Receptor Species-dependent Desensitization Controls KCNQ1/KCNE1 K+ Channels as Downstream Effectors of Gq Protein-coupled Receptors*

Activation of Gq protein-coupled receptors (GqPCRs) might induce divergent cellular responses, related to receptor-specific activation of different branches of the Gq signaling pathway. Receptor-specific desensitization provides a mechanism of effector modulation by restricting the spatiotemporal activation of signaling components downstream of Gq. We quantified signaling events downstream of GqPCR activation with FRET-based biosensors in CHO and HEK 293 cells. KCNQ1/KCNE1 channels (IKs) were measured as a functional readout of receptor-specific activation. Activation of muscarinic M1 receptors (M1-Rs) caused robust and reversible inhibition of IKs. In contrast, activation of α1B-adrenergic receptors (α1B-ARs) induced transient inhibition of IKs, which turned into delayed facilitation after agonist withdrawal. As a novel finding, we demonstrate that GqPCR-specific kinetics of IKs modulation are determined by receptor-specific desensitization, evident at the level of Gαq activation, phosphatidylinositol 4,5-bisphosphate (PIP2) depletion, and diacylglycerol production. Sustained IKs inhibition during M1-R stimulation is attributed to robust membrane PIP2 depletion, whereas the rapid desensitization of α1B-AR delimits PIP2 reduction and augments current activation by protein kinase C (PKC). Overexpression of Ca2+-independent PKCδ did not affect the time course of α1B-AR-induced diacylglycerol formation, excluding a contribution of PKCδ to α1B-AR desensitization. Pharmacological inhibition of Ca2+-dependent PKC isoforms abolished fast α1B receptor desensitization and augmented IKs reduction, but did not affect IKs facilitation. These data indicate a contribution of Ca2+-dependent PKCs to α1B-AR desensitization, whereas IKs facilitation is induced by Ca2+-independent PKC isoforms. In contrast, neither inhibition of Ca2+-dependent/Ca2+-independent isoforms nor overexpression of PKCδ induced M1 receptor desensitization, excluding a contribution of PKC to M1-R-induced IKs modulation.

The canonical signaling pathway of activated G␣ q subunits comprises stimulation of PLC␤ 2 isoforms, hydrolysis of PIP 2 to release inositol trisphosphate (IP 3 ) and diacylglycerol (DAG), and subsequent Ca 2ϩ release from internal stores and DAG-mediated activation of protein kinase C (PKC) (1). More recent data (reviewed in Ref. 2) indicate that the linear pathway from G q PCR activation to G q -induced PLC stimulation is inadequate to explain the broad range of (sometimes divergent) cellular responses. Functional diversity upon stimulation of G q -coupled receptors might reflect differential coupling to G␣ q family members, different G q efficacies for PLC␤ activation, or activation of pathways independently of phosphoinositide hydrolysis. Furthermore, receptor-specific targeting of effector molecules has been shown to depend on the spatial proximity of G q PCR and effector protein (e.g. an ion channel) and on the mobility of signal molecules in the plasma membrane (3).
It is conceivable that G q PCRs can be distinguished by their efficiency to activate different branches of the G␣ q signaling pathway, either IP 3 -Ca 2ϩ or DAG-PKC, as recently shown for KCNQ2/3 channel modulation by P2Y 2 -R and M 1 -R (4,5). Stimulation of distinct G q -coupled receptors induces compartment-specific targeting of G q effector proteins. As reviewed in Refs. 6 -9, compartment-dependent translocation of PKC isoforms to the plasma membrane, the nuclear membrane, or the Golgi complex determines specific intracellular responses by placing PKC isoforms in proximity to their interaction partners. By using genetically encoded FRET-based sensors of organelle-specific PKC activity, recent studies provide evidence that location-specific DAG production enables recruitment of PKC isoforms to different intracellular membranes (10,11). Spatially restricted activation of components downstream of G q has been shown to modulate several types of ion channels in a receptor species-dependent fashion, e.g. N-type Ca 2ϩ channels (12), G protein-activated inward-rectifying K ϩ (GIRK) channels (3), and KCNQ1/KCNE1 channels (13).
Apart from spatial organization of signaling components, temporal aspects of receptor activation or G protein-effector interactions determine receptor-specific kinetics of G q PCR signaling. Diversity of GPCR-induced cellular effects might result from activation of multiple G proteins with varying efficacy and kinetics, inducing either fast cellular responses of limited extent through one type of G protein or a slow, but extended response through another (14). Moreover, as reviewed in Ref. 15, the affinity of activated G protein subunits to their specific effectors and their interaction kinetics determine the equilibrium of active and inactive trimeric G proteins and the dynamics of the G protein cycle.
Receptor-specific desensitization might provide a mechanism of effector modulation by restricting the spatiotemporal activation of downstream G q signaling components. Receptor desensitization in the continuous presence of an agonist terminates the responsiveness of a cell by limiting second messenger formation. A recent study, investigating the role of PKC activation in regulating TRPC6 channel activity, provides evidence that desensitization of the histamine H1 receptor reduces TRPC6 channel activity by rapid termination of DAG production (16). However, whether receptor species-dependent differences in desensitization determine the time course of channel modulation during stimulation of distinct G q PCRs has not been elucidated. Rapid termination of second messenger production by receptor desensitization delimits activation of downstream effector proteins. Thus, it is conceivable that receptor desensitization reduces the efficiency to activate certain effector proteins but favors activation of other branches of G q signaling pathways with either faster activation kinetics or higher second messenger affinity.
