Proc. Natl. Acad, Sci. USA Vol. 74, No. 12, pp. 5524-5528, December 1977 Biochemistry Muscarinic acetylcholine receptors of the developing retina (synapse formation/receptor localization /3-quinuclidinyl benzilate) HIROYUKI SUGIYAMA, MATHEW P. DANIELS, AND MARSHALL NIRENBERG Laboratory of Biochemical Genetics, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda. Maryland 20014 Contributed by Marshall Nirenberg, October 7, 1977 ABSTRACT Six- and 13-day chicken embryo retinas contain 10 and 320 fmol per mg of protein of specific binding sites for 343H|quinuclidinyl benzilate, a ligand of muscarinic acetyl- choline receptors. Most of the receptors of 13-day embryo retina were found, by autoradiography, to be localized in two sharp bands within the inner synaptic layer of the retina. In the adult, the receptors were found almost exclusively in three bands in the inner synaptic layer of the retina. A possible mechanism for generating sets of stratified or columnar neurons and relating one set to another is proposed. The vertebrate retina provides a model system for synapse formation because synaptic circuits may be assembled with relatively few types of cells and because cultured neurons dis- sociated from retina form synapses in profusion in vitro (1, 2). Biochemical (3-7), histological (8, 9}, and electrophysiological (10-13) evidence strongly suggests that acetylcholine (ACh) functions as a neurotransmitter in the retina. Developmental] and histological studies of chicken retina acetylcholinesterase (EC 3.1.1.7 AChE) (8), ACh (14), choline acetyltransferase (EC 2.3.1.6) (2), and nicotinic ACh receptors (15, 16) have been reported. In this report, the properties of muscarinic ACh receptors of chicken retina, the number of receptors, and their distribution within the retina during embryonic development are de- scribed. MATERIALS AND METHODS Homogenate Preparations. Neural retinas of White Leg- horn chicken embryos were homogenized in 50 mM sodium phosphate buffer, pH 7.4 (buffer A). In some experiments, homogenates were diluted several times with buffer A and centrifuged at 17,300 X g for 20 min at 3°. The pellet was suspended in buffer A (membrane fraction). All experiments were performed with freshly prepared homogenates or mem- branes. Binding Assay. (3-+)-Quinuclidiny] benzilate (QNB), a gift from Hoffman-La Roche, Inc., was labeled by catalytic 9H exchange and purified as described by Yamamura and Snyder (17); the specific activity was 8.4 Ci/mmol. 3(+)-[3-3H}] QNB used in some experiments was obtained from Amersham /Searle (13 Ci/mmol). [3H|QNB binding was measured by a modification of the method of Yamamura and Snyder (17). Homogenates were combined with [7HJQNB in buffer A and incubated for various periods. Each 100- to 150-ul portion of the reaction mixture (usually containing 100-200 yg of protein) then was diluted into 5 ml of ice-cold buffer A, immediately filtered, and washed three times, each with 5 ml of buffer A. Binding kinetics were The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U. S. C. §1734 solely to indicate this fact. . 5924 measured at 25° by using Whatman glass fiber GF/C filters. The concentration of (+)-[(3HJQNB in the reaction mixture was 0.5-1.0 nM. When the effects of competing ligands were tested, homogenates were incubated with desired concentrations of ligands for 5-10 min at 25° and then mixed with (SH]QNB solution containing the same concentrations of the ligands. Equilibrium studies were performed at 4° with Millipore HAWF filters or, in some cases, GF/C filters (results were es- sentially the same). For the determination of nonspecific binding, homogenates were incubated with 0.4-10 uM atropine sulfate for 10 min in ice and then mixed with [HJQNB solution containing the same concentration of atropine. The number of [SH|QNB binding sites was determined by Scatchard analysis in some experiments but more often was determined at one saturating concentration of (+)-[7H]QNB (6-10 nM). [7HJQNB Autoradiography. Neural retinas were dissected in cold Dulbecco’s phosphate-buffered saline with Ca?+ and Mg?