Report

Requirement of a Macromolecular Signaling Complex for β Adrenergic Receptor Modulation of the KCNQ1-KCNE1 Potassium Channel

See allHide authors and affiliations

Science  18 Jan 2002:
Vol. 295, Issue 5554, pp. 496-499
DOI: 10.1126/science.1066843

Abstract

Sympathetic nervous system (SNS) regulation of cardiac action potential duration (APD) is mediated by β adrenergic receptor (βAR) activation, which increases the slow outward potassium ion current (I KS). Mutations in two humanI KS channel subunits, hKCNQ1 and hKCNE1, prolong APD and cause inherited cardiac arrhythmias known as LQTS (long QT syndrome). We show that βAR modulation of I KSrequires targeting of adenosine 3′,5′-monophosphate (cAMP)–dependent protein kinase (PKA) and protein phosphatase 1 (PP1) to hKCNQ1 through the targeting protein yotiao. Yotiao binds to hKCNQ1 by a leucine zipper motif, which is disrupted by an LQTS mutation (hKCNQ1-G589D). Identification of the hKCNQ1 macromolecular complex provides a mechanism for SNS modulation of cardiac APD throughI KS.

Cardiac function is regulated by the SNS; stimulation of βARs increases heart rate and contractility and shortens APD (1). βAR-mediated regulation ofI KS is particularly important because patients with KCNQ1 or KCNE1 mutations have a form of LQTS (2, 3) in which fatal cardiac arrhythmias are precipitated by increased SNS activity associated with exercise and startling (4–6).

Because βAR signaling results in PKA activation, we sought to determine whether PKA and a phosphatase were components of an hKCNQ1 macromolecular signaling complex. In immunoprecipitations of hKCNQ1 from cardiac homogenates, we found that PKA catalytic and regulatory (RII) subunits, protein phosphatase 1 (PP1) (but not PP2A), and the PKA- and PP1-targeting protein yotiao [which also targets PKA and PP1 to N-methyl-d-aspartate (NMDA) receptors in the brain (7)] coimmunoprecipitated with hKCNQ1 from human cardiac homogenates (Fig. 1A). Thus, the PKA catalytic and regulatory subunits, PP1, and yotiao are components of the hKCNQ1 macromolecular complex.

Figure 1

Presence of PKA, RII, PP1, and the targeting protein yotiao in the KCNQ1 macromolecular complex. (A) KCNQ1 immunoprecipitated from human cardiac homogenates using anti-KCNQ1 (22). Positive control for immunoblotting was human cardiac homogenate; negative control for immunoprecipitation was anti-KCNQ1 preabsorbed with the antigenic peptide (lane 3). (B) For pull-down assays, either microcystin- or cAMP-Sepharose beads were incubated with cardiac homogenates from wild-type (WT) or TG+ mice (expressing hKCNQ1-hKCNE1) (23). Bound proteins were immunoblotted with anti-hKCNQ1. Specificity of the microcystin pull-down was demonstrated using free microcystin LR. “+control” is anti-hKCNQ1 immunoblot of TG+ cardiac homogenate that recognizes the hKCNQ1-hKCNE1 fusion product of the transgene, which migrates at ∼125 kD. (C) Proteins were immunoprecipitated from homogenates of WT or hKCNQ1-hKCNE1 TG+ mouse hearts with anti-hKCNQ1 and immunoblotted with anti-hKCNQ1, anti-PKA (catalytic subunit), anti-RII (PKA regulatory subunit), anti-PP1 (catalytic subunit of PP1), and anti-yotiao. “+ control” is TG+ cardiac homogenate. (D) Proteins from cardiac homogenates (200 μg) that associated with GST-hKCNQ1 fusion proteins (24) were immunoblotted with anti-yotiao, anti-PKA, and anti-PP1 (25). “+ control” is TG+ cardiac homogenate. Data shown are representative of three or more experiments.

