Abstract
The metabotropic glutamate (mGlu) receptors mGlu1 and mGlu5 mediate distinct inositol 1,4,5-trisphosphate (IP3) and Ca2+ signaling patterns, governed in part by differential mechanisms of feedback regulation after activation. Single cell imaging has shown that mGlu1 receptors initiate sustained elevations in IP3 and Ca2+, which are sensitive to agonist concentration. In contrast, mGlu5 receptors are subject to cyclical PKC-dependent uncoupling and consequently mediate coincident IP3 and Ca2+ oscillations that are largely independent of agonist concentration. In this study, we investigated the contribution of Gq/11α protein expression levels in shaping mGlu1/5 receptor-mediated IP3 and Ca2+ signals, using RNA interference (RNAi). RNAi-mediated knockdown of Gq/11α almost abolished the single-cell increase in IP3 caused by mGlu1 and mGlu5 receptor activation. For the mGlu1 receptor, this unmasked baseline Ca2+ oscillations that persisted even at maximal agonist concentrations. mGlu5 receptor-activated Ca2+ oscillations were still observed but were only initiated at high agonist concentrations. Recombinant overexpression of Gqα enhanced IP3 signals after mGlu1 and mGlu5 receptor activation. It is noteworthy that although mGlu5 receptor-mediated IP3 and Ca2+ oscillations in control cells were largely insensitive to agonist concentration, increasing Gqα expression converted these oscillatory signatures to sustained plateau responses in a high proportion of cells. In addition to modulating temporal Ca2+ signals, up- or down-regulation of Gq/11α expression alters the threshold for the concentration of glutamate at which a measurable Ca2+ signal could be detected. These experiments indicate that altering Gq/11α expression levels differentially affects spatiotemporal aspects of IP3 and Ca2+ signaling mediated by the mGlu1 and mGlu5 receptors.
Activation of the phospholipase C (PLC) pathway via coupling of G protein-coupled receptors (GPCRs) to G proteins of the Gq/11 family results in inositol 1,4,5-trisphosphate (IP3) production and mobilization of intracellular calcium (Ca2+i). Receptor activation can initiate spatially and temporally unique Ca2+ signals and thereby regulate an array of cellular processes (Berridge et al., 2000). In this study, we have investigated the contribution of Gq/11α protein expression in shaping receptor-initiated IP3 and Ca2+ signaling patterns.
The group I metabotropic glutamate receptors mGlu1a and mGlu5a are family C GPCRs. mGlu1a and mGlu5a receptors couple preferentially to the major isoforms of the Gq family, Gqα and G11α, and are capable of activating distinct patterns of Ca2+ mobilization (Hermans and Challiss, 2001). Stimulation of the mGlu1a receptor predominantly initiates a peak-and-plateau Ca2+ response, whereas mGlu5a receptor activation mediates a Ca2+ oscillatory response over a broad range of stimulus strengths (Nash et al., 2002). Previous studies proposed that mGlu5 receptor-mediated Ca2+ oscillations occur because of feedback inhibition via PKC-dependent phosphorylation of a single threonine residue (Thr840; Kawabata et al., 1996); however, very recent studies have shown that Thr840 is not phosphorylated but instead plays a structural/permissive role for the phosphorylation of an adjacent residue (Ser839) by PKC (Kim et al., 2005). Similar PKC-dependent Ca2+ oscillations have also been described after glutamate activation of astrocytes (Codazzi et al., 2001) and activation of another family C GPCR, the Ca2+-sensing receptor (Young et al., 2002). Use of the pleckstrin homology domain of phospholipase Cδ tagged with enhanced green fluorescent protein (eGFP-PHPLCδ) has enabled IP3 oscillations underlying mGlu5a receptor-activated Ca2+ oscillations to be observed (Nash et al., 2001, 2002; Nahorski et al., 2003). These PKC-dependent Ca2+ oscillations (referred to as “dynamic uncoupling”) are distinct from regenerative Ca2+-induced Ca2+-release (CICR), which is generated through an intrinsic property of the IP3 receptor (Thomas et al., 1996; Taylor and Thorn, 2001). CICR oscillations can be maintained with a relatively low steady-state increase in IP3, as observed after activation of the M3 muscarinic acetylcholine (mACh) receptor with a low agonist concentration in the same cell background (Nash et al., 2001). Our previous studies exploring the determinants of mGlu5 receptor signaling led us to propose a model in which Ca2+ oscillation frequency is dependent on receptor expression levels but is largely independent of agonist concentration (Nash et al., 2002). However, the importance of receptor-G protein coupling efficiency in group I mGlu receptor-mediated Ca2+ signaling has not yet been investigated. It is clear that regulation and localization of Gq/11α proteins could be a key, influencing factor in shaping the Ca2+ signals produced.