+ (PBS). Pieces of retina from 13-day embryos or adult chickens were incubated for 90 min in 5 ml of PBS containing 4 or 2 nM (+)-[SH|QNB (13 Ci/mmol), respectively, and then washed eight times, each with 5 ml of PBS. In control experi- ments, pieces of retina were preincubated in 5 ml of PBS con- taining 0.4 4M atropine sulfate for 10 min, followed by incu- bation in 5 ml of [SHJQNB solution in PBS containing 0.4 uM atropine sulfate for 90 min. The tissue then was washed twice with 5 ml of PBS containing 0.4 uM atropine sulfate and six times with 5 ml of PBS. Samples were kept in an ice bath at each step. Both experimental and control retinas were washed for 25 min (all washes). Retinas then were sandwiched between two pieces of mouse liver and frozen quickly in liquid Freon cooled in liquid nitrogen. Frozen pieces were sectioned (12 wm thick) and thaw-mounted onto glass slides coated with Kodak NTR-2 nuclear emulsion. To minimize diffusion of [2H|QNB, mounted sections were immediately dried under a stream of nitrogen gas. Slides were stored in the dark at 4° with a desiccant. Autora- diographs were developed, fixed, and then immersed in 2.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.0, for | hr at room temperature. Some slides were stained with 0.02% tolu- idine blue for 5 min at room temperature. Retina Cell Cultures. Cells were prepared from 8-day em- bryos and cultured in rotating petri dishes as described (1) with minor modifications: 1.5 X 107 cells in 3 ml of medium (95% Eagle’s basal medium with Earle’s salts and 5% fetal bovine serum) were cultured in a bacterial petri dish (35 mm; Falcon no. 1008) placed on a rotary shaker with an excursion of 2.6 cm (75-80 rpm) in a 37° incubator in a humidified atmosphere of 5% CO./95% air. Half of the medium was replaced each day. Abbreviations: ACh, acetylcholine, AChE, acetylcholinesterase: QNB. 3-quinuclidiny] benzilate; PBS, phosphate-buffered saline with Ca? and Mg?*. Biochemistry: Sugiyama et al. ool DRATES OF GNB BINDING JBQNS CONCENTRATION, 2 [AND RELEASE TOTAL | 9 100 a b @ 80 mam Zz Co Tt mou vy g Oa a aT OPE SA SEO MINUTES nV BH (+) ONB Fic. 1. QNB binding to receptors in homogenates of 13-day chicken embryo retina. (A) Kinetics of QNB binding to and release from sites on membrane preparations at 25°. The ordinate represents specific binding of [SH]QNB per 0.273 mg of protein (0.15 and 3.0 ml reaction mixtures for on and off reactions, respectively). The initial concentration of (+)-{3H]QNB was 2.0 nM. For the release kinetics experiment, the reaction mixture was incubated for 3 min and then diluted with 19 volumes of buffer A containing 0.4 uM atropine. Fifty-seven min after the first dilution, the reaction mixture was again diluted 20-fold with buffer A with 0.4 uM atropine. No release of QNB was observed during further incubation (not shown). (B) QNB con- centration curve and Scatchard plot (Inset). Specific binding (@) is the difference between total binding (0) and nonspecific binding (O) in the presence of 0.4 4M atropine. Each point represents the mean of triplicate values. B and F correspond to concentrations (nM) of specifically bound QNB and free (—)-QNB, respectively. The line without points represents the concentration of active isomer (~)-QNB added, The concentration of protein in the reaction mixture was 2.86 mg/m. Tubes were incubated at 4° for 100 min. RESULTS Receptor Properties. The rates of (7H|QNB binding to and release from receptors in homogenates of 13-day chicken em- bryo retina are shown in Fig. 1A. [H|QNB bound rapidly to retina membranes; in the presence of 2 nM (+)-(3HJQNB, half maximal binding was achieved in 2 min and maximal binding, in approximately 15 min. Some, but not all, of the reactions are reversible. The addition of 0.88 uM atropine and dilution of reaction mixtures 20-fold resulted in the dissociation of ap- proximately 50% of the (3H|QNB-receptor complex. QNB association and dissociation reactions both exhibited biphiasic kinetics with fast and slow association reactions and dissociation reactions. The kinetics will be discussed elsewhere; however, the rate constants (k) for fast and slow QNB association reactions were estimated, by assuming bimolecular irreversible reactions asa first approximation, to be 2.7 