Agonist Dose-dependent Phosphorylation by Protein Kinase A and G Protein-coupled Receptor Kinase Regulates β2 Adrenoceptor Coupling to Gi Proteins in Cardiomyocytes*

Adrenoceptors receptors (ARs) play a pivotal role in regulating cardiovascular response to catecholamines during stress. β2ARs, prototypical G protein-coupled receptors (GPCRs), expressed in animal hearts, display dual coupling to both Gs and Gi proteins to control the adenylyl cyclase-cAMP dependent protein kinase A (PKA) pathway to regulate contraction responses. Here, we showed that the β2AR coupling to Gi proteins was agonist dose-dependent and occurred only at high concentrations in mouse cardiac myocytes. Both the β2AR-induced PKA activity, measured by fluorescence resonance energy transfer (FRET) imaging, and the increase in myocyte contraction rate displayed sensitivity to the Gi inhibitor pertussis toxin (PTX). Further studies revealed that activated β2ARs underwent PKA phosphorylation at a broad range of agonist concentrations. Disruption of the PKA phosphorylation sites on the β2AR blocked receptor/Gi coupling. However, a sufficient β2AR/Gi coupling was also dependent on the G protein-coupled receptor kinase (GRK)-mediated phosphorylation of the receptors, which only occurred at high concentrations of agonist (≥100 nm). Disruption of the GRK phosphorylation sites on the β2AR blocked receptor internalization and coupling to Gi proteins, probably by preventing the receptor's transportation to access Gi proteins. Furthermore, neither PKA nor GRK site mutated receptors displayed sensitivity to the Gi-specific inhibitor, GiCT. Together, our studies revealed distinct roles of PKA and GRK phosphorylation of the β2AR for agonist dose-dependent coupling to Gi proteins in cardiac myocytes, which may protect cells from overstimulation under high concentrations of catecholamines.

and ␤ 2 ARs are prototypical GPCRs expressed in animal hearts, which mediate the increases in cardiac contraction upon agonist stimulation through the G s -adenylyl cyclase-cAMPdependent PKA pathway (1). ␤ 2 ARs are also known to uniquely couple to inhibitory G i proteins, which inhibits adenylyl cyclases to reduce cardiac contraction and initiates anti-apoptotic and cell growth signaling (2,3). Although ␤ 2 AR coupling to G i has been studied in reconstituted systems and in fibroblasts (4,5), little is known about the regulation process and the circumstances under which this coupling occurs in cardiac cells, along with its physiological consequences.
Various studies indicate that phosphorylation of ␤ 2 AR plays a critical role in regulating differential G protein coupling. The ␤ 2 AR phosphorylation by PKA mediates the switch of coupling from G s to G i (6,7), presumably operating in a feedback after receptor activation. Our previous studies have revealed that ␤ 2 AR/G i coupling is also dependent on receptor internalization and recycling (8,9). Meanwhile, Wang et al. have shown that direct inhibition of GRK2 prevents ␤ 2 AR/G i coupling in mouse cardiac myocytes (8), supporting a role of the ␤ 2 AR phosphorylation by GRK in receptor/G i coupling.
Studies have also revealed that ␤ 2 ARs are differentially phosphorylated by PKA and GRK in fibroblasts in an agonist dosedependent manner (10 -13). Although the PKA-mediated phosphorylation is sensitive to stimulation with subnanomolar concentrations, the GRK-mediated phosphorylation happens only when ␤ 2 ARs are stimulated with saturated concentrations (10 -13). This differential regulation of receptor phosphorylation by different kinases and their effects on subsequent receptor trafficking indicate that the ␤ 2 AR/G i coupling may be finely regulated in an agonist concentration-dependent manner.
