Activation of Endoplasmic Reticulum-Localized Metabotropic Glutamate Receptor 5 (mGlu₅) Triggers Calcium Release Distinct from Cell Surface Counterparts in Striatal Neurons

Abstract

Metabotropic glutamate receptor 5 (mGlu₅) plays a fundamental role in synaptic plasticity, potentially serving as a therapeutic target for various neurodevelopmental and psychiatric disorders. Previously, we have shown that mGlu₅ can also signal from intracellular membranes in the cortex, hippocampus, and striatum. Using cytoplasmic Ca²⁺ indicators, we showed that activated cell surface mGlu₅ induced a transient Ca²⁺ increase, whereas the activation of intracellular mGlu₅ mediated a sustained Ca²⁺ elevation in striatal neurons. Here, we used the newly designed ER-targeted Ca²⁺ sensor, ER-GCaMP6-150, as a robust, specific approach to directly monitor mGlu₅-mediated changes in ER Ca²⁺ itself. Using this sensor, we found that the activation of cell surface mGlu₅ led to small declines in ER Ca²⁺, whereas the activation of ER-localized mGlu₅ resulted in rapid, more pronounced changes. The latter could be blocked by the Gq inhibitor FR9000359, the PLC inhibitor U73122, as well as IP₃ and ryanodine receptor blockers. These data demonstrate that like cell surface and nuclear mGlu₅, ER-localized receptors play a pivotal role in generating and shaping intracellular Ca²⁺ signals.

1. Introduction

As a universal signaling molecule, calcium (Ca²⁺) is a critical intracellular second messenger [1,2,3]. Cytosolic Ca²⁺ signals are generated via Ca²⁺ release from intracellular stores like the endoplasmic reticulum (ER) or via entry from the extracellular space. In the former, activation of cell surface receptors such as G protein-coupled receptors (GPCRs) or receptor tyrosine kinases (RTKs) can lead to the activation of phospholipase C (PLC), the cleavage of phosphatidyl inositol 4, 5 bisphosphate (PIP₂) and the production of inositol 1,4,5 trisphosphate (IP₃) [4]. IP₃ binding to IP₃ receptors on the ER membrane allows Ca²⁺ to be released from luminal stores, affecting many cytoplasmic signaling systems. Ca²⁺ within the ER lumen also plays a signaling role [5,6,7], affecting the ability of the ER to communicate with other organelles as well as supporting critical ER functions such as protein folding and lipid biosynthesis. Not surprisingly, disruption of ER Ca²⁺ homeostasis can lead to ER stress, loss of mitochondrial function, apoptosis, and cell death [8]. Thus, transient increases of cytoplasmic Ca²⁺ as well as perturbation of luminal Ca²⁺ can both lead to a variety of cell biological phenomena critical to cellular homeostasis.

Metabotropic glutamate receptors (mGlu receptors) are a family of GPCRs that modulate neuronal excitability and synaptic transmission in the central nervous system [9,10]. In particular, the mGlu subtype, mGlu₅, plays a fundamental role in synaptic plasticity primarily by regulating intracellular Ca²⁺ levels via release from ER Ca²⁺ stores. Many studies have shown that activated plasma membrane mGlu₅ couples to G_q/11 and phosphatidylinositol (PI)-PLC to generate IP₃-mediated release of Ca²⁺ from both IP₃ and ryanodine intracellular receptors. Depending upon the cell type in which the receptor is expressed, activation of cell surface mGlu₅ can generate a rapid, transient Ca²⁺ signal or a slowly decaying oscillatory response [11,12,13].

Over the last two decades, work from this lab has also shown that mGlu₅ is one out of a growing number of GPCRs [14,15,16] that can signal from intracellular membranes [17,18,19,20,21,22]. Specifically, more than 90% of mGlu₅ is present on intracellular membranes such as the ER and the nuclear membrane, where it can be activated by ligands that can be transported across cell membranes [17]. Like mGlu₅ present on the cell surface, activation of endogenous mGlu₅ receptors expressed on the inner membrane of isolated striatal nuclei also generate IP₃, leading to the IP₃-mediated release of Ca²⁺ in the nucleus [17,23]. Interestingly, in the striatum, mGlu₅-activated nuclear response patterns yield long prolonged Ca²⁺ responses (>10 min; 17). These data showed that the nucleus could function as an autonomous organelle independent of signals originating in the cytoplasm, and that nuclear mGlu₅ receptors play a dynamic role in mobilizing Ca²⁺ in a specific, localized fashion.

