Characterization of a Glutamic Acid ... - Semantic Scholar

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1984 by The American Society of Biological Chemists, Inc

Vol. 259, No. 20, Issue of October 25, pp. 12756-12762 1984 Printed in il..S.A.

Characterization of a Glutamic Acid Neurotransmitter Binding Siteon Neuroblastoma Hybrid Cells* (Received for publication, February 14, 1984)

Alfred T. MaloufS, Ronald L. Schnaarg, and Joseph T. Coylell From the Departments of Pharmacology and Neuroscience and the llDivision of Child Psychiatry, Departments of Psychiatry and Pediatrics, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205

Glutamate is thought to be a major excitatory neu- the major excitatory neurotransmitters in the brain (1, 2). rotransmitter in the central nervous system. To study Hayashi was the first to demonstrate that glutamate has the glutamate receptor andits regulation under care- potent neuroexcitatory properties whenapplied to cortical fullycontrolledconditions, the specificbinding of neurons (3). Subsequently, an abundanceof electrophysiolog[‘Hlglutamate wascharacterizedin washed mem- ical and neurochemical evidence has accumulated to support branes isolated from a neuroblastoma X retina hybrid the candidacy of glutamate as a neurotransmitter for many cell line, NIB-RE-105. [‘HJGlutamate bound in a sat- neuronal pathways throughout the central nervous system (1, urable and reversible fashion with an apparent disso- 2). More recent electrophysiologic studies using specific agociation constant, KO,of 650 nM and a maximum bind- nists and selective antagonists have delineated three distinct ing capacity, B,,,, of 16 pmol/mg of protein. Pharmacologic characterization of the site indicates that it excitatory amino acid receptors which bind glutamate: the N-methyl-o-aspartate-reclosely resembles the Na+-independent bindingsite for quisqualate-responsive site,the sponsive site, and the kainate-responsive site (4). The quisglutamate found on brain membranes and thought to qualate-responsive site appears to be the major neurotransbe an excitatory amino acid neurotransmitter receptor. Thus, whilekainate, N-methyl-DL-aspartate, and non- mitter receptor for glutamate in the brain and is the central electrophysiological aminoacidligandsdidnotdisplace [‘Hlglutamate, concern of our studies. In addition to the quisqualate and ibotenate were potent inhibitors of studies referenced above, the quisqualate-responsive site has specific binding.Furthermore, this binding site is reg- been identified by direct bindingof [3H]glutamate toisolated ulated by ions in a manner which resembles that de- brain membranes (5-9). Inhibition of [3H]glutamate binding by glutamate analogs correlates well with scribed in the hippocampus (Baudry, M., and Lynch, to these membranes G. (1979)Nature (Lond.) 282,748-750). Calcium (10 their electrophysiological potency at the quisqualate-responmM) increased the number of binding sites 2.6-fold sive site. with no change in receptor-ligand affinity. Lanthanum Interestinthisneurotransmitterbindingsitehas been (1mM) was theonly other cation added which enhanced stimulated recently by data suggesting that its regulation is (%fold) the bindingof [’Hlglutamate. Monovalent cat- involved in a simple electrophysiologicalcorrelate tolearning: ions resulted in a decrease in the number of glutamate long-termpotentiation.Repeated highfrequencyelectrical binding sites. Incubation of membranes in the presencestimulation of certain pathways in the hippocampus results of chloride ions caused a marked increased in [3H] in a long-lasting increase in synaptic efficiency (perhaps via glutamate binding, an effectwhich was synergistic increased depolarization for equivalent electrical stimuli) in with that of calcium incubation. Thus, N18-RE-105 postsynaptic cells. This increased postsynaptic excitability is cells possess a binding sitefor [‘Hlglutamate pharma- defined as LTP’ (10). Since there is ampleevidence that the cologically similar to an excitatory neurotransmitter hippocampus is involved in learning and memory, it is thought binding site in brain and which exhibits regulatory that LTP may representone possible cellularmechanism properties resembling those previously described in hippocampal membranes, providing an excellent model involved in these phenomena. A relationship between regulation of the [3H]glutamate receptor and LTP in thehippofor mechanistic studies. campus has been reported by Baudry and Lynch (11-13). Theydemonstrated,in hippocampal slices, that high frequency electrical stimulation caused an increase in [3H]gluMuch evidence suggests that L-glutamic acid, in addition tamate binding, while low frequency stimulation, which did to itsrole in metabolism and protein structure, acts as one of not cause LTP, did not alter binding. Since glutamate has been reported to be a major excitatory neurotransmitter in * This work was supported by National Institutes of Health Grants the hippocampus, it is possible that the persistent up-regulaRSDA MH-00125, NS-13584, and HD-14010, and the Surdna Foun- tion of glutamate receptors could be responsible for LTP. In dation. The costs of publication of this article were defrayed in part addition,theformation of LTP in hippocampalslices is by the payment of page charges. This article must thereforebe hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 dependent on the presence of calcium ions in the external medium and [3H]glutamate binding to membranes from this solely to indicate this fact. $ Supported by Training Grant GM07626 from the National Insti- tissue is sharply increased by calcium in vitro (14). This tutes of Health. This work was performed in partialfulfillment of the calcium-induced up-regulation was found on isolated memrequirements for thedegree of Doctor of Philosophy from The Johns branes from areas of the brain which demonstrate LTP but Hopkins University. Present address, Division of Preclinical Neuronot those in which L T P is absent. Thus, a correlation exists science and Endocrinology, Scripps Clinic and Research Foundation, La Jolla, CA 92037. 5 Recipient of American Cancer Society Junior Faculty Research Award JFRA-51. To whom correspondence should be addressed.

