Adrenergic Regulation of HCN4 Channel Requires Protein Association with β2-Adrenergic Receptor*

Background: Protein complexes often play critical roles in signal transduction. Results: The HCN4 channel binds the β2-adrenergic receptor to form a macromolecular complex. Disruption of this channel-receptor complex abolishes adrenergic modulation of pacemaker currents and spontaneous contraction rates in sinoatrial nodal cells. Conclusion: The channel-receptor association is critical for cardiac chronotropic regulation. Significance: Channel-receptor complexes are the fundamental form of channel regulation. β1- and β2-adrenergic receptors utilize different signaling mechanisms to control cardiac function. Recent studies demonstrated that β2-adrenergic receptors (β2ARs) colocalize with some ion channels that are critical for proper cardiac function. Here, we demonstrate that β2ARs form protein complexes with the pacemaker HCN4 channel, as well as with other subtypes of HCN channels. The adrenergic receptor-binding site was identified at a proximal region of the N-terminal tail of the HCN4 channel. A synthetic peptide derived from the β2AR-binding domain of the HCN4 channel disrupted interaction between HCN4 and β2AR. In addition, treatment with this peptide prevented adrenergic augmentation of pacemaker currents and spontaneous contraction rates but did not affect adrenergic regulation of voltage-gated calcium currents. These results suggest that the ion channel-receptor complex is a critical mechanism in ion channel regulation.

␤ 1 -and ␤ 2 -adrenergic receptors utilize different signaling mechanisms to control cardiac function. Recent studies demonstrated that ␤ 2 -adrenergic receptors (␤ 2 ARs) colocalize with some ion channels that are critical for proper cardiac function. Here, we demonstrate that ␤ 2 ARs form protein complexes with the pacemaker HCN4 channel, as well as with other subtypes of HCN channels. The adrenergic receptor-binding site was identified at a proximal region of the N-terminal tail of the HCN4 channel. A synthetic peptide derived from the ␤ 2 AR-binding domain of the HCN4 channel disrupted interaction between HCN4 and ␤ 2 AR. In addition, treatment with this peptide prevented adrenergic augmentation of pacemaker currents and spontaneous contraction rates but did not affect adrenergic regulation of voltage-gated calcium currents. These results suggest that the ion channel-receptor complex is a critical mechanism in ion channel regulation.
Adrenergic receptors (ARs) 2 exert essential control on physiological cardiac function. In particular, ␤AR activation increases heart rate and contractility through production of a second messenger, cAMP. The cAMP pathway is a classical model of a soluble messenger that diffuses to remote target effectors and exhibits physiological regulation. However, accumulating evidence sheds light on a new perspective on this model. In contrast to the classical view of a global increase in cAMP, a local increase in cAMP within small subcellular regions often plays a critical role in physiological responses (1). These local events form signaling hotspots or microdomains that provide fine-tuning of hormonal control for cardiac function. Prototypical examples are ␤ 1 AR and ␤ 2 AR in the heart. Both of these ARs activate the cAMP pathway. However, stimulation of these receptors shows distinct subcellular cAMP increases; stimulation of ␤ 1 AR raises global cAMP levels, whereas stimulation of ␤ 2 AR raises local cAMP levels (2,3). Accordingly, stimulation of these receptors evokes distinct cardiac responses. It is not fully understood how these differences are generated, but one mechanism would be receptor-specific protein complexes that tether relevant signaling enzymes and molecules (4).
A voltage-gated calcium channel (Ca v 1.2) has been shown to form a protein complex with ␤ 2 AR, which provides local regulation of channel activity in neurons (5). We have previously demonstrated that the neuronal M-type voltage-gated potassium channel KCNQ2 requires a signaling complex containing the m 1 muscarinic acetylcholine receptor for its regulation in neurons (6). These studies suggest that molecular association of the ion channel and G-protein-coupled receptor is a fundamental form of a signaling mechanism.
