Identification and Localization of Muscarinic Acetylcholine Receptor Proteins in Brain with Subtype-Specific Antibodies

The Journal of Neuroscience, October 1991, 1 f(10): 32183228

Identification and Localization of Muscarinic Acetylcholine Receptor Proteins in Brain with Subtype-specific Antibodies

Allan I. Levey,132,5 Cheryl A. Kitt,2 William F. Simonds,4 Donald L. Price,1v2.3 and Mark R. Brann5,a Departments of ‘Neurology, 2Pathology, and 3Neuroscience, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, 4Metabolic Diseases Branch, The National Institute of Diabetes, Digestive, and Kidney Diseases, and 5Laboratory of Molecular Biology, National Institute of Neurological Diseases and Stroke, National Institutes of Health, Bethesda, Maryland 20892

mRNAs encoding five genetically distinct muscarinic ACh characterization of receptor subtypes, including cellular and receptors are present in the CNS. Because of their phar- subcellular localization, will be of great value in defining the macological similarities, it has not been possible to detect targets for the development of more effective and specific the individual encoded proteins; thus, their physiological therapeutic agents. functions are not well defined. To characterize the family of proteins, a panel of subtype-selective antibodies was gen- In the CNS, muscarinic ACh receptorsmediate many functions erated against recombinant muscarinic receptor proteins and (e.g., memory, learning, arousal,and motor control), and drugs shown to bind specifically to each of the cloned receptors. acting at these sites have potential value in the treatment of a Using immunoprecipitation, three receptor proteins (m,, mz, number of neurological disordersinvolving choline@ systems, and m,) accounted for the vast majority of the total solubi- including Alzheimer’s diseaseand Parkinson’sdisease. Recent- lized muscarinic binding sites in rat brain. These receptor ly, five genesthat encode highly related muscarinic receptor subtypes had marked differences in regional and cellular proteins (m,-m,) have been identified (Kubo et al., 1986; Bonner localization as shown by immunocytochemistry. The m,-pro- et al., 1987, 1988; Peralta et al., 1987). The proteins have seven tein was present in cortex and striatum and was localized putative membrane-spanningregions that are characteristic of to cell bodies and neurites, consistent with its role as a major the superfamily of G-protein coupledreceptors and exhibit func- postsynaptic muscarinic receptor. The m,-receptor protein tional differences:m,, m,, and m, preferentially activate phos- was abundant in basal forebrain, scattered striatal neurons, pholipaseC via a pertussistoxin-insensitive G-protein; m2and mesopontine tegmentum, and cranial motor nuclei; this dis- m4 inhibit adenylyl cyclase activity via a pertussistoxin-sen- tribution is similar to that of cholinergic neurons and sug- sitive G-protein (Peralta et al., 1988; Hulme et al., 1990; Wess gests that m2 is an autoreceptor. However, m2 was also pres- et al., 1990). Similarly, the receptors selectively modulate a ent in noncholinergic cortical and subcortical structures, variety of ion channels(Fukuda et al., 1988; Joneset al., 1988a,b; providing evidence that this subtype may presynaptically Higashida et al., 1990; Hulme et al., 1990). The five mRNAs modulate release of other neurotransmitters and/or function have distinct distributions in brain (Buckley et al., 1988; Weiner postsynaptically. The ma-receptor was enriched in neostri- et al., 1990), further highlighting the possiblefunctional differ- atum, olfactory tubercle, and islands of Calleja, indicating encesamong the subtypes. However, direct characterization of an important role in extrapyramidal function. These results the receptor proteins in the nervous system hasbeen limited by clarify the roles of these genetically defined receptor pro- the complex pharmacologicalproperties of the subtypes(Peralta teins in cholinergic transmission in brain. Since the nonse- et al., 1987; Akiba et al., 1988; Buckley et al., 1989; Dorje et lective muscarinic drugs used in the treatment of patients al., 1991). For example, even the most selective ligandsavail- with neurological disease produce many side effects, the able, such as pirenzepine, which have been usedto define pharmacologically distinct binding sites(Hammer et al., 1980), have Received Mar. 20, 1991; revised May 14, 1991; accepted May 16, 1991. similar affinities for multiple cloned receptor subtypes(Peralta We thank Drs. Solomon Snyder and Joseph Coyle for reviewing the manuscript; et al., 1987; Buckley et al., 1989; Hulme et al., 1990; Dorje et Craig Heilman, Helen Fedor, Sharon Edmunds, and David Gdula for expert al., 199 1). Therefore, the relative abundance, localization, and technical assistance; Frank Doje for evaluating antisera; and Norman Nash for functions of the native muscarinic receptor proteins in the ner- synthesis of oligonucleotides. This work was supported by grants from the U.S. Public Health Service (NIH AG 03359, AG 05146) The Robert L. and Clara G. vous system are presently unknown. A better understandingof Patterson Trust, and The Metropolitan Life Foundation. D.L.P. is the recipient this gene family may lead to the development of more effective of a Javits Neuroscience Investigator Award (NIH NS 10580) and a Leadership and Excellence in Alzheimer’s Disease (LEAD) award (NIA AG 079 14). A.I.L. is and specific therapies for many neurologic and psychiatric dis- the recipient of a Clinical Investigator Development Award (NIH NS 0.1387) and eases. a Faculty Scholar Award from the Alzheimer’s Disease and Related Disorders We have recently developed a strategy using molecular and Association. Correspondence should be addressed to Dr. Allan Levey, Department of Neu- immunological techniques to produce subtype-specificantisera roloev. The Johns Hopkins University School of Medicine, 509 Pathology Build- to recombinant muscarinicreceptor proteins(Levey et al., 1990). ing, 600 North Wolfe-Street, Baltimore, MD 2 1205. In the present study, we have produced solublemuscarinic re- = Present Address: Department of Psychiatry, University of Vermont, Burling- ton, VT 05405. ceptor fusion proteins in bacteria, incorporating large regions Copyright 0 1991 Society for Neuroscience 0270-6474/91/l 13218-09$05.00/O of the nonconservedthird cytoplasmic (i3) loops of the human The Journal of Neuroscience, October 1991, 17(10) 3219

