 (1 µmol/L). Interestingly, the subsequent application of this selective GIRK1/4 inhibitor did not result in pronounced current inhibition (Fig. 7C, E), suggesting that homomeric GIRK4 (I229L) channels are rather insensitive to XAF-1407. Coexpression of GIRK4 (S143T) subunits resulted in ACh-sensitive heteromeric GIRK4 (S143T)/GIRK4 (I229L) channels (Fig. 7B) that were almost completely blocked by 1 µmol/L XAF-1407 (Fig. 7D, E).

Fig. 7.

GIRK4 (I229L) subunits form homomeric GIRK channels. A. Representative current trace of ACh (10 µmol/L)-evoked GIRK4 (I229L) currents in a CHO cell (GTP (25 µmol/L)-containing pipette solution). Dotted line indicates zero current level. B. representative membrane recording of GIRK currents in a M₂-R/GIRK4 (I229L)/GIRK4 (S143T)-expressing cell. Application of ACh (10 µmol/L) is indicated by bars. Upward deflections represent membrane currents in response to depolarizing voltage ramps (-120 mV to + 60 mV, duration 500 ms applied every 20 s). C-D. Representative GIRK currents during coapplication of ACh (10 µmol/L) and XAF-1407 (1 µmol/L) from a CHO cell expressing M₂-R and GIRK4 (I229L) subunits (C) or from a cell additionally transfected with a cDNA encoding for GIRK4 (S143T) subunits (D). 0.5 mmol/L GTPγ-S was included in the pipette solution. E. Summarized data of the XAF-1407-induced inhibition of homomeric GIRK4 (I229L) currents or heteromeric GIRK4 (I229L)/GIRK4 (S143T) currents. Single data points are indicated by diamonds. Statistical significance was tested with a Mann-Whitney test. Number of experiments as indicated. P-value < 0.01 was marked by two asterisks

Although XAF-1407 is an established inhibitor of GIR1/4 channels [11], the exact mechanism of its inhibitory action has not been described in the literature. It was beyond the scope of the present study to elucidate the mechanism of XAF-induced GIRK inhibition, but our data suggest that the high PIP₂ affinity of the GIRK4 (I229L) subunit reduces the XAF-sensitivity of the channel mutant.

In a next series of experiments, we aimed to quantify the Phe-induced GIRK inhibition of homomeric GIRK4 (I229L) channels in comparison to homomeric GIRK4 (S143T). During stimulation of α_1B-ARs with 10 µmol/L Phe, GTPγ-S-activated currents through homomeric GIRK4 (S143T) channels were inhibited by about 74% compared to 73% for GIRK4 (I229L) channels (Fig. 8A and B, see also the summarized data in Fig. 8E). However, since 10 µmol/L Phe induced an almost complete inhibition of GIRK4 (I229L) channels, slight impediments of the PKC-induced channel inhibition are hardly to detect. We therefore reduced the phenylephrine concentration to 1 µmol/L and determined the extent of current inhibition of GIRK4 (S143T) and GIRK4 (I229L) channels under these conditions. As shown in the representative current recording in Fig. 8C µmol/L Phe reduced GIRK4 (S143T) channel activity by about 50%, indicating that this Phe concentration is appropriate to induce submaximal current inhibition. The enhanced PIP₂-channel interaction of the I229L mutant should alleviate the PKC-induced channel inhibition, but as shown in Fig. 8D µmol/L phenylephrine also blocked homomeric GIRK4(I229L) channels by about 50%. The summarized data in Fig. 8E demonstrate that the extent of GIRK current reduction by 1 or 10 µmol/L Phe was almost identical in GIRK4 (S143T)- and GIRK4 (I229L)-expressing cells. Interestingly, the application of a saturating ACh concentration often induced a slight initial activation of GIRK4 (I229L) channels, followed by a slow current increase (see Fig. 8B). Although it was beyond the scope of the present study to investigate the time course of GIRK4 (I229L) channel activation, future studies might be performed to delineate the underlying mechanism, e.g. by manipulating the expression levels of putative candidate proteins that transduce GPCR stimulation to GIRK current activation.

Fig. 8.

