Diversity of Calcium Signaling by Metabotropic Glutamate Receptors*

During prolonged application of glutamate (20 min), patterns of increase in intracellular Ca2+concentration ([Ca2+] i ) were studied in HEK-293 cells expressing metabotropic glutamate receptor, mGluR1α or mGluR5a. Stimulation of mGluR1α induced an increase in [Ca2+] i that consisted of an initial transient peak with a subsequent steady plateau or an oscillatory increase in [Ca2+] i . The transient phase was largely attributed to Ca2+mobilization from the intracellular Ca2+ stores, but the sustained phase was solely due to Ca2+ influx through the mGluR1α receptor-operated Ca2+ channel. Prolonged stimulation of mGluR5a continuously induced [Ca2+] i oscillations through mobilization of Ca2+ from the intracellular Ca2+ stores. Studies on mutant receptors of mGluR1α and mGluR5a revealed that the coupling mechanism in the sustained phase of Ca2+ response is determined by oscillatory/non-oscillatory patterns of the initial Ca2+response but not by the receptor identity. In mGluR1α-expressing cells, activation of protein kinase C selectively desensitized the pathway for intracellular Ca2+ mobilization, but the mGluR1α-operated Ca2+ channel remained active. In mGluR5a-expressing cells, phosphorylation of mGluR5a by protein kinase C, which accounts for the mechanism of mGluR5a-controlled [Ca2+] i oscillations, might prevent desensitization and result in constant oscillatory mobilization of Ca2+ from intracellular Ca2+ stores. Our results provide a novel concept in which oscillatory/non-oscillatory mobilizations of Ca2+ induce different coupling mechanisms during prolonged stimulation of mGluRs.


During prolonged application of glutamate (20 min), patterns of increase in intracellular Ca 2؉ concentration ([Ca
] i ) were studied in HEK-293 cells expressing metabotropic glutamate receptor, mGluR1␣ or mGluR5a. Stimulation of mGluR1␣ induced an increase in [Ca 2؉ ] i that consisted of an initial transient peak with a subsequent steady plateau or an oscillatory increase in [Ca2؉] i. The transient phase was largely attributed to Ca 2؉ mobilization from the intracellular Ca 2؉ stores, but the sustained phase was solely due to Ca 2؉ influx through the mGluR1␣ receptoroperated Ca 2؉ channel. Prolonged stimulation of mGluR5a continuously induced [Ca2؉] i oscillations through mobilization of Ca 2؉ from the intracellular Ca 2؉ stores. Studies on mutant receptors of mGluR1␣ and mGluR5a revealed that the coupling mechanism in the sustained phase of Ca 2؉ response is determined by oscillatory/non-oscillatory patterns of the initial Ca 2؉ response but not by the receptor identity. In mGluR1␣-expressing cells, activation of protein kinase C selectively desensitized the pathway for intracellular Ca 2؉ mobilization, but the mGluR1␣-operated Ca 2؉ channel remained active. In mGluR5a-expressing cells, phosphorylation of mGluR5a by protein kinase C, which accounts for the mechanism of mGluR5a-controlled [Ca2؉] i oscillations, might prevent desensitization and result in constant oscillatory mobilization of Ca 2؉ from intracellular Ca 2؉ stores. Our results provide a novel concept in which oscillatory/non-oscillatory mobilizations of Ca 2؉ induce different coupling mechanisms during prolonged stimulation of mGluRs.
Ca 2ϩ can transduce many diverse cellular processes. Such diversity may be achieved by different amplitude and distinct spatial and temporal patterns of Ca 2ϩ response (1). In B lymphocytes, the amplitude and duration of Ca 2ϩ signaling controls differential activation of pro-inflammatory transcriptional regulators (2). In differentiating neurons, the frequency of [Ca2ϩ] i oscillations affects expression of specific neuronal phenotypes such as channel maturation and neurotransmitter expression (3). Compartmentalization of Ca 2ϩ signaling is also important in different cellular processes. For example, cytosolic Ca 2ϩ signals activate c-fos gene transcription through the serum response element, but nuclear Ca 2ϩ signals activate it through cyclic AMP response element (4).