sequences. Antisera to these fusion proteins are shown to react incomplete Freund’s adjuvant, and then were boosted monthly with 50 highly specifically with each of the five cloned receptor subtypes. pg of protein. Blood was obtained by venipuncture at 3 and 4 weeks postboost, and the sera were stored at - 70°C. For the present studies, Using these antisera, the genetic subtypes of muscarinic receptor antisera taken 10 weeks post-primary immunization [m,, AL14; m2, proteins in rat brain are directly characterized by immunopre- AL1 7; m3, AL18; m4, AL28 or AL7 (Levey et al., 1990); and m,, AL221 cipitation and immunocytochemistry. were used for immunoprecipitation studies and affinity purified for immunocytochemistry. Affinity purification was performed as described previously (Harlow and Lane, 1988) using the respective purified fusion Materials and Methods proteins (24 mg) conjugated to Affi-Gel (Sigma). Sera were preadsorbed Construction ofexpression plasmids. Segments of the human muscarinic with 250 &ml of purified glutathione S-transferase, and immunopre- receptor genes m,, m2, ma, and m, were subcloned into the bacterial cipitates were clarified prior to affinity purification in order to remove exoression vector DGEXZT (Smith and Johnson. 1988). as described antibodies reactive with the nonmuscarinic portions of the fusion pro- previously for the‘ m, receptor (Levey et al., 1990). dligonucleotide teins. primers (5 1 bases), synthesized on an Applied Biosystems DNA syn- Immunoprecipitation studies. Stable CHO-Kl cell lines individually thesizer, were used in polymerase chain reactions (PCR) in the presence expressing the human muscarinic receptor subtypes m,-m, or dissected of cloned templates to amplify DNA sequences encoding the i3 loops regions of rat brain were used for these experiments. Production of the of each receptor subtype as described previously (Levey et al., 1990). cell lines and characterization of the receptor binding properties have Restriction sites for BamHI and EcoRI were incorporated into the prim- been previously described (Buckley et al., 1989; Dorje et al., 199 1). The ers at the 5’ and 3’ ends of the gene fragments, respectively, except for total number of receptors/mg of membrane protein (determined by m,, which contains an internal BamHI site and was thus replaced by a saturation binding) for each subtype is m,, 25 18 fmol; m2, 747 fmol; BglII site at the 5’ end. In addition, stop codons were placed in the 3’ m,, 1831 fmol; m4, 1778 fmol; and m,, 954 fmol. Animal care was oligonucleotide to ensure that no additional amino acids were translated. performed in accordance with institutional guidelines. Two male Sprague- The sequences of the oligonucleotide primers were m, (sense), 5’-AGG Dawley rats were decapitated, and the brain regions were dissected, TCC GGA TCC GAG ACG CCA GGC AAA GGG GGT GGC AGC pooled, and assayed the same day. Dissected regions included (1) cortex, AGC AGC AGC TCA; m, (antisense), 5’-AGG TCC GAA TTC ACT including the cortical mantle (minus olfactory bulb) and hippocampus TCCGCTTGGCAGCTGCTCCTTTCCACGGGGCTTCTG; (dorsal and ventral): (2) striatum, including caudate-putamen, accum- mz (sense), 5’-CGA TCG AGA TCT GTT GCC AAC CAA GAC CCC bens, olfactory tubercle, globus pallidus, and basal forebrain; (3) thal- GTT TCT CCA AGT CTG GTA; m2 (antisense), 5’-AGG TCC GAA amus, including most of the dorsal thalamus and habenula; and (4) TTC ACT GCT TTT CAT CTC CAT TCT GAC CTG AAG ACC hindbrain, including pons, medulla, and cerebellum. Membranes from CCA CTA; m4 (sense), 5’-AGG TCC GGA TCC CCG AAG GAG AAG the cell lines and brain were prepared as described previously (Luthin AAA GCC AAG ACG CTG GCC TTC CTC AAG: m, (antisense). et al., 1988). Receptors were solubilized by resuspending membranes 5-AGG TCC GAA TTC AGC GCA CCT GGT TGC GAG CGA TGC (to 1.O mg/ml protein) in 10 mM Trisll .O mM EDTA (TE) buffer con- TGG CGA ACT TGC GGG, m, (sense), 5’-AGG TCC GGA TTC AAA taining 1.0% digitonin and 0.2% cholic acid and labeled’with 3H-N- GCT GAG AAG AGA AAG CCA GCT CAT AGG GCT CTG TTC; methylscopolamine (3H-NMS, 1.O nM). Receptors (195 11) were mixed and m, (antisense), 5’ AGG TCC GAA TTC ACA TTT GAT GGC with 5 ~1 ofantisera to m,-m,i3 fusion proteins and coprecipitated with TGG GGT TGG GAT TGA GGC CTT TCG TTG. PCR amplification goat antirabbit immunoglobulin, and radioactivity in the immunoprecip- of the DNA, isolation and restriction digestion of the fragments, and itates was determined (Levey et al., 1990). Nonspecific trapping of re- subcloning were all performed using standard procedures (Sambrook et ceptors in immunoprecipitates