Imaging of Muscarinic Acetylcholine Receptor Signaling in Hippocampal Neurons: Evidence for Phosphorylation-Dependent and -Independent Regulation by G-Protein-Coupled Receptor Kinases

Abstract

We used the inositol 1,4,5-trisphosphate (IP₃) biosensor, the pleckstrin homology (PH) domain of PLCδ1 (phospholipase C) tagged with enhanced green fluorescent protein (eGFP-PH_PLCδ), to examine muscarinic acetylcholine (mACh) receptor regulation of phospholipase C/IP₃ signaling in intact single hippocampal neurons in “real time.” Initial experiments produced a pharmacological profile consistent with the presence of a predominant M₁ mACh receptor population coupled to the IP₃ response. To investigate M₁ mACh receptor regulation, neurons were stimulated with approximate EC₅₀ concentrations of the mACh receptor agonist methacholine before (R1) and after (R2) a short (60 sec) exposure to a high concentration of agonist. This resulted in a marked attenuation in the R2 relative to R1 response. Inhibition of endogenous GRK6 (G-protein-coupled receptor kinase) activity, by the introduction of catalytically inactive ^K215RGRK6, partially reversed the attenuation of agonist-induced responsiveness, whereas overexpression of wild-type GRK6 increased receptor desensitization. Manipulation of endogenous GRK2 activity through introduction of either wild-type or catalytically inactive GRK2 (^K220RGRK2) almost completely inhibited agonist-stimulated IP₃ production, implying a phosphorylation-independent regulation of M₁ mACh receptor signaling, most probably mediated by a GRK2 N-terminal RGS-like (regulator of G-protein signaling) domain interaction with GTP-bound Gα_q/11. Together, our data suggest a role for both phosphorylation-dependent and -independent regulation of M₁ mACh receptors in hippocampal neurons.

Introduction

Neuronal G-protein-coupled receptors (GPCRs) play important roles in many aspects of brain function. In addition to modulating neurotransmission at CNS synapses, GPCRs are also instrumental in different forms of synaptic plasticity and regulate processes such as gene transcription (Berridge, 1998; Greengard, 2001; Katz and Clemens, 2001). A fundamental property of GPCRs is an ability to adapt to different patterns of stimulation. Prolonged or recurrent activation of GPCRs results in a consequent attenuation of signaling. Classically, this type of GPCR regulation is initiated by phosphorylation of the receptor by agonist-dependent (homologous) or -independent (heterologous) mechanisms (Hausdorff et al., 1990; Ferguson, 2001), brought about by GRKs (G-protein-coupled receptor kinases) and second-messenger-regulated kinases, respectively. Receptor phosphorylation may attenuate signaling per se and/or may facilitate recruitment of arrestin proteins (Krupnick and Benovic, 1998). The phosphorylated receptor-arrestin complex can be internalized and may also act as an adaptor scaffold to recruit other signaling pathways (Ferguson, 2001).

GRKs and second-messenger-regulated kinases are expressed at high levels in neurons (Arriza et al., 1992; Erdtmann-Vourliotis et al., 2001; Grange-Midroit et al., 2002); however, we still know relatively little about whether GPCRs, endogenously expressed in neurons, are regulated by GRK-dependent mechanisms. McConalogue et al. (1998) used an immunocytochemical approach to provide evidence for a GRK2/3-dependent desensitization-internalization of NK₁ neurokinin receptors in guinea pig myenteric neurons, whereas Kouznetsova et al. (2002) have reported that CB₁ cannabinoid receptor desensitization can be attenuated by the expression of a dominant-negative GRK2 mutant in rat hippocampal neurons. Evidence for the involvement of GRK4, GRK5, and GRK6 in neuronal GPCR regulation has also been presented. Thus, De Blasi and colleagues have made a strong case for GRK4 involvement in the desensitization of type 1 metabotropic glutamate receptors in Purkinje cells (Sallese et al., 2000; Iacovelli et al., 2003), and targeted GRK5 (Gainetdinov et al., 1999) and GRK6 (Gainetdinov et al., 2003) gene knock-out by homologous recombination has provided evidence for an involvement of these GRKs in muscarinic acetylcholine (mACh) receptor and dopamine receptor regulation in vivo, respectively.