To directly monitor mGlu₅-mediated changes in ER Ca²⁺ itself, we used the newly designed and optimized ER-targeted Ca²⁺ sensor, ER-GCaMP6-150 [24], as a robust, specific approach to bypass less interpretable changes in cytosolic Ca²⁺ versus changes in ER Ca²⁺ content. Using this tool, we found that activation of cell surface mGlu₅ using DHPG led to a modest decline in somal ER Ca²⁺, whereas activation of intracellular mGlu₅ using Quis resulted in more pronounced soma ER Ca²⁺ changes. In either case, the effects were almost two-fold larger when measured in neurites. ER-localized mGlu₅-mediated Ca²⁺ responses could also be blocked by the Gq inhibitor, FR9000359, and the PLC inhibitor, U73122, but not by U73343. Finally, both IP₃ and ryanodine receptor blockers prevented ER mGlu₅-mediated decreases in luminal Ca²⁺. Collectively, these studies show that like cell surface and nuclear receptors, activated ER mGlu₅ receptors couple to G_q/11 and PLC to generate the IP₃-mediated release of Ca²⁺ from Ca²⁺-release channels in the ER. Thus, the ER can also function as an autonomous organelle in mobilizing Ca²⁺ in a specific, localized fashion.

2. Materials and Methods

2.1. Materials

(+)-α-Amino-3,5-dioxo-1,2,4-oxadiazolidine-2-propanoic acid, quisqualate (Quis), (S)-3,5-dihydroxyphenylglycine (DHPG), 2-methyl-6-(phenylethynyl)pyridine (MPEP), 7-(hydroxyimino)-cyclopropan[b]chromen-1a-carboxylate ethyl ester (CPCCOEt), (±)-4-(4-aminophenyl)-1,2-dihydro-1-methyl-2-propylcarbamoyl-6,7-methylenedioxyphthalazine (SYM2206), 2-Aminoethoxydiphenylborane (2-APB), 1-[6-[[(17β)-3-Methoxyestra-1,3,5(10)-trien-17-yl]amino]hexyl]-1H-pyrrole-2,5-dione (U73122), and 1-[6-[[(17β)-3-Methoxyestra-1,3,5(10)-trien-17-yl]amino]hexyl]-2,5-pyrrolidinedione (U73343) were purchased from Tocris (Bio-Techne Corporation, Minneapolis, MN, USA). (4R,6S,8S,10Z,12R,14R,16E,18R,19R,20S,21S)-11,19,21-trihydroxy-4,6,8,12,14,18,20-heptamethyl-22-[(2S,2′R,5S,5′S)-octahydro-5′-[(1R)-1-hydroxyethyl]-2,5′-dimethyl [2,2′-bifuran]-5-yl]-9-oxo-10,16-docosadienoic acid (Ionomycin) and 1-[[[5-(4-nitrophenyl)-2-furanyl]methylene]amino]-2,4-imidazolidinedione, monosodium salt (Dantrolene) were purchased from Cayman Chemical (Ann Arbor, MI, USA). N-methyl-D-aspartic acid (NMDA) was purchased from Sigma-Aldrich (St. Louis, MO, USA). 2-Amino-2-(3-cis/trans-carboxycyclobutyl)-3-(9H-thioxanthen-9-yl) propionic acid (LY393053) was obtained from Lilly Research Laboratories, Eli Lilly and Company (Indianapolis, IN, USA). (3R)-N-acetyl-3-hydroxy-L-leucyl-(αR)-α-hydroxybenzenepropanoyl-2,3-didehydro-N-methylalanyl-L-alanyl-N-methyl-L-alanyl-(3R)-3-[[(2S,3R)-3-hydroxy-4-methyl-1-oxo-2-[(1-oxopropyl)amino]pentyl]oxy]-L-leucyl-N,O-dimethyl-L-threonine, (7→1)-lactone (FR900359) [25] was a gift from Dr. Ken Blumer (Washington University School of Medicine, St. Louis, MO, USA).

