Degradation of the G Protein-coupled Receptor Kinase 2 by the Proteasome Pathway*

GRK2 is a ubiquitous member of the G protein-cou-pled receptor kinase (GRK) family and has been shown to play a key role in determining the desensitization and resensitization patterns of a variety of G protein-cou-pled receptors. In this report, we show that GRK2 is actively degraded by the proteasome proteolytic pathway, unveiling a new mechanism for the rapid regulation of its expression levels. Interestingly, activation of b 2 -adrenergic receptors ( b 2 AR) markedly increases GRK2 ubiquitination and degradation through the proteasome pathway. In addition, blocking GRK2 degradation notably alters b 2 AR signaling and internalization, consistent with a relevant physiological role for GRK2 proteasomal degradation. Activity-dependent modulation of GRK2 cellular levels emerges as an important mechanism for modulating the cellular response to agonists acting through G protein-coupled receptors. Agonist-promoted desensitization of G protein-coupled receptors (GPCR) 1 involves rapid phosphorylation of the activated receptor by a family of specific G protein-coupled receptor kinases (GRK), followed by binding to the phosphorylated receptor of regulatory proteins termed arrestins, leading to uncoupling from G proteins (1, 2).

GRK2 is a ubiquitous member of the G protein-coupled receptor kinase (GRK) family and has been shown to play a key role in determining the desensitization and resensitization patterns of a variety of G protein-coupled receptors. In this report, we show that GRK2 is actively degraded by the proteasome proteolytic pathway, unveiling a new mechanism for the rapid regulation of its expression levels. Interestingly, activation of ␤ 2 -adrenergic receptors (␤ 2 AR) markedly increases GRK2 ubiquitination and degradation through the proteasome pathway. In addition, blocking GRK2 degradation notably alters ␤ 2 AR signaling and internalization, consistent with a relevant physiological role for GRK2 proteasomal degradation. Activity-dependent modulation of GRK2 cellular levels emerges as an important mechanism for modulating the cellular response to agonists acting through G protein-coupled receptors.
Agonist-promoted desensitization of G protein-coupled receptors (GPCR) 1 involves rapid phosphorylation of the activated receptor by a family of specific G protein-coupled receptor kinases (GRK), followed by binding to the phosphorylated receptor of regulatory proteins termed arrestins, leading to uncoupling from G proteins (1,2). GRK2 is the most ubiquitous member of this kinase family and has been shown to play a key regulatory physiological role in the heart and other tissues (3,4). GRK2 may also play a role in facilitating the transient agonist-induced GPCR internalization allowing receptor dephosphorylation and recycling (5)(6)(7). Thus, GRK2 contributes to determine the rate and extent of both GPCR desensitization and resensitization. This fact, together with the recently suggested participation of GRKs and arrestins in the GPCR-mediated mitogen-activated protein kinase cascade (8) and the ability of GRK2 to phosphorylate tubulin (9), further stresses the relevance of GRK2 in GPCR signaling.
Changes in GRK2 expression may therefore alter the efficacy of GPCR signal transduction systems. In this regard, GRK2 expression levels have been reported to increase in human heart failure (10), hypertension (11) or upon lymphocyte activation (3), or to decrease rapidly in several rat tissues during the perinatal period. 2 Despite recent advances in the understanding of the mechanisms regulating the activity and subcellular distribution of GRK2 (1,2,6,12,13), very little is known about the mechanisms that regulate the kinase cellular complement. In this study, we sought to characterize the degradation process of GRK2 protein and its potential modulation. Our results clearly show that GRK2 is rapidly degraded by the proteasome pathway and that kinase degradation is modulated by activation of ␤ 2 -adrenergic receptors (␤ 2 AR). Regulation of GRK2 stability may provide an important mechanism for modulating the cellular levels of this kinase and contribute to the regulation of desensitization/resensitization patterns of GPCR, as suggested by the fact that blockade of GRK2 degradation alters ␤ 2 AR signaling as well as receptor internalization.