was determined using control antisera al., 1989). The fragments were subcloned into the BamHI-EcoRI mul- [nonspecific rabbit serum (NRS)] raised against a fusion protein without tiple cloning site on pGEX2T. Recombinant plasmids were verified to a muscarinic receptor moiety. Total soluble receptors were determined be in frame by double-stranded dideoxy sequencing (Kraft et al., 1988). by gel filtration on G-25 Sephadex and were corrected for nonspecific The pGEX2T vector encodes the 27.5 kDa glutathione S-transferase binding in the presence of 1 PM atropine. (GST; EC 2.5.1.18, from Schistosoma japonicum) fused to the N-ter- Immunocytochemistty. Fourteen male albino rats (Charles River, 250- minus of the muscarinic i3 loops. Transcription of the fusion sequences 350 gm) were deeply anesthetized with 4% chloral hydrate and perfused are under tat promoter regulation, and inducible with isopropylthio- intracardially with 0.9% saline, followed by 0.1 M phosphate-buffered galactoside (IPTG). 3% paraformaldehyde, pH 7.6. The brains were immediately removed Induction and puriJication of muscarinic i3 fusion proteins in Esche- and cryoprotected in 30% sucrose in 0.1 M phosphate buffer (pH 7.6) richia coli. Strain BLZl(DE3)pLysS was used to express the fusion pro- overnight at 4°C frozen on dry ice, and sectioned at 40 pm on a freezing teins as described (Smith and Johnson, 1988; Levey et al., 1990). Briefly, sliding microtome. Tissue sections were processed for immunocyto- the bacteria were transformed with either a recombinant or parent vector chemistry using the peroxidase-antiperoxidase method as previously and grown in 400 ml of LB broth containing 100 &ml amuicillin and described (Kitt et al., 1984). Affinity-purified antibodies were used at 50 &ml chloramphenicol to an optical density .of 0.4, and then the 0.5 &ml. Immunocytochemical controls consisted of adsorption of the fusion proteins were induced with 1 mM IPTG for 34 hr. Cultures were antibodies with 100 j&ml of GST or the muscarinic i3 fusion proteins harvested, and the cell pellets were washed with 10 ml of buffer A (50 for 20 min prior to staining. An additional series of sections was in- rnM Tris, pH 8.0, 10 mM EDTA, 25% sucrose), frozen and thawed, cubated in normal rabbit immunoglobulin (1 .O &ml). All sections were resuspended in 12 ml of buffer A containing 40 mg of lysozyme, and dehydrated through a graded series of alcohols and xylenes and cov- incubated on ice 1 hr. The suspension was centrifuged, and the cell pellet erslipped for microscopic examination. was resuspended in 10 ml of buffer B (10 mM Tris, 1 mM EDTA, 1 mM dithiothreitol) containing a protease inhibitor mixture (1 mM phenyl- methylsulfonyl fluoride, 0.2 mg/ml aprotinin, 1 &ml leupeptin, and 1 Results &ml pepstatin), repeatedly frozen and thawed, and mixed with 25 ml Muscarinic i3 loop fusion proteins of buffer C (20 mM HEPES, pH 7.6, 100 mM KCI, 0.2 mM EDTA, 20% glycerol) containing protease inhibitors, and 4 ml of 10% Triton X- 100. Soluble fusion proteins incorporating the i3 loops of the five muscatinic receptors were obtained in high yields (1 O-25 mg/ The solubilized fusion oroteins were affinitvI - muitied bv batch urocedure using 4 ml of reduced glutathione-agarose beads (5b% slurry; Sigma liter) from cultures of E. coli transfected with the recombinant Chemical Co.) equilibrated in buffer C, mixed for 2 hr at 4°C and washed plasmids (Fig. 1). Two or more protein bands were seen with five times with 20 ml of cold buffer C containing protease inhibitors. The fusion proteins were eluted with buffer C containing 5 mM reduced each ofthe purified preparations; smaller proteins may represent glutathione, pH 7.5, and the fractions were analyzed by SDS-PAGE, degradation products. We have also observed shortened prod- pooled, and dialyzed into buffer B using Centriprep chambers (Milli- ucts with expression of other recombinant pGEX2T plasmids pore). (A. I. Levey, unpublished observations). The proteins migrated Immunization of rabbits and affinity purification of antisera. Each in the range of mobilities expected after fusion with the 27.5 fusion protein was injected into two female New Zealand White rabbits. Initially, animals received 200 PLgof affinity-purified fusion protein (in kDa GST, given that the cloned segments of the i3 loops encode 0.75 ml) emulsified in an equal volume of comolete Freund’s adiuvant, the following number of amino acids: m,, 126; m2, 135; m,, followed by secondary injections at 3 weeks-with the same dose in 193; m4, 152; m,, 99. 3220 Levey et al. - Muscarinic Receptor Subtype Proteins in Brain