Stimulation of two metabotropic glutamate receptor sub-types, mGluR1␣ and mGluR5a, triggers the release of Ca 2ϩ from the intracellular stores through inositol 1,4,5-triphosphate (InsP 3 ) 1 formation (InsP 3 /Ca 2ϩ pathway) (5)(6)(7)(8). We recently reported that transient application (1-60 s) of glutamate induces single-peaked intracellular Ca 2ϩ mobilization in mGluR1␣-transfected cells but elicits [Ca2ϩ] i oscillations in mGluR5a-transfected cells (9). The response patterns of the [Ca2ϩ] i increase depend upon the identity of a single amino acid, aspartate (at position 854) or threonine (at position 840), located within the G-protein-interacting domains of mGluR1␣ and mGluR5a, respectively. Phosphorylation of threonine (840) of mGluR5a by PKC interferes with the signal transduction through mGluR5a. We hypothesized that repetitive phosphorylation and dephosphorylation of mGluR5a could induce [Ca2ϩ] i oscillations by signaling on and off. In mGluR1␣, nonphosphorylation at aspartate (854) produces a non-oscillatory and PKC activator-resistant Ca 2ϩ response (9). This previous study provides the first evidence that an agonist can produce oscillatory/non-oscillatory patterns of Ca 2ϩ response by stimulating different receptor subtypes. However, it remained uncertain whether and how these two mGluRs control different cellular processes depending on their oscillatory/non-oscillatory Ca 2ϩ responses. We report here that prolonged stimulation of mGluR1␣ induced an increase in [Ca2ϩ] i that consisted of an initial transient peak and a subsequent steady plateau or an oscillatory increase in [Ca 2ϩ ] i . The transient phase was largely attributed to Ca 2ϩ release from intracellular Ca 2ϩ stores, but the sustained phase was solely due to Ca 2ϩ influx through a mGluR1␣ receptor-operated Ca 2ϩ channel. On the other hand, prolonged stimulation of mGluR5a continuously induced [Ca 2ϩ ] i oscillations through mobilization of Ca 2ϩ from the intracellular Ca 2ϩ stores. The coupling mechanism in the sustained phase of Ca 2ϩ response is determined by oscillatory/non-oscillatory patterns of the initial Ca 2ϩ response but not by the receptor identity. Thus, during prolonged stimulation of mGluRs, oscillatory/nonoscillatory patterns of Ca 2ϩ response lead to different coupling mechanisms in Ca 2ϩ signaling.
For construction of the mutant receptors, mGluR1␣(T) and mGluR5a(D), aspartate (854) of mGluR1␣ and threonine (840) of mGluR5a were changed into threonine and aspartate, respectively, as described previously (9). The cDNA encoding rat mGluR1␣, mGluR5a, * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ʈ To whom correspondence should be addressed: Neuroscience Research Laboratory, Institute for Drug Discovery Research, Yamanouchi Pharmaceutical Co., Ltd., 21 Miyukigaoka, Tsukuba-shi Ibaraki 305, Japan. Tel.: 298-52-5111; Fax: 298-56-2515. or a mutant receptor was inserted into the eukaryotic expression vector, pEF-BOS. After transfection of the above plasmids, HEK-293 cells expressing mGluR1␣, mGluR5a, or a mutant receptor were selected with 400 g/ml geneticin and isolated by a single cloning step (9). These cells were loaded for 45 min with Fura-2/AM (6 M) dissolved in balanced salt solution containing 135 mM NaCl, 5.4 mM KCl, 1.8 mM CaCl 2 , 0.9 mM MgCl 2 , and 10 mM HEPES (pH 7.4). After incubation, the coverslips were mounted in a laminar flow chamber with a flow rate of 2 ml/min at 32°C. The chamber was mounted on a Nikon inverted stage microscope. [Ca 2ϩ ] i measurement was started at 15 min after superfusion of balanced salt solution. Light from a Xenon lamp was filtered through either of two different band-pass filters (340 nm or 380 nm) in the excitation path and conducted to the specimen on the microscope stage through a diachronic mirror. The excitation wavelength was constantly switched between 340 nm and 380 nm. The fluorescence emitted from the cells was passed through a band-pass filter (510 nm). The video images were obtained using an intensified charge-coupled camera. Output from the camera was digitized and stored by a computerized imaging system (Hamamatsu, Argus 50). Ratios of sequential 340/380-nm excitation image pairs were compared with a standard curve for free Ca 2ϩ constructed from shallow solutions of known Ca 2ϩ and Fura-2 concentration. EGTA (1 mM) was included instead of CaCl 2 in all experiments using a Ca 2ϩ -free extracellular buffer. The application of reagent or change of external medium is indicated in each graph by a bar or set of bars.

