A’euroscrmcr Vol. 35, No. 3, pp. 653-668, 1990 0306-4522/90 $3.00 + 0.00 Printed in Great Britain Pergamon Press plc ii‘ 1990 IBRO
CHARACTERIZATION AND REGIONAL DISTRIBUTION OF STRYCHNINE-INSENSITIVE [3H]GLYCINE BINDING SITES IN RAT BRAIN BY QUANTITATIVE RECEPTOR AUTORADIOGRAPHY
J. W. MCDONALD,* J. B. PENNEY,? M. V. JOHNSTON~ and A. B. YOUNG~$ *Neuroscience and Medical Scientist Training Program, TDepartment of Neurology, University of Michigan, Ann Arbor, MI 48104, U.S.A. iDepartments of Pediatrics and Neurology, Johns Hopkins University, Baltimore, MD, U.S.A.
Abstract-Recent evidence suggests that a strychnine-insensitive glycine modulatory site is associated with the N-methyl-D-aspartate receptor-channel complex. A quantitative autoradiographic method was used to characterize the pharmacological specificity and anatomical distribution of strychnine-insensitive [iH]glycine binding sites in rat brain. [‘H]Glycine binding was specific, saturable, reversible, pH and temperature-sensitive and of high affinity. [‘H]Glycine interacted with a single population of sites having a KD of approximately 200 nM and a maximum density of 6.2 pmol/mg protein (stratum radiatum, CAI). Binding exhibited a pharmacological profile similar to the physiologically defined strychnine-insensitive glycine modulatory site. Binding was stereoselective; the rank order of potency of simple amino acids as displacers of binding was: glycine > D-serine > D-alanine > L-serine > L-alanine > L-valine > D-Vahe. Binding was not altered by the inhibitory glycine receptor ligand, strychnine, by the glutamate agonists, quisqualate and kainate, or by GABA receptor selective ligands. Most competitive agonists or antagonists of the N-methyl-u-aspartate recognition site were ineffective displacers of glycine binding. The exceptions were the aminophosphono series of antagonists, D-alpha-aminoadipate, gamma-D-glutamylglycine and beta-D-aspartylaminomethylphosphonic acid. However, the inhibition of [‘Hlglycine binding produced by the aminophosphono compounds could be accounted for by the level of glycine contamination present in these compounds. The non-competitive NMDA receptorxhannel blockers, phencyclidine, its thienyl derivative, and MK-801 did not alter glycine binding. Kynurenate, glycine methylester, L-serine-O- sulfate, L-homocysteic acid, and several glycine-containing dipeptides were effective displacers of glycine binding. Structureeactivity relations of agonists and antagonists of the strychnine-insensitive glycine binding site are discussed. The distribution of strychnine-insensitive [‘Hlglycine binding was heterogeneous with the following rank order of binding densities: hippocampus > cerebral cortex z caudate-putamen > thalamus > cerebellum > brain stem. This distribution of binding was correlated with N-methyl-D-aspartate-sensitive [‘HIglutamate binding (r’ = 0.77; P < 0.001; Pearson product-moment) and [‘H]thienylcyclohexylpiperidine binding (rZ = 0.72; P < 0.001). These observations are consistent with the hypothesis that the strychnine-insensitive glycine binding site is closely associated with the N-methyl-D-aspartate receptor The N-methyl-D-aspartate (NMDA)-subtype of distinct sites for magnesium and zinc, and a channel glutamate receptors is part of a receptor-ion channel binding site (phencyclidine, PCP receptor) for dis- complex with multiple regulatory sites including sociative anesthetics and {( + )-S-methyl- lo,1 1-dihy- dro-5H-dibenzo[a,d]cycIohepten-5,10-imine maleate} (MK-801). 2’ 22.34The complex also has a strychnine- $To whom correspondence should be addressed. insensitive glycine binding site.23.27,58The NMDA Abbrmiutions: AP5, DL-2-amino-5-phosphonopentanoic acid; operated channel is gated by magnesium in a voltage- AP7, DL-2-amino-7-phosphonoheptanoic acid; APB, DL- 2-amino-4-phosphonobutyric acid; ASP-AMP, beta-D- dependent manner.32.36.45 Magnesium also alters aspartyl-aminomethylphosphonic acid; DGG, gamma- the binding characteristics of [‘Hltrichloroacetic acid D-glutamylglycine; GAMS, gamma-D-glutamylamino- ([‘HITCP) and [3H]MK-801 to the NMDA associated methylsulfonate; GDEE, L-glutamyl-diethylester; channel.8,52,M Zinc selectively antagonizes NMDA- HA-966, 3-amino-I-hydroxypyrollid-2-one; HPLC, high pressure liquid chromatography; LGG, gamma-L-glu- mediated responses at a site distinct from that tamylglycin& L-SOS, L-ser&e:O&lfate; ?$K-801, {(y )- of magnesium. 49.6’ Zinc also inhibits [3H]glutamate 5-methyl- IO,1 I-dihydro-SH-dibenzo[a,d]cyclohepten- binding to NMDA recognition sites and reduces 5,10-imine maleate}; NAAG, N-acetyl-L-aspartyl-L- [3H]MK-801 binding to the channel.““2 glutamic acid: NMDA, N-methyl-D-aspartate; PCP. PCP receptor ligands block NMDA responses phencyclidine; TCP, N-( I-[2-thienyl]cyclohexy1)3,4- piperidine; THA, tetrahydroamino acridine; TLC, thin in a non-competitive, activity-dependent manner layer chromatography. probably by binding within the ion channel.‘,62 653 654 .I W. M(.Do~IAI.I) ~a, t/i. Modulation of NMDA receptor activation regulates Michigan (Ann Arbor. MI. U.S.A.). All other compound\ [‘H]TCP and [3H]MK-801 binding.’ I’ However, were purchased from Sigma. reciprocal interactions have not been demon- strated.“‘.‘“,’ I)I.-AP~. IIL-AP7, kynurenate. glycine methyl ester, Johnson and Ascher” demonstrated, in outside- n-alpha-aminoadipic acid. L-homocysteic acid, ASP AMP. out patches, that glycine potentiates NMDA- and five glycine-containing dipeptides were tested for poss- induced increases in cation conductance in a ible glycinc contamination by thin layer chromatography (TLC) with acidic (n-butanol:acetic acid:water, I?: 3: 5) strychnine-insensitive manner. Following this and basic (ethanol: ammonia : water. 20: 1: 4) solvenl sys- observation. several reports have indicated that tems. Ninhydrin was used to visualize glycine. Additionally. glycine and structurally-related analogs also poten- I>I.-APS. rj~-AP7. kynurenate. NMDA and glutamate tiate NMDA-mediated events in other systems: were tested for glycine contamination using high pressure liquid chromatography (HPLC). Electrochemical detection cnhancemcnt of NMDA-stimulated electrophysio- of glycine was carried out using a modification of previously logical responses,J,“.2’ [‘H]TCP and [‘H]MK-801 described o-phthaldialdehyde derivitization.‘J;‘.“‘,’ Con- binding,’ ‘i ‘i hi and [‘Hlnorepinephrine release.“’ pounds were examined by HPLC at concentrations Furthermore. reciprocal interactions between the (100 nM I mM) that allowed detection of at least 0.001”‘0 NMDA recognition site and the glycinc modulatory glycine contamination. Multiple sources of compounds were tested when available. site have been described.’ “.“.Q Glycine can stim- ulatc agonist binding to the NMDA recognition site in membrane homogenates. Furthermore, NMDA Adult malt Sprague Dawley albino rats (three months receptor agonists stimulate [‘Hlglycine binding old) were killed by decapitation. The brains were quickly whereas antagonists inhibit [‘Hlglycine binding. removed. frozen on dry ice, and mounted on cryostat chucks. Mounted brains were allowed to equilibrate to the However, additional studies have demonstrated no tcmperaturc of the cryostat (- 12 to -20’C) and 20-/(M interaction between glycine and NMDA recognition horizontal sections were cut and thaw-mounted onto sile~,‘.‘.“l,l gelatin-coated slides. Sections were stored at -20 C for The anatomical localization of strychnine- less than 24 h. Preliminary experiments determined that [‘Hjglycine binding was not altered by frozen storage for insensitive [‘Hlglycine sites and the regional relation- 34h. ship between this site and NMDA recognition and PCP sites have not been described in detail. It is not clear whether these binding sites are always Tissue sections were preincubated in 50 mM Tris citrate associated or if they can exist independently of one (pH 7.4) at 4 C for 30 min to remove endogenous com- pounds and then dried under a stream of cool air. In another. preliminary experiments this length of preincubation was Although there have been previous reports determined to maximize [‘Hlglycine binding while maintain- of strychnine-insensitive glycine binding in rat mg tissue integrity. Longer preincubation periods of up to brain.