PGEX mi m2 m3 m4 m5 were also precipitated with a high degree of efficiency. Other experiments with fewer receptors demonstrated almost com- 97- plete recovery of m2 (87% of 24 fmol) and m4 (97% of 9 fmol). These experiments demonstrated that antibodies generated 66- against recombinant i3 loops were able to bind quantitatively the functional muscarinic receptors with virtually complete subtype specificity.

45- Immunoprecipitation of muscarinic receptor subtypes from rat brain The direct identification of the genetically defined muscarinic 31- receptor subtypes in rat brain and their regional distributions were determined by immunoprecipitation (Table 2). Three receptors, m,, m,, and m4, the major subtypes solubilized from brain, showed marked regional variability. For example, cortical 22- tissues contained higher levels of m,- and m,-receptors (40% each) and lower levels of the m,-subtype (15%). There were no Figure 1. Coomassie blue-stained 10% SDS-PAGE analysis ofaffinity- major differences between neocortex and hippocampus when purified muscarinic receptor i3 loop fusion proteins. Approximately 1 assayed separately (data not shown). The striatum (a composite fig of GST [encoded by the pGEX2T @GEX) parental plasmid] or 10 of neostriatum, nucleus accumbens, globus pallidus, basal fore- pg of the i3 loop fusion proteins ml-m5 were loaded in the lanes as brain, and olfactory tubercle) showed similar amounts of m,-, marked. Migration of molecular weight markers (in kDa) is indicated on the left. m2-, and m,-receptor proteins. The m,-receptor predominated in thalamus, with relatively fewer m,- and m,-proteins; m, was the major muscarinic subtype solubilized from hindbrain. The Characterization of subtype-specific antisera experiments were performed using limiting numbers of total The specificities of antisera raised against the muscarinic i3 loop receptors to ensure quantitative binding of the subtypes (deter- fusion proteins were tested by immunoprecipitation of cloned mined using the cloned receptors). The m,- and m,-receptor muscarinic receptor subtypes expressed in stably transfected proteins were not immunoprecipitated from brain, although both CHO-Kl cells (Table 1). The receptors were individually sol- subtypes were detected in cells transfected with the respective ubilized with detergent, labeled with 3H-NMS, a ligand that cDNAs using identical assay conditions (Table 1). Furthermore, binds with high affinity to all five subtypes, and immunopre- we have detected the native m,-receptor protein in glandular cipitated with antisera raised against the fusion proteins. Each tissue using the same conditions and reagents (Dorje et al., in receptor subtype was immunoprecipitated by a single antiserum press). corresponding to the homologous fusion protein used for immunization. Heterologous antiserum-receptor pairs showed lit- Immunocytochemical localization of muscarinic receptor tle or no binding above background levels (nonspecific trapping subtypes in control immunoprecipitates using antisera raised against a The three most abundant muscat-uric receptor proteins (m,, m,, fusion protein not containing a muscarinic receptor polypep- and m,) characterized by immunoprecipitation were localized tide). The relatively small differences (three- to fourfold) in total with cellular and subcellular resolution using the subtype-spe- receptors per assay were due to differences in levels of expression cific antibodies (Fig. 2). In general, immunoreactivity typically in the cell lines. Antibody binding did not inhibit receptor- appeared to be concentrated in the surface membranes of cells ligand interaction since the reversibly bound 3H-NMS was di- or neuritic processes, although the precise localization was dif- rectly recovered in the immunoprecipitates; other experiments ficult to evaluate light microscopically. Cytoplasmic staining also showed no inhibition of the antibodies in a soluble receptor was also occasionally evident and was frequently adjacent to binding assay using the same radioligand. The m,- and m,- the nucleus in the presumed Golgi apparatus. No staining of receptors were quantitatively recovered, and the other subtypes nuclei or glia was detected with any of the antibodies. Specificity