Ca 2ϩ Responses in HEK-293 Cells Expressing mGluR1␣-In
HEK-293 cells expressing mGluR1␣, prolonged application of 30 M glutamate (20 min) induced an initial transient peak response followed by an oscillatory or a steady plateau Ca 2ϩ response ( Fig. 1, a and b). 44% of cells showed [Ca 2ϩ ] i oscillations, 39% showed steady plateau. [Ca 2ϩ ] i oscillations observed in the sustained phase of mGluR1␣ stimulation are characterized by base-line spikes of relatively constant amplitude. A minor population of cells showed three types of responses; a single transient peak response, an initial transient peak with a subsequent response best described as spike plateau, and an initial peak slowly descending to the base line followed by a base-line spiking type of [Ca 2ϩ ] i oscillations (data not shown). The transient peak of Ca 2ϩ response in mGluR1␣-expressing cells was largely attributed to Ca 2ϩ mobilization from the intracellular Ca 2ϩ stores, because Ca 2ϩ response was only slightly reduced in the absence of external Ca 2ϩ (Fig. 5a). In contrast, [Ca 2ϩ ] i oscillations during the sustained phase in mGluR1␣-expressing cells were completely abolished in the absence of external Ca 2ϩ (Fig. 1c). The oscillatory Ca 2ϩ response in the sustained phase was blocked by 30 M SK&F96365, a receptor-operated Ca 2ϩ channel blocker (11) but not by 10 M nimodipine, a voltage-gated Ca 2ϩ channel blocker (12) (Fig. 2, a and b). Steady plateau Ca 2ϩ response and other types of responses in the sustained phase of mGluR1␣ stimulation were also fully dependent on extracellular Ca 2ϩ and blocked by SK&F96365 but not by nimodipine (data not shown). Ca 2ϩ influx in the sustained phase of mGluR1␣ stimulation is mGluR1␣ receptor-operated, because mGluR1␣ antagonist, 1a-(N-phenyl)carbamoyl-1a,7a-dihydro-7(1H)-hydroxyiminocyclopropa[b]chromen (10), blocked Ca 2ϩ influx in the sustained phase of Ca 2ϩ response (Fig. 2c).
In mGluR1␣-expressing cells, in the absence of external Ca 2ϩ , 100 M carbachol could still mobilize Ca 2ϩ from intracellular Ca 2ϩ stores during prolonged application of 30 M glutamate (Fig. 3), indicating that during prolonged stimulation of mGluR1␣, the InsP 3 /Ca 2ϩ pathway is turned off before Ca 2ϩ stores are depleted. These results show that during prolonged stimulation of mGluR1␣, the InsP 3 /Ca 2ϩ -pathway is selectively desensitized, and Ca 2ϩ entry from SK&F96365-sensitive Ca 2ϩ channel solely contributes to the [Ca 2ϩ ] i increase in the sustained phase of mGluR1␣ stimulation.