There is substantial evidence for an M₁ mACh receptor population in the hippocampus and in particular a potential role of this GPCR subtype in mechanisms underlying longer-term regulation of synaptic function. In the present study, we examined potential roles of GRK2, GRK5, and GRK6 in the regulation of M₁ mACh receptor signaling in single hippocampal neurons using an IP₃ biosensor imaging approach (Stauffer et al., 1998; Nahorski et al., 2003). Our data reveal a potential role for both phosphorylation-dependent and -independent regulation of mACh receptor signaling on the basis of manipulations of GRK2 and GRK6 activities through the introduction of wild-type or dominant-negative GRK mutants.

Materials and Methods

Cell culture and transfections. Hippocampal neurons from 1-d-old Lister hooded rat pups were isolated as described previously (Schell et al., 2001). Briefly, isolated hippocampi were dissociated with Pronase E (0.5 mg/ml) and thermolysin (0.5 mg/ml) in HBSS (in mm: 130 NaCl, 10 HEPES, 5.4 KCl, 1.0 MgSO₄, 25 glucose, and 1.8 CaCl₂, pH 7.2) for 30 min. Tissue fragments were further dissociated by trituration in HBSS containing DNase I (40 μg/ml). After centrifugation and further trituration, cells were plated onto poly-d-lysine (50 μg/ml)-treated 25 mm glass coverslips. For the first 72 hr, cells were cultured in Neurobasal medium (Invitrogen, Paisley, UK), supplemented with B27 and 10% fetal calf serum. Cytosine arabinoside (5 μm) was added after 24 hr, and, after 72 hr, cells were transferred to serum-free medium. Cultured cells were transfected on day 5 with a 3:1 ratio of either vector control or GRK constructs, to eGFP-PH_PLCδ, respectively using the Lipofectamine 2000 reagent (Invitrogen) according to the instructions of the manufacturer.

Measurement of IP₃ in single cells and assessment of mACh receptor desensitization. Translocation of eGFP-PH_PLCδ [the pleckstrin homology (PH) domain of PLCδ1 (phospholipase C) tagged with enhanced green fluorescent protein (eGFP)] was visualized using an Olympus Optical (Europa, UK) FV500 scanning laser confocal IX70 inverted microscope. Cells were incubated at 37°C using a temperature controller and microincubator (PDMI-2 and TC202A; Burleigh Instruments) and perfused at 5 ml/min with Krebs' buffer (in mm: 119 NaCl, 4.7 KCl, 1.2 MgSO₄, 1.2 KH₂PO₄, 4.2 NaHCO₃, 10 HEPES, 11.7 glucose, and 1.3 CaCl₂, pH 7.4). Images were captured using an oil immersion 100× objective and 4.5× optical zoom. Cytosolic IP₃ levels were measured as the relative change in fluorescence detected in an area of interest as described previously (Nash et al., 2002). This approach detects IP₃ accumulation rather than phosphatidylinositol 4,5-bisphosphate (PIP₂) depletion, because selective removal of IP₃ (by coexpression of the enzyme IP₃ 3-kinase) completely prevents translocation of eGFP-PH_PLCδ (Nash et al., 2002; Nahorski et al., 2003). Drugs were applied via perfusion lines. Desensitization of the mACh receptor was assessed in hippocampal neurons transfected with eGFP-PH_PLCδ on day 5 in vitro (DIV) and used experimentally between days 7 and 10 in vitro. All experiments were undertaken in the presence of tetrodotoxin (500 nm) to block action potential-dependent synaptic activity. Desensitization was assessed in single cells using a slightly modified protocol to that published previously for the M₃ mACh receptor in SHSY5Y neuroblastoma cells (Willets et al., 2003a). Neurons were challenged with an approximate EC₅₀ concentration of the mACh receptor agonist methacholine (MCh) for 30 sec (termed R1), followed by a 5 min washout to allow recovery of PIP₂, [Ca²⁺]_i, and eGFP-PH_PLCδ fluorescence to basal levels. After this, a maximal concentration of MCh (100 μm) was applied for 1 min to induce receptor desensitization. The washout period after desensitization was varied before rechallenge with the same approximate EC₅₀ concentration of MCh (termed R2). Receptor desensitization was determined as the reduction in peak IP₃ formation in R2 when compared with R1.