Table 1. Immunoprecipitation of muscarinic receptor subtypes with antisera to the i3 loop fusion proteins

Receptor Total subtype receptors Immunoprecipitates(fmol/pellet) (fmoVassay) NRS m, n-b m, m, m, ml 44.2 + 2.1 2.9 f 0.6 49.6 + 5.4 2.8 rf- 0.4 2.8 ZIZ0.1 2.4 rt 0.5 2.9 rt 0.1 m2 38.0 + 0.3 1.3 -t 0.5 1.8 zk 0.2 28.1 k 3.4 1.5 + 0.2 1.3 + 0.3 2.2 k 0.1 m3 11.6 + 1.9 0.8 + 0.1 0.9 f 0.1 0.9 + 0.1 13.4 f 0.7 0.8 + 0.1 0.9 + 0.1 m4 40.6 + 0.9 1.5 f 0.3 1.8 zk 0.1 1.7 Ik 0.4 1.7 IL 0.2 27.8 + 1.5 1.5 + 0.4 m5 15.7 f 1.4 2.5 f 0.1 2.0 k 0.3 2.3 k 0.5 1.8 + 0.2 2.1 + 0.4 12.4 + 2.6 Functional receptors were detergent solubilized from stable CHO-KI cell lines transfxted with each of the five human muscarinic receptor cDNAs, labeled with ‘H-NM& and immunoprecipitated with control serum (NRS) or antisera to the muscarinic receptor i3 loops m,-m,. Shown are the mean k SD of triplicate samples. The homologous receptor- antiserum pairs are shown in italics. The Journal of Neuroscience, October 1991, 1 f(10) 3221

Table 2. The distribution of muscarinic receptor subtypes in rat brain as determined by immunoprecipitation with subtype-specific antibodies

Region Total lmmunoprecipitates (fmol/pellet) receotors (fmol/assay) NRS ml m2 m3 m4 ms Cortex 47.4 + 2.4 1.2 k 0.2 20.4 f 2.3 18.7 k 3.1 1.6 f 0.2 8.6 f 1.6 1.4 k 0.3 (40) (37) (0) (15) (0) Striatum 48.8 f 2.0 1.2 + 0.2 16.6 f 2.6 15.2 -+ 1.8 1.4 f 0.2 15.4 -f 4.0 1.4 * 0.2 (31) (29) (0) (29) (0) Thalamus 41.8 +- 1.3 1.1 f 0.1 3.4 +- 0.3 21.5 + 2.3 1.3 -t 0.2 7.5 * 1.7 1.3 + 0.3 (6) (49) (0) (15) (0) Brainstem 20.6 * 2.4 0.9 f 0.2 1.3 + 0.3 18.2 f 1.8 1.0 + 0.2 1.2 * 0.2 1.1 -c 0.2 (2) (84) (0) (1) (1) Total receptors represent solubilized ‘H-NMS binding sites as determined independently by gel exclusion chromatog- raphy. Shown are the mean + SD (n = 6) from a representative experiment. Numbers in parentheses are the percentages of total receptors specifically bound (minus background values in control immunoprecipitates with NRS concentrations). was demonstrated by complete inhibition of staining after cumbens (Fig. 3). These findings are in register with the local- preadsorption of the antisera with the respective i3 loop fusion ization of m,-, m,-, and m,-mRNAs in striatum and olfactory proteins (Fig. 2D’-F’). Preadsorption of the antiserawith GST, tubercle (Buckley et al., 1988; Weiner et al., 1990; Vilaro et al., the protein common to each of the muscarinic fusion proteins, 199 1). Although immunoprecipitation studies demonstrated had no observableeffect on staining (data not shown). The m,- approximately equal levels of the three proteins in striatum, and m,-receptors were not detected reliably by immunocyto- immunocytochemistry showed that much of the ml-protein in chemistry, in agreementwith the immunoprecipitation studies. this region was due to enrichment in basalforebrain, including This might be due to methodological problems (e.g., denatur- the septum (Fig. 2) and more caudal aspectsof cholinergic cell ation of the epitopes by fixation), or possibly lower levels of groups(Ch l-4). The m,-immunoreactivity was localized in neu- expressionsince both mRNAs are present in brain. ronal processesand, to a lesserextent, appeared to be on the In cortex, m,-, m,-, and m,-immunoreactivities were differ- surfaceof cell bodies.The m,-receptor was presentat high levels entially localized. The m,-immunoreactivity was presentin most in the anterior and intralaminar nuclei of the thalamus,anterior cortical neuronsand was particularly densein neuropil in layers pretectal nucleus, superior and inferior colliculi, pontine nuclei, II/III and VI (Fig. 2A.D); the laminar distribution of the protein many regions of the tegmentum, including pedunculopontine is consistentwith the cellular localization of m, mRNA (Buckley tegmental nucleus and laterodorsal tegmental nucleus, and all et al., 1988; Weiner and Brann, 1989). The m,-protein wasdense motor nuclei of the cranial nerves. in layer IV and the border of layers V/VI (Fig. 2B,E). Although initial studiesfailed to reveal m,-mRNA in cortex (Buckley et Discussion al., 1988) more sensitive methods have localized this subtype Muscarinic ACh receptors, like other neurotransmitter recep- to cell bodies in a pattern similar to the m,-protein (Weiner and tors, including nicotinic ACh, catecholamine, glutamate, and Brann, 1989); however, the m,-protein was associatedmostly GABA, are encoded by a family of genes.The high degreeof with fibers and presumptive terminals and only occasionally relatednesswithin each family has made it difficult, if not im- perikarya. The m,-immunoreactivity, considerably less dense possible, to use the small pharmacological differencesamong than the other subtypes in cortex, was localized in the neuropil the subtypes to identify the proteins in tissues.In the present of layers II/III and patchesin layer IV (Fig. 2C,F), and in scat- study, we have developed a panel of antibodies that bind the tered perikarya (mostly layer V). These findings are consistent five cloned muscarinic receptorswith virtually completesubtype with the localization of m,-mRNA (Buckley et al., 1988; Weiner specificity. These antibodies have provided an unprecedented and Brann, 1989). The perikaryal immunoreactivity was best opportunity to characterize a family of genetically defined neu- visualized with another antibody raised to the denatured form rotransmitter receptor proteins. Using immunoprecipitation, of m,i3 (Levey et al., 1990) (Fig. 3E). Regional differenceswith three receptor proteins, m,, m2, and ma, were shown to account these subtypes were also evident; for example, m,-immuno- for the vast majority of solubilized muscarinic receptor binding reactivity was more abundant in frontoparietal regions than sites in brain. The precisecellular and subcellular localization retrosplenial cortex, whereasthe converse was true of m2. Sim- of theseproteins by immunocytochemistry hasgreatly clarified ilarly, in hippocampus, m, was particularly densein CA1 and the roles of the muscarinic receptor subtypesin cholinergic func- dentate gyrus, and m, was more abundant in CA2-4. tion. The m,-, m2-, and m,-receptors were differentially localized The specificity of the antibodies hasbeen establishedby many in the striatum. Immunoreactivities for the m,- and m,-recep- independent criteria: (1) the regions of the i3 loops selectedfor tors were particularly denseand patchy in the neostriatum and the fusion proteins showvirtually no sequencehomology among nucleus accumbens(Fig. 2A.C). Discrete perikaryal localization muscarinic subtypes;(2) immunoblotting experiments have ver- of both receptors may have been obscured by the denseim- ified that the i3 loops are antigenically unrelated (Levey et al., munoreactivity associated with the neuropil. The m,-protein 1990); (3) eachantiserum bound a singlecloned receptor subtype was alsopresent in the substantianigra. In the olfactory tubercle, by immunoprecipitation; (4) the distribution of receptor pro- m, and m4 were present in the neuropil; m4 was particularly teins was similar in both immunoprecipitation and immuno- densein the islandsof Calleja. The m,-receptor was localized cytochemistry experiments; (5) antibody binding was blockable to scattered large neurons in caudate-putamenand nucleus ac- with the homologousi3 loop fusion proteins, but not the pa- /