X 108 M7! min7! and 1.4 X 108 M~! min7}, respectively. Both fast and slow QNB-receptor dissociation reactions were first-order reactions with rate con- stants of 1.2 min7! and 0.041 min7!, respectively. The relationship between [3H|QNB concentration and binding to receptors in retina is shown in Fig. 1B. The binding of the pharmacologically active isomer, (—)-[7H|QNB, to retina receptors is a saturable process. In the presence of 0.4 4M at- ropine, relatively little nonspecific (3H|QNB binding was found with homogenates prepared from 13-day chicken embryo retina; however, nonspecific (3H |QNB binding was markedly increased when homogenates are prepared from >15-day chicken embryo retina or posthatched retina. The dissociation constant (Kp) determined by Scatchard analysis (Fig. 1B and inset) was 0.12 nM (—)-[(3H|QNB. However, we consistently observed higher apparent dissociation constants (~0.4 nM) with homogenates from retina of chickens 2 weeks after hatching and of adult chickens. The calculated concentration of specific ONB binding sites in 13-day chicken embryo retina is 325 Proc. Natl. Acad. Sci. USA 74 (1977) 5525 —T tT ia MUSCARINE OF INTIAL RATE OF QNB BINDING on ° —O 4 L S LOG [B/Bpg-B)] % ° v T NX \ t \ . . ‘N =, 1 J 4 1 4 o oo OO @ 6 [LIGAND] (M) Fic. 2. (A) Inhibition of [[H]QNB binding by various com- pounds. (B) Hill plot. Thirteen-day embryo retinas were used. Whole homogenates and membrane fractions were used and no significant difference was noted. The final concentration of (+)-[PHJQNB was 0.5 nM (1.0 nM in some experiments). Protein concentrations in re- action mixtures were 0.5-2.6 mg/ml. Receptor concentrations were 0.1-1.0 nM. Initial rates of binding (usually 0 to 3-4 min) were fitted to a model of bimolecular irreversible reaction mechanism, and the bimolecular association rate constant was calculated. The apparent rate constant in the presence of protecting drugs was expressed as the percentage of the control value in A. Hill plots were obtained by as- suming that the percentage decrease of [*H]QNB binding rate rep- resents the percentage of the receptor sites occupied by unlabeled ligands which corresponds to B/Bmax- When ACh was tested the ho- mogenate was preincubated with 3 uM eserine for 30 min at 25° before addition of ACh. Symbols: 0, scopolamine; ©, atropine sulfate; ¢, oxotremorine; @, AChCI; a, carbamylcholine chloride; +, pilocarpine; * muscarine. fmol/mg of protein and each retina contained 818 fmol per retina (4.9 X 10!! sites per retina) of specific QNB binding sites. The apparent Hill coefficient is 1.0 (plot not shown), which suggests that QNB binds to independent, noninteracting re- ceptors. Although only one population of QNB binding sites was detected by Scatchard analysis, kinetics of the QNB binding to and release from receptors show that (7H|QNB-receptor complexes are heterogeneous. The effects of different concentrations of unlabeled ligands known to activate or inhibit muscarinic ACh receptors on the initial rate of (3H}QNB binding to receptors in homogenates prepared from 13 day chick embryo retina are shown in Fig. 2A. [SHIQNB binding was markedly decreased in the presence of antagonists of muscarinic ACh receptors such as scopolamine or atropine or receptor activators such as oxotremorine, ACh, carbamylcholine, pilocarpine, or muscarine at expected physiological concentrations. (7H|]QNB binding was not af- fected by prior incubation of homogenates with 10 nM a- bungarotoxin for 3 hr at 25° (not shown). Thus, the specificity of QNB binding sites for ligands closely resembles that of muscarinic ACh receptors. As shown in Fig. 2B, the apparent Hill coefficients of acti- vators of the muscarinic ACh receptor such as oxytremorine, ACh, and carbamylcholine were 0.6 to 0.8, whereas those of receptor antagonists were approximately 1. These results agree well with those of Birdsall et al. (18). The apparent Hill coef- ficients of pilocarpine and muscarine were approximately 1. Pilocarpine has been shown to be both an activator and an an- 5526 Biochemistry: Sugiyama et al. Table 1. Apparent dissociation constants and Hill coefficients of ligands for muscarinic acetylcholine receptors of 