Research ResourceBiochemistry

In Vivo Phosphoproteomics Analysis Reveals the Cardiac Targets of β-Adrenergic Receptor Signaling

See allHide authors and affiliations

Science Signaling  04 Jun 2013:
Vol. 6, Issue 278, pp. rs11
DOI: 10.1126/scisignal.2003506

Abstract

β-Blockers are widely used to prevent cardiac arrhythmias and to treat hypertension by inhibiting β-adrenergic receptors (βARs) and thus decreasing contractility and heart rate. βARs initiate phosphorylation-dependent signaling cascades, but only a small number of the target proteins are known. We used quantitative in vivo phosphoproteomics to identify 670 site-specific phosphorylation changes in murine hearts in response to acute treatment with specific βAR agonists. The residues adjacent to the regulated phosphorylation sites exhibited a sequence-specific preference (R-X-X-pS/T), and integrative analysis of sequence motifs and interaction networks suggested that the kinases AMPK (adenosine 5′-monophosphate–activated protein kinase), Akt, and mTOR (mammalian target of rapamycin) mediate βAR signaling, in addition to the well-established pathways mediated by PKA (cyclic adenosine monophosphate–dependent protein kinase) and CaMKII (calcium/calmodulin-dependent protein kinase type II). We found specific regulation of phosphorylation sites on six ion channels and transporters that mediate increased ion fluxes at higher heart rates, and we showed that phosphorylation of one of these, Ser92 of the potassium channel KV7.1, increased current amplitude. Our data set represents a quantitative analysis of phosphorylated proteins regulated in vivo upon stimulation of seven-transmembrane receptors, and our findings reveal previously unknown phosphorylation sites that regulate myocardial contractility, suggesting new potential targets for the treatment of heart disease and hypertension.

Introduction

Stimulation of β-adrenergic receptors (βARs) is an essential component of the fight-or-flight response in human physiology, which results in enhanced cardiac output mediated by increased contractile force and heart rate (1). βARs are a class of seven-transmembrane receptors, and they bind to catecholamines, such as adrenaline (epinephrine), and activate adenylate cyclases, resulting in the increased synthesis of cyclic adenosine monophosphate (cAMP), which stimulates cAMP-activated protein kinases. This signaling pathway eventually leads to the increased contractility of cardiomyocytes through the phosphorylation of key excitation-contraction coupling proteins; however, most of the phosphorylation targets downstream of βARs are unknown. Tight regulation of the βAR response is critical because sustained activation of βARs promotes pathologic cardiac remodeling (2). Moreover, chronic therapy with βAR inhibitors in patients with hypertension or heart failure prevents the progression of cardiac dysfunction (3). Further insights into βAR signaling pathways are thus of great importance to understand how heart diseases, such as heart failure, develop and how to optimize clinical treatment.

Quantitative phosphoproteomics is a powerful technology to investigate signaling pathways in cellular systems (4), and recent advances have enabled the mapping of phosphorylation sites in tissue samples (57). Here, we combined approaches to perform quantitative phosphoproteomics in vivo, and we identified specific phosphorylation sites (phosphosites) on proteins from murine hearts that were regulated in response to the stimulation of βARs. There are two types of βARs in the human heart: β1AR, which constitutes ~80% of βARs, and β2AR, which constitutes the remaining ~20% (8). Because both receptors couple to distinct intracellular signaling pathways (9), we investigated cardiac phosphorylation changes in response to the activation of each individual pathway. Our approach enabled the identification of hundreds of phosphorylation sites that are regulated by β1AR activation.

Results

Regulation of phosphosites in response to βAR stimulation

We targeted the βARs pharmacologically in three groups of mice (Fig. 1A). Mice in the control group were treated with β1AR- and β2AR-specific antagonists to inhibit endogenous βAR activity. To understand β1AR and β2AR signaling pathways individually, we used combinations of specific agonists and antagonists to activate one receptor while inhibiting the other. We prioritized compounds with high specificity over compounds with large potency in our study design to circumvent potential cross-activation. We applied the agonist intravenously during electrocardiogram recordings 10 min after the antagonist was administered. The hearts were excised and snap-frozen 10 min after agonist application when the elevated heart rates had stabilized (fig. S1). Next, we extracted and digested cardiac proteins, enriched for phosphopeptides by titanium dioxide (TiO2) chromatography (10, 11), and analyzed the samples by nanoflow liquid chromatography–tandem mass spectrometry (LC-MS/MS). We identified a total of 4246 proteins with 8518 phosphosites that could be mapped to a specific residue. Evaluation of the acquired data (fig. S2); lists of all identified proteins, phosphopeptides, and modified specific peptides (tables S1 to S3); and all of the raw mass spectrometric files and annotated phosphopeptide MS/MS spectra are presented in the Supplementary Materials.

Fig. 1 βAR-mediated regulation of cardiac protein phosphorylation.

(A) βAR antagonists or agonists were administered to three groups, each consisting of three mice at points 1 (t = 0 min) and 2 (t = 10 min). Cardiac proteins were extracted from triplicate samples for each heart and then subjected to phosphopeptide enrichment and LC-MS/MS analysis. The total number of identified proteins and phosphosites localized to a specific residue and the number of phosphoproteins and phosphopeptides regulated upon stimulation of β1AR are summarized. (B) P values from statistical analysis (by t test) of phosphopeptide intensities (β1AR-stimulated mice versus control mice) were plotted as a function of the ratios of the peptide intensities. The hyperbolic curve represents an FDR of 0.01 and separates phosphopeptides that were statistically significantly different between the two groups (yellow) from those that were not (blue). A few well-characterized phosphosites are highlighted.

We normalized the peptide MS signal intensities across experiments to enable us to compare the three groups: control mice, β1AR-stimulated mice (the β1 group), and β2AR-stimulated mice (the β2 group). Each group was highly reproducible between biological replicates (correlation: average R2 = 0.93, fig. S3). We did not expect changes in protein abundance in response to acute βAR stimulation; consistent with this, we did not detect differences in nonphosphorylated peptide intensities among the three groups (fig. S4, A and B, permutation-based t test). To determine differences in the extent of protein phosphorylation among the groups, we evaluated phosphopeptide signal intensities from the β1AR and β2AR groups relative to the control group with permutation-based t test statistics (12). We identified 523 phosphopeptides from 353 different proteins that were statistically significantly regulated in mice upon stimulation of β1AR [Fig. 1B, permutation-based false discovery rate (FDR) < 0.01]. The phosphopeptides encompass a total of 670 unique phosphorylation sites (table S4), thereby increasing the number of known targets of β1AR stimulation from tens to hundreds. For mice stimulated with a β2AR agonist, none of the identified phosphopeptides were statistically significantly different from those of the control mice (fig. S4C, permutation-based t test statistics), which is likely because stimulation of β2AR elicited a weaker cardiac response than did stimulation of β1AR, which was reflected in the smaller enhancement of heart rate. However, comparison between hearts stimulated through β1AR and β2AR suggested cross talk between both pathways because phosphopeptides regulated upon β1AR stimulation had substantially higher intensity ratios than did all other identified phosphopeptides when comparing β2AR-stimulated and control hearts (Fig. 2, Wilcoxon rank-sum test). Our data set of 670 regulated phosphosites in response to β1AR activation is the first global determination of targets downstream of β1AR stimulation in the heart.