“.Y_‘: “,~(, we now provide the first detailed 2 h at 4 C did not further enhance binding. In regional distribution studies. sections were then incubated in the characterization of the regional distribution and same buffer (pH 7.4) containing 1OOnM [‘Hlglycine for pharmacology of glycine binding sites in rat brain 35 min at 4 C (final volume = 8 ml); under these conditions, s&ions using quantitative autoradiography. The less than 5% of free ligand was bound and “zone A” topography of strychnine-insensitive glycinc binding conditions were maintained. In preliminary experiments. sites is compared with the distribution of NMDA- [‘Hlglycine binding in rat forebrain was found to be insen- sitive to the addition of 100pM strychnine. In saturation sensitive [‘H]glutamate, and [‘HITCP binding sites. studies. the concentration of [‘Hlglycine was 40 nM and Some of this work has been presented in preliminary varying amounts of unlabeled glycine were included in the form.“’ incubation buffer to give final concentrations of glycine ranging from 41 nM to 2.5 PM. Non-specific binding was determined in the presence of I mM unlabeled glycine and represented less than 10% of total binding at a concen- EXPERIMENTAL PROCEDURES tration of 100 nM [?H]glycine.” Following the incubation, sections were rinsed three time, Mtrtwial.\ with 2 ml ice-cold buffer followed by a final rinse with 2 ml [‘H]Glycine was obtained from Amersham (Arlington Ice-cold glutaraldehyde (49%): acetone mixture (1: 19 v/v) Heights, IL. U.S.A., specific activity = 13.3-19 Cijmmol). and quickly dried under a warm stream of air to minimize [IL-2-Amino-4-phosphonobutyric acid (APB). uL-2-amino- uneven ligand dissociation. Preliminary experiments deter- 5.phosphonopentanoic acid (AP5), uL-2-amino-7-phospho- mined that this rinse procedure optimized the ratio of ndheptanoic &d (AP7), gamma-o-glutamylamino-methyl- specific to non-specific binding. Maximum rinse time wa\ sulfonate (GAMS), gamma-u-glutamylglycine (DGG). less than 5 s. gamma-L-giutamylglyci~e (LGG), L-gl;ta&yl-diethylester Tissue sections were apposed to tritium-sensitive tilm (GDEE), beta-D-aspartyl-aminomethylphosphonic acid (LKB Ultrofilm ‘H) for two to six weeks. A set of radio- (ASPS-AMP), u-homocysteine sulfinic acid, 7-chlorokyn- active standards (American Radiochemicals, ARC c’l4) urenate, and 3-amino-1-hydroxypyrollid-2-one (HA-966) calibrated against whole brain pastes with known amounts were obtained from Tocris Neuramin (Essex, U.K.). of protein--tritium was exposed with each film. Protein N-Acetvl-L-aspartyl-L-glutamic. ..- acid (NAAG) and glycyl- values of whole brain paste sections were determined by the I.-glutamic acid were purchased from Bachem (Torrance, Lowry method. Quantitative analysis of the resulting auto- CA. U.S.A.). MK-801 was a rift from Dr P. Anderson radiograms was performed densitometrically using a micro- (Merck, Sharp & Dohme, West Point. PA. U.S.A.). TCP computer based video densitometer system (Imaging and PCP were gifts from Dr E. Domino. University of Research. St Catherines. Ontario). Optical density values Strychnine-insensitive glycine binding sites 655 were converted to pmol/mg protein using a computer gener- ated polynomial regression analysis which compared film densities produced by the tissue sections to those of radioac- tive standards. 47 Ten readings per area, bilaterally, were averaged in triplicate sections from at least three animals. Values are expressed relative to protein and represent the mean k S.E.M. unless otherwise stated. N-MethL’I-o-aspartate-sensitive [‘H]glutamate binding NMDA receptors were labeled with [3H]glutamate under conditions which select for binding to NMDA receptors.” Briefly, following a 30-min pre-rinse, sections were incu- bated with 200 nM [ZH]glutamate plus 2.5 PM quisqua- late in 50 mM Tris-acetate, pH 7.4, at 4°C. Sections were rinsed three times with 2ml of buffer and once with a Fig. 2. Time dependence and reversibility of strychnine-in- glutaraldehydeeacetone solution. sensitive [3H]glycine binding in stratum radiatum of the CA1 area in rat brain sections. Sections were incubated at 4°C [3H] Thienylcyclahexylpiperidine binding with 100 nM [3H]glycine for various times. At 60min, dissociation was initiated by “infinite dilution” (arrow). PCP receptors were labeled with 20 nM [)H]TCP in Data represent specific binding (binding displaceable by 50mM Trisacetate in the presence of I mM Mg2+ at 4°C I mM unlabeled glycine) and are expressed as for 45 min, following a 30-min pre-rinse.23 One milh- mean & S.E.M. (n = 3) in terms of pmol/mg protein. molar Mg’+ maximizes [jH]TCP binding in frozen rat brain sections.8 Using this method, binding equilibrium is achieved by approximately 45 min in the presence of Specific binding decreased faster at acidic rather than I mM Mg’+. Sections were rinsed 3 x 1min with ice-cold alkaline pH values; there was no specific binding at buffer. a pH value of 5. In contrast, at a pH value of 10 Data analysis binding was reduced by 30%. ~q,, values for competitors were calculated by logit-log The effects of three buffers on strychnine-insensi- analysis. K, values were determined by the method of Cheng tive glycine binding were evaluated: Triscitrate, and Prusoff’ using the formula K, = rc,/(l + ([L]/K,)). Tri-acetate and Tris-HCl (pH 7.4). Tris-base Glycine, NMDA and PCP receptor binding distributions were compared with the Pearson product-moment co- (50 mM) was used in each of the three buffer systems. efficient and linear regression analysis. Specific equilibrium binding in stratum radiatum of CA1 was greatest in Tris-citrate buffer and was, in RESULTS comparison, reduced by 11 f 5% in Tris-acetate buffer and by 19 + 4% in HCl buffer. The qualitative Effects of’ temperature, pH and incubation bufer on distribution of glycine binding under each of the strychnine-insensitive [‘Hlglycine binding buffer conditions was similar. Binding to the strychnine-insensitive glycine site in Kinetics of‘ strychnine-insensitive [3H]glycine binding hippocampal area CA1 was reduced by raising the temperature from 4 to 37°C. Relative to 4°C binding At a concentration of 100 nM [3H]glycine, binding at 23°C was reduced by 52 f 6%. Incubation at 37°C in the stratum radiatum of the CA1 area in rat brain reduced binding by 68 + 8%. reached equilibrium rapidly at 4°C and remained at Strychnine-insensitive glycine binding was very equilibrium for at least 2 h (Fig. 2). Dissociation of sensitive to changes in pH (Fig. 1). Specific binding glycine binding (by infinite dilution) was very rapid of [3H]glycine was studied at 10 pH values ranging with a rate constant of 0.76/min. The half-time for from 4.0 to 10.0. Binding in the stratum radiatum of receptor-ligand dissociation was approximately CA1 was optimal in the narrow pH range 7.0-7.5. 0.92 min. Since the dissociation of glycine was very rapid, the association rate constant was determined to be 2.05 x 106/M per min assuming second order kinetics. These rate constant values indicate a dissoci- ation constant (K,) value of 369 f 41 nM in the stratum radiatum of hippocampal area CAl. Saturation of strychnine-insensitive glycine binding Saturation experiments were performed over the range of 40nM-2.5pM glycine (Table 1; Fig. 3). Scatchard transformation of the saturation isotherm demonstrated maximal binding (B,,,) of 6.2 f 0.04 pmol/mg protein (stratum radiatum of CAl) to a single population of sites with a KD of 196 f 14 nM Effect of pH on strychnine-insensitive [‘Hlglycine (Fig. 3 inset). The Hill coefficient from this data was binding in stratum radiatum of the CA1 hippocampal subfield. Rat brain sections were incubated with 1OOnM 0.96 & 0.02 indicating no cooperativity between sites. [Hlglycine (4°C) in 50 mM Trisscitrate at the pH indicated. The discrepancy between K, determination by kinetic Data represent specific binding (mean k S.E.M., n = 3). experiments and the equilibrium KD value (Fig. 3) 656 .1. W. MCDO\AI 1) 1’1 (11 .Table I. Regional comparison of parameters of strychnine- 130, insensitive [‘Hlglycine bindmg . he&?+ ‘al- 0 zn2+ B 111.1, 110- Region K,, (nM) ( pmol:mg protein) Hill IOO- SR CAI 196 * IO 6.2 i 0.04 0.96 * 0.02 SK CA3 I98 * 73 4.6 i 0.4 0.97 & 0.05 90 SMDG l69i 17 4.3 * 0.2 I .o I * 0.03 80 Mcd-CPu 200 * IO 2.9 & 0.04 I .04 * 0.07 (‘ICX 250 * Ih 4.9 * 0.3 I .03 * 0.03 70 AV THAL 213 i 2X 3.6 + 0.3 0.96 + 0.01 60 Parameters of strychnine-insensitive [‘Hlglycine equilibrium -6~.