The abbreviations used are: LTP, long-term potentiation; B,, maximum binding capacity; KD,apparent dissociation constant.

12756

Characterization of a Glutamate Receptor Neuroblastoma on

Cells

12757

volume of 1 ml/flask of cells collected unless otherwise stated. After sonication for 10 s in ice-cold buffer with a Sonifier cell disruptor (Ultrasonics; setting 6),the suspension was centrifuged at 30,000 x g for 10 min a t 4 "C and the supernatant fluid wasdiscarded. The pellet (membrane fraction) was resuspended by sonication in the same volume of ice-cold Tris-HC1 and recentrifuged as above. The pellet was finally resuspended as above in a volume of ice-cold Tris-HC1 equal to 1 ml/flask, which yielded a suspension having 1-2 mgof protein/ml. For studies in which membranes were preincubated with different ions, the appropriate salts weredissolved at the desired concentrations in Tris-HC1 (or other buffers as indicated) and the membrane pellet was resuspended by sonication in these solutions. Routinely, for each binding determination, 100 pl of the membrane suspension was added to a 1.5-ml Eppendorf (polypropylene) microcentrifuge tube containing 100 pmol of [3H]glutamate (adjusted to approximately 5,000 cpm/pmol with unlabeled glutamate) and other additives (see below) dissolved in a total volume of 900 pl of TrisHCl. In contrast, for saturation isotherms, the final concentration of radioligand ranged from 50 nM-1 pM. To measure nonspecific binding, unlabeled L-glutamic acid was added to a final concentration of 0.1 mM. Potential competitors and inhibitors of binding were added at concentrations indicated under "Results." Binding incubations were conducted for 20 min at 37 "C and then centrifuged for 10 min at 30,000 X g (4 "C). The supernatant was aspirated, 1 mlof ice cold buffer was added tothe tube without disturbing the pellet and immediately aspirated, and the pellet (with bound radioligand) was EXPERIMENTAL PROCEDURES dissolved in 0.5 ml of Protosol and subjected to liquid scintillation spectroscopy in 10 mlof Econofluor (with 0.5% (v/v) acetic acid Materials added). Enzyme Assays-One flask of N18-RE-105 hybrid cells was harThe following materials were obtained from the indicated sources: phosphate-buffered saline and tissue culture medium (No. 430-2100) and fetal calf serum (virus- vested in 10 mlof (Ca2+-Mg2+)-free screened, mycoplasma-tested) from Grand Island Biological Co. cells were collected by centrifugation at 250 X g for 5 min. The cells (GIBCO); dibutyryl cyclic adenosine 3':5'-monophosphate, eserine were resuspended in 1.0 ml of Tris-HC1 supplemented to 0.02% (v/v) sulfate, thymidine, hypoxanthine, aminopterin, bovine serum alhu- with Triton X-100 detergent. The pellet was triturated 5 times and min, trifluoperazine, a-aminoadipic acid, DL-a-aminopimelic acid, placed in a Potter-Elvehjem glass homogenizer with a Teflon pestle. glutamate diethyl ester, N-methyl-DL-aspartic acid, kainic acid, y- Cells were homognized (40 strokes) at 0 "C. Portions (50 pl) of the aminobutyric acid, and quisqualic acid from Sigma; aminophosphon- homogenate were assayed for tyrosine hydroxylase (17), choline aceobutyric acid from Calbiochem; ibotenate from Regis Chemical Co., tyltransferase (181, and glutamic acid decarboxylase (19) activities. Milwaukee, WI; ~-[3,4-~H]glutamic acid (44.1 Ci/mmol), ~ - [ 2 , 3 - ~ H ] Uptake Studies-High affinity [3H]glutamate and [3H]4-aminoaminobutyric acid (36.1 Ci/mmol), tissue solubilizer (Protosol), and butyric acid transport were studied using N18-RE-105 cells which liquid scintillation fluors from New England Nuclear; tissue culture had firmly attached to the bottom of Costar 24-well plates. The plasticware (75-cm2flasks No. 25116 and 35-mm dishes No. 25000) procedure was adapted from the method ofCoyle andEnna as described for rat brain synaptosomal preparations (20). Growth mefrom Corning Glass Works, Corning, NY. All other reagents