Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Distinct primary structures, ligand‐binding properties and tissue‐specific expression of four human muscarinic acetylcholine receptors.

Distinct primary structures, ligand‐binding properties and tissue‐specific expression of four... The EMBO Journal vol.6 no.13 pp.3923 -3929, 1987 Distinct primary structures, ligand-binding properties and tissue- specific expression of four human muscarinic acetylcholine receptors Ernest was initially discerned through the discovery of selective muscar- G.Peralta, Avi Ashkenazi'2, John W.Winslowl, inic agents which exhibit different Douglas H.Smith, J.Ramachandran"2 and affinities for receptors in Daniel various tissues (Hammer et al., 1980; Birdsall and Hulme, 1983; J.Capon Cohen and Sokolovsky, 1987). Studies with selective antagonists Departments of Molecular Biology and 'Developmental Biology, Genentech, such as pirenzepine, which discriminates between the high affinity Inc., 460 Point San Bruno Boulevard, South San Francisco, CA 94080 and MI receptors in certain regions of the brain and the low affinity 2Department of Biochemistry and Biophysics, University of California, San Francisco, CA 94143, USA M2 receptors of heart, have permitted tentative proposals of func- tional differences between subtypes in their abilities to regulate Communicated by M.Birnstiel specific biochemical effector systems (Gil and Wolfe, 1985; To investigate the molecular basis for the diversity in mus- Brown et al., 1985). Molecular cloning of the porcine cerebral carinic cholinergic function, we have isolated the genes encod- (MI) and atrial (M2) mAChRs has established that the major ing the human Ml and M2 muscarinic receptors (mAChR) receptor subtypes of brain and heart are indeed different poly- as well as two previously undiscovered mAChR subtypes, peptides encoded by distinct genes which share 38% overall designated HiM3 and HM4. The amino acid sequence of each amino acid homology (Kubo et al., 1986a,b; Peralta et al., 1987). reflects a structure of subtype consisting seven, highly con- Evidence for the existence of previously uncharacterized genes served transmembrane a segments and large intracellular encoding novel mAChR subtypes has been provided by studies region unique to each subtype, which may constitute the with other selective muscarinic compounds (Giraldo et al., 1987; ligand-binding and effector-coupling domains respectively. Korc et al., 1987), as well as hybridization analysis of mam- differences in for Significant affinity muscarinc ligands were malian genomic DNAs which revealed additional sequences hom- detected in individual mAChR subtypes produced by trans- ologous to but distinct from the M1 and M2 mAChR genes fection of mammalian cells. Each exhibited subtype multiple (Peralta et al., 1987). To provide a basis for investigating mAChR states for differences in the affinity agonists; among subtypes functional diversity, we have isolated the genes for the human affinities and of such sites the proportions suggest capacity HM1 and HM2 mAChRs, as well as two novel mAChR sub- to with of mAChR subtypes interact differentially the cellu- types, herein designated HM3 and HM4. Characterization of the mRNA lar effector-coupling apparatus. Subtype-specific ex- primary structures of these four human mAChR subtypes, their in and a pression was observed the heart, pancreas neuronal agonist and antagonist binding properties, and their patterns of cell that the of mAChR ex- line, indicating regulation gene expression in various tissues reveals that all four mAChRs indeed pression contributes to the differentiation of cholinergic ac- represent functional proteins encoded by distinct, actively tivity. transcribed genes, and suggests unique biological functions for Key words: muscarinic receptors/receptor subtypes/ligand bind- each receptor subtype. ing/gene regulation Results and discussion Molecular human Introduction cloning of muscarinic receptors reveals two novel subtypes transmission muscarinic Synaptic by acetylcholine receptors To isolate the genes encoding human mAChRs, we employed is the central and (mAChRs) employed throughout peripheral a based strategy upon sequences conserved between the porcine nervous to elicit a and diverse of neuro- systems large array MI and M2 mAChR subtypes, as well as similarities with other actions. In the autonomic nervous mAChRs physiological system known receptors which transduce signals through interaction with the force and rate of heart muscle the regulate contraction, nucleotide guanine binding (G) proteins. Since the protein coding of smooth and the of contraction muscle, secretory activity of the M2 mAChR and hamster sequences 3-adrenergic receptor which receive innervation glands parasympathetic (Taylor, 1985). are each contained within a exon et single (Dixon al., 1986; mAChRs also constitute the of majority cholinergic synapses Peralta et a was screened for to be al., 1987), genomic library receptor within the central nevous where are system, they thought clones with the that human mAChR would also as and expectation genes to such neural important processes learning memory lack introns within their coding regions. The and of mAChR functional porcine M1 An (Nathanson, 1987). important aspect M2 mAChRs exhibit the greatest degree of amino acid identity the multitude of biochemical and elec- is reflected diversity by in first and their five transmembrane domains connecting loops actions evoked to trophysiological by acetylcholine binding et therefore low (Peralta al., 1987); stringency hybridizations include the of intracellular levels of which mAChRs, regulation were carried out with a this to fragment containing region identify and inositol and the or cGMP cAMP, phospholipids, opening related human sequences. In addition, the MI and M2 mAChRs of the calcium and ion channels found chloride closing potassium, as as to share the most to each other well similarity ,3-adrenergic tissues Given its role in certain (Nathanson, 1987). ubiquitous in their first receptors and the visual rhodopsins a in the cytoplasmic loop in neuronal signal transduction, major question study transmembrane a site of G and second domain, function relates to the molecular mechanisms which possible protein of cholinergic et Dixon et Nathans et interaction (Zuker al., 1985; of muscarinic al., 1986; the underly biological specialization activity. Yarden et Peralta et To isolate between mAChRs al., 1986; al., 1986; al., 1987). for structural The potential heterogeneity Oxford, IRL Press Limited, England E.G.Peralta et al. H2N- * HM1 1 - - - - - - - - - - - - - - I- - - - - - - - - - - - - - - - - - - - - - - - - - - - MNTSAPP M TL H*NANST TS PL FPANI SS SW IH SP SD AG LP PG TV TH FG S YN S R AA F s HM4 MNGSS as1~ HM3 1-~~~~~~~~~~~~~-MANFTP - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --MANFTP H - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - M NNSTN HM2 1-21 HM1 8A V S P T V L APGKPWVAFIGI TGLLSLATTL S FKVNTE HM4 5 1 SP D G[T D D P L GH TVWOVVFIAFLTGILF;]L V T I I GN I L VS F KV NO