PKD Phosphorylation as Novel Pathway of K_V11.1 Regulation

The potassium channel K_V11.1 is encoded by KCNH2 and also known as human-ether-a-go-go related gene type 1 (hERG). The tetrameric channel protein is expressed in a variety of tissues including brain, heart, pancreas, vasculature, and a number of tumors [1]. It is best known for its role in the heart, where it is the molecular correlate of the rapid delayed rectifying current I_Kr [2, 3]. Mutations in the KCNH2 gene are associated with several cardiac arrhythmias such as Long QT and Short QT syndromes and atrial fibrillation [1, 2, 4].

Several posttranscriptional and posttranslational mechanisms have been described to finetune expression levels and/or biophysical parameters of K_V11.1 [5-7]. Protein kinase A (PKA) phosphorylates K_V11.1 directly at four consensus sites in vitro [8, 9]. Protein kinase C (PKC) effects have been investigated in different expression systems and experimental models (reviewed in [1]). Despite numerous studies, it is still unresolved whether PKC directly phosphorylates the channel protein. Identification of 18 consensus sites by prediction algorithms and individual ablation of these sites did not provide any evidence for a direct phosphorylation at these sites [10, 11]. Cockerill et al. provided evidence for a role of the channel’s N-terminus for PKC mediated phosphorylation [11]. It has been suggested that an atypical PKC phosphorylation site might be involved in the PKC effects [1, 11].

Noteworthy, multiple lines of evidence indicate that PKC activity is indispensable for activation of protein kinase D (PKD). Especially, the novel PKCs Ɛ and η give robust activation of PKD [12] but also classical PKCα can activate PKD [13]. PKD is expressed in a variety of tissues and its function has been implicated in health and disease (reviewed in [14-16]). Several studies found that a variety of stimuli induce PKD activation through the α-adrenergic pathway, and in many cases this could be attenuated by PKC inhibition [17-19]. Here, we hypothesized that PKC effects on K_V11.1 are mediated through PKD1.

Preparation of brain samples. All procedures involving animals conformed to Danish animal welfare regulations. Briefly, mouse or rat brain was cut in smaller pieces, homogenized in buffer [in mM: 50 Tris/HCl pH 7.4, 2.5 EGTA, 5 EDTA, 10 NaCl, 10 KCl, 0.5 pefabloc, 0.1 NaF, and 320 sucrose with a protease inhibitor cocktail (in µg/mL: 2 aprotinin, 1 leupeptin, 2 antipain, 10 benzamidine, and one tablet of phosSTOP/10 mL)] and centrifuged. The supernatant was retained, centrifuged again, and pellets were resuspendend in 2.5 mL homogenizing buffer for every gram of brain.

Preparation of HEK293 cell samples. HEK293 cells were transfected with K_V11.1 alone or co-expressed with PKD1. For mass spectrometry the cells were grown in a special media, immunoprecipitated and digested to peptides. For lysates cells were harvested in solubilization buffer without detergent [in mM: 50 Tris pH 8.5, 150 NaCl, 5 EDTA] and supplemented with 8 µM leupeptin, 0.4 mM pefabloc and one tablet of phosSTOP pr. 10 mL, centrifuged (2 min, 100 x g) and incubated with solubilization buffer containing 1% NP-40 on a spinning wheel for 3 hours. Samples were centrifuged at > 10.000 x g for 10 min and supernatants were aliquoted, frozen in liquid nitrogen and stored until further use.

Sample preparation for mass spectrometry. Sample lysates were precleared with Protein G-dynabeads (Invitrogen). Then primary antibody (2 μg/1 mg total protein in the sample) was added followed by overnight incubation. After addition of dynabeads in lysis buffer [in mM: 50 Tris/HCl, pH 8.5, 5 EDTA, 150 NaCl] containing 1% NP-40, 8 µM leupeptin (Sigma), 0.4 mM pefabloc (Sigma), one tablet of phosSTOP pr. 10 mL (Roche) and incubation for 2 h, multiple wash steps were performed. At the end beads were resuspended in Laemmli buffer (containing 100 mM DTT) and incubated for 3 min at 70°C.