were of the highest quality and were obtained from standard sources. Trifluo- dium was aspirated and cells were washed twice in Krebs phosphate perazine dihydrochloride was the gracious gift of Smith, Kline and buffer. The cells were preincubated for 10 min at 37 "C in 950 p1 of French Laboratories, Philadelphia, PA. Aminophosphonoheptanoic, the same buffer after which 50 p1 of [3H]glutamate (specific activity aminophosphonohexanoic, and aminophosphonovaleric acids were diluted to 0.44 Ci/mmol with unlabeled glutamate) or [3H]4-aminothe gifts of J. F. Collins, Department of Chemistry, City of London butyric acid (specific activity diluted to 0.36 Ci/mmol with unlabeled Polytechnic, London, United Kingdom. Mouse neuroblastoma X ret- 4-aminobutyric acid) were added to give a final concentration of 2 x M. After incubating the cells for 4 min (with glutamate) or 2 min ina hybrid cell lines were kindly made available by Dr. Marshall Nirenberg and Chinese hamster embryonic neural retina X fibroblast (with 4-aminobutyric acid), the cells were washed twice with ice cold cell lines were kindly provided by Dr. David Trisler, National Insti- buffer. One ml of 1 M NH40H was then added to each well and the suspension was triturated 10 times (removing all cells from the plate) tutes of Health, Bethesda, MD. and transferred to a scintillation vial, 10 mlof scintillation fluor (New England Nuclear Formula 947) and 60 pl of glacial acetic acid Methods were added, and radiolabel was quantified by liquid scintillation Cell Culture-N18-RE-105 hybrid cells (mouse neuroblastoma spectroscopy. Since the high affinity uptake processes are sodiumclone N18TG-2 X Fisher rat 18-day embryonic neural retina) were dependent, control wells were treated as above except that the NaCl grown essentially as described for other neuroblastoma hybrid cell in the Krebs buffer was replaced with choline chloride. Protein was lines (16). Briefly, stock cultures (passage 5-28) were maintained at measured by the method ofLowry et al. (21) with bovine serum 37 "C in 75-cm2 tissue culture flasks in 95% Dulbecco's Modified albumin as standard. Eagle's medium (supplemented with sodium bicarbonate (44mM), Statistical Analysis-All biochemical data were tested for signifithymidine (16 p ~ )hypoxanthine , (100 p ~ )and , aminopterin (1p ~ ) ) cance using the parametric Student's two-tailed t-test. and 5% fetal calf serum in a humidified atmosphere of 10% COz,90% air. Cellswere subcultured by removing the growth medium and RESULTS replacing it with 10 ml of (Cazf-Me)-freephosphate-buffered saline which contained 137 mM NaCl, 5.22 mM KCl, 0.168 mM Na2HPO4, Neuronal Cell Lines with Glutamate Neurotransmitter Bindand 0.22 mM KH2P04adjusted to pH7.4. After 10 min, the cells were removed from the flasks by agitation, collected by centrifugation at ing Sites-Seven clonal cell lines were screened for specific under conditions that were opti250 X g for 5 min, and reseeded in the above growth medium at a binding of ~-[~H]glutamate density of 300,000 cells/flask. All media were sterilized by filtration mal for labeling these sites inrat brain membranes (Table I). (0.22-pm pore cellulose acetate/nitrate filters, Millipore). The media Three cell lines derived from a hybridization of Chinese were normally replaced on the 3rd, 5th,and 6thday after seeding and hamster embryonic neural retina and fibroblasts (A2d, Ala-3, cells were subcultured on the 7th day. A2a-1) did not exhibit detectable specific binding. The NG[3HlGlutamate Binding-N18-RE-105 cells (or other neuronal cell lines), grown in 75-cmZ flasks, were harvested a t confluence and 108-15 line, a mouse neuroblastoma x rat glioma hybrid with collected by centrifugation as described above. The cell pellet was cholinergic characteristics, exhibited modest specific binding. resuspended in ice-cold 50 mM Tris-HC1, pH 7.1 (Tris-HCl), at a Of three mouse neuroblastoma x rat embryo retina hybrids