In the present study, we quantitatively analyzed signaling events downstream of distinct G q PCRs by means of FRETbased biosensors in a stable KCNQ1/KCNE1-transfected CHO cell line and in HEK 293 cells. I Ks amplitude during G q PCR stimulation was measured as a functional readout of receptor species-dependent activation of G q effectors. Our study demonstrates for the first time significant receptor-dependent differences in the time course of G q protein activation, DAG production (as a prerequisite of PKC activation), and PIP 2 hydrolysis upon stimulation of either adrenergic ␣ 1B or muscarinic M 1 receptors. These fundamental differences are attributed to different time courses of receptor desensitization and, as a consequence, induce different time courses of I Ks modulation. Our data provide evidence that receptor-specific desensitization controls PIP 2 reduction and recruitment of different PKC isoforms, thus resulting in different modes of fine-tuning of I Ks activity.

Results
Receptor Species-dependent Desensitization at the Level of G␣ q Activation-Activation of G q PCRs can be measured and quantified by using a FRET-based biosensor that monitors the G q protein cycle (17). The expression of the G q protein biosensor and either the ␣ 1B -AR or M 1 -R in HEK 293 cells allowed us to compare receptor species-dependent temporal properties of G q protein activation.
As illustrated by the representative recordings in Fig. 1, A and D, increasing concentrations of phenylephrine (Phe) or acetylcholine (ACh) (ranging from 0.1 nM to 1 M) result in an incremental increase of G q activation as monitored by the stepwise decrease of the FRET ratio. The decrease in the FRET ratio (⌬F) was normalized to ⌬F max (the FRET signal obtained during a single application of 1 M Phe or ACh), yielding concentrationresponse curves for phenylephrine (␣ 1B -AR)-and acetylcholine (M 1 -R)-induced G q activation (Fig. 1, B and E). Based on the concentration-response curves (revealing EC 50 values of 37 nM (Phe, ␣ 1B -AR) and 14 nM (ACh, M 1 -R)), we applied 1 M Phe as a saturating agonist concentration for ␣ 1B -AR activation and 10 M ACh for maximal M 1 -R-induced G q activation. During sustained ␣ 1B -adrenergic receptor activation by phenylephrine, a gradual increase in the FRET ratio, i.e. a decline in G q activation, was observed (Fig. 1C). The amount of desensitization was quantified by the ratio FRET 30 s after peak /FRET peak , indicating a reduction of the FRET ratio of 33 Ϯ 4.5% (n ϭ 8, Phe 1 M) for the ␣ 1B -AR (see also supplemental Fig. S1). In contrast, during stimulation of M 1 receptors (n ϭ 6, ACh 10 M), the FRET ratio remained stable (Fig. 1F).
It is conceivable that the receptor-specific differences in G protein activation reflect either different rates of desensitization of ␣ 1B and M 1 receptors or less efficient coupling of ␣ 1B receptors to G␣ q subunits as compared with M 1 receptors. We therefore investigated whether receptor species-dependent differences occurred downstream of G q activation, using FRET sensors for DAG formation and PIP 2 depletion.
Analysis of Receptor Species-dependent Desensitization by Monitoring the Time Course of DAG Formation-To analyze receptor species-dependent aspects of signaling downstream of the G q protein, we analyzed the time course of DAG formation in HEK cells expressing either ␣ 1B -AR or M 1 -R and the fluorescent biosensor DAGR. DAGR reports formation of DAG by an increase in FRET ratio, and the time course of DAG formation can be used to analyze the rate of receptor desensitization during agonist exposure (18). The representative recordings in Fig.  2 show effects of G q PCR stimulation on DAG production during application of Phe (1 M) and ACh (10 M). Stimulation of both receptor species caused a comparable, rapid increase in FRET ratio, reflecting activation of PLC and formation of DAG at the plasma membrane. However, DAG production induced by ␣ 1B receptors rapidly decayed during stimulation ( Fig. 2A), an effect that was not observed for the M 1 -R (Fig. 2B). The summarized data, expressed as the ratio FRET 30 s after peak / FRET peak in Fig. 2C, indicate a more than 50% reduction of the DAG signal during activation of ␣ 1B receptors. In contrast, during stimulation of M 1 receptors, we observed only a 10% reduction in DAG.
Receptor species-dependent differences in the time course of DAG reduction persisted across agonist concentrations. In the presence of nonsaturating agonist concentrations (e.g. 50 and 200 nM Phe and ACh), we observed a rapid decline of the DAG signal in ␣ 1B -AR-, but not in M 1 -R-expressing cells (see supplemental Fig. S2). The different time courses of DAG production in ␣ 1B -and M 1 -R-expressing HEK cells are likely to reflect receptor species-dependent differences in desensitization. Alternatively, differences in receptor expression levels may affect the time course of desensitization.
If this would be the case, any changes of G q PCR expression levels, controlled by adjusting the amount of transfected cDNA, might either increase or decrease the rate of desensitization of M 1 or ␣ 1B receptors. We therefore transfected HEK 293 cells with different amounts of receptor cDNA and subsequently monitored the time course of DAG formation during receptor stimulation. As shown in Fig. 2, D-I, neither increasing the expression level of M 1 receptors nor decreasing the expression of ␣ 1B receptors significantly affected the receptor-specific desensitization properties.
