Proc. Nat!. Acad. Sci. USA Vol. 85, pp. 8737-8741, November 1988 Neurobiology

A regulates Ca2" mobilization in hippocampal (excitatory amino acids/calcium stores) SHAWN N. MURPHY AND RICHARD J. MILLER* Department of Pharmacological and Physiological Sciences, University of Chicago, Chicago, IL 60637 Communicated by Stephen J. Benkovic, August 15, 1988t

ABSTRACT We investigated the effect of various excita- MATERIALS AND METHODS tory amino acids on intracellular free Ca24 concentration ([Ca24]i) in single mouse hippocampal neurons in vitro by using Embryonic mouse C57BL/6NHSD hippocampal neurons the Ca24-sensitive dye fura-2. In normal physiological solution, were cultured as monolayers as described (2) and used from glutamate, kainate, N-methyl-D-aspartate, and quisqualate all 3 to 14 days in culture. Cells were loaded with fura-2 in pentakis(acetoxymethyl ester) and checked for complete produced increases [Ca24]i. When all extracellular Ca2+ was deesterification with excitation spectra (2). The cells were removed, kainate and N-methyl-D-aspartate were completely bath-perfused and bath volume was exchanged every 7 sec. ineffective, but quisqualate and glutamate were able to produce Emissions from excitation at 340 nm and 380 nm alternating a spike-like Ca24 transient, presumably reflecting the release at 60 Hz were stored every second (17). Each wavelength was of Ca24 from intracellular stores. Ca24 transients of similar inspected for drug autofluorescence and changes in dye shape could also be produced by the a1-adrenergic agonist fluorescence. Fura-2 fluorescence typically decayed 10-20%o phenylephrine. After the production of a Ca24 transient a in 1 hr, which represented mostly dye leakage from cells second addition of quisqualate was ineffective unless intracel- when compared to nonirradiated controls. Estimation of lular stores were refilled by loading the cell with Ca24 following [Ca2e]i could be obtained by using calibration solutions (2) or depolarization in Ca24-containing medium. None of the con- by whole-cell calibration (8). Rmin, Rmax, and K were ob- ventional excitatory amino acid receptor antagonists inhibited tained by each method [0.367 ± 0.06, 14.5 ± 2.6, 3200 ± 160 the Ca24-mobilizing effects of quisqualate. Furthermore a- and 0.389 ± 0.03, 8.47 ± 1.3, 2360 ± 570, respectively (mean amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA) + SD, n = 3)]. The former method allowed a more precise was unable to produce Ca24 mobilization in Ca24-free me- determination of the constants and was therefore used for dium, although it could produce Ca24 influx in Ca24- display of the data. All experiments were performed in containing medium. Thus, glutamate can produce mobilization Hepes-buffered Hanks' balanced salt solution (pH 7.45; of Ca24 from intracellular stores in hippocampal neurons by components, in mM: NaCl, 137; KCI, 5.4; MgSO4, 0.41; acting on a quisqualate-sensitive but AMPA-insensitive recep- MgCl2, 0.49; CaC12, 1.26; KH2PO4, 0.44; Na2HPO4, 0.64; tor. This receptor is therefore distinct from the quisqualate NaHCO3, 3; glucose, 5.5; Hepes, 20). Ca2+-free solutions receptor that produces cell depolarization. The possibility that were achieved by adding 20 ,uM EGTA to nominally Ca2+- this Ca24-mobilizing effect is mediated by inositol triphosphate free medium. Equimolar N-methyl-D-glucamine was substi- production is discussed. tuted for Na+ in Na4-free solutions. Quisqualate, a-amino- 3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA), and It is currently believed that the excitatory ibotenate were purchased from Cambridge Research Bio- glutamate produces its effects by acting at at least three chemicals (Harston, U.K.) and found to be <1% glutamate separate types ofreceptors (1). This action leads to the gating by thin-layer chromatography. of a variety of ion channels that exhibit different permeabil- ities to monovalent cations and Ca2". An important conse- RESULTS quence ofthese actions is the influx ofCa2" into neurons and an increase in the intracellular concentration of free Ca2" Experiments were performed with single mouse hippocampal (2, 3). This increase is thought to be essential neurons in vitro (n = 84). [Ca2`]j was measured by micro- [Ca2+]i for fluorimetry using the Ca2+-sensitive dye fura-2 (17). We promoting glutamate-induced changes in neuronal activity (4) examined the effects of excitatory amino acids with the and (5) as well as cell death in pathological specific intention ofdetermining whether they could mobilize situations (6). Previous work in our own and other laborato- Ca2+ from intracellular stores. As we have previously dem- ries has demonstrated that glutamate stimulates Ca24 entry onstrated that activation of excitatory amino acid receptors into neurons through both N-methyl-D-aspartate (N-Me-D- produces Ca2+ influx due to the activation ofion channels (2), Asp)-gated ion channels and voltage-sensitive Ca2+ channels it was first necessary to establish conditions under which (2, 3). We now report that glutamate can also increase [Ca2+]i such effects were not observed. Fig. 1 demonstrates the by mobilizing intracellular Ca2" stores. This effect is inde- pendent of the presence of external Ca2` and is mediated by Abbreviations: AMPA, a-amino-3-hydroxy-5-methyl-4-- a glutamate receptor that may not be linked to an ion channel propionate; [Ca2+]i, intracellular concentration of free Ca2+; N-Me- and whose specificity appears distinct from those of other D-Asp, N-methyl-D-aspartate; InsP3, inositol trisphosphate. glutamate receptors. *To whom reprint requests should be addressed at: Department of Pharmacological and Physiological Sciences, University of Chi- cago, 947 East 58th Street, Chicago, IL 60637. The publication costs of this article were defrayed in part by page charge tCommunication of this paper was initiated by Emil Thomas Kaiser payment. This article must therefore be hereby marked "advertisement" and, after his death (July 18, 1988), completed by Stephen J. in accordance with 18 U.S.C. §1734 solely to indicate this fact. Benkovic. 8737 Downloaded by guest on September 25, 2021 8738 Neurobiology: Murphy and Miller Proc. Natl. Acad. Sci. USA 85 (1988) 1200 r