EXPERIMENTAL PROCEDURES
Cell Treatments and Metabolic Labeling-HEK-293 cells (American Type Culture Collection, Manassas, VA) were transfected with pCDNA3-GRK2, pCDNA3-GRK2-K220R, and/or pws636-SF ␤ 2 AR by the calcium phosphate precipitation method and selected for stable expression as described (5). In some experiments, cells were transiently transfected as reported (14) with different combinations of GRK2 and ␤ 2 AR, as well as constitutively active ␤ 2 AR (15) or a COOH-terminal truncated mutant (amino acids 1-354) of the mouse ␤ 2 AR. 3 For metabolic labeling, cells were kept for 2 h in methionine and cysteine-free Dulbecco's modified Eagle's medium (DMEM) and then incubated for 15-30 min in DMEM supplemented with 250 Ci/ml of [ 35 S]methionine and [ 35 S]cysteine labeling mixture (NEN Life Science Products).The plates were washed with phosphate-buffered saline and chased for the indicated times in DMEM plus 10% fetal bovine serum. The proteolysis inhibitors lactacystin (Calbiochem, La Jolla, CA), ALLN (Sigma), and leupeptin (Boehringer Mannheim) were added 90 min before metabolic labeling and maintained during the chase periods (16). Pulse-labeled cells expressing ␤ 2 AR were challenged with isoproterenol (Sigma) or vehicle during chase periods in DMEM supplemented with 20 mM HEPES (pH 7.5) and 1 mM ascorbic acid (Sigma).
Western Blotting-Cells lysates were obtained exactly as described (12) and resolved in 7% SDS-PAGE. Proteins were transferred to nitro-* This work was supported by Grant PM95-0033 from the Ministerio de Educación y Cultura, Grant AE 00213/95 from the Comunidad de Madrid, and Grant BMH4 -98-3566 from the European Union (to F. M.) and grants from Comision Interministerial Ciencia y Tecnología and Fundación Ramón Areces (to J. G. C.). The Centro de Biología Molecular holds an institutional grant from the Fundación Ramón Areces. 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.
cellulose membranes and probed with GRK2 Ab9 polyclonal antibody raised against recombinant GRK2 as described (5,12), and the blot was developed using a chemiluminiscent method (ECL, Amersham Pharmacia Biotech). In other experiments, blots were analyzed using the GRK2 polyclonal antibodies AbFP2 or affinity-purified Ab1 (17), raised against different regions of bovine GRK2 (see "Results"). Densitometric analysis of the blots was performed as above.
In Vitro GRK2 Degradation Assays-Bovine GRK2 was overexpressed in Sf9 cells using the baculovirus expression system and purified as described previously (12). Purification of a 20 S proteasome (MCP) from rat liver was as described (18,19). The in vitro proteolytic assay was performed by incubating 2 g of purified MCP with 1.25 g of GRK2 exactly as reported (18,19). Proteins were resolved in 10% SDS-PAGE and analyzed by Coomassie Blue staining. Alternatively, GRK2 samples (10 g) undigested or following digestion with proteasome for 60 min were injected on a gel filtration HPLC column (TSK 5000, Beckman, Palo Alto, CA) (18). Fractions in the ϳ80-kDa range were tested for either the presence of GRK2 and proteolytic fragments in Western blot or for GRK2 activity as detailed below.
Determination of GRK2 Activity-GRK2 samples obtained by gel filtration on an HPLC column (see above) were diluted 1:10, and GRK2 activity was assessed by the rhodopsin phosphorylation assay as described (12). Relative GRK2 activity was estimated by densitometric analysis of the autoradiography.
Ubiquitination of GRK2-HEK-293 cells were transiently transfected by the calcium phosphate precipitation method with different combinations of expression vectors for GRK2, ␤ 2 AR, and HA-ubiquitin (20). Cells were treated with isoproterenol or vehicle for 30 min, and ubiquitin conjugates were purified by immunoprecipitation with a specific monoclonal antibody against the HA epitope (Boehringer Mannheim). The presence of ubiquitinated GRK2 was subsequently analyzed by Western blotting with Ab9 antibody as described above. Similar GRK2 and ubiquitin expression levels were confirmed in cell lysates by dot blot and Western blot analysis using their specific antibodies. ␤ 2 AR levels were determined by radioligand binding as described (5).