Simultaneous Measurements of DAG Production and I Ks Modulation in Stably Transfected KCNQ1/KCNE1 CHO Cells during Stimulation of ␣ 1B -AR and M 1 -R-Temporal aspects of G q PCR signaling were further analyzed in CHO cells stably transfected with KCNQ1/KCNE1 and transiently transfected with either ␣ 1B or M 1 receptors and the DAG sensor DAGR. We simultaneously measured DAG production and modula-tion of I Ks amplitude as a functional readout of G q PCR activation. Previous studies have reported conflicting results on I Ks modulation by different G q -coupled receptor species.
Both inhibition and activation of KCNQ channels have been reported (13,19,20). Some of these receptor speciesdependent differences in I Ks modulation might be attributed to activation of different branches of downstream G q signaling, e.g. enhanced PIP 2 depletion or activation of different PKC isoforms.
The representative recordings in Fig. 3 show the relation between G q PCR desensitization (assessed with the DAGR FRET signal) and I Ks current amplitudes in simultaneous recordings. Application of Phe (1 M) in ␣ 1B -AR-expressing cells resulted in I Ks inhibition caused by depletion of membrane PIP 2 (13), followed by a sustained increase in current amplitude after agonist withdrawal (Fig. 3, A and C). This delayed increase in current was lacking after M 1 -R stimulation. On the other hand, the initial inhibition of I Ks during agonist application appeared to be more pronounced upon stimulation of M 1 receptors as compared with ␣ 1B -adrenergic receptors (Fig. 3, A and C). The corresponding DAG dynamics in ␣ 1B -AR-express-ing cells showed a rapid decline during agonist application and had almost returned to basal levels when the ␣ 1B -AR-induced I Ks increase occurred (Fig. 3, B and D). In contrast, stimulation of M 1 receptors induced a more pronounced I Ks inhibition (about 30% of the initial current amplitude in the absence of agonist) as compared with stimulation of ␣ 1B receptors (about 15%, Fig. 3C). DAG production persisted in the presence of acetylcholine (Fig. 3, B and D   Periods of time plotted in blue correspond to times of exposure to agonists. Note the rapid decay of the FRET ratio in the presence of agonist in ␣ 1B -AR-but not in M 1 -R-expressing cells. C, summarized data of DAG reduction during agonist application (determined by the ratio FRET 30 s after peak /FRET peak ), indicating a more than 50% reduction of the DAG signal during activation of ␣ 1B receptors (n ϭ 10), but only a minor reduction (less than 10%) during stimulation of M 1 receptors (n ϭ 12). Significant differences are indicated by asterisks. D and E, representative FRET recordings of HEK 293 cells transfected with 0.25 g (n ϭ 7) or 0.5 g ␣ 1B -AR cDNA (n ϭ 7) and DAGR during application of phenylephrine (1 M). As shown by the summarized data in F comparing DAG reduction (measured 30 s after FRET peak ), decreasing or increasing the expression level of ␣ 1B receptors does not affect the rapid desensitization properties (0.1 g (n ϭ 4) or 1 g of ␣ 1B -AR cDNA (n ϭ 7)). G and H, representative FRET recordings of HEK 293 cells transfected with 0.5 g (G, n ϭ 7) or 1 g of M 1 -R cDNA (H, n ϭ 7) and DAGR. Kinetics of DAG formation during application of acetylcholine (10 M) were not affected by decreasing or increasing the expression level of M 1 receptors (see also summarized data in I; 0.1 g, n ϭ 6, 0.25 g, n ϭ 5). *, p Ͻ 0.01. n.s. ϭ not significant. Error bars indicate mean Ϯ S.E. of n cells.
agonist withdrawal. (Fig. 3, A and C). These data suggest that receptor species-dependent differences in G q effector signaling account for divergent I Ks modulation, supporting the idea that desensitization of different receptor species determines temporal aspects of downstream G q signaling pathways.
Qualitatively, the temporal aspects of DAG formation during G q PCR activation are identical in CHO and HEK cells (compare representative traces in Figs. 2 and 3B; see also supplemental of the cellular background. Moreover, if the receptor-specific time course of the DAG signal reflects intrinsic receptor properties, receptor species-dependent desensitization should be evident at other branches of the G q signaling pathway, e.g. the level of PLC activation and depletion of membrane PIP 2 .

Stimulation of ␣ 1B Receptors in CHO and HEK Cells Induced a Membrane PIP 2 Depletion That Rapidly Recovered in the
Presence of Phenylephrine-We investigated the depletion of membrane PIP 2 following ␣ 1B or M 1 receptor stimulation in CHO and HEK cells by utilizing a biosensor that directly reports the depletion of membrane PIP 2 during G q PCR/PLC activation with a decrease in FRET ratio (21,22). As illustrated in a representative FRET recording in Fig. 4A, stimulation of ␣ 1B receptors in CHO cells caused a rapid decrease in FRET ratio, reflecting membrane PIP 2 depletion. This ␣ 1B -AR-mediated depletion showed a biphasic time course: during application of phenylephrine (1 M), FRET slowly decreased for about 20 s, followed by a rapid decay. Remarkably, the FRET ratio rapidly returned to baseline levels, indicating PIP 2 replenishment during sustained receptor stimulation. In contrast, exposure of M 1 -R-transfected CHO cells to acetylcholine (10 M) resulted in a fully reversible decrease in the FRET ratio ( Fig. 4B) with a rapid onset of PIP 2 depletion that was persistent in the presence of the agonist. Comparing membrane PIP 2 depletion in HEK 293 cells during ␣ 1B and M 1 receptor stimulation yielded analogous results, i.e. transient reduction and replenishment of PIP 2 in the presence of phenylephrine and persistent PIP 2 depletion during application of acetylcholine (Fig. 4, D and E). The summarized FRET 30 s after peak /FRET peak ratios in Fig. 4, C and F, indicate a reduction of 56% of the PIP 2 signal in CHO cells and a reduction of 32% in HEK 293 cells during activation of ␣ 1B receptors and a minor reduction (less than 5%) during stimulation of M 1 receptors in both cell types.