between LTP andglutamate receptor regulation. Studies onthe molecular mechanisms involved in glutamate receptor regulation are hampered by the cellular heterogeneity of brain tissue. Therefore, we initiated a study to identify a homogeneous clonal neuronal cell line which both possessed the pharmacologically relevant glutamate receptor and demonstrated calcium-dependent up-regulation. One such cell line, a neuroblastoma X retina hybrid designated N18-RE105, wasfound to possess a high density of such receptors. In this report, the binding of [3H]glutamate to membranes isolated from N18-RE-105 cells is characterized in terms of its kinetics, pharmacology, and regulation by monovalent and divalent ions. In addition, these binding characteristics are compared to those of hippocampal membranes. In theaccompanying report (15),the regulation of this glutamate receptor in intact cells will be described. Together, these two studies demonstrate agonist and ion up-regulation of glutamate receptors in a homogeneous neuronal cell line, providing an excellent system for characterization of the molecular mechanisms which may underlie long-term potentiation and help provide a molecular model for learning or memory.

12758

Characterization of a Glutamate Receptor Neuroblastoma on TABLE I to neuronnl cell lines

Cells

sine hydroxylase, cholineacetyltransferase,andglutamate decarboxylase revealed specific activities less than 1%of that Washed membranes were prepared from the indicated cell lines as in adult mouse brain. In addition, sodium-dependent uptake described under“Methods.” Specific [3H]glutamatebinding was mea- of ~-[~H]glutamate [‘H]4-aminobutyric or acid was quite low, suredin the presence of 600 nM radioligand (20-min incubation, indicating thecells use neither glutamic acid nor 4-aminobu37 “C). Membranes were collected from both control cells and cells grown for 5-7 days in the presence of 1 mM dibutyryl cyclic AMP, an tyric acid as neurotransmitter. Binding Characteristics of the Glutamate Receptor on N18agent used to induce differentiation in neuronal tissue culture cell lines. Nonspecific binding was determined by incubating membranes RE-105 Cells-Equilibrium binding analysisof [3H]glutamate with the radioligand in the presence of 100 p~ unlabeled glutamate to N18-RE-105 membranesdemonstrates a high ratio of and the resulting value wassubtracted to generate the specific binding specific to nonspecific binding (Fig. 2). While binding data reported. NCE-Sa-K, N18-RE-103, and N18-RE-105 are mouse neu- varied somewhat between cell preparations, analysis of 14 roblastoma X embryonic neural retina hybrids, NG-108-15 is a mouse separate isotherms with 2-6 determinations a t each concenneuroblastoma X rat glioma hybrid, and A2d, Ala-3, and A2a-1 are tration revealed a n average KDof 521 & 57 nM (X f S.E.) and embrvonic Chinese hamster neural retina X fibroblast hvbrids. a n average BmSxof 19.8 f 2.8 pmol/mg of protein (X f S.E.). [SH]Clutamatebinding to All binding was conducted at 37 “C and at membrane concenCell line Dibutyryl trations within the linear range of the assay. While no eviContra' CAMP-treated cells denceformultiplesites was detected, the possibility of a pmol/mg protein pmol/mg protein second site with a much lower KO cannot be ruled out from 0 0 A2d a single these data.Nevertheless, our further analysis assumes 0 0 Ala-3 site of binding based on the data from the isotherms per0 0 A2a-1 formed. 3.7 ND” NCE-Sa-K Inhibition isotherms for both linear and conformationally 3.7 6.3 N18-RE-103 6.2 7.8 N18-RE-105 restrictedexcitatory analogs of glutamate were performed 3.0 2.0 NG-108-15 (Fig. 3 and Table 11). The results of this structure-activity 7.2 Rat cerebellum study correlatedclosely with similar datacollected using these 18.2 Rat hippocampus compounds at the glutamate binding site of rat cerebellar ND, not determined. membranes (Table 11). Quisqualic acid was the most potent competitive blocker of [3H]glutamate binding washed to N18RE-105 membranes closely followed by L-glutamate itself. Of 17 analogs tested,14 exhibited K , values comparable to those reported in rat cerebellum, while only three compounds, Dglutamate, L-aspartate, and aminophosphonovalerate exhibited significantdiscrepancies, with only L-aspartate exhibiting a major shift in rank order of potency (Table 11). Analogs which have very low affinities for the cerebellar binding site have equally low affinities for the neuroblastoma binding site. Most notably,N-methyl-DL-aspartic acid and kainic acid, two potent glutamate analogs that act at excitatory amino acid sites distinct from thequisqualate-responsive site, have negligible affinity for either the brainor neuroblastoma binding site. Lastly, agents which bind to other characterized receptors, suchas glutathione, 4-amino butyric acid, and carbamyl choline, have negligible affinity for the [3H]glutamate binding FIG.1. N1S-RE-105 neuroblastoma X retina hybrid cells in site on the neuroblastomahybrid membranes. Regulation of Glutamate Binding by Cations-To characterculture. N18-RE-105 cells weregrownunder conditions described under “Methods” and the photomicrograph was prepared using an ize the direct effects of cations on glutamate receptorsin Olympus IMT phase microscope fitted with an Olympus OM-2 cam- isolated membranes, N18-RE-105 cells were grown in normal era body at a magnification of x 33 (to the film). Bar = 100 pm. growthmedium and their membranes were harvested and split into aliquots and centrifuged. Control pellets were resusexamined, the N18-RE-105 clone consistently exhibited the pendedin buffer(Tris-HC1) alone while the experimental highest level of specific binding of ~-[~H]glutamate. The spe- pellets were resuspendedin the same buffer supplemented cific binding in the N18-RE-105 clone was comparable to that with various concentrations (10 pM-100mM)of mono- or of the cells with divalent cations. Control and experimental membrane susmeasured in rat brain membranes. Treatment dibutyryl cyclic AMP, often used to induce differentiation in pensions were preincubated for 10 mina t 37 “C and then were neuronal cell lines, did not augment specific binding by these assayed for [3H]glutamate binding without changing theion cells. concentrations. A significant increase in binding was meaAs shown in Fig. 1, the N18-RE-105 cell line exhibited a sured after treatment withcalcium ions (Fig. 4). Thebinding high degree of neuronal morphology: rounded soma withlong reached a maximum of 2.6-fold of control when 10 mM CaClz branched neurites containing many varicosities. Althoughthis was included but declined back toward control levels when line contains dense core vesicles2 and complex ganglioside^,^ membranes were exposed to higher concentrations. LaC13 its neurotransmitter characteristics have not been defined. produced a similar pattern of stimulation, reaching a maxiAssays for the neurotransmitter synthesizing enzymes tyro- mum binding of 3.0-fold of control binding a t 1 mM (Fig. 4). Dose-response curves for monovalentcationsandMgS04 * M.Nirenberg,Laboratoryof Biochemical Genetics, National Heart,Lung and Blood Institute, National Institutes of Health, revealed no enhancementof [3H]glutamatebinding at concentrations up to 150 mM. Incontrast,thesesalts causeda personal communication. decrease in [3H]glutamatebinding, similar to resultsreported R. Schnaar, unpublished data.