C The Journal of Neuroscience, October 1991, 1 I(10) 3223 rental fusion protein (minus the i3 loop); (6) the localization of and Palacios, 1986; Mash and Potter, 1986; Spencer et al., 1986; the receptor proteins agreed with the localization of the respec- Regenold et al., 1987), has been extensively studied by ligand tive mRNAs (Buckley et al., 1988; Weiner et al., 1990; Vilaro autoradiography. Recently, other subtypes have been identified et al., 1991), although, as expected, there were important dif- in brain homogenates using pharmacological or kinetic analyses ferences in the subcellular distribution of proteins and mRNAs; (Doods et al., 1987; Ehlert and Tran, 1990; Waelbroeck et al., and (7) receptor localization (m,) was very similar using anti- 1990). The relation of muscarinic receptor binding sites, defined peptide antibodies directed to a completely different region of by various ligands, to the genetically distinct receptor proteins the protein (A. I. Levey, W. F. Simonds, and M. R. Brann, is complicated and uncertain (Peralta et al., 1987; Akiba et al., unpublished observations). Thus, the rigorous evaluation of the 1988; Buckley et al., 1989; Dorje et al., 1991). For example, antibodies provides a firm basis for analyzing the native receptor Ml-selective ligands, such as pirenzepine, bind with high affinity subtypes. to m,- and ma-cloned receptors, and with intermediate affinity Immunological characterization of the muscarinic receptor to m3 and m,. M,-selective ligands, such as AFDX-116, also proteins complements and extends previous in situ hybridiza- bind with high affinity to multiple subtypes (rn? and m,). There- tion studies (Buckley et al., 1988; Weiner and Brann, 1989; fore, the binding sites localized using these and other selective Weiner et al., 1990). Since mRNAs are localized primarily to compounds probably correspond to a minimum of two receptor cell bodies and proximal dendrites, in situ hybridization studies proteins, although the exact composition would be expected to provide clear evidence for the distribution and identity of neu- vary depending on the receptors expressed in the tissue and the rons that synthesize the receptor subtypes. On the other hand, binding conditions. Because of the vast amount of information this approach yields no information about the abundance or obtained using these pharmacological approaches, it is impor- subcellular distribution of the receptors; these considerations tant to compare the distributions of the pharmacologically de- are important because of possible differences in translational fined binding sites with the localization of the gene products, efficiencies, turnover, and transport of the proteins. The im- as summarized in Table 3. munological methods we have developed provide direct infor- M, receptors, the most abundant pharmacological subtype in mation about the relative abundance and precise localization cortex and striatum (Cortes and Palacios, 1986; Mash and Pot- of the subtypes. As summarized in Table 3, for most of the ter, 1986; Spencer et al., 1986), are a composite of the m,- and subtypes there is an excellent correspondence between the dis- m,-proteins; however, the differential cellular localizations of tributions of mRNA and protein. However, two of the mus- m, and m4 suggest that the receptors have important functional carinic receptor proteins, m, and m,, whose mRNAs have been differences. Both proteins are abundant in these regions (al- identified in brain (Buckley et al., 1988; Weiner and Brann, though m, is relatively higher in cortex than striatum and m4 1989; Weiner et al., 1990), were not reliably detected in this is relatively higher in striatum than cortex), and both have an investigation. These data suggest that m,- and m,-proteins may M,-like pharmacology. Moreover, Luthin and Wolfe (Luthin et be expressed at very low levels in brain. The subcellular distri- al., 1988) have shown that pirenzepine (M,) binding sites in bution of the receptors also appears to be highly regulated. In forebrain are immunoprecipitated by anti-peptide antibodies some regions, the proteins were clearly localized to perikarya against m, and that other unidentified receptor proteins are also (e.g., m, in cortical neurons and m, in striatum), and as described labeled by this ligand. Using the same strategy, we have im- in Results, these findings were consistent with the distributions munoprecipitated 3H-pirenzepine sites and recovered the ligand of the mRNAs. However, in other instances (e.g., m, in cortex with both m,- and m,-receptors (A. I. Levey, unpublished ob- and m4 in substantia nigra), the proteins were predominantly servations). In cortex, activation of M, receptors leads to phos- localized in neurites, fibers, or presumptive terminals. In these phatidylinositol hydrolysis (Brown et al., 1984) and inhibition cases, knowledge of the distribution of mRNA helps interpret of the M-current (McCormick and Prince, 1985). The prefer- the cellular origin of the subtypes; for example, m,-protein in ential coupling of the cloned ml-protein, but not m4, to these cortex might be derived from either cortical or basal forebrain signal transduction pathways, together with the relative abun- cells that express m,-mRNA. Similarly, m4 is possibly associated dance of m, in cortex, suggests that m,-protein accounts for with striatonigral terminals, since the m,-mRNA is present in these effects of cortical M,-receptor activation. The immuno- medium-sized striatal (projection) neurons and not in substantia cytochemical localization of m,, including high levels in super- nigra (Weiner et al., 1990). Ultrastructural and lesion analysis ficial laminae of cortex, also matches the pattern of M,-binding will be valuable to evaluate further the compartmentation of sites (Mash and Potter, 1986; Spencer et al., 1986). The role of the receptors, including synaptic details. the m,-receptor in cortical function, which is localized primarily The localization of muscarinic receptors in brain (Yamamura in the neuropil of superficial layers of cortex and perikarya in and Snyder, 1974; Kuhar and Yamamura, 1975), including M,- layer V, is presently unclear. Since anticholinergics cause cog- and M,-pharmacological subtypes (Wamsley et al., 1984; Cortes nitive dysfunction in humans (Drachman and Leavitt, 1974) c Figure 2. Immunocytochemical localization of m,, m2, and m4 receptor proteins in rat cortex, striatum, and basal forebrain. Shown are coronal sections through the forebrain demonstrating muscarinic receptor immunoreactivity using affinity-purified antibodies reactive with the i3 loops of m, (A), m, (B), and m4 (C’). Note the differential distribution of the three receptors in cortex (ctx), caudate-putamen (cp), and olfactory tubercle. The m2 protein is particularly dense in the medial septum (ms), where many cholinergic neurons are located. The arrow points to m4 receptor in an island of Calleja; UC represents anterior commissure. Higher magnification of parietal cortex demonstrates localization of the three receptors (D, m,; E, m,; F, m,) in cortical laminae as marked. Immunological specificity was confirmed by complete inhibition of staining after preadsorption of the antibodies on the respective fusion proteins (D’, m,i3; E’, m,3; F’, m,i3). Scale bars: A-C, 500 pm; and D-F, 220 pm.