Studies examining the role of Gq/11α proteins generally, and in conjunction with mGlu receptor signaling, have been facilitated greatly by the generation of Gq/11α knockout mice (for review, see Offermanns, 2003). However, gene deletion studies are limited by the mortality of Gqα/G11α double-knockout mice and also by the possibility that the phenotype of the cells studied may adapt to compensate for the loss of a particular Gα, as observed for the deletion of other G protein subtypes, including Goα (Greif et al., 2000) and G15α (Davignon et al., 2000). Other investigators have successfully used antisense methods to reduce Gq/11α expression; however, these studies often rely on microinjection, making an accurate determination of endogenous Gqα and G11α protein expression difficult to ascertain (Macrez-Lepretre et al., 1997; Haley et al., 1998). Determining the relative Gqα and G11α expression levels after knockdown is clearly desirable in this type of study. Distinct roles of Gqα and G11α in mGlu1a receptor-mediated Ca2+ signaling in Purkinje neurons have recently been shown to result from differential expression levels of these two isoforms (Hartmann et al., 2004). Therefore, it was shown that Gqα was solely required for mGlu receptor-dependent synaptic transmission, whereas both Gqα and G11α contributed to long-term depression in Purkinje neurons.
In the current study, we have used Gq/11α-RNAi and Gqα overexpression in combination with single-cell IP3 and Ca2+ imaging as a novel approach to investigate the role of Gq/11α expression in GPCR-mediated signaling. RNAi-knockdown of Gq/11α protein expression was initially characterized in HEK cells stably expressing recombinant M3 mACh receptor. We then demonstrated the effects of RNAi and recombinant Gqα expression on IP3 and Ca2+ signals generated by mGlu1 and mGlu5 receptors expressed recombinantly in CHO cells. By altering Gq/11α expression levels, the agonist concentration-dependencies of these GqPCRs were changed. Furthermore, the temporal profiles of Ca2+ signals generated indicate a central role for Gq/11α in defining the nature of the response observed.
Materials and Methods
Cell Culture and Plasmid Transfection. CHO cells stably expressing the human mGlu1a or mGlu5a receptor under the control of the inducible LacSwitch-II system (Stratagene, La Jolla, CA) were maintained as described previously (Hermans et al., 1998; Nash et al., 2002) and are denoted as CHO-lac-mGlu1 or CHO-lac-mGlu5. HEK cells stably expressing the M3 mACh receptor (HEK-m3) were created and maintained as described previously (Tovey and Willars, 2004). Plasmid containing the fusion construct between eGFP and the pleckstrin homology domain of PLCδ1 (eGFP-PHPLCδ) was kindly donated by T. Meyer (Stanford University, Stanford, CA).
For single cell imaging experiments, CHO-lac-mGlu1/5 or HEK-m3 cells were grown on 25-mm coverslips and cotransfected 72 h before experimentation with 1.8 μg of Gq/11α-RNAi, control RNAi, or full-length human Gqα and 0.2 μg of eGFP-PHPLCδ (for IP3 imaging) or eGFP (for Ca2+ imaging) using 6 μl of GeneJuice (Novagen/EMD Biosciences, Madison, WI) per coverslip. For induction of maximal mGlu receptor expression in CHO cells, the medium was replaced with fresh culture medium containing 100 μM IPTG 18 to 20 h before experimentation. For standard SDS-PAGE immunoblotting, HEK-m3 cells were transfected in six-well plates 72 h before experimentation with 2 μg of Gq/11α-RNAi or control RNAi using Lipofectamine 2000. CHO cells were transfected in flasks (175 cm2) with 10 μg of Gq/11α-RNAi or control RNAi using 30 μl of GeneJuice and after 24 h, cells were seeded into six-well plates for a further 48 h. For receptor biotinylation, real-time PCR and 6 M urea SDS-PAGE analysis, CHO cells were transfected using the Nucleofection system (Amaxa Biosystems, Gaithersburg, MD), according to the manufacturer's optimized protocol. In brief, 5 × 106 cells were transfected with 2 μg of control- or Gq/11α-RNAi and Program U-23 on the Nucleofector, before seeding cells into six-well plates 72 h before experimentation.
RNAi Design and Preparation. To design an RNAi plasmid expressing Gq/11α-specific small interfering RNA (siRNA), the mRNA sequences for human Gqα (GenBank accession number NM_002072) and G11α (GenBank accession number NM_002067) were aligned to identify potential target sequences. Candidate 19-base pair sequences, homologous for both human Gqα and G11α genes (and containing a G/C content of 40–60%) were identified, and gene specificity was checked using the BLASTN algorithm to search the GenBank sequence database. RNAi-expressing constructs for five selected sequences were created according to the manufacturer's instructions [using the pSilencer 1.0-U6 expression system; Ambion (Austin, TX)], and the knockdown of recombinantly expressed CFP-labeled Gqα was used to assess the effectiveness of these constructs (data not shown). The target sequence selected for Gq/11α silencing was 5′-GATGTTCGTGGACCTGAAC-3′, corresponding to positions 932 to 950 relative to the start codon of human Gq and G11 (denoted “Gq/11-RNAi”). Furthermore, an additional control RNAi construct was generated (using the nucleotide sequence 5′-GCTGACCCTGGAGAGTTCATC-3′), and is denoted “control RNAi”.
