c-Src tyrosine kinase binds the beta 2-adrenergic receptor via phospho-Tyr-350, phosphorylates G-protein-linked receptor kinase 2, and mediates agonist-induced receptor desensitization.

The nonreceptor tyrosine kinase Src has been implicated in the switching of signaling of beta2-adrenergic receptors from adenylylcyclase coupling to the mitogen-activated protein kinase pathway. In the current work, we demonstrate that Src plays an active role in the agonist-induced desensitization of beta2-adrenergic receptors. Both the expression of dominant-negative Src and treatment with the 4-amine-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d]pyrimidine (PP2) inhibitor of Src kinase activity blocks agonist-induced desensitization. Agonist triggers tyrosine phosphorylation of the beta2-adrenergic receptor and recruitment and activation of Src. Because phosphorylation of the Tyr-350 residue of the beta2-adrenergic receptor creates a conditional, canonical SH2-binding site on the receptor, we examined the effect of the Y350F mutation on Src phosphorylation, Src recruitment, and desensitization. Mutant beta2-adrenergic receptors with a Tyr-to-Phe substitution at Tyr-350 do not display agonist-induced desensitization, Src recruitment, or Src activation. Downstream of binding to the receptor, Src phosphorylates and activates G-protein-linked receptor kinase 2 (GRK2), a response obligate for agonist-induced desensitization. Constitutively active Src increases GRK phosphorylation, whereas either expression of dominant-negative Src or treatment with the PP2 inhibitor abolishes tyrosine phosphorylation of GRK and desensitization. Thus, in addition to its role in signal switching to the mitogen-activated protein kinase pathway, Src recruitment to the beta2-adrenergic receptor and activation are obligate for normal agonist-induced desensitization.

The nonreceptor tyrosine kinase family of Src functions in a wide spectrum of cell signaling (1). The recruitment and activation of Src kinases is obligate for the mediation of Ras activation by G-protein-linked receptors (GPLRs) 1 (2,3). GPLRs display four well known responses to the stimulation by agonists: namely, activation, desensitization, internalization, and eventual resensitization (4,5). The activation of adenylylcyclase in response to agonist stimulation of ␤2adrenergic receptors results in elevation of intracellular cyclic AMP levels, activation of protein kinase A, and later phosphorylation of the receptor by protein kinase A and G-protein-linked receptor kinases (GRKs) (6). Phosphorylation leads to desensitization and receptor sequestration, dependent upon the binding of ␤-arrestin to the phosphorylated receptor as well as to clathrin (7,8). Internalization occurs via clathrin-mediated processes (9), and resensitization/dephosphorylation follows by the action of the protein phosphatase 2A (10) or 2B (11). These activities are organized by the protein kinase A-anchoring protein AKAP250 or gravin (12). Gravin acts as a scaffold for interactions among protein kinase A, protein kinase C, PP2B, and the ␤2-adrenergic receptor (13).
Src has been shown to participate in the formation of ␤2adrenergic receptor-Src complexes, in association with ␤-arrestin, to switch the signaling from adenylylcyclase to activation of Ras and its downstream effector, the mitogen-activated protein kinase (3). Src has been reported to phosphorylate and activate GRKs (14), observations confined to systems in which the elements are overexpressed at high levels or examined in reconstituted systems, leaving unanswered the question whether or not Src interacts directly with the ␤2-adrenergic receptor and functions in desensitization. The proposed model for Src action defines a temporal sequence in which GRK2 phosphorylates the ␤2-adrenergic receptor, and an Src-arrestin complex then targets to the receptor, facilitating internalization and signaling to Ras (3). In the current work we reveal that Src targets a conditional SH2-binding site of the ␤2-adrenergic receptor, is recruited to the phospho-Tyr-350 receptor, is activated, and then phosphorylates GRK2. This temporal sequence is shown to be obligate for agonist-induced desensitization.
Suppression via Antisense Oligodeoxynucleotides-Antisense, sense, and control missense oligodeoxynucleotides with the same base composition, but in scrambled order, were synthesized and purified to cell * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Epifluorescence Imaging-Microscopy of live cells stably transfected to express a green fluorescent protein-tagged ␤-adrenergic receptor fusion protein was performed on the Eclipse TE300 (Nikon) inverted microscope equipped with a ϫ 40 objective and a set of filters (13). Images were acquired using MicroMAX Imaging System (Princeton Instruments Inc.) and WinView32 software. Fluorescent dyes were imaged sequentially in frame-interface mode to eliminate spectral overlapping between the channels (12).