Fig. 2 Evaluation of phosphopeptides identified from the β2AR-stimulated hearts focusing on peptides regulated in the β1AR-stimulated mice.

All of the phosphopeptides identified in the mice stimulated with the β2AR agonist were sorted into two groups, depending on whether they were identified as being regulated in the β1AR-stimulated mice. For each of the two groups of phosphopeptides, the ratio of their intensities between the β2AR-stimulated mice and the control mice was calculated. The box plots represent the log2-transformed intensity ratios for the phosphopeptides that were not identified as being regulated by β1AR stimulation (left, n = 7166 phosphopeptides) and the ratios for phosphopeptides that were identified as being regulated by β1AR stimulation (right, n = 387 phosphopeptides). Evaluation of the intensity ratios represented in both box plots showed that they are statistically significantly different (Wilcoxon rank-sum test, P < 0.001). Thus, the phosphopeptides regulated upon β1AR stimulation have statistically significantly higher intensity ratios than other phosphopeptides when comparing β2AR-stimulated hearts and control hearts. This indicates that at least a subset of the phosphosites regulated in response to β1AR stimulation is also regulated in response to β2AR stimulation.

Kinases activated in the β1AR response

To address whether the phosphorylated proteins (phosphoproteins) regulated in response to β1AR stimulation fell into clusters of interconnected proteins, we generated protein-protein interaction networks based on the InWeb database (13) with our identified phosphoproteins as seeds. We evaluated direct and indirect connections with DAPPLE software (14), which compares the generated network to matched, randomized networks. There was statistically significant interconnectivity among the regulated phosphoproteins (P < 0.001 for the direct connections, within-degree node-label permutation test). Pathway enrichment analysis (15) showed that the proteins phosphorylated upon β1AR stimulation are involved in physiologically important pathways, such as striated muscle contraction, hypertrophic cardiomyopathy, and tropomyosin and actin binding. This analysis also highlights serine and threonine kinase activity, particularly emphasizing mammalian target of rapamycin (mTOR)–, Akt-, and p38 mitogen-activated protein kinase (MAPK)–mediated signaling (Fig. 3A), which is suggestive of regulatory roles for many of the identified phosphosites.

Fig. 3 Characteristics of phosphoproteins and phosphosites regulated by β1AR stimulation.

(A) Overrepresented pathways (orange) and gene ontology (GO) terms (purple) for phosphoproteins regulated in the β1AR-stimulated mice. For each pathway and GO term, the fraction of entities covered by our data set is indicated in parentheses, and the number of regulated phosphoproteins is listed next to the bar. (B) Sequence motifs revealed by analyzing amino acid residues flanking phosphosites regulated in the β1AR-stimulated hearts against residues flanking phosphosites that are not regulated with iceLogo (iceLogo static reference method, t test: P < 0.01).

To further investigate the kinases mediating the β1AR phosphorylation response, we used our large data set of regulated phosphosites to determine whether the amino acid residues flanking the phosphosites regulated by β1AR stimulation conformed to a sequence motif that differed from those of other phosphosites. Analysis of the regulated sites indeed revealed a sequence motif: they exhibited preference for arginine in the −3 position, R-X-X-pS/T, as well as an underrepresentation of proline in the +1 position (Fig. 3B). For the same phosphosites, we used the integrative computational approach NetworKIN (16) to predict kinase-substrate relations based on consensus sequence motifs and physical protein-protein interactions. Comparison of the score distributions for kinase predictions for the set of phosphosites regulated upon β1AR stimulation versus those of all other phosphosites highlighted 22 kinases (P < 0.05, Kolmogorov-Smirnov test; Table 1). Nineteen kinases were associated with greater activity in the β1AR-stimulated mice compared to the control mice; these include the two known kinases in the cardiac βAR response, cAMP-dependent protein kinase (PKA) and CamKIID, as well as the kinases AMP-activated protein kinase (AMPK), Akt, and Rho-activated kinase 1 (ROCK1), which explains the overrepresentation of arginines at position −3 of the regulated phosphosites. The kinases glycogen synthase kinase-3α (GSK-3α), cyclin-dependent kinase 2 (CDK2), and MAPK1 had greater activity in the control mice than in the β1AR-stimulated mice, which is consistent with the underrepresentation of proline at position +1 of the regulated phosphosites.

Table 1 Kinases predicted to be regulated by β1AR stimulation based on the identified phosphosites.

The table summarizes our NetworKIN analysis of all identified phosphosites. Kinases with statistically significantly different distribution scores in the data set of regulated versus nonregulated phosphosites from the comparison of β1AR-stimulated mice and control mice are listed (P < 0.05, two-sided Kolmogorov-Smirnov test). The name of the kinase is provided as is the two-sided Kolmogorov-Smirnov P value. Using one-sided Kolmogorov-Smirnov tests, we evaluated whether the kinase was more active in the β1AR-stimulated mice or in the control mice, and the source of the score is also provided.

View this table:

Phosphorylation of the activation loop of a kinase is a proxy for its activity (17). These phosphosites therefore represent a complementary approach to evaluate kinase activities compared to the substrate-oriented strategy described earlier. On the basis of the direct measurement of 642 phosphorylation sites on 144 kinases (table S5), we found that β1AR stimulation activated Akt3, MAPK3, p38α, and mTOR, whereas it inactivated GSK3A. These data consolidate the findings from the statistical approach that focused on kinase substrates. The increased activity of Akt and the reduced activity of GSK3 upon β1AR stimulation were further confirmed by Western blotting analysis (fig. S5). In conclusion, the increased activities of PKA, CamKIID, AMPKs, Akts, mTOR, p38α, ROCK1, and MAPK3, as well as the reduced activity of GSK3, are key contributors to the β1AR-dependent phosphorylation response in the heart (Fig. 4).

Fig. 4 Cardiac proteins with phosphosites regulated upon stimulation of β1AR.

Blockade of the β2ARs with ICI-118-551 followed by stimulation of the β1ARs by dobutamine resulted in activation of cAMP-activated kinases, such as CamK2D, AMPK2, PKA, and mTOR, as well as ERK1 and other MAPKs. These kinases in turn regulate the phosphorylation of proteins involved in controlling cellular excitability and contractility. Cellular depolarization is initiated by increased Na+ influx through the NaV1.5 channel, which is followed and potentiated by Ca2+ influx through the CaV1.2 channel, which in turn activates the RyR2 ryanodine receptors, resulting in excitation-contraction coupling. Contraction is controlled by proteins of the contractile filament, for example, troponin T (TNNT1 and TNNT2), myosin-binding protein C (MYBPC), and troponin I (TNNI3). Repolarization is mediated by K+ efflux, partly through the KV7.1 channel, and removal of Ca2+ and Na+ from the cytosol is controlled by the ion transport regulator proteins PLN and PLM. Phosphosites increased or decreased in intensity are indicated by yellow “P” and green “dP,” respectively. Except for ion channels, all proteins are represented by their gene names. Kinases are in yellow, and channels and transporters are in blue.