Immunoblot Analysis. Levels of endogenous Gq/11α protein expression in HEK-m3 and CHO-lac-mGlu1a cells were determined by a standard Western blot protocol (Willets and Kelly, 2001) using a Gq/11α-specific antibody at 1:5000 dilution (CQ, kindly donated by G. Milligan, University of Glasgow, Glasgow, UK). Antibodies against Gi1–3α (1:2000; Santa Cruz Biotechnology, Santa Cruz, CA), G12α (1:1000; Santa Cruz Biotechnology) and γ-tubulin (1:10,000; Sigma, St. Louis, MO) were used as controls for RNAi specificity and protein loading. Antibodies against the C-terminal region of the mGlu1 receptor (1:1000; Chemicon International, Temecula, CA) and the C-terminal region of the mGlu5 receptor (1:1000; Upstate Biotechnology, Lake Placid, NY), respectively, were used to detect mGlu1a and mGlu5a receptor expression. All primary antibody incubations were made at room temperature for 2 h or overnight at 4°C. Resolution of both Gqα and G11α with an antibody (CQ) was achieved using SDS-PAGE gels containing 12.5% acrylamide and 6 M urea, as described previously (Milligan, 1993). The relative mobility of each Gα subunit was confirmed using cell lysates obtained from CHO-lac-mGlu cells recombinantly expressing human Gqα or G11α using equivalent gels.
Cell-Surface Biotinylation of mGlu1a and mGlu5a Receptors. CHO-lac-mGlu1a and -mGlu5a cells were transiently transfected with Gq/11α-RNAi, control RNAi, or full-length human Gqα using the Amaxa Nucleofection system, seeded into six-well plates, and induced 48 h later. After 72 h, cells were washed twice with PBS at room temperature and labeled with membrane-impermeant EZ-Link Sulfo-NHS-biotin (1 mM in PBS; Pierce, Rockford, IL) for 30 min at room temperature. Cells were washed twice with ice-cold PBS, once with 500 mM Tris/HCl, pH 7.4, and twice more with ice-cold PBS. Cells were then lysed for 10 min with solubilization buffer (10 mM Tris/HCl, 150 mM NaCl, 1 mM EDTA, 500 μM EGTA, 1% Igepal, and 0.1% SDS, pH 7.4) and centrifuged at 14,000g for 5 min. The cleared supernatant (900 μl) was then incubated with 200 μl of streptavidin/agarose beads (diluted 1:10 in solubilization buffer) for 2 h with constant rotation at 4°C. The remaining supernatant was retained to assess receptor expression in whole-cell lysates (see below). Beads were recovered by centrifugation during two washes with 1 ml of solubilization buffer and two washes with 1 ml of PBS. Immunocomplexes were dissociated with 50 μl of 2× sample buffer (125 mM Tris/HCl, 50 mM dithiothreitol, 4% SDS, 20% glycerol, 0.01% bromphenol blue, pH 6.8), heated at 90°C for 5 min, and then resolved by SDS-PAGE (as described above). For total cell extracts, 20 μl of the retained supernatant was diluted (1:1) with 2× sample buffer before SDS-PAGE.
PCR Amplification of the RNAi Target Region from CHO Cells. To determine the sequence of the RNAi target region of Gqα and G11α derived from CHO cells, total RNA was isolated from CHO cells using the RNAeasy kit (QIAGEN, Valencia, CA). Samples (200 ng of RNA) were reverse-transcribed into cDNA using the Omniscript RT-PCR kit (QIAGEN). Primers specific for Gqα and G11α based on rat sequence information (see Table 1) were used for PCR to amplify a 160-base pair fragment containing the RNAi target region using KOD HiFi DNA Polymerase (Novagen) as described in the manufacturer's instructions. PCR amplification was carried out over 30 cycles consisting of 98°C for 20 s, 50°C or 52°C (for Gqα and G11α primers, respectively) for 20 s, and 72°C for 20 s. PCR reaction products were subsequently purified using a QIAquick PCR purification kit (QIAGEN); after gel electrophoresis, the PCR product was extracted and purified using a QIAEX II Gel extraction kit (QIAGEN). PCR products were sequenced directly using the same primers.
Quantitative Real-Time PCR. After RNAi transfection, total RNA and cDNA were prepared from samples in triplicate, as described above. An RT-negative control was included for each triplicate to control for genomic DNA contamination. Real-time PCR using SYBR-green fluorescence (Applied Biosystems, Foster City, CA) was carried out using an ABI PRISM 7700 sequence detection system as described previously (Medhurst et al., 2000). PCR parameters were 50°C for 2 min, 95°C for 10 min, 40 cycles of 95°C for 15 s, and 60°C for 1 min. Real-time PCR data were captured using Sequence Detector software (Applied Biosystems) to obtain threshold cycle values for the gene of interest. Values were normalized against a housekeeping gene, cyclophilin, and expressed as percentage of control. All measurements were performed in triplicate for three separate transfections using the Amaxa Nucleofection system.
Single Cell Imaging of IP3. Increases in cellular IP3 were detected by measuring the translocation of eGFP-PHPLCδ from the plasma membrane to the cytosol as described previously (Nash et al., 2002; Young et al., 2003). Cells were transfected as described above, and coverslips were mounted on the stage of an Olympus IX70 inverted epifluorescence microscope and perfused (5 ml/min) at 37°C with Krebs-Henseleit buffer (KHB): 118 mM NaCl, 4.7 mM KCl, 1.2 mM MgSO4, 1.3 mM CaCl2, 1.2 mM KH2PO4, 4.2 mM NaHCO3, 10 mM HEPES, and 11.7 mM glucose, pH 7.4) using a Gilson Minipuls 2 pump (Gilson, Inc., Middleton, WI). Confocal images were collected after excitation at 488 nm using an Olympus FV500 laser scanning confocal microscope at a scan rate of 1.5 to 2.5 Hz. eGFP-PHPLCδ translocation was measured by creating a region of interest in the cytosol and plotting the average pixel intensity in that region versus time. Data are expressed in relative fluorescent units (RFU) by subtraction of background fluorescence followed by dividing the fluorescent intensity at a given time by the initial fluorescence within each region of interest (F/F0). Drug applications were made using the perfusion line as indicated.