Radioligand Binding Assay-The number of ␤ 2 -adrenergic receptors was determined by radioligand binding. Approximately 10 6 stably transfected CHO cells were incubated with 30 pM [ 125 I]iodocyanopindolol (PerkinElmer Life Sciences) in the absence or presence of 10 M propranolol (for defining the amount of nonspecific binding) at 23°C for 90 min. The incubation buffer contained 50 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , and 150 mM NaCl (16). The cells were collected on GF/C binderfree glass giber filter membranes at reduced pressure and washed rapidly. The radioligand bound to the washed cell mass retained by the filter was quantified by use of a ␥-counter. Nonspecific binding was less than 5% of radioligand binding. Each experimental point was performed in triplicate.
Cyclic AMP Assay-One day prior to the analysis, A431 cells and CHO clones stably transfected with ␤2-adrenergic receptor and ␤2adrenergic receptor-GFP-pcDNA3 were seeded in 96-well plates at the density of 20,000 cells/well. Cells were washed and challenged with or without 10 M isoproterenol in 50 l of HEM buffer (20 mM Hepes, pH 7.4, 135 mM NaCl, 4.7 mM KCl, 1.2 mM MgSO 4 , 2.5 mM NaHCO 3 , containing Ro-20 -1724 (0.1 mM; Calbiochem)) and adenosine deaminase (0.5 unit/ml) for the times indicated in the figure legends. The first challenge with agonist was for 0 -60 min. For some studies, the cells first challenged with agonist were washed free of agonist (three times) at the end of the first incubation and then challenged again for 5 min with agonist. The incubations were stopped by addition of 100 l of ice-cold ethanol. The agonist-stimulated cyclic AMP production was determined as described elsewhere (15).

RESULTS
The nonreceptor tyrosine kinase Src has been shown to play a prominent role in linking G-protein-linked receptors to Ras and downstream signaling to the mitogen-activated protein kinase network (2,3). For the ␤2-adrenergic receptor, Src has been shown to act as a switch from receptor regulation of Gs and adenylylcyclase to activation of Ras following the normal progression of receptor activation, desensitization, and sequestration. The role of Src in desensitization is far less clear. Src has been proposed to target to GPLRs via its interaction with ␤-arrestin, suggesting in a temporal sense that Src would be functioning in post-receptor desensitization. To test the hypothesis that Src may be participating in events prior to receptor sequestration and Ras activation, we focused on an analysis of ␤2-adrenergic receptor-mediated activation and desensitization, using the accumulation of intracellular cyclic AMP as the readout.
Human epidermoid A431 carcinoma cells are a widely used model of ␤2-adrenergic receptor action (18,19). Challenging A431 cells with the ␤2-adrenergic agonist isoproterenol (10 M) leads to peak accumulation of cyclic AMP at 5 min, which declines to approximately half-maximal levels within 30 -60 min, reflecting ongoing agonist-induced desensitization (Fig. 1). The cyclic AMP response of A431 cells to agonist, following a 30 or 60 min first challenge with isoproterenol, a washout of agonist, and a second challenge with isoproterenol for 5 min, is rectified to the level of naïve cells that have been challenged with agonist for 5 min. To explore what role, if any, Src may have in agonistinduced desensitization of ␤2-adrenergic receptors, we performed the same experiments in A431 cells pretreated for 30 min with the Src family tyrosine kinase inhibitor PP2 (Fig. 1). In the presence of the Src inhibitor, agonist-induced desensitization was abolished, and intracellular cyclic AMP levels continued to increase over the period of 60 min of challenge with agonist. These data suggest that Src activity is important not only for subsequent downstream signaling to Ras but also for agonistinduced desensitization itself. To test the observations by an independent approach that targets Src only, A431 cells were treated for 4 days with ODNs antisense to Src to suppress Src levels. Treatment with ODNs antisense to Src also abolished agonist-induced desensitization (Fig. 1). Suppression of Src levels by antisense ODNs yields increased cyclic AMP accumulation in A431 cells challenged with isoproterenol for up to 60 min. Treatment of cells with ODNs sense or missense to Src, in contrast, was without effect (data not shown).