f 3HJGlutamatebinding

~

~

Characterization of a Glutamate Receptor on Neuroblastoma Cells

12759

1 I

FIG. 2. Association isotherm for the binding of ['Hlglutamate to N18-RE-105neuroblastoma membranes. Washed membranes, prepared from confluent cell cultures grown in 75cm2tissue culture flasks, were incubated with varying concentrations of [3H]glutamate at 37 "C and bound radiolabel was measured using the centrifugation assay described under "Methods." Nonspecific binding was measured in the presence of 100 PM unlabeled glutamate. Each point represents the average of six determinations. The inset presentsa Scatchard plot of the data from the saturation isotherms. The KD is expressed as a nanomolar concentration and the Bmaxas picomoles/mg of protein.

c

SPECIFIC

BOUND (pmol/mg protein)

200

400

600

800

1000

[GLUTAMATE](nM)

TABLEI1 Affinities of amino acid agonists and antagonists for the ['H]glutamic acid binding site on isolated N18-RE-105 and rat brain membranes Washed membranes were prepared from rat cerebellum (7) or N18RE-105 cells and incubated with 100 nM [3H]glutamate in the presence of 10 nM to 1 mM concentrations of the substances listed. ICw values were extrapolated from the logit-log plots of the resulting data and inhibition constants were calculated using the following formula: KI = ICm/(l + f/KD); where f = the free ligand concentration (100 nM) and KD = 654 nM (see Fig. 2). KI Analog

N18-RE-105

Linear agonists 0.8 L-Glutamic acid 27.8 3.4 L-Aspartic acid 7.1 10.9 Cysteine sulfonic acid \ O ' O 28.6 66.4 D-Aspartic acid 0 I I I I I 83.0 D-Glutamic acid -7 -6 >100.0 L-Glutamine -4 -3 -5 -8 >100.0 N-Methyl-DL-aspartic acid Inhibitor Concentration (log MI Cyclic agonists FIG. 3. Competition curves for the displacement of [3H]glu0.6 Quisqualic acid tamate by excitatory amino acids. Washed membranes prepared 8.5 Ibotenic acid from confluent cell cultures grown in 75-cm2 tissue culture flasks >100.0 Kainic acid were incubated with 100 nM [3H]glutamate inthe presence of various Antagonists quisqualate (W), DL-a-Aminoadipicacid concentrations of L-glutamate (M), 3.7 amino phosphonobutyrate (M), N-methyl-DL-aspartate 3.9 L-a-Aminoadipic acid (U and ) kainate , (A-A). Specific glutamate binding is de4.7 D-a-Aminoadipic acid termined as described under "Methods." Each data point is the mean 8.6 Aminophosphonobutyric acid of triplicate determinations from four separate experiments. 9.9 L-Aminopimelicacid 13.6 Aminophosphonoheptanoic acid 39.7 Glutamate-y-hydroxamate using hippocampal membranes (22). NaCl was about a 1059.2 AminoDhosDhonovalericacid fold more potent inhibitor than other monovalent cations, Aminophosphonohexanoic acid >100.0 >100.0 producing a 50% decrease in binding when added at a concenGlutamic acid diethyl ester >100.0 tration of 10 mM (Fig. 5). Other Scatchard analysis of the [3H]glutamate binding data colGlutathione >100.0 4-Aminobutyric acid B100.0 lected using N18-RE-105 membranes preincubated with 10 Carbamyl choline >100.0 mM CaC12 or 10 mM NaCl for 10 min at 37 revealed that

. .