Here, using a fluorescence resonance energy transfer (FRET) assay with a biosensor for PKA activity and a physiological cardiac contraction rate assay, we show for the first time that ␤ 2 AR/G i coupling occurs only at a saturated concentration of isoproterenol (Iso) in cardiac cells. Stimulation induces the PKA-mediated phosphorylation of the ␤ 2 AR over a wide range of agonist concentrations, whereas the GRK-mediated phosphorylation occurs only at sufficiently high concentrations (Ն100 nM) of Iso and is necessary for receptor internalization. We further demonstrate that mutations of the specific PKA or GRK phosphorylation sites on the ␤ 2 AR abolish receptor coupling to G i protein. Our data suggest that the PKA-mediated phosphorylation is necessary for ␤ 2 AR/G i coupling, but a sufficient receptor/G i coupling is also dependent on the GRK-mediated phosphorylation and subsequent trafficking of the receptors.

Site-directed Mutagenesis and Recombinant Adenoviruses-
pcDNA3.1-FLAG-tagged murine ␤ 2 AR was used as a template for mutagenesis. A protein kinase A mutant (PKAmut) ␤ 2 AR lacking protein kinase A (PKA) phosphorylation sites (serines 261, 262, 345, and 346) was generated by replacing the four serines with alanines. A GRK mutant (GRKmut) ␤ 2 AR lacking the GRK phosphorylation sites (serines 355, 356, and 358) was also created by replacing serines with alanine. These phosphorylation sites are determined by Tran et al. (13). Both ␤ 2 AR mutants were sequenced and transformed into adenovector for producing adenoviruses, as described previously (14). The titers of the viruses were assessed by comparison of receptor expression levels with both Western blot and ligand binding, as described previously (15). The GFP-G i CT virus used in contraction assays was a kind gift from Dr. Walter Koch (Thomas Jefferson University, Philadelphia, PA).
Cardiac Myocyte Contraction Assay-Measurement of spontaneous myocyte contraction was performed as described previously (16). Neonatal cardiac myocytes were isolated from mice lacking either ␤ 1 AR (␤ 1 AR-knock-out; ␤ 1 AR-KO) or both ␤ 1 and ␤ 2 AR genes (␤ 1 ␤ 2 AR-KO). ␤ 1 ␤ 2 AR-KO myocytes were infected with murine ␤ 2 AR, PKAmut ␤ 2 AR, or GRKmut ␤ 2 AR at a multiplicity of infection of 100 and stimulated with 10 M Iso. Cardiac myocytes were treated with G i inhibitor pertussis toxin (PTX) at 0.3 g/ml, 3 h before stimulation. GFP-G i CT virus was co-infected at a multiplicity of infection of 100 or, as indicated, along with the wild type or mutant ␤ 2 AR viruses. Contraction rate assays were performed the next day as described previously (16).
Receptor Phosphorylation-Neonatal cardiac myocytes were infected with FLAG-tagged wild type or mutant ␤ 2 ARs at a multiplicity of infection of 100. After 48 h of expression, cells were serum-starved for 2 h and stimulated with different doses of Iso for 5 min. For the time course experiments, cells were stimulated with 10 M Iso for different times as indicated. Cells were placed into lysis buffer (10 mM Tris, pH 8.0, 1% Nonidet P-40, 150 mM NaCl, 2 mM EDTA, and protease and phosphatase inhibitor mixture (Thermo Scientific)) for 30 min at 4°C. Samples were centrifuged, and the supernatant was resolved by SDS-PAGE for Western blotting. The ␤ 2 AR phosphorylation by PKA at Ser(P) 261/262 was detected with a monoclonal antibody from Dr. Richard Clark (University of Texas Health Science Center, Houston, TX). The ␤ 2 AR phosphorylation by GRK at Ser(P) 355/356 was detected with a polyclonal antibody from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). IRDye 680CW goat antimouse and IRDye 800CW goat anti-rabbit secondary antibodies were used to detect Ser(P) 261/262 and Ser(P) 355/356 phosphorylation, respectively. Western blots were visualized with an Odyssey scanner (Li-cor Biosciences, Lincoln, NE) and quantified before normalization against the base-line levels. A and B, ␤ 1 AR-KO cardiac myocytes expressing the PKA biosensor AKAR2.2 were stimulated with 10 nM (A) or 10 M (B) Iso. Myocytes were treated with G i protein inhibitor PTX for 2 h before stimulation to examine the effect on PKA FRET activity induced by agonist. C and D, ␤ 1 AR-KO cardiac myocytes were stimulated with 10 nM (C) or 10 M (D) Iso with and without PTX pretreatment, and the increases in contraction rate over base-line levels were measured as a function of time. E, ␤ 1 AR-KO myocytes were stimulated with 10 nM or 10 M Iso with and without PTX treatment. The phosphorylation of phospholamban was detected and normalized against total phospholamban (n ϭ 4). **, p Ͻ 0.01 when compared with the group without PTX treatment by two-way ANOVA. ***, p Ͻ 0.001 when compared with the group without PTX treatment by Student's t test.