L K HM3 16 V R L VS S S H NRYETV V F ATVTGSLSLVTFV VG N I L V S IK V N R O L HM2 7 [ S N SL A L T S PY K TF E VV L GSLSLVTIIGNILVMVSL I GLES M N L Y T Y M HH WA L G T L A C D L WL A L D Y V F IL|SL A A DL 8 T N NY HM1 L A C A D L I I WL AD Y V T N Y F LLS G I MHWAL GNL ACDL V N S MN L TTY I HM1 HM04 10 T V N N Y FLLSLACADLI I GSMS MNLTYII KGYW_LGNACDLWLALDYV 0013 65 TVNNYFLFSLACADLIIGAFSMNLYTVyIIGY4I LGIAVVCDLWLALDYV 0M2 56 TVNNYFLFSLACADLII G V F SMNLYTTL V I GY LPVVVCDLWLALDYV 13-41 V HM1 108 ASNASVMNLLLI SFDRYFSVTFRDPL SYRTAKRTPPR R|AKARLLM II G LIA UL V|S F V L U HM4 151 ASNASVMNLLVI S F D R Y FSITRP L T YRA K R T T KIRA GVMI GLA WVIS F V L U HM3 115 VSNASVMNLL I I SFDRYFCVTKPLTYPART T K MA GL MI AAAWV L S F V L U HM2 106LVSNASVMNLL ISFDRYFCVTKPLTYPKR T T KMAGMMIAAAWVLSFnIL- 4-5o v HM1 15s T MLAGQCYIQFLSOPIIT FGTAMA A F Y L P V TV TL APAILFWFOIYLVGERL 00144 201 A P A I L F V G KRTVP G E CF I F LSEPTT F G T A I A A F YP V T I MTIL WOLYF P A I L F R T V P_ a 0013 165 APAILFWQFVVGKRTVPDNQCFIOFLSNPAVTFGTAIAAFYLPVVIMTVL FLSNPAVTFGTAIAAFY|PVVIM D CJ WQFVVGK HM3 2051A HM2 156 AA P ILFWOFIVGROFFSNA GTAIAAFYLPVII MT V L J5 -61 HM1 208 YWR I Y RETEN R ARELAL GSETPGGGG - - - - - - - - - S S R P G A HM4 251 YWRIWR IRYTEKT LAIG L0A N7,HPTG SHSSRSCSSY L aLIQ S M E[AfEMTE HM3 215 Yl SLARS RI HKHVHKHRP[E]GPK[KKIAK[TILA KSJLKSPL MOSQ1VKKPP GEAAR HM2 206 Y I S RAS K EKDKK PVANODPVSP SL G R I PNNNNM- S D D G 001 249 []G P E T P PGRCCCCRA[RLLQAY- - - - - - - - -IWKE E E E E DEG[M[S L T HM4 301 K R NRRKYRC HFWFTEKSWKP SS[QMDQDHSSSDSWNNNDAAASLENSA HM3 265 L R N GKL E PPTALPIPPP VA D T ESSSGS -T Q[;TKERPA REL HM2 255 O ANPRKD P VEE NCVQGEEG S S DSV V A S MR DDE I W D LEJHRKI HM1 290 E G E P VV I KMP- - - - - - - - DI HM4 351 SSXDE I I TRAIYSIVLKL GHSTILN[TKLPSSD LOVPEEELGMVD HM3 314 TT EA T PAMPAPPLQPRALN ASRWSKIQIVT KOQGNE AIEI IP - HM2 305 E NTV STS LGHSKDE NFsK Q T C I R I GT KT PK2DSCTP N ----WTVEVVLGS - HM1 309 P[AO QP T K P P R S P N - - - - - - - - - - - - - - - .-. - - - - - - - - - - - - - - - - HM4 401 LW RK DKL AQK SVDDGGSFPKSFSKLPI QLESAVDTAKTSDVNSSVGKS HM3 0012 350 VI HM1 325 R TKKGRDRAGKGQKPRGKEOL T F KEKKAARTLSAILLAF HM4 451 TALPL sF E3EAT K L KFRSQ K KR S L EKKA TLSAILLAF HM3 362 - KSI RRNOQVKKQ RRTRI ]EIKVT IFAI L L A F PA[MRPALAJNVAR 0M2 350 -S GQN[DEE3QNIVARK K MTPKL- PAKKK P P S REKKVTRTILAI L A F VIl 16-70 375 |IL TWTPYNI MVLV KEE[L EILCGY WLCYLNSTINP|M|CYALCN|KB HM1 HM4 501 IITWTPYNIMVLVNTF DSCIDCIPKTF!NLGYWL CYI NSTIN P CY A L CNT HM3 410 I| TWTPYNVMVLVNTFCQSCIPDITVUASI G Y WL CYFINSTINPACYAL C N A T 397 IIT AACPYN A IPNT V TII GYWLCYINSTINPACYAL C N A T HM2 -COOH HM1 425 RDTR DKRRw I PKRPGSV HRTPS OC HM4 551 FRTTFK LLCQCDKKKRR QY QYQORQSVIFHK APEQAL HM3 460 F K K T FH L LLCOYRNI TAIR 447 F K K T F K H L LMCHYKNI GAT HM2 Fig. 1. Primary sequence homology between the four human muscarinic receptor subtypes. Since maximum homology occurs between HM1 and HM4 or HM3 and HM2, the sequence alignment is presented in this order. Gaps were introduced to maximize significant sequence identity and to compensate for the large size of the cytoplasmic loop 5-6 of HM4. Identities are indicated by the boxed amino acid residues. Asterisks denote potential N-linked glycosylation sites in the amino-terminal domains. Regions I-VII are putative transmembrane domains (Peralta et al., 1987) and are indicated by lines over the hydrophobic sequences. The hydrophilic loops are numbered according to the transmembrane regions which they connect and their orientations are indicated with respect to the extracellular (outer, o) or cytoplasmic (inner, i) side of the plasma membrane. The amino-terminal region is indicated by H2N- and is postulated to be extracellular, while the carboxy-terminal region is indicated by -COOH and is predicted to reside on the cytoplasmic side of the plasma membrane. Numbers at the left denote the position within each protein sequence of the first residue in each line. 3924 Human muscarinic acetycholine receptors In contrast to the virtual identity between the human and por- clones with this highly specific local identity, candidates were cine homologues of the Ml or M2 mAChRs, which suggests then characterized for their ability to hybridize with two sets of severe functional constraints on the structures of these pairs of oligonucleotide probes based upon the corresponding porcine M2 subtypes, all four human subtypes diverge dramatically in the mAChR sequences (amino acids 55-62 and 64-70, detailed in large loop connecting transmembrane domains 5 and 6 (Figure Materials and methods) (Peralta et al., 1987). Eighteen of 23 1). The length of this region differs significantly between the four clones hybridizing with the M2 mAChR restriction fragment were mAChR subtypes, consisting of 156, 181, 184 and 241 residues also found positive with both sets of oligonucleotides, and could in HM 1, HM2, HM3 and HM4 respectively. By analogy with be divided into four classes based on their ability to hybridize rhodopsin, the 5-6 loop region should reside on the cytoplasmic preferentially with individual members of each pool of side of the plasma membrane, suggesting that the exceptional oligonucleotides. Since the porcine Ml and M2 mAChRs are degree of divergence between mAChR subtypes in this region strikingly dissimilar in their large third cytoplasmic loop (Peralta may be related to their capacity for differential coupling to distinct et al., 1987), probes consisting of these subtype-unique sequences biochemical effectors or ion channels (Nathanson, 1987). Limited were used to identify one class of the human clones as the hom- ologue of the porcine cerebral Ml mAChR, and a second class homology occurs in the 5-6 loop region adjacent to the trans- membrane domains, as well as the serine, threonine and acidic as the homologue of the porcine atrial M2 mAChR. Nucleotide sequence analysis of the MI homologue predicts a polypeptide residue-rich middle region of the loop, between HM1 and HM4 of 460 amino acids sharing 98.9% identity with the porcine Ml and between HM2 and HM3 (21%), suggesting greater (22%) mAChR (Kubo et al., 1986a), while the M2 homologue sequence functional similarity between these pairs of mAChR subtypes reveals a protein of 466 amino acids with 97.4 % identity to the 1). Southern blot analysis of genomic DNA utilizing (Figure derived from the 5-6 loop of each subtype indicates that porcine M2 mAChR subtype (Peralta et al., 1987) (Figure 1). probes the human does not encode other sequences highly hom- The virtual sequence identity between each pair of human and genome mAChR subtypes, together with their ligand bind- to these unique mAChR domains (data not shown). porcine ologous ing characteristics (see below), suggests that these clones encode Limited amino acid conservation is also evident in tue amino- terminal of the four subtypes, which vary in length from the human Ml and M2 mAChR subtypes, thus designated HM1 regions and HM2 respectively. HMl and HM2, like the porcine Ml and 22 to 67 residues. However, a conserved feature of the amino terminus of all four subtypes is the presence