2.2. Plasmid Constructs

ER-GCaMP6-150 (Addgene plasmid #86918; 24) was a gift from Dr. Ghazaleh Ashrafi (Washington University School of Medicine) with permission from Dr. Timothy Ryan (Weill Cornell Medicine, New York, NY, USA).

2.3. Primary Neuronal Culture and Transfection

Primary striatal cultures using neonatal 1-day-old rat pups were prepared and maintained as previously described [17,26]. The cells were plated onto poly-d-lysine-coated, glass-inserted P35 dishes (14 mm; 60,000/glass, Cellvis, Mountain View, CA, USA) for immunostaining or Ca²⁺ real-time imaging. Cells were cultured in humidified air with 5% CO₂ at 37 °C. The cultures were transfected with ER-GCaMP6-150 at Div (days in vitro) 11 using Lipofectamine 2000 (Thermo Fisher Scientific, Waltham, MA, USA) and immunostained or imaged in real time 4–7 days after transfection as previously described [18].

2.4. Immunocytochemistry

After transfection, primary striatal cultures at DIV 15–18 were fixed and stained as described previously [18]. Primary antibodies included mouse monoclonal anti-calnexin (1:50; BD Bioscience., Becton, NJ, USA) and chicken polyclonal anti-GFP (1:2000; Aves Labs., Tigard, OR, USA). Secondary antibodies included goat anti-mouse Cy3 (1:300; Jackson Immunoresearch, West Grove, PA, USA) and goat anti-chicken Alexa 488 (1:500; Molecular Probes, Eugene, OR, USA).

2.5. Measurement of ER Ca²⁺ Dynamics

ER-targeted, low-affinity GCaMP6-150 was used to measure ER Ca²⁺ dynamics. Transfected striatal neurons were preincubated with control medium (125 mM NaCl, 5 mM KCl, 20 mM Hepes, 1.5 mM CaCl₂, 1.5 mM MgCl₂, 10 mM glucose, pH 7.4) containing mGlu₁ antagonist, CPCCOEt (20 µM), and AMPA receptor, antagonist SYM2206 (25 µM), for 30 min at 37 °C before adding DHPG (100 µM) to measure cell surface mGlu₅ activation. In addition, the impermeable, non-transported mGlu₅ antagonist, LY393053 (LY53, 20 µM) [17], was included in the incubation buffer before adding Quis (20 µM) to measure specific intracellular mGlu₅ activation. To study the pathways involved in intracellular mGlu₅ activation, U-73122 (phospholipase C inhibitor, 10 µM), U73343 (inactive analog of U73122, 10 µM), 2-APB (IP3 receptor inhibitor, 100 µM), Dantrolene (ryanodine receptor inhibitor, 10 µM), or FR900359 (G_q/11 inhibitor,1 µM) was also included in the incubation buffer before adding Quis. In all cases, drugs at 100x concentrations were added to the side of the dish by hand-held pipette and allowed to diffuse over the cells.

2.6. Confocal Microscopy and Data Analysis

Fluorescent Measurements of ER Ca²⁺ were performed and quantitated as described previously [22]. Briefly, 4–7 days after transfection with GCaMP6-150, ER Ca²⁺ imaging in striatal neurons was conducted using a laser confocal microscope (Olympus BX 50WI, Center Valley, PA, USA) Fluoview1200 with an Olympus LUMPlanFl/lR 40×/0.80w objectives. Cultures were preincubated with a control medium and treated as described above. The real-time ER Ca²⁺ images were conducted at 3–5 sec intervals and collected by an Olympus Fluoview FVX Confocal Laser Scanning system using Fluoview 4.2 acquisition software (https://www.olympus-lifescience.com/en/downloads/detail-iframe/?0[downloads][id]=847249651) (accessed on 11 February 2021). Images were processed with MetaMorph (version 7.7) (https://www.moleculardevices.com/products/cellular-imaging-systems/acquisition-and-analysis-software/metamorph-microscopy) (accessed on 14 April 2023) Professional Image Analysis software, produced by Universal imaging corporation (Downingtown, PA, USA). Immunofluorescence was analyzed around the soma area or in proximal dendrites at a distance of 40 µm. The average intensity across all images in soma and neurites was calculated at each time point for each category treated with different agonists or antagonists and then compared. Fluorescent ER Ca²⁺ signals, in response to electrical activity (ΔF), were normalized to the resting fluorescence (F₀). The F₀ value was additionally corrected for background autofluorescence measured in a nearby non-transfected region. Separate controls were performed with each experiment, and Student’s t test was used to determine statistical significance. Half maximal response times (T½) were assessed from the slopes of drug-induced ER Ca²⁺ loss curves derived from the first significant point of deflection outwards to 100 s. The statistical analysis was performed by GraphPad prism 10.2 (GraphPad software, Boston, MA, USA). Aggregated data (N = 12 for each condition) were analyzed via one-way of variance (ANOVA) followed by Tukey’s multiple comparison test.