~.7 -6 -4 -3 -2 -1 binding in six brain regions were determined by Log[lon].M Scatchard analysis of saturation isotherms. Rat brain sections were incubated with 40 nM [‘Hlglycine in the Fig. 4. EfTects of magnesium and zmc ions on strychnine- presence of I nM to 2.5,~M unlabeled glycine at 4 C insensitive [‘Hlglycine binding in stratum radiatum of the for 35 min as described in Experimental Procedures. CA I hippocampal area. Data are mean + S.E.M. (17 : 3). Values represent specific binding and are presented as mean + S.E.M. (II = 3). There were no significant differ- cnces between regional K,, values. SR. stratum radiatum: SMDG. stratum moleculare dentate gyrus; Med-CPU. medial caudate putamen: CICX. layers 1. II cingulate cortex: AV THAL, anterior tentral thalamus. All compounds were tested at a single concen- tration of 100 /J M (unless stated otherwise) for their probably reflects the ditfculty in accurately determin- ability to alter strychnine-insensitive [‘Hlglycinc ing the association rate constant due to fast dissocia- binding (100 nM [‘Hlglycine). Values represent the tion of glycine. Similar K, values were obtained in mean & S.E.M. of three animals and were taken from other brain regions (Table I ). stratum radiatum of hippocampal area CAI (Fig. 4 and Tables 2 and 3). The level of glycine contamination in commer- cial compounds as assessed by either thin layer Distinct binding sites on the NMDA rcccptor~ chromatography or high pressure liquid chromatog- channel complex have been described for the cations. raphy varied considerably between compounds. magnesium and zinc. The diff‘erential effects of these Less than 0. I % glycine contamination was present two ions on specific strychnine-insensitive glycine in the majority of compounds tested. However, binding is illustrated in Fig. 4. Magnesium enhanced I)L-AP5 (from two sources) contained approximately [‘Hlglycine binding over the concentration range, 1% glycine. NMDA and glutamate were also con- 0. I IO mM (P < 0.05, ANOVA). Binding was maxi- taminated with low levels of glycine (O.Ol-0.1%). mally increased to I20 + 7% of control at IO mM magnesium. Binding was slightly reduced at IO ii M magnesium. In contrast. zinc inhibited glycine bind- ing over the concentration range of 10 PM-- IO mM (P < 0.005, ANOVA). Ten millimolar zinc maximally reduced binding to 68 k 3% of the control value. Table 2. Competition for strychnine- insensitive (‘Hlglycine binding sites in stratum radiatum of CA1 by glycinc and structurally related amino acids Displacer K, (I’M) S.E.M. Glycine 0. I3 0.02 0.Serinc 0.i.s 0.02 I>-Alanine 0.53 0.14 I.-Serinc 16.37 3.21 Fig. 3. Representative saturation isotherm and Scatchard I -Alaninc 46.12 x.43 analysis (inset) of strychnine-insensitive glycine equilibrium L-Valinc 4.24 mM I .42 binding in stratum radiatum of the CAI hippocampdl rl-Valine > IO mM subtield. Rat brain sections were incubated with 40nM [‘Hlglycine in the presence of I nM to 2.5 /cM unlabeled Values represent mean t S.E.M. glycine at 4 C for 35 min as described in Experimental 01 = 3). Assays were performed Procedures. Each point represents specific binding (total using 100 nM [‘Hlglycine as binding minus binding present in adjacent sections incu- described in Experimental Pro- bated in the presence of I mM unlabeled glycine). The cedures. K, values were determined experiment was repeated six times in six animals with similar assuming a K,, of I98 nM for results. Averaged data represent mean + S.E.M. glycine. Strychnine-insensitive glycine binding sites 657 Table 3. Pharmacological profile of strychnine-insensitive [lH]glycine binding Percentage inhibition or stimulation of strychnine-insensitive [-‘H]glycine binding Compounds tested (100 PM) Mean S.E.M. *P-value Inhibitory glycine receptor hgands beta-alanine -44.43 6.15 0.001 Non-active compounds Strychnine NMDA receptor antagonists UL-APV -61.96 4.67 0.001 DL-APH -35.04 2.22 0.002 THA 43.44 8.49 0.038 Non-active compounds CPP PCP TCP MK-801 Dextromethorphan Glutamate receptor antagonists Kynurenate -83.07 5.1 0.00 I uL-2-Amino-4-phosphonobutyric acid (APB) - 26.48 6.2 0.014 Non-active compounds gamma-o-Glutamylaminomethylsulfonate (GAMS) L-Glutamyl-diethyl ester (GDEE) Glutamate receptor agonists Non-active compounds L-Glutamic acid D-Aspartic acid L-Aspartic acid NMDA lbotenic acid Quinolinic acid Quisqualic acid Kainic acid Amino acid-containing compounds u-alpha-Aminoadipic acid - 51.33 4.41 0.007 Glycine methyl ester -100 1.29 0.001 L-Homocysteic acid -45.34 11.96 0.031 L-Serine-O-sulfate (L-SOS) -63.07 8.12 0.003 0-Phospho-oL-tyrosine -45.58 24.44 0.03 beta-u-Aspartyl-aminomethylphosphonic acid (ASP-AMP) -60.37 8.49 0.005 gamma-o-Ghttamylglycine -80.21 6.41 0.001 gamma-L-Glutamylglycine -99.07 3.59 0.001 gamma-L-Aspartylglycine -93.97 3.5 0.001 Glycylglycine -62.07 4.6 0.001 Non-active compounds o-Asparagine L-Asparagine o-Aspartic acid L-Aspartic acid L-Arginine L-Cysteic acid L-Cysteine L-Cysteine sulfonic acid L-Glutamic acid Homocarnosine sulfate o-Homocysteine sulfinic acid o-Homocysteic acid o-Phenylalanine L-Phenylalanine L-Proline Taurine N-Acetyl-L-aspartic acid N-Acetyl-L-glutamic acid N-Acetyl-L-aspartyl-L-glutamic acid (NAAG) alpha-L-Glutamyl-L-glutamic acid Glycyl-L-glutamic acid Continued 658 J. W. MCDOYAI.I)(‘I t/l. Percentage inhibiti or stimu~ltion of strychnine-insensitive [‘H]giycine binding Compounds tested ( IO0p M) Mea11 S.E.M. *P -VdlK GABA receptor ligands Non-active compounds GABA delta-.~mino-.~\;-villeric acid ‘Y-Benzovl-d-amino valeric acid t)L-2.4-Diaminobutyric acid Data represent mean _+ S.E.M. (II = 3) and are expressed as percentage stimulation or inhibition (-) of binding in the presence of displacer relative to the amount of binding in the absence of displacer. All readings were taken from stratum radiatum of the CA 1 hippocampa~ area. Binding assays were performed as described in Experimental Procedures using 100 nM t’H]gfycine. All displacers were present at single lOO/lM concentrations. Statistical comparisons were made by one-way ANOVA, binding in presence of displacer vs total binding (non-active compounds produced non-significant alterations in [“Hlglycine binding. i.c. P z 0.05). See Refs 24. 37. 46. 60 for review of the pharmacology of some 01 these compounds. The potencies of glycine and both the u- and Strychnine, an antagonist at the inhibitory glycine L-isomers of serine, alanine and valine are presented receptor, did not alter glycine binding at 100 HIM. in Table 2. Of these amino acids, giycine was the most In contrast, beta-alanine significantly reduced strych- potent inhibitor of strychnine-insensitive glycine njne-insensitive gtycine binding by 44 I): 6% at a binding with a K, value of 0.13 k 0.02 I’M. In general concentration of 100 FM (P < 0.001 vs total binding, the D-amino acid isomers were more potent than the ANOVA). corresponding L-forms. The rank order of potency in displacing glycine binding was: glycine > D-serine > n-aianinc > L-serine > I.-alanine > I.-valine > Of the ligands that bind to the NMDA rcccptor- o-vaiine. All these amino acids produced complete operated ion channel, only tetrahydroam~noacridine inhibition of glycine binding with the exceptions (THA) significantly affected strychnine-insensitive of u- and L-valine which were only evaluated up glycine binding; 100 PM THA enhanced glycine to concentrations of IOmM. All Hill coefficients binding by 43%. In contrast, similar concentrations were near unity. The potencies of these ammo acids of the dissociative anesthetics PCP and its thienyl as displacers of [‘H]glycine binding in our assay derivative, TCP. did not alter glycine binding. An wcrc highly corrclatcd with their ability to inhibit effect of MK-801 was also absent, The selective, high strychnine-insensitive [‘Hlglycine binding in homo- affinity, competitive NMDA antagonist, CPP. did genate studies’X.Fh.“’ (r2 > 0.82; P < 0.01: Pearson not alter binding. In contrast, the aminophosphono product-moment). enhance NMDA-stimulated series of compounds significantly reduced glycine tXH]TCPSh (t.? = 0.98; P < 0.01) and [‘H]MK-~01~~ binding. The rank order of effectiveness of these (I.’ = 0.92; P < 0.01) binding in membranes, and compounds in inhibiting glycine binding, at a single stimulate NMDA-mediated [‘Hlnorepinephrine re- lease in whole brain sections”’ (r’ = 0.83; P < 0.01). 120 However. much higher concentrations (eight to 40 2 100 times greater) were necessary to enhance NMDA- 0s mediated [~H]norepinephrine release.“’ = 80 2 m2 60 [u g.S 40 c? HA-966 and 7-chiorokynurenate. two competi- rt 20 m tive antagonists of the strychnine-insensitive glycine 9 o 0 site.‘“.‘h inhibited ]‘H]glycine binding in a dose- 8 7 -6 -5 -4 -3 -2 dependent manner (Fig. 5). In stratum radiatum of Log[Competitor].M CA I. 7-chlorokynurenate was 41 times more potent than HA-966 in inhibiting [‘Hfglycine binding; K, Fig. 5. Dispiacement of strychnine-insensitive [rH]glycine values were 0.27 + 0.03 !tM (7-chlorokynure~te) binding by two competitive antagonists. 7-chlorokynurenate and I I. 11 i 0.89 ,UM (HA-966). Hill coefficients were and HA-966. Sections were incubated with IOOnM [‘Hlglycine at 4 C in the presence of various concentrations near unity indicating that these ligands are acting at of competitors. Measurements were made in stratum radia- ;L single population of glycine sites under the present turn of hippocampal area CAI. Data are mean & S.E.M. conditions. (If = 4). Strychnine-insensitive glycine binding sites 659 100 p M concentration, was DL-APS > DL-API > in cellular layers. In general, binding decreased DL-AP4 (-62 & 5%, -35 + 2%, -26 * 6%; vs total at caudal levels of the neuroaxis. Within the tel- binding, P < 0.05, ANOVA). Additionally, the encephalon, the rank order of binding density was amino acid-containing NMDA antagonists, D-alpha hippocampal formation > cerebral cortex > olfactory aminoadipate, ASP-AMP, and gamma-D-glutamyl- nuclei > basal ganglia > diencephalic nuclei. Mid- glycine all produced greater than 50% inhibition brain structures exhibited the least amount of binding of glycine binding. Of these compounds, gamma-D- and the cerebellum was intermediate. glutamylglycine was the most effective displacer of Within the olfactory bulb, a distinct laminar distri- binding (binding was reduced by 80%). The L-isomer bution of glycine binding was present (Fig. 6b); the of glutamylglycine completely displaced glycine greatest density of binding was in the external plexi- binding. The non-selective glutamate receptor antag- form layer (43% of maximum), intermediate levels in onist, kynurenate, significantly antagonized glycine the internal granule layer, and lowest density in the binding, displacing 83% of binding at a 100 LAM glomerular layer. In addition to the high binding concentration (P < 0.001, ANOVA). Two additional density in the olfactory bulb, the medial and lateral antagonists with selectivity favoring non-NMDA olfactory nuclei also exhibited a high degree of receptors, CAMS and GDEE, did not alter binding. binding (75% of the maximum density of glycine sites). lT&ct of’ glutamate receptor agonists Cerebral cortical regions demonstrated moderate Neither L-glutamate nor D- and L-aspartate to high levels of binding. The density of binding affected strychnine-insensitive [‘Hlglycine binding varied widely within cortical layers with greatest significantly although trends suggested that the variation between different cortical regions (up to D-form of aspartate, in comparison to the L-isomer, two-fold). Generally, binding density decreased from produced a greater reduction in binding. None of superficial to deep cortical layers with the homo- the prototypic selective glutamate receptor-subtype geneous distribution pattern present in the entorhinal agonists altered glycine binding. However, the cortex being the exception. Of the cortical regions, the NMDA agonists, L-homocysteic acid and L-serine-O- cingulate had the highest binding density and the sulfate, both significantly reduced binding by 45 and frontal and frontoparietal regions exhibited lower 63%. respectively (P < 0.03, ANOVA). levels of glycine binding. The basal ganglia also revealed a heterogeneous l?flkct of amino acid-containing compounds pattern of binding. Binding was greatest in the cau- With the exception of the simple amino acids dateeputamen, moderate in nucleus accumbens, and described above, none of the other amino acids tested low in globus pallidus. Very little binding was present altered glycine binding. Of note, taurine, which has in the entopeduncular nucleus. Marked variation in been demonstrated to inhibit high affinity glycine binding was also observed within these structures. In uptake in the rat forebrain, does not displace strych- the corpus striatum. 69% of maximal CAI binding nine-insensitive [‘Hlglycine binding. The methyl ester was present in the ventral posterior region, 55% derivative of glycine completely inhibited glycine ventroanteriorally, 44% dorsomedially, and 38% binding. dorsolaterally. In addition, binding densities in the In addition to the amino acid-containing com- posterior globus pallidus were nearly three times that pounds above, several dipeptides also significantly present more anteriorally. The lateral septum exhib- reduced binding. Glutamyl- and aspartyl-glycine ited 43% of maximal CA1 binding while less binding both reduced binding by greater than 80% while was present in the other basal forebrain structures. N-glycylglycine only reduced binding by 62% In general, the diencephalic nuclei had moderate (P < 0.001. ANOVA). levels of [‘Hlglycine binding. Various nuclei were readily distinguished by the amount of glycine bound. Efltict of GABA receptor ligands The medial dorsal nucleus had the greatest density of None of the GABA derivatives tested significantly binding, and the habenula contained the fewest bind- altered strychnine-insensitive glycine binding at con- ing sites. The anterior ventral, ventral lateral and centrations of 100 PM. ventral posterior medial nuclei possessed equivalent levels of binding (46% maximum). The medial and Regional distribution qf‘ strychnine -insensitice glycine lateral geniculate were readily distinguishable with binding sites slightly higher binding in the former. Binding in the Strychnine-insensitive glycine binding to rat brain nucleus reunions was only 35% of maximum. sections exhibited a marked regional heterogeneity, The hippocampal formation exhibited a very with the highest binding density present in the distinctive laminar distribution of glycine binding stratum radiatum of the CA1 hippocampal subfield (Fig. 6a) that was similar to the topography of (Table 4 and Fig. 6). Essentially all brain regions NMDA-sensitive glutamate binding (Fig. 7).‘xJ’JZ showed strychnine-insensitive glycine binding. Within The maximal density of binding sites observed any- laminated structures, dendritic zones generally where in the brain was found in the stratum radiatum had the greatest density of binding with less binding of area CAI. Within the hippocampus, levels were 660 .I. W Mc Do\\I I) (21 Table 4. Regronal distribution of strychnine-lnseclsitive [‘Hlglycine binding sltcs 1n rat braln [‘H]Glycine bound (fmol:mp protein) Relative to Mean stratum radiatum Brain region Abbrcwations (II = 4) S.E.M. of CAI (‘56) Olfactory region Glomerular layer GL 371 I 6 I5 External plexiform layer EPL IO96 43 33 Internal granule layer IGL 603 1x ‘4 Medial anterior olfactory nucleus Med AON 188X 7’) 74 Lateral anterior olfactory nucleus Lat AON I901 X6 75 Primary olfactory cortex PO< I910 X7 75 Cortex Entorhinal. layers I and II I .Z ENT 95 I I42 37 Entorhinal, layer IV 4 ENT IO49 I31 41 Entorhinal. layers V and VI 5.6 ENT 1020 95 40 Frontoparietal. layers I and II I .2 FrP;i I550 IO0 61 Frontoparietal. layer IV 4 FrPa I296 95 51 Frontoparietal. layers V and VI 5.6 FrPa IOXX X4 43 Frontal. layers 1 and II I.2 FRCX I506 I67 59 Frontal. layer IV 4 FRCX 13.51 96 53 Frontal. layers V and VI 5.6 FRCX 955 56 JX Cingulate. layers 1 and II I.2 CICX 1910 1% 76 C‘ingulate. layers V and VI 5.6 CICX I.566 I30 h? Basal ganglia Caudate~ putamcn. medial Med-CPu ll2l 75 44 Caudate putamen. lateral Lat-CPu 955 97 3X Caudate~ putamen, anterior Ant-CPU I396 52 55 Caudatc putamen, posteriol Post-CPU 174X 70 60 Globus pallidus. anterio! An-GP 285 7 I I Globus pallidus. posterior