To address whether internalization of the ␣ 1B receptor causes its fast desensitization, we analyzed the time courses of arrestin binding as well as of subsequent receptor internalization. Binding of arrestins represents the initial step to prime a GPCR for endocytosis. Of note, the recruitment of arrestins to the receptor prevents G protein binding (23). Thus, recruitment of arrestins could contribute to desensitization of G q signaling. To estimate the kinetics of arrestin binding for stimulated ␣ 1B -ARs and M 1 -Rs, we used a FRET assay that monitors the recruitment of Turquoise-labeled ␤-arrestin to receptors that were labeled with YFP at their C terminus (Fig. 5A) (24). The biosensor revealed that recruitment of arrestin occurred fast and with similar kinetics for both receptor subtypes (Fig. 5, B and C). Therefore, arrestin binding cannot explain the differences in desensitization kinetics of both receptors. Furthermore, the internalization of ␣ 1B -AR during long-term exposure to Phe was even slower; in the absence of agonist, we observed a proper membrane localization of ␣ 1B -AR-YFP-IL3 (Fig. 5D,  panel a). During exposure to 10 M Phe, internalization of the receptor was evident by a reduction in plasma membrane fluorescence and formation of intracellular punctae, starting between 15 and 30 min of incubation time (Fig. 5D, panels d-g).  This is in line with the time course for internalization of this receptor in a previous study (25) and is significantly slower (up to 15 min) than the time course of acute desensitization observed in our experiments (e.g. see Fig. 2 with complete desensitization of ␣ 1B receptors within 60 s of agonist application). These results confirm that neither recruitment of arrestins nor an internalization of the ␣ 1B -AR caused its acute desensitization. We further investigated recovery from ␣ 1B -AR desensitization by consecutive applications of phenylephrine (1 M) with variable time intervals. The second Phe-induced DAG signal was normalized to the first DAG signal and plotted against recovery time (defined as the time interval between termination of the first DAG signal and increase of the second signal). As shown in Fig. 5E, up to 80% recovery of the DAG signal was achieved within 300 s. Although we did not determine the molecular mechanism of recovery from desensitization, its rapid time course excludes receptor endocytosis/ recycling as the underlying mechanism. Thus, desensiti- zation and recovery from desensitization are most likely related to direct modifications of the receptor protein, such as phosphorylation/dephosphorylation.
Contribution of PKC Isoforms to Acute ␣ 1B Receptor Desensitization-Receptor phosphorylation has been shown to be the earliest biochemical event in homologous and heterologous ␣ 1B receptor desensitization either by G protein-coupled receptor kinases (GRK) or by PKC (26). Both conventional PKC isoforms (cPKCs, Ca 2ϩ -and DAG-dependent) and novel PKC isoforms (nPKCs, DAG-dependent, but Ca 2ϩ -independent) are downstream effectors of G q -coupled receptor signaling pathways (27), but their contribution to homologous desensitization of ␣ 1B -R is still undefined. We aimed to investigate the contribution of different PKC isoforms to homologous G q PCR desensitization either by isoform-specific pharmacological PKC inhibition or by manipulating PKC expression levels.
In a first series of experiments, the contribution of PKC to homologous ␣ 1B receptor desensitization was analyzed in HEK cells cotransfected with ␣ 1B -AR and DAGR by application of the protein kinase inhibitor staurosporine. In the continuous presence of staurosporine (100 nM, 2-h incubation time, with staurosporine in the recording solution), the acute phase of ␣ 1B receptor desensitization was completely abrogated (see the representative FRET recordings in Fig. 6, A and B). As indicated by the summarized data in Fig. 6G, comparing the ratio FRET 30 s after peak /FRET peak in the presence or absence of staurosporine, reduction of the DAG signal during agonist application was markedly diminished by staurosporine.
Staurosporine is a nonspecific protein kinase inhibitor that does not discriminate between PKC isoforms. To evaluate the contribution of different PKC isoforms to homologous desensitization of ␣ 1B receptors, we either inhibited cPKCs (Fig. 6, E and F) or coexpressed the nPKC isoform PKC␦ as WT PKC␦ (Fig. 6C) or as the inactive mutant PKC␦ DN (Fig. 6D). Activation of ␣ 1B receptors in the presence of either WT PKC␦ or PKC␦ DN resulted in DAG formation that rapidly declined to a similar rate as compared with control conditions (see also Fig.  6G), suggesting that PKC␦ does not contribute to the acute phase of ␣ 1B receptor desensitization.
To reduce activation of cPKCs, we dialyzed ␣ 1B -AR/DAGRcotransfected HEK cells with the Ca 2ϩ chelator BAPTA (5 mM) via the patch pipette and measured the time course of DAG decay in the presence of 1 M Phe. As depicted in Fig. 6, E and G, buffering the increase in [Ca 2ϩ ] i by BAPTA abolished the acute phase of ␣ 1B receptor desensitization. The PKC inhibitor Gö6976 inhibits the cPKC isoforms PKC␣ and PKC␤I with an IC 50 of 2.3 and 6.2 nM without affecting the activity of PKC␦ even at high concentrations in the micromolar range (28), thus providing a reliable pharmacological tool to disrupt the activity of cPKCs. In the presence of 10 nM Gö6976 (2-h incubation plus Gö6976 in the recording solution), application of Phe induced DAG formation that did not decline in the presence of agonist, indicating that inhibition of cPKCs eliminated acute desensitization (see also Fig. 6G).