To further test this newly discovered role for Src in agonistinduced desensitization of ␤2-adrenergic receptors, we tested the effects of stable transfection of A431 cells with an expression vector harboring either the constitutively active form (Y527F) or the dominant-negative form (K295R/Y527F) of c-Src (Fig. 2). Stable expression of the dominant-inhibitory form of Src abolishes agonist-induced desensitization of ␤2-adrenergic receptor with respect to cyclic AMP accumulation. The accumulation of intracellular cyclic AMP increased throughout the time course of ␤-adrenergic stimulation, for up to 60 min, in the clones expressing the dominant-negative form of Src. This observation confirms the results obtained with cells treated with the Src family inhibitor PP2 and with ODNs antisense to suppress Src levels ( Fig. 1). Thus, inhibition of Src activity, suppression of Src expression, and expression of a dominantinhibitory form of Src all lead to a loss of agonist-induced desensitization of ␤2-adrenergic receptor-mediated activation of adenylylcyclase. The cyclic AMP responses to isoproterenol in the cells with Src suppression or inhibition are the greatest cyclic AMP responses that we have observed with these cells to date. Expression of the constitutively active form (Y527F) of Src, in contrast, had no demonstrable effect on agonist-induced desensitization (Fig. 2). Although we expected that the expression of the constitutively active Src would further amplify the desensitization response, no such alteration was noted. This observation suggests that endogenous Src may be fully activated with respect to desensitization only when the ␤2-adrenergic receptors first have been activated. Under these circumstances, the introduction of constitutively active Src without proper targeting to ␤ 2 -adrenergic receptor, as in the basal state, may have little effect on desensitization.
We tested the hypothesis that agonist treatment of A431 cells activated endogenous Src activity in A431 cells (Fig. 3). Cells were challenged with isoproterenol (10 M) for periods up to 60 min. At the conclusion of the activation, whole-cell extracts were prepared, and the extracts were subjected to an immune precipitation "pull down" assay using antibodies against the ␤2-adrenergic receptor (antibody CM04). The immune precipitations were subjected to SDS-PAGE analysis, and the resultant immunoblots were stained with antibodies to c-Src. Some Src was found in association with the ␤2-adrenergic receptor under basal, unstimulated cell culture conditions, probably reflecting the observation that growth factors such as insulin catalyze the phosphorylation of the ␤2-adrenergic receptor on residue Tyr-350, which then constitutes a canonical SH2-binding site (20 -22). In response to challenge with isoproterenol, the level of Src association increased by more than 3.0-fold within 30 min and was maintained until at least 60 min. To determine whether the Src associated with the ␤2adrenergic receptor was activated, we performed immunoblotting of the Src associated with the ␤2-adrenergic receptor using an antibody that detects only Src phosphorylated at the Tyr-416 position (anti-phospho-Src (Tyr-416)). The results of the blotting for activated Src revealed that the Src associated with the ␤2-adrenergic receptor following the challenge with isoproterenol was phosphorylated at Tyr-416 and therefore activated. Thus, not only were Src expression and activity necessary for agonist-induced desensitization of ␤2-adrenergic receptors, but association and activation by this G-protein-linked receptor were also increased sharply in response to agonist.
The increase in the amount of Src associated with the ␤2adrenergic receptor, in tandem with the increase in Src activation, prompted the question what the basis was for this increased association and activity of Src with the receptor. The simplest hypothesis to explain the agonist-induced increase in Src association and activation was that the ␤2-adrenergic receptor might display increased phosphorylation of Tyr-350, a conditional canonical SH2-binding site. To test this possibility, we challenged A431 cells with isoproterenol (10 M) for periods up to 60 min and performed immune precipitation pull-downs of ␤2-adrenergic receptors from the whole-cell extracts (Fig. 4). The immune precipitates were subjected to SDS-PAGE, and the resultant immunoblots were stained with anti-phosphotyrosine antibody (PY99). These data show conclusively and un- c-Src-␤2-Adrenergic Receptor Signaling Complexes expectedly that the phosphotyrosine content of the ␤2-adrenergic receptor increased upon challenge of the cells with isoproterenol. Within 5 min of challenge with isoproterenol, the content of phosphotyrosine more than doubled, increasing to 2.5-and then 3.0-fold by 30 and 60 min, respectively. We confirmed these data using metabolic labeling of an independent hamster smooth muscle cell line (DDT1-MF2 vas deferens) with [ 32 P]P i overnight and stimulation with 10 M isoproterenol for 5 or 15 min (data not shown).