Post-GP 686 I? 27 Accumbens nucleus Acb I 0x5 66 43 Entopeduncular nucleus EPN 126 IS Basal forebrain Ventral limb diagonal band VLDB 417 25 I 6 Bed nucleus stria termtnahs BNST 617 47 24 Lateral septum Lat-SEP I I63 I25 40 Diencephalic nuclc~ Medial dorsal MD-THAL 1306 91 55 Anterior ventral AV-THAL I IX3 XI 47 Ventral lateral VL-THAL 115.5 74 46 Ventral posterior medial VPM-THAL II72 62 46 Habenula l{B 99 55 Lateral gcniculate LG I I64 71 46 Medial geniculate MG 1312 91 57 Reunions RE XX7 76 35 Hippocampal formatton CA I. stratum lacunosum moleculare SLM-CA I I791 133 71 CA I. stratum radiatum SR-CA I ‘536 I34 IO0 c‘4 I. stratum pyramidale SP-CA I 163.5 141 64 CA I 1 stratum oriens SO-CA 1 2061 I IO Xl C‘A3. stratum radiatum SR-CA3 I707 IO0 67 CA3, stratum pyramidale SP-CA3 741 75 29 CA3. stratum oricns SO-CA3 1566 I00 62 CA4 CA4 796 x7 31 Dentate gyrus, stratum moleculare SMDG 7360 142 x9 Brainstem Central grak C‘G 60 I6 Inferior coliiculus I(’ 19 Substantia nigra SNR 37 Cerebellum Molecular Iaver Mol-Cb 54 C;rdllUk Iaye; Gr-C‘b X9 30 Values arc presented as the mean k S.E.M. (II = 4) and as a percentage relative to stratum radiatum of the CAI hippocampal subfield. [‘H]Glycine concentration was 100 nM. Autoradiography and quantitation of autoradiograms were performed as described in Experimental Procedures. greater in CAI than in CA3 and lowest in the hilus. oriens > stratum pyramidale. Eighty-nine per cent The rank order of binding densities within CAI of maximal binding was present in the stratum mol- and CA3 subfields was stratum radiatum > stratum cculare of the dentate gyrus. Strychnine-insensitive glycine binding sites 661 E c 8 a 0 f J. W. M~~Do\.\I I> <‘I tri. Fig. 7. Comparison of representative autoradiograms of strychnine-insensitive [JH]glycine binding (A) and NMDA-sensitive glutamate binding (B). Glycine binding was carried out using IOOnM [‘Hlglycine as described in Experimental Procedures. NMDA-sensitive [3H]glutamate binding was performed with 200 nM [3H]glutamate in the presence of 2.5 PM quisqualate as described in detail previously.” Qualitatively, the distribution of the two binding sites was nearly identical. Brainstem structures exhibited low levels of bind- Comparison of the distribution of binding sites com- ing with the exception of moderate levels present prising the N-methyl-D-uspnrtate receptor-c~hunnrl in the central gray, the inferior colliculus and in complex the substantia nigra. In the cerebellum, distinct laminar binding was observed. Density levels were A representative comparison of the distribution low in the molecular layer which is the Purkinje of strychnine-insensitive [3H]glycine binding sites and cell dendritic zone. Binding in the granule cell layer NMDA-sensitive [‘HIglutamate binding sites in rat of the cerebellum was 30% of maximal binding brain is illustrated in Fig. 7. Overall, the regional density (seven times greater than the molecular distribution of the two sites was qualitatively very layer). similar. Linear regression analysis comparing the Strychnine-insensitive glycine binding sites 663 distribution of strychnine-insensitive [3H]glycine tally defined site. 23~19The equilibrium dissociation binding, NMDA-sensitive [3H]glutamate binding, constant (&,) derived in this study corresponds well and [3H]TCP binding in 30 brain regions revealed with not only the lcsO value for glycine inhibition a marked degree of concordance using single con- of [3H]glycine binding in the present study, but also centration equilibrium binding values (r* = 0.77, with EC,, values for glycine enhancement of NMDA- P < 0.001, Pearson product-moment, [‘Hlglycine vs mediated [3H]TCP and [3H]MK-801 binding.5’,55 The [3H]glutamate; r’ = 0.73, P < 0.001, [3H]glycine vs KD value of 198 nM calculated from rat brain sections [‘HITCP). in the present study is nearly identical to values As we have previously reported,33 the distribution reported in membrane homogenate studies.6.‘8 of NMDA-sensitive [‘HIglutamate binding and Strychnine-insensitive glycine binding sites also [)H]TCP binding sites in rat brain is nearly identical exhibit a ligand specificity indicative of the physio- (u’ = 0.91, P < 0.001) with the exception of the cer- logically defined strychnine-insensitive glycine recep- ebellar granule layer. However, greater variation is tor. The rank order of potency of glycine and the observed between the distribution of these two bind- D- and L-isomers of serine, alanine and valine as ing sites and the distribution of strychnine-insensitive displacers of strychnine-insensitive [3H]glycine bind- glycine binding sites. In stratum radiatum of the CA1 ing in this study is highly correlated with the ability hippocampal subfield, the ratio of maximum binding of these ligands to stimulate NMDA-mediated densities obtained from saturation experiments is [‘Hlnorepinephrine release” and enhance [‘H]TCP approximately 1: 3.4 : 4, TCP : glycine : glutamate. and [‘H]MK-801 binding in biochemical studies.5h.63 Comparison of single point equilibrium binding A similar amino acid profile has been observed density ratios reveals a general pattern among brain for enhancement of NMDA-induced conductances regions. Glycine and glutamate binding ratios are and NMDA-stimulated calcium influx.“,5’ Further- presented relative to a TCP binding value of one. In more, the threshold concentration of glycine neces- contrast to the lower ratio of glycine to glutamate sary to displace [3H]glycine binding is very similar binding present in the hippocampal formation and to the glycine levels sufficient to enhance NMDA- cerebral cortical regions, the ratio of glycine to mediated currents in patch clamp studies.‘3,‘9 Except glutamate binding in most other brain regions is for glycine, serine, alanine and valine, none of the reversed: olfactory structures, 8.5: 7; basal ganglia, other simple amino acids were effective displacers 12: 6.6; basal forebrain structures, 10: 8.6; thalamic of [‘Hlglycine binding in agreement with patch nuclei, 12.6: 8.6; brainstem structures, 16:9. The clamp studies and with previous biochemical cerebellar granule layer exhibits the largest ratio of reports. 5.‘3.50.5’~63The ability of 7-chlorokynurenate glycine and glutamate to TCP binding (60: 50: I). The and HA-966 to inhibit [‘Hlglycine binding in glycine: glutamate binding ratio also varies within our assay is consistent with their ability to block structures. The ratio in stratum pyramidale of CA1 is NMDA-responses and limit NMDA-mediated brain 17: 8, compared with a ratio of 7.6: 8.2 in the same injury. “~‘6.4’ The ability of several dipeptides to layer of area CA3. inhibit [3H]glycine binding may be related to their ability to antagonize NMDA responses.60 Kainate- and quisqualate-type glutamate receptor and GABA DISCUSSION receptor selective ligands failed to displace specific strychnine-insensitive glycine binding in agreement Correspondence to physiologically dejned strychnine- with previous biochemical and electrophysiological insensitive glycine site studies.3,‘” In rat brain sections, [3H]glycine interacts with a It is unlikely that [3H]glycine binding represents single population of sites in a saturable and reversible an enzymatic or transport binding site*’ or simple manner. The linear Scatchard plots of saturation data sequestration since strychnine-insensitive glycine with corresponding Hill coefficients near unity and binding was maximal at 4°C and was greatly reduced the displacement studies support binding to a single at 23 or 37-C. In addition, the KD determined by receptor population. The inability of strychnine and kinetic experiments was similar to equilibrium K, the relative inability of beta-alanine and taurine to suggesting that sequestration is unlikely to play a role displace [3H]glycine binding in our assay distinguishes in binding. Binding was highest in areas where glycine these sites from the classical inhibitory glycine recep- uptake is lowest. Furthermore, inhibitors of high tor which is present in high concentrations in spinal affinity glycine uptake described in rat forebrain, cord. Furthermore, the kinetics, pharmacological such as taurine, are ineffective displacers of binding. selectivity, narrow pH dependency, and regional Taken together with the evidence described above, brain distribution of [3H]glycine binding suggest that the high degree of concordance between the distribu- the sites labeled in this study correspond to the tion of strychnine-insensitive glycine binding sites physiologically defined strychnine-insensitive glycine and the distributions of [3H]TCP and NMDA- sites. sensitive glutamate binding sites strongly suggest that Strychnine-insensitive glycine binding sites exhibit the site labelled by [3H]glycine in this study represents the appropriate kinetics expected for the physiologi- the physiologically defined receptor. brain regions the ratio of TCP and NMDA-sensitive Two distinct binding sites on the NMDA recep- glutamate binding densities is relatively constant.” 