α_1B-AR-induced inhibition of homomeric GIRK4 (S143T) and GIRK4 (I229L) channels. A. representative membrane recording of a GTPγ-S (0.5 mmol/L)-activated GIRK current during coapplication of ACh (10 µmol/L) and Phe (10 µmol/L) from a CHO cell transfected with M₂-R, GIRK4 (S143T) and α_1B-AR. B. GIRK currents from a CHO cell expressing GIRK4 (I229L) subunits, M₂-R and α_1B-AR. C/D. Representative GTPγ-S-activated GIRK currents during application of ACh (10 µmol/L) and Phe (1 µmol/L) from CHO cells transfected with M₂-R, GIRK4 GIRK4 (S143T) (C) or GIRK4 (I229L) subunits (D) and α_1B-AR. E. Summarized data of Phe-induced GIRK inhibition (mean ± S.E.) of GIRK4 (S143T) and GIRK4 (I229L) channels. Single data points are indicated by diamonds. Significance was evaluated with a Kruskal-Wallis-Anova and Dunn´s post-hoc test. Number of experiments as indicated. P-values < 0.05 are marked by an asterisk. n.s. = not significant

We next compared the Ang II-induced inhibition of homomeric GIRK4 (S143T) and GIRK4 (I229L) channel mutants. To achieve different levels of current inhibition, a submaximal Ang II concentration of 100 nmol/L and a saturating concentration of 1 µmol/L were applied. Figure 9A and B show representative recordings of GTPγ-S-activated GIRK4 (S143) or GIRK4 (I229L) currents during exposure to ACh (10 µmol/L) and Ang II (1 µmol/L). Stimulation of AT₁-receptors with 1 µmol/L Ang II reduced both GIRK4 (S143T) and GIRK4 (I229L) channel activity by about 70% (see also the summarized data in Fig. 9E).

Fig. 9.

High PIP₂ sensitivity of homomeric GIRK4 (I229L) channels does not impede AT₁-R-induced current inhibition. A/B. representative membrane recordings of GTPγ-S (0.5 mmol/L)-activated GIRK currents during coapplication of ACh (10 µmol/L) and Ang II (1 µ nmol/L) from CHO cells transfected with M₂-R, AT₁-R and GIRK4 (S143T) (A) or GIRK4(I229L) (B)

According to the lower angiotensin sensitivity of GIRK4 (S143T) compared to wildtype GIRK1/4 channels, a concentration of 100 nmol/L induced an inhibition of GTPγ-S-activated GIRK4 (S143T) currents of about 40% (Fig. 9C). Application of 100 nmol/L Ang II resulted in an almost equal AT₁-R-induced inhibition of GTP-γ-S activated currents in GIRK4 (I229L)-expressing CHO cells (Fig. 9D and summarized data in E). These experiments demonstrate that pronounced GIRK channel inhibition was evident during stimulation of both α_1B-AR and AT₁-receptors, indicating that the high PIP₂-affinity of mutant GIRK channels does not prevent inhibitory PKC effects per se.

C/D. membrane recordings of GTPγ-S (0.5 mmol/L)-activated GIRK4 (S143T) currents (C) or GIRK4 (I229L) currents (D) during coapplication of ACh (10 µmol/L) and Ang II (100 nmol/L).

E. Summarized data of Ang II-induced GIRK inhibition (mean ± S.E.). Single data points are indicated by diamonds. Significance was evaluated with a One-Way Anova (F-value = 7.46) and Holm-Sidak post-hoc test. Number of experiments as indicated. n.s. = not significant. P-value < 0.01 was marked by two asterisks, P < 0.001 by three asterisks.

Discussion

In the present study, we investigated the functional importance of the serine residue S¹⁹¹ in the GIRK4 subunit for the receptor-specific modulation of GIRK channel activity and identified S¹⁹¹ as a target for the Ang II-activated PKCε. The novel findings of the present study in terms of receptor-specific phosphorylation of S¹⁹¹ and its contribution to GIRK inhibition are summarized in Fig. 10.

Fig. 10.

Scheme of the G_qPCR-specific activation of different PKC isoforms and their targeted residues within the GIRK4 subunit

By using specific pharmacological PKC inhibitors, previous studies from our group identified the receptor-dependent PKC isoforms that contribute to the α_1B-AR- and AT₁-R- induced GIRK inhibition as PKCα and PKCε [16, 29]. In these studies, we provided evidence that the receptor-species dependent activation of different PKC isoforms differentially shaped the cellular response during G_qPCR stimulation with either facilitative or inhibitory effects on GIRK currents. Facilitative effects were attributed to the PKCε-mediated phosphorylation of S⁴¹⁸ whereas the phosphorylation sites that contribute to α_1B-AR- and AT₁-R- induced GIRK inhibition were not identified.