Microcystin Sepharose beads (which bind to PP1) and cAMP Sepharose beads (which bind to the PKA regulatory subunit) specifically associated with hKCNQ1 when incubated with cardiac homogenates from a transgenic mouse (TG+) expressing an hKCNQ1-hKCNE1 fusion protein in the heart. No such association was detected in homogenates from wild-type (WT) mouse hearts, which lack I KS because their heart rate (∼500 beats per minute) requires K+ channels with faster kinetics to control APD (Fig. 1B). PKA catalytic and regulatory (RII) subunits, PP1, and yotiao were also associated with hKCNQ1-hKCNE1 from TG+ mouse hearts (Fig. 1C).

The binding of kinases and phosphatases to ion channels through targeting proteins can be mediated by leucine zipper (LZ) motifs (8). We tested whether a LZ motif in the COOH-terminus of hKCNQ1 (amino acid residues 588 to 616) functioned in targeting PKA and PP1 to hKCNQ1 through yotiao. GlutathioneS-transferase (GST)–KCNQ1 fusion proteins were incubated with cardiac muscle homogenates from human or TG+(hKCNQ1-hKCNE1) mouse hearts. GST–KCNQ1-LZ, containing the LZ motif, coprecipitated PP1, PKA, and yotiao (Fig. 1D). This interaction was specific, as neither GST alone nor GST-KCNQ1 fragments from other intracellular regions without the LZ motif coprecipitated PP1, PKA, or yotiao (Fig. 1D). GST–KCNQ1-LZ did not directly bind recombinant PP1 or PKA catalytic subunits (9), indicating that the binding of PP1 and PKA to hKCNQ1 requires yotiao. Moreover, a mutant peptide (KCNQ1-LZm) containing alanine substitutions (Leu602 → Ala, Ile609 → Ala) that disrupt the LZ motif did not coprecipitate PP1, PKA, or yotiao. Yotiao contains multiple coiled coils that can form LZs but has not been reported to bind to targets via LZs. Kinase and phosphatase binding to ion channels through specific targeting proteins provides a mechanism regulating phosphorylation at the subcellular level (10,11). For example, PP1 is targeted to the type 2 ryanodine receptor–calcium release channel on the cardiac sarcoplasmic reticulum through spinophilin (8) and is now shown to interact through yotiao with hKCNQ1 in the plasma membrane.

Addition of cAMP, in the absence of exogenous PKA, induced phosphorylation of immunoprecipitated KCNQ1 (Fig. 2A). Using an alanine substitution, we identified Ser27 as the unique site of PKA phosphorylation on hKCNQ1 (Fig. 2B). We did not detect AKAP 79 or AKAP 15-18 (other PKA targeting proteins) with KCNQ1. Indeed, our data show that yotiao is the only protein that targets PKA and PP1 to hKCNQ1 complex, because incubation of cardiac homogenates with a peptide containing the KCNQ1 LZ motif (which binds yotiao) prevented coimmunoprecipitation of yotiao, PKA, and PP1 with KCNQ1 (Fig. 2C). The corresponding mutant peptide containing a disrupted LZ (GST–KCNQ1-LZm) did not disrupt complexes of KCNQ1 with yotiao, PKA, or PP1 (Fig. 2C).

Figure 2

Phosphorylation of Ser27 on KCNQ1 by PKA requires an intact LZ motif in the COOH-terminus of KCNQ1. (A) hKCNQ1-hKCNE1 immunoprecipitated from TG+ mouse heart was incubated with cAMP (10 μM) or exogenous PKA (5 units) and [γ-32P]ATP. The immunoprecipitates were size-fractionated, followed by autoradiography. The specificity of the PKA phosphorylation was established using the PKA inhibitor PKI (500 nM). (B) Identification of Ser27 as the PKA phosphorylation site on hKCNQ1 by alanine substitution (S27A). WT hKCNQ1 and hKCNQ1-S27A expressed in CHO cells were immunoprecipitated with anti-hKCNQ1 and incubated with PKA (5 units) and [γ-32P] ATP. The immunoprecipitates were size-fractionated and phosphorylation was detected by autoradiography. Equivalent amounts of hKCNQ1 and hKCNQ1-S27A were present in immunoprecipitates, as demonstrated by immunoblotting (lower panel). Data are representative of three experiments. (C) A fusion peptide containing the COOH-terminal LZ motif of hKCNQ1 (hKCNQ1-LZ) competes off yotiao, PKA, and PP1 from the hKCNQ1-hKCNE1 complex. TG+ cardiac homogenates (200 μg) were incubated with glutathione beads containing hKCNQ1-LZ or hKCNQ1-LZm. Supernatants from these reactions were immunoprecipitated with anti-hKCNQ1 and the precipitates were blotted for yotiao, PKA, and PP1. The “+” control is TG+cardiac homogenate. (D) WT hKCNQ1, mutant channels with disrupted LZ motifs, hKCNQ1-LZm, or hKCNQ1-G589D were coexpressed with yotiao in CHO cells. Equal amounts of WT and mutant hKCNQ1, yotiao, PKA, and PP1 were shown by immunoblotting (first three lanes). Proteins were immunoprecipitated from lysates with anti-KCNQ1 and immunoblotted. Immunoprecipitated hKCNQ1 from CHO cells without yotiao is shown in the last lane. Data shown are representative of three or more experiments.