Earlier studies demonstrated both in vivo and in vitro that phosphorylation of the ␤2-adrenergic receptor at Tyr-350 generates an SH2-binding site that Grb2, 1-phosphatidylinositol 3-kinase, or dynamin can bind via SH2 domains (23,24). We tested whether Tyr-to-Phe substitution of the Tyr-350 residue of the ␤2-adrenergic receptor would alter association of the receptor with Src (Fig. 5). These studies were performed in CHO-K cells that express very low levels of endogenous ␤2-adrenergic receptors. Stably transfected CHO clones were created that express either the wild-type ␤2adrenergic receptor or the Y350F mutant form of the receptor. Clones were selected that express comparable levels of expression of ␤2-adrenergic receptors, 0.25 pmol of receptor/mg of membrane protein, as measured using the high affinity, radiolabeled ␤2-adrenergic antagonist ligand iodocyanopindolol. CHO clones were challenged with isoproterenol (10 M) for periods up to 60 min and then harvested for whole-cell extracts. The extracts were subjected to immune precipitation pull-down assays with antibodies to the ␤2adrenergic receptor (CM04). The immune precipitates were subjected to SDS-PAGE, and the resultant blots were stained with antibodies against Src as well as with a second antibody (CM02) to ensure equivalent loading of samples. As noted above with pull-down assays from whole-cell extracts prepared from A431 cells (Fig. 3), the amount of Src associated with the wild-type ␤2-adrenergic receptor increased in response to challenge of the cells with isoproterenol (Fig. 5). In sharp contrast to the situation observed for pull-down assays in cells expressing the wild-type receptor, the pull-down assays performed with whole-cell extracts of CHO clones expressing the Y350F mutant form of the ␤2-adrenergic receptor do not reveal an increase in Src association in response to agonist challenge. In fact, contrary to the increase shown in Src-␤2-adrenergic receptor association for the wild-type receptor, Src association with the Y350F mutant form of ␤2-adrenergic receptor actually declines in response to challenge with isoproterenol. Thus, the likely basis for the increase in Src-␤2-adrenergic receptor association and Src activation in response to agonist challenge is Src association with the phosphorylated Tyr-350 residue of the receptor that creates an SH2-binding site.
To test whether or not the agonist-induced increase in phosphotyrosine content of the ␤2-adrenergic receptor is indeed a reflection of the phosphorylation of the Tyr-350 residue, we challenged CHO clones stably expressing either the wild-type receptor or the Y350F mutant form of the receptor with isoproterenol for periods up to 60 min. Whole-cell extracts were prepared from both clones and then employed for pull-down assays for the ␤2-adrenergic receptor using the CM04 antireceptor antibody that recognizes an exofacial epitope of the receptor not influenced by the Tyr-350 residue localized to the cytoplasmic, C-terminal tail of the ␤2-adrenergic receptor. The immune precipitates were subjected to SDS-PAGE and immunoblotting and stained with either anti-phosphotyrosine PY99 antibody or anti-receptor CM02 antibody to establish equivalent loading of samples. For the clones expressing wild-type receptor, the phosphotyrosine content increased (Fig. 6), as observed above. In the cells expressing the Y350F mutant form of the ␤2-adrenergic receptor, however, phosphotyrosine content at 30 and 60 min post-challenge with isoproterenol was reduced in comparison with that obtained for wild-type receptors. An increase in phosphotyrosine content was observed at 5 min in the clones expressing the Y350F mutant form of the receptor, which presumably reflects increased phosphorylation of either Tyr-132 and Tyr-141 in the second intracellular loop or Tyr-354 and/or Tyr-364 in the cytoplasmic, C-terminal tail of the hamster ␤2-adrenergic receptor. Each of these tyrosine residues has been characterized earlier with respect to phosphorylation, but only the Tyr-350 site creates an SH2-binding site upon phosphorylation (20).