Desensitization and Adenylyl Cyclase Assays-HEK-293 cells stably expressing GRK2 and ␤ 2 AR or transiently transfected with GRK2 and wild-type ␤ 2 AR or COOH-terminal truncated ␤ 2 AR were subcultured (6 ϫ 10 5 cells) for 12 h before preincubation with 30 M lactacystin or vehicle (90 min, 37°C) and maintained in its presence through the following steps. Subsequently, cells were incubated with 10 M isoproterenol or vehicle (30 min, first challenge). Afterward, cells were exten-sively washed with phosphate-buffered saline and incubated with DMEM for 30 min (resensitization period) before adding 10 M isoproterenol (10 min, second challenge). Lysates were obtained and assessed in triplicate for cAMP levels as described (21). Data were corrected according to the total protein amount in each sample and compared with values obtained in control conditions (not treated with lactacystin and not undergoing the first challenge with agonist).
Internalization Assays-Following preincubation with 30 M lactacystin or vehicle (control) in DMEM for 3 h, cells were challenged with 0.1 or 1 M isoproterenol for 2 min at 37°C, washed by centrifugation, and labeled at 4°C with M1 FLAG antibody. Receptor sequestration was then quantitated by flow cytometry essentially as described (5). The fraction of sequestered receptors was calculated by comparing the signal obtained in the absence or presence of agonist.

GRK2 Is Degraded by the Proteasome in Vivo and in Vitro-
The GRK2 degradation rate was measured by pulse and chase analysis using HEK-293 cells that stably express GRK2. In such cells, neither the subcellular localization pattern of the kinase nor the ratio between subcellular compartments was altered when compared with endogenous GRK2 (14). A rapid decay in the 35 S-labeled GRK2 immunoprecipitated by a specific kinase antibody is shown in Fig. 1A, indicating that GRK2 is a short-lived protein with a half-life estimated at ϳ60 min. A similar value was obtained for the endogenous GRK2 in identical experiments performed in Jurkat cells (data not shown), ruling out the possibility that overexpression might alter the normal rate of kinase proteolysis. To ascertain which mechanism was responsible for the rapid GRK2 degradation, pulse and chase experiments were performed with specific inhibitors of different proteolytic pathways (Fig. 1B). Treatment with the cysteine protease inhibitor ALLN clearly prevents GRK2 degradation after 4 h of chase when compared with control conditions, at concentrations (50 M) in which it inhibits the proteasome pathway but not at doses (1 M) in which only nonproteasomal proteases are blocked (22), thus excluding the involvement of the latter. The presence of leupeptin, an inhib- itor of protein degradation in lysosomes, did not alter GRK2 degradation. These results suggested that the proteasome pathway was involved in the rapid GRK2 proteolysis. Consistently, the addition of the highly specific proteasome inhibitor lactacystin fully blocks kinase degradation in intact cells (Fig. 1B).