of two to five poten- M2 mAChR subtypes, share 39% overall amino acid identity. tial sites for N-linked glycosylation (Figure 1), a feature common Sequence analysis of the remaining two classes of human clones to rhodopsins and the 3-adrenergic receptor (Zuker et al., 1985; with homology to the porcine mAChRs revealed that each class Dixon et al., 1986; Nathans et al., 1986; Yarden et al., 1986). encodes a polypeptide highly homologous to, yet quite distinct The high degree of native mAChR glycosylation (Nathanson, from the HM1 and HM2 subtypes (Figure 1). The first of these, 1987), as well as an absence of such sites in the other predicted designated HM3, predicts a protein of 479 amino acids and Mr extracellular regions of each subtype, suggest the functional 53 049 sharing 40 and 55% overall amino acid identity with HM1 of these otherwise unrelated amino-terminal sequences. similarity and HM2 respectively. The other novel class of clones, designated encodes a significantly larger protein of 590 amino acids HM4, The four human muscarinic receptor subtypes exhibit distinct and 66 127 displaying 43, 35 and 37% overall identity with M, binding properties antagonist HM2 and HM3 respectively (Figure 1). Analysis ofcDNA HM1, that the molecular clones isolated in this study encode clones corresponding to each class of mAChR indicates that all To confirm and to determine whether different human four human mAChR subtypes lack introns in their respective functional mAChRs, exhibit distinct pharmacologic characteristics coding regions (unpublished data). Accuracy of the initiation and mAChR subtypes in a similar cellular context, a transient mam- termination codon assignments shown in Figure 1 is further when presented et al., 1986) was utilized to indicated by the ability to express muscarinic ligand-binding malian expression system (Eaton of each to bind muscarinic compounds. utilizing these predicted signals (see below). study the ability subtype activity carried out with intact trans- Saturation binding experiments human muscarinic receptor contains a large Each or cell using the muscarinic antagonist fected cells homogenates domain unique to each subtype cytoplasmic benzilate revealed that all four human [3H]quinuclidinyl (QNB) the method of Kyte and Doolittle analysis by Hydropathicity encode polypeptides which specifically bind this mAChR genes each of the four human mAChRs contains indicates that (1982) with affinity (Table I). The apparent dissoci- radioligand high seven hydrophobic, potential transmembrane domains, suggesting constants determined by Scatchard analysis of the ation (KD) that all four subtypes adopt a structure within the plasma mem- from 22.8 to 112.2 for intact saturation data ranged pM binding to that for rhodopsin and (3-adrenergic brane similar proposed and 16.6 to 173.5 for cell homogenates (Table I). The cells pM extracellular amino terminus and intracellular with an receptors, for was lower in the HM4 notably of subtype [3H]QNB affinity carboxy terminus (Figure 1) (Zuker et al., 1985; Dixon et al., than that of the other human both whole cells and homogenates et Yarden et 1986; Peralta et al., Nathans al., 1986; al., 1986; whereas less variation is typically mAChR (Table I), subtypes The high degree of identity between all four subtypes 1987). of this receptors expressed in seen in the binding compound by segments and the short connect- within these membrane-spanning of animal tissues 1987). Non-transfected a (Nathanson, variety our previous proposal that (65%) (Figure 1) supports ing loops while trans- <200 mAChRs/cell, cells endogenous expressed well occur in the cleft created upon of acetylcholine may binding 7000 -90 000 depending upon fected cells receptors, expressed of these domains into the lipid bilayers insertion hydrophobic and transfection. the specific subtype particular et Two acid residues are conserved al., 1987). aspartic (Peralta experiments as the dis- [3H]QNB binding using Competition all four membrane in the otherwise highly hydro- among subtypes were carried out to determine the affinity of placeable ligand and third transmembrane domains suggesting that second phobic for each of the recombinant drugs additional anti-muscarinic residues directly in the these charged may participate negatively 2; Table I). The classic muscarinic miAChR subtypes (Figure of the neurotransmitter acetylcholine, a endogenous binding bound with to all four human affinity amine. high atropine antagonist quaternary 3925 Human muscarinic acetylcholine receptors 00- -// 0 X I. dissociation for the of Table Apparent constants (KD) binding antagonists 8--O and to the four human mAChR agonists subtypes at. pz. AF. 5 0- so- oxo. carb. HM1 HM2 HM3 HM4 HM1 HM1 Antagonist-binding I. o- if0 1 I 1 I1 [3H]QNB 10 6 7 6 9 5 4 3 8 7 6 5 4 3 2 homogenate KD (pM) 26.5 37.5 16.6 173 -[ I cells KD (pM) 22.8 83.3 39.1 112 3.3 16.6 Atropine KD (nM) 1.2 1.1 12.5 2.7 Pirenzepine KD (#M) 0.5 1.2 5 s 19] AF-DX1 16 KHa (nM) 50.0 20.0 - 300 ;0-~~ HM2 HM2 KLb (PM) 6.8 0.8 2.6 1.1 %Hc 13.5 40.7 19.5 0 I / m 10 9 6 7 6 5 4 AI 8 7 6 5 4 3 2 Agonist-binding Carbachol KHa 7.3 0.1 8.1 la (AM) a 130 200 KLb 470 560 (AM) - 6 %Hc 30.1 28.4 6.5 Oxotremorine KHa (nM) - 8.0 22.0 LHM3 HM3 KLb 4.1 5.3 4.1 4.0 (AM) %Hc - 20.0 5.1 0 - 0 II -Se, , , 10 9 6 7 6 5 4 3 6 766 54 3 2 The KD values for [3H]QNB were determined independently by Scatchard of analysis saturation binding curves obtained from the same population of intact cells (n = 4) or = cell homogenates (n 2) that were utilized for or antagonist agonist binding studies respectively. The statistical variation of each value and the mean receptor numbers determined from Scatchard analysis, expressed as [3H]QNB binding sites/cell and fmol/mg homogenate are described in Materials and methods. All protein competition displacement data were fitted non linear least with one by squares regression 5 3 9 8 7 6 5 4 3 8 7 6 4 or two site models and the best fit was selected of by analysis variance -log[agonlstJ -log[antagonlst] (Munson and Rodbard, 1980). aKD for the high affinity binding state. 