Receptor Internalization and Recycling-After expressing the FLAG-␤ 2 AR, the ␤ 1 ␤ 2 AR-KO myocytes were serum-starved for 2 h and stimulated with 10 nM or 10 M Iso for different times. For receptor recycling, cells were stimulated with Iso for 10 min, rinsed, and refed with serum-free medium for different times. Cells were then fixed, permeabilized, and stained with anti-FLAG antibody, which was revealed with an Alexa-488 conjugated goat anti-mouse antibody (Invitrogen). Fluorescence images were taken with a CCD camera on a Zeiss microscope with Metamorph software. HEK293 cells were transfected with plasmids expressing the FLAG-␤ 2 AR, PKAmut ␤ 2 AR, and GRKmut ␤ 2 AR. Receptor internalization and recycling was quantified in HEK293 cells by using a fluorescence-linked immunosorbent assay (FLISA) to determine cell surface receptor levels as previously described (17). In brief, the FLAG-tagged receptors were labeled with anti-FLAG M1 antibody, which was revealed by the Alexa-488 goat-anti-mouse secondary antibody and detected by Spectramax M2 fluorometer.
PKA Activity Measurement through FRET Assay-␤ 1 AR-KO or ␤ 1 ␤ 2 AR-KO cardiac myocytes were infected with virus expressing A-kinase activity reporter (AKAR2.2) as previously described (18). For ␤ 1 ␤ 2 AR-KO myocytes, cells were also infected with receptor virus for overnight expression. The procedure for PKA FRET activity was carried out as previously described (18).
Statistical Analysis-Curve fitting and statistical analyses were performed using Prism software (GraphPad Software, Inc. San Diego, CA).

Stimulation of the ␤ 2 AR Induces an Agonist Concentration-dependent Receptor/G i Coupling in Cardiac
Myocytes-By activation of the endogenous ␤ 2 AR with Iso in ␤ 1 AR-KO cardiac myocytes, we analyzed the receptor-induced PKA activity through a FRET assay using the fluorescent biosensor AKAR2.2. Activation of the ␤ 2 AR with 10 nM Iso induced a transient peak in PKA activity, which returned to base-line levels within 5 min (Fig. 1A). Pretreatment with PTX, which specifically inhibits G i proteins, did not significantly change the ␤ 2 AR-induced FRET response. However, a saturated concentration of 10 M Iso resulted in a more sustained PKA FRET response, which was further enhanced with PTX pretreatment (Fig. 1B). This finding indicates that 10 M Iso, but not 10 nM Iso, selectively induces ␤ 2 AR/G i coupling for PKA activity in cardiac myocytes. We then used a cardiac contraction rate assay to examine the physiological consequences of this dose-dependent receptor/G i protein coupling. We found that at 10 nM Iso, the ␤ 2 AR-induced increase in contraction rate was not PTXsensitive (Fig. 1C), whereas PTX significantly enhanced the increase in contraction rate induced by 10 M Iso (Fig. 1D). These data again suggest that the ␤ 2 AR couples to G i proteins only after stimulation with high concentrations of agonist. Consistent with this notion, the PKA-mediated phosphorylation of phospholamban, a PKA substrate that is critical for the ␤AR-induced cardiac contractile response, was enhanced by PTX only after stimulation with 10 M Iso (Fig. 1E). Previous reports have suggested that the PKA-mediated phosphorylation of the ␤ 2 AR enhances the receptor coupling to G i protein (6,7), and the GRK-dependent phosphorylation of the ␤ 2 AR occurs only after stimulation with high concentrations of agonist (13), making the phosphorylation of ␤ 2 AR a potential regulator of this dose-dependent G i coupling in cardiac myocytes.