Detection of endogenously expressed GRKs. After 7 d in culture, hippocampal neurons were lysed, and GRK expression was detected by Western blotting using specific rabbit polyclonal anti-GRK2, GRK3, GRK5, or GRK6 (Santa Cruz Biotechnology, Santa Cruz, CA) as described previously (Willets and Kelly, 2001).

Data analysis. Data were analyzed using one-way ANOVA (Excel 5.0; Microsoft, Redmond, WA), followed by Student's t test. Significance was accepted when p < 0.05.

Results

Characterization of mACh receptor desensitization in hippocampal neurons

Transfection of 5 DIV hippocampal cultures yielded up to 5% of cells expressing eGFP-PH_PLCδ, an example of which can be seen in Figure 1A. When stimulated with the mACh receptor agonist MCh (100 μm), the eGFP-PH_PLCδ construct translocated to the cytoplasm, producing a rapid peak, followed by a slowly fading plateau phase that rapidly decreased to prestimulation levels after addition of atropine (1 μm) (Fig. 1B) or agonist washout. Concentration-dependent increases in cytosolic eGFP_PLCδ fluorescence were seen for MCh (Fig. 1C). Pirenzepine (300 nm) and the M₁-selective toxin MT7 (100 nm) (Fig. 1D) fully reversed the MCh-induced increase in IP₃ in >75% cells investigated. These pharmacological data indicate the presence of a predominant or exclusive M₁ mACh receptor population in the majority of neurons studied.

Figure 1.

Imaging IP₃ in hippocampal neurons. A, Image of a hippocampal neuron expressing eGFP-PH_PLCδ. B, Representative trace showing IP₃ production, indicated by translocation of eGFP-PH_PLCδ from the membrane to the cytoplasm, in response to MCh (100 μm) and subsequent atropine (1 μm) addition. C, Representative trace (and hippocampal images) showing concentration-dependent increases in IP₃ generation in response to 30 sec additions of MCh (1, 3, 10, 30, and 100 μm) interspersed by 4 min washout periods. D, Representative traces showing MCh (100 μm, 30 sec)-stimulated IP₃ generation, which is inhibited in the same hippocampal neuron after MT7 (100 nm, 30 min preincubation) treatment. Data are representative of 10 separate experiments.

Consistent with the minimal mACh receptor desensitization in the continuing presence of 100 μm MCh (Fig. 1B), repeated applications of MCh (100 μm) for 1 min, interspersed with 5 min washes, also resulted in only a slight attenuation of the IP₃ response (Fig. 2A). The experimental protocol was modified to use a 30 sec application of a submaximal concentration of MCh before and after a near-maximal concentration of MCh (100 μm) applied for 1 min (Fig. 2B). Comparison of the responses before (R1) and after (R2) the 60 sec pulse of MCh revealed a clear reduction (∼50%) in the second response (Fig. 2B). Using R1 and R2 concentrations of MCh of 3, 10, or 30 μm produced similar degrees of desensitization (40-50%) (Fig. 2C). The extent of the reduction in mACh receptor signaling observed was also dependent on the washout period before the R2 challenge (Fig. 2C). Thus, the desensitization (R2/R1 ratio) was similar at 3 or 5 min after removal of 100 μm MCh but was decreased (i.e., R2 approaches R1 response) after 10 min of washout (Fig. 2C), suggesting that receptor resensitization can occur over this timescale. In all additional experiments, 5 min was used as the standard washout period between agonist additions.

Figure 2.