2. of human mAChR bKD for the low state. Fig. Ligand binding properties affinity binding subtypes expressed by transfection of mammalian cells. cThe amount of high as a of Competition displacement of affinity receptors percentage the total number [3H]QNB muscarinic of receptors. binding by antagonists (left panel) was studied in intact cells transfection following with each of the four muscarinic clones; receptor was determined with cell agonist binding (right panel) homogenates with HM3 and HM4 similar subtypes, HM1, displaying KD as described in Materials and methods. The prepared studied antagonists values 1.2 and 1.1 nM In contrast to 0 (KD 3.3, respectively). were atropine ( ), pirenzepine (A) and AF-DX1 16 The (e). agonists were carbachol and oxotremorine studied The the small variations in for (A) (0). apparent regional atropine affinity reported constants dissociation of the different are summarized in Table (KD) ligands I. various tissues (Birdsall and Hulme, the HM2 1983), subtype bound atropine with significantly lower 16.6 affinity (KD nM) HM2 and HM3 = 0.8 and subtypes (KD 1.1 respectively), AM than the other three human mAChR as well as the natural subtypes whereas the low state of HM1 was lower affinity significantly porcine atrial M2 mAChR (KD 1.0 and nM) (Schimerlik = HM2 (KD 6.8 ,uM). the of displayed largest proportion high Searles, 1980) or a recombinant porcine M2 mAChR 4.1 (KD affinity AF-DX1 16 binding sites (40.7% of the total receptor nM) (Peralta et al., 1987). Substantial differences were also sites) as well as the highest 20 of the four affinity (KD nM) detected in the abilities of the four human mAChRs to bind the human for mAChRs this ligand (Table I). Thus, the HM2 subtype selective muscarinic antagonist pirenzepine, which distinguishes is most similar to M2 receptors of porcine and rat atria in display- between MI and M2 mAChR subtypes (Hammer et al., 1980), the ing highest overall affinity for AF-DX1 16 (Giraldo et al., and the cardioselective AF-DX1 16 antagonist (Giraldo et al., Korc et 1987; al., 1987). 1987). The affinity of pirenzepine for HM1 was 25-fold higher Multiple agonist binding states are a than for the HM2 a difference affinity property of subtype, comparable with that each muscarinic receptor subtype observed for MI receptors of rat cerebral cortex and hippocampus and M2 of rat receptors atria, confirming the assignment of the In contrast to the of single class binding sites recognized by most HMl and HM2 subtypes (Table I). Notably, the affinity ofHM3 muscarinic antagonists, agonist binding generally reveals the and HM4 for is pirenzepine much closer to that of HMl than existence of multiple affinity binding states (Nathanson, 1987). HM2 (Table I), suggesting that these mAChR subtypes may have The proportion of high and low agonist affinity sites can be altered been characterized as Ml receptors by previous investigators. by guanine nucleotides suggesting that heterogeneous agonist Of the four recombinant human mAChRs, only HM3 exhibited binding reflects conformation states of receptor which arise homogeneous binding to the antagonist 16 (Figure AF-DX1 2). through interactions with G proteins (Nathanson, 1987). How- Binding of AF-DX1 16 to the other three mAChR subtypes was since ever, individual subtypes may bind a given agonist with heterogeneous and best fit to computer models assuming two different and affinities, respond differently to guanine nucleotide classes of sites, indicating that this drug distinguishes between addition, the of presence multiple mAChR subtypes within a high and low affinity states for each of these receptors. The value single tissue could possibly give the appearance of heterogeneous calculated for the state of the HM3 agonist binding independent of the effects ofG proteins. To deter- single affinity subtype (KD = 2.6 was with the low mine the state of the agonist binding properties of individual human tM) mAChR comparable affinity 3926 E.G.Peralta et al. Z ID0 ZIO 0 Z I o C Tissue-specific expression of muscarinic receptors Binding studies with selective muscarinic antagonists have sug- gested that mAChR subtypes are differentially expressed in various mammalian tissues (Hammer et al., 1980; Birdsall and ,w r .Hulme, 1983; Giraldo et al., 1987). To the directly investigate 28S- U s expression of mAChRs in tissues in which mAChR subtypes have been pharmacologically defined, RNA isolated from rat heart, 18S- whole brain, cerebral cortex, pancreas and NG108-15 rat/mouse neuroblastoma x glioma cells was subjected to Northern blot analysis under stringent hybridization conditions utilizing probes derived from the subtype-specific sequences contained in the 5-6 loop region of each human mAChR. Each probe detected a dis- crete species of mRNA within a subset of the tissues or cells analyzed (Figure 3); the estimated sizes of the mRNAs corre- HM1 HM2 HM3 HM4 sponding to eachsubtypewere3.1 kb forHM1, 5.2 kb forHM2, Fig. 3. Tissue specificity of mAChR subtype gene expression. Polyadenylated 3.3 kb for HM3 and 4.5 kb for HM4. Notably, expression of mRN. A was isolated from the indicated rat tissues, denatured, separated by each rat mAChR subtype mRNA was detected in whole brain, electriophoresis through a 1.1% agarose-formaldehyde gel (Maniatis et al., providing definitive evidence for the transcription of the novel 1982) transferred to nitrocellulose and hybridized with uniformly 32p_ HM3 and HM4 genes, and suggesting a role for each subtype labeleA restriction fragments specific for each subtype. mRNA samples central nevous system function. In contrast, subtype-specific in includle: NG108 (NG), heart (H), cerebral cortex (C), pancreas (P) and whole brain (WB); each lane contained 5yg of each type of mRNA, except expression was observed in other tissues. Heart mRNA was found cerebiral cortex samples which contained 3 The positions of 28S to hybridize only with the HM2 probe, supporting the M2 sub- Ag/lane. and 1 8S ribosomal RNA are indicated. Hybridization of the filter with a type assignment of this human mAChR clone. With pancreatic 32p_er nd-labeled oligonucleotide encoding a portion of the mouse alpha-actin RNA was observed with the HM4- significant hybridization only codinl region (Alonso et al., 1986) was performed to confirm that specific probe, suggesting that this subtype corresponds to the equiv. alent amounts of each type of mRNA were present in the s corres sponding lanes (not shown). 'glandular' mAChR subtype described in earlier studies (Figure 3) (Giraldo et al., 1987; Korc et al., 1987). This observation (pes, [3H]QNB competition binding analysis was carried out is consistent with the significantly larger size previously detected subty the muscarinic agonists oxotremorine and carbachol, a close for the mAChR found in pancreatic acinar cell preparations, as with tural relative of the natural neurotransmitter acetylcholine. expected for HM4 (Hootman et al., 1985). Interestingly, ex- struc len in Figure 2, heterogeneous binding was readily apparent pression of mAChR RNA in NG108-15 cells is restricted to As se IMl,HM2 andHM4 with carbachol, and for HM2 and HM3 HM3, suggesting that biochemical events mediated by muscarinic for oxotremorine. In each case, computer analysis of the bind- agonists in this neuron-derived cell line may largely reflect bind- with dlata eave a statisticallv better fit with a two site than a one ing to this subtype. inz Biological role of muscarinic receptor subtype diversity site model (Table I). Oxotremorine exhibited substantially higher The studies presented here demonstrate that the human gen- affinity than carbachol for both classes of sites for a given human ome contains a family of at least four evolutionarily conserved mAChR subtype (Table I), consistent with binding studies per- mAChRs which differ strikingly in their primary structures, their formed with tissue preparations (Schimerlik and Searles, 1980; ability to bind various ligands, and in their pattern of tissue- Peterson and Schimerlik, 1984). The low affinity binding states specific