2.7. Animal Studies

All animal procedures were performed according to NIH guidelines and approved by the Washington University Institutional Animal Care and Use Committee under protocols 21-0052 and 22-0228. Animals were under the care of the Washington University School of Medicine Division of Comparative Medicine.

3. Results

ER-GCaMP6-150 as an ER Ca²⁺ indicator: Previously, we demonstrated that ER-localized mGlu₅ receptors would be oriented such that the ligand-binding domain is within the ER lumen. As such, agonists must cross both the cell surface lipid bilayer as well as the ER membrane for receptor activation [17,18,23]. Mechanistically, agonist transport is achieved via the sodium-dependent glutamate transporter and/or the cystine, glutamate xCT exchanger [17,18]. Using radiolabeled glutamate and Quis, as well as uptake studies and knock-out animals, we showed that the mGlu₅ agonist DHPG is a non-transported, non-permeable agonist, whereas glutamate and Quis enter the cell via active transport [17]. Although we see similar results in cortex, hippocampus, and spinal cord dorsal horn neurons [17,18,19,20], we typically use primary dissociated striatal neurons because >70% of the neurons are medium spiny neurons that express mGlu₅ [17]. Inasmuch as the G protein binding portion of mGlu₅ is in the cytoplasm, after ligand binding, G_q/11-coupled receptors like mGlu₅ activate PLC which, in turn, hydrolyzes the membrane PIP₂ into the second messengers IP₃ and diacylglycerol (DAG). IP₃ can then activate IP₃ receptors on the ER membrane, leading to the release of ER Ca²⁺. A schematic representation of these interactions is shown in Figure 1. The release of ER Ca²⁺ is frequently measured as an increase in cytoplasmic Ca²⁺ assessed by the Ca²⁺ indicator dye, Oregon Green^TM 488 BAPTA-1, AM. Alternatively, and more directly, decreases in ER Ca²⁺ content can be directly monitored by the loss of luminal fluorescence using the genetically encoded Ca²⁺ sensor, ER-GCaMP6-150 [24], that is restricted to the ER throughout the neuron (Figure 2A) and co-localizes with the ER marker calnexin (Figure 2B). Real-time imaging of somatic ER-GCaMP6-150 before and after the addition of either DMSO as a treatment control or the Ca²⁺ ionophore, ionomycin, are shown in Figure 2C and quantified in Figure 2D. Consistent with published reports using this sensor [24], ionomycin (4 µM) treatment led to a rapid decrease in ER Ca²⁺ levels (Figure 2C,D).