The kinase mTOR is a well-known regulator of cell growth (18), and mTOR activation increases protein synthesis by phosphorylating key translational regulators. This is evidenced in our data set, in which we found that eukaryotic translation initiation factor 4B (EIF4B) and eukaryotic translation initiation factor 4E–binding protein 1 (EIF4EBP1) were activated by site-specific phosphorylation (Table 2). Because adult myocardial cells do not proliferate, hypertrophy induced by sustained βAR activation (19) is primarily caused by the increased size of the individual cells (1). It will be important to investigate whether mTOR activation is maintained during chronic activation of the β1AR. Such activation would suggest a molecular mechanism for induced cardiac hypertrophy as a consequence of long-term activation of the β1AR pathway.

Table 2 Identified regulated phosphosites that have known functional roles.

A number of the phosphosites that we identified as regulated upon stimulation of β1AR have functional effects in systems other than that of βAR signaling. The table columns contain the name of the encoding gene, the amino acid position of the regulated phosphosite, the ratio of the abundance of the phosphopeptide in β1AR-stimulated mice to that in control mice, the P value for the change (by Student’s t test), and whether the association of the phosphosite with the βAR response was not previously characterized, as well as what functions have been previously described for the phosphosite in the literature.

View this table:

Site-specific regulation of ion channels with emphasis on KV7.1

The main role of the kinases activated as a result of β1AR stimulation is to target proteins with key roles in excitation-contraction coupling to enhance cardiac output. This includes increasing the influx of Ca2+ and Na+ to increase the intracellular Ca2+ concentration for enhanced cardiac contractility and to augment K+ efflux to repolarize myocardial cells faster for increased heart rate. We identified the exact residues that are modulated by phosphorylation on key proteins involved in these processes (Fig. 4). For a subset of the phosphosites, functional effects have previously been described (Table 2). Phosphorylation of Ser16 and Thr17 of phospholamban (PLN) facilitates enhanced Ca2+ re-uptake into the sarcoplasmic reticulum at high heart rates by removing the inhibitory effect of PLN on the sarcoplasmic-endoplasmic reticulum calcium ATPase (SERCA) (20). We found that the extent of phosphorylation of both sites was more abundant in the β1AR-stimulated mice than in the control mice. Moreover, we estimated the fractional phosphorylation site occupancies of the sites because we had ratio measurements from both phosphorylated and nonphosphorylated peptides (21). The phosphorylation occupancy of PLN Ser16 and Thr17 increased from ~1% in hearts from control mice to ~60% in hearts from the β1AR-stimulated mice, thereby underscoring the pivotal role of PLN in the βAR response. The extent of phosphorylation of phospholemman (PLM) at Ser83, which relieves the inhibition of the sarcolemmal Na+/K+-ATPase, was also greater in the β1AR-stimulated mice than in the control mice. This dual release from inhibition by phosphorylated PLN and PLM would be expected to increase the extrusion of Ca2+ and Na+ from the cytosol after myocardial contraction and represents a potential mechanism to counteract the augmented influxes through ion channels.

Ca2+ enters the cytosol through L-type calcium channels (CaV1.2) and ryanodine receptors (RyR2), and Na+ enters through sodium channels (NaV1.5). In vitro studies pointed to phosphorylation of CaV1.2 Ser1928 as the mechanism underlying enhanced Ca2+ current upon β1AR stimulation, but experiments with a mouse model carrying the channel mutation S1928A could not confirm this in vivo (22). We found that Ser1928 was equally phosphorylated in hearts from control and β1AR-stimulated mice, whereas the phosphorylation abundance was increased at two previously uncharacterized phosphosites: Ser1700 on CaV1.2 and Ser200 on its accessory β2 subunit. We also found that the extent of phosphorylation of RyR2 at Ser2693 and Ser2696 was substantially increased in β1AR-stimulated mice compared to control mice, in addition to that of the previously described phosphosite Ser2808. For NaV1.5, we identified increased phosphorylation at residues Ser484, Ser667, and Thr670 in response to β1AR stimulation. The enhanced K+ efflux that ensures faster repolarization is mediated by voltage-gated K+ channels. Upon β1AR stimulation, the threshold for KV7.1 channel activation is shifted to more negative potentials, and the maximal current is increased (23), which is presumably mediated by phosphorylation of KV7.1 Ser27 (24). Because of sequence differences between murine and human KV7.1, we did not cover Ser27 in the mouse experiments, because the relevant tryptic peptide was too short. Surprisingly, we still identified regulated KV7.1 phosphopeptides, suggesting that Ser27 is not the only phosphosite that regulates KV7.1 upon β1AR stimulation.

Phosphorylation can affect both the conduction and trafficking of ion channels (25). Hence, we designed experiments to simultaneously address which phosphorylation sites on human KV7.1 channels were regulated by βAR stimulation and whether the phosphorylation events could be involved in forward trafficking of the channel (Fig. 5A). Taking advantage of the fact that the subcellular localization of KV7.1 can be controlled in Madin-Darby canine kidney (MDCK) cells by a so-called calcium switch assay (see Materials and Methods for details) (26), we expressed and trapped human KV7.1 channels in either the endoplasmic reticulum (ER) or the plasma membrane (fig. S6A) in MDCK cells grown in SILAC (stable isotope labeling by amino acids in cell culture) media. Lysates from cells expressing the channel at the plasma membrane (cultured in light SILAC medium) were mixed either with lysates from cells expressing the channel in the ER (cultured in heavy SILAC medium) or with lysates from cells also expressing the channel at the plasma membrane but incubated with a βAR agonist (also cultured in heavy SILAC medium). The MDCK cells were stimulated with a general βAR agonist because they have more β1AR than β2AR (fig. S6B). We immunoprecipitated KV7.1 from these SILAC mixtures, and, after in-gel digest, tryptic peptides were analyzed by LC-MS/MS. Our data reveal that the extent of phosphorylation of human KV7.1 Ser27 and Ser92 was greater at the plasma membrane than in the ER, and that the extent of phosphorylation of both sites at the plasma membrane was increased upon βAR stimulation (Fig. 5, B to D). We confirmed that the increased extent of KV7.1 Ser92 phosphorylation was also specific to β1AR stimulation in MDCK cells by applying the same treatment strategy as in the animal studies (fig. S6C).

Fig. 5 βAR regulation of the KV7.1 ion channel.