Assessment of mACh receptor desensitization in hippocampal neurons. A, Representative trace showing IP₃ responses to repeated 1 min additions of 100 μm MCh (horizontal bars) interspersed by 5 min washout periods. B, A representative trace (and hippocampal images) showing how the response to an approximate EC₅₀ concentration of MCh (R1/R2, 10 μm, 30 sec) is attenuated by the application of a near-maximal MCh addition (MCh_max; 100 μm, 1 min). C, Mean data using the protocol shown in B. The R2/R1 ratio was calculated for experiments in which 3, 10, or 30 μm MCh was used as the approximate EC₅₀ concentration. After the near-maximal MCh addition, the washout time before addition of the R2 pulse was varied (3, 5, or 10 min). Data are shown as means ± SEM for five to eight separate experiments on at least three different hippocampal preparations.

Detection of endogenous GRK expression

Western blot analysis of lysates prepared from 7 DIV hippocampal neurons showed the presence of immunoreactivity to GRK2, GRK3, GRK5, and GRK6 antibodies (Fig. 3A). A similar pattern of expression was observed in lysates prepared from hippocampus of 7- to 8-d-old rats (Fig. 3A).

Figure 3.

Effects of wild-type and dominant-negative GRK5 and GRK6 constructs on mACh receptor desensitization in hippocampal neurons. A, Western blot detection of GRK2, GRK3, GRK5, and GRK6 protein in 7 DIV hippocampal cultures (lane 1) or hippocampal homogenates (lane 2) prepared from 7- to 8-d-old rats. B, Representative trace of hippocampal neurons cotransfected with a 1:3 ratio of eGFP-PH_PLCδ and ^K215RGRK6. Additions (shown by the horizontal bars) were 10 μm MCh for 30 sec (R1), 100 μm MCh for 1 min (MCh_max), and 10 μm MCh for 30 sec (R2), interspersed by 5 min washout periods. C, Desensitization of mACh receptor-mediated IP₃ production in single cells transfected with the vector control (pcDNA3; n = 21 cells), ^K215RGRK5 (DN5; n = 21), wild-type GRK5 (GRK5; n = 6), ^K215RGRK6 (DN6; n = 18), or wild-type GRK6 (GRK6; n = 8). Data are shown as means ± SEM for the number of neurons indicated above taken from least three different hippocampal preparations. ^K215RGRK6 expression significantly attenuated desensitization (^**p < 0.01), whereas overexpression of GRK6 caused a significantly greater change in the R2/R1 ratio compared with controls (^*p < 0.05).

Effects of inhibiting GRK5 and GRK6 activities on mACh receptor desensitization

Hippocampal neurons were transfected with catalytically inactive, dominant-negative GRK5 and GRK6 constructs, which are mutated at a conserved lysine residue (K215R) required for ATP binding at the catalytic domain (Willets et al., 2002). To determine whether ^K215RGRK5 and ^K215RGRK6 were expressed, hippocampal neurons were transfected with eGFP-tagged versions and detected by confocal microcopy. Expression could still be detected 10 d after transfection (data not shown). Transfection with empty vector (pcDNA3), ^K215RGRK5, or ^K215RGRK6 had no effect on acute MCh-stimulated IP₃ production. Transfection of hippocampal neurons with pcDNA3 did not affect the extent of mACh receptor desensitization (Fig. 3C). In ^K215RGRK6-transfected neurons, the R2/R1 ratio difference was reduced, indicating that inhibition of endogenous GRK6 in hippocampal neurons attenuates mACh receptor desensitization (Fig. 3B,C). However, transfection with the closely related ^K215RGRK5 had no effect on mACh receptor responsiveness (Fig. 3C). Wild-type GRK6 was also coexpressed in neurons, and this resulted in a small but statistically significant increase in the R2/R1 ratio, suggesting that overexpression of GRK6 could further increase agonist-driven mACh receptor desensitization (Fig. 3C). In contrast, coexpression of wild-type GRK5 had no effect on M₁ mACh receptor desensitization (Fig. 3C).