The Journal of Neuroscience, October 1991, 17(10) 3225

Table 3. Comparison of muscarinic receptor subtype proteins (immunocytochemistry), mRNAs (in siiu hybridization), and pharmacological - binding sites (autoradiography) in select regions of rat brain

Brain region Immunocytochemistry In situ hybridization Autoradiography m, m, m, m, m2 m, m4 m, M, M, Neocortex +++ ++ + +++ ++ ++ + - +++ ++ Hippocampus +++ ++ + +++ + ++ ++ + +++ ++ Basalforebrain - +++ - - +++ + + - - +++ Striatum +++ + +++ +++ ++ - +++ - +++ + Substantianigra - + + - + - - ++ + + Amygdala ++ + + +++ + + + + ++ + Thalamus + +++ + + +++ ++ + + + ++ Motor neurons - +++ - - +++ - - - - t++

Shown are the relative densities (-, undetectable; +, low; + +, moderate; + + +, high) of the muscarinic receptor proteins (present results), mRNAs (Buckley et al., 1988; Weiner and Brann, 1989; Weiner et al., 1990; Vilaro et al., 1991; D. M. Weiner, A. I. Levey, and M. R. Brann, unpublished observations), and binding sites (Cartes and Palacios, 1986; Mash and Potter, 1986; Spencer et al., 1986). The range of densities represent general differencesin staining intensity among the brain regions for each subtype. Becauseof the nonquantitative methods used to visualize the receptors, comparison among receptors (i.e., between columns) should be made with caution. and in particular, M,-selective antagonistsimpair memory in ceptor protein, as previously describedfor the mRNA (Buckley animals (Messer et al., 1990), the m,- and m,-proteins must be et al., 1988), is present in many noncholinergic cells and fibers consideredas candidates for mediating theseeffects. Similarly, in cortex, pontine nuclei, and cerebellum. Therefore, at these either or both of these receptor proteins in striatum may be sites the m2 may function either postsynaptically or on non- responsiblefor the therapeutically important effects of M,-se- cholinergic nerve terminals to modulate releaseof other neu- lective antagonists, such as trihexiphenidyl, in patients with rotransmitters. Parkinson’sdisease and other movement disorders. Therefore, In conclusion, this investigation demonstratesthe distribu- drugs targeted to only m, or m4 have considerablepotential for tions of the three most abundant muscarinic receptor subtype more effectively and specifically treating patients with Alzhei- proteins in rat brain. Our combined molecular and immuno- mer’s diseaseand Parkinson’s disease. logical approach has allowed the selective localization of mus- The localization of m,-protein provides very strong evidence carinic receptor subtypes;this has not been possiblewith avail- for the role of this subtype as an autoreceptor. Although mus- able pharmacological methods. Immunocytochemistry also carinic binding siteswith low affinity for pirenzepine (M,) were allows a much higher degreeof cellular resolution than possible previously shown to be associatedwith cholinergic neuronsand with receptor autoradiography and can be extended to the ul- projections (Mash et al., 1985; Mash and Potter, 1986; Spencer trastructural level. Knowledge of the distributions of thesere- et al., 1986) and to inhibit ACh releasein cortex and hippo- ceptors has helped to clarify the functional properties of genet- campus(Raiteri et al., 1984; Meyer and Otero, 1985; Quirion ically defined subtypes in neuronal circuits in brain. This et al., 1989; Marchi et al., 1990), the genetic subtype responsible information, together with the development of more selective for this important function has not been well defined. At least muscarinic drugs, should lead to more rational and effective two receptor proteins, m2and m4,bind Ml-selective ligandswith treatments of many neurological and psychiatric diseasesas- similar affinities (Buckley et al., 1989; Dorje et al., 1991); both sociated with dysfunction of cholinergic systems. proteins also have common signal transduction mechanisms that may be important in regulation of neurotransmitter release References (Higashida et al., 1990). Moreover, the pharmacology of the Akiba I, Kubo T, MaedaA, Bujo H, Nakai J, Mishina M, NumaS autoreceptors has been variously describedas M, (Raiteri et al., (1988) Primarystructure of porcinemuscarinic acetylcholine receptor III and antagonistbinding studies. FEBS Lett 235:257-26I. 1984; Meyer and Otero, 1985; Quirion et al., 1989) or M, (Mar- BonnerTI, BuckleyNJ, Young AC, BrannMR (1987) Identification chi et al., 1990). In the present study, m,-receptor protein was of a family of muscarinicacetylcholine receptor genes. Science 237: associatedwith central choline& systems(Mesulam et al., 1983) 527-532. in basalforebrain, striatum, mesopontine tegmentum, and cra- BonnerTI, Young AC, Brann MR, Buckley NJ (1988) Cloningand expression ofthe human and rat m5 muscarinic acetylcholine receptor nial motor neurons; theseresults correlate extremely well with genes.Neuron 1:4034 10. the expression of m,-mRNA (Buckley et al., 1988) and M,- Brown E, Kendall DA, Nahorski SR (1984) Inositol phospholipid binding sites (Mash and Potter, 1986; Spenceret al., 1986). It hydrolysis in rat cerebral cortical slices. 1. Receptor characterization. is also important to note that our results indicate that m,-re- J Neurochem 42: 1379-l 387. t Figure 3. Examples of muscarinic receptor subtype immunoreactivity in various regions of rat brain. A parasagittal section through the rat thalamus, brainstem, and cerebellum is shown in A. The m, receptor was the predominant subtype localized in these regions, in agreement with immunoprecipitation studies (Table 2). Note the intense immunoreactivity in the hippocampus (hi), anteroventral nucleus of the thalamus (t/z), anterior pretectal nucleus (pt), superior (SC) and inferior (ic) colliculi, pontine nuclei @n), cerebellum (cb), and trigeminal (q and facial (VII) motor nuclei. Examples of cellular localization of m, receptor are shown in the striatum (B) and basal forebrain (0). The receptor appears to be localized predominantly in the perikaryal and dendritic membranes and, in addition, is present in the neuropil in both regions. The m, receptor (C) is present in many neurons in cortex. The m4 receptor (E) appears localized at the membrane surface (arrow) of scattered neurons in cortex (predominantly layer V) using an antibody reactive with denatured m,i3 (Levey et al., 1990) and silver intensified (Kitt et al., 1984). Scale bars: A, 350 pm; B, 14 Frn; Cand E, 16 pm; D, 18 pm. 3226 Levey et al. * Muscarinic Receptor Subtype Proteins in Brain