mGlu₅ agonists trigger ER Ca²⁺ release in soma and neurites of striatal neurons: Once the correct targeting of ER-GCaMP6-150 to the ER was verified, endogenous mGlu₅ receptor-mediated Ca²⁺ responses were determined using pharmacological isolation of either cell surface or ER-localized receptors as we have described [18]. Consistent with the previous data, bath application of the transported agonist, Quis, generated a pronounced loss of ER Ca²⁺ in both neuronal cell bodies (−29.46 ± 4.59 S.E.M, n = 3) and neurites (−47.92 ± 4.88 S.E.M, n = 3) 100 sec after treatment (Figure 3A–C). These changes were entirely due to activation of ER-restricted mGlu₅ since they occurred in the presence of the impermeable, non-transported antagonist, LY393053, which we have shown to block all cell surface mGlu₅ receptor contributions [18] (Figure 3A–C). In contrast, pretreatment with the cell-permeant, mGlu₅ antagonist, MPEP, blocked all Quis responses (Figure 3C,E). To directly assess cell surface mGlu₅‘s ability to release ER Ca²⁺, sibling cultures transfected with ER-GCaMP5-150, were treated with the impermeable, non-transported agonist, DHPG. DHPG led to an ER Ca²⁺ loss of 18.85% ± 2.13 S.E.M, n = 3 in the soma and a loss of 25.35% ± 5.37 S.E.M, n = 3 in the neurites. Notably, DHPG had a more modest effect in triggering ER Ca²⁺ release than Quis (~66% of a Quis somal response and 55% of a Quis neurite response; Figure 3D,E). In comparison to ionomycin, which, at 100 sec, has maximally released ER Ca²⁺ levels from the soma (Figure 2D; (−78.50% ± 0.24 S.E.M, n = 3) or neurites (−83.79% ± 7.6 S.E.M, n = 3). At that same time point, DHPG released only 25–30% of ER Ca²⁺ measured in striatal soma or neurites, respectively. Quis activation of ER-localized mGlu₅ receptors led to the release of ~40% of somal ER Ca²⁺ in comparison to ionomycin, but >50% of maximal ER Ca²⁺ release in striatal neurites (Figure 3E). Taken together, these data unequivocally establish that ER-localized mGlu₅ is functionally active and is the major contributor to the increased cytoplasmic Ca²⁺ levels following receptor activation.

Quis triggers more rapid ER Ca²⁺ release than DHPG in striatal neurons: Using this same preparation of dissociated striatal neurons, we previously showed that cell surface-localized and intracellular mGlu₅ are associated with distinct patterns of Ca²⁺ release such that cell surface receptors exhibited rapid transient Ca²⁺ responses, whereas intracellular mGlu₅ exhibited sustained Ca²⁺ signals [18,21]. Examining the slopes of ER Ca²⁺ release over time following Quis or DHPG treatment in Figure 3C,D, it is clear that Ca²⁺ is being released more rapidly following Quis application in the soma and the neurites than is the case following DHPG addition. To quantify these differences, we calculated the T½ after drug application based on experimental curves such as those represented in Figure 2D and Figure 3C,D. In keeping with the data shown in Figure 2 and Figure 3, Quis triggered a more rapid ER Ca²⁺ release than DHPG at 100 sec in striatal cell bodies (29.86 s ± 4.18 SD) and especially neurites (20.33 s ± 6.23 SD; Figure 4). Ionomycin was 13.95 s ± 4.37 SD in cell bodies and 6.40 s ± 2.32 SD in neurites. DHPG exhibited a T½ of 43.06 s ± 8.02 SD in cell bodies and 39.83 s ± 9.83 SD in neurites. In striatal neurites, Quis was ~60% of the ionomycin rate of ER Ca²⁺ release, whereas DHPG was only 30% (Figure 4). Even at 5 s, Quis was >50% of the ionomycin release rate, while DHPG was not significantly different than the DMSO control (Figure 2D and Figure 3D). Collectively, these data indicate that the activation of the ER-localized mGlu₅ receptor releases more total ER Ca²⁺ at a faster rate than the activation of cell surface receptors.