Agonist Stimulation Induces a Dose-dependent Distinct PKA and GRK Phosphorylation for ␤ 2 AR Internalization and Signaling in Cardiac
Myocytes-Analysis of the phosphorylation of ␤ 2 AR in cardiac myocytes showed that the PKA-dependent phosphorylation at serines 261/262 occurred at subnanomolar concentrations of Iso and displayed a dose-dependent increase (Fig. 2, A and B). In contrast, the ␤ 2 AR phosphorylation by GRK at serines 355/356 occurred only at 100 nM Iso or at higher doses (Fig. 2, A and C). After stimulation with saturated 10 M Iso, the kinetics of the GRK-and PKA-mediated phosphorylation of the ␤ 2 AR differed (Fig. 2D). The PKA-mediated phosphorylation showed a rapid increase, followed by a slow decrease, whereas the GRK-mediated phosphorylation was slower but more sustained over time (Fig. 2, E and F). It has been well documented that the GRK-mediated phosphorylation is involved in the endocytosis of ␤ 2 AR in fibroblasts (19,20), prompting us to explore the role of ␤ 2 AR phosphorylation in receptor internalization in cardiac myocytes. The activated ␤ 2 ARs showed little punctuate staining upon stimulation with 10 nM Iso in cardiac myocytes (Fig. 3A). However, stimulation with 10 M Iso induced a time-dependent clustering of ␤ 2 ARs in the cells, similar to those reported previously (8), which is a characteristic of receptor internalization (Fig. 3A). Moreover, the removal of Iso after 10 min of stimulation allowed the internalized receptor to return to the cell surface (Fig. 3A). These observations were further confirmed by quantification with a FLISA assay (Fig. 3B).
We then directly tested the role of phosphorylation in receptor internalization and signaling by using mutant ␤ 2 ARs with the PKA or GRK phosphorylation sites replaced by alanines. The expressions of both PKAmut and GRKmut were quantified and equalized through Western blot and ligand binding (supplemental Fig. S1). Western blots of the mutant ␤ 2 ARs also confirmed abolishment of the specific receptor phosphorylation events upon stimulation with 10 M Iso (supplemental Fig.  S2). Immunostaining of the wild type and PKAmut ␤ 2 ARs FIGURE 3. Stimulation of the ␤ 2 AR induces agonist dose-dependent internalization in cardiac myocytes. A, the ␤ 1 ␤ 2 AR-KO cardiac myocytes expressing FLAG-␤ 2 AR were stimulated with 10 nM or 10 M Iso for 10 or 30 min or 10-min Iso stimulation followed by drug removal for 30 or 60 min. The FLAG-␤ 2 ARs were visualized through immunofluorescence imaging. These data are representative of at least three experiments. B, the cell surface FLAG-␤ 2 ARs were quantified by FLISA in HEK293 cells after stimulation up to 10 or 30 min or after 10 min Iso stimulation followed by drug removal for 30 or 60 min. Receptor levels were normalized against base-line levels (n ϭ 5). **, p Ͻ 0.01 by Student's t test. Con, control.