Using the cyclic AMP response to isoproterenol stimulation as a readout, we examined in parallel the effects of the Y350F mutation of the ␤2-adrenergic receptor on agonist-induced desensitization in these CHO clones (Fig. 7). Clones stably expressing either the wild-type or the Y350F mutant forms of the receptor were challenged with isoproterenol for up to 60 min. In some cases (30 and 60 min) the clones were then washed free of the agonist and then challenged a second time for 5 min with isoproterenol, and intracellular cyclic AMP accumulation was c-Src-␤2-Adrenergic Receptor Signaling Complexes measured. Agonist-induced desensitization in the CHO clones expressing the wild-type ␤2-adrenergic receptor was essentially complete within 30 min. Following the sharp increase in cyclic AMP accumulation observed in response to isoproterenol at 5 min post-challenge, the clones displayed desensitization. The cyclic AMP levels of CHO cells expressing the wild-type ␤2-adrenergic receptors declined in the continued presence of agonist. By 30 or 60 min of challenge with agonist, cyclic AMP levels of these clones declined to that of unstimulated CHO cells. Washout of the agonist, followed by a second 5-min challenge with agonist, enabled resensitization, the restimulation provoking a sharp rise in intracellular cyclic AMP accumulation, albeit a somewhat smaller response than that observed with the naïve cells.
Expression of the Y350F mutant form of the ␤2-adrenergic receptor does not alter the ability of the receptor to activate adenylylcyclase, because the cyclic AMP response of the mutant receptor is essentially equal to that observed for the clones expressing the wild-type receptor (Fig. 7). What is striking, however, is that the clones expressing the Y350F mutant form of the ␤2-adrenergic receptor display a desensitization response that is substantially reduced in magnitude from that observed in clones expressing the wild-type receptor. At both 30 and 60 min post-challenge with isoproterenol, clones expressing the wild-type receptor display complete desensitization, whereas the CHO clones expressing the Y350F mutant retain ϳ50% of the cyclic AMP response observed following a single 5-min challenge with agonist. Although the fundamental observations obtained from the studies of the A431 cells and CHO clones stably expressing wild-type receptors are the same, the cellular context does play a role in the magnitude of the biological readout of cyclic AMP accumulation.
Because phosphorylation of the ␤2-adrenergic receptor by protein kinase A and GRK2 precedes association of the phospho-receptor with ␤-arrestin and because ␤-arrestin acts as an adaptor for clathrin-mediated internalization of the ␤2-adrenergic receptor, we explored the phosphorylation state of GRK2 in A431 cells in the basal and isoproterenol-stimulated states (Fig. 8). Src has been shown to be capable of phosphorylating and increasing the activity of GRK2 (14). A proposed model of Src activity, however, suggests that Src is targeted to the receptor, not directly, as suggested by the prominent role of Tyr-350 and the SH2-binding site of the ␤2-adrenergic receptor created upon phosphorylation, but rather via ␤-arrestin, following the phosphorylation of the ␤2-adrenergic receptor by GRK2. As an alternative, we hypothesize that Src associates directly with, and is activated by, the phospho-Tyr-350 ␤2adrenergic receptor. Src then is activated through association with the ␤2-adrenergic receptor and phosphorylates GRK2, driving the desensitization response. The ␤2-adrenergic receptor phosphorylated by Src-activated GRK2 then favors association of ␤-arrestin and clathrin with the receptor complex. This hypothesis was tested by measuring the amount of phosphotyrosine content (phosphorylation) in GRK2 under the same conditions that were employed to test agonist-induced desensitization and sequestration. A431 wild-type and stably transfected clones were either untreated or challenged with isoproterenol (10 M) for periods of up to 60 min, and the whole-cell extracts were prepared for immune precipitation pull-down assays using antibodies specific for GRK2 (Fig. 8). The immune precipitates were subjected to SDS-PAGE, and the resultant immunoblots were stained with either anti-phosphotyrosine antibodies (PY99) or antibodies against GRK2 to establish equivalent loading. For the wildtype cells, GRK2 appears to have some phosphotyrosine content in the basal state. Stimulation with ␤2-adrenergic agonist leads to a sharp increase in phosphotyrosine content that peaks at 5 min (Ͼ5.0-fold) and then declines by 30 to 60 min postchallenge at a level about 3-fold greater than basal. Pretreatment (30 min) with the Src kinase family inhibitor PP2 (Fig. 8,  ϩPP2) was associated with a complete loss of the ␤2-adrenergic agonist-induced increase in the phosphotyrosine content of GRK2. In clones expressing the constitutively activated Src c-Src-␤2-Adrenergic Receptor Signaling Complexes mutant (Fig. 8, CA-Src), the phosphotyrosine content of GRK2 was sharply elevated in the basal state and equivalent to that observed in response to stimulation with agonist. In contrast, clones expressing the dominant-negative form of Src (Fig. 8, DN-Src) showed very little phosphotyrosine content in GRK2, either in the basal state or in response to stimulation with ␤2-adrenergic agonist.