Involvement of GRK2 in mACh receptor regulation

We and others have reported that GRK2 and GRK3 are able to inhibit receptor-mediated IP₃ formation independent of their receptor kinase activity (Carman et al., 1999; Willets et al., 2001, 2003a) but dependent on direct binding of the RGS-like (regulator of G-protein signaling) domain of GRK2/3 to Gα_q-GTP. To date, most of this work has been undertaken in cell lines with overexpressed kinases and/or receptors. In this study, we examined whether a similar phenomenon can be observed in hippocampal neurons. Transfection of wild-type GRK2 into hippocampal neurons markedly inhibited MCh-stimulated eGFP-PH_PLCδ translocation (Fig. 4). An equally marked inhibition of signaling was seen in neurons transfected with the catalytically inactive, dominant-negative ^K220RGRK2 (Fig. 4). These data indicate that the GRK2-mediated inhibition of mACh receptor signaling is not dependent on receptor phosphorylation.

Figure 4.

Effects of wild-type and dominant-negative GRK2 on mACh receptor-stimulated IP₃ production. A, Representative traces of hippocampal neurons cotransfected with a 1:3 ratio of eGFP-PH_PLCδ and empty vector (pcDNA3), wild-type GRK2 (GRK2), or dominant-negative/kinase-dead GRK2 (^K220RGRK2). The response to a single addition of MCh (100 μm, 60 sec) is shown.B, Inhibitory effects of GRK2 and ^K220RGRK2 overexpression on mACh receptor-mediated IP₃production in single neurons. Data are shown as means ± SEM for 6-11 neurons taken from at least three different hippocampal preparations. Both GRK2 and ^K220RGRK2 expression significantly attenuated the response to MCh (^**p < 0.01) compared with controls.

Discussion

The cholinergic innervation of the hippocampus is widespread and derives mainly from the septal nuclei (Rouse et al., 1999). Lesions to this pathway, or blockade of mACh receptors, can lead to memory and attentional deficits (Bartus et al., 1982; Power et al., 2003). Likewise, there is evidence for extensive expression of mACh receptors, particularly of the M₁ subtype, in both neuronal soma and dendrites in rat (Levey et al., 1995) and human (Shiozaki et al., 2001) hippocampal pyramidal cells. The M₁ mACh receptor subtype preferentially couples via Gα_q/11-proteins to activation of phospholipase C, and recent studies have revealed that agonist-stimulated Gα_q/11-[³⁵S]GTPγS binding is abolished in hippocampal membranes from M₁ mACh receptor knock-out mice (Porter et al., 2002). Furthermore, activation of mACh receptors on CA1 pyramidal neurons leads to IP₃-dependent Ca²⁺ waves that propagate from the dendrites to the soma, in which they invade the nucleus (Power and Sah, 2002). This action, and indeed the activation of extracellular signal-regulated kinases selectively by the M₁ mACh receptor (Berkeley and Levey, 2003), could be significant in cholinergic-induced changes in hippocampal synaptic plasticity.

We used a fluorescent biosensor, eGFP-PH_PLCδ, to image for the first time IP₃ generation stimulated by M₁ mACh receptor activation in single hippocampal neurons in culture. In particular, we focused on the potential regulation of mACh receptors by GRKs, and this has been facilitated by the ability to cotransfect specific GRK constructs with the biosensor and thus allow GRK/biosensor coexpressing neurons to be imaged. Continuous or repeated maximal stimulation of M₁ mACh receptors only led to a slight desensitization. However, in view of the receptor reserve observed for this subtype in hippocampus with respect to agonist-stimulated Gα_q/11-[³⁵S]GTPγS binding (Porter et al., 2002), we adopted a different protocol comparing submaximal agonist responses assessed before and after a 60 sec maximal stimulation. Under these conditions, we observed a significant attenuation of IP₃ generation in hippocampal neurons during rechallenge.