"c

1.0

15.0

>.100.0 0.6 3.1 >-100.0 3.4 2.4 4.6 5.7 14.3

>100.0

Characterization of a GlutamateReceptor on NeuroblastomaCells I

I

I

I

I

60 n

0 I

LL

40

..

m

20

10

20

30

40

GLUTAMATE BOUND (prnol/rng Proleml

1

I

I

-5

-4

-3

1

I

-2 Collon Concentrollon (log M)

-I

FIG. 6. Scatchard analysis of ['Hlglutamate binding to N18RE-105 cell membranes preincubated with calcium chloride or sodium chloride. Washed membranes were prepared from confluent cell cultures and resuspended in Tris-HC1 buffer (control, 0, seeFig. 2) or the same buffer containingeither 10 mM calcium chloride (A)or 10 mM sodium chloride (M). Specific [3H]glutamate binding was determined as described under "Experimental Procedures" and the data were plotted by the method of Scatchard. The KDvalues are expressed as a nanomolar concentration and the B,, values as picomoles/mg of protein. Each data pointis the mean of triplicate determinations from three experiments.

FIG. 4. Effect of calcium chloride and lanthanum chloride preincubation on ['Hlglutamate binding to isolated N18-RE105 membranes. Washed membranes from N18-RE-105 cells were resuspended in Tris-HC1 buffer containing the indicated concentrations of calcium chloride (0)or lanthanum chloride (0)and preincubated for 10 min at 37 "C before being assayed for [3H]glutamate binding as described in the text. The data are expressed as theamount lular calcium concentrations. Trifluoperazine, aphenothiof glutamate binding compared to thatin Tris-HC1 buffer alone. Each azine widely used as a specific inhibitor of calmodulin-medidata point is the mean of triplicate determinations from each of four ated events (23), was therefore added to membrane homogeto eight separate experiments.

I00

1

-5

-4

-3

-2

-I

Callon Concentraiion (log M )

FIG. 5. Effects of monovalent cations and magnesium on ['Hlglutamate binding to isolated N18-RE-105 membranes. Washed membranes from N18-RE-105 cells were resuspended in Tris-HC1 buffer containing the indicated concentrations of sodium chloride (O),potassium chloride (A),lithium chloride (M), or magnesium sulfate (0) and preincubated for 10 min at 37 "C before assaying for [3H]glutamate binding. The data are expressed as the amount of glutamate binding compared to that in Tris-HC1 buffer is the mean of triplicate determinations from alone. Each data point each of four to eight separate experiments.

the respective increase and decrease in [3H]glutamatebinding was the result of a change in receptor number relative to control values (Fig. 6). The Bmaxvalues for control, 10 mM CaCi2-,and 10 mM NaC1-treated membranes were 15.5 pmol/ mg of protein, 46.9 pmol/mg of protein, and 6.97 pmol/mg of protein, respectively, while KO values were 654 nM, 523 nM and 435nM, respectively. These data reveal a remarkable range of regulation of the glutamate receptor concentration by mono- and divalent cations, perhaps indicative of a link between binding sites and ion channels (22). One possible explanation for the rapid calcium-induced increase inbinding observed in N18-RE-105 and hippocampal membranes is that theglutamate receptor can be activated by coupling with calmodulin in response to increasing intracel-

nates along with CaC12during the preincubation period. Concentrations of 0.01-10 1M trifluoperazine (which did not affect binding when added alone) were ineffective at blocking CaC12 stimulation of binding when the two compounds were added together (data not shown). Since the half-maximal dose for trifluoroperazine inhibition of calmodulin's action is usually around 1 PM, it appears unlikely that calmodulin mediates the CaC12effect on glutamate binding. the Regulation of Glutamate Binding by Anions-Allof binding experiments described above were performed in 50 mM Tris-HC1 buffer. In order to assess possible binding regulation by anions, the membranes were washed and resuspended in either Tris-HC1 (as above), 50 mM Tris acetate buffer, pH 7.1 (Tris-Ac), or in 50 mM Tris citrate, pH 7.1 (Tris-Cit). In each of the three buffer systems, glutamate binding was performed in the appropriate buffer alone (control) or after a 10-min preincubation at 37"C in the same buffer supplemented with either 1mM Ca(Ac)2,1 mMKC1, or 1 mM CaC12. Control binding in Tris-Ac or Tris-Cit buffers was only about one-tenth of that measured using Tris-HC1 buffer (Fig. 7). The addition of low concentrations of potassium chloride caused a marked increase in binding in the chloride-free buffer background. In Tris-Ac buffer, addition of calcium acetate also caused a sharp increase in glutamate binding, somewhat less than that caused by the addition of KC1. The addition of both calcium and chloride ions (CaC12) resulted in a synergistic effect leading to a 25-fold overall increase in glutamate binding. These data demonstrate that both calcium and chloride ions act to increase glutamate binding. In Tris-Cit buffer, calcium acetate did not increase glutamate binding, probably because of citrate's ability to chelate calcium. In this buffer, only the chloride effect was seen when CaClz was added. One mM CaCL (2 mM Cl-) was no more effective at increasing glutamate binding than 1mM KC1 (1 mM Cl-), suggesting that the chloride effect is saturated a t 1 mM concentration. The data provide further evidence that chloride ions can act in addition to calcium ions to up-regulate the glutamate receptor. DISCUSSION

Among several homogeneous neuronal cell lines tested, the neuroblastoma x rat embryonic retina cell line, N18-RE-105,

Characterization of Glutamate a Receptor

on Neuroblastoma Cells

12761

~

@ FIG. 7. Calcium and chloride regulation of [*H]glutamatebinding to N18-RE-106membranes. N18-RE105 cells were collected from confluent 75-cm2 flasks and sonicated, and membranes were collected by centrifugation in one of three buffers: Tris-HC1 (as in previous experiments, A ) , 50 mM Tris citrate ( B ) , or Tris acetate ( C ) . Membranes were preincubated for 10 min at 37 “C in the indicated buffer or in the same buffer supplemented with 1 mM concentrations of calcium acetate, potassium chlroide or calcium chloride. [3H] Glutamate binding was determined on all membrane fractions as described under “Methods.” Each point is the mean of triplicate determinations from three experiments. *p < 0.005; **p < 0.001 (compared to control in each panel).