Buckley NJ, Bonner TI, Brann MR (1988) Localization of a family of Mash DC, Potter LT (1986) Autoradiographic localization of M 1 and muscarinic receptor mRNAs in rat brain. J Neurosci 8:46464652. M2 muscarine receptors in the rat brain. Neuroscience 19:551-564. Buckley NJ, Bonner TI, Buckley CM, Brann MR (1989) Antagonist Mash DC, Flynn DD, Potter LT (1985) Loss of M2 muscat-me recep- binding properties of five cloned muscarinic receptors expressed in tors in the cerebral cortex in Alzheimer’s disease and experimental CHO-Kl cells. Mol Pharmacol 35469-476. cholinergic denervation. Science 228: 1115-l 117. Cortes R, Palascios JM (1986) Muscarinic cholinergic receptor sub- McCormick DA, Prince DA (1985) Two types of muscarinic response types in the rat brain. I. Quantitative autoradiographic studies. Brain to acetylcholine in mammalian cortical neurons. Proc Nat1 Acad Sci Res 3621227-238. USA 82:6344-6348. Doods HN, Mathy MJ, Davidesko D, van Charldorp KJ, de Jonge A, Messer WS, Bohnett M, Stibbe J (1990) Evidence for a preferential van Zwieten PA (1987) Selectivity of muscarinic antagonists in involvement of M 1 muscarinic receptors in representational memory. radioligand and in vivo experiments for the putative M 1, M2 and M3 Neurosci Lett 116:1&t-189. receptors. J Pharmacol Exp Ther 242:257-262. Mesulam MM, Mufson EJ, Wainer BH, Levey AI (1983) Central Dorje F, Wess J, Lambrecht G, Tacke R, Mutschler E, Brann MR cholinergic pathways in the rat: an overview based on an alternative (199 1) Antagonist binding profiles of five cloned human muscarinic nomenclature (Ch l-Ch6). Neuroscience 10: 1185-l 20 1. receptor subtypes. J Pharmacol Exp Ther 256:727-733. Meyer EM, Otero DH (1985) Pharmacological and ionic characteri- Dorje F, Levey AI, Brann MR (in press) Immunological detection of zations of the muscarinic receptors modulating [3H]acetylcholine re- muscarinic receptor subtype proteins (ml-m5) in rabbit peripheral lease from rat cortical synaptdsomes. J Neurosci 5-l 202-l 207. tissues. Mol Pharmacol, in press. Peralta EG. Ashkenazi A. Winslow JW. Smith DH. Ramachandran J. Drachman DA, Leavitt JL (1974) Human memory and the cholinergic Capon DJ (1987) Distinct primary structures, l&and-binding prop- system. A relationship to aging? Arch Neurol 30: 113-l 2 1. erties and tissue-specific expression of four human muscarinic ace- Ehlert FJ, Tran LLP (1990) Regional distribution of M 1, M2, and tylcholine receptors. EMBO J 6:3923-3929. non-M 1, non-M2 subtypes of muscarinic binding sites in rat brain. Peralta EG, Ashkenazi A, Winslow JW, Ramachandran J, Capon DJ J Pharmacol Exp Therm225: 1148-l 157. (1988) Differential regulation of PI hydrolysis and adenylyl cyclase Fukuda K. Hiaashida H. Kubo T. Maeda A, Akiba I. Buio H. Mishina by muscarinic receptor subtypes. Nature 334:434-437. M, Numa S I( 1988) Selective coupling with K+ currentsofmuscarinic Quirion R, Aubert I, Lapchak PA, Schaum RP, Teolis S, Gauthier S, acetylcholine receptor subtypes in NG108-15 cells. Nature 335:355- Araujo DM (1989) Muscarinic receptor subtypes in human neu- 358. rodegenerative disorders: focus on Alzheimer’s disease. Trends Phar- Hammer R, Berrie CP, Birdsall NJM, Burgen ASV, Hulme EC (1980) macol Sci [Suppl] 10:80-84. Pirenzepine distinguishes between subclasses of muscarinic receptors. Raiteri M, Leardi R, Marchi M (1984) Heterogeneity of presynaptic Nature 283:90-92. muscarinic receptors regulating neurotransmitter release in the rat Harlow D, Lane D (1988) Antibodies: a laboratory manual. Cold brain. J Pharmacol EXD Ther 228:209-2 14. Spring Harbor, NY: Cold Spring Harbor Laboratory. Regenold W, Arujo D, Q&ion R (1987) Direct visualization of brain Higashida H, Hashii M, Fukuda K, Caulfield MP, Brown DA (1990) M2 muscarinic receptors using the selective antagonist [‘H]AF-DX 116. Selective coupling of different muscarinic acetylcholine receptors to Eur J Pharmacol 144:417-419. neuronal calcium currents in DNA-transfected cells. Proc R Sot Lond Sambrook J, Fritsch EF, Maniatas T (1989) Molecular cloning: a [Biol] 242:68-74. laboratory manual. Cold Spring Harbor, NY: Cold Spring Harbor Hulme EC, Birdsall NJM, Buckley NJ (1990) Muscarinic receptor Press. subtypes. Annu Rev Pharmacol Toxic01 30:633-73. Smith DB, Johnson KS (1988) Single-step purification of polypeptides Jones SV, Barker JL, Bonner TI, Buckley NJ, Brann MR (1988,) Elec- expressed in Escherichia colias fusions with glutathione S-transferase. trophysiological characterization of cloned m 1 muscarinic receptors Gene 67:3140. expressed in A9 L cells. Proc Nat1 Acad Sci USA 85:4056-4060. Spencer DG Jr, Horvath E, Traber J (1986) Direct autoradiographic Jones SV, Barker JL, Buckley NJ, Bonner TI, Collins RM, Brann MR determination of M 1 and M2 muscarinic acetylcholine receptor dis- (1988b) Cloned muscarinic receptor subtypes expressed in A9 L cells tribution in rat brain: relation to cholinergic nuclei and projections. differ in their couolina to electrical resoonses. Mol Pharmacol 34: Brain Res 380:59-68. 421426. - - Vilaro MT, Wiederhold K-H, Palacios JM, Mengod G (1991) Mus- Kitt CA, Price DL, Struble RG, Cork LC, Wainer BH, Mobley WC carinic choline@ receptors in the rat caudate-putamen and olfactory f 1984) Evidence for cholineraic neurites in senile _olaaues. _ Science tubercle belong predominantly to the m4 class: in situ hybridization 22611443-1445. and receptor autoradiography. Neuroscience 40: 159-l 67. Kraft RJ, Tardoff KS, Krauter KS, Leinwand LA (1988) Using mini- Waelbroeck M, Tastenoy M, Camus J, Christophe J (1990) Binding prep plasmid DNA for sequencing double-stranded templates with of selective antagonists to 4 muscarinic receptors (M 1 to M4) in rat Sequenase. Biotechniques 6:544-546. forebrain. Mol Pharmacol 38:267-273. Kubo T, Fukuda K, Mikami A, Maeda A, Takahashi H, Mishina M, Wamsley JK, Gehlert DR, Roeske WR, Yamamura HI (1984) Mus- Haga T, Haga K, Ichiyama A, Kangawa K, Kojima M, Matsuo H, carinic antagonist binding site heterogeneity as evidenced by auto- Hirose T, Numa S (1986) Cloning, sequencing and expression of radiography after direct labeling with [3H]( -)QNB and [)H]pirenzepine. complementary DNA encoding the muscarinic acetylcholine receptor. Life Sci 34: 1395-1402. Nature 323:411-416. Weiner DM, Brann MR (1989) Distribution of ml-m5 muscarinic Kuhar MJ, Yamamura HI (1975) Light microscopic autoradiographic receptor mRNAs in rat brain. Trends Pharmacol Sci [Suppl] 10: 115. localization of muscarinic receptors in rat brain by specific binding Weiner DM, Levey AI, Brann MR (1990) Expression of muscarinic of a potent antagonist. Nature 253:560-56 1. acetylcholine and dopamine receptor mRNAs in rat basal ganglia. Levey AI, Stormann TM, Brann MR (1990) Bacterial expression of Proc Nat1 Acad Sci USA 87:7050-7054. human muscarinic receptor fusion proteins and generation of subtype- Wess J, Bonner TI, Dorje F, Brann MR (1990) Delineation of mus- specific antisera. FEBS Lett 275:65-69. carinic receptor domains conferring selectivity of coupling to guanine Luthin GR. Harkness J. Artvmvshvn RP. Wolfe BB (1988) Antibodies nucleotide-binding proteins and 2nd messengers. Mol Pharmacol 38: to a synthetic peptide can be’used to distinguish between muscarinic 5 17-523. acetylcholine receptor binding sites in brain and heart. Mol Pharmacol Yamamura HI, Snyder SH (1974) Muscarinic cholinergic binding in 341327-333. rat brain. Proc Nat1 Acad Sci USA 7 1: 1725-l 729. Marchi M, Ruelle A, Andrioli GC, Raiteri M (1990) Pirenzepine- insensitive muscarinic autoreceptors regulate acetylcholine release in human neocortex. Brain Res 520:347-350.