Having established that Src plays an important role in agonistinduced desensitization of ␤2-adrenergic receptors, we also sought to explore whether Src participated in agonist-induced sequestration. A431 cells were stably transfected with a GFP-␤2adrenergic receptor fusion protein, so that receptor localization could be monitored in live cells by epifluorescence microscopy. The clones stably expressing the GFP-tagged ␤2-adrenergic receptor were stimulated for 30 min with isoproterenol (10 M) and examined for localization of the epifluorescence signal (Fig. 9). In the unstimulated, control situation, the GFP-tagged ␤2adrenergic receptor is localized predominantly to the plasma membrane (Fig. 9a). The cell membranes are well defined with GFP-tagged receptor (see arrows, Fig. 9a). After 30 min of stimulation with ␤-adrenergic agonist, the receptor distribution is largely intracellular, although some plasma membraneassociated receptor is still obvious (Fig. 9b). When the clones are treated with the Src family tyrosine kinase inhibitor PP2 for 30 min (Fig. 9c), two changes in the localization of the receptors are noted. First, there appears a greater level of receptor localized to the cytoplasmic, perinuclear regions of the cell. Second, and more striking, is the marked thickening of the band of GFP-tagged receptors in the plasma membrane region (see arrows, Fig. 9c). Challenging the PP2-treated cells with agonist only increases the thickening of the band of GFP-tagged receptor localized about the plasma membrane. Washout of the PP2 in the continued presence of isoproterenol leads eventually to a normal pattern of perinuclear receptor localization (data not shown). These data suggest that Src activity is important not only to desensitization but also to the subsequent phase of receptor biology in which the desensitized receptors are trafficked to the cytoplasmic, perinuclear regions of the cells for resensitization and eventual transit back to the plasma membrane. Inhibition of Src activity stalls the agonist-induced internalization at some point early in vesicular trafficking from the plasma membrane. DISCUSSION The Src family of nonreceptor tyrosine kinases have been shown to play important roles in the biology of G-proteinlinked receptors. In response to agonist binding, GPLRs typically demonstrate several discrete, temporally linked responses, the first of these being activation of the receptor, transducing activation of the cognate heterotrimeric G-protein(s) to which the receptor is coupled. Following activation, virtually all GPLRs display agonist-induced desensitization, sequestration/internalization, and a resensitization phase. Recently, the ␤2-adrenergic receptor was shown to display signaling "switching" in which the desensitization of its well known cyclic AMP response was accompanied by a second wave of responses involving a new cast of G-protein partners and signaling to the mitogen-activated protein kinase pathway via Ras (3,4). The latter response to mitogen-activated protein kinase has been shown to involve internalization in some cell lines (26) and a scaffold role of the ␤2-adrenergic receptor for mitogen-activated protein kinase signaling elements in other cell lines (25). Thus, much of what we know about the role of Src in GPLR generally and ␤2-adrenergic receptors specifically involves agonist-induced responses that occur temporally after desensitization.
Herein we explored the possible role of Src in events that precede rather than follow desensitization. The results demonstrate an important role of Src in the ability of the ␤2-adrenergic receptor to undergo agonist-induced desensitization. The primary results on which this claim is made are as follows.