Mass spectrometry. Immunoprecipitated proteins were separated by SDS-PAGE, fixed, stained, bands excised and proteins proteolytically cleaved to peptides as previously described [20, 21]. Peptide separation was performed by reversed-phase C₁₈ HPLC on an Easy nLC system (Thermo Fisher Scientific). The effluent from the HPLC was directly electrosprayed into an LTQ Orbitrap Velos mass spectrometer (Thermo Fisher Scientific, Slangerup, Denmark) through a nano-spray ion source. The peptide mixture was analyzed by full-scan MS spectra (m/z 300-2000, resolution 30, 000) in the Orbitrap analyzer. For every full-scan the most intense peptide ions were sequentially isolated, fragmented by higher energy collisional dissociation (HCD), and fragments were recorded by the Orbitrap mass analyzer. The acquired data was processed by MaxQuant (version 1.1.1.25) (Max-Planck Institute of Biochemistry, Department of Proteomics and Signal Transduction, Munich, Germany).

Recordings. All recordings of transiently transfected HEK293 cells were performed in the whole-cell patch clamp configuration using an EPC9 patch clamp amplifier (HEKA Elektronik, Lambrecht, Germany). Data acquisition was done with the Pulse software (HEKA Electronik). The series resistance recorded in the whole cell configuration was kept below 10 MΩ and compensated (80%). The extracellular solution contained [in mM: 140 NaCl, 4 KCl, 2 CaCl₂, 1 MgCl₂, and 10 HEPES, pH=7.4]. The intracellular solution contained [in mM: 110 KCl, 10 EGTA, 5.17 CaCl₂, 1.42 MgCl₂, 4 K₂ATP, and 10 HEPES, pH=7.2]

Protocols. Current–voltage (I/V) relations were obtained from the step-protocol by plotting the outward current at the end of the 1-second test pulse as a function of the test potential. The voltage-dependence of activation was determined by fitting the amplitude of peak tail currents versus test potential to a two-state Boltzmann distribution of the form I(V) = 1/(1+exp [(V_½ - V)/a]), where V_½ = potential for half-maximal activation, and a = slope factor. Other kinetic parametres were analysed as described in detail (for all online suppl. material, see www.karger.com/doi/10.1159/000491007) in the Supplementary Methods.

Data analysis. Data analysis was performed using Excel (Microsoft), Igor Pro 4.04 (Wavemetrics, OR, USA) and Prism 4 (GraphPad Software, CA, USA). All values are given as mean ± SEM. The number of independent experiments is indicated by n. Statistical comparisons of results were performed using one-way analysis of variance (ANOVA) or two-way ANOVA. Following ANOVA, Bonferroni’s method of multiple comparisons was used to identify significant differences between means. A value of P < 0.05 was considered statistically significant.

Primary antibodies: anti-K_V11.1, APC-062 (Alomone, Jerusalem, Israel, 1: 500); anti-PKD1, H00005587-A01 (Abnova, Roskilde, Denmark, 1: 500); anti-β-tubulin, MAB3408 (Millipore, Hellerup, Denmark, 1: 4000).

Secondary antibodies: For western blot analyses; peroxidase conjugated affinipure F(ab)fragment donkey anti-rabbit and anti-mouse (Jackson ImmunoResearch, Suffolk, UK), donkey anti-mouse 680RD, 926-68072 (LI-COR Biosciences, Copenhagen, Denmark); goat anti-rabbit 800CW, 926-32211 (LI-COR Biosciences). For confocal imaging; Alexa Fluor®488-conjugated donkey anti-rabbit IgG (1: 200), Alexa Fluor®568-conjugated donkey anti-mouse IgG (1: 200). Alexa®Fluor 647 Phalloidin (1: 200) was used to stain actin filaments and DAPI (1: 300) was used to stain the nucleus. Secondary antibodies were purchased from Invitrogen (Glostrup, Denmark).

Point mutations in K_V11.1 (GenBank Acc. No. NM_000238) were introduced by standard site-directed mutagenesis. All constructs were verified by complete DNA sequencing of the cDNA insert (Macrogen Inc., Seoul, Rep. of Korea). Human protein kinase D1 (PKD1, NM_002742) expression plasmid was kindly provided by Dr. A. Hausser (University of Stuttgart, Germany) [22].

First, we mapped the in vivo phosphorylation sites of K_V11.1 using tissue samples, where K_V11.1 channels from both mouse and rat brain tissue were purified by immunoprecipitations and submitted to analysis by LC-MS/MS mass spectrometry. Our analysis led to identification of 14 in-vivo phosphorylation sites (Table 1). In Table 1 we assigned the identified residues from mouse and rat to the nomenclature of the human K_V11.1 residues by alignment of mouse, rat, and human sequences (see online suppl. material, Suppl. Fig. S1). Our data confirmed phosphorylation at residue S283 which had been reported as a site for PKA-mediated phosphorylation previously [9]. None of the identified residues were PKC consensus sites; however, S284 was a strong candidate for PKD phosphorylation.