expression, revealing a heretofore unappreciated degree calculated for either agonist were similar for each of the four of specialization in muscarinic cholinergic function. Despite the mAChR subtypes, with KD values ranging from 130 to 560yM substantial amino acid divergence between the four mAChR for carbachol, and 4.0 to 5.3 ,uM for oxotremorine. The high subtypes, 151 of 241 residues (65%) contained within the seven affinity carbachol binding state was notably higher for HM2 (KD predicted membrane-spanning domains and their short connecting = 0.1 M) than for HM1 and HM4 (KD = 7.3 and 8.1 AM loops are identical in each. The majority of amino acid differences respectively), and in the range of the high affinity carbachol sites among the four receptor subtypes within these regions correspond reported for rat cerebral Ml and porcine atrial M2 mAChRs to conservative substitutions, supporting the notion that these (KD = 1.41 and 0.83 respectively) (Fisher et al., 1983; IAM sequences constitute the receptor ligand-binding domain. The Peterson and Schimerlik, 1984; Collins and Crankshaw, 1986). most unique structural feature of each subtype is the large loop The high affinity binding states of HM2 and HM3 for oxotre- between the fifth and sixth transmembrane regions, which is morine were comparable (KD = 8.0 and 22.0 nM respective- predicted to reside on the cytoplasmic side of the plasma mem- ly) and similar to the porcine atrial M2 mAChR (KD = 2.3 nM) brane, and is thus likely to play an important role in specifying (Peterson and Schimerlik, 1984). Significantly, the appearance which effector system(s) will couple to a given subtype. This of multiple affinity states for either oxotremorine or carbachol possibility is supported by studies with (3-adrenergic receptor upon transfection of each human mAChR subtype cannot be deletion mutants showing that deletions of the corresponding attributed to endogenous mAChRs since their level (< 200 recep- domain of this receptor abolish its ability to regulate adenylyl tors/cell) cannot account for even the smallest population of high cyclase activity without impairing ligand binding (Dixon et al., agonist binding sites observed in these experiments. Fur- affinity 1987). A greater degree of similarity within the 5-6 loop region the variable proportion of high and low affinity (or thermore, is evident between the HM1 and HM4 subtypes, and the HM2 of high affinity) binding sites detected for each human absence and HM3 subtypes, suggesting that both of these pairs of sub- mAChR subtype suggests that the four subtypes may differ in types share greater functional similarity. their abilities to couple efficiently with the G proteins present The ligand binding properties of human mAChR subtypes ex- within these cells. 3927 Human muscarinic acetylcholine receptors mammalian as well as the thus suggest that the HM3 be pressed by transfected cells, pattern subtype may preferentially coupled to adenylyl cyclase. The pancreas-specific expression of HM4 of gene expression detected for the homologous rat mRNAs, clues to the relationship between the subtypes suggests that this mAChR be to the provide important subtype may coupled poly- identified by earlier phosphoinositide turnover and calcium mobilization associated characterized here and those investigators. analysis of mRNA from rat tissues revealed that with the carbachol stimulated secretory activities of pancreatic Northern blot are in the heart acinar cells et Sekar et only the HM2 and HM4 homologues expressed (Korc al., 1987; al., 1987). corre- In tissues in which different mAChR subtypes may be co- and pancreas, respectively, suggesting that these subtypes 'glandular' human mAChR the contribution of to the exist- spond to the 'cardiac' and subtypes. expressed, subtype heterogeneity subtype is also supported by ence of multiple agonist binding affinity states cannot easily be The assignment of the HM2 ligand HM2 discerned. The here for cells trans- binding analysis demonstrating that the recombinant poly- binding experiments presented a fected with individual subtypes conclusively demonstrate that exhibits low affinity for pirenzepine, and high propor- peptide sites for AF-DX1 16, two distinguishing multiple agonist affinity states are the property of all four mAChR tion of high affinity M2 receptors in mammalian atria (Hammer et polypeptides. High affinity agonist binding by mAChRs is be- characteristics of al., 1980; Giraldo et al., 1987). In addition, HM2 shares virtual lieved to reflect interactions with G proteins since the proportion with the M2 mAChR protein purified from of such high affinity sites can be substantially decreased by the amino acid identity atria (Peralta et al., 1987). The identity of HM1 is in- addition of guanine nucleotides or by pertussis porcine toxin-catalyzed affinity for pirenzepine and expression in ADP-ribosylation, treatments which uncouple receptors from G dicated by its high HM1 exhibits virtual identity with proteins (Stryer and Boume, 1986). The differences in the cerebral cortex. Furthermore, affinity MI clones corresponding to the major subtype porcine mAChR and proportion of the multiple agonist affinity states detected for porcine cerebral cortex (Kubo et al., 1986a). While individual mAChR subtypes expressed purified from in a single cell type may HM1 the highest pirenzepine affinity of the four human reflect the capacity of individual subtypes to interact differentially exhibits ability of the HM3 and HM4 subtypes to bind with G proteins, and thus to couple specifically with the various mAChRs, the with relatively high affinity, and their comparable biochemical effector systems present. pirenzepine to HM1 in brain, suggests that previous studies of In conclusion, our results demonstrate that the human mAChR abundance Ml displaying high pirenzepine affinity may have gene family is composed of at least four members; specific struc- brain mAChRs to discriminate between these three mAChR subtypes. tural features distinguish each receptor and are manifested in the failed In to the preferential expression of the HM2 and HM4 distinct ligand-binding properties displayed by individual contrast sub- respectively, all four mAChR sub- types expressed by transfection of mammalian cells. This strategy subtypes in heart and pancreas extents in whole rat brain. This will be further employed to study the biochemical properties of types are expressed to similar nature of synaptic signalling each mAChR and thereby continue the investigation of difference may reflect the complex subtype, in nevous system compared with the per- the physiological significance of receptor subtype multiplicity. that occurs the central indicate a physiological require- ipheral nervous system, or may Materials and methods receptor subtypes which may regulate different ment for multiple expression of mAChRs within effector systems. Subtype-specific Isolation of molecular clones central nervous is suggested by previous studies, which the system A human genomic library (Lawn et al., 1978) was screened under low