ER Ca²⁺ release triggered by Quis is blocked by U-73122, FR900359, 2-APB, and Dantrolene: ER-resident GPCRs appear to use a variety of G proteins to activate downstream pathways. In particular, release of ER Ca²⁺ has been linked to pertussis toxin-sensitive pathways, suggesting G_i/o-driven mechanisms [27]. However, we have previously shown that, in HEK293 cells stably expressing mGlu₅ and/or endogenous receptors expressed in striatal neurons, pertussis toxin does not affect mGlu₅-mediated Ca²⁺ changes [23,28]. Rather, in either case, mGlu₅ couples with G_q/11/PLC/IP₃ to release cytoplasmic and nucleoplasmic Ca²⁺. Similarly, we hypothesized that ER-localized mGlu₅ couples to G_q/11 proteins to activate downstream signaling components. To confirm whether activated ER-localized mGlu₅ coupled to PLC to generate IP₃-mediated release of Ca²⁺, we used the same real-time imaging assay of somatic ER-GCaMP6-150 before and after the addition of Quis in striatal neurons preincubated with various inhibitors such as 1 µM FR900359 (G_q/11 inhibitor), 10 μM U73122 (a PLC inhibitor), 10 μM U73343 (an inactive analog of U73122), 100 μM 2-APB (an IP₃ receptor inhibitor), and 10 µM Dantrolene (ryanodine receptor inhibitor). Results showed that ER Ca²⁺ release triggered by Quis was blocked by FR900359 (−2.74 ± 1.59 S.E.M), U-73122 (−4.37 ± 0.48 S.E.M), 2-APB (−2.94 ± 0.87 S.E.M), and Dantrolene (−3.72 ± 0.75 S.E.M). It was not blocked by U733343 (−25.85 ± 2.34 S.E.M; Figure 5A,C). These data confirm that, just like cell surface and nuclear receptors, ER-localized mGlu₅ also couples to G_q/11 to activate cytoplasmic PI-PLC, leading to the hydrolysis of PIP₂ and generation of IP₃. The latter leads to the release of Ca²⁺ from the ER into the cytoplasm. Collectively, these data show that ER-localized mGlu₅ can function independently of signals originating at the cell surface and thus plays a dynamic role in mobilizing Ca²⁺ in a specific, localized manner.

NMDA does not release Ca²⁺ from the ER in striatal medium spiny neurons: Inasmuch as glutamate not only activates mGlu receptors but also ionotropic glutamate receptors such as NMDA, cytoplasmic Ca²⁺ changes could also be due to Ca²⁺ flux via these channels. Although we pharmacologically block these channels as well as other mGlu receptors, NMDA receptors often co-localize with metabotropic glutamate receptors to influence neuronal Ca²⁺ dynamics and synaptic plasticity [29,30,31,32,33]. To test whether NMDA affected intracellular mGlu₅-mediated Ca²⁺ responses in striatal neurons, we quantified changes in ER Ca²⁺ upon treatment with NMDA (10 µM). Changes in ER Ca²⁺ 100 s after applying Quis or NMDA are shown in Figure 5C. Bath application of NMDA had no significant effect on ER Ca²⁺ levels (−2.29 ± 0.60 S.E.M), whereas Quis treatment of sibling cultures did (Figure 5B,C). Therefore, despite previous reports, NMDA does not influence agonist effects on intracellular mGlu₅′s release of ER Ca²⁺.

4. Discussion

A growing body of data has established that GPCRs not only signal from the cell surface but also from intracellular compartments. One such receptor is mGlu₅ which, over 20 years ago, was shown to be present and active not only on the cell surface but also on isolated nuclei [28]. Since mGlu₅ couples to G_q/11, these experiments used Ca²⁺ indicators such as Oregon Green BAPTA-AM or Fura2 to measure real-time changes in cytosolic Ca²⁺ levels as a proxy for release of ER Ca²⁺ [17,21]. Data from those experiments indicated that functional activity is generated by at least two separate pools of mGlu₅—on the cell surface and on the inner nuclear membrane. The ER-GCaMP6-150 probe allowed us to directly visualize a third pool of mGlu5 receptors, namely those located on the ER, an organelle difficult to isolate in a functionally intact state. Similar to mGlu₅ activity in purified striatal nuclei, the activation of ER-restricted mGlu₅ produced faster, larger, and longer effects visualized with the ER-GCaMP6-150 probe than did the activation of the cell surface receptor. At peak response times, cell surface responses were ~64% of the mGlu₅ ER response in cell bodies and 53% in neurites (Figure 3C–E). In keeping with ER peak responses being larger, the amount of Ca²⁺ released was much greater, being ~2-fold more within the cell body and ~4-fold more in neurites, measuring the area of the Ca²⁺ curve from initiation outwards to 100 s (n = 12 neurons/each condition) [20,22]. ER-restricted, mGlu₅-mediated Ca²⁺ responses were blocked by the Gq inhibitor, FR9000359, the PLC inhibitor, U73122, as well as both IP₃ and ryanodine receptor blockers. Thus, activated ER-restricted mGlu₅ receptors signal via the same canonical G_q/11/PLC/IP₃ pathway that is found at the plasma membrane as well as the inner nuclear membrane [34,35,36,37,38,39,40]. Taken together, these data reinforce the notion that mGlu₅ can signal from different membrane platforms to generate downstream sequelae with unique spatiotemporal profiles.