When hKCNQ1 and yotiao were coexpressed in Chinese hamster ovary (CHO) cells, PKA and PP1 were coimmunoprecipitated with hKCNQ1 (Fig. 2D). Without yotiao, PKA and PP1 did not coimmunoprecipitate with hKCNQ1 (Fig. 2D). Disrupting the hKCNQ1 LZ with alanine substitutions (KCNQ1-LZm) or with the LQTS-associated KCNQ1-G589D mutation (Gly589 is the first “e” position in the COOH-terminal hKCNQ1 LZ motif) prevented binding of yotiao, PKA, or PP1 to the channel (Fig. 2D). Thus, the LZ motif is required for yotiao-mediated targeting of PKA and PP1 to hKCNQ1.

The membrane-permeant cAMP analog 8 Br-cAMP (300 μM) significantly increased hKCNQ1-hKCNE1 tail currents after +60 mV prepulses 1.8 ± 0.1–fold (n = 9) (Fig. 3A) in myocytes from mice expressing the human proteins. cAMP has similar effects on nativeI KS recorded from guinea pig ventricular myocytes (12). hKCNQ1-hKCNE1 channel activity (assayed from tail currents after +60 mV prepulses) was significantly enhanced (3.6 ± 0.7 fold, n = 5) (Fig. 3A) by cAMP (300 μM) plus okadaic acid (OA, 1 μM), a PP1 inhibitor that binds to the phosphatase catalytic site. OA alone did not significantly increase hKCNQ1-hKCNE1 currents (n = 5) (Fig. 3A). Thus, kinase (PKA) activity potently regulates I KS and inhibiting phosphatase (PP1) activity enhances this regulation.

Figure 3

Requirement of Ser27, the KCNQ1 LZ motif, and yotiao for PKA-dependent regulation ofI KS. (A) Whole-cell patch clamp currents recorded from WT and TG+ (hKCNQ1-hKCNE1) murine cardiomyocytes (26). Tail current after +60 mV pulses was significantly increased by cAMP (P < 0.01, Student'st test, pre-cAMP versus cAMP, n = 9) and by cAMP plus OA (P = 0.01, Student'st test, pre-cAMP + OA versus cAMP + OA,n = 5). Currents shown in upper traces were activated by a series of 2-s voltage pulses from –40 mV to +80 mV (20-mV increments, 0.067 Hz) from a –65 mV holding potential and deactivated by returning to –40 mV. hKCNQ1-hKCNE1 channel current is evident as slowly activating outward current during pulses and deactivating current “tails” upon repolarization detectable in TG+(arrow) but not TG cells. Myocytes were treated with cAMP and OA (bottom two traces) and hKCNQ1-hKCNE1 currents were activated by depolarization (+60 mV). Currents recorded before (open circles) and after (solid circles) a 5-min external application of 8-Br-cAMP (300 μM) or 8-Br-cAMP (300 μM) plus OA (1.0 μM). Scale: 20 pA/pF and 1 s. (B) Requirement of yotiao for cAMP and OA dependent modulation of hKCNQ1-hKCNE1 channels in CHO cells. Shown are mean currents (n = 5) (+60 mV pulse, –40 mV return) as well as plots of mean tail current ± SEM vs. pulse voltage. Tail currents (after +60 mV) were significantly different by ANOVA P < 0.05. CHO cells were transfected with hKCNQ1-hKCNE1 without (left) or with (right) yotiao. Internal solutions without cAMP (control); with 200 μM cAMP (cAMP); or with 200 μM cAMP plus 0.2 μM OA (cAMP/OA) were dialyzed at room temperature for 13 min before measurements were made. Open symbols, no yotiao; solid symbols, with yotiao. Squares, control; triangles, cAMP; circles, cAMP/OA. Scale: 100 pA/pF, 1 s. (C) Substituting Ala for Ser27 in KCNQ1.I KS was measured in CHO cells cotransfected with yotiao, hKCNE1, and hKCNQ1-S27A (n = 6). Scale: 100 pA/pF, 1 s. (D) Disruption of the KCNQ1 LZ motif by Ala substitution. CHO cells were transfected with yotiao, hKCNE1, and hKCNQ1-LZm (n = 5). Scale: 100 pA/pF, 1 s. (E) Disruption of the KCNQ1 LZ motif with the LQTS mutation (G589D). CHO cells were transfected with yotiao, hKCNE1, and hKCNQ1-G589D (n = 5). Scale: 50 pA/pF, 1 s. In (C) to (E), cells were dialyzed with cAMP/OA-containing (solid squares) or control (open squares) internal solutions. Currents and plots as in (B).