stringency hybridization conditions with a 680-bp uniformly 32P-labeled restriction that M I and M2 mAChRs occupy postsynaptic and pre- fragment indicate encoding the first five transmembrane domains of the porcine M2 muscarinic within cholinergic nerve terminals synaptic locations respectively receptor described as (Peralta et al., 1987). Positive clones were screened with of the cerebral cortex (Mash et al., 1985). Moreover, the effects 32P-end-labeled oligonucleotide pools, each containing 64 degeneracies, encoding of acetylcholine on ion conductances within the central nervous either amino acids 55-62 (Gln-Thr-Val-Asn-Asn-Tyr-Phe-Leu) or 64-70 (Ser- system have been used to correlate mAChR subtypes with specific Leu-Ala-Cys-Ala-Asp-Leu) of the porcine atrial M2 muscarinic receptor (Peralta et al., 1987) and with oligonucleotides encoding unique regions of the porcine electrophysiological functions; M2 mAChRs appear to mediate M1 (amino acids 259-286) and M2 (amino acids 240-263) mAChRs. Restriction an increase in potassium conductance within brain cells, while fragments from each genomic clone which hybridized with the oligonucleotide MI mAChRs are associated with an inhibition of potassium per- probes described above were subcloned into M13 vectors and sequenced by the meability (Egan and North, 1986). The predominant expression chain termination method (Sanger et al., 1980; Messing et al., 1981). Nucleotide sequence data will be submitted to the EMBL/GenBank Data Libraries. of a single mAChR subtype, HM3, in NG108-15 neuronal cells and Nirenberg, 1975) suggests the interesting possi- (Matsuzawa Protein sequence homologies that of mAChR subtype expression may be highly bility regulation Sequence comparisons were first determined by separate alignment of each pair of receptors (Needleman and Wunsch, 1970; Fitch and Smith, 1983). This analysis differentiated at the level of the individual neuron as well as in revealed a high degree of identity within the predicted transmembrane domains gross regions of the brain. and short hydrophilic loops of each pair of mAChR and thus allowed the alignment Earlier studies have revealed correlations between the occu- of all four sequences by visual inspection. The large cytoplasmic domain joining pation of a given receptor subtype which has been defined by the putative fifth and sixth transmembrane domains was analyzed separately for ligand specificity, and the activation of a specific cellular effector. both amino acid and nucleotide homology. The scoring parameters are those of Dayhoffet al. (1983) with a deletion penalty identical to that of Lipman and Pearson For example, preferential coupling of the MI subtype to poly- (1985). phosphoinositide turnover in rat brain and mouse anterior pituitary Expression of human mAChRs cells, and of the M2 subtype to inhibition of adenylyl cyclase Each mAChR coding sequence was inserted into a derivative of pF8CIS9080 in rat heart or mouse anterior pituitary cells and activation of containing EcoRI, SmaI and BamHI linker sites replacing the Factor VIII coding the inward potassium channel of myocardium have been reported region (Eaton et al., 1986). The following coordinates indicate the nucleotide (Gil and Wolfe, 1985; Pfaffinger et al., 1985; Watson et al., sequences of each receptor that were inserted into the expression vector: positions 1986). The pattern of expression observed for the HM3 and HM4 -66 to + 1418, HM1; bp -42 to + 1467, HM2; bp -81 to + 1598, HM3; bp -7 to + 1795, HM4 (Figure 2). For transient expression of mAChRs, 60-mm mAChRs suggests specific roles in effector coupling for these plates of human embryonic kidney cells (50% confluent) were transfected as subtypes as well. Carbachol stimulation of NG108-15 cells results described et method (Wigler previously (Eaton al., 1986) by the calcium phosphate in with no effect on inhibition, adenylyl cyclase et hours after PBS contain- apparent phos- al., 1979). Sixty transfection cells were harvested in and our 5 mM EDTA and tested in phoinositide hydrolysis Harden, ligand-binding assays. (Hughes 1987); findings ing 3928 E.G.Peralta et al. Ligand-binding analysis and 118- Pharm. 7her., 240, Korc,M., Ackerman,M. Roeske,W. (1987) Exp. cells or with cell as Binding studies were carried out with intact homogenates, intact Mishina,M., earlier (Peralta et al., 1987). Incubations were 75 min at 37°C for Kubo,T., Fukuda,K., Mikami,A., Maeda,A., Takahashi,H., Haga, described Hirose,T. and Numa.S. cells and 60 min at 20-22°C for cell homogenates. Non-specific binding was K., Ichiyama,A., Kangawa,K., Kojima,M., Matsuo,H., 411-416. determined in the presence of 10 atropine and was always < 10% of the (1986a) Nature, 323, /sM total [3H]QNB binding. The mean KD values for [3H]QNB binding and receptor Takahashi,H., Kubo,T., Maeda,A., Sugimoto,K., Akida,I., Mikami,A., Haga,K., Hirose,T. and Fed. numbers expressed per transfected cell were determined by Scatchard analysis Numa.S. (1986b) Ichiyama,A., Kangawa,K., Matsuo,H., Biochem. Soc. 367-373. of [3H]QNB saturation binding by intact cells (n = 4) utilizing the LIGAND Eur. Lett., 209, J. Mol. 105-132. program (Munson and Rodbard, 1980) and are as follows: HM1, 22 A 8 pM, and Kyte,J. Doolittle,R. (1982) Biol., 57, and Maniatis,T. 1157- 9000 ± 2100 sites/cell; HM2, 83 + 41 pM, 48 600 + 1400; 39 6 Cell, 15, HM3, + pM, Lawn,R., Fritsch,E., Parker,R., Blake,G. (1978) 27 400 ± 7600; HM4, 112 i 55 pM, 89 600 21 700. Similar results for and 1435-1441. [3H]QNB-binding were obtained for cell homogenates (n = 2) with the range Lipman,D. Pearson,W. (1985) Science, 227, and Sambrook,J. Molecular A Labora- of values as follows: 17 pM, 13 fmol/mg homogenate protein (n = 1); HMI, Maniatis,T., Fritsch,E.F. (1982) Cloning: Manual. Cold Harbor New York. HM2, 26 10 pM, 155 ± 31 fmol/mg; HM3, 37 29 pM, 44 A20 fmol/mg; L tory Spring Laboratory Press, and 115-117. HM4, 172 i 73 pM, 332 241 fmol/mg. Antagonist competition displacement Mash,D., Flynn,D. Science, Potter,L. (1985) 22?, and Proc. Acad. Sci. experiments were performed with a subset of the transfections described above Matsuzawa,H. Natl. USA, 72, Nirenberg,M. (1975) 3472-3476. and the mean KD values for [3H]QNB binding and receptor numbers are: HMl, Nucleic Acids 309-321. and 19.5 ± 1 pM, 7100 ± 3300 sites/cell (n = 2); HM2, 95 i 60 pM, 58 100 Res., 9, Messing,J., Crea,R. Seeburg,P. (1981) and Anal. Biochem., 220-239. + 13 800 (n = 3); HM3, 45 ± 9 pM, 16 400 + 3200 (n = 3); 112 Munson,P. 107, HM4, Rodbard,D. (1980) and 193-200. + 83 pM, 82 900 28 800 (n = 3). The agonist competition Science, displacement Nathans,J., Thomas,D. Hogness,D. (1986) 232, Annu. Rev. 195-236. experiments were performed with cells from three transfections and the values Neurosci., KD Nathanson,N. (1987) 10, and J. Mol. 443-453. for [3H]QNB binding and receptor numbers for one Biol., representative experiment Needleman,S. Wunsch,C. (1970) 48, are as follows: HM1, 42 pM, 13 250 sites/cell; 57 78 Ramachandran,J., HM2, pM, 000; HM3, Peralta,E., Winslow,J., Peterson,G., Smith,D., Ashkenazi,A., and 600-605. 