We assessed a possible role of PKD1 by employing whole-cell patch clamp recordings. We transiently transfected HEK293 cells with K_V11.1 alone or co-expressed with PKD1 cDNA (Fig. 1). To investigate the potassium currents, we applied a standard step protocol with 1 s test pulses ranging from -100 to +60 mV in 10 mV increments. Fig. 1A shows representative current recordings of K_V11.1 and K_V11.1 + PKD1. As expected, the current/voltage relationship (IV curve) resulted in the characteristic bell-shape due to inactivation of the channel [3]. Of note, a high degree of cell to cell variability of current levels was seen in line with literature [23]. The maximum outward current amplitude was normalized to the cell capacitance and used to construct the current-voltage (I-V) relationship of K_V11.1 (n=7) and K_V11.1 co-expressed with PKD1 (n=7) (Fig. 1B). The maximum outward current amplitude at +20 mV for K_V11.1 was 29.5±9.5 pA/pF and significantly reduced by 56.2 % upon co-expression with PKD1 (12.9±1.7 pA/pF, P=0.018). The tail current amplitude was also reduced by PKD1 co-expression (Fig. 1C). K_V11.1 current reduction has earlier been assigned to a shift of the voltage-dependent activation to more depolarized potentials [24]. Therefore, the normalized peak tail currents were fitted to a Boltzmann function to determine the V_½ values of voltage-dependent activation. No difference of the voltage-dependent activation upon co-expression of PKD1 (Fig. 1D, Table 2) was observed. Deactivation, inactivation and recovery from inactivation properties (Table 2) were unchanged.

We assessed whether PKD1 co-expression would affect K_V11.1 protein expression or subcellular localization. Western blot experiments revealed that there was no change in protein levels (see online suppl. material, Suppl. Fig. S2). PKD1 did not alter the localization of K_V11.1 in HEK293 cells, and K_V11.1 and PKD1 were both localized in the plasma membrane as assessed by confocal microscopy (see online suppl. material, Suppl. Fig. S3). Attempts to test a possible direct interaction between K_V11.1 channel and PKD1 by co-immunoprecipitation remained unsuccessful (data not shown).

To analyse whether S284 and other residues would be phosphorylated by PKD1, we immunoprecipitated K_V11.1 channels co-expressed with PKD1 in HEK293 cells to map the phosphorylation sites by mass spectrometry. We identified seven residues that were phosphorylated upon co-expression with PKD1. This indicated that PKD1 could affect K_V11.1 phosphorylation in a heterologous expression system by direct phosphorylation of several residues. MS/MS spectra for all identified sites with evidence for site localization are provided (see online suppl. material) in Suppl. Fig. S4-9. A few residues did stand out from the data; Residue S284 appeared to be phosphorylated in HEK293 cells co-expressed with PKD1. In addition, it was identified in vivo and motif-wise PKD1 was the most likely kinase to phosphorylate this site, making S284 our prime candidate for PKD1 phosphorylation. Also amino acids S263 and S266 were phosphorylated. S266 was identified to be phosphorylated by itself, whereas S263 was only identified in peptides where S266 was also phosphorylated and we decided to address their combined effect in the subsequent functional analysis. S322 was phosphorylated and found in a cluster together with S318, T319, and S320. Three of the four were found phosphorylated in vivo. We chose to address the combined effects of the four residues in the cluster. Finally, residue T353 appeared to be phosphorylated in HEK293 cells and in vivo.

We sought to address the functional impact of the selected amino acids after ablation of the serine residues to alanine (A) hereby mimicking the dephosphorylated state. To this end, we generated four K_V11.1 mutants (K_V11.1-S284A, K_V11.1-S263/S266A, K_V11.1-S318/T319/S320/S322A, and K_V11.1-T353A). We expressed the mutant channels in HEK293 and analysed biophysical parameters by patch-clamp electrophysiology.