The HCN4 channel is a hyperpolarization-activated channel that is expressed predominantly in sinoatrial nodal (SAN) cells. The HCN4 channel generates the funny current, I f (7), which is widely believed to be the pacemaker current at the sinoatrial node. However, it is still unclear as to how other mechanisms contribute to the pacemaking activity because sinoatrial nodes from HCN4-deficient mouse embryos can still produce regular heartbeats (7)(8)(9)(10)(11). Nonetheless, accumulating evidence indicates that HCN4 plays a critical role in generating regular physiological heartbeats (12,13).
Adrenergic stimulation increases the heart rate through augmentation of I f . Not surprisingly, the HCN4 channel contains the cAMP-binding module, and HCN4 channel activity is facilitated by cAMP (14). Interestingly, recent studies show that ␤ 2 AR colocalizes with the HCN4 channel at caveolae (15,16), and ␤ 2 AR exhibits predominant control of this channel (17). We reasoned that HCN4 and ␤ 2 AR would be good candidates to test for ion channel-receptor complex formation and its physiological relevance. In this study, we demonstrate that 1) the N-terminal cytosolic tail of HCN4 is the ␤ 2 AR-binding site, 2) the peptide derived from the binding site disrupts channelreceptor binding, and 3) the peptide prevents ␤-adrenergic regulation of the pacemaker current and chronotropic function in SAN cells.

EXPERIMENTAL PROCEDURES
Expression Plasmids and Materials-The HCN4 channel in pcDNA3 was obtained from Martin Biel (Ludwig-Maximilians University). V5 epitope-tagged HCN4 was generated by PCR and subcloned using a pcDNA3/V5 TOPO directional kit (Invitrogen). HCN1-V5 was obtained by RT-PCR from mouse brain cDNA according to the sequences in the GenBank TM Data Bank (AF028737) and subcloned into the pcDNA3/V5 vector. HCN2-GFP was obtained from Tallie Baram (University of California, Irvine). FLAG-tagged ␤ARs were obtained from Robert Lefkowitz (Duke University) via Addgene. ␣ARs were obtained from the Missouri S&T cDNA Resource Center. FLAG-tagged ␣AR constructs were generated by PCR and subcloned into the pcDNA3 vector. The A-kinase activity reporter (AKAR) plasmid has been described (18). All PCR-derived constructs were verified by sequencing. The N-terminal myristoyl peptides were synthesized at GenScript (Piscataway, NJ) according to sequences GVNKFSLRMFGSQKAVEREQERVK-SAGFWIIHPYSD for HBAR and SVVFRAREYKEVGWG-GHAMRQLSFDSIFNKKQPSEI for shHBAR. The peptides were dissolved in water as 227 M stock solutions and kept at Ϫ80°C.
Cell Preparation and Cultures-HEK293 and HeLa cells were grown in DMEM with 10% fetal bovine serum. Serum-free cultures of rat neonatal ventricular myocytes were used for the calcium current recordings. Acquisition of tissues and primary cells was under the regulation of the Institutional Animal Care and Use Committee at the University of California, Irvine. Preparation of rat neonatal ventricular myocytes was described previously (18). Isolated myocytes were cultured in a serumfree medium containing M199 medium with Earle's salts, 2.2 mg/ml sodium bicarbonate, 25 mM HEPES, 2 mg/ml BSA, 2 mM L-carnitine, 5 mM creatine, 0.1 M insulin, 5 mM taurine, 100 IU/ml penicillin, and 100 g/ml streptomycin. Cells were used within 6 days after plating. Rat SAN cells were prepared basically according to the procedure described (9). Adult rats were deeply anesthetized with isoflurane and decapitated. Hearts were removed, and SAN regions were isolated. SAN regions were cut into tissue strips and used for enzymatic dissociation. The enzymatic solution contained 2 mg/ml collagenase type 1A (Sigma-Aldrich), 2 mg/ml elastase (Worthington), and 1 mg/ml BSA in a low calcium solution (140 mM NaCl, 5.4 mM KCl, 1 mM MgCl 2 , 0.2 mM CaCl 2 , 5.5 mM D-glucose, and 5 mM HEPES (pH 7.4)). After transferring the tissue strips using a flame-forged Pasteur pipette into 70 mM L-glutamic acid, 20 mM KCl, 80 mM KOH, 10 mM (ϩ)D-␤-OH-butyric acid, 10 mM KH 2 PO 4 , 10 mM taurine, 1 mg/ml BSA, and 10 mM HEPES (pH 7.4), cells were dissociated by passing through a strainer several times and collected by centrifugation. Cells were kept on ice until used in the patch-clamp experiments.