111 tor-channel complex have been described for the contrast, ratios of strychnine-insensitive glycine bind- prototypic cations magnesium and zinc. Magnesium ing to NMDA-sensitive glutamate binding and TCP gates the NMDA channel in a voltage-dependent binding arc much more variable between brain struc- mannerlX’!_‘h.~5 and has also been shown to both turcs. Using single concentration equilibrium binding enhance and inhibit [‘H]TCP and [‘H]MK-801 bind- values, TCP: glycine : glutamate binding ratios varied ing depending on the concentration of magnesium from I : 7: 9 in area CA I to I : I3 : 9 in basal ganglia and the state of channel activation.“,“.“’ Magnesium. and thalamus, and I : 16:‘) in the brainstem. The however, does not altcr NMDA-sensitive gluta- largest discrepancy occurred in the cerebellum mate binding.‘,“’ WC found that concentrations OI ( I : 60: 50). Since the K,s of these binding sites do magnesium from 0.1 to IOmM stimulate glycine not vary considerably between brain regions (see binding. This concentration range is consistent with Results and Table I). the single concentration equi- the concentrations needed to block NMDA associ- librium values reflect mainly alterations in the ated ion channels and alter [?H]TCP and [3H]MK-801 maximum number of binding sites and are unlikely binding. IO reflect different endogenous concentrations of Zinc also selectively antagonizes the NMDA reccp- D(llycinc among various brain regions. The absolute tor channel in electrophysiological. biochemical, and stoichiometric numbers arc only estimates of the ncurotoxicity studies but at a site which is distinct actual receptor stoichiometry since the single ligand from the magnesium binding site.‘i.““” “.” Monahan concentration : K,) ratios differed between rcceptol and MicheYJ have shown that I mM zinc antagon- binding assays [concentration : K,, ratios were: TCP izes NMDA-sensitive glutamate binding in mcm- ( I : I ). glutamate (I : I ). glycinc (1: 2)]. Howcvcr, the branc homogenates. The concentration range of zinc different ligand concentration : K, ratios do not that inhibits strychnine-insensitive glycine binding alter the basic conclusion that the receptor stoi- ( IO jcM~- IO mM) is in the range of concentrations that chiometry differs between brain regions. These data attenuate NMDA neuroexcitation (5 ~~M~~lmM) and suggest there may be heterogeneity to the NMDA NMDA-mediated neuronal injury (30 /[M-0.5 mM) receptor-channel complex as has been seen with In cortical and hippocampal cell cultures and reduce the GABA receptors-channel complexs7.“’ and/or that [‘H]MK-801 binding (I icM_I mM) to rat brain some glycine sites may not be associated with the membranes, i’):i%I,J’) ‘2 hi The antagonistic action of zinc NMDA receptor were all effective displacers of glycine binding as a single methyl group while all compounds with were several dipeptides (especially those containing aspartate length side chains are not active. Of com- glycine). However, the inhibition of [3H]glycine bind- pounds with glutamate-length side chains, the order ing by the amino phosphono compounds can be of activity with respect to the terminal acid group is accounted for by the level of glycine contamination SO,H > PO, H, > COOH. present in these compounds. HPLC analysis of The number of atoms separating the amino and glycine content indicated that APS (from two carboxyl terminals is also an important factor in sources; Tocris Neuramin, Sigma) contained approx- determining activity at the strychnine-insensitive imately 1% glycine. At IOOpM, DL-APS inhibited glycine receptor. In comparison to glycine, inter- [“Hlglycine binding by 62%. At this concentration terminal chain lengths greater than one carbon of AP5, 1 PM glycine contamination would be reduce (beta-alanine, two carbons) or eliminate present. Since the I&, of glycine is 200 nM, this level (GABA, three carbons) activity. This is also true of glycine contamination would be expected to pro- for activity of ligands at the NMDA recognition duce approximately 80% inhibition of [3H]glycine site.46,60 binding. The contamination could account for Substitutions of the amino and carboxyl terminals the level of [‘Hlglycine displacement produced by are also important determinants of activity. Replace- 100 p M AP5. AP4 and AP7 contained less than 0. I % ment of the carboxyl terminal of beta-alanine with a glycine and this level of glycine contamination would sulfonic group (taurine) eliminates activity. In con- be expected to inhibit [‘Hlglycine binding by less than trast, addition of a methylester moiety to the carboxyl 30% at 1OOpM AP4 or AP7 (accounting for the terminal of glycine (glycine methylester) maintains degree of inhibition of binding produced by these two pronounced activity. Methylation of the primary compounds). In contrast, the level of glycine present amine of glycine (N-methyl-glycine) eliminates activ- in the glycine containing dipeptides (Iess than 0.1%) ity at the glycine binding site.5’.56Of note, methyl- cannot explain the inhibition of PHjglycine produced ation of aspartate (N-methyl-aspartate) increases by these dipeptides since they all produced 60-100% activity at the NMDA receptor, whereas larger inhibition of binding at 100 PM concentrations. Most alkylations decrease activity.4h other effective displacers of 13H]glycine binding as Addition of heteroatoms in the carbon chain con- well as inactive compounds (NMDA, glutamate) necting the acidic groups appears to be well tolerated. contained low levels of glycine contamination which Replacement of the methylene group of homocysteic could not account for their effects on [jH]giycine acid with an oxygen atom (serine-O-sulfate) main- binding. tains activity. Of the dipeptides with glycine at the The problem of glycine contamination of commer- gamma-carboxy terminal the order of potency of the cial compounds, water supplies, and glassware may acidic terminal of glycine is COOH > PO,H, > complicate the interpretation of studies examining SO,H. Furthe~ore, dipeptides with glycine at the the function of excitatory amino acid receptor modu- N-terminal are less effective than those with glycine lators. In autoradiographic experiments, endogenous at the C-terminal and non-glycine containing dipep- glycine in the tissue sections is always an issue despite tides are not active at all. extensive washes. In some patch clamp or binding Several common structural features of reported studies using extensively washed membranes where antagonists of the strychnine-insensitive glycine site glycine contamination is thought to be minimal, the imply that comformationally restricted and hetero- levels of glycine in commercial materials becomes cyclic structures are important determinants of antag- problematic. The purity of compounds used in stud- onists of the glycine receptor. These include the ies of this modulator site should be considered in the cyclic anhydric analog of GABA, HA-966,‘4.h0cyclo- interpretation of the results. leucine,5Sa kynurenate, its analog 7-chlorokynuren- ate,lh and two heterocyclic quinoxaline derivatives Structural requirements of ~tr~~h~~ne-insensitive &cyano-7-nitroquinoxaline-2,3-dione (CNQX) and givcitw receptor ligands 6,7-dinitroquinoxaline-2.3-dione (DNQX).