[Ca 2ϩ ] i Oscillations in HEK-293 Cells Expressing mGluR1␣ or mGluR5a-Prolonged application of glutamate (20 min) constantly elicited sinusoidal [Ca 2ϩ ] i oscillations in cells expressing mGluR5a (Fig. 4a). In the late phase of stimulation, these [Ca 2ϩ ] i oscillations were not abolished in the absence of external Ca 2ϩ (Fig. 4b). Thus, prolonged stimulation of mGluR5a continuously mobilizes Ca 2ϩ from intracellular stores. In mGluR5a-expressing cells, the frequency of oscillations was lowered by the removal of external Ca 2ϩ . Thus, a source of extracellular Ca 2ϩ is also involved in the generation of [Ca 2ϩ ] i oscillations, which is in good agreement with the notion that the lack of Ca 2ϩ influx lengthens the time required for the refilling of intracellular Ca 2ϩ stores (13)(14)(15).
It has been proposed that [Ca 2ϩ ] i oscillations can be classified into two types, base-line spiking and sinusoidal oscillations, and can be driven by several different mechanisms (15)(16)(17). treatment with 100 nM PMA (18) reduced but did not abolish the Ca 2ϩ response induced by transient application (20 s) of glutamate (Fig. 5b). This PMA-resistant Ca 2ϩ response is fully dependent on the influx of extracellular Ca 2ϩ , because the Ca 2ϩ response was abolished in the absence of external Ca 2ϩ (Fig. 5c). These results show that activation of PKC is responsible for desensitization of InsP 3 /Ca 2ϩ -pathway, but the mGluR1␣ receptor-operated Ca 2ϩ -permeable channel is resistant to desensitization by PKC in mGluR1␣-expressing cells. Dissociation between InsP 3 formation and Ca 2ϩ influx strongly suggests that neither the formation of inositol phosphate me-tabolites nor Ca 2ϩ mobilization is needed for the activation of the mGluR1␣ receptor-operated Ca 2ϩ channel.
Ca 2ϩ Responses in HEK-293 Cells Expressing Mutant mGluRs-Different patterns of Ca 2ϩ response in mGluR1␣expressed and mGluR5a-expressed cells elicited by transient application (20 s) of glutamate results from a single amino acid substitution, aspartate of mGluR1␣ (position 854) or threonine of mGluR5a (position 840) (9). In cells expressing mGluR1␣(T), prolonged application of 30 M glutamate (20 min) elicited constant [Ca 2ϩ ] i oscillations that were not abolished in the absence of external Ca 2ϩ (Fig. 6a). On the other hand, in cells expressing mGluR5a(D), prolonged application of 30 M glutamate (20 min) induced an initial transient peak response followed by a steady plateau or an oscillatory Ca 2ϩ response (Fig.  6, b and c), both of which are identical with those in mGluR1␣expressing cells. 20% of cells showed [Ca 2ϩ ] i oscillations, 63% showed steady plateau. The sustained phase of Ca 2ϩ response in mGluR5a(D)-expressing cells was abolished in the absence of external Ca 2ϩ (Fig. 6, b and c). These results indicate that the coupling mechanism in the sustained phase of Ca 2ϩ response is determined by oscillatory/non-oscillatory patterns of the initial Ca 2ϩ response but not by the receptor identity.
In this study, we found that mGluR1␣ and mGluR5a show distinct coupling mechanisms during prolonged stimulation by glutamate (20 min). In the initial phase of stimulation, both mGluR1␣ and mGluR5a trigger the release of Ca 2ϩ from the intracellular stores through InsP 3 /Ca 2ϩ pathway. In the sustained phase of stimulation, in mGluR1␣-expressing cells, InsP 3 /Ca 2ϩ -pathway is desensitized, and Ca 2ϩ entry solely contributes to [Ca 2ϩ ] i increase. In contrast, in mGluR5a-expressing cells, InsP 3 /Ca 2ϩ pathway is not desensitized, and Ca 2ϩ is mobilized continuously from the intracellular Ca 2ϩ stores during the prolonged stimulation. In mGluR5a-expressing cells, Ca 2ϩ entry is also involved in the generation of [Ca 2ϩ ] i oscillations, because the frequency of oscillations is lowered by the removal of external Ca 2ϩ . At present, it is unclear whether or not the Ca 2ϩ -permeable channels, which are activated during stimulation of mGluR1␣ or mGluR5a, are the same. However, the coupling mechanisms during prolonged stimulation of these two mGluRs are clearly distinct in that mGluR5a continuously couples to InsP 3 /Ca 2ϩ pathway, but mGluR1␣ does not.