To further address the role of the proteasome in regulating GRK2 stability, in vitro degradation experiments were conducted with recombinant GRK2 and purified 20 S proteasome from rat liver (18). The purified MCP preparations were judged homogeneous by SDS-PAGE analysis ( Fig. 2A). Under the conditions tested, recombinant GRK2 (indicated as the 80-kDa protein in Fig. 2A) is indeed rapidly degraded, whereas GRK2 incubated for 2 h without MCP remains unaltered. To exclude the possibility that a minor contaminant protease present in the incubation could contribute to GRK2 proteolysis, we tested different MCP preparations for their ability to degrade GRK2, obtaining essentially the same results (data not shown). Moreover, different doses of the specific proteasome inhibitor lacta-cystin clearly prevent GRK2 cleavage as compared with control (Fig. 2B). Altogether, our data indicate that the kinase is a target for the proteasome, in vivo as well as in vitro. Interestingly, our in vitro GRK2 degradation experiments reveal the presence of a ϳ74-kDa band that appears as GRK2 decays (Fig.  2, A and C). This protein appears to represent an intermediate proteolytic product (also degraded by MCP following longer incubation times), whose formation is blunted in the presence of lactacystin (Fig. 2C). In a first attempt to determine whether the MCP initially digests the NH 2 -or the COOH-terminal region of GRK2 to yield the 74-kDa protein, we performed immunoblot analysis of GRK2 preparations incubated with or without MCP, using different GRK2 antibodies. Both intact GRK2 and the 74-kDa band are detected efficiently by the polyclonal antibody Ab9 raised against recombinant GRK2 and AbFP2, a polyclonal antibody against the COOH terminus (residues 436 -689) (Fig. 2C). Moreover, antibody Ab1, which recognizes a more restricted epitope (GRK2 residues 648 -655) that would be eliminated in an initial COOH-terminal cleavage, yields the same results (Fig. 2C, right panel), strongly suggesting that GRK2 is initially digested at its NH 2 terminus.
The generation of the 74-kDa band as a result of GRK2 degradation in vitro raised the question of whether this cleavage product could be also observed in vivo. Fig. 3A indicates that the 74-kDa band is indeed immunodetected when measuring steady-state protein levels in cellular lysates from HEK-293 cells stably overexpressing GRK2 alone or with ␤ 2 AR (Fig.  3A, lanes 4 and 5), whereas it is undetectable in untransfected cells (Fig. 3A, lane 1). Notably, the 74-kDa fragment seems to be more abundant in cells coexpressing GRK2 and a constitutively activated ␤ 2 AR mutant (15) (lane 2 versus lane 3 in Fig.  3A). Moreover, we are able to immunoprecipitate the 74-kDa protein with a GRK2-specific antibody (Fig. 3B) but not with a preimmune serum (data not shown), when cells are metabolically labeled for periods longer than those used in pulse and chase experiments. These data indicate that the partial digestion of GRK2 by the proteasome takes place in situ and suggest that receptor activation might play some role in the process to render this product. We next determined the functional activity of the 74-kDa fragment by testing its ability to phosphorylate rhodopsin, a well known GRK2 substrate (12). For this purpose, kinase samples, undigested or digested with MCP, were purified by HPLC chromatography to remove MCP prior to the phosphorylation assay. The eluted fractions were resolved by SDS-PAGE and analyzed in Western blot (Fig. 3C). Most GRK2 was cleaved by MCP to yield the 74-kDa fragment. As can be seen in Fig. 3D, MCP-digested fractions (open circles) have less capacity to phosphorylate rhodopsin than undigested fractions (filled circles). Samples treated with MCP still display undigested GRK2 (ϳ27%), which may account for the remaining activity. The decrease in activity (a 44% reduction) is lower, however, than that expected in such a case (a 73% reduction) indicating that the 74-kDa fragment can indeed phosphorylate rhodopsin, although with less efficiency than GRK2.
␤ 2 AR Activation Increases GRK2 Ubiquitination and Degradation-To explore whether receptor activation can modulate kinase turnover, we treated HEK-293 cells stably coexpressing GRK2 and ␤ 2 AR with the ␤-agonist isoproterenol (10 M) or vehicle during different chase periods after 35 S metabolic labeling (Fig. 4A). Under control conditions (Fig. 4A, ϪISO), the levels of GRK2 decay essentially as described in cells expressing the kinase alone (see also Fig. 1A). Agonist stimulation (Fig. 4A, ϩISO) clearly enhances the GRK2 degradation rate as compared with control (27% GRK2 remaining after 1 h of ISO treatment versus 54% in untreated cells; p Ͻ 0.05, data indicate the means Ϯ S.E. of three to five independent experiments). GRK2 half-life thus decreases from 60 min (basal conditions) to ϳ30 min (stimulated turnover). The presence of lactacystin or ALLN completely abrogates GRK2 degradation in isoproterenol-treated cells (Fig. 4B), as they do in unstimulated cells. This ruled out the possibility that the increase in kinase processing triggered by receptor stimulation might arise from the activation of an additional proteolytic pathway other than proteasome-dependent degradation.