Ca2+ Imaging. Cells were transfected as described, and then loaded with fura-2 AM (Invitrogen) in 1 ml of KHB (final concentration, 5 μM) for 60 to 90 min. The cells were mounted on an inverted epifluorescence microscope (Diaphot; Nikon, Tokyo, Japan) with an oil immersion objective (40×) and then excited at 340 and 380 nm (for fura-2 AM) and 488 nm (for eGFP) using a Spectramaster II monochromator (PerkinElmer Life and Analytical Sciences, Boston, MA) at a sample rate of 0.7 Hz. GFP-containing cells were identified (as a marker of Gq/11α-RNAi/control RNAi/Gqα-transfected cells), and sequential images were then captured from GFP-transfected cells at wavelengths above 510 nm after drug applications made using the perfusion line as indicated. Ca2+ signals are expressed as 340 nm/380 nm ratios.
Data Analysis. Curve fitting of data and calculation of EC50 values was carried out using Prism 3.0 (GraphPad Software, Inc., San Diego, CA). Statistical differences between data sets were determined by one-way analysis of variance for multiple comparisons, followed by Bonferroni's multiple-range test at p < 0.05 (using Prism 3.0 software), or Student's t test (unpaired; p < 0.05 being considered significant).
Results
RNAi-Mediated Knockdown of Gq/11α Protein Expression and Function in HEK Cells Stably Expressing the M3 mACh Receptor. RNAi constructs designed using 19-base pair sequences targeting homologous regions of human Gqα and G11α protein expression were initially tested in HEK-m3 cells. Immunoblotting experiments revealed robust silencing of Gq/11α 72 h after transfection with Gq/11α-RNAi compared with control RNAi. In these experiments, the transfection efficiency typically achieved was ∼50%, suggesting a high degree of Gq/11α silencing in individual HEK cells. Antibodies detecting γ-tubulin and G protein α-subunits, Gi1–3α, confirmed equal protein loading (Fig. 1A).
Single-cell IP3 imaging, using cells transfected with eGFP-PHPLCδ alone, showed a concentration-dependent increase in cytosolic fluorescence in response to 30-s applications of methacholine, with a best-fit EC50 of 0.9 μM. This was unaffected by cotransfection with control RNAi but was inhibited by Gq/11α-RNAi, as demonstrated by a rightward shift in the concentration-response curve (EC50 6 μM) and a suppressed maximal response (2-fold, compared with 3.5-fold over basal in control cells, shown in Fig. 1B). Representative images (Fig. 1B) showed a clear suppression of eGFP-PHPLCδ translocation from the plasma membrane to the cytosol in Gq/11α-RNAi-containing cells compared with control RNAi and also indicate that these cells display no obvious morphological changes.
RNAi-Mediated Knockdown of Gq/11α Protein Expression in CHO Cells. Having identified a functional Gq/11α-RNAi construct in HEK cells, we tested the same construct in CHO cell lines stably and inducibly expressing mGlu1 and mGlu5 receptors. Immunoblotting showed a significant inhibition of endogenous Gq/11α expression 72 h after transfection with Gq/11α-RNAi compared with control RNAi cells (Fig. 2A). The transfection efficiency typically achieved in CHO cells was 40 to 50% (determined using green fluorescent protein; data not shown), again suggesting highly efficient Gq/11α protein silencing. It is noteworthy that Gi1–3α, G12α, and γ-tubulin protein expression levels were unaffected (Fig. 2A). Further experiments confirmed that induction of mGlu1 and mGlu5 receptor expression was unaffected by RNAi treatment (Fig. 2B).
After the identification of an RNAi construct with Gq/11α silencing activity in both human and CHO cell backgrounds, the 19-base pair RNAi target sequence within Gqα and G11α derived from the CHO cell line was investigated. No sequence information for Chinese hamster Gq/11α was available in the GenBank/EMBL databases. We therefore amplified the appropriate regions from CHO cells by PCR with Gqα- and G11α-selective primers, based on rat sequence information (Table 1). Sequencing of the amplicons from CHO cells indicated two base pair differences (corresponding to C→T changes at positions 9 and 21 in Fig. 3A) in G11α compared with CHO Gqα and human Gq/11α.