(A) SILAC experiment with KV7.1 expressed in MDCK cells with channels localized to the plasma membrane (light), the ER (heavy), or the plasma membrane and stimulated with βAR agonist (heavy). (B and C) MS/MS spectra verifying the phosphorylation of KV7.1 residues Ser27 and Ser92. (D) SILAC heavy/light ratio plot for quantified KV7.1 peptides. Ratios for peptides isolated from channels localized at the plasma membrane (in cells stimulated with βAR agonist versus unstimulated cells) were plotted as a function of ratios for peptides isolated from channels localized at the plasma membrane as compared to the ER (n = 3 experiments for each condition). Each dot represents a KV7.1 peptide. Regulated phosphopeptides are in yellow, nonregulated phosphopeptides are in purple, and nonregulated, nonphosphorylated peptides are in blue. (E) Electrophysiological recordings from oocytes coexpressing KCNE1 and wild-type KV7.1 (n = 20 oocytes) or KV7.1_S92A (n = 24 oocytes). Currents were measured before and after stimulation with cAMP by applying the depicted voltage-step protocol. After stimulation with cAMP, KV7.1 conducted statistically significantly more current than did KV7.1_S92A. Representative current traces are shown. The Western blot shows the comparable abundances of KV7.1 and KV7.1_S92A (see fig. S8 for quantification). *P < 0.01 by Student’s t test.

To address the functional implications of Ser27 and Ser92 phosphorylation for KV7.1 channels, we investigated KV7.1 channels in which phosphorylation of Ser27 or Ser92 was either abolished or mimicked by site-specific mutagenesis. Trafficking of KV7.1 from the ER to the plasma membrane was not affected by substitution of Ser27, Ser92, or both with either alanine or aspartate (fig. S7), thus supporting our finding that KV7.1 is phosphorylated after it reaches the plasma membrane. We next investigated channel conduction, focusing on the residue Ser92 that we found was regulated in both murine and human KV7.1 because Ser27 has previously been investigated in this context (24). To mimic the slowly activating delayed rectifier K+ current (IKs) of the heart, we coexpressed wild-type human KV7.1 or its S92A or S92D mutants together with its interaction partner, potassium voltage-gated channel subfamily E member 1 (KCNE1), in Xenopus laevis oocytes (27). Currents were elicited and measured by voltage-step protocols before and after stimulation with cAMP, where cAMP was used to initiate the downstream effects of βAR stimulation. After stimulation, KV7.1 S92A conducted significantly less current than did the wild-type channel at all potentials tested (Fig. 5E, Student’s t test). Western blotting analysis verified that the abundances of the wild-type and S92A channel proteins were similar (Fig. 5E and fig. S8). The fold change in current amplitude induced by cAMP stimulation was significantly smaller for both the KV7.1 S92A and KV7.1 S92D mutant channels (fig. S8). Hence, KV7.1 Ser92 was phosphorylated upon βAR stimulation when the channel was localized at the plasma membrane, and this phosphorylation event resulted in increased IKs current. Phosphorylation of KV7.1 Ser27 and Ser92 thereby presents a mechanism for amplified K+ efflux at faster heart rates, facilitating an adequate diastolic interval for myocardial filling and perfusion.

Discussion

We identified 353 proteins that are regulated at 670 phosphosites in response to β1AR stimulation. These findings reveal signaling pathways not previously associated with the response, a potential mechanism that may explain cardiac hypertrophy, and phosphorylation sites on channels and transporters that regulate cardiac excitability. Several specific kinases are highlighted. For example, we showed that acute β1AR stimulation led to the activation of Akt3 and the inactivation of GSK3, which is further supported by the general observation that sites on known substrates of Akt were phosphorylated, whereas those of GSK3 were dephosphorylated. Delineating the particular kinases activated downstream of the receptor identified potential therapeutic targets circumventing the adverse effects of chronic βAR regulation. Accordingly, mice with constitutive GSK3 activity do not develop hypertrophy after sustained βAR stimulation (28). Because activated GSK3 inhibits mTOR (29), and because we showed that acute β1AR stimulation resulted in both reduced GSK3 activity and increased mTOR activity, it is more likely that mTOR is the kinase underlying hypertrophy given its role in controlling cell growth. The general value of high-resolution MS to quantitatively and qualitatively investigate signaling pathways in vivo is underscored with our investigation of the cardiac βAR response. Our data set provides a depiction for the entire heart of cardiac proteins regulated by the response and presents specific mechanisms to be investigated further. This study expands our molecular view of the βAR response substantially and demonstrates the utility of quantitative phosphoproteomics screens in tissue and organs for exploring signaling pathways. Moreover, our study identifies potential targets for drug discovery and disease treatment strategies that might substitute for or complement the widely prescribed β-blockers.

Materials and Methods

Animal experiments

Our investigation conforms to the Guide for the Care and Use of Laboratory Animals published by the U.S. National Institutes of Health and the Directive 2010/63/EU of the European Parliament. Nine male mice (C57BL6; body weight, 19 ± 1 g) were anesthetized with 1.5% isoflurane. Core body temperature was maintained at 37°C for the entire experimental procedure. Tygon tubing (inside diameter, 0.4 mm; outer diameter, 0.6 mm; dead space, <20 μl) was used for cannulation of the right jugular vein. Six-lead surface electrocardiogram was continuously recorded. After baseline stabilization (>5 min), drugs or 0.9% saline was administered twice with a 10-min interval. Drugs were dissolved in dimethyl sulfoxide (DMSO) and diluted in saline to a uniform administered volume of 10 μl per gram of body weight. The final DMSO concentration was 1%. Metoprolol (4.0 mg/kg, Sigma), ICI-118-551 (2.5 mg/kg, Sigma), dobutamine (1.5 mg/kg, Sigma), and salbutamol (2.5 mg/kg, Sigma) were infused intravenously over 1 min, and the tubing was flushed with 0.1 ml of saline. Ten minutes after the last dose, the thorax was quickly opened, and the heart was explanted, quickly rinsed in saline, and snap-frozen in liquid nitrogen. The β1AR and β2AR agonists and antagonists were chosen on the basis of receptor affinity data (metoprolol: KiADRB1 = 14 nM, KiADRB2 = 2455 nM; ICI-118-551: KiADRB1 = 135 nM, KiADRB2 = 0.6 nM; dobutamine: KdADRB1 = 3 μM, KdADRB2 = 25 μM; salbutamol: KdADRB1 = 22 μM, KdADRB2 = 0.8 μM) (http://www.bindingdb.org). Other ligands may exhibit larger βAR selectivity based on in vitro assays; however, the combined use of agonist and antagonist reduced the potential functional bias from activating both systems. Salbutamol is a partial agonist for β2AR and has only partial efficacy at the receptor relative to a full agonist. Hence, maximal receptor stimulation was not an objective and may not have been reached in the present study.