:

TRIS-HCL

*

2

*

@

TRIS-CITRATE

TRIS-ACETATE

**

T

2.0 5

E,

7 2

5

*T

I .O

$ 3

None

CaAc,

KC1

CaCI,

None

CaAc,

KC1

CaCl,

None

CoAc,

KC1

CaCI,

Assay Buffer Additions

was shown to possess the highest concentration of specific membrane glutamate binding sites. Saturation isotherms reveal a single class of saturable sites having a & of 650 nM and a B,, of 16 pmol/mg of protein (Fig. 2), in excellent agreement with analysis of glutamate bindingto ratcerebellar membranes. The pharmacology of these glutamate binding sites is also remarkably similar to that in rat brain (Table 11). Most notably,quisqualate is the most potent inhibitor of [3H]glutamate binding to N18-RE-105 membranes closely followed by unlabeled glutamate itself (Fig. 3). N-Methyl-Daspartateandkainate showed negligible ability to inhibit binding of labeled glutamate to these neuroblastoma membranes (Fig. 3), supporting the hypothesis that they bind to sites which are pharmacologically distinct from the quisqualate/glutamate site defined here. In addition, the diaminodicarboxylic acid analogs such as D-a-aminoadipate are potent inhibitors of glutamate binding to both brain and N18-RE105 membranes, as are thephosphono acid derivatives, all of which show a rank order of potency which mirrors that reported in electrophysiologic studies in brain. These data demonstrate a remarkable kineticand pharmacologic similarity between the glutamate binding sites on N18-RE-105 cell membranes and thoseon ratbrain,constitutingthe first report of such excitatoryamino acid receptors on clonal neuronal cells in culture. Glutamate receptors on N18-RE-105 cell membranes are subject to cation regulation similar to thatreported for receptors on hippocampal membranes (14) and on forebrain synaptosomes (24). Membrane suspensions which are preincubated with calcium for 10 min at 37 “C prior to being added to the glutamate binding assaymixture show sharply increased ligand binding. This effect is maximized using 10 mM Ca2+ anddeclines at higher concentrations. Lanthanum ions, which are known to bind to calcium channels, produce a similar typeof binding stimulation which is maximal at 1 mM La3+. As with hippocampal and forebrain membranes, the calcium-induced increase in bindingwas shown, by Scatchard analysis, to be due to an increase in the number of receptors available to interact with free ligand. Although hippocampal membrane up-regulation is maximal at somewhat lower calcium concentrations (approximatelyl mM), the effect appears to be very similar to that on N18-RE-105 membranes. When chloride ions were eliminated from the basic binding assay buffer, a stimulation of glutamate binding by added

chloride ions was also revealed (Fig. 7). The stimulatory effect was even greater than thatproduced by calcium. As with brain membranes (13,24), calcium and chloride produce a synergistic increase inglutamate binding on neuroblastoma membranes. However, while calcium appears to require the presence of chloride ions to produce an effect on hippocampal and forebrain membranes, it is capable of stimulating glutamate binding to neuroblastoma membranes in the absence of chloride ions. Monovalent cations and magnesium inhibited [3H]glutamate binding on neuroblastoma membranes apparently by decreasing the B,,,. As reported for hippocampal membranes (13), NaCl is approximately 10-fold more potent than LiCl, KC1, or MgS04 at reducing glutamate binding. This finding also supports the pharmacologic data demonstrating that [3H]glutamate is binding to a neurotransmitter receptor and not anuptake site,since sodium has been shown to stimulate binding to uptake sites while it inhibits binding in our system. Furthermore, amino acid transport studies in the N18-RE-105 clone did not reveal the avid uptake of ~-[~H]glutamate observed inbrain synaptosomal preparations. Thus, cation regulation of neuroblastoma glutamate receptors is remarkably similar to the regulation of these receptors on hippocampal membranes. Baudry and Lynch (13) have speculated that theinhibition of glutamate binding by monovalent cations may reflect an association between the receptor and sodium channels. Electrophysiological studies have in fact shown that glutamate produces its neuroexcitatory action by opening sodium channels and it is entirely reasonable that some form of autoregulation exists to limit damage which might result from overstimulation. Since glutamate also stimulates the conductance of calcium ions, an increased intracellular calcium concentration may trigger the up-regulation of the glutamate receptor which, in turn,is responsible for producing LTP in hippocampal neurons. As discussed in the introduction, there is a strong correlation between the formation of LTP and the calcium regulation of this receptor. We have begun to probe the possible biochemical mechanisms underlying the described glutamate regulation. A calcium binding protein, calmodulin, regulates many calciummediated cellular events, and it seemed a likely candidate for involvement in the glutamate receptor regulatory mechanism. The use of trifluoperazine, a phenothiazine tranquilizer which has been shown to specifically inhibit the interaction of