In view of the contrasting reports of RNAi specificity in the literature, it was imperative to assess the knockdown of Gqα relative to G11α. Using the Nucleofection system to achieve optimal plasmid transfection (>80% determined using GFP; data not shown) and real-time PCR with Gqα- and G11α-specific primers (Table 1), a robust knockdown (∼75%) of Gqα mRNA in both mGlu1 and mGlu5 receptor-expressing CHO-lac cells was observed (Fig. 3B). We were surprised to observe G11α knockdown in the same experiments. Gqα and G11α knockdown was normalized to cyclophilin mRNA levels, which were unaffected after Gq/11α-RNAi transfection. To complement the mRNA analysis, SDS-PAGE gels containing 6 M urea were used to resolve Gqα and G11α proteins in these cells (see Materials and Methods). Under these conditions, G11α migrates farther than Gqα, an observation confirmed using lysates obtained from CHO-lac-mGlu cells recombinantly expressing human Gqα and G11α in equivalent gels (Fig. 3C). If the antibody immunoreactivity for both Gα subunits is comparable, these data suggest that expression levels of Gqα and G11α proteins under control conditions in CHO-lac-mGlu1a cells are similar. The relative reduction of Gqα and G11α protein expression was similar to the reduction in Gqα and G11α mRNA levels determined by real-time PCR. A reduction of Gqα mRNA by ∼75% corresponded to an almost complete knockout of Gqα protein; similarly, a reduction in G11α mRNA by ∼60% led to a slightly less efficient knockdown of G11α protein expression. The relationship between mRNA and protein knockdown is dependent on the turnover of Gqα and G11α proteins. A previous study in CHO cells reported that the half-time of Gq/11α is 18 h (Mitchell et al., 1993). Levels of Gi1–3α, G12α, and γ-tubulin protein expression in these experiments were unaffected (data not shown). These experiments demonstrate the effectiveness of RNAi in silencing both Gqα and G11α and so provide evidence that mismatches in the RNAi target sequence are tolerated, at least for these closely related gene products.
Effects of Manipulating Gq/11α Expression on Cell-Surface Receptor Expression. To examine the possibility that up- or down-regulation of Gq/11α protein expression can influence mGlu1/5 receptor trafficking to the plasma membrane, we used a cell-surface protein biotinylation strategy. In both CHO-lac-mGlu1a (Fig. 4A) and mGlu5a (Fig. 4B) cells, Gq/11α protein expression was increased or decreased by transient transfection with human recombinant Gqα or Gq/11α-RNAi, respectively. For mGlu1a and mGlu5a receptors, doublet immunoreactive bands were evident at ∼150 kDa. mGlu1a and mGlu5a receptor cell-surface biotinylation indicated that the mature, higher molecular weight protein (top band) is enriched at the plasma membrane (Fig. 4, A and B). It is noteworthy that altering Gq/11α expression levels did not affect cell-surface mGlu1a and mGlu5a receptor, or indeed total receptor expression in whole-cell extracts, indicating that the trafficking/localization of these receptors to the plasma membrane is independent of Gq/11α protein expression.
The Effects of Manipulating Gq/11α Protein Expression on Single Cell IP3 Levels. Measurements of IP3 production in cells transfected with Gq/11α-RNAi, control RNAi, or full-length human Gqα were made possible by cotransfection of the IP3 biosensor eGFP-PHPLCδ. In all of the control RNAi CHO-lac-mGlu1a cells examined, a graded increase in eGFP-PHPLCδ translocation was observed in response to incremental glutamate concentrations, each applied for 30 s (Fig. 5), with a best-fit EC50 of 8.4 μM for the peak IP3 response. In CHO-lac-mGlu1a cells transfected with Gq/11α-RNAi, responses to glutamate were almost abolished, with no translocation of eGFP-PHPLCδ observed in several individual cells. In contrast, when cells were transfected with full-length human Gqα, the EC50 for the peak IP3 response left-shifted from 8.4 to 2.0 μM glutamate (p < 0.05), whereas the maximal responsiveness was not significantly affected (p > 0.05; Fig. 5).
Single-cell mGlu5a receptor-induced IP3 responses underlying oscillatory Ca2+ signals have been studied extensively in CHO-lac-mGlu5a cells (Nash et al., 2002; Nahorski et al., 2003; Young et al., 2003). To examine the effect of manipulating Gq/11α-protein levels, cells were transfected with Gq/11α-RNAi, control RNAi, or full-length human Gqα. Responses to a single maximal concentration (30 μM) of glutamate were subsequently examined over a 300-s time period for each treatment. Within a population of control cells, mGlu5a receptor-induced IP3 increases in response to glutamate eGFP-PHPLCδ translocation were not detected in 38% of cells (denoted “non-responders” in Fig. 6, A and B). In control RNAi cells in which eGFP-PHPLCδ translocation was measurable, the predominant response was a peak elevation of IP3 followed by repetitive oscillations until agonist washout (denoted as oscillatory in Fig. 6, A and B). In several cells (24% of the total number of cells), no detectable change in IP3 was detected after the initial peak increase (denoted as a single spike in Fig. 6A). In CHO-lac-mGlu5a cells transfected with Gq/11α-RNAi, repetitive oscillations in IP3 were still observed. However, the number of nonresponders and single spike responses increased to 86% (from 62%) of the total number of cells. In contrast, after full-length Gqα transfection, the number of nonresponding cells decreased to 25% of the total population, and only a relatively low number of cells (8%) exhibited IP3 oscillations. Instead, in 50% of Gqα-transfected cells, IP3 signals were manifested as an initial peak followed by a sustained plateau response until agonist washout (denoted as saturating in Fig. 6, A and B). In addition, the mean increase in peak IP3 production was at least 2-fold higher in Gqα-transfected cells (0.37 ± 0.07 RFU; p < 0.05), compared with control (0.17 ± 0.03 RFU) and RNAi-transfected cells (0.13 ± 0.05 RFU) (Fig. 6C).