Overexpression of WT PKC␦ Induced Rapid Desensitization of ␣ 1A -AR-induced DAG Formation-The consistent kinetics of ␣ 1B -AR-induced DAG signaling upon overexpression of WT PKC␦ might exclude a contribution of PKC␦ to acute ␣ 1B recep-tor desensitization. To exclude inadequate PKC␦ overexpression above the endogenous protein levels, we tested the impact of PKC␦ overexpression on the function of the closely related ␣ 1A -AR. We investigated DAG kinetics in ␣ 1A -AR-expressing HEK cells either under control conditions (Fig. 7A) or during coexpression of the WT PKC␦ (Fig. 7B) or during expression of the catalytically inactive mutant PKC␦ DN (Fig. 7C). The representative FRET recording in Fig. 7A shows a slow decline of the DAG signal during stimulation of ␣ 1A -AR with Phe (1 M). The summarized data in Fig. 7F, expressed as the ratio FRET 30 s after peak /FRET peak , indicate a minor reduction of about 12% during activation of ␣ 1A receptors under control conditions, but a rapid decline of DAG production upon overexpression of WT PKC␦ (Fig. 7B, see also Fig. 7F). The pronounced reduction of DAG signaling (more than 40%) reflects PKC␦-induced ␣ 1A -AR desensitization, which was only observed in the presence of coexpressed PKC␦. To exclude an artifact of protein overexpression, we cotransfected ␣ 1A -ARexpressing cells with the PKC␦ DN and monitored DAG production. In the presence of the mutant PKC␦ DN, DAG kinetics resemble those observed under control conditions (Fig. 7, C and F), supporting the previous finding that the inactive kinase PKC␦ DN specifically inhibits PKC␦-induced signaling pathways (29).
To investigate the contribution of endogenous PKC isoforms to ␣ 1A -AR desensitization, we monitored the time course of the ␣ 1A -AR-induced DAG signal in the continuous presence of staurosporine (Fig. 7D) or Gö6976 (Fig. 7E). As shown by the summarized data in Fig. 7F, comparing the ratio FRET 30 s after peak /FRET peak , reduction of the DAG signal during agonist application was significantly diminished by staurosporine, but not by Gö6976.
These data indicate a receptor species-dependent and PKC isoform-specific regulation. Inhibition of cPKCs with Gö6976 abrogates desensitization of ␣ 1B -ARs, but not of ␣ 1A -ARs. Furthermore, although ␣ 1A -AR signaling was regulated by PKC␦, there was no direct evidence for a regulation of ␣ 1B -AR by PKC␦.
Signaling of M 1 -R Was Not Affected by Intracellular Kinases-Agonist-induced desensitization of muscarinic M 1 receptor requires GRK-dependent receptor phosphorylation and recruitment of ␤-arrestins, resulting in clathrin-dependent receptor internalization (30) (for review, see Ref. 31). Stimulation of PKC has been shown to exert opposing effects on M 1 receptor activity, e.g. enhanced acute receptor desensitization (32), a loss of cell surface M 1 receptors, or even no effect on M 1 receptor desensitization or internalization (31). Analogous to the experiments on ␣ 1B -R/DAGR-cotransfected cells (Fig. 6), we investigated the contribution of different PKC isoforms to M 1 -R signaling by pharmacological inhibition of PKC or coexpression of the wild-type PKC␦. As depicted in Fig. 8, the time course of DAG formation during ACh exposure (10 M) in the presence or absence of the PKC inhibitors staurosporine and Gö6876 and upon overexpression of PKC␦ was indistinguishable, indicating that M 1 -R signaling at the level of PLC activation is not modulated by PKC. We did not investigate whether increasing the PKC␦ expression level had long-term effects on the number of M 1 receptors at the cell surface. However, during DECEMBER   our experiments (with a duration of 300 -600 s), receptor internalization is unlikely to occur because we did not observe a PKC-induced reduction in M 1 receptor responsiveness.

Pharmacological PKC Inhibition Reduced the Acute Desensitization of ␣ 1B -AR and Modulated I Ks Facilitation-Recent
studies on G q PCR-induced modulation of I Ks suggest fundamental differences in the biophysical channel modulation by Ca 2ϩ -dependent and Ca 2ϩ -independent PKC isoforms (13,20), but to date, temporal aspects of cPKC and nPKC activation and their contribution to I Ks modulation are mainly elucidated from kinetic models (13).
Our data suggest that receptor species-dependent differences in desensitization account for a receptor-specific recruitment of different PKC isoforms downstream of PIP 2 depletion and a PKC-induced feedback modulation of receptor activity. Apart from modulating homologous receptor desensitization, the recruitment of different PKC isoforms during G q PCR activation might shape the temporal signaling properties of downstream G q signaling effectors.
To address this, we analyzed the modulation of I Ks amplitude in KCNQ1/KCNE1 CHO cells and monitored ␣ 1B -AR function simultaneously (DAGR-FRET biosensor) in the presence of PKC inhibitors. As shown previously in Fig. 3A, the ␣ 1B -ARinduced inhibition of I Ks in the absence of PKC inhibitors was transient and turned into delayed facilitation after agonist withdrawal. Nonspecific inhibition of PKC isoforms with staurosporine reduced the rapid decay of DAG formation in the presence of phenylephrine (Fig. 9, B and D). Furthermore, PKC inhibition with staurosporine increased ␣ 1B -R-induced I Ks reduction and abolished I Ks facilitation (Fig. 9, A and C).