Immunoprecipitation-HEK cells were transfected using TransIT-LT1 reagent (Mirus Bio LLC, Madison, WI). Thirty hours after transfection, cells were harvested and lysed in buffer A (150 mM NaCl, 5 mM EGTA, 5 mM EDTA, 20 mM HEPES (pH 7.4), and 1% Triton X-100) with Complete protease inhibitor mixture (Roche Applied Science). Replacement of Triton X-100 with 1% n-dodecyl-␤-D-maltopyranoside (Affymetrix, Santa Clara, CA) or 1% CHAPS (Sigma-Aldrich) produced identical results. After centrifugation at 22,000 ϫ g for 30 min, supernatants were further precleared by incubation with protein G resin. Immunoprecipitations were performed with 10 l of anti-FLAG antibody-conjugated resin (Sigma-Aldrich) or protein G-Sepharose together with 1 g of anti-GFP antibody (Abcam, Cambridge, MA). Following overnight incubation at 4°C, the immunoprecipitates were washed twice with buffer A, twice with buffer A containing 650 mM NaCl, twice with buffer A, and once with 10 mM Tris (pH 7.6) and 1 mM EDTA and then analyzed by immunoblotting.
Surface Protein Labeling-HCN4-V5 channels were transiently expressed in HEK cells. Cells were pretreated with 1 M HBAR or 1 M shHBAR for 3 h at 37°C. Cells were washed twice with ice-cold PBS, followed by incubation with sulfo-NHS-LC-biotin (Thermo Scientific) for 30 min at 4°C. The treated cells were then washed twice with PBS containing 100 mM glycine. Cells were lysed in buffer A with Complete protease inhibitor mixture. Biotinylated proteins were purified by NeutrAvidin resin (Thermo Scientific) and detected by immunoblotting using anti-V5 antibody.
Electrophysiological Measurements-Patch-clamp recordings were performed on isolated cells using an Axopatch 200B patch-clamp amplifier (Molecular Devices, Sunnyvale, CA). Signals were sampled at 2 kHz, filtered at 1 kHz, and acquired using pCLAMP software (Version 7, Molecular Devices). For measuring I f , regularly beating single spindle-shaped SAN cells were visually selected for patch clamping. Patch pipettes with 2-4 megohms were filled with a solution containing 130 mM potassium aspartate, 10 mM NaCl, 5 mM EGTA, 2 mM CaCl 2 , 2 mM MgCl 2 , 2 mM ATP, 0.1 mM GTP, 5 mM creatine phosphate, and 10 mM HEPES (pH 7.2). The extracellular perfusion solution contained 110 mM NaCl, 30 mM KCl, 1.8 mM CaCl 2 , 1 mM MgCl 2 , 5.5 mM D-glucose, and 5 mM HEPES (pH 7.4). Cells were held at Ϫ50 mV. For step hyperpolarization, 1-s voltage steps from Ϫ50 to Ϫ150 mV were applied. For voltage ramping, voltage steps ranging from ϩ10 to Ϫ150 mV with a duration of 1.4 s were applied every 10 s. Cells showing an increasing linear leak were excluded from analyses.