Tissue homogenization and protein digestion

The sample preparation was performed as described previously (6). Tissue samples were transferred to a urea solution with phosphatase inhibitors [6 M urea, 2 M thiourea, 10 mM Hepes (pH 8.0), 1 mM orthovanadate, 5 mM sodium fluoride, 5 mM β-glycerophosphate) in 5 μl of extraction buffer per milligram of tissue. The tissues were homogenized by ceramic beads (Precellys 24, Bertin Technologies) followed by microtip sonication on ice and centrifugation at 16,000g for 20 min at 4°C. The protein concentrations of the retrieved supernatants were measured (Quick Start Bradford Dye Reagent X1, Bio-Rad), and for each tissue lysate, three aliquots of 7-mg protein were reduced (final concentration of 1 mM dithiothreitol, 750 rpm, 30 min in a thermomixer) and alkylated (final concentration of 5.5 mM chloroacetamide, 750 rpm, 20 min in darkness in a thermomixer) before being digested with 35 μg of endoproteinase Lys-C (Wako) (750 rpm, 3 hours in a thermomixer), diluted with 25 mM ammonium bicarbonate to lower the urea concentration below 2 M, and then further digested with 35 μg of modified trypsin (sequencing grade, Promega) (750 rpm, 8 hours in a thermomixer). Trypsin digestion was quenched by lowering the pH to ~2 with trifluoroacetic acid (TFA). Samples were centrifuged at 16,000g for 20 min, and supernatants were desalted and concentrated on Sep-Pack C18 cartridges (Waters) and eluted with 60% acetonitrile (MeCN) in 1% TFA. At this point, 10 μg was taken from each sample and stored on in-house packed C18 STAGE tips for downstream measurement of nonphosphorylated peptides.

Phosphopeptide enrichment

Phosphopeptides were enriched with TiO2 beads essentially as described by Olsen et al. (4). One milligram of TiO2 beads (GL Sciences Inc.) was suspended in 5 μl of 2,5-dihydroxybenzoic acid (DHB) [0.02 g DHB/ml 80% MeCN, 0.5% acetic acid (AcOH)], mixed for 15 min at 30 rpm, and added to each of the samples, which were then incubated with gentle rotation for 15 min at 30 rpm. The TiO2 beads were quickly spun down, and the supernatants were transferred to new reaction tubes and incubated with a second round of TiO2 beads and incubated for another 15 min at 30 rpm. After spin-down, the beads were washed with 100 μl of 5 mM KH2PO4, 30% MeCN, 350 mM KCl, followed by 100 μl of 40% MeCN, 0.5% AcOH, 0.05% TFA, and then were resuspended in 50 μl of 80% MeCN, 0.5% AcOH. The resuspended beads were loaded onto in-house packed C8 STAGE tips in 200-μl pipette tips preconditioned with 80% MeCN, 0.5% AcOH, washed once with the same buffer, and eluted with 2 × 10 μl of 5% ammonia and 2 × 10 μl of 10% ammonia, 25% MeCN. Ammonia and organic solvents were evaporated with a vacuum centrifuge. The peptides were then acidified in 1% TFA, 5% MeCN and loaded onto in-house packed C18 STAGE tips, which were preconditioned with 20 μl of MeOH, 20 μl of 80% MeCN, 0.5% AcOH, and 2 × 20 μl of 1% TFA, 3% MeCN. After loading, the STAGE tips were washed with 2 × 20 μl of 8% MeCN, 0.5% AcOH and 1 × 50 μl of 0.5% AcOH.

LC-MS/MS analysis

Peptides were eluted with 2 × 10 μl of 40% MeCN, 0.5% AcOH, and the organic solvents were removed in a vacuum centrifuge. Peptides were reconstituted in 2% MeCN, 0.5% AcOH, 0.1% TFA and analyzed by online reversed-phase C18 nanoscale LC-MS/MS on an LTQ-Orbitrap Velos mass spectrometer (Thermo Electron) with the top 10 higher-energy collisional dissociation (HCD) fragmentation method (30), essentially as previously described (6). The samples were analyzed in randomized order. The LC-MS analysis was performed with a nanoflow Easy–nLC system (Proxeon Biosystems) connected through a nanoelectrospray ion source to the mass spectrometer. For the phosphopeptide-enriched samples, peptides were separated by a linear gradient of MeCN in 0.5% AcOH for 180 min in a 15-cm fused-silica emitter in-house packed with reversed-phase ReproSil-Pur C18-AQ 3 μm resin (Dr. Maisch GmbH). For the samples that were not enriched for phosphopeptides, a 70-min gradient was used. Full-scan MS spectra were acquired at a target value of 1 × 106 and a resolution of 30,000, and the HCD MS/MS spectra were recorded at a target value of 5 × 104 and with a resolution of 7500 with a normalized collision energy of 40%.

Peptide identification

Raw MS files were processed with MaxQuant software (ver.1.0.14.7, Max Planck Institute of Biochemistry, Department of Proteomics and Signal Transduction, Munich, Germany) by which the precursor MS signal intensities were determined and HCD MS/MS spectra were deisotoped and filtered such that only the 10 most abundant fragments for each 100 mass/charge ratio (m/z) range were retained. Proteins were identified with the Mascot search algorithm (http://www.matrixscience.com) by searching all MS/MS spectra against a concatenated forward and reverse version of the rat and mouse International Protein Index v.3.37 protein sequence database supplemented with protein sequences of common observed contaminants, such as human keratins and porcine trypsin. The HCD MS/MS spectra were searched with a fixed modification of carbamidomethyl-cysteine, and we allowed for variable modifications of oxidation (M), acetylation (protein N-term), Gln→pyro-Glu, and phosphorylation (STY). Search parameters were set to an initial precursor ion tolerance of 7 parts per million (ppm) with an MS/MS tolerance at 0.02 daltons and requiring strict tryptic specificity with a maximum of two missed cleavages. Peptides were filtered on Mascot score (>10), precursor mass accuracy, and peptide length (>6). To achieve an FDR < 0.01 with phosphopeptide identifications, we required quantitation in at least two-thirds of the experiments from the β1 group, the β2 group, or the control group. In general, the reporting of our mass spectrometry data acquisition, processing, and search results as well as sharing of all MS raw files has been done according to the Molecular and Cellular Proteomics Guidelines. Raw mass spectrometric files in Thermo Scientific’s *.raw format as well as annotated MS2 spectra for phosphopeptides are available for download through our Web site (http://cpr1.sund.ku.dk/datasets/proteomics/). A combined zip folder with all annotated MS2 spectra for all phosphopeptides can be downloaded from http://cpr1.sund.ku.dk/datasets/proteomics/Raw_files_BetaAdrenergicSignaling.zip. A combined zip folder with all MS raw files for phosphopeptide enriched samples can be downloaded from http://cpr1.sund.ku.dk/datasets/proteomics/RawFiles_Phospho_Experiments.zip. A combined zip folder with all MS raw files for proteome samples can be downloaded from http://cpr1.sund.ku.dk/datasets/proteomics/RawFiles_Nonphospho_Experiments.zip.