The studies on mutant receptors of mGluR1␣ and mGluR5a demonstrate that desensitization of the InsP 3 /Ca 2ϩ pathway occurs when the initial Ca 2ϩ response is non-oscillatory. In contrast, the sinusoidal [Ca 2ϩ ] i oscillations seen in mGluR5aor mGluR1␣(T)-expressing cells prevent the InsP 3 /Ca 2ϩ pathway from desensitization during prolonged stimulation of these receptors. The precise mechanism by which the sinusoidal [Ca 2ϩ ] i oscillations avoid InsP 3 /Ca 2ϩ pathway desensitization is unclear; however, our earlier study of mGluR5a-controlled [Ca 2ϩ ] i oscillations may give a hint (9). In that study, we showed that PKC inhibitors eliminate [Ca 2ϩ ] i oscillations and convert Ca 2ϩ response from an oscillatory to non-oscillatory pattern in mGluR5a-expressing cells. In contrast, the PKC activator, PMA, abolishes the Ca 2ϩ response in mGluR5a-expressing cells (9). We suggested that phosphorylation of mGluR5a by PKC inactivates mGluR5a, thus resulting in the decrease of [Ca 2ϩ ] i , whereas subsequent dephosphorylation of mGluR5a restores the signal transduction through mGluR5a, thus regenerating the [Ca 2ϩ ] i increase. We had then proposed that repetitive cycles of phosphorylation and dephosphorylation of mGluR5a generate [Ca 2ϩ ] i oscillations (9). Although not proven by experiments, continuous cycling of phosphorylation/dephosphorylation of mGluR5a may be provided by oscillations in the activity of PKC. During mGluR5a stimulation, not only [Ca 2ϩ ] i , but also the cellular level of diacylglycerol, the other bifurcating limb of phosphoinositide pathway (19), would oscillate. It is conceivable that PKC activity that is known to be affected by Ca 2ϩ and diacylglycerol would also oscillate (20). In the present study, we found that PKC is responsible for desensitization of InsP 3 /Ca 2ϩ pathway during mGluR1␣ stimulation. If PKC activity, which is first incremented by mGluR5a stimulation, decreases rapidly before the InsP 3 /Ca 2ϩ pathway is desensitized, such an oscillating PKC activity would prevent the InsP 3 /Ca 2ϩ pathway from desensitization in mGluR5a-expressing cells.
In cultured astrocytes, glutamate induces [Ca 2ϩ ] i oscillations through mGluR5 (21,22). Similar to the observations in mGluR5a-expressed HEK-293 cells (9), the PKC activator abolishes Ca 2ϩ response in cultured astrocytes (21,23). Moreover, both PKC inhibitor and PP1/PP2A phosphatase inhibitor convert Ca 2ϩ response from an oscillatory to non-oscillatory pattern, suggesting that the same mechanism underlies the generation of [Ca 2ϩ ] i oscillations in mGluR5-expressed heterologous and native cells (21). Thus, it may also occur in native cells that oscillatory/non-oscillatory mobilizations of Ca 2ϩ result in distinct coupling mechanisms during prolonged stimulation of mGluRs. Although it remains to be elucidated whether these different coupling mechanisms would indeed establish diverging cellular processes, our results provide a new insight into Ca 2ϩ signaling when long lasting stimuli are evoked in cells.