27 pM, 34 200; HM4, 16 pM, 133 000. The KD Science, values for competing antagon- Schimerlik,M. Capon,D. (1987) 236, and 33-74. Biochem., 14, ists and agonists were determined from three transfections and displacement exper- Peterson,G. Schimerlik,M. (1984) Prep. and Hille,B. Nature, iments performed in duplicate (with the exception of cholinergic binding to HM 1 Pfaffinger,P., Martin,J., Hunter,D., Nathanson,N. (1985) 536-538. 317, which was analyzed in duplicate in one experiment); equivalent results were and J. Mol. Roe,B. Biol., obtained and the values of one representative experiment are presented in Table Sanger,F., Coulson,A., Barrel,B., Smith,A. (1980) 161-178. I. Untransfected cells expressed <200 muscarinic receptors/cell (< 1.3 fmol/mg 143, and 3407-3413. 19, homogenate protein). Schimerlik,M. Searles,R. (1980) Biochemistry, Dixon,J. and J. Biol. 340-344. Chem., 262, Sekar,M., Hokin,L. (1987) Northern hybridization analysis and Annu. Rev. Cell 391-419. Biol., 2, Stryer,L. Bourne,H. (1986) Subtype-specific restriction fragment probes encoded the following unique regions The In and Taylor,P. (1985) Goodman,L. Gilman,A. (eds), Pharmacological of the large cytoplasmic domain amino acids 265- of each receptor: HM1, Basis New 100. of 7herapeutics. Macmillan, York, p. 363; HM2, amino acids 268-381; HM3, amino acids 265-394; HM4, amino and Watson,M., Roeske,W., Vickroy,T., Smith,T., Warmsley,J. Yamamura,H.I. acids 292-390. with Each fragment was uniformly radiolabeled [a-32P]dCTP Trends Pharmacol. 7 46-55. (1986) Sci., (Suppl.), (Maniatis et al., 1982). Stringent hybridization conditions for Southern and North- and Wigler,M., Pellicer,A., Axel,R., Urlaub,G. Chasin,L. (1979) Silverstein,S., ern filters are as previously described (Peralta et al., 1987). Proc. Acad. Sci. 1373-1376. Natl. USA, 76, Brandt,D., Burnier,J., Harkins,R., Yarden,Y., Rodriguez,H., Wong,S., May,D., and Proc. Natl. Acad. Ross,E. Chen,E., Ramachandran,J., Ullrich,A. (1986) Acknowledgements Sci. 6795-6799. USA, 83, for Dr P.Moore for We thank Drs J.Witkin and C.Gorman helpful discussions, 851-858. and Cell, Zuker,C., Cowman,A. Rubin,G. (1985) 40, advice on computer analysis of ligand-binding data, S.Marsters for tissue culture assistance, M.Vasser, P.Jhurani and P.Ng for oligonucleotide synthesis, Drs on 1987 Received September 14, for mRNA Drs and A.Rosenthal and C.Logsdon providing samples, P.Seeburg and C.Morita and S.Peters for P.Braun for communicating unpublished data, prep- added in Note proof aration and AF-DX1 16 < of figures. Pirenzepine 11-[[2-(diethylamino)methyl] of this Bonner et al. - the submission 1-piperidinyllacetyl > -5,1 1-dihydro-6H-pyrido[2,3,6][1,4]benzodiazepine- Following original (Science, 235, manuscript the and of three 6-one > Drs G.Trummlitz and W.Reuter. were kindly provided by R.Hammer, 527-532, 1987) reported cloning partial sequence analysis Rml Rm3 and Hm4 of their Inc. and CA16417 to J.R. mAChR This work was supported by Genentech, NIH grant subtypes. (rat), (human). Comparison (rat) here indicates that Hm4 is to with the equivalent subtypes mnAChRs reported Rml and Rm3 are rat of HM1 and HM4 and that HM3, suggests homologues References the rat warrants further since of respectively. Assignment subtypes investigation Southern that there be five additional Alonso,S., Minty,A., and J. Mol. Evol., 23, Bonner et al. showed Bourlet,Y. Buckingham,M. (1986) by analysis may rat 11-22. uncharacterized muscarinic genes. Birdsall,N. and Trends Pharmacol. Sci., 4, 459-463. Hulme,E. (1983) Brown,J.H., and Mol. Pharmacol., 27, 525-531. Goldstein,D. Masters,S. (1985) Cohen,S. and Trends Pharmacol. Sci., 8, 1-4. Sokolovsky,M. (1987) Collins,S. and Am. J. 251, 546-552. Crankshaw,D. (1986) Physiol, Dayhoff,M., Barker,W. and Methods 91, 524-545. Hunt,L. (1983) Enzymol., Dixon,R., Kobilka,B., Dolhman,H., Frielle,T., Bolan- Strader,D., Benovic,J., owski,M., Bennett,C., Mumford,R., Slater,E., Rands,E., Diehl,R., Sigal,I., Caron,M., Lefkowitz,R. and Nature, 321, 75-79. Strader,C. (1986) Dixon,R., Sigal,I., Candelore,M., Blake,A. and Strader,C. Rands,E., Register,R., (1987) Nature, 326, 73-77. Eaton,D.L., Wood,W., Eaton,D., Wion,K., Mather,J., Hass,P., Hollingshead,P., Lawn,R., Vehar,G. and Gorman,C. (1986) Biochemistry, 26, 8343-8347. Egan,T. and North,R. (1986) Nature, 319, 405-407. Fisher,S., Klinger,P. and Agranoff,B. (1983) J. Biol. Chem., 258, 7358-7363. Fitch,W. and Smith,T. (1983) Proc. Natl. Acad. Sci. USA, 80, 1382-1386. Gil,D. and Wolfe,B. (1985) J. Phann. Exp. Ther., 232, 608-616. Giraldo,E., Hammer,R. and Ladinsky,H. (1987) Life Sci., 40, 833-840. Hammer,R., Berrie,C., Birdsall,N., Burgen,A. and Hulme,E. (1980) Nature, 283, 90-92. Hootman,S., Picardo-Leonard,T. and Burnham,D. (1985) J. Biol. Chem., 260, 4186-4194. Harden,T. J. Pharm. Ther., 237, 173-178. Hughes,A. and (1987) Exp. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The EMBO Journal Springer Journals

Distinct primary structures, ligand‐binding properties and tissue‐specific expression of four human muscarinic acetylcholine receptors.

Download PDF
7 pages

Loading next page...
 
/lp/springer-journals/distinct-primary-structures-ligand-binding-properties-and-tissue-s8n5MZh00g

References (4)

Publisher
Springer Journals
Copyright
Copyright © European Molecular Biology Organization 1987
ISSN
0261-4189
eISSN
1460-2075
DOI
10.1002/j.1460-2075.1987.tb02733.x
Publisher site
See Article on Publisher Site

Abstract

The EMBO Journal vol.6 no.13 pp.3923 -3929, 1987 Distinct primary structures, ligand-binding properties and tissue- specific expression of four human muscarinic acetylcholine receptors Ernest was initially discerned through the discovery of selective muscar- G.Peralta, Avi Ashkenazi'2, John W.Winslowl, inic agents which exhibit different Douglas H.Smith, J.Ramachandran"2 and affinities for receptors in Daniel various tissues (Hammer et al., 1980; Birdsall and Hulme, 1983; J.Capon Cohen and Sokolovsky, 1987). Studies with selective antagonists Departments of Molecular Biology and 'Developmental Biology, Genentech, such as pirenzepine, which discriminates between the high affinity Inc., 460 Point San Bruno Boulevard, South San Francisco, CA 94080 and MI receptors in certain regions of the brain and the low affinity 2Department of Biochemistry and Biophysics, University of California, San Francisco, CA 94143, USA M2 receptors of heart, have permitted tentative proposals of func- tional differences between subtypes in their abilities to regulate Communicated by M.Birnstiel specific biochemical effector systems (Gil and Wolfe, 1985; To investigate the molecular basis for the diversity in mus- Brown et al., 1985). Molecular cloning of the porcine cerebral carinic cholinergic function, we have isolated the genes encod- (MI) and atrial (M2) mAChRs has established that the major ing the human Ml and M2 muscarinic receptors (mAChR) receptor subtypes of brain and heart are indeed different poly- as well as two previously undiscovered mAChR subtypes, peptides encoded by distinct genes which share 38% overall designated HiM3 and HM4. The