Although DHPG has been widely used to simulate a specific mGlu₅ receptor response (e.g., “chemical” LTD in hippocampal slices) [41], it is glutamate that is released at the synapse and is transported into the cell where, as we have shown, it can activate the large intracellular pool of mGlu₅ receptors [19]. In striatal, cortical, and spinal cord dorsal horn neurons, we have found that DHPG activation of cell surface mGlu₅ elicited a rapid, transient Ca²⁺ response, whereas Quis (or glutamate) activation of intracellular mGlu₅ produced sustained Ca²⁺ responses [18,20,21]. Not surprisingly, these unique spatiotemporal differences led to markedly different signaling outcomes. For example, we found that intracellular mGlu₅ primarily uses protein synthesis-dependent MEK/ERK pathways to generate striatal LTD, whereas cell surface mGlu₅ uses mammalian target of rapamycin complex 2 (mTORC2) [22]. Thus, it is critical to recognize that glutamate is not just activating its plethora of ionotropic and metabotropic receptors on the cell surface but, at least for mGlu₅ and mGlu₁, it is activating intracellular components as well.

Although generally considered to be a continuous organelle, within neurons, the ER itself is highly compartmentalized, ranging from tubules and cisternae in dendrites to sheetlike cisternae in the soma and then very narrow tubules running down the axons [42]. As the major intracellular Ca²⁺ reservoir, ER regulation of Ca²⁺ signaling is highly involved in many functions including neurotransmitter release at synapses, protein synthesis, folding, and transport throughout the neuron, especially in somatodendritic regions. ER Ca²⁺ signaling is also responsible for integrating cellular interactions with other organelles like the mitochondria and plasma membranes to maintain homeostasis [43,44,45].

Many processes have been implicated in the ER release of Ca²⁺ store regulation including Ca²⁺ entry via NMDA receptors which, in turn, can trigger ryanodine receptor activation and further ER Ca²⁺ release [29,30,31,32,33]. Activation of NMDA receptors in striatal cultures did not release Ca²⁺ from the ER (Figure 5B,C), nor did we find any evidence of ER fission in striatal neurons in which several groups have reported to be associated with NMDA-mediated ER Ca²⁺ release [33]. It has also been reported that AMPAR-associated Ca²⁺ influx might contribute to this process. Given that we included an AMPA receptor inhibitor in the culture media prior to mGlu₅ agonist treatments, it seems doubtful that either ionic glutamate receptor is playing a role in this paradigm. Moreover, in the past, we have directly tested for AMPAR agonist effects in both wild-type and mGlu₅ KO cultures and have seen no change in baseline Ca²⁺ levels using cytoplasmic Ca²⁺ fluorophores [18]. As an additional control in those experiments, we also tested the NMDA inhibitor, APV, and saw no effect either [18]. Thus, we concluded that neither NMDARs nor AMPARs are contributing to the release of Ca²⁺ from the ER in our dissociated striatal culture paradigm.

Taken together, all of our data are consistent with activation of intracellular mGlu₅ directly releasing ER Ca²⁺. Collectively, these data underscore the importance of intracellular mGlu₅ in the cascade of events underlying sustained synaptic transmission.

5. Conclusions

Although past experiments have used caged ligand and real-time imaging to show that intracellular mGlu₅ is functional in dendrites, as the largest intracellular organelle, the ER extends throughout the cell, forming a complex network of tubules and sheets. Since it is impossible to isolate the entire organelle, the newly developed Ca²⁺ indicator, ER-GCaMP6-150, allows for the study of ER-localized mGlu₅ function in situ. Using this tool, we found that activation of ER-localized mGlu₅ resulted in pronounced somal Ca²⁺ declines compared to the cell surface receptor. Akin to the cell surface receptor, decreases in luminal Ca²⁺ were coupled to G_q/11 and PLC to generate IP₃-mediated release of Ca²⁺ from Ca²⁺ release channels in the ER. ER-localized mGlu₅ effects were twice as large and more rapid than those triggered by cell surface mGlu₅, especially in striatal neurites. Thus, the ER can function as an autonomous organelle mobilizing Ca²⁺ in a specific, localized fashion.