We now demonstrate that the phosphorylation site S¹⁹¹ is differentially addressed by α_1B-ARs and AT₁-Rs, resulting in receptor-species dependent effects of G_qPCR-induced GIRK channel inhibition. Mutation of the GIRK4 phosphorylation site S¹⁹¹ alleviates the inhibitory effect of Ang II on GIRK channel activity in both Ad-GIRK4 (S191A)-infected atrial myocytes (Fig. 2) and in GIRK4 (S191A)-coexpressing CHO cells (Fig. 4), suggesting that this phosphorylation site is targeted at least by the Ang II-activated PKCε isoform. In contrast, GIRK4 (S191A) inhibition was not abrogated during stimulation of α_1B-ARs, indicating that S¹⁹¹ is of minor importance for Phe-induced GIRK inhibition.

The question might arise if mutated GIRK4 (S191A) subunits are effectively incorporated into heteromeric GIRK1/4 channels by substituting wildtype GIRK4 subunits in atrial myocytes and CHO cells. Although mutant GIRK4 (S191A) subunits did not form functional homomeric channels (Fig. 3), the high efficiency of cell transfection or adenoviral infection should result in a substantial amount of heteromeric channels that contain mutant GIRK4 subunits.

Grasser and coworkers demonstrated that the subunit stoichiometry of GIRK1/GIRK4 channels in Xenopus oocytes depends on the ratio and amount of injected RNA encoding for the different subunits [13]. Thus, upon a 25-fold excess of GIRK1 over GIRK4, a “non-conventional” stoichiometry of 3:1 GIRK1/GIRK4 subunits was observed. Injection of equal amounts of RNA resulted in both GIRK1/GIRK4 and homomeric GIRK4 channels [13]. In the present study, we used a ratio of 2:1:1.5 for cDNAs encoding for GIRK1, GIRK4 and GIRK4 (S191A) subunits to avoid a vast express of GIRK1 over the other GIRK subunits.

Wildtype α_1B-receptors inhibit GIRK currents through the PKCα isoform. The Phe-activated PKCα does not phosphorylate S¹⁹¹ but targets other, yet unidentified, residues in the GIRK4 subunit to induce GIRK channel inhibition. Stimulation of AT₁-receptors activates PKCε that induces GIRK current inhibition by targeting at least S¹⁹¹ within the GIRK4 subunit.

Although the whole-cell IK_(ACh) in atrial cells is predominantly carried through GIRK1/4 channels, biochemical experiments on bovine atrial membranes indicated that a substantial fraction of GIRK4 protein forms homotetrameric channel complexes [9, 13]. The contribution of endogenous atrial GIRK4 currents to the whole-cell IK_(ACh) appears to be negligible, but high expression levels of monomeric or concatenated dimeric GIRK4 subunits reduced the probability of heteromeric channel formation and promoted the formation of homomeric GIRK4 channels [2]. The IK_(ACh) of GIRK4-overexpressing atrial myocytes displayed specific characteristics such as reduced inward rectification, loss of desensitization and slow activation kinetics, indicating that currents through functional GIRK4 channel complexes dominate the macroscopic IK_(ACh) [2, 28]. Accordingly, in the present study, the IK_(ACh) in Figs. 1G and 2B shows the characteristic features of GIRK4 monomeric channels (slower activation, loss of desensitization). We have further evidence that adenoviral-induced expression of GIRK4 subunits results in a high number of homomeric GIRK channels (see Online resource Supplemental Fig. 4). We evaluated the amplitudes of basal and ACh-activated currents in non-treated cells, GFP-, GIRK4- and GIRK4 (S191A)-infected atrial myocytes (data correspond to the experiments shown in Fig. 1) and found a significant reduction of GIRK currents in GIRK4-infected cells. The reduced amplitude of GIRK4 currents is in line with a previous study from our group that characterized homomeric GIRK4 channel in rat atrial myocytes [2]. As demonstrated by the original traces in that study, the amplitude of GIRK4 currents is usually smaller than IK_(ACh) in control cells.