K_V11.1-S284A displayed a significantly higher maximum outward current amplitude than wild-type K_V11.1 at +20 mV; S284A: 58.2±6.9 (n=11), WT: 29.9±3.5 pA/ pF (n=8, P< 0.001) respectively; (Fig. 2A,B) indicating that this residue mediates at least some of the PKD1 phosphorylation effect on the channel. Also the tail current amplitude was significantly higher for the mutant (Fig. 2C). We did not observe any significant shift in the voltage-dependent activation for K_V11.1-S284A compared with wild-type (Fig. 2D, Table 3). The slow time constant of deactivation was significantly faster at -60 and -50 mV for the S284A mutant compared to wild-type. Other voltage-dependent and kinetic parameters were unaffected by the alanine mutation (Table 3). We analysed protein levels of S284A by western blotting experiments and found that expression levels were similar to wild-type channels (see online suppl. material, Suppl. Fig. S2). K_V11.1-S263/S266A, K_V11.1-S318/T319/ S320/S322A, and K_V11.1-T353A behaved similarly to wild-type channels (see online suppl. material, Suppl. Table S1).

K_V11.1 is an important potassium channel which is manifested in its association with inherited and acquired cardiac arrhythmia where even mild changes of the conducted current may have deleterious effects. The current conducted by K_V11.1 is regulated by a number of physiological stimuli mediated by G-protein coupled receptors, e.g. α- and β-adrenergic, muscarinic, and angiotensin II receptors [11, 25-29]. Hence, understanding the molecular basis of the regulation is crucial for elucidating disease mechanisms [1, 30]. Here, we provide evidence for involvement of protein kinase D based on in vivo and in vitro mass-spectrometry findings together with functional analysis of wild-type and mutant K_V11.1 channels.

LC-MS mass spectrometry revealed several phosphorylated residues in vivo and suggested PKD as a strong candidate. Co-expression of K_V11.1 with PKD1 reduced current amplitudes; however, we were unable to identify significant changes in kinetic parameters that could explain the decrease in I_max. Protein levels were unaffected. Due to the lack of specific pharmacological tools, i.e. specific activators or blockers of PKD1 which would allow for an assessment of PKD1 effects independent of PKC in patch-clamp recordings [14], we instead chose four candidate K_V11.1 mutants for further functional analysis based on the results obtained from HEK293 cell experiments and in vivo data. Ablation of the putative PKD phosphorylation site in the mutant S284A increased the maximal current.

Our data might help mitigating a long-standing controversy in the field regarding PKC regulation of the channel. Most studies have investigated the acute effects of different pathways upstream of PKC activation with variable results ranging from reports indicating that neither PKC nor PKA are involved [31], that PKA and PKC show composite and partial effects with complex cross-talk [32], or that either convential PKC (cPKC) [10] or novel PKC (nPKC) [25] are exclusively involved. Other pivotal points of discussion have been whether PKC phosphorylates the channel directly [10, 11, 27] and whether PKC mediated phosphorylation increases [32] or decreases current K_V11.1 (or I_Kr) amplitudes [10, 11, 24, 26, 27]. However, the majority of studies found a positive shift in the voltage-dependence of activation and an acceleration of deactivation, with different levels of reduction of the K_V11.1 current amplitude (e.g [10, 11, 27, 33].)

Our results support the notion that PKC mediated functional effects are independent of the previously suggested consensus sites [10]. They also confirm a role of the N-terminal part of the protein as proposed by other studies [11, 34]. Cockerill and coworkers found that residues 2-354 of the K_V11.1 N-terminus are critical for both cPKC phosphorylation and functional modulation of the channel [11]. Of note, both cPKC [13] and nPKC [12] have been reported to activate PKD1. While we cannot rule out that PKC exerts a variety of direct and indirect effects, we propose that PKD1 is a new player, either activated independently or downstream of PKC, targeting S284 in the N-terminus of the channel.

Recent attempts to resolve the identity of the involved PKC isoform by using inhibitory peptidomimmetics indicated a predominant role of PKCε, with minor contributions of PKCβ, PKCα, and PKCη [35]. Interestingly, Haworth and coworkers reported that in adult cardiomyocytes PKCε is the predominant activator of PKD and this activation was somewhat counteracted by PKA activation indicating a crosstalk between the two kinase pathways [18]. This crosstalk might mainly be a PKC effect on the cAMP level [36, 37], but the molecular mechanism behind this is currently unclear. Therefore, we did not investigate the candidates that are already known to be phosphorylated by PKA.

Our data indicate that PKD1 contributes to K_V11.1 regulation by targeting S284 in the N-terminus of the channel. This pathway could therefore present an explanation to years of divergent results regarding the regulation of I_Kr or K_V11.1 by protein kinases.

This work was supported by the Danish National Research Foundation.

The authors declare to have no conflict of interests.

A. Lundby and N. Schmitt contributed equally to this work.