In the presence of Gö6976, the DAG signal during stimulation of ␣ 1B receptors was significantly prolonged (see FRET ratio in Fig. 9B and summarized data in 9D), indicating a reduction of homologous ␣ 1B -AR desensitization. As depicted in Fig. 9E (showing the time courses of ␣ 1B -ARinduced DAG signals on an expanded time scale), the attenuation of ␣ 1B -AR desensitization became evident during sustained agonist application. For example, after 40 s of agonist   DECEMBER 16, 2016 • VOLUME 291 • NUMBER 51 JOURNAL OF BIOLOGICAL CHEMISTRY 26419 application ( Fig. 9E at t ϭ 145 s), the Phe-induced DAG signal was decreased to 38% in the absence, but only to 65% in the presence of Gö6976. Moreover, pharmacological inhibition of Ca 2ϩ -dependent PKC isoforms increased the ␣ 1B -AR-induced I Ks reduction (Fig. 9C), but did not abolish I Ks facilitation after agonist withdrawal. These data indicate a contribution of different PKC isoforms to the time course of ␣ 1B -AR-induced I Ks modulation. Activation of Ca 2ϩ -dependent PKC isoforms induces rapid ␣ 1B -AR desensitization, which delimits PIP 2 depletion, resulting in moderate I Ks inhibition. Disruption of cPKC activation prolongs ␣ 1B -AR-induced PIP 2 depletion, thus augmenting I Ks inhibition (compare the extent of current inhibition indicated by dotted lines in Figs. 3C, left panel,  and 9C). However, a contribution of cPKCs to the current increase after agonist withdrawal can be excluded, because I Ks facilitation was still observed during pharmacological inhibition of cPKCs. In contrast, inhibition of cPKCs and nPKCs by staurosporine abolished I Ks facilitation, suggesting that activation of Ca 2ϩ -independent PKC isoforms mediates I Ks increase.

Desensitization Controls I Ks as an Effector of G q PCRs
The Rapid Decline of DAG Formation Is Reduced in HEK Cells Expressing the ␣ 1B /␣ 1A -CT chimera as Compared with ␣ 1B -AR-induced DAG Signals-Phosphorylation sites at the C terminus of ␣ 1B -AR have been shown to be critically involved in receptor desensitization (33)(34)(35) and receptor endocytosis (36). These phosphorylation sites are not conserved among the other ␣-adrenergic receptor subtypes. As shown previously, ␣ 1A -ARs are less sensitive to agonist-induced desensitization and are phosphorylated to a lesser extent as compared with ␣ 1B -AR (34). Because former studies were not designed to analyze dynamic temporal aspects of desensitization, we coexpressed chimeric ␣ 1B -AR carrying the C terminus of ␣ 1A -AR (␣ 1B / ␣ 1A -CT chimera) together with DAGR in HEK 293 cells and compared the agonist-induced DAG dynamics among wildtype ␣ 1B -AR, ␣ 1A -AR, and the ␣ 1B /␣ 1A -CT chimera (Fig. 10). The representative FRET recordings in Fig. 10 show the effects of G q PCR stimulation on DAG dynamics during application of Phe (1 M). DAG production in HEK 293 cells rapidly decayed during stimulation of ␣ 1B receptors (Fig. 10A), but was significantly prolonged in ␣ 1A -AR-and ␣ 1B /␣ 1A -CT chimera-expressing cells (Fig. 10, B and C). The summarized data, expressed as the ratio FRET 30 s after peak /FRET peak in Fig. 10E, indicate a more than 50% reduction of the DAG signal during activation of ␣ 1B receptors in line with the data shown in Fig. 2, and a significantly smaller reduction during stimulation of ␣ 1A receptors (about 15%, the same data as depicted in Fig. 7D) and ␣ 1B /␣ 1A -CT chimera receptors (about 27%). During sustained agonist application (Ͼ60 s), the DAG signal declined almost to baseline in ␣ 1B -AR-expressing cells, but decayed with a slower time course during stimulation of ␣ 1A -AR-and ␣ 1B /␣ 1A -CT chimera-AR. As indicated by the summa-rized data in Fig. 10F, comparing the ratio FRET 60 s after peak / FRET peak , the desensitization characteristics of ␣ 1B /␣ 1A -CT chimera-expressing cells resemble those observed in wild-type ␣ 1A -AR-expressing cells, although desensitization of ␣ 1B / ␣ 1A -CT chimera-AR occurred slightly faster than those of wildtype ␣ 1A -AR (Fig. 10F). Furthermore, as for the wild-type ␣ 1A -AR, overexpression of WT PKC␦ induced pronounced desensitization of ␣ 1B /␣ 1A -CT chimera receptors as indicated by the rapid decline of the DAG signal (Fig. 10D).
Because the time course of desensitization induced by the chimera is an intermediate between the ␣ 1B -AR and ␣ 1A -AR desensitization phenotypes, it is proposed that additional regions, apart from the C terminus, contribute to ␣ 1B receptor desensitization. Although the identification of additional mechanisms of ␣ 1B -AR desensitization is beyond the scope of the present study, the significant reduction of desensitization in chimeric ␣ 1B receptors carrying the ␣ 1A -AR C terminus emphasizes that the number of phosphorylation sites (i.e. the degree of receptor phosphorylation) is directly related to the extent of acute desensitization of the receptors.