Data analysis

The summed extracted ion chromatogram (XIC)–based MS signal intensities of all detected charge-states for every modification-specific peptide in each LC-MS raw file were normalized with the “normalizeQuantiles” function from the Bioconductor R package LIMMA. This is a nonlinear method that performs intensity-dependent normalization. The function normalizes the peptide intensities such that each quantile for each sample is set to the mean of that quantile across the data set. This results in peptide intensity distributions that are empirically identical. The method is equivalent to the method used for normalization of microarray intensities between arrays (31). Downstream data analysis was performed with Microsoft Office Excel and Perseus (32) (Max Planck Institute of Biochemistry, Department of Proteomics and Signal Transduction, Munich, Germany) software. For the control distribution of nonphosphorylated peptides, only nonphosphorylated peptides from the normalized, modification-specific peptides were included in the analysis, and peptide identification was required in a minimum of three experiments in one group. All intensities were log2-transformed, and missing values were imputed by normal distributed values (width 0.3, down-shift 2.2). For the comparison of β1AR-stimulated mice versus controls as well as of β2AR-stimulated mice versus controls, a two-sided t test was performed with a permutation-based FDR cutoff of 0.01, with a bend of the curve of S0 = 0.5 and 250 randomizations. For the phosphorylated peptides, normalized intensities of phosphorylated, modification-specific peptides were used in the analysis. Peptide identification was required in at least three samples in one group, and the intensities were log2-transformed. For the comparison of β1AR-stimulated mice versus controls as well as of β2AR-stimulated mice versus controls, a two-sided t test was performed with a permutation-based FDR cutoff of 0.01, with a bend of the curve of S0 = 0.5 and 250 randomizations. Phosphosite occupancy was calculated on the basis of a modified version of the absolute phosphorylation site stoichiometry equations that we previously described for a global, quantitative phosphoproteomics analysis of cell cycle progression with SILAC quantitation (21). All stoichiometry calculations were performed pairwise between the control group and the β1AR-stimulated group, with all subsequent equations being based on ratios defined as sample B divided by sample A (B/A ratio). The proportion between a phosphorylated peptide and its corresponding nonphosphorylated peptide counterpart (termed a for sample A and b for sample B) was calculated from the normalized, label-free XIC-based ratio for the phosphorylated peptide (x), the label-free ratio for the same nonphosphorylated peptide (y), and the overall ratio of the protein based on its iBAQ ratio (z), with the constraint that the sum of phosphorylated peptide intensities and nonphosphorylated peptide intensities divided by the number of protein molecules must remain constant between the control and β1AR-stimulated groups. The proportion between a phosphorylated peptide and its corresponding nonphosphorylated peptide counterpart in sample A was calculated as a = (zy)/(xz), and in sample B as b = [x(zy)]/[y(xz)]. We therefore calculated the fractional phosphorylation site stoichiometry in samples A and B as a/(1 + a) and b/(1 + b), respectively. Sequence pattern analysis was made for all of the phosphorylation sites identified in mice that were stimulated with the β1AR agonist. All phosphopeptides whose abundances were statistically significantly regulated in the mice treated with the β1AR agonist were used as input for the amino acid sequences of the experimental data set, and all of the phosphopeptides whose abundances were not significantly regulated in the β1AR-stimulated mice were used as the reference data set. Amino acid sequences of the experimental data set were analyzed against the reference data set with iceLogo (33), with percentage difference as the scoring system and a P value cutoff of 0.01. Pathway and GO enrichment analyses were performed with proteins that had phosphorylation sites regulated upon β1AR stimulation as input with the innateDB software tool (15). Enriched pathways and GO terms for molecular function were determined on the basis of their corrected P values, requiring at least three input genes in a group and coverage greater than 5%. P values were calculated with a hypergeometric test and corrected for multiple testing with the Benjamini-Hochberg method, applying an FDR <0.01.

NetworKIN analysis

We used the integrative computational NetworKIN algorithm to predict kinase-substrate relationships (16). The NetworKIN prediction is based on an integrated evaluation of kinase consensus motifs, which are extracted from the NetPhorest atlas (34), with contextual information for kinase-substrate relationships to disambiguate in case of ambiguous motifs. The context modeling is based on probabilistic kinase-substrate associations derived from protein-protein interaction data from the STRING database (35). We used all of our identified phosphosites as NetworKIN queries, and we calculated prediction scores for the likely kinase that phosphorylated any given site. We filtered the identified kinases for those that we detected in the mouse hearts, such that we restrained the analysis to only those kinases for which we had experimental evidence. We thus considered 144 different kinases to be relevant. For kinase groups with only one member after filtering, we used the NetPhorest score. For kinase groups containing multiple kinases, we used the calibrated NetworKIN score. For each kinase, we compared the distribution of prediction scores of regulated phosphorylation sites to the distribution of prediction scores for phosphosites that were not regulated by β1AR stimulation by the nonparametric Kolmogorov-Smirnov test. Kinases with a two-sided Kolmogorov-Smirnov P value <0.01 were considered to have predicted activities that were statistically significantly different in the two data sets. On the basis of one-sided Kolmogorov-Smirnov tests, we evaluated whether the kinase was predicted to be more active in the data set represented by regulated phosphosites than in the data set represented by nonregulated phosphosites, thereby evaluating whether the kinase was more active in the β1AR-stimulated mice than in the control mice.

Protein-protein interaction network

The InWeb protein-protein interaction database that we used is a public database of protein-protein interactions. InWeb combines reported protein-protein interactions from MINT, BIND, IntAct, PPrel, ECrel, Reactome, and other sources, and, in addition to human interactions, it includes interactions in orthologous protein pairs that pass a strict threshold for orthology and assigns a probabilistic score to each interaction based on the neighborhood of the interaction, the scale of the experiment in which the interaction was reported, and the number of different publications in which the interaction is cited (13). We used DAPPLE to build and analyze a network with the regulated phosphoproteins as seeds (14) (http://www.broadinstitute.org/mpg/dapple). We considered direct and indirect connections, and DAPPLE evaluated the statistical significance of the network by comparing it to 1000 random, matched networks generated with a within-degree node-label permutation.

Western blotting analysis of kinases

After protein concentration determination was performed, 100 μg of cardiac lysates from control or β1AR-stimulated mice (lysates from three mice for each condition) was resolved by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto a nitrocellulose membrane (Protran, Biosciences). Proteins of interest were visualized with specific antibodies and an enhanced chemiluminescence kit (Invitrogen). The following commercial antibodies (all from Cell Signaling Technology) were used: anti–phospho-Ser308-Akt (cat. no. 9275), anti–total Akt (cat. no. 9272), anti–phospho-Ser21-GSK3A (cat. no. 9331), and anti–total GSK3A (cat. no. 9338). Quantification of blots was performed with ImageJ.