Treating cells with the cell-permeable PP2 inhibitor of Src kinases impairs desensitization; suppression of the expression of Src specifically leads to a loss in agonist-induced desensitization for the ␤2-adrenergic receptor; and expression of a dominant-negative mutant (K295R/Y527F) of Src nullifies agonistinduced desensitization and enables an impressive cyclic AMP response for periods up to 60 min post-challenge with agonist. These data implicate Src itself, rather than other Src family members, in the desensitization of ␤2-adrenergic receptors in response to agonist challenge. Other Src kinase family members, however, may play analogous roles for GPLRs other than ␤2-adrenergic receptors.
How Src is recruited to the receptor complex is not understood. Based upon pull-down assays, it has been suggested that Src associates with at least one well known member of the desensitization complex, ␤-arrestin. The interaction between ␤-arrestin and Src appears to be either a weak SH3 motif found in the N terminus of ␤-arrestin and/or a binding site that interacts with the catalytic domain of Src. Because the current hypothesis for agonist-induced desensitization suggests that GRKs first are recruited to the activated GPLR, followed by Ser/Thr phosphorylation of the receptor and eventual recruitment of ␤-arrestin to the phosphorylated GPLR, the proposed association of Src with ␤-arrestin as a targeting mechanism seems untenable. In addition, it has been shown that Src both phosphorylates and activates GRK2 itself (demonstrated in vivo in the current work), arguing that the proper sequence of Src association with the c-Src-␤2-Adrenergic Receptor Signaling Complexes receptor would be in advance of (and not following) GRK-catalyzed phosphorylation of the GPLR (Fig. 10).
What other mechanism might be proposed and tested to explain the association of Src with the ␤2-adrenergic receptor in advance of GRK association? The ␤2-adrenergic receptor harbors a putative, canonical SH2-binding domain flanking the Tyr-350 residue found in the cytoplasmic, C-terminal tail of the receptor. Upon phosphorylation, the Tyr-350 residue constitutes a functional SH2-binding domain that can recruit SH2 domain-containing partners to the ␤2-adrenergic receptor, including the adaptor molecule Grb2, the GTPase involved in trafficking dynamin, and 1-phosphatidylinositol 3Ј-kinase. The interaction of the Tyr-350-phosphorylated ␤2-adrenergic receptor with Grb2 has been shown to be dependent upon the integrity of the SH2, but not SH3, domains of Grb2 (24). The presence of a potential SH2-binding site in the cytoplasmic, C-terminal region of the ␤2-adrenergic receptor provoked us to test the hypothesis that this site may become phosphorylated in response to challenge with agonist. This working hypothesis is displayed in Fig. 10. The results displayed in Fig. 6 clearly demonstrate that agonist stimulates the phosphorylation of the Tyr-350 residue and that a Tyr-to-Phe substitution at this residue largely attenuates the increase in phosphotyrosine content found in the ␤2-adrenergic receptor upon agonist treatment (Fig. 6). Earlier it was shown that the ␤2-adrenergic receptor is a substrate for tyrosine phosphorylation in response to activation of several growth factors, including insulin and insulin-like growth factor 1 (21)(22)(23). Insulin treatment in vivo leads to the phosphorylation of Tyr-350, as well as Tyr-354 and Tyr-364. Phosphorylation of the Tyr-350 residue of the ␤2adrenergic receptor leads to recruitment of Grb2, dynamin, and PI3 kinase to the receptor (17,20,23,24). Although the nature of the tyrosine kinase catalyzing the phosphorylation of Tyr-350 in response to ␤2-adrenergic agonist remains to be established, the Tyr-350 site is shown in the current work to be critical for agonist-induced desensitization.