Discussion
Spatial and temporal organization of signaling components downstream of G␣ q activation is related to distinct cellular responses upon activation of different G q PCRs (2). Receptorspecific cellular events might be induced by promiscuous coupling of GPCRs to G␣ q family members, interaction of G␣ q with GRK2 or regulators of G protein Signaling (RGS) proteins, or different affinities of G␣ q effector activation (2). Furthermore, functional diversity upon stimulation of G q PCRs can be determined by spatial proximity of G q PCRs and signal molecules (3),   by activating different branches of the signaling pathway downstream of G␣ q (4,5), or by connecting signaling enzymes to their substrates in specific microdomains (8).
The aspect of receptor species-dependent desensitization that determines temporal aspects of downstream G protein signaling has not been elucidated so far. In our study, we investigated the receptor-specific desensitization of ␣ 1B and M 1 receptors as paradigmatic G␣ q -coupled receptors (37,38).
We provide evidence that receptor-specific desensitization shapes the kinetics of G q signaling, resulting in receptor species-dependent modulation of effectors. By using a variety of FRET-based biosensors, we demonstrated receptor species-dependent differences in desensitization on all levels of downstream G q PCR signaling, including G protein activation (Fig. 1), PIP 2 depletion upon activation of PLC (Fig. 4), and DAG formation (Fig. 2).
Several lines of evidence support the idea that receptor-specific desensitization reflects intrinsic receptor properties rather than being related to insufficient receptor-G protein coupling or insufficient effector activation. First, receptor-specific desensitization of ␣ 1B and M 1 receptors was not abolished upon decreasing or increasing their respective expression levels ( Fig.  2) and was present at saturating and moderate agonist concentrations at a given receptor expression level (compare 100 nM versus 1 M responses in Fig. 1). Second, as yielded by the concentration-response curves for ␣ 1B -AR-and M 1 -R-induced G q activation (Fig. 1), the agonist concentrations used in the present study induced full activation of G proteins and downstream G␣ q components. Third, similar temporal aspects of ␣ 1B and M 1 receptor signaling were evident in CHO cells stably expressing KCNQ1/KCNE1 and HEK 293 cells, excluding cell line-dependent differences in rate and extent of receptor desensitization (Fig. 4).
Transient ␣ 1B -R activity induces a rapid decline of signaling events downstream of ␣ 1B receptor activation, e.g. on the level of DAG production. The biosensor DAGR has been validated to report the kinetics of G q PCR regulation by dynamic DAG formation (39), which correlates with the time course of receptor desensitization. Furthermore, it has been demonstrated that manipulations that impair receptor desensitization (e.g. altering receptor expression levels, phosphorylating GPCRs, or silencing GRK2 by siRNA) lead to sustained DAG formation (39).
GPCR desensitization involves biochemical processes with different time frames, including receptor phosphorylation by G protein-coupled receptor kinases or protein kinase C as the earliest biochemical event, receptor internalization, and with a longer time course, changes in the receptor expression level (26). By using site-directed mutagenesis, desensitization-associated serine residues, which are targeted either by GRK2 or by PKC, have been identified in the C terminus of the ␣ 1B receptor (35). Although truncation studies demonstrated that homologous desensitization of the ␣ 1B -AR depends unequivocally on PKC-induced phosphorylation of serine residues in the C terminus (35,40), conflicting results were reported upon the contribution of GRK to ␣ 1B -AR phosphorylation and desensitization (41,42), depending on the cell expression system.
Mutational studies on hamster ␣ 1B -AR indicate that PKC activation and phosphorylation of residues within the C-tail promote rapid receptor internalization (43). More recent studies propose that interactions of ␣ 1B -AR with ␤-arrestin and the clathrin adaptor complex AP2 (36) or with Rab4 and Rab5 proteins that are associated with early endosomes (25) regulate receptor internalization. All receptor modifications stated above may result in termination of receptor signaling by internalization of the receptors.
PKC phosphorylation sites are not conserved among the other ␣-adrenergic receptor subtypes, and differences in receptor desensitization and internalization between ␣ 1B -AR and ␣ 1A -AR (with modest internalization of ␣ 1A -AR as compared with ␣ 1B -AR) are mainly related to receptor subtype-dependent differences in ␤-arrestin binding (36). A recent study (44) describes that differences in the ␣-AR subtype internalization account for activation of different, subtype-specific downstream G q PCR effectors. As revealed by our data (Fig. 5), receptor internalization (with a time course of Ն15 min) does not contribute to the acute phase of ␣ 1B -AR desensitization (see also Fig. 2), which was almost complete within 60 s of agonist application. In addition, we were able to elicit DAG signals upon repetitive agonist application (Fig. 2), excluding a significant reduction in the number of receptor molecules on the cell surface. Furthermore, the fast recovery from desensitization (Fig. 5) suggests that direct modification of the receptor protein instead of receptor endocytosis is the underlying mechanism of receptor desensitization.