Control of KV7.1 channel localization in MDCK cells with a calcium switch and transient expression of KV7.1 mutants

MDCK cells stably expressing hKV7.1-eGFP (enhanced green fluorescent protein) (36) were plated on glass coverslips (12 mm in diameter, Thermo Scientific) for immunostaining or in 75-cm2 flasks for cell lysates. For SILAC experiments, SILAC DMEM (Dulbecco’s modified Eagle’s medium)–based medium deficient in arginine and lysine (PAA) was supplemented with 10% dialyzed fetal bovine serum, l-glutamine, streptavidin, sodium pyruvate, penicillin, and streptomycin as well as stable isotope-labeled arginine and lysine (Sigma-Aldrich). For “light” medium, we used l-[12C6, 14N4]arginine (Arg0) and l-[12C6, 14N2]lysine (Lys0). The cells were grown for 2 days in heavy or light SILAC medium until they reached confluency. The medium was then changed to heavy or light SILAC medium supplemented with 4 mM EGTA for 9 hours to depolarize the cells, and then again changed back to heavy or light SILAC medium to start the polarization process of the cells. The cells were allowed to polarize for up to 27 hours. During this polarization process, KV7.1 changes its intracellular localization. The channel is localized to the ER after 3 hours and is then released from the ER later in the polarization process, which results in strong basolateral expression after 24 hours (26). After 4 hours (when KV7.1 is localized in the ER) or 27 hours (when KV7.1 is localized in the basolateral membrane), the cells were treated with 100 nM isoproterenol for 20 min. The cells were then either fixed in 3% paraformaldehyde (for immunostaining) or harvested to prepare cell lysates. MDCK cells were transiently transfected with plasmids encoding wild-type hKV7.1 or its Ser27 or Ser92 mutants with Lipofectamine Plus reagent (Invitrogen) according to the manufacturer’s instructions. The cells were plated on glass coverslips and allowed to grow for 3 days to fully polarize before being fixed and subjected to confocal analysis. For targeted analysis of the KV7.1 Ser92 mutant channel, we analyzed polarized MDCK cells expressing KV7.1-eGFP at three different stimulation conditions, each prepared in five biological replicates. Control cells were treated with 100 nM metoprolol and 10 nM ICI-118-551 for 40 min; cells exposed to specific β1AR activation were treated with 10 nM ICI-118-551 for 20 min followed by treatment with 25 μM dobutamine for an additional 20 min; and cells exposed to specific β2AR activation were treated with 100 nM metoprolol for 20 min followed by treatment with 25 μM salbutamol for 20 min.

Immunostaining and confocal imaging

Transiently transfected or stably expressing cells grown on glass coverslips were fixed in 3% paraformaldehyde in phosphate-buffered saline (PBS) for 30 min at room temperature. Quenching was performed by a 30-min incubation with 0.2% fish skin gelatin in PBS supplemented with 0.1% Triton X-100 (PBST). The cells were incubated for 1 hour with primary antibodies diluted in PBST. Secondary antibodies were diluted in PBST and applied for 45 min. The coverslips were mounted in ProLong Gold (Invitrogen). The samples were analyzed by laser scanning confocal microscopy with either a Leica TCS SP2 system equipped with argon and helium-neon lasers or the Zeiss LSM 710 confocal system. Images were acquired with a 63× water immersion objective, numerical aperture (NA) 1.2 (Leica TCS SP2) or a 63× oil immersion objective, NA 1.4 (Zeiss LSM 710 confocal system), both with a pinhole size of 1 airy unit and a pixel format of 1024 × 1024. Line averaging was used to reduce noise in the obtained pictures. For double- and triple-labeling experiments, sequential scanning was used to enable separation of signals from the individual channels. The acquired images were processed with Photoshop (Adobe) or Zen 2009 (Zeiss). The antibodies used for immunostaining were goat polyclonal anti-KV7.1 (1:50, clone C-20, Santa Cruz Biotechnology) and Alexa Fluor 488–conjugated donkey anti-goat immunoglobulin G (1:200). Alexa Fluor 647 phalloidin (1:200) was used to stain F-actin filaments, and 4′,6-diamidino-2-phenylindole (1:300) was used to stain nuclei. All secondary antibodies were purchased from Invitrogen.

Preparation of cell lysates and immunoprecipitations

Unless otherwise noted, all chemicals were obtained from Sigma-Aldrich. MDCK cells were harvested in solubilization buffer [50 mM tris-HCl (pH 7.4), 10 mM NaCl, 10 mM KCl, 10 mM NaF, 1% Triton X-100, 0.5% sodium deoxycholate, 8 μM leupeptin, 0.4 μM Pefabloc, 1 mM orthovanadate, 5 mM sodium fluoride, 5 mM β-glycerophosphate] for 3 hours. The samples were then centrifuged at >10,000g at 4°C for 10 min, and the supernatant was collected. Protein concentrations were measured with the DC protein assay (Bio-Rad Laboratories) according to the manufacturer’s instructions and were adjusted to 1.5 μg/μl. The heavy and light SILAC samples were mixed in a 1:1 ratio. KV7.1-eGFP channels were immunoprecipitated with agarose beads coupled to a GFP-binding protein (Chromotek). Precipitates were resolved by SDS-PAGE, after which the gel was fixed and stained, and bands corresponding to the channel protein were excised and subjected to tryptic in-gel digest, as previously described (37). The peptides were analyzed by LC-MS/MS on a LTQ Orbitrap Velos Instrument with a 150-min gradient. For each experiment, the SILAC ratios were normalized to the median ratio for all nonphosphorylated peptides. For each peptide, the median-normalized SILAC ratio was calculated. For targeted analysis of KV7.1 Ser92, MDCK cells were harvested, KV7.1 channels were immunoprecipitated, and the channel proteins were subjected to in-gel digest as described earlier. All samples were analyzed by LC-MS/MS on a Q Exactive mass spectrometer targeting the phosphorylated Ser92 peptide VS(ph)IYSTR at m/z 453.2101 (2+) for HCD fragmentation. For quantitation, we extracted the ion chromatograms of the four most abundant fragment ions with ±10 ppm mass tolerance: the x6-98 ion at m/z 736.3630, the y6-98 ion at m/z 708.3681, the y5 ion at 639.3466, and the y4 ion at m/z 526.2625. For each experiment, the four fragment ion intensities were normalized relative to those of the unmodified peptides. The fold change in Ser92 phosphorylation abundance was calculated from the normalized intensities for each of the four fragment ions in each experiment relative to the median for the individual fragment ions in the control condition. The fold change reported is the median fold change for the four fragment ions.

Molecular biology

Plasmids containing complementary DNAs (cDNAs) encoding human KV7.1 (GenBank accession no. NM_000218) (in pXOOM for expression in X. laevis oocytes or mammalian cells) and KCNE1 (NM_000219) (in pGEM-HEJuel for expression in X. laevis oocytes) have been described previously (38). The mutations S27A, S27D, S92A, and S92D and the double mutants S27A/S92A and S27D/S92D were introduced by mutated oligo extension polymerase chain reaction assay with Pfu Turbo polymerase (Stratagene) from the plasmid template containing KV7.1 wild-type cDNA, digested with Dpn I (Fermentas), and used to transform Escherichia coli XL1 Blue cells. All constructs were verified by complete DNA sequencing of the cDNA insert (Macrogen Inc.). Complementary RNA (cRNA) for injection was prepared from linearized wild-type or mutant KV7.1 and KCNE1 with the T7 mMessage mMachine kit (Ambion) according to the manufacturer’s instructions and stored at −80°C until required for injection. RNA concentrations were quantified by UV spectroscopy, and quality was checked by gel electrophoresis.