Wildtype heteromeric GIRK1/4 channels might still exist in atrial myocytes and in CHO cells cotransfected with GIRK1, GIRK4 and GIRK4 (191 A) subunits, but the reduced PMA sensitivity of mutant GIRK4 subunit-expressing atrial cells (Fig. 1) and the receptor-specific differences in the extent of G_qPCR-induced IK_(ACh) inhibition (Figs. 2I and 4G) indicate that the whole IK_(ACh) is dominated by the properties of mutant GIRK channels.

The coexpression of GIRK4 (S143T) and GIRK4 (S191A) subunits in CHO cells resulted in a significant expression of functional heteromeric GIRK4 (S143T)/GIRK4 (S191A) channels. The incorporation of S191A subunits into GIRK4 (S143T) channels is evident from the receptor-dependent differences in G_qPCR-induced GIRK modulation (Fig. 6). The impediment of the AT₁-R-induced GIRK inhibition supports the notion that the reducing effect of GIRK4 (S191A) is transferred to heteromeric GIRK channels.

It is worth mentioning that a significant part of the G_q-induced current inhibition remains after the elimination of the S¹⁹¹ site. During stimulation with G_qPCR-agonists, both PKC activation and PIP₂ depletion contribute to GIRK inhibition. Thus, the remaining part of current inhibition might be attributed to PIP₂-depletion and/or to the phosphorylation of other residues. In a previous study ofMao and coworkers [25], a double mutation of GIRK1 (S185A)/GIRK4 (S191A) did not completely eliminate the current inhibition by PMA. The authors concluded that other PKC sites are present in the GIRK1/4 channel and suggested serine²²¹ in GIRK1 and serine ²²⁷ in GIRK4 as additional phosphorylation sites. Since the phosphorylation-deficient mutants GIRK1 (S221A) and GIRK4 (S227A) subunits did not express functional channels in Xenopus oocytes, further attempts are required to prove the contribution of these residues to receptor-specific GIRK inhibition.

The phosphorylation site S¹⁹¹ in the GIRK4 subunit (see cartoon in Fig. 1A) is located in the linker region between the transmembrane domain (TMD) and the large cytoplasmic domain (CTD), adjacent to several PIP₂-binding sites [35]. Therefore, it might be suspected that mutations of S¹⁹¹ interfere with the interaction of PIP₂ and GIRK channels, potentially resulting in altered PIP₂-sensitivity of GIRK channels. A previous study demonstrated that PKC-induced and PIP₂-dependent inhibition of GIRK1/4 channels expressed in HEK 293 cells are inversely interdependent [18]. PKC-induced GIRK inhibition is increased at low membrane PIP₂ levels whereas high levels of membrane PIP₂ significantly attenuate the PKC-induced GIRK1/4 channel inhibition. The authors postulated that activation of PKC decreases the PIP₂ affinity of GIRK channels and destabilizes the PIP₂-channel interactions. On the contrary, stabilization of the PIP₂-channel interaction alleviates PKC-induced channel inhibition. It is conceivable that the mutation S191A enhances the PIP₂-channel interaction and impedes the PKC-induced channel inhibition, resulting in reduced sensitivity of GIRK currents towards PMA. However, a study of Adney and coworkers [1] demonstrated that phosphorylation-deficient GIRK2 (S196Q) and phosphomimetic GIRK2 (S196E) channel mutants displayed similar sensitivities to PIP₂ depletion as wildtype GIRK2 channels. Since S¹⁹⁶ in the GIRK2 subunit is the PKC phosphorylation site that corresponds to S¹⁹¹ in the GIRK4 subunit, it can be assumed that mutation of S¹⁹¹ does not affect the channel PIP₂-interactions. It was beyond the scope of the present study to investigate the PIP₂-affinity of the S191A channel, but the receptor-dependent differences in the extent of mutant GIRK channel inhibition argue against a higher PIP₂-sensitivity of GIRK4 (S191A) subunits.