13 dav chick embryo retina App Kp,* Ligands nM h Antagonists QNB 4° 0.12 1.0 25° 0.44 1.0 0.29" — Atropine 0.69 1.0 Scopolamine 0.17 1.1 Activators Oxotremorine 130 0.7 Acetylcholine 1,100 0.8 Carhamylcholine 1,700 0.6 Muscarine 8,700 Ll Pilocarpine 1,100 1.0 Local anesthetics Tibucaine 30,000 1.0 Tetracaine 21,000 1.4 * Values at 25° except otherwise specified. The Kp values for QNB were obtained by determining the binding of |"H}QNB at equilib- rium. The Kapp values for the other antagonists and activators represent the concentrations that result in 50% inhibition of the initial rate of ["H]QNB binding; values were not corrected for h< 1. The K app values for local anesthetics are estimated from experi- ments where the retina homogenate with or without different con- centrations of a local anesthetic were incubated for 60 min at 25° in the absence of [*H|QNB, then 0.50 nM (+)-[7H]QNB was added and the reaction mixtures were incubated for an additional 5 min. + 0.29 and 4.4 nM (—)-QNB are the dissociation constant values de- termined from rate constants for slow and fast association and dissociation reactions, respectively. tagonist of the muscarinic ACh receptor; although muscarine is an activator of the muscarinic ACh receptor in other organ- isms, the apparent Hill coefficient with chicken embryo retina receptors resembles that of a receptor antagonist. The apparent Hill coefficients of <1 observed with oxytremorine, ACh, and carbamylcholine can be interpreted in various ways such as negative cooperativity, heterogeneity of binding sites, or de- sensitization of the ACh receptor. The demonstration by W. Klein in this laboratory that heterogeneity of muscarinic ACh receptors can be detected by kinetic studies was confirmed. Muscarinic ACh Receptors during Embryonic Develop- ment. The concentration and number of muscarinic ACh re- ceptors in chicken embryo retina are shown in Fig. 3 A and B, respectively, as a function of developmental age. Values re- ported for nicotinic ACh receptors (16) also are shown for comparison. Muscarinic ACh receptors were detected in 5.5- day embryo retina, but the concentration was relatively low (10 fmol/mg of protein). Between the 6th and 14th days of em- bryonic development, the concentration of specific QNB binding sites increased 30-fold. In the 5.5- to 9-day chicken embryo retina, specific QNB binding sites accumulated expo- nentially with a doubling time of approximately 26 hr; in 9- to 14-day embryo retina, the doubling time was approximately 60 hr. The maximal concentration of QNB binding sites (320 fmol/mg of protein) was attained in the retina of the 13- to 14-day embryo. No further change was detected during later embryonic development; however, receptor concentrations were lower in retina 2 weeks after hatching and in adult retina. Although the concentration of receptors decreased after hatching, the number of receptors per retina increased slightly after hatching (Fig. 3B). The adult chick retina contained 1200 Proc. Natl. Acad. Sci. USA 74 (1900) RET RORLNE RECEPTOR ROETYIOHOUNE RECEPTORS, yey 2O000F CONCENTRATION PER RETINA ee 2000 ® ® o e 1ooob + F487 1000 sooF NN NO “pauses 4 500 CULTURE x L ey, t t ‘ ‘ 100 NICOTINIC 4 100 go oO tT eu oO fMOL QNB OR a-BT BOUND/RETINA HATCH fMOLS QNB OR a-BT BOUND /mg PROTEIN ° 14 sail Shitty yt tye) 5 05 10 15 20 4ADLTS 10 18 20) ADULT DAYS Fic. 3. The accumulation of QNB binding sites in chicken em- brvo retina as a function of development age. The number of specific sites per mg protein (A) or per retina (B) is shown. For comparison, «-bungarotoxin (a-BT) binding to nicotinic ACh receptors (16) is shown (broken lines). Circles. intact retina in vivo; triangles, cultured cells dissociated from 8-day chicken embryo retina. Filled circles are values obtained by Scatchard analysis. Open symbols represent the specific binding obtained at 6-10 nM (+)-[SH]QNB. Tubes were in- cubated at 4° for 60 min. In some cases, the retina dissection was not complete; the amount of protein per retina then was estimated from the published values (19). Each point represents the mean of at least three determinations. fmol of specific QNB binding sites per retina(7.2 