Electrophysiology

X. laevis stage V and VI oocytes were injected with KV7.1 wild-type or KV7.1-S92A cRNA (5 to 10 ng per oocyte) and KCNE1 in a 1:1 molar ratio between channel subunits and accessory subunits. Oocytes from three different batches were tested, and currents were measured by two-electrode voltage clamp recordings as described previously (38). Oocytes were incubated in Kulori medium [90 mM NaCl, 1 mM KCl, 1 mM MgCl2, 1 mM CaCl2, 5 mM Hepes (pH 7.4)]. All experiments were performed at room temperature 24 hours after injection. Currents were recorded with a two-electrode voltage clamp amplifier (Dagan CA-1B) with electrodes pulled from borosilicate glass capillaries on a horizontal patch electrode puller with tip resistances between 0.3 and 1.5 megaohms when filled with 2 M KCl. During experiments, oocytes were placed in a small chamber (of 200-μl volume), and channel activity was measured in Kulori medium. After measurement, the oocytes were again placed in the incubator for 1 hour with 100 μM 8-Br-cAMP added to the Kulori medium. After incubation, currents were measured again. Data analysis was performed with Excel and PRISM software, with all values given as means ± SEM. Current-voltage (I/V) relations were obtained by plotting the outward current at the end of the 2-s test pulse as a function of the test potential. The number of independent experiments is indicated in the figure legends. Comparison of the biophysical properties was performed with an unpaired t test. P < 0.05 was considered statistically significant.

Protein preparation from oocytes and Western blotting analysis

Unless otherwise noted, experiments were performed at 4°C, and chemicals were obtained from Sigma-Aldrich. Total protein samples were prepared from oocytes either before or after 1 hour of stimulation with 100 μM 8-Br-cAMP at 19°C. Oocytes where homogenized by trituration in 1 ml of HbA homogenization buffer containing protease inhibitors [5 mM MgCl2, 5 mM NaH2PO4, 1 mM EDTA, 80 mM sucrose, 20 mM tris-HCl (pH 7.4), with 0.008 mM leupeptin, 1 PhosSTOP tablet per 10 ml of buffer, and 0.4 mM Pefabloc]. Samples were centrifuged at 250g for 10 min to remove cell debris, and the supernatant was transferred to a new tube and centrifuged at 16,000g for 20 min. The pellet was resuspended in 1× sample buffer, heated to 65°C for 15 min, and then frozen until further use. Western blotting analysis was performed with precast 4 to 15% SDS-PAGE gels (Bio-Rad) and nitrocellulose membranes. Primary antibody was diluted in PBS containing 5% milk and applied overnight at 4°C. Anti-KV7.1 (1:500, APC-022, Alomone) and anti-actin (1:1000, MAB1501, Millipore) were used as primary antibodies. Horseradish peroxidase–conjugated AffiniPure F(ab) fragment donkey anti-rabbit and anti-mouse antibodies (1:10,000, Jackson ImmunoResearch) were used as secondary antibodies. The membranes were stripped by incubation with 0.2 N NaOH for 30 min at room temperature and reblocked in 5% milk, and then new primary antibody was applied. Bands corresponding to KV7.1 and actin were quantified with ImageJ.

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/6/278/rs11/DC1

Fig. S1. Electrophysiological effects of pharmacological βAR modulation on murine hearts.

Fig. S2. Evaluation and assessment of MS and MS/MS data quality.

Fig. S3. Correlation plot analyses of biological replicates of phosphopeptide measurements.

Fig. S4. Statistical analysis of identified peptides.

Fig. S5. Western blotting analysis of Akt and GSK-3α in control and β1AR-stimulated mice.

Fig. S6. Localization of KV7.1 channels and presence of βARs in MDCK cells.

Fig. S7. Trafficking of KV7.1 Ser27 and Ser92 mutant channels in MDCK cells.

Fig. S8. Testing of KV7.1 wild-type, KV7.1 S92A, and KV7.1 S92D mutant channels coexpressed with KCNE1 in X. laevis oocytes.

Table S1. List of all identified modification-specific peptides.

Table S2. List of all identified class 1 phosphorylation sites.

Table S3. List of all identified proteins.

Table S4. List of all phosphorylation sites regulated by β1AR stimulation.

Table S5. List of all identified phosphorylated kinases.

References and Notes

Acknowledgments: We thank C. Frøkjær Jensen, H. Borger Rasmussen, and A. Kirby for valuable discussions; R. Graham, S.-P. Olesen, and G. Duker for discussions of βAR pharmacology; A. Mujezinovic for technical assistance with oocyte measurements; S. Grubb for input with electrophysiological experiments; L. Rossin for input on DAPPLE; and D. B. Bekker-Jensen for graphical presentations. Funding: This work was supported by The National Danish Council for Independent Research for Medical Sciences (research career program Sapere Aude to A.L. and J.V.O. and a research grant to M.B.T.), Marie Curie IEF and EMBO Long-Term postdoctoral fellowships (to C.F.), and the EU 7th Framework grant PRIME-XS (Contract no. 262067). The Center for Protein Research is partly funded by the Novo Nordisk Foundation. Author contributions: A.L. conceived and coordinated the project; designed the experiments; performed sample preparation, phosphoproteomics, proteomics, and immunoprecipitations; analyzed MS and electrophysiology data; and wrote the paper. J.V.O. designed the experiments, analyzed MS data, and wrote the paper. M.B.T. performed animal experiments and edited the manuscript. M.N.A. performed MDCK cell and imaging experiments. C.D.K. performed R-based analyses. C.F. performed Western blotting analysis. H.H. performed NetworKIN analysis. L.J.J. performed NetworKIN analysis and edited the paper. N.S. performed molecular biology experiments. A.B.S. performed molecular biology and electrophysiology experiments. Competing interests: J.V.O. consults for Evotec Munich, which has an interest in phosphoproteomics. Data and materials availability: Raw mass spectrometric files in Thermo Scientific’s *.raw format as well as annotated MS2 spectra for phosphopeptides are available for download through our Web sites at http://cpr1.sund.ku.dk/datasets/proteomics/Raw_files_BetaAdrenergicSignaling.zip, http://cpr1.sund.ku.dk/datasets/proteomics/RawFiles_Phospho_Experiments.zip, and http://cpr1.sund.ku.dk/datasets/proteomics/RawFiles_Nonphospho_Experiments.zip.
View Abstract

Navigate This Article