900 F

CM 600 F Cu 0

300 F

-j QL 50 mM K+ 1 0 gm Q 300 M KA - Ca+ 0 10 20 30 40 Time, min FIG. 1. Quisqualate (Q) but not kainate (KA) increases [Ca2+], in Ca2l-free medium. In this hippocampal , depolarization with 50 mM K+ produced a large increase in [Ca2+],. In Ca2+-free medium (-Ca2+), kainate was ineffective but quisqualate produced a Ca2+ transient. When the neuron was returned to normal medium, kainate produced a large increase in [Ca2+],. On switching back to Ca2+-free medium, [Ca2+], fell and then rose again when Ca2' was reintroduced. effects of kainate in both normal and Ca2"-free media. In pleted in a time-dependent manner when the external Ca2" Ca2"-free medium kainate produced no effect. However, in concentration is reduced. Because of this lability, we always normal medium kainate produced a large increase in [Ca2+]i challenged cells with agonist in Ca2"-free medium 3 min after similar in form to increases in [Ca2"]J produced by depolar- Ca2" loading (except under special circumstances; see Fig. 2 ization with 50 mM K+. When the external Ca2" was A and B). Application of quisqualate to glial cells never removed during the kainate application, [Ca2+] immediately produced an increase in [Ca21]i. In neurons, Ca2+ transients fell back toward basal levels only to increase once more on could be produced in Ca2+-free medium by glutamate (Fig. the reintroduction of Ca2". In Ca2--free medium, depolar- 2C) (EC50 = 30 ,uM) and by ibotenate (EC50 = 100 ,uM) but ization of cells with 50 mM K+ or addition of N-Me-D-Asp not by kainate (Fig. 1) or N-Me-D-Asp (Fig. 2D) even at (Fig. 2) also produced no increase in [Ca2+]i. These results extremely high concentrations. Furthermore, coapplication indicate that under the Ca2+ free conditions used in these of either kainate or N-Me-D-Asp (Fig. 2D) did not reduce the studies, activation of Ca21 influx is not observed. effects of quisqualate. Quisqualate proved to be a potent In contrast to the above observations, Figs. 1 and 2A stimulus (EC50 = 200 nM). No response was normally illustrate the effect of quisqualate on [Ca2+]i in Ca2`-free observed at 10 nM quisqualate, but clear responses were seen medium. Quisqualate was equally effective in the presence of at 100 nM, with maximal effects at 10 ,uM (Fig. 2E). The tetrodotoxin (1 AM) or when all external Na+ had been absolute magnitude ofthe effect produced by quisqualate was replaced by N-methyl-D-glucamine. Quisqualate produced a quite variable, however: A[Ca2+]i = 163 ± 26 nM (mean ± clear spike-like Ca2+ transient that, under these conditions, SEM, n = 29). An interesting feature of the effect produced presumably reflects Ca2+ release from an intracellular store. by quisqualate is that it showed only minor desensitization. A second challenge with quisqualate produced no detectable In Fig. 2F we compare the effect of multiple additions of effect, indicating that the stores had become depleted