For measuring voltage-gated calcium currents, the pipette solution contained 140 mM CsCl, 2 mM MgCl 2 , 10 mM EGTA, 3 mM ATP, and 10 mM HEPES (pH 7.2). The extracellular solution contained 15 mM BaCl 2 , 145 mM tetraethylammonium chloride, 10 mM D-glucose, and 10 mM HEPES (pH 7.4). Cells were held at Ϫ80 mV. To obtain voltage-current relationships, 500-ms step depolarizations were applied in 10-mV steps from Ϫ80 to ϩ50 mV. To monitor adrenergic regulation of calcium current, 100-ms step depolarizations to Ϫ10 mV were applied every 10 s. HBAR and shHBAR were added to the cell suspension or culture plate at a final concentration of 1 M and treated for 30 min to 2 h before recording. The peptides were included in the patch solution at 1 M in relevant experiments. All electrophysiological measurements were performed at room temperature.
Spontaneous Contraction Measurement-Phase-contrast live cell images of regularly beating SAN cells (spindle and muscle types) were obtained using an Olympus IX-81 inverted microscope and a ϫ20 phase-contrast objective lens equipped with a Hamamatsu Photonics ImagEM CCD camera controlled by MetaMorph 7.6.3 (Molecular Devices). Images were acquired at 285 frames/10 s. Regions of interest were selected such as to detect clear differences for spontaneous contractions. The cell chamber was maintained at 35°C.
FRET Experiments-HeLa cells were transfected with an expression plasmid containing AKAR. One day after transfection, transfected cells were replated onto 18-mm glass coverslips. Measurements of FRET signals were performed as described (19). Briefly, fluorescence emission was acquired using a microscope imaging station as described above. The excitation light was generated by Lambda LS (Sutter Instrument Co., Novato, CA) and passed through an S436/10ϫ or S500/20ϫ filter. The light intensity was reduced to minimize photobleaching. Dual-emission images were obtained through a dual-view module (Photometrics, Tucson, AZ). The exposure time was 100 ms; images were taken every 10 s. Fluorescence intensities from each cell were background-subtracted, and the YFP/CFP ratios were calculated. Cells were washed and observed in a solution containing 144 mM NaCl, 5 mM KCl, 2 mM CaCl 2 , 0.5 mM MgCl 2 , 10 mM glucose, and 10 mM HEPES (pH 7.4). All data are presented as means Ϯ S.E.

RESULTS
The V5 epitope-tagged HCN4 channel and four subtypes (␣ 1a , ␣ 2b , ␤ 1 , and ␤ 2 ) of FLAG-tagged ARs were cotransfected in HEK293 cells. These cells were used for immunoprecipitation with anti-FLAG antibody to purify the four AR subtypes. Strong co-purification of the HCN4 channel was detected with ␤ 2 AR but not with the other AR subtypes (Fig.  1A). This selective receptor preference of the HCN4 channel caught our attention. We tested whether other subtypes of HCN channels can interact with ␤ 2 AR: HCN1-V5 and HCN2-GFP were coexpressed with the four AR subtypes, and their interaction was examined by immunoprecipitation. Both HCN1 and HCN2 channels interacted with ␤ 2 AR but not with the other AR subtypes (Fig. 1, B and C). These results suggest that HCN channels have a common mechanism for interacting with ␤ 2 AR.
To elucidate the molecular mechanism of binding, mapping analyses were performed on the HCN4 channel (Fig. 2). Various GFP-tagged HCN4 fragments were coexpressed with ␤ 2 AR and tested for binding by immunoprecipitation. As an initial step, we determined whether the N-or C-terminal tail was responsible for the binding. Only the GFP fusion protein containing the N-terminal tail bound ␤ 2 AR to a similar extent as that containing the full-length HCN4 channel (Fig. 2, A and B). Further mapping analyses demonstrated that a fragment containing amino acid residues 225-260 of the HCN4 protein was sufficient to bind ␤ 2 AR (Fig. 2, C and D). However, we often observed that longer fragments exhibited stronger binding (Fig.  2, C and D), thus suggesting that the surrounding regions of this binding domain contribute to or stabilize the interaction. The identified ␤ 2 AR-binding region is conserved within HCN channel subtypes, as expected from the immunoprecipitation results (Fig. 2E). A computational secondary structure analysis (20) predicted that the domain contains an amphipathic ␣-helix (amino acids 237-249) sandwiched by short extended strands on both sides.