Our hippocampal cultures express GRK2, GRK3, GRK5, and GRK6, and manipulation of the expression-activity of these kinases had profound effects on M₁ mACh receptor-PLC signaling. Perhaps surprisingly, overexpression of GRK2 and the catalytically inactive ^K220RGRK2 mutant both markedly suppressed mACh receptor-mediated IP₃ generation. This strongly implies a phosphorylation-independent mechanism for this kinase in hippocampal neurons, consistent with several examples of such regulation at a number of GPCRs (Willets et al., 2003b). Currently, the most likely mechanism relates to the direct binding of GRK2 to activated GTP-bound Gα_q/11 through an RGS-like domain present in the N terminus of GRK2 (Sallese et al., 2000; Sterne-Marr et al., 2003). Recent findings using mutations in the RGS domain have revealed a novel sequence in GRK2/3, termed the C-site, which avidly binds Gα_q/11 and is absent from other GRKs (Sterne-Marr et al., 2003). Indeed, the crystallographic structure of GRK2 reveals three domains occupying the vertices of an essentially equilateral triangle with spacing such that each of the three domains could potentially interact simultaneously with the GPCR, Gα_q/11, and Gβγ subunits (Lodowski et al., 2003).

To our knowledge, our data provide the first report of phosphorylation-independent regulation of G_q/11 signaling in neurons by GRKs. Although GRK2 is thus able to suppress PLC-coupled receptor signaling independently of receptor phosphorylation, we cannot discount the fact that endogenous GRK2 may also phosphorylate the M₁ mACh receptor. Suppression of endogenous GRK2 through antisense or RNA interference techniques should allow additional assessment of the role of GRK2 in M₁ mACh receptor regulation in hippocampal neurons. However, the dramatic and rapid suppression of signaling by GRK2 independently of its kinase activity remains a powerful and potential regulator of GPCR action in vivo. Such a phenomenon may extend beyond the GRK2/3 subfamily, because a recent paper by Perroy et al. (2003) revealed a similar phosphorylation-independent regulation of G_i/o-coupled GABA_B receptor signaling mediated by GRK4 in cerebellar granule neurons.

In addition to a GRK2-mediated phosphorylation-independent suppression of M₁ mACh receptor signaling, we show here that manipulations of GRK6 activity can also affect mACh receptor responsiveness. Indeed, inhibition of endogenous GRK5 or GRK6 by dominant-negative constructs provided evidence only for GRK6 playing a role in M₁ mACh receptor signaling in the hippocampus. Considering that neither GRK5 nor GRK6 possess a C-site RGS domain, and are thus unable to bind Gα_q, we reason that endogenous GRK6 is likely to regulate M₁ mACh receptor signaling through a phosphorylation-dependent mechanism. This conclusion is consistent with our previous findings on the regulation of the endogenously expressed M₃ mACh receptor in SH-SY5Y neuroblastoma cells by GRK6-dependent receptor phosphorylation (Willets et al., 2001, 2002, 2003a). Unfortunately, because of the heterogeneous nature of hippocampal cultures and difficulties in culturing sufficient neurons, direct examination of GRK6- and GRK2-mediated phosphorylation of the M₁ mACh receptor will have to be performed in cell lines endogenously or recombinantly expressing the M₁ mACh receptor. The relatively small changes induced by GRK6 and ^K215RGRK6 suggest the potential involvement of other kinases in the regulation of M₁ mACh receptor signaling; however, pharmacological inhibition or downregulation of PKC activities fail to affect agonist-induced M₁ mACh receptor desensitization (data not shown).

In conclusion, this study has revealed regulation of a GPCR in hippocampal neurons using the eGFP-PH_PLCδ biosensor to follow “real time” changes in IP₃ in individual neurons. Furthermore, we showed the potential for phosphorylation-dependent and -independent regulation of M₁ mACh receptor signaling in hippocampal neurons through selective actions of GRK6 and GRK2, respectively. Whether the actions of these GRKs play a role in the acute regulation of M₁ mACh receptor function in the hippocampus or whether they also initiate longer-term changes in synaptic activity, perhaps associated with cholinergic activation of mitogen-activated kinases, remain to be established.