X 10!! sites per retina), These results show that genes for muscarinic ACh receptors are expressed early in the development of the retina and suggest that some neurons synthesize muscarinic ACh receptors but not neuroblasts as reported for nicotinic ACh receptors (16). The number of muscarinic and nicotinic ACh receptors increase more than 30-fold and the receptors accumulate at similar rates between the sixth and ninth days in embryos. The maximal concentration of muscarinic ACh receptors is attained in the retina of the 13-day embryo, whereas nicotinic ACh receptors continue to increase until hatching. The concentration of spe- cific QNB binding sites in retina of the 6- to 13-day embryo is 2- to 3-fold higher than the concentration of a-bungarotoxin binding sites; however, this ratio is reversed in the adult retina. Thus, the ratio of muscarinic to nicotinic ACh receptor changes markedly during retina development. Cells dissociated from 8-day chicken embryo retina were cultured for various times in rotating petri dishes. At various times, homogenates were assayed for specific binding of (3HJQNB (Fig. 3A). The concentration of QNB binding sites increased from 50 fmol of specific QNB binding sites per mE of protein in retina of the 8-day embryo to 225 fmol/mg ol protein after 4 days of culture. Thus, the accumulation 0! muscarinic ACh receptors in cultured retina cells resembled that in the intact retina. Receptor Distribution in Retina. The distribution of [SH|QNB binding sites in intact 13-day chicken embryo retina and in adult retina is shown in Fig. 4. In 13-day embryo retina. most of the silver grains were localized in two narrow bands within the inner synaptic layer of the retina (also termed “inner plexiform layer”) (Fig. 4 A and B). In adult retina (Fig. 4 C and Biochemistry: Sugiyama et al. SS ee ve Fic. 4. Autoradiography of sections of chicken retina incubate with ["H|QNB in the absence of atropine. (A) Phase-contrast view and (B) dark-field view of stained section of 13-day embryo retina exposed for 390 days. (C) Phase-contrast view and (D) dark-field view of stained section of adult chicken retina exposed for 173 days. (Bars represent 100 um.) Lines at the left of each photograph represent the boundaries of layers: R, photoreceptor layer; O, outer synaptic layer; IN, inner nuclear layer; IS, inner synaptic layer: G, ganglion cell layer: A, ganglion axon layer. D), two or three bands of silver grains could be seen within the inner synaptic layer of the retina. Histograms relating the density of silver grains on autora- diographs that had been exposed for shorter times with grain location over the retina are shown in Fig. 5. Two sharply de- fined bands of silver grains of equal density can be seen within the inner synaptic layer of 13-day embryo retina. Fewer silver grains were associated with the lower portion of the inner nu- clear layer (cell bodies of amacrine and bipolar neurons and Muller cells) and with ganglion neuron soma and axons but were not associated with other regions of the retina. The average number of silver grains over the entire retina incubated in the absence or presence of 0.4 4M atropine (nonspecific (RH|QNB binding) was 3.83 and 0.87 grain per 100 um?, respectively. 25 & ~ 1.5 2 eS 2 S.0Fr Oo 2 a =. 2 a 2.97 0.5 S$ a AQ ~ uy | UONSPECIFIC _° : | Loa iC a IT 4 qe | ~~ eL LOX a = io 2 io on Ceo Tr I 0.5 \ I pepe Oo jeg Joey tieins po 0 RO N S Ga RON S Fig. 5, Histograms showing the grain distribution in ]H|QNB autoradiographs of sections of 13-day chicken (A) or adult chicken (B) retina. Sections of retina of both ages, treated for both total binding and nonspecific binding, were subjected to autoradiography for 65 days. Grains were counted at X600 magnification by using a camera lucida. Specific binding was obtained by subtracting non- specific from total binding. Number of grains counted were: 543 and 220 for total and nonspecific binding, respectively, for 13-day embryo retina: 1592 and 847 for total and nonspecific binding, respectively, tor adult retina. Abbreviations are as in Fig. 4. Proc. Natl. Acad. Sci. USA 74 (1977) 5527 EMBRYO ADULT O DAY \ Ar ~ ob) acre acre ud \ oT 4 s O 20 oe tT >