showed time-dependent internalization of the receptors after stimulation with 10 M Iso (Fig. 4A). However, the GRKmut ␤ 2 ARs failed to sufficiently internalize under Iso stimulation. Washout of Iso after 10 min of stimulation showed that the internalized receptors underwent recycling to the cell surface (Fig. 4B). Quantification through the FLISA assay confirmed that the wild type and PKAmut ␤ 2 ARs but not the GRKmut ␤ 2 ARs were significantly internalized upon 10 min of stimulation with 10 M Iso (Fig. 4C), supporting the necessary role of GRK phosphorylation in the internalization of the ␤ 2 AR in cardiac myocytes. Together, the characterization of the mutant ␤ 2 ARs revealed distinct actions of PKA and GRK on the receptor phosphorylation and trafficking in cardiac myocytes upon agonist stimulation, deeming the phosphorylation-deficient mutants suitable for closer examination on the effects on the receptor signaling in these cells.
By introducing either wild type or mutant ␤ 2 AR into ␤ 1 ␤ 2 AR-KO cardiac myocytes, we were able to examine the physiological effects of the phosphorylation of the ␤ 2 AR upon receptor activation. Stimulation of the PKAmut ␤ 2 AR induced a stronger contraction rate increase over that of the wild type (Fig. 5A). Stimulation of the GRKmut ␤ 2 AR showed a contraction rate increase similar to that of the wild type, but this increase was much more sustained (Fig. 5A). Although the base-line contraction rates between the wild type and mutant ␤ 2 ARs were not significantly different (Fig. 5B), the maximal

. Stimulation of wild type and mutant ␤ 2 ARs induces different receptor internalization and recycling.
A, the FLAG-tagged ␤ 2 AR, PKAmut ␤ 2 AR, and GRKmut ␤ 2 AR expressed in cardiac myocytes were stimulated with 10 M Iso for different times before they were visualized by immunofluorescence imaging. B, the wild type and mutant ␤ 2 ARs were visualized for recycling after stimulation for 10 min, followed by drug removal for the indicated times. These data are representative of at least three experiments. C, the cell surface wild type and mutant ␤ 2 ARs were quantified through FLISA after 10 min of Iso stimulation and recycling after drug removal for 30 and 60 min in HEK293 cells (n ϭ 3). Receptor levels were normalized against base-line levels. **, p Ͻ 0.01 by Student's t test. Con, control. contraction rate increase induced by the PKAmut ␤ 2 AR was significantly higher than that of wild type (Fig. 5C). These data strongly suggest that the phosphorylation of the ␤ 2 AR has significant consequences in regulating myocyte contraction rate responses.

PKA and GRK Phosphorylation Differentially Regulate ␤ 2 AR/G i Coupling in Cardiac Myocytes-By
analyzing PKA activity in cardiac myocytes induced by these mutant ␤ 2 ARs, we were able to observe the kinetics of upstream signaling, which leads to contraction induced by the receptors. Stimulation of the wild type, PKAmut ␤ 2 AR, or GRKmut ␤ 2 AR with 10 M Iso induced a stronger FRET increase in PKA activity, although the increase induced by the GRKmut ␤ 2 AR was less than those of the wild type and PKAmut receptors (Fig. 6A). Inhibition of G i proteins with PTX pretreatment significantly enhanced the PKA FRET increase induced by activation of the wild type ␤ 2 AR (Fig. 6B). However, the FRET response induced by the PKAmut ␤ 2 AR was insensitive to PTX (Fig.  6C), indicating that this receptor was unable to couple to G i protein.
Interestingly, the FRET response induced by the GRKmut ␤ 2 AR was also PTX-insensitive (Fig. 6D), indicating that the GRK-dependent phosphorylation was necessary for ␤ 2 AR coupling to G i protein.
Extending the PKA FRET results into the contraction rate assay reinforced the physiological effects of this phosphorylation-dependent G protein-coupling phenomenon. Although pretreatment with PTX significantly enhanced the wild type ␤ 2 AR-induced increase in contraction rate upon stimulation with 10 M Iso (Fig. 7, A and E), it did not significantly affect the PKAmut ␤ 2 AR-induced increase in contraction rate (Fig. 7, B and E). Moreover, PTX did not enhance the GRKmut ␤ 2 AR-induced increase in contraction rate; instead, it lowered the contraction rate increase (Fig. 7, C  and E), which could be in part due to the enhanced base-line contraction rate by PTX treatment (Fig. 7D). These data suggest that both PKA and GRK phosphorylation are required for ␤ 2 AR coupling to G i protein, which thereafter affects myocyte contraction rate response.