Cells expressing the Y350F mutant form of the ␤2-adrenergic receptor fail to display normal agonist-induced desensitization (Fig. 7). The phosphotyrosine content of the ␤2-adrenergic receptor in response to stimulation by agonist is sharply attenuated when the Y350F mutant receptor is expressed (Fig. 6). Furthermore, the ability of the ␤2-adrenergic receptor to recruit Src in response to agonist stimulation is abolished by the Tyr-to-Phe substitution of the Tyr-350 residue of the receptor (Fig. 5). These data provide a compelling case for the role of the Tyr-350 residue as a docking site for Src binding and for Src activation upon phosphorylation. A working model that captures these features of Src association with the ␤2-adrenergic receptor in the process of agonist-induced desensitization is displayed (Fig. 10). According to this model, it is the phosphorylation of Tyr-350 by some tyrosine kinase in response to agonist that initiates the process of desensitization. Upon phosphorylation, the Tyr-350 residue recruits Src to the plasma membrane-localized receptor, both docking and activating Src. Activated Src, in turn, phosphorylates and activates GRK2. GRK2 then catalyzes the further phosphorylation of the ␤2adrenergic receptor on Ser/Thr residues, which in turn creates the docking site for ␤-arrestin, largely completing the desensitization phase of the receptor biology. Acting as an adaptor between the phosphorylated receptor and clathrin, ␤-arrestin then initiates the sequestration/internalization of the ␤2-adrenergic receptor, a process essential for resensitization and relocalization to the plasma membrane to occur. Based upon the model (Fig. 10), one can ask, if the Tyr-350 phosphorylated receptor is necessary to recruit Src in advance of GRK2, what the outcome is of inhibiting Src on the subsequent activation of GRK. The answers to this question were quite clear. Inhibition of Src with PP2, expression of a dominant-negative version of Src, or suppression of Src levels by treatment with antisense oligodeoxynucleotides all blocked agonist-induced desensitization as well as the agonist-induced increase in the phosphotyrosine content (i.e. activation) of FIG. 10. Model of ␤2-adrenergic receptor-mediated recruitment and activation of Src, which activates GRK2 and leads both to desensitization and to ␤-arrestin-mediated targeting of the receptor-Src complex with clathrin. The phosphorylation of Tyr-350 of the ␤2-adrenergic receptor creates an SH2-binding site to which Src is recruited upon activation of receptor by ␤-adrenergic agonist. Src association with the receptor SH2-binding site activates Src and catalyzes its phosphorylation/activation of GRK2, a Ser/Thr kinase that phosphorylates the agonist-activated receptor. The Ser/Thr-directed phosphorylation of the ␤2-adrenergic receptor recruits ␤-arrestin to the complex to complete agonist-induced desensitization and commence clathrin-mediated internalization. Inhibition of Src kinase with PP2, by expression of a dominantnegative mutant of Src or the Y350F mutation of the ␤2-adrenergic receptor, abolishes the ability of Src to be recruited/activated in response to ␤-adrenergic agonist and thereby precludes both desensitization and eventually full internalization of the receptor-Src complex. Not shown for the sake of simplicity is the scaffold molecule gravin that nucleates the complex of the receptor with several additional kinases and phosphatases. ␤AR or bAR, ␤-adrenergic receptor; TK, tyrosine kinase.

c-Src-␤2-Adrenergic Receptor Signaling Complexes
GRK2. Similarly, inhibiting Src activity with PP2 was shown to block agonist-induced changes in receptor biology, including desensitization and receptor trafficking. Using autofluorescent protein-tagged ␤2-adrenergic receptors and epifluorescence, we demonstrated a marked thickening of the band of receptor localized in close proximity to the plasma membrane in response to agonist in the PP2-treated cells. This observation made in live cells is consistent with the notion that Src activity is obligate both for desensitization and for proper trafficking of the ␤2-adrenergic receptor in response to agonist.
Taken together, these results reveal a novel, important role for Src in the regulation of ␤2-adrenergic receptors that may well extend to many other members of the GPLR superfamily. Src is critical for agonist-induced desensitization. For the ␤2-adrenergic receptor, recruitment of Src requires phosphorylation of a tyrosine residue in the cytoplasmic, C-terminal tail of the receptor, which creates a conditional, canonical SH2-binding site for Src. Through this early receptor-Src association and Src activation, Src is positioned to phosphorylate incoming GRK2 and to facilitate the subsequent phosphorylation of the receptor on Ser/Thr residues of the cytoplasmic, C-terminal tail that recruit ␤-arrestin and ultimately clathrin. The interaction of Src with ␤-arrestin, observed when both signaling molecules are overexpressed (3), may reflect a later role of Src in receptor internalization or signaling to Ras. These studies are the first to illuminate the role of Src in agonist-induced desensitization.