Expression of yotiao was required to reconstitute cAMP-dependent regulation of the hKCNQ1-hKCNE1 channel in a heterologous expression system (CHO cells). In the absence of yotiao, there was no significant effect of intracellular cAMP (200 μM) on hKCNQ1-hKCNE1 currents either in the absence or presence of OA (0.2 μM n = 5) (Fig. 3B). With yotiao, hKCNQ1-hKCNE1 tail current (after +60 mV conditioning pulses) was doubled by cAMP [without cAMP, 24.3 ± 7.5 pA/pF; with cAMP (200 μM), 50.7 ± 9.2 pA/pF;n = 5] and OA (0.2 μM) plus cAMP (200 μM) increased the hKCNQ1-hKCNE1 current ∼4-fold (89 ± 15.6 pA/pF,n = 5) (Fig. 3B).

Substituting Ala for Ser27 (S27A) in KCNQ1, which is phosphorylated in response to cAMP, eliminated cAMP-dependent enhancement of hKCNQ1-hKCNE1 current (n= 6, Fig. 3C). Expression in CHO cells of hKCNE1 and yotiao with mutant hKCNQ1 (hKCNQ1-LZm) or the G589D mutant linked to LQTS (13), both of which disrupt the KCNQ1 LZ motif (Fig. 2D), ablated cAMP-dependent regulation of I KS(n = 5) (Fig. 3, D and E). Thus, assembly of the macromolecular complex via LZ-mediated binding of yotiao to hKCNQ1, as well as PKA phosphorylation of Ser27, are necessary to reconstitute PKA- and PP1-dependent regulation of hKCNQ1-hKCNE1 channels.

The finding that the hKCNQ1-G589D mutation prevents cAMP-dependent regulation of I KS suggests that these mutant channels may not respond to βAR-mediated signaling in patients. Along with the reduced expression ofI KS in carriers of the hKCNQ1-G589D mutation (13), uncoupling of the channel from SNS modulation may exacerbate the defect in APD shortening and further increase the risk of exercise-induced ventricular tachyarrhythmias that cause sudden cardiac death (SCD) in some patients (6). Indeed, SCD is associated with SNS stimulation in 81% of individuals with the KCNQ1-G589D mutation in whom a trigger event can be identified (13), in contrast to other K+ channel mutations in which only ∼10% of SCD cases are triggered by SNS activation (4).

Taken together, our data show that the regulation of hKCNQ1 by PKA-dependent phosphorylation requires a macromolecular complex that includes PKA, PP1, and the targeting protein yotiao and provide a mechanistic link between the sympathetic nervous system and modulation of the cardiac APD through I KS.

  • * These authors contributed equally to this report.

  • To whom correspondence should be addressed. E-mail: rsk20{at}columbia.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article