after quisqualate to similar challenges with the a1-adrenergic the initial challenge (Fig 2A). However, we found that agonist phenylephrine. Multiple responses to quisqualate quisqualate-sensitive stores could be rapidly refilled if we could be obtained in a single cell under the appropriate loaded cells with Ca2+ by depolarizing them with 50 mM K+ conditions, but the response to phenylephrine in some cells in Ca2+-containing medium. This allows Ca2+ entry through was rapidly desensitized in spite of attempts to refill phenyl- voltage-sensitive Ca2+ channels (2). On switching back to ephrine-sensitive stores by the same procedure that proved Ca2+-free medium a further quisqualate-induced transient successful with quisqualate. We attempted to antagonize the could now be obtained (Fig. 2A). Thus, to ensure maximal effects of quisqualate with a variety of excitatory amino acid filling of intracellular Ca2` stores, we routinely applied a antagonists (Fig. 2G). However, when 1 ,uM quisqualate was 2-min depolarization to cells to load Ca2+ into intracellular used, y-tD-glutamylaminomethylsulfonate, L- stores before challenging them with the relevant agonist in diethyl ester, and DL-2-amino-4-phosphonobutyrate all failed Ca2+-free medium. Longer Ca2` loading did not increase the to produce antagonism even at concentrations as high as 1 magnitude of quisqualate-induced Ca2` release from intra- mM. Further experiments also indicated that the glutamate/ cellular stores. In Ca2+-free medium, quisqualate-sensitive quisqualate receptor involved in producing intracellular Ca2` intracellular stores are apparently quite labile (Fig. 2B). Thus, mobilization was distinct from that linked to cell depolariza- following a switch to Ca2+-free medium, quisqualate-induced tion. We previously demonstrated (2) that both quisqualate transients became progressively smaller the longer the time and the quisqualate analogue AMPA produce Ca2` influx that had elapsed prior to a challenge with quisqualate. Clearly into neurons (2). This influx results in a sustained increase in therefore some mechanism exists by which intracellular [Ca2+]i that persists while agonist is present. In the present stores can sense extracellular Ca2+, and they become de- study we found examples of cells in which quisqualate Downloaded by guest on September 25, 2021 Neurobiology: Murphy and Mifler Proc. Natl. Acad. Sci. USA 85 (1988) 8739

300E 1000 B

750 200 F

500 H

100 H 250 F

K..50, mM K+ - 100 tM Q X%-.A..J- 0 0 L -_10jiM Q Ca - Ca2+ 0 10 20 30 10 20 30 40 50 60 70

150 400r 150FE C 0

300 H 100 F 100 C cJ 200 H + 04CM 50F 100 H Em4 1~ 10ARM 10S1IQ 10 100 M G 1 0 _gM o - 300 iM 300 gM NMDA 0L,10M 10nM 4M 10p Q LII 57 62 66 71 14 19 24 29 34 39 6 11 16 21 26 31 36 41

600 200G

600

400 100

200

1 gM 1 gM 1 M 1 M Q ~K ~t < _ L%~J4~AJ~J 50 mM K+ 0 PorQ ol 1 mM AP4 1mM GAMS %ad7-Un 0 10 20 30 40 50 24 29 34 39 44 49 54 59 Time, min