We further employed a specific inhibitor of the G i protein consisting of the C-terminal portion of G␣ i2 (G i CT) (21) to probe receptor/G i coupling by either wild type or mutant ␤ 2 AR.

FIGURE 5. Activation of wild type and mutant ␤ 2 ARs induces distinct responses in contraction rates in cardiac myocytes.
A, the ␤ 1 ␤ 2 AR-KO cardiac myocytes expressing ␤ 2 AR, PKAmut ␤ 2 AR, or GRKmut ␤ 2 AR were stimulated with 10 M Iso for contraction rate measurement. The base-line levels of contraction rates (B) and the maximal increases in contraction rate after stimulation (C) were plotted. ***, p Ͻ 0.001 when compared with the wild type ␤ 2 AR by two-way ANOVA. *, p Ͻ 0.05 by Student's t test. FIGURE 6. Wild type and mutant ␤ 2 ARs differentially couple to G i proteins to regulate agonist-induced PKA FRET responses. A, the ␤ 1 ␤ 2 AR-KO cardiac myocytes expressing the PKA FRET biosensor AKAR2.2 together with the ␤ 2 AR, PKAmut ␤ 2 AR, or GRKmut ␤ 2 AR were stimulated with 10 M Iso as indicated. The PKA FRET ratio was measured. B-D, myocytes were also treated with PTX to access the receptor/G i coupling in regulating PKA FRET response after Iso stimulation. **, p Ͻ 0.01; ***, p Ͻ 0.001 when compared with the wild type receptor by two-way ANOVA. YFP, yellow fluorescent protein; CFP, cyan fluorescent protein.
We first determined that infection with G i CT virus at a multiplicity of infection of 100 optimally inhibited receptor/G i coupling and enhanced the ␤ 2 AR-induced increase in contraction rate (Fig.  8A). We then introduced G i CT along with wild type or mutant ␤ 2 AR into ␤ 1 ␤ 2 AR-KO cardiac myocytes. Overexpressing G i CT significantly enhanced the wild type ␤ 2 AR-induced increase in contraction rate (Fig. 8, B and F), which was similar to that obtained with PTX treatment. However, G i CT did not affect the PKAmut ␤ 2 AR-induced increase in contraction rate (Fig. 8, C and  F). Similar to that of PTX treatment (Fig. 7C), inhibition of G i protein with G i CT also reduced the GRKmut ␤ 2 AR-induced increase in contraction rate (Fig. 8, D and F). Overexpression of G i CT did not affect the base-line contraction rates (Fig. 8E). Together, we showed that inhibition of G i with G i CT significantly enhanced the maximal contraction rate induced by the wild type but not by the PKAmut and GRKmut ␤ 2 ARs.

DISCUSSION
By measuring PKA activity and cardiac contraction rate increase, we show for the first time that ␤ 2 AR activation of G i proteins occurs selectively at high concentrations but not at low concentrations of agonist. Moreover, we show that the receptor coupling to G i proteins is differentially regulated by the PKA-and GRK-mediated phosphorylation of the activated ␤ 2 ARs. At both low and high concentrations of agonist, the activated ␤ 2 ARs undergo the PKA-mediated phosphorylation. In comparison, only high concentrations of agonist induce the GRK-mediated phosphorylation of the ␤ 2 ARs for subsequent internalization, which is also necessary for sufficient receptor coupling to G i proteins (8,9). Our studies link together various components of the ␤ 2 AR, including receptor phosphorylation, receptor trafficking, and differential receptor/G protein coupling in cardiac cells, which may allow the activation of the ␤ 2 AR signaling pathway to function as either a stimulatory or protective mechanism for cardiac cells under different levels of stress.