amino acid sequence of each amino acid homology (Kubo et al., 1986a,b; Peralta et al., 1987). reflects a structure of subtype consisting seven, highly con- Evidence for the existence of previously uncharacterized genes served transmembrane a segments and large intracellular encoding novel mAChR subtypes has been provided by studies region unique to each subtype, which may constitute the with other selective muscarinic compounds (Giraldo et al., 1987; ligand-binding and effector-coupling domains respectively. Korc et al., 1987), as well as hybridization analysis of mam- differences in for Significant affinity muscarinc ligands were malian genomic DNAs which revealed additional sequences hom- detected in individual mAChR subtypes produced by trans- ologous to but distinct from the M1 and M2 mAChR genes fection of mammalian cells. Each exhibited subtype multiple (Peralta et al., 1987). To provide a basis for investigating mAChR states for differences in the affinity agonists; among subtypes functional diversity, we have isolated the genes for the human affinities and of such sites the proportions suggest capacity HM1 and HM2 mAChRs, as well as two novel mAChR sub- to with of mAChR subtypes interact differentially the cellu- types, herein designated HM3 and HM4. Characterization of the mRNA lar effector-coupling apparatus. Subtype-specific ex- primary structures of these four human mAChR subtypes, their in and a pression was observed the heart, pancreas neuronal agonist and antagonist binding properties, and their patterns of cell that the of mAChR ex- line, indicating regulation gene expression in various tissues reveals that all four mAChRs indeed pression contributes to the differentiation of cholinergic ac- represent functional proteins encoded by distinct, actively tivity. transcribed genes, and suggests unique biological functions for Key words: muscarinic receptors/receptor subtypes/ligand bind- each receptor subtype. ing/gene regulation Results and discussion Molecular human Introduction cloning of muscarinic receptors reveals two novel subtypes transmission muscarinic Synaptic by acetylcholine receptors To isolate the genes encoding human mAChRs, we employed is the central and (mAChRs) employed throughout peripheral a based strategy upon sequences conserved between the porcine nervous to elicit a and diverse of neuro- systems large array MI and M2 mAChR subtypes, as well as similarities with other actions. In the autonomic nervous mAChRs physiological system known receptors which transduce signals through interaction with the force and rate of heart muscle the regulate contraction, nucleotide guanine binding (G) proteins. Since the protein coding of smooth and the of contraction muscle, secretory activity of the M2 mAChR and hamster sequences 3-adrenergic receptor which receive innervation glands parasympathetic (Taylor, 1985). are each contained within a exon et single (Dixon al., 1986; mAChRs also constitute the of majority cholinergic synapses Peralta et a was screened for to be al., 1987), genomic library receptor within the central nevous where are system, they thought clones with the that human mAChR would also as and expectation genes to such neural important processes learning memory lack introns within their coding regions. The and of mAChR functional porcine M1 An (Nathanson, 1987). important aspect M2 mAChRs exhibit the greatest degree of amino acid identity the multitude of biochemical and elec- is reflected diversity by in first and their five transmembrane domains connecting loops actions evoked to trophysiological by acetylcholine binding et therefore low (Peralta al., 1987); stringency hybridizations include the of intracellular levels of which mAChRs, regulation were carried out with a this to fragment containing region identify and inositol and the or cGMP cAMP, phospholipids, opening related human sequences. In addition, the MI and M2 mAChRs of the calcium and ion channels found chloride closing potassium, as as to share the most to each other well similarity ,3-adrenergic tissues Given its role in certain (Nathanson, 1987). ubiquitous in their first receptors and the visual rhodopsins a in the cytoplasmic loop in neuronal signal transduction, major question study transmembrane a site of G and second domain, function relates to the molecular mechanisms which possible protein of cholinergic et Dixon et Nathans et interaction (Zuker al., 1985; of muscarinic al., 1986; the underly biological specialization activity. Yarden et Peralta et To isolate between mAChRs al., 1986; al., 1986; al., 1987). for structural The potential heterogeneity Oxford, IRL Press Limited, England E.G.Peralta et al. H2N- * HM1 1 - - - - - - - - - - - - - - I- - - - - - - - - - - - - - - - - - - - - - - - - - - - MNTSAPP M TL H*NANST TS PL FPANI SS SW IH SP SD AG LP PG TV TH FG S YN S R AA F s HM4 MNGSS as1~ HM3 1-~~~~~~~~~~~~~-MANFTP - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --MANFTP H - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - M NNSTN HM2 1-21 HM1 8A V S P T V L APGKPWVAFIGI TGLLSLATTL S FKVNTE HM4 5 1 SP D G[T D D P L GH TVWOVVFIAFLTGILF;]L V T I I GN I L VS F KV NO L K HM3 16 V R L VS S S H NRYETV V F ATVTGSLSLVTFV VG N I L V S IK V N R O L HM2 7 [ S N SL A L T S PY K TF E VV L GSLSLVTIIGNILVMVSL I GLES M N L Y T Y M HH WA L G T L A C D L WL A L D Y V F IL|SL A A DL 8 T N NY HM1 L A C A D L I I WL AD Y V T N Y F LLS G I MHWAL GNL ACDL V N S MN L TTY I HM1 HM04 10 T V N N Y FLLSLACADLI I GSMS MNLTYII KGYW_LGNACDLWLALDYV 0013 65 TVNNYFLFSLACADLIIGAFSMNLYTVyIIGY4I LGIAVVCDLWLALDYV 0M2 56 TVNNYFLFSLACADLII G V F SMNLYTTL V I GY LPVVVCDLWLALDYV 13-41 V HM1 108 ASNASVMNLLLI SFDRYFSVTFRDPL SYRTAKRTPPR R|AKARLLM II G LIA UL V|S F V L U HM4 151 ASNASVMNLLVI S F D R Y FSITRP L T YRA K R T T KIRA GVMI GLA WVIS F V L U HM3 115 VSNASVMNLL I I SFDRYFCVTKPLTYPART T K MA GL MI AAAWV L S F V L U HM2 106LVSNASVMNLL ISFDRYFCVTKPLTYPKR T T KMAGMMIAAAWVLSFnIL- 4-5o v HM1 15s T MLAGQCYIQFLSOPIIT FGTAMA A F Y L P V TV TL APAILFWFOIYLVGERL 00144 201 A P A I L F V G KRTVP G E CF I F LSEPTT F G T A I A A F YP V T I MTIL WOLYF P A I L F R T V P_ a 0013 165 APAILFWQFVVGKRTVPDNQCFIOFLSNPAVTFGTAIAAFYLPVVIMTVL FLSNPAVTFGTAIAAFY|PVVIM D CJ WQFVVGK HM3 2051A HM2 156 AA P ILFWOFIVGROFFSNA GTAIAAFYLPVII MT V L J5 -61 HM1 208 YWR I Y RETEN R ARELAL GSETPGGGG - - - - - - - - - S S R P G A HM4 251 YWRIWR IRYTEKT LAIG L0A N7,HPTG SHSSRSCSSY L aLIQ S M E[AfEMTE HM3 215 Yl SLARS RI HKHVHKHRP[E]GPK[KKIAK[TILA KSJLKSPL MOSQ1VKKPP GEAAR HM2 206 Y I S RAS K EKDKK PVANODPVSP SL G R I PNNNNM- S D D G 001 249 []G P E T P PGRCCCCRA[RLLQAY- - - - - - - - -IWKE E E E E DEG[M[S L T HM4 301 K R NRRKYRC HFWFTEKSWKP SS[QMDQDHSSSDSWNNNDAAASLENSA HM3 265 L R N GKL E PPTALPIPPP VA D T ESSSGS -T Q[;TKERPA REL HM2 255 O ANPRKD P VEE NCVQGEEG S S DSV V A S MR DDE I W D LEJHRKI HM1 290 E G E P VV I KMP- - - - - - - - DI HM4 351 SSXDE I I TRAIYSIVLKL GHSTILN[TKLPSSD LOVPEEELGMVD HM3 314 TT EA T PAMPAPPLQPRALN ASRWSKIQIVT KOQGNE AIEI IP - HM2 305 E NTV STS LGHSKDE NFsK Q T C I R I GT KT PK2DSCTP N ----WTVEVVLGS - HM1 309 P[AO QP