We anticipated that a synthetic peptide derived from this ␤ 2 AR-binding domain of HCN4 would function as a disrupter peptide for HCN4-␤ 2 AR interaction. To this end, a peptide (designated the HBAR peptide) was synthesized according to the amino acid sequence of the ␤ 2 AR-binding domain of HCN4 (amino acids 225-260). In addition, a shuffled peptide, the shHBAR peptide, was also synthesized as a control peptide. When 1 M HBAR peptide was included in the immunoprecipitation procedure, it reduced the coprecipitation of the HCN4 channel, indicating interference in the binding of ␤ 2 AR and the HCN4 channel as expected (Fig. 3A). In contrast, the control shHBAR peptide did not alter the coprecipitation of the HCN4 protein. Reduction of the ␤ 2 AR-bound HCN4 channel in the presence of HBAR peptides was not due to degradation or internalization of the HCN4 channel because the overall pro- tein and surface protein levels of the HCN4 channel were not altered (Fig. 3B). Next, we examined whether HBAR interferes with receptor binding of HCN1 or HCN2 because the ␤ 2 ARbinding site is conserved among HCN channels. The identical conditions used for the HCN4 treatment reduced both HCN1 and HCN2 binding to ␤ 2 AR (Fig. 3, C and D).
Because the HBAR peptide successfully disrupted HCN-␤ 2 AR interactions, we used this peptide to study the functional role of the HCN channel-␤ 2 AR complex. We dissociated pacemaker cells from rat sinoatrial nodes and measured I f , which is predominantly generated by the HCN4 channel. Regularly beating spindle-shaped cells were visually selected for patchclamp measurements. These cells exhibited slowly activated inward currents that were activated by hyperpolarization (Fig.  4A), which is a characteristic feature of I f . Application of 1 M isoproterenol (ISO) enhanced I f (Fig. 4, A and B), which was associated with a shift in the activation curve (Fig. 4C) as reported previously (21). We also observed that repetitive strong step hyperpolarizations often broke the tight seals of patch pipettes, which made it difficult to monitor I f continuously by this measurement. Thus, we used a voltage ramp protocol to monitor time-dependent changes. Slow voltage ramps from ϩ10 to Ϫ130 mV with a duration of 1.4 s were applied every 10 s. This hyperpolarizing voltage ramp activated an inward current at V Ͻ Ϫ70 mV (Fig. 4D). This hyperpolarization-activated current was inhibited by the HCN channel blocker ZD7288 (10 M) (Fig. 4D), thus confirming that the activated current was indeed I f . As observed by step depolarizations, application of ISO augmented voltage ramp-activated I f (Fig. 4E). Surprisingly, I f increased within 10 s of ISO application, and the increased channel activity was maintained thereafter (Fig. 4F). Next, we examined how the HBAR and shHBAR peptides affect I f . HBAR treatment did not alter the size or current densities of SAN cells (Table 1). In addition, the incomplete Boltzmann analysis suggested that there were no differences in steady-state gating parameters for I f . On the other hand, HBAR treatment abolished adrenergic responses of I f  . Synthetic peptide HBAR disrupts the HCN channel-␤ 2 AR interaction. A, pretreatment with HBAR but not shHBAR reduced co-immunoprecipitation of the HCN4 channel. HCN4-V5 and FLAG-␤ 2 AR were coexpressed in HEK cells and subjected to immunoprecipitation (IP) using anti-FLAG antibody. B, pretreatment with HBAR did not alter HCN4 protein amount at the plasma membrane (surface HCN4) and in the whole cell lysate. C, pretreatment with HBAR reduced ␤ 2 AR-bound HCN1. D, pretreatment with HBAR reduced ␤ 2 AR-bound HCN2. Error bars show S.E. JULY 6, 2012 • VOLUME 287 • NUMBER 28 when stimulated by 1 M ISO (Fig. 5, A, E, and F). In contrast, shHBAR-treated cells maintained equivalent adrenergic responses compared with the untreated cells (Fig. 5, B, E, and  F). To further evaluate the physiological relevance for the HCN channel-␤ 2 AR complex, we measured spontaneous contrac-tion rates in isolated SAN cells (Fig. 6). Control shHBARtreated SAN cells showed 244 Ϯ 26 beats/min (Fig. 6, A and C). Application of 1 M ISO increased the rate to 125 Ϯ 9.0% (n ϭ 15) of the control. In contrast, HBAR-treated cells did not respond to ISO (91.7 Ϯ 3.7%, n ϭ 12) (Fig. 6, B and C). These

TABLE 1 Electrophysiological characteristics of untreated, HBAR-treated, and shHBAR-treated SAN cells
Cell capacitances were obtained by a built-in function of Axopatch 200B during seal tests. I(⌬Ϫ40 mV) is the difference between I(Ϫ110 mV) and I(Ϫ70 mV) obtained by the voltage ramp protocol as described for Fig. 4E. Values are means Ϯ S.E. ANOVA, analysis of variance; pF, picofarads. findings suggest that the association of HCN channels and ␤ 2 AR is critical for the adrenergic regulation of the cardiac chronotropic function.