Most proteins targeted to the proteasome pathway undergo polyubiquitination (23,24). To investigate the possible role of ubiquitination in basal and induced degradation of GRK2 in intact cells, we transiently transfected HEK-293 cells with different combinations of a tagged ubiquitin construct, GRK2, and ␤ 2 AR. In cells in which ubiquitin was overexpressed with the kinase and the receptor, a band smear corresponding to multi-ubiquitinated GRK2 (marked with a line in the right side of the upper panel in Fig. 4C) was clearly detected upon receptor stimulation with isoproterenol, whereas GRK2 ubiquitination was much lower in untreated cells (Fig. 4C, upper panel). Taken together, our data suggest that the extent of GRK2 ubiquitination is stimulation-dependent and probably accounts for the enhanced agonist-induced kinase degradation by the proteasome pathway.
Given that ␤ 2 AR occupancy by agonist promotes very rapid translocation of GRK2 to the plasma membrane and kinase activation (1, 2, 6, 13), we tested whether the activity of the kinase might also affect its turnover. To this end, HEK-293 cells stably expressing ␤ 2 AR and a dominant-negative GRK2 mutant (GRK2-K220R) (25) were metabolically labeled and chased in the presense (Fig. 4D, lanes I) or absence (Fig. 4D, lanes C) of isoproterenol. Agonist stimulation promotes no change in mutant kinase levels (Fig. 4D). More surprisingly, the amount of GRK2-K220R does not appear to decay even after 8 h of chase in the absence of agonist (Fig. 4D, upper  panel), indicating that the mutant protein is far more stable than the wild type.
Blocking GRK2 Degradation Alters ␤ 2 AR Signaling-The rapid and agonist-modulated degradation of GRK2 by the proteasome would provide a mechanism to regulate the cellular complement of GRK2 and, therefore, GPCR desensitization patterns. In this regard, we used HEK-293 cells stably expressing both ␤ 2 AR and GRK2 to test the effects of blocking GRK2 degradation with the specific proteasome inhibitor lactacystin on the subsequent cell response to the ␤-adrenergic agonist isoproterenol and on the ␤ 2 AR desensitization process. In control cells, agonist challenge promotes a 1.8 Ϯ 0.1-fold increase in cAMP levels over basal values. Pretreatment with isoproterenol for 30 min induces the typical decrease in responsiveness to a subsequent agonist challenge (23 Ϯ 7% decrease) (Fig.  5, left panel) and also a decrease in GRK2 levels (45 Ϯ 9% of control), consistent with our previous data. Interestingly, inhibition of GRK2 degradation with lactacystin results in a clear increase in GRK2 protein levels both in control and isoproterenol-pretreated cells (Fig. 5, inset, lanes 2 and 4) and reduces ␤ 2 -adrenergic receptor signaling by 33 Ϯ 2 and 44 Ϯ 1%, respectively (Fig. 5, left panel). Because cellular ␤ 2 AR levels remain unchanged with lactacystin treatments (see legend to Fig. 5), these data suggest that alteration of the kinase/receptor ratio clearly affects signal transduction. Other factors may also contribute to the reduction in ␤ 2 AR signaling in the presence of lactacystin, because regulatory subunits of cAMP-dependent protein kinase (26,27) and other components of the transduc- tion cascade could be degraded by the proteasome pathway. However, adenylyl cyclase activation by forskolin is not modified in the presence of lactacystin, and the same results were obtained in the presence of cAMP phosphodiesterase inhibitors (data not shown). Moreover, lactacystin treatment does not affect cAMP production by a transiently transfected truncated ␤ 2 AR, which lacks phosphorylation sites for GRK2 but retains those for cAMP-dependent protein kinase and its ability to modulate adenylyl cyclase (Fig. 5, left panel), indicating that changes in GRK2 levels have a direct and crucial effect in the observed alteration in ␤ 2 AR responsiveness.