12762

Characterization of Glutamate a

calmodulin with effector protein (23), was used to test this possibility. Glutamate binding to neuroblastoma membranes preincubated with CaClz plus various concentrations of TFP was not significantly different from membranes preincubated with CaCl, alone. These data argue against the involvement of calmodulin in the CaC1, stimulation of glutamate binding and further experiments will be necessary to clarify the biochemical mechanisms involved in this calcium-induced upregulation. In summary, the N18-RE-105 somatic cell hybrids (and the other hybrids reportedhere) are the first neuronaltissue culture cells shown to possess saturable bindlng sites for [3H]glutamate on their membranes. The binding sites have pharmacological selectivity for amino acid analogs remarkably similar to that of rat brain membranes. These sites on N18-RE-105 cells are also highly regulated by mono- and divalent cations and anions in a manner which mirrors the way these ions regulate glutamate receptors on hippocampal membranes. Since ion regulation of glutamate binding may underlie LTP, N18-RE-105 cells provide an in vitro model system for dissecting the molecular mechanisms of this important phenomenon. Ultimately, since LTP is thought to be a cellular physiological event involved in learning, the N18RE-105 line may provide a useful in vitro model system for dissecting the role of receptor regulation in higher brain function. Acknowledgments-We wish to thank Linda Anderson, Dr. David Trisler, and Dr. Marshall Nirenberg at the National Institutes of Health for providing the hybrid cell lines, Patti Swank-Hill for technical assistance, and Dorothea K. Stieff for manuscript preparation. REFERENCES 1. Roberts, P. J., Storm-Mathisen, J. and Johnston, G. A. R., eds (1981) Glutamate: Transmitter in the Central Nervous System,

Receptor on Neuroblastoma Cells John Wiley and Sons, Chichester, England 2. Di Chiara, G., and Gessa, G. L., eds (1981) Advances in Biochemical Psychopharmacology,Vol. 27, Raven Press, New York 3. Hayashi, T. (1954) Keio J. Med. 3 , 183-192 4. Watkins, J. C., and Evans, R. H. (1981) Annu. Reu. Pharmacol. 2 1,165-204 5. Michaelis, E. K., Michaelis, M. L., and Boyarsky, L. L. (1974) Biochim. Biophys. Acta367,338-348 6. Roberts, P. J. (1974) Nature (Lond.) 252,399-401 7. Slevin, J., Collins, J., Lindsley, K., and Coyle, J. T. (1982) Brain Res. 249,353-360 8. DeRobertis, E., and De Plazas, S. F. (1976) J. Neurochem. 2 6 , 1237-1243 9. Foster, A. C., and Roberts, P. J. (1978) J. Neurochem. 3 1 , 14671477 10. Lynch, G., and Baudry, M. (1984) Science 224,1057-1063 11. Baudry, M., and Lynch, G. (1982) Neurosci. Res. Prog. Bull. 20, 663-671 12. Baudry, M., Oliver, M., Creager, R., Wieraszko, A., and Lynch, G. (1980) Life Sei. 27,325-330 13. Baudry, M., and Lynch, G. (1980) Proc. Natl. Acud. Sci. U. S. A. 77,2298-2302 14. Dunwiddie, T. V., and Lynch, G. (1979) Brain Res. 169,103-110 15. Malouf, A. T., Coyle, J. T., and Schnaar, R.L. (1984) J. Biol. Chem. 2 5 9 , 12763-12768 16. Nelson, P., Christian, C., and Nirenberg, M. (1976) Proc. Natl. Acud. Sci. U. S.A. 73, 123-127 17. Coyle, J. T. (1972) Biochem. Pharmacol. 2 1 , 1935-1944 18. Bull, G., and Oderfeld-Nowak,B. (1971) J. Neurochem. 18,935941 19. Wilson, S. H., Schrier, B. K., Farber, J. L., Thompson, E. J., Rosenberg, R. N., Blume, A. J., and Nirenberg, M. W. (1972) J. Biol. Chem. 2 4 7 , 3159-3169 20. Coyle, J. T., and Enna, S. J. (1976) Brain Res. 1 1 1 , 119-133 21. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. Chem. 1 9 3 , 265-275 22. Baudry, M., and Lynch, G. (1979) Nature ( L o n d . ) 282,748-750 23. Weiss, B., and Levin, R. M. (1978) Adu. Cyclic Nucleotide Res. 9 , 285-303 24. Fagg, G. E., Foster, A. C., Mena, E. E., and Cotman, C. W. (1982) J. Neurosci. 2 , 958-965