It has been well documented that the PKA-mediated phosphorylation of ␤ 2 AR enhances the receptor coupling affinity to G i protein in a reconstituted system, and mutation of the PKA phosphorylation sites blocks receptor coupling to G i in HEK293 fibroblasts (6,7). In this study, we have further resolved the necessity of the PKA-dependent phosphorylation of ␤ 2 AR in coupling to G i proteins in cardiac myocytes. Activation of the murine ␤ 2 AR by both low and high concentration of isoproterenol leads to phosphorylation of the receptor by PKA. At high concentrations of isoproterenol, the phosphorylation levels of the ␤ 2 AR by PKA are further enhanced in comparison with those at low concentrations. This is probably due to dissociation of phosphodiesterase 4D isoforms from the activated receptors (data not shown), which leads to higher local PKA activity for receptor phosphorylation. However, under stimulation with low concentrations of isoproterenol, the activated ␤ 2 ARs do not couple to G i , even if the PKA-mediated phosphorylation occurs (Figs. 1 and 2). However, disruption of the PKA phosphorylation sites on the ␤ 2 AR with mutagenesis blocks the receptor coupling to G i (Figs. 6 -8). Together, these data indicate that PKA phosphorylation is necessary but not sufficient for the activated ␤ 2 AR coupling to G i in cardiac myocytes.
In contrast, the ␤ 2 ARs undergo GRK-dependent phosphorylation only under stimulation with high concentrations of Iso (Fig. 2). Disruption of the GRK sites prolongs the Iso-induced FIGURE 7. Wild type and mutant ␤ 2 ARs differentially couple to G i proteins to regulate agonist-induced contraction rate increases. A-C, the ␤ 1 ␤ 2 AR-KO cardiac myocytes expressing the ␤ 2 AR, PKAmut ␤ 2 AR, or GRKmut ␤ 2 AR were stimulated with 10 M Iso in the presence or absence of PTX. The response in myocyte contraction rate was measured. D and E, the base-line levels of contraction rates and the maximal increases in contraction rates after stimulation in the presence or absence of PTX were plotted. ***, p Ͻ 0.001 when compared with the group without PTX treatment by two-way ANOVA. *, p Ͻ 0.05 when compared with the group without PTX treatment in Student's t test.
myocyte contraction rate response (Fig. 5A), which is consistent with the notion that the phosphorylation of ␤ 2 AR by GRK is required for rapid receptor desensitization (22). We do not observe obvious change in the overall PKA activity induced by the GRKmut (Fig. 6A), probably because the change is small and occurs in the local environment of the receptor and thus is not sensed by PKA probes presented throughout the cytosol. However, similar to those of the PKAmut, the PKA FRET and contraction rate responses induced by the GRKmut ␤ 2 AR do not show sensitivity to the inhibition of G i protein (Figs. 6 -8), indicating that the GRK-mediated phosphorylation is required for sufficient coupling to G i . One possibility is that the GRK-phosphorylated receptors undergo internalization to promote the access of the receptor to G i protein. Another possibility is that the GRK-mediated phosphorylation of the ␤ 2 AR directly enhances the binding affinity of the receptor to G i protein.