FIG. 2. (A) Depletion and refilling of quisqualate (Q)-sensitive Ca2+ stores. A Ca2' transient was produced by quisqualate in Ca2+-free medium. A second challenge by quisqualate was without effect. After reintroduction of Ca2' and depolarization-stimulated Ca2' influx, quisqualate produced another response in Ca2+-free medium. (B) Lability of quisqualate-sensitive Ca2+ stores. Cell was challenged with quisqualate in Ca2'-free medium at 3, 5, 7, 9, and 3 min following depolarization-induced Ca2" influx in normal medium. (C) Glutamate (G) produces Ca2' transients in Ca2+-free medium (same neuron as in A). Each glutamate challenge was preceded by depolarization-induced Ca2+ influx in normal medium. These have been edited out of the trace for clarity and the numbers on the axis represent real time. (D) N-Me-D-Asp (NMDA) produces no increase in [Ca2+]i in Ca2+-free medium nor does it block the effects of quisqualate. Experiments with N-Me-D-Asp were run in Mg2+-free medium augmented with (10 jIM) (9). (E) Effect ofvarious concentrations ofquisqualate on [Ca2+]i in Ca2+-free medium. (F) The response to quisqualate shows little desensitization. Several responses to quisqualate in Ca2+-free medium could be obtained from a single cell. Phenylephrine (P) also produced Ca2+ transients in Ca2+-free medium. However the response to phenylephrine exhibited significant desensitization. (G) Lack of effect of excitatory amino acid antagonists on Ca2' transients produced by quisqualate. AP4, DL-2-amino4 phosphonobutyrate; GAMS, y-D-glutamylaminomethylsulfonate. Downloaded by guest on September 25, 2021 8740 Neurobiology: Murphy and Miller Proc. Natl. Acad. Sci. USA 85 (1988) produced Ca2" influx in Ca2"-containing medium but no hippocampal neurons corresponds exactly to any of these mobilization ofintracellular Ca2" in Ca2"-free medium, cells previously described receptors. However, in some cases in which quisqualate produced Ca2" mobilization but no quisqualate-induced InsP3 synthesis has been shown to be A[Ca2"Ii in Ca2"-containing medium without reloading Ca2" effectively blocked by 2-amino-4-phosphonobutyrate, al- stores, and cells in which both types of responses occurred. though this may be somewhat dependent on stage of devel- Furthermore, the concentration of quisqualate (-1 WM) opment (10, 12). 2-Amino-4-phosphonobutyrate was com- required to activate Ca2+ influx is greater than that required pletely ineffective in the present studies, indicating that the for mobilization of intracellular Ca2+ stores. In addition, receptor involved may not precisely correspond to any de- AMPA was never observed to produce Ca21 mobilization but scribed previously. Nevertheless, stimulation of InsP3 syn- could frequently be observed to produce Ca2+ influx. Fig. 3 thesis must still be considered as the most probable event illustrates an example of this. In this cell AMPA failed to mediating the effects of quisqualate. Neurons appear to produce a mobilization response in Ca2`-free medium but contain a variety of Ca2" stores, as ado other cell types (16). quisqualate was effective. On switching to normal Ca2 - One ofthese is sensitive to InsP3, whereas others are sensitive containing medium, AMPA produced Ca2` influx. A second to drugs such as caffeine and ryanodine and may be involved quisqualate response could now be obtained that can be seen in phenomena such as Ca2+-induced Ca2" release (16-18). superimposed on the AMPA response. This transient quis- Hippocampal neurons appear to contain both types of stores. qualate response is characteristic of Ca2+ mobilization. However, depletion of the caffeine-sensitive store does not alter the quisqualate response, indicating that only the InsP3- sensitive store is important in this case (unpublished obser- DISCUSSION vations). The results presented here demonstrate that a glutamate/ Glutamate appears to be able to increase [Ca2+]i as a result quisqualate-specific receptor is linked to Ca2+ mobilization Qf several distinct processes following the activation of in hippocampal neurons. This receptor appears to be different kainate-, quisqualate-, and N-Me-D-Asp-specific receptors. from the quisqualate-specific glutamate receptor that is These include Ca2+ influx through voltage-sensitive Ca2' responsible for cell depolarization. In addition to the lack of channels as a result of cell depolarization (2, 3), Ca2+ influx effect of AMPA in the present studies, we have also found through receptor-operated channels (2, 3), and, as shown differences in the ability ofantagonists to block the two types here, Ca2+ mobilization from intracellular stores. Little is of receptors (unpublished observations). It seems likely that known about the physiological consequences of Ca2+ mobi- the Ca2+-mobilizing responses we have observed are medi- lization from intracellular stores in central neurons. How- ated through the synthesis of inositol trisphosphate (InsP3). ever, Ca2+ has been implicated in the control of many Excitatory amino acid-induced InsP3 synthesis has been important neuronal functions. Of particular interest is the demonstrated in several areas of the brain, including the potential role of Ca2+ in excitatory amino acid-induced hippocampus (10-12), and probably also occurs following the changes in synaptic efficacy such as long-term depression expression of quisqualate receptors in frog oocytes (13). In (19) and potentiation (5). Indeed, in some cases of long-term addition, at the neuromuscular synapse ofthe lobster walking depression the involvement of a quisqualate receptor has leg, glutamate hyperpolarizes the presynaptic membrane, been specifically indicated. Apparently the effects of quis- another response that could also involve InsP3 synthesis and qualate on hippocampal InsP3 synthesis are dependent on Ca2` mobilization (14). Several ofthese responses have been developmental factors and seem to be greatest during periods shown to involve the participation of a pertussis toxin- of synaptogenesis. Determining the role of glutamate- sensitive guanine nucleotide-binding protein (13-15). It is not induced Cat+ mobilization in such processes will be of completely clear whether the receptor we have described in considerable interest. 500