‘” The The structure-activity relations of glycine receptor common structural features of these glycine receptor ligands that can be determined from the inhibition antagonists include planar conformationat restriction studies of [‘H]glycine binding in Tables 1 and 2 and functional groups up to at least four atoms from confirm and extend results from previous studies the alpha-amino acid carbon. of structureeactivity relations.s6 The D-forms of Several other putative antagonists of the NMDA simple alpha-amino acids are in general more potent recognition site also inhibit strychnine-insensitive displacers than the L-forms (about 80 times) of glycine binding (gamma-D-glutamylglycine, ASP- strychnine-insensitive glycine binding. This stereo- AMP. D-alpha-aminoad~pate, kynurenatef. z4.4ft.h”In specificity is similar to the structural requirements fact, gamma-D-glutamylglycine is more effective of antagonists of the NMDA recognition site.46 A in inhibiting glycine binding than NMDA-sensitive hydrogen on the alpha carbon (glycine), a methylene- glutamate binding (unpublished observation). The hydroxyl group (serine), or a methylene-O-sulfate effects of these compounds on [‘Hlglycine binding group L-serine-O-sulfate (L-SOS) is more active than cannot be explained entirety by non-competitive, allosteric interations with NMDA recognition sites The data extend earlier descriptions of the pharma- &cause the selective NMDA antagonist, CPP, did cology of strychnine-insensitive [‘Hlglycine binding not reduce (‘Hlglycine binding. Furthermore. the and demonstrate that this labeled site corresponds to NMDA agonists. L-SOS and L-homocysteic acid. also the physiologically defined site. The results also indi- inhibited [‘Hlglycine binding. Glycinc contamination cate that the stoichiometry between the modulatory cannot account for the inhibition of[‘H]glycine bind- sites comprising the NMDA rcccptor complex var> ing produced by these compounds. The structural among brain regions. The preliminary studies of the features of the dipeptide-type NMDA antagonists structure--activity relations of’ strychnine-inscnsitivc indicated above arc consistent with the features ol glycine receptor ligands could be of considerable glycinc receptor ligands. particularly the antagonists. value in the development of selective agonists and It is the&ore likely that the structural requirements antagonists for the glycine modulatory site. Antag- for antagonists of both NMDA and glycine sites onists of the strychnine-insensitive glycine site may partially overlap. prove to be an effective treatment in inherited juvcnilc non-ketotic hyperglycinemia”’ as well as other ncuro- logic disorders.“,‘X,“‘J’ ” ” CON(‘I.lISION .~1c~Xllo~l~l~~tl~c~l?ll~tl~\ Supported by USPHS grant The results presented in this paper provide the first lPOINSl9613. J.W.M. is a recipient of a MSTP fellowship detailed information about the regional brain distri- (5 T33 6M07863-07). We thank S. M. Scholler and bution of strychnine-insensitive [jH]glycine binding. K. O’Mara for technical assistance. REFERENC‘ES I. AIM N. A.. Berry S. C.. Burton N. R. and Lodge D. (19X3) The dissociative anesthetics. ketamine and phencychdine. selectivelv reduce excitation of central mammalian neurons by N-methyl-u-aspartate. Br. J. Pharmac. 79, 565 575. 7. Assaf S. -Y. and Chung S. (1984) Release of endogenous Zn’ from brain tissue during activity. Nurure 388, 734 736. 3. Benavides J., Rivy J., Carter C. and Scatton B. (1988) Differential modulation of [‘H]TCP binding to the NMDA receptor by L-glutamate and glycine. E~rr. J. Phurnuc~. 149, 67 72. 4. Bertolino M., Vicini S.. Mazzetta J. and Costa E. (1988) Phcncyclidinc and glycine modulate NMDA-activated high conductance cationic channels by acting at different sites. /Vru,-oc,i. Le/r. 84, 351 355. 42 Birch P. J.. Grossman C. J. and Hayes A. G. (198X) 6.7.Dinitroquinoxaline-2.3.dion and 6.nitro. 7.cyano-quinoxalirlc- ?.3-dion antagonize responses to NMDA in the rat spinal cord via an action at the strychnine-insensitive glycinc receptor. Eur. .I. Phrrmtcrc 156, 177 180. 5. Bonhaus D. W.. Burg B. C. and McNamara J. 0. (1987) Biochemical evidence that glycine allosterically regulatca an NMDA receptor-coupled ion channel. Grr. J. Pharmuc. 142, 489-490. 52 Bonhaus D. W.. Ych G.-C., Skaryak L. and McNamara J. 0. (19X9) Glycine regulation of the ,Y-methyl-I,-aspartatc receptor-gated ion channel in hippocampal membranes. Mo/e(,. Pharnwc. 36, 273- 279. 6. Bristow D. R.. Bowery N. G. and Woodruff Cr. N. (1986) Light microscopic autoradiographic localization of [‘H]glycinc and [‘HIstrychnine binding sites in rat brain. Eur. J. Pharmw. 126, 303 307. 7. Cheng Y. C. and PrusolT W. H. (1973) Relationship between the inhibition constant (K,) and the concentratton of Inhibitor which causes 50 per cent inhibition (I,(,) of an enzymatic reaction. Bioc~hml. Phurnwc. 22, 3099 310X. x. Chu D. C. M., Maragos W. F.. Penney J. B. and Young A. B. (1988) Magnesium ions increase binding aflinity 01. dissociative anesthetic receptors but not of NMDA receptors in rat brain. In Frontier.\ ,,t Esc,itcrtor:,- .4mim ,4c,itl Rcwcrrd~ (eds Cavalheiro E. A., Lehmann J. and Turski L.). Vol. 46. pp. 567 570. 9. Cotman C. W.. Monaghan D. T.. Ottersen 0. P. and Storm-Mathisen J. (1987) Anatomical organization of cxcttatory amino acid receptors and their pathways. Trcn& Nruroc(~i. 10, 273 279. IO. Deneris E. S.. Connolly J.. Boulter J.. Wada E.. Wada K.. Swanson L. W.. Patrick J. and Heinemann S. (198X) Primary structure and expression of beta 2: a novel subunit of neuronal nicotinic acetylcholine receptors. Nrztrojz I. 45 54. I I. Fadda E., Danysr W.. Wroblewskt J. T. and Costa E. (1988) Glycinc and t)-serene mcreasc the alfimty of ,v’-methyl-t,-aspartatc sensitive glutamate binding sites in rat brain synaptic membranes. N~~uropharnloc~ol~~~~~ 27, II83 11x5. II. Fagg G. E. (1987) Phencyclidine and related drugs bind to the activated NMDA receptor channel complex in rat bram membranes, Nwrmci. Le//. 76, 221 227. I?. Flatman J. A., Schwindt P. C.. Grill W. E. and Afstrom S. 7‘. (19X3) Multiple actions of .t’-methyl-l,-aspartste on cat neocortical neurons irl ritr-o. Bruin Rec. 226, I69 173. 14. Fletcher E. J. and Lodge D. (1988) Glycine reverses antagomsm of N-methyl-t,-aspartate (NMDA) by I-hydroxy-i- aminonvrrolidone-2 (HA-966) but not by t)-2.amino-5-phosphonovalcrate (I,-AP~) on rat cortical slices Eur. ./. Phtrrn&. 151, I61 162. I5 Forsvthc I. D.. Westbrook G L. and Mavcr M. L. (1988) Modulatton of excitatory synaptic transmission by gtycinc and zinc m cultures of mouse hippocampal neurons. .I. Neuwsci. 8, 3733 3741. 16. Gerritsen T.. Kaveggia E. and Waisman H. A. (1981) A new type of idiopathic hyperglycmcmia vvith hype-oxaluria. Pdiutrics 36. 882 X9 I. 17. Goldman D.. Deneris E.. Luytcn W.. Kochhar A., Patrick J. and Heinemann S. (1987) Members of a mcottntc acetylcholine receptor gent family arc expressed in different regions of the mammalian contra1 nervous system. (‘cl/ 48, 965 973. IX. Greenamyre J. T, Olson J. M. M.. Penney J. B. and Young A. B. (1985) Autoradiographtc characterization of .V-methyl-I,-aspartatc-. quisqualatc- and kainate-sensitivac glutamate binding sites. .I. Phcmnuc VY~. Ther. 233, 254 263. Strychnine-insensitive glycine binding sites 667 19. Greenamyre J. T. (1986) The role of glutamate in neurotransmission and in neurologic disease. Arch. Neurof. 43, 1058-1063. l9a. Greenberg D. A. and Marks S. S. (1988) Cation interactions with putative NMDA receptor-gated channels labeled by [‘HIMK-801 in rat cerebral cortex. Neurosci. Lelt. 95, 236-240. 20. Hannuniemi R. and Dja S. S. (1981) Uptake of leucine, lysine. aspartic acid. and glycine into isolated neurons and astrocytes. Neurochem. Res. 8, 873-884. 21. Huettner J. E. and Bean B. P. (1988) Block of N-methyl-D-aspartate-activated current by the anticonvulsant MK-801: selective binding to open channels. Proc. natn. Acuti. Sci. U.S.A. 85, 13077131 I. 22. Javitt D. C., Jotkowitz A., Sircar R. and Zukin S. R. (1987) Non~competitive regulation of phencyclidine~sigma-r~ep- iors by the N-methyl-p-aspartate receptor antagonist D-(-)-2-amino-5-phosphonovaleric acid. Neurosci. L&l. 78, 1933198. 23. Johnson J. W. and Ascher P. (1987) Glycine potentiates the NMDA response in cultured mouse brain neurons. Nufure 325, 529953 I. 24. Jones A. W., Smith D. A. S. and Watkins J. C. (3984) Stucture-activity relations of dipeptide antagonists of excitatory amino acids. ~~~rQ.