HBAR-treated SAN cells (n) shHBAR-treated SAN cells (n)
To eliminate the possibility that HBAR treatment disturbs the cAMP pathway, we measured PKA activity using a FRETbased protein kinase A activity reporter, AKAR, in HeLa cells (18,22). AKAR-expressing cells were first stimulated with a ␤ 2 -agonist, terbutaline, followed by ISO, which also activated ␤ 1 AR. Application of terbutaline evoked a 5.6 Ϯ 0.5% (n ϭ 32) increase in the relative YFP/CFP ratio of AKAR (Fig. 7A). Subsequent application of ISO further increased the YFP/CFP ratio, which showed a peak increase of 10.1 Ϯ 0.7% (n ϭ 32) (Fig. 7A). ISO application alone resulted in an 11.1 Ϯ 2.1% (n ϭ 8) increase in the relative CFP/YFP ratio compared with the control, which is consistent with reported AKAR responses to ISO (22). These AKAR responses were suppressed by pretreatment with a ␤ 2 AR-selective inhibitor, ICI-118551. The pretreatment reduced terbutaline-induced AKAR responses to 19.8 Ϯ 8.6% of the control (n ϭ 14) and ISO-induced responses to 63.5 Ϯ 8.0% (n ϭ 14) of the control (Fig. 7A). These results confirmed the activation of ␤ 2 AR by terbutaline and of both ␤ 1 AR and ␤ 2 AR by ISO. We then examined whether HBAR had any effect on PKA activity. Pretreatment with either HBAR or shHBAR did not alter terbutaline-and ISO-induced AKAR responses (Fig. 7, B and C). These results suggest that HBAR does not interfere with the general cAMP pathway or PKA activation.
To further examine the effect of HBAR on other channels, we measured adrenergic regulation of voltage-gated calcium channels from ventricular myocytes. Voltage-gated calcium currents were measured in cultured ventricular myocytes using barium as a charge carrier. Application of 1 M ISO induced a 2.7 Ϯ 0.2-fold increase (n ϭ 6) in the calcium current (Fig. 8, A  and B). Incubation with HBAR or shHBAR resulted in equivalent ISO responses compared with the untreated myocytes (Fig.   8B). These experiments indicate that HBAR treatment does not disturb adrenergic regulation of the calcium channel.

DISCUSSION
This study emphasizes the importance of protein interactions between ion channels and G-protein-coupled receptors in ion channel regulation. Our results suggest that tethering of the input (receptor) and the effector (ion channel) as a protein complex is critical for signal transduction. A similar functional relevance of channel-receptor complexes has been identified  . PKA activation is not disturbed by HBAR treatment. A, PKA activity was measured using AKAR. Application of a ␤ 2 AR-selective agonist, terbutaline, was followed by application of a nonselective agonist for ␤ 1 AR ϩ ␤ 2 AR, ISO. YFP/CFP ratios calculated from fluorescence intensities measured at the cytoplasm are plotted against time. Ratios were normalized to that at t ϭ 0. Application of 5 nM ICI-118551 suppressed most of the terbutalineinduced PKA activity. B, PKA activity in HBAR-or shHBAR-treated cells. C, summary of PKA activity shown in A and B. ICI-118551 significantly suppressed AKAR responses induced by both terbutaline and ISO, but neither HBAR nor shHBAR treatment altered AR-mediated PKA activation. **, Ͻ 0.01. Error bars show S.E.