Gq/11α Protein Expression Levels Influence Stimulus Strength and the Temporal Profile of mGlu Receptor-Mediated Ca2+ Mobilization. Previous studies from our laboratory have made real-time concurrent measurements of mGlu1a receptor-generated IP3 and Ca2+i in single cells (Nash et al., 2001). During prolonged activation, these receptors give an initial transient peak in IP3 production followed by a sustained plateau phase that is closely mirrored by a peak and sustained plateau Ca2+ response. In contrast, the mGlu5a receptor produces oscillatory IP3 and Ca2+ signals (Nash et al., 2001, 2002; Nahorski et al., 2003). In the current study, the effects of manipulating Gq/11α protein levels on the glutamate concentration-dependence and temporal profile of mGlu1a and mGlu5a receptor-mediated Ca2+ signaling were investigated. Receptors were therefore exposed to incremental increases in glutamate (1 to 100 μM, 200 s at each concentration) to determine the role of Gq/11α expression levels on complex calcium signals.
In control RNAi and Gqα-expressing CHO-lac-mGlu1a cells, ∼60% of transfected cells responded to glutamate with an initial peak elevation in Ca2+ followed by a sustained plateau phase. After Gq/11α-RNAi transfection, however, 84% (66 of 79) of cells elicited transient peak increases in Ca2+ mobilization without a sustained phase even at maximal concentrations of glutamate (Fig. 7C, denoted as “non-saturating”). These nonsaturating responses most often occurred as a single baseline spike coincident with agonist addition (as illustrated with the representative trace in Fig. 7A). On occasion, however, additional intermittent spikes were evident, perhaps signifying oscillations occurring as a result of regenerative CICR. The few cells that actually initiated a sustained peak and plateau did so only at high concentrations of agonist (7 of 66 cells over four experiments). Changing levels of Gq/11α-protein expression altered the glutamate concentration threshold for the onset of mGlu1a receptor-induced Ca2+ responses (Fig. 8A). Thus, in control RNAi cells, the initial onset of Ca2+ mobilization occurred at either 3 or 10 μM glutamate (35 and 50% of the total number of cells, respectively), which correlates with the IP3 measurements made in these cells (Fig. 5). In Gq/11α-RNAi transfected cells, a greater agonist concentration was required to produce significant increases in IP3 (around 10 μM glutamate), reflecting a shift to higher agonist concentrations for the onset of Ca2+ mobilization (only 14% responded to 3 μM but 53% responded to 10 μM). Furthermore, 25% of Gq/11α-RNAi-transfected cells compared with 8% of control cells responded only to concentrations greater than 10 μM glutamate. The shift in stimulus threshold was further exemplified in experiments using Gqα overexpression, where 3 μM glutamate mediated a large increase in IP3 (Fig. 5), so that 74% (25 of 34) of cells at this concentration of glutamate were able to initiate a Ca2+ signal compared with 35% of control cells (Figs. 7 and 8A). In the mGlu1a model system, Gq/11α expression levels clearly regulate the efficacy and the nature of agonist-mediated receptor activation. This is manifested as altered levels of IP3 production and hence differential Ca2+ signals in response to increasing agonist concentrations. In contrast, mGlu5a receptor signaling is relatively independent of stimulus-strength over a wide agonist concentration range (Nash et al., 2002), and it was not clear what effect alterations in Gq/11α protein levels would have.
In control RNAi mGlu5a receptor-expressing cells, increasing concentrations of glutamate produced nonsaturating Ca2+ oscillations (Fig. 7, B and D). Although in many cells the frequency of Ca2+ oscillations was initially sensitive to the increase in agonist concentration, in all cases, the Ca2+ frequency quickly became insensitive to agonist concentration, which is indicative of oscillations driven by PKC-dependent feedback on the receptor (see Nash et al., 2002; Fig. 7B, representative trace). Similar observations were made from mGlu5a receptor-expressing cells transfected with Gq/11α-RNAi, but in the majority of cells, the onset of response occurred at higher concentrations of agonist (Fig. 8B; 10 μM versus 3 μM glutamate in control cells); consequently, nonsaturating Ca2+ oscillations in most Gq/11α-RNAi-transfected cells could not be identified until concentrations of ≥10 μM were used (Fig. 7, representative trace). In CHO-lac-mGlu5a cells transfected with Gqα 75% (52 of 69 cells) cells responded at 1 μM glutamate. A key effect of Gqα overexpression was the observation of mGlu5 cells exhibiting saturating Ca2+ responses to increasing concentrations of glutamate (Fig. 7, B and D). Thus, although oscillatory responses were observed at low glutamate concentrations, these rapidly converted to a sustained plateau response upon increases in agonist concentrations. In this way, the Ca2+ response mirrored the IP3 signals produced by the mGlu5a receptor in recombinant Gqα-expressing cells (Fig. 6B).