Previous reports have provided evidence that GRK2 plays a facilitatory role in receptor internalization (5,6,28). Consistently, the blocking of kinase degradation with lactacystin markedly enhanced (up to 2.6-fold) receptor sequestration after a short term (2 min) challenge with isoproterenol (Fig. 5, right  panel). These results are in agreement with our previous report (5) and further indicate that the impairment of the normal kinase turnover has important functional consequences on receptor desensitization and internalization processes. DISCUSSION Changes in the cellular complement of GRK2 can modulate the desensitization/resensitization processes of a variety of GPCR (7, 29) and may prove critical in determining the speci-ficity of regulation of certain receptors (4,7,30). In this report, we show that GRK2 displays a rapid turnover through the ubiquitin-dependent proteasome pathway. This proteolytic pathway is responsible for the rapid degradation of an increasing number of cell cycle regulators and key signaling proteins (see Ref. 31 for a recent review). Changes in the stability of proteins as a consequence of dysregulated proteasome complex activity can lead to an impairment in the cellular functions mediated by these proteins (31). In this line, our data indicate that GRK2 degradation is modulated by GPCR activation and that blocking of kinase degradation strongly affects its role as regulator of GPCR function.
Proteins targeted to the proteasome pathway usually undergo polyubiquitination (23,24), although this is not an absolute requirement (16,32,33). We show that GRK2 ubiquitination and proteolysis by the proteasome pathway is increased upon agonist activation of GPCR, although the kinase also exhibits weak ubiquitination under unstimulated conditions, suggesting that common mechanisms can regulate signal-dependent and basal GRK2 degradation in contrast to I␤␣ degradation (16). On the other hand, the ability of purified 20 S proteasome to digest GRK2 indicates that a direct interaction between the kinase and proteasome subunit(s) can occur. It is feasible that an undetermined structural motif in GRK2 con- tributes to target the protein to this degradation machinery, in addition to ubiquitination.
In contrast to most proteasome substrates that undergo a complete and processive degradation, proteolytic processing of GRK2 by the proteasome yields a 74-kDa fragment that transiently accumulates in vivo as well as in vitro. The generation of this product could only be a "marker" of the initial steps of GRK2 degradation, like the 42-kDa fragment of cyclin B (34), and/or have an unknown physiological role. Consistent with the latter possibility, the 74-kDa fragment indeed displays kinase activity toward a G protein-coupled receptor substrate. Interestingly, its ability to phosphorylate rhodopsin is clearly lower than that of the full-length kinase. In this regard, it has been recently reported that GRK2 is also able to phosphorylate other proteins different from G protein-coupled receptors, such as tubulin (9). It is tempting to speculate that the proteolytic removal of a NH 2 -terminal domain of GRK2 by the proteasome both alters the kinase efficacy toward GPCR and modifies kinase activity toward other cellular substrates.
We provide evidence that ␤ 2 AR activation promotes an increase in GRK2 ubiquitination and degradation. Several hypotheses can be considered to explain this observation. First, GRK2-mediated phosphorylation promotes arrestin binding to the agonist-activated receptor, leading to receptor uncoupling and internalization (1,5,28). Data suggest the formation of a multimolecular complex (receptor/GRK2/␤-arrestin) at specialized locations of the plasma membrane (5, 35-37) that will be internalized after recruitment of different proteins of the endocytic machinery. It is possible that interaction of GRK2 with molecules of this multimolecular complex plays a role in its signal-dependent degradation by locating the kinase in close proximity with the molecules responsible for its ubiquitination and/or allowing conformational changes that unmask kinase domains directly targeted for ubiquitination. In this regard, it is worth noting that the yeast G protein-coupled receptor Ste2p is ubiquitinated at the plasma membrane even if endocytosis is blocked (38, 39). Alternatively, the higher level of GRK2 ubiquitination due to agonist receptor activation might result from second messenger-mediated mechanisms or as a consequence of stimulated GRK2 activity leading to phosphorylation of unknown factors different from GPCR or to transient increases in kinase autophosphorylation, improving GRK2 modification by ubiquitination. Data obtained with the GRK2-K220R mutant are consistent with a strong relationship between the stability of GRK2 and its cellular function; the slow turnover of GRK2-K220R as compared with wild-type GRK2 is largely unaffected by receptor activation, the mutant is deficient in kinase activity, and it strongly impairs ␤-arrestin binding and the subsequent internalization (1,28,40).