It was not the aim of the present study to investigate ␣ 1A -AR desensitization, but previous studies showed significant differences in the extent of phosphorylation between ␣ 1A -AR and ␣ 1B -AR (34). ␣ 1A -ARs are phosphorylated to a lesser extent as compared with ␣ 1B -ARs and display only minor desensitization. These data are in line with our observation that overexpression of the wild-type PKC␦, but not of the inactive mutant PKC␦ DN, rapidly terminated DAG signaling during ␣ 1A -AR stimulation (Fig. 7). Furthermore, these experiments demonstrate the efficiency of PKC␦ overexpression. Apparently, overexpression of the wild-type PKC␦ augments ␣ 1A -AR desensitization. However, the contribution of different PKC isoforms to endogenous ␣ 1A -AR signaling is evident from our experiments shown in Figs. 6 and 7. Although staurosporine abrogates desensitization of both ␣ 1A and ␣ 1B receptors, a contribution of cPKCs to ␣ 1A -AR desensitization can be excluded. Inhibition of cPKCs with Gö6976 abolishes acute desensitization of ␣ 1B but not of ␣ 1A receptors (Figs. 6 and 7), indicating receptor-specific regulation by different PKC isoforms.
Moreover, our experiments investigating the kinetics of ␣ 1B / ␣ 1A -CT chimera-induced desensitization indicate that the origin of the ␣-AR subtype C terminus determines the time course and kinetics of receptor desensitization. These data are in line with a previous study (34) on chimeric ␣ 1A and ␣ 1B receptors, carrying the mutual C-terminal region of each receptor subtype. Chimeric ␣ 1A -AR displayed marked basal and agonistinduced phosphorylation (in contrast to the wild-type ␣ 1A -AR), whereas the opposite was observed in chimeric ␣ 1B -AR (34). The rapid decline of the ␣ 1B /␣ 1A -CT-induced DAG signal upon overexpression of WT PKC␦ supports the idea that the C ter-that is based on CFP-and YFP-labeled PH domains of PLC␦1 (22). To monitor the production of the second messenger DAG, we used receptor species as indicated (0.5) and the biosensor DAGR (0.5), which reports conformational changes of a CFP/YFP-labeled DAG-binding domain of protein kinase C (PKC␤2) (18). DAGR was kindly provided by Dr. Alexandra Newton (Addgene plasmid number 14865). For some experiments, cells were cotransfected with 0.5 g of a plasmid encoding for WT PKC␦ or PKC␦ DN (29), kindly provided by Dr. Bernard Weinstein via Addgene (WT, plasmid number 16386, and DN, plasmid number 16389). Functionality of WT PKC␦ and PKC␦ DN was confirmed in experiments on KCNQ1/ KCNE1-expressing CHO cells measuring I Ks modulation during ␣ 1B receptor stimulation (see supplemental Fig. S4). Both cell lines were transfected using either polyethyleneimine as described in Ref. 22 or Lipofectamine (Invitrogen) according to the manufacturer's instructions. Prior to experiments, cells were seeded on sterile, poly-L-lysine-coated glass coverslips and analyzed 24 h (cells expressing DAGR) or 48 h after transfections.
Fluorescence Microscopy and Imaging-All experiments were performed using single cells at ambient temperature. Fluorescence was recorded using an inverted microscope (Zeiss Axiovert 200, Carl Zeiss AG, Göttingen, Germany) equipped with a Zeiss oil immersion objective (100ϫ/1.4), a Polychrome V illumination source, and a photodiode-based dual emission photometry system suitable for CFP/YFP FRET (TILL Photonics/FEI GmbH, Munich, Germany). For FRET measurements, single cells were excited at 435-nm wavelength with light pulses of variable duration (10 -50 ms) at a frequency of 5 Hz to minimize photobleaching. Corresponding emitted fluorescence from CFP (F 480 or F CFP ) or from YFP (F 535 or F YFP ) was acquired simultaneously, and FRET was defined as ratio F YFP /F CFP . Fluorescent signals were recorded and digitized using a commercial hardware/software package (EPC10 amplifier with an integrated D/A board and Patchmaster software, HEKA, HEKA Elektronik, Lambrecht/Pfalz, Germany). Details on optical filters and beam splitters of the setup are given in Ref. 22. The individual FRET traces for obtaining concentration-response curves were normalized to the maximal response of the G protein biosensor at saturating agonist concentrations (FRET/ FRET 10 M ), denoted as FRET/FRET max ). All other traces were normalized to the initial ratio value before agonist application (FRET/FRET 0 ). For receptor internalization experiments, YFP was excited at 500 nm, and fluorescence images were acquired with a Zeiss AxioCam MRm epifluorescence camera and corresponding AxioVision software. After application of phenylephrine, consecutive pictures were taken after the incubation times as indicated.
Current Measurement-Membrane currents were measured using whole-cell patch clamp technique. Pipettes were fabricated from borosilicate glass and filled with the solution listed below (direct current resistance, 4 -6 megaohms). Currents were measured by means of a patch clamp amplifier (LM/EPC 7, List Electronics, Darmstadt, Germany). Signals were filtered (corner frequency, 1 KHz), digitally sampled at 1 KHz, and stored on a computer equipped with a hardware/software package (ISO2, MFK, Frankfurt/Main, Germany) for voltage control and data acquisition. Experiments were performed at ambient temperature (23-26°C). For combined patch clamp and FRET measurements, standard patch clamp equipment was attached to the optical setup. Application of different solutions was performed by means of a custom-made solenoid-operated flow system. Whole-cell I Ks was routinely measured during depolarizing pulses (to ϩ 60 mV, duration 5 s, applied every 20 s) from a holding potential of Ϫ80 mV (see also supplemental Fig. S5).
Statistical Analysis-All data are presented as individual observations or summarized data (mean Ϯ S.E. of n cells). Student's t test was used to compare the means between two groups. p values less than 0.05 were considered statistically significant.