Discussion
Complex Gq/11-coupled receptor-mediated Ca2+ signals enable a wide range of cellular processes to be regulated (Berridge et al., 2000). Low levels of receptor occupancy are known to induce small steady-state increases in IP3 that evoke oscillatory Ca2+ signals, the frequency of which is sensitive to agonist concentration. Higher levels of IP3 generation caused by more intense receptor stimulation can lead to larger, peak-and-plateau–type Ca2+ responses. Studies using the IP3 biosensor eGFP-PHPLCδ have since shown that Ca2+ oscillations can also occur in synchrony with IP3 oscillations, as described in ATP-stimulated canine kidney epithelial cells (Hirose et al., 1999) or stimulation of mGlu5 receptors in CHO cells (Nash et al., 2001). Oscillations in IP3 are likely to occur as a result of negative feedback regulation by PKC (Codazzi et al., 2001; Nash et al., 2002; Young et al., 2002) or RGS proteins (Luo et al., 2001) but also by a positive feedback effect of Ca2+ to enhance PLC activity (Young et al., 2003). For the mGlu5a receptor, synchronous IP3 and Ca2+ oscillations occur via a PKC-dependent dynamic uncoupling of the receptor from its G protein by phosphorylation of a specific residue on the receptor (Kawabata et al., 1996; Kim et al., 2005). Because the substrate for PKC is the receptor, this provides a mechanism by which the frequency of Ca2+ oscillations are regulated by receptor density in addition to agonist concentration, as shown for the mGlu5 receptor (Nash et al., 2002). The mGlu1 receptor lacks this critical consensus sequence and is not subject to this feedback regulation. Therefore, mGlu1 receptor activation can initiate sustained peak-and-plateau Ca2+ responses. These studies suggest that Ca2+ signals initiated by Gq/11-coupled GPCRs are susceptible to regulation by a number of factors, including receptor-G protein coupling efficiency, receptor density, agonist concentration and sensitivity to feedback mechanisms.
The fine-tuning of Ca2+ signals by these factors may account for contrasting patterns of mGlu5 receptor-driven Ca2+ oscillations in different cell backgrounds (Codazzi et al., 2001; Dale et al., 2001; Nash et al., 2002). In the current study, we have extended these investigations by assessing the contribution of Gq/11α protein expression levels in generating distinct IP3 and Ca2+ signaling patterns mediated by group I mGlu receptors. Using RNAi or recombinant Gqα expression in combination with an IP3 biosensor or Ca2+ sensitive dye, we have been able to study the concentration-dependence and the temporal profile of IP3 and Ca2+i in single cells in real-time.
The use of RNAi as an approach to gene silencing is now extensive, although its mechanism is not yet fully understood (Dykxhoorn et al., 2003; Hannon and Rossi, 2004; Meister and Tuschl, 2004). The enormous interest in this technique has also highlighted potential limitations, in particular with respect to RNAi specificity (Hannon and Rossi, 2004; Snove and Holen, 2004). Here, we have shown a robust RNAi-induced silencing of Gq/11α protein expression in HEK-m3 cells that resulted in a suppression of the methacholine-induced IP3 signal. Very recently, effective silencing of Gq/11α proteins in HEK cells has been reported in two other studies, using an RNAi sequence identical to the one used here. Using an siRNA approach, the authors established a role for Gq/11α proteins in stress fiber formation (Barnes et al., 2005), but not chemotaxis (Hunton et al., 2005), after angiotensin II AT1A receptor activation.
Effective Gq/11α silencing was also observed in CHO-lac cell lines stably expressing mGlu1 or mGlu5 receptors, with no effect on the expression levels of Gi1–3α and G12α proteins or the receptors themselves. Despite the presence of a 2-base-pair mismatch in the RNAi target region of G11α compared with Gqα in the Chinese hamster cell model, real-time PCR and a pan-Gq/11α antibody revealed knockdown of both Gqα and G11α. Several reports have described similar tolerance of mismatches between siRNA and target mRNA, especially when these occur at the periphery of the target sequence (Amarzguioui et al., 2003; Vickers et al., 2003; Snove and Holen, 2004). Other studies also indicate that siRNAs containing mismatches can act as endogenous micro-RNAs to inhibit translation, but not to cause significant mRNA degradation (Doench et al., 2003; Saxena et al., 2003). However, this was not the case here, in that we observed a clear reduction in mRNA and protein expression. The effect therefore seems to be specific to Gqα and G11α as a result of the close homology of the two sequences in the target region, instead of generic knockdown of unrelated proteins. In this study, we have therefore been able to use the RNAi construct to assess the effects of combined Gq/11α knockdown, but in doing so also provide evidence for a “cross-reactivity” of siRNA targeting closely related genes.
Knockdown of Gq/11α protein almost completely eliminated mGlu1a receptor-mediated IP3 responses in single cells at maximal concentrations of agonist, whereas increased Gqα expression produced a significant enhancement in agonist sensitivity. Despite the ability of mGlu1a receptors to couple to other G proteins, such as Giα, which can contribute to phosphoinositide turnover (Hermans and Challiss, 2001), our current data suggest that Gq/11α coupling alone accounts for the majority of PLC activation in this expression system. Reducing the expression level of Gq/11α clearly affected the signaling properties of mGlu1a receptor, so that the subsequent Ca2+ response no longer resembled the characteristic sustained peak-and-plateau signal typical of this receptor subtype (Kawabata et al., 1996; Nash et al., 2001). Instead, the observed baseline Ca2+ spiking was likely to be the result of regenerative CICR associated with small increments in IP3 (Bootman et al., 1996; Thomas et al., 1996; Nash et al., 2001; Young et al., 2003). This finding supports the notion that the temporal response initiated by the mGlu1a receptor is dependent on stimulus strength and is likely to be influenced by receptor and G protein expression levels.