We also investigated the receptor-dependent modulation of GIRK4 (I229L) channels that display a high affinity for PIP₂. In contrast to previous studies on Xenopus oocytes [36, 38], we did not observe reduced PKC-sensitivity (Figs. 8 and 9) of GIRK4 (I229L) channels. Stimulation of α_1B-ARs and AT₁-Rs with submaximal or saturating agonist concentrations induced GIRK current reductions that were not alleviated in GIRK4 (I229L)-expressing cells compared to GIRK4 (S143T)-transfected cells. This discrepancy might originate from cell-type specific differences in the spatial organization of GIRK channel signaling networks. A previous study from Cui and coworkers [10] attributed receptor-dependent GIRK channel modulation in mouse atrial myocytes to different signaling microdomains for PIP₂. Close proximity of Endothelin receptor A (ET_A-R) and GIRK channels in caveolae, combined with a restricted mobility of PIP₂ in this microdomain, ensured a strong inhibition of GIRK channels. In contrast, stimulation of Bradykinin receptors (B₂-R) that are excluded from GIRK channel domains, induced only minor inhibition of atrial GIRK currents. Although the spatial organization of α_1B-ARs or AT₁-Rs and GIRK channels in CHO cells is unknown, the concept that the signaling efficiency of PIP₂ depends on its restricted diffusion may explain the cell-type specific differences of the PKC sensitivity of GIRK4 (I229L) channels. In case that GIRK4 (I229L) channels have limited access to local PIP₂ pools, detection of subtle PIP₂ changes is impeded. Despite the strong interaction of PIP₂ with the GIRK4 (I229L) subunit, pronounced PKC-induced channel inhibition might be attributed to the limited efficiency of PIP₂ signaling.

In addition, specific differences in the experimental conditions of our study and the study of Zhang and coworkers [36, 38] have to be considered. We investigated the modulation of GIRK4 (I229L) channel activity during stimulation of different G_qPCRs that activate preferentially PKCα and PKCε [16] . In contrast, Zhang and coworkers applied PMA that non-specifically activates both classical and novel PKCs by mimicking DAG [4].

In Xenopus oocytes, differences in the PKC sensitivity of wildtype GIRK4 and GIRK4 (I229L) channels became evident after a recording time of about 10 min. PKC-induced channel reduction fully developed within 20–30 min after PMA exposure [38]. In our study, the steady-state level of agonist-dependent GIRK4 (I229L) channel reduction in CHO cells developed within 3 to 5 min (Figs. 8 and 9). Thus, it is conceivable that the receptor-activated PKC isoforms PKCα and PKCε rapidly inhibit GIRK channels by direct phosphorylation of serine/threonine residues within the GIRK4 subunit. The reduced PKC sensitivity of GIRK4 (I229L) channels during long-term experiments [38] might be attributed to temporally delayed effects of the strong PIP₂-channel interaction.

To summarize, we described receptor-specific regulation of the phosphorylation-deficient GIRK4 (S191A) channel mutant that is evident in both native atrial myocytes and CHO cells. The receptor-species dependent effects of G_qPCR-induced GIRK4 (S191A)-inhibition suggest that the phosphorylation site S¹⁹¹ is differentially targeted by α_1B-ARs and AT₁-Rs. However, although our data indicate a receptor-species-dependent targeting of S¹⁹¹ by different PKC isoforms, the contribution of other GIRK4 phosphorylation sites to specific PKC-induced GIRK inhibition remains to be determined.

The G_qPCR-induced inhibition of mutant GIRK4 (I229L) channels with high PIP₂-affinity was not impeded under our experimental conditions, suggesting that receptor-activated PKC isoforms directly phosphorylate residues in the GIRK4 subunit rather than weakening PIP₂ channel interaction.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Participated in research design: Marie-Cecile Kienitz, Leonie Inderwiedenstraße. Conducted experiments and performed data analysis: Leonie Inderwiedenstraße, Marie-Cecile Kienitz. Wrote the manuscript: Marie-Cecile Kienitz.

Funding

Open Access funding enabled and organized by Projekt DEAL. This research was funded by the German Research Foundation (Deutsche Forschungsgemeinschaft (DFG)), project number 441909735.

Data availability

All data supporting the findings of this study are available within the paper and its Online Resource (supplementary information). The origin of cDNAs used in the present study is indicated in the method section. Original electrophysiological data are available from the authors upon reasonable request.

Declarations

Ethical approval

Rats were killed following protocols in accordance with the guidelines of the European Community (86/609/EEC) and approved by the animal welfare officer of the Ruhr-University Bochum.

Consent for publication

All authors confirm that they have read and approved the final version of this manuscript for publication.

Competing interests

The authors declare no competing interests.