HCN4 Channel Complex with ␤ 2 -Adrenergic Receptor
for other ion channels, such as the Ca v 1.2 calcium channel and ␤ 2 AR (5), as well as the KCNQ2 potassium channel and m 1 acetylcholine receptor (6).
In this work, we demonstrated that multiple subtypes of HCN channels form protein complexes with ␤ 2 AR. When HCN channel-␤ 2 AR complexes were disrupted by HBAR, I f did not respond to ISO. On the other hand, adrenergic regulation of the cardiac ventricular calcium current, which is generated by Ca v 1.2 (23), was not affected by HBAR. These results indicate that HBAR treatment is selective to the HCN4 channel complex. Furthermore, HBAR treatment did not alter PKA activation mediated by both ␤ 1 AR and ␤ 2 AR. These results indicate that HBAR does not interfere with the cAMP signaling pathway per se. Thus, loss of the adrenergic response of I f in the presence of HBAR would be due to the disruption of the channel-receptor interaction.
Adrenergic modulation of I f is a critical mechanism in heart rate regulation governed by the sympathetic nerve system. Despite the wide belief that ␤ 1 AR is the major AR controlling cardiac function (24), Barbuti et al. (15) have demonstrated that the HCN4 channel colocalizes with ␤ 2 AR at caveolae in SAN cells. Subsequently, they demonstrated that I f is regulated mainly by ␤ 2 AR and that disruption of caveolae prevents ␤ 2 AR-mediated regulation (17). Because our results indicate that disruption of ␤ 2 AR signaling extinguishes the response to ISO, our findings are consistent with the observation of Barbuti et al. that ␤ 2 AR is the dominant subtype for the regulation of I f in SAN cells.
Interestingly, Ca v 1.2 shows many similarities to HCN4 channels: both types of channels 1) are located in caveolae (15,25), 2) form a protein complex with ␤ 2 AR (Ref. 5 and this study), and 3) bind caveolin-3 (5,16,25). However, these two channel types show distinct adrenergic regulations, namely the Ca v 1.2 channel in cardiac ventricular myocytes is regulated by both ␤ 1 AR and ␤ 2 AR (23,25), and disruption of caveolae prevents only ␤ 2 AR-mediated regulation while sustaining ␤ 1 AR-mediated regulation (25). On the other hand, as mentioned above, the HCN4 channel in SAN cells is regulated predominantly by ␤ 2 AR, whereas ␤ 1 AR does not couple to the HCN4 channel unless caveolae are disturbed (17). These results suggest that the caveolae of ventricular myocytes and SAN cells have distinct cAMP compartments.
We demonstrated in Fig. 4F that I f reached its maximal adrenergic response within 10 s after agonist application. In contrast, global PKA phosphorylation measured using AKAR required 60 s to reach the peak response (Fig. 7A). We think that this rapid adrenergic augmentation of I f is further evidence for channel-receptor complexes because the minimal time requirement for cAMP to reach the channel suggests their contiguous location. We previously demonstrated a similar case with PKC phosphorylation, where m 1 acetylcholine receptorinduced PKC phosphorylation was accomplished within 20 s of agonist application when PKC was tethered adjacent to the PKC substrate, whereas global PKC phosphorylation took 60 s (26). Thus, we speculate that the HCN4-␤ 2 AR complex would be an underlying mechanism for rapid heart rate change upon adrenergic stimulation.
In summary, we have demonstrated that HCN-␤ 2 AR association is critical for regulation of I f . We therefore propose that protein complexes containing both the ion channel and G-protein-coupled receptor are a fundamental mechanism in controlling cellular responses.