However, evidence so far argues that the GRK phosphorylationdependent receptor trafficking is more likely the key for sufficient receptor/G i coupling. To support this notion, either blocking receptor internalization or disrupting receptor recycling through point mutation on the PDZ motif of the ␤ 2 AR blocks the receptor/G i coupling, which yields enhanced myocyte contraction responses (9,15). Moreover, Wang et al. (8) have shown that direct inhibition of GRK2 activity blocks the receptor coupling to G i proteins in myocytes. Furthermore, the GRK phosphorylation-dependent recruitment of arrestin-PDE4D complexes to the activated ␤ 2 AR has been implicated in receptor/G i coupling (23). These studies, together with our data, suggest that the GRKmediated phosphorylation of ␤ 2 AR and subsequent receptor trafficking are critical for sufficient access of the receptor to G i proteins. Consistent with this notion, a G i -specific inhibitor, G i CT, selectively blocks the receptor coupling to G i without affecting the receptor coupling to G s . These data suggest that the G i CT peptide did not compete against the unphosphorylated receptors in binding to G s protein but blocks the phosphorylated receptor in binding to G i protein, which reinforces the idea that ␤ 2 AR/G i coupling is dependent on the agonist-induced phosphorylation of the receptor. Together, our data suggest that although the PKA-mediated phosphorylation is required for high affinity binding to G i protein, the GRK-mediated phosphorylation is necessary for receptor trafficking that appears to control the accessibility of ␤ 2 AR to G i protein in cardiac myocytes.
Cardiac contraction in response to adrenergic stimulation is a tightly controlled process to ensure steady increases in cardiac output without overstimulation. Our observation of an agonist concentration-dependent switch of the activated ␤ 2 AR coupling from G s and G i may be critical for such fine regulation. At low concentrations of catecholamines, the activated receptors will couple only to G s for a stimulatory effect on cAMP/PKA activities for increasing cardiac contraction. However, in the presence of high concentrations of catecholamines, the activated ␤ 2 ARs will switch from G s to G i to reduce cAMP/PKA activities in cardiac tissues while preventing the overstimulation of cardiac contraction. The mechanism for how high concentrations of agonists can induce receptor phosphorylation by different kinases for subsequent switch of G protein coupling is currently not known. Structural studies reveal that higher concentrations of agonists induce conformational changes distinct FIGURE 8. G i CT inhibits ␤ 2 AR/G i coupling for myocyte contraction rate response upon agonist stimulation. A, the maximal responses in contraction rate induced by the ␤ 2 AR were enhanced by expression of G i CT along with the wild type ␤ 2 AR in ␤ 1 ␤ 2 AR-KO cardiac myocytes. B-D, the ␤ 1 ␤ 2 AR-KO cardiac myocytes expressing the ␤ 2 AR, PKAmut ␤ 2 AR, or GRKmut ␤ 2 AR were stimulated with 10 M Iso in the presence or absence of G i CT. The response in myocyte contraction rate was measured. E and F, the base-line levels of contraction rates and the maximal increases in contraction rates after stimulation in the presence or absence of G i CT were plotted. **, p Ͻ 0.01; ***, p Ͻ 0.001 when compared with the wild type receptor by two-way ANOVA. *, p Ͻ 0.05 when compared with the group without G i CT in Student's t test.
from those induced by low concentrations (17), which may enhance the receptor binding affinity to GRK. Alternatively, low concentrations may selectively activate the G s protein-precoupled pool of receptors with high affinity binding, whereas higher concentrations of agonists activate both precoupled and non-coupled pools of receptors. The latter pool of receptors may be involved in coupling to G i proteins. The mechanism thus remains to be addressed. This protective mechanism by receptor/G i coupling may be essential for preventing a high frequency of contractions, such as fibrillation during stress. In addition, ␤ 2 AR/G i coupling can play a protective role against the apoptotic effects induced by adrenergic/G s stimulation (24 -27).
More than a simple extension of the biochemical characterization of the ␤ 2 AR, this study provides insightful information on the physiological effects of receptor phosphorylation by PKA and GRK. More importantly, it opens up questions about the nature of phosphorylation and its role in receptor desensitization, a concept that has been established in the last couple of decades. Undoubtedly, phosphorylation seems to play a role in diminishing the effects of receptor activation in cardiac myocytes, since abolishing those phosphorylation sites enhanced the contraction rate over the wild type upon stimulation (Fig.  4). However, we find that receptor phosphorylation appears to be critical in regulating contraction through coupling to different G proteins to protect cardiac cells from overwhelming agonist stimulation. Our study provides a clear step toward understanding the nature of receptor regulation by PKA-and GRK-mediated phosphorylation in a physiological context.