~e~e?z~e 13, 573-58 I. 24a. Joseph M. and Davies P. (1983) Electrochemical detection of ~~-phthaldia~dehyde beta-mercaptoethanol derivatives of amino acids. J. Chromat. 277, 125-136. 25. Kemp J. A., Priestly T. and Woodruff G. N. (1986) MK-801, a novel, orally active anticonvulsant is a potent, non-competitive N-methyl-D-aspartate receptor antagonist. Br. J. Pharmac. Proc. (Suppl.) 89, 535. 26. Kemp J. A.. Foster A. C., Leeson P. D., Priestly T.. Tridgett R. and Iversen L. L. (1988) 7-Chlorokynurenic acid is a selective antagonist at the glycine nlodu~tory site of the ~-methyl-~-aspartate receptor complex. Proc. nam. Acad. Sci. U.S.A. 85, 654776550. 27. Kessler M., Terramani T., Lynch G. and Baudry M. (1989) A glycine site associated with NMDA receptors: characterization and identification of a new class of antagonists. /. Neurochem. 52, 1319-1328. 28. Kishimoto H., Simon J. R. and Aprison M. H. (1981) Determination of the equilibrium dissociation constants and number of glycinc binding sites in several areas of the rat central nervous system, using a sodium-independent system. f. Nrurochrm. 37, IOIS-1024. 29. Kleckner N. W. and Dingledine R. (1988) Requirement for glycine inacti~~ation of NMDA receptors expressed in Xcnopus oocytes. S&me 241, 835837. 30. Koh J.-Y. and Choi D. W. (1988) Zinc alters excitatory amino acid neurotoxicity on cortical neurons. J. Neurosci. 8, 2164-2171. 30a. Lindroth P. and Mopper K. (1979) HPLC determination of sub-picomolar amounts of amino acids by pre-column fluorescence derivatization with o-phthaldialdehyde. Ana&. Chem. 51, 166771674. 31. Loo P., Braunwalder A., Lehmann J. and Williams M. (1986) Radiohgand binding to central phencyclidine recognition sites is dependent on excitatory amino acid receptor agonists. J. ~~urrn~~. 123, 467.-468. 32. MacDonald J. F., Porietis A. V. and Wojtowicz J. M. (1982) L-Glutamic acid induces a region of negative slope conductance in the current voltage relationship of cultured spinal cord neurons. Bruin Res. 237, 248.--253. 33. Maragos W. F., Penney J. B. and Young A. B. (1988) Anatomic correlation of NMDA and ‘H-TCP-labeled receptors in rat brain. J. Neumsci. 8, 493.--501. 34. Martin D. and Lodge D. (1985) Ketamine acts as a non-competitive N-methyl+aspartate antagonist on frog spinal cord in ritro. ~~rupjlffrin~~ulog~ 24, 99991003. 35. Marvizon J. C. G. and Skolnick P. (1988) [‘H]Glycine binding is modulated by Mg?+ and other ligands of the NMDA receptor-cation channel complex. Eur. J. Pharmac. 151, 1577158. 36. Mayer M. L., Westbrook G. L. and Guthrie P. B. (1984) Voltage-dependent block by Mg’+ of NMDA responses in spinal cord neurons. Nature 309, 261-263. 37. Mayer M. L. and Westbrook G. L. (1987)The physiology of excitatory amino acids in the vertebrate central nervous system. Prog. New&d. 28, 1977276. 38. McDonald J. W., Silverstein F. S. and Johnston M. V. (1987) MK-801 protects the neonatal brain from hypoxic- ischemic damage. Eur. J. Pharmuc. 140, 3599361. 39. McDonald J. W., Penney J. B., Johnston M. V. and Young A. B. (1988) Quantitative autoradiography of [‘Hlglycine binding to the glycine receptor associated with the NMDA receptor operated channel. Sot. Neurosci. Ahsrr. 14, 484. 40. McDonald J. W., Silverstein F. S. and Johnston M. V. (1989) Neuroprotective effects of MK-801, TCP, PCP, and CPP against N-methyl-p-aspartate induced neurotoxicity in an in ciro perinatal rat model. Bruin Res. 490, 33-40. 41. McDonald J. W., Uckele J., Silverstein F. S. and Johnston M. V. (1989) HA-966 (I-bydroxy-3-aminopyrrolidone-2) selectively reduces ~-methyl-~-aspartate (NMDA)-mediate brain injury. ~e~~#~cj. Lptf. 104, 167-170. 42. Monaghan D. T. and Cotman C. W. (1985) Distribution of N-methyl-o-aspartate-sensitive L-[7H]glutamate-binding sites in rat brain. J. Neurosci. 5, 290992919. 43. Monaghan D. T., Olverman J. H., Nguyen L., Watkins J. C. and Cotman C. W. (1988) Two classes of iv-methyl-o-aspartate recognition sites: differential distribution and differential regulation by glycine. Proc. nafn. Acud. Sci. L’.S.A. 85, 983669840. 44. ~oI~ahan J. B. and Michel J. (1987) Identi~cation and c~racterization of an ~-methyl-~-aspartate-specific L-[~H]glutamate recognition site in synaptic plasma membranes, J. ~e~ro~~e~z. 48, 1699-1708. 44a. Murphy D. E., Schneider J., Boehm C., Lehmann J. and Williams M. (1987) Binding of [?H]3-(2-carboxypiperazin-4- yl)propyl- I-phosphonic acid to rat brain membranes: a selective, high-affinity ligand for N-methyl-p-aspartate receptors. J. Pharmar. erp. Ther. 240, 778-784. 45. Nowak L.. Bregestovski P.. Ascher P., Herbert A. and Prochiants A. (1984) Magnesium gates glutamate-activated channels in mouse central neurons. Nufure 307, 462-465. 46. Olverman H. J., Jones A. W., Mewett K. N. and Watkins J. C. (1988) Structure/activity relations of N-n~ethyl-~- aspartate receptor ligands as studied by their inhibition of ~~H]D-2-amino-5-phosphonopentano~c acid binding in rat brain membranes. Neuroscience 26. 17-3 1. 47. Pan H. S., Frey K. A., Young A. ‘B. and Penney J. B. (1983) Changes in [‘HJmuscimol binding in substantia nigra, entopeduncular nucleus, globus pallidus and thalamus after striatal lesions as demonstrated by quantitative auto- radiography. J. Neuroxi. 3, 1189-l 198. 48. Perez-Clausell J. and Danscher G. (IY85) Intraves~cular localization of zinc m rat tclencephalic bouton<. A h,stochcmical study. Brcr;tl Re.\. 337, YI 9X 40. Pelers S.. Koh J. and Choi W. (19X7) Zinc selectively blocks the action of .~-methyl-I,-aspart~tte on cortical wuronb. .%icww 236, 589 593. 50. Ransom R. W. and Deschena N. L. (IYXX) NMDA-induced hippocampal [‘H]norepinephrinc rclcase IS modulated h) glycinc. Eur. J. Fhurrwc~. 156, 149 155. 51. Reynolds I. J.. Murphr\ S. N. and Miller R. J. (1987) ‘H-Labclcd MK-801 binding to the cxcitatnry amino aad reccplor complex from rat braIn is enhanced by glycinc. Prw. /w/11. A~,trcl. SC,;. C;.S.A. 84, 7744 774X. 52. Reynolds 1. J. and Miller R. J. (1988) [7H]MKXOI binding to the NMDA receptor,ionophorc complex is rcgulatcd bq divalent cations: evidence for multiple regulatory sites Elrr. J. f’l~ur~ncrc~.151, I03 II?. 53. Rothman S. M. and Olney J. W. (lY87) Excitotoxiclt> and the NMDA rcccptor. 7wztlv !Vt,rrro.cc.i. IO, 299 302. 54. Simon R. P.. Swan J. H.. Giftiths T. and Mcldrum B. S. (IYX4) Blockade of :V-methyl-I)-asp~irtate receptors may protect .qginst ischemic damage in the brain. .%iewc, 226, X50 X5:! -55 Sncll L. c‘.. Morter R. S. and Johnson K. M. (19X7) Glyc~nc potcnt~ateh ,V-methyl-I,-aapartule-induced [‘HITC‘P binding to rat cortical membl-ants. .I;cwo~ 1. f.v//. 83, 317 !I 7. 55a. Snell I.. D. and Johnston K. M. ( IYXX) C‘yclolcucmc competitively antagonizes the strychnine-insensitlvc glycinc rcccptor. tlrr. .I flwwclc~. 156, IO5 I IO Sncll L. D.. Mortcr R. S. and Johnson K. M. (IYXX) Structural rcqmremcnts for actwation of the glycine receplor that modulates the N-methyl-u-aspartate operated ion channel. Ew. .I. Phurrtwc~. 156, I05 I IO. Squires R. 1:. ( 19x3) Bewodiarepine receptor multiplicity. .V’~rroplfu”t~“f/~,~~/~~~~,22, 1443 14.50. Tlxmas J. W., Hood W. I-., Monahan J. B., Contreras I’. C and O’Donohue T. L. (19Xx) Glycine modulation of the phcncylclidine binding site in mammalian hra~n. Brcrir~ Km. 442, 396 398. llnncratall J. R.. Kuhar M. J.. Niehoff I>. L. and Palacio\ J. M. (19X1) Bcnzodiazeplne receptors arc coupled to a subpopulation of gamma-aminobutyric acid (GABA) reccptcjrs’ evidence from a quantitative autoradiographic stud?. .I. Phtrumrc~.c’_v,I. Th 218, 797 X04 Wutkin\ J. C‘. and Olverman l-1. J. (I YX7) ,Agonist and antagonists for excitatory amino acld receptors. Trcvd~ .Vcwrowc. 10. 265 272. Wcsthrook G. L. and Mayer M. L (lYX7) Micromolar concentration\ of Zn” antagoni7c NMDA and GABA rcsponsea of hippocampal neurons $‘trt~rr’r 328, 640 647 Wang E. [{. F.. Kemp J. A.. Prlesll) T., Knight A. R.. Woodrut?’ F. N. and Iversen L. L. (IYX6) The ant~convulsant MK-XOI is ;I potent ,~-melhyl-l,-asparrale antagonist. I’rr~ ,INIII. .4& .%i. l’.S.A. 83, 7104 7108. Wang E. 11. F.. Knight A. R. and Ransom R. (19X7) Glyclne modulates [‘HJMK-XOI binding lo the NMDA rcccptor in rat brain. Elo-. J. Phtrrwtr~. 142, 4X7 4XX. Wang E. H. F.. Kni&l A. R. and Woodruff<;. N. (19Xx) [‘HIMK-XOI labels a silt on the h-methyl-I)-aspartate receptor channel complex rn rat brain memhrancs. ./. ~V~wwc~/rcw.SO, 274 281