400

c 300 F

+ 0sCM Con 200 F

100 F

0 50 mM K+ 1 0 gum Q 100 LM AMPA II I -Ca2+ o 4 5 10 15 20 25 30 35 Time, min FIG. 3. AMPA is unable to produce increases in [Ca2i] in Ca2l-free medium although quisqualate (Q) is effective. In normal medium AMPA stimulated Ca2" influx. Quisqualate then produced an additional response that was superimposed upon the AMPA response. Downloaded by guest on September 25, 2021 Neurobiology: Murphy and Miller Proc. Nati. Acad. Sci. USA 85 (1988) 8741

We thank Dr. Stanley Thayer for extensive discussions. This 9. Reynolds, I. J., Murphy, S. N. & Miller, R. J. (1987) Proc. research was supported by Public Health Service Grants DA02121, Nati. Acad. Sci. USA 84, 7744-7748. DA02575, and MH40165 and grants from Miles and Marion Labo- 10. Nicoletti, F., ladarola, M. J., Wroblewski, J. T. & Costa, E. ratories. S.N.M. was also supported by GM07151. (1986) Proc. Nati. Acad. Sci. USA 83, 1931-1935. 11. Sladeczek, F., Pin, J.-P., Recasens, M., Bockaert, J. & Weiss, 1. Mayer, M. C. & Westbrook, G. L. (1987) Prog. Neurobiol. 28, S. (1985) Nature (London) 317, 717-719. 197-276. 12. Schoepp, D. D. & Johnson, B. G. (1988) J. Neurochem. 50, 2. Murphy, S. N., Thayer, S. A. & Miller, R. J. (1987) J. Neu- 1605-1613. rosci. 7, 4145-4158. 13. Sugiyama, H., Ito, I. & Hirono, C. (1987) Nature (London) 325, 3. MacDermott, A., Mayer, M. L., Westbrook,. G. L., Smith, 531-533. S. J. & Barker, J. C. (1986) Nature (London) 321, 519-522. 14. Miwa, A., Kawai, N. & Ui, M. (1987) Brain Res. 416, 162-165. 4. MacDermott, A. B. & Dale, N. (1987) Trends Neurosci. 10, 15. Nicoletti, F., Wroblewski, J. T., Fadda, E. & Costa, E. (1988) 280-284. Neuropharmacology 27, 551-556. 5. Abrams, T. W. & Kandel, E. R. (1988) Trends Neurosci. 11, 16. Thayer, S. A., Perney, T. M. & Miller, R. J. (1988) J. Neuro- 128-135. sci., in press. 6. Choi, D. W. (1987) J. Neurosci. 7, 369-379. 17. Thayer, S. A., Sturek, M. & Miller, R. J. (1988) Pflugers Arch. 7. Kano, M. & Kato, M. (1987) Nature (London) 325, 276-279. 412, 216-223. 8. Scanlon, M., Williams, D. A. & Fay, F. S. (1987) J. Biol. 18. Lipscombe, D., Madison, D. V., Poenie, M., Reuter, H., Chem. 262, 6308-6312. Tsien, R. W. & Tsien, R. Y. (1988) Neuron 1, 355-365. Downloaded by guest on September 25, 2021