Ca2+ signals activated by the mGlu5 receptor were also sensitive to changes in Gq/11α expression. Using recombinant Gqα expression, we have revealed an ability of mGlu5 receptor to overcome PKC-dependent feedback allowing the receptor to mediate plateau IP3 and Ca2+ responses. In cells transfected with recombinant Gqα, peak IP3 responses were clearly enhanced, and Ca2+ mobilization was achieved at lower threshold concentrations of agonist. This supports our previous proposal (Nash et al., 2002) and suggests that increased PLC activation reduces the period of PKC-dependent uncoupling and favors transition to sustained Ca2+ mobilization. With increased coupling and PLC activation, the mGlu5 receptor therefore mirrors M3 mACh receptor signaling in lacrimal acinar cells, where PKC feedback occurs over a defined agonist concentration range (Bird et al., 1993). It is likely that sustained activation occurs when IP3 receptors are saturated, leading to store depletion and capacitative Ca2+ entry. Indeed, previous studies have demonstrated mGlu5 receptor-activated peak-and-plateau responses in the CHO-lac-mGlu5a cell-model after thapsigargin-induced depletion of intracellular Ca2+ stores (Nash et al., 2002). Using RNAi, we have shown that a reduction in mGlu5 receptor stimulus-strength restricts IP3 production, even at maximal concentrations of agonist, and constrains Ca2+ responses to baseline oscillations that require higher concentrations of agonist for initiation. Although PKC dependence was not tested here, previous studies in the same cells have shown that PKC down-regulation can prevent IP3 oscillations, such that maximal concentrations of agonist initiate sustained peak-and-plateau Ca2+ responses in a similar manner to the mGlu1a receptor (Nash et al., 2002). It is noteworthy that at lower concentrations of agonist, PKC-independent Ca2+ oscillations were seen that were sensitive to agonist concentration and may reflect the Ca2+ oscillations observed in some cells after RNAi treatment in the present experiments.
In conclusion, we have shown that altering Gq/11α expression markedly alters the temporal profile and agonist concentration-dependencies of IP3 and Ca2+ signals generated after either mGlu1a or mGlu5a receptor activation. This work also demonstrates that regulating Gq/11α expression levels can fundamentally alter the signaling properties of the mGlu5 receptor subtype. For the mGlu5 receptor, it seems that the predominant dynamic uncoupling mechanism linking receptor activation to Ca2+ signaling instills particular emergent properties. Thus, whereas the receptor expression level primarily determines the Ca2+ oscillatory frequency, the transitions between CICR, dynamic uncoupling, and peak-and-plateau behaviors are modulated by the level of Gq/11 expression. It will be interesting to establish the effects of changing the levels of other intermediates in the signaling pathway (e.g., PLCβ expression) or the presence of Homer proteins (Fagni et al., 2000; Kiselyov et al., 2003) on mGlu1/5 receptor signaling. It is also interesting to speculate on whether translocation of mGlu5 receptors to the plasma membrane from an intracellular locus (Hubert et al., 2001) and/or direct or indirect (e.g., through changes in RGS protein expression) changes in G protein expression/function can fundamentally alter neuronal Ca2+ signaling by this receptor subtype under altered physiological or pathophysiological conditions.
Acknowledgments
We thank Drs. Melanie Robbins, Isabel Benzel, and Andrew Calver (GlaxoSmithKline, Harlow, UK) for expert technical guidance on RT-PCR and helpful discussion. We also thank Dr. Mark Nash (Novartis Institute of Medical Sciences, London, UK) for initial input and ongoing intellectual contribution to this research project.
Footnotes
- Received April 26, 2005.
- Accepted October 18, 2005.
This work was supported by the Wellcome Trust of Great Britain (grant 062495) and by a joint Biotechnology and Biological Sciences Research Council and GlaxoSmithKline Ph.D. studentship (to P.J.A.).
Article, publication date, and citation information can be found at http://molpharm.aspetjournals.org.
doi:10.1124/mol.105.014258.
ABBREVIATIONS: PLC, phospholipase C; GPCR, G protein-coupled receptor; IP3, inositol 1,4,5-trisphosphate; PKC, protein kinase C; eGFP, enhanced green fluorescent protein; eGFP-PHPLCδ, pleckstrin homology domain of PLCδ1 tagged with enhanced green fluorescent protein; CICR, Ca2+-induced Ca2+-release; mGlu, metabotropic glutamate; mACh, muscarinic acetylcholine; RNAi, RNA interference; CHO, Chinese hamster ovary; HEK, human embryonic kidney; siRNA, small interfering RNA; PAGE, polyacrylamide gel electrophoresis; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; RT, reverse transcription; KHB, Krebs-Henseleit buffer; RFU, relative fluorescent units; AM, acetoxymethyl ester; RGS, regulator of G protein signaling.
- The American Society for Pharmacology and Experimental Therapeutics