Several recent reports suggest that both the extent and rate of receptor internalization as well as the pattern of desensitization/resensitization processes may vary depending on the cellular content of GRK2 and ␤-arrestin (7,30,35). Our data are consistent with these reports and, further, suggest that a tight regulation of GRK2 cellular levels by agonists is needed for normal receptor function and modulation. Indeed, the rise in GRK2 complement by blockade of its normal degradation pathway clearly enhanced the initial rate of receptor sequestration, according to the facilitatory role of GRK2 previously reported in this process (5,28). More interestingly, receptor response to agonists is markedly affected in these conditions, indicating that in our cellular system a moderate increase of GRK2 over its steady-state levels is enough to evoke a strong change in GPCR signaling, putting forward the relevance of an adequate ratio between the receptor and its regulatory proteins. In this context, it is tempting to suggest that the increased down-regulation of kinase levels observed after sustained GPCR stimulation would help to preserve some degree of response to additional activation by agonists. In this regard, it is worth noting that GRK2 levels rapidly decrease in several tissues during the rat perinatal period, 2 a physiological situation characterized by a dramatic and prolonged surge in plasma catecholamines. On the other hand, the impaired protein degradation by the proteasome pathway described in pathological conditions such as hypothyroidism (41) could explain the increased GRK2 levels detected by our laboratory in several  Fig. 4 or cells transiently transfected with GRK2 and a COOH-terminal truncated mouse ␤ 2 AR were preincubated with lactacystin (Lac) or vehicle (control), pretreated with isoproterenol or vehicle (control) for 30 min, resensitizated, and finally challenged with isoproterenol for 10 min to both assess receptor-dependent adenylyl cyclase activity and GRK2 levels by immunoblot (inset). Data are expressed as fold stimulation of cAMP levels over basal values in control cells (18 Ϯ 4 pmol/mg protein). Fold stimulation in control conditions was taken as 100%. Treatments did not significantly alter total receptor levels in stable or transfected cells (28.6 Ϯ 0.6 and 1.2 Ϯ 0.3 pmol/mg protein, respectively). Data are the means Ϯ S.E. from two or three experiments performed in duplicate. Right panel, HEK-293 cells stably transfected with GRK2 and an epitope-tagged ␤ 2 AR receptor were incubated with lactacystin or vehicle (control) for 3 h and receptor internalization in response to a 2-min challenge with the indicated concentrations of agonist analyzed by flow cytometry (see "Experimental Procedures"). Receptor internalization under control conditions was taken as 100%. Results are the means Ϯ S.E. of three independent experiments performed in duplicate. rat hypothyroid tissues 4 and contribute to the reported attenuation of adrenergic responses in this situation.
In summary, our results indicate that receptor activation leads not only to changes in GRK2 activity and subcellular localization but also promotes enhanced GRK2 proteolysis through a ubiquitin-dependent proteasome pathway. Activitydependent regulation of GRK2 stability may provide an important mechanism for modulating its expression levels, contributing to the regulation of desensitization/resensitization patterns of GPCR. Changes in GRK2 levels promoted by the activation of a given receptor may also modify the efficacy of the cellular response to subsequent challenges by other agonists acting through different GPCR. This new pathway of receptor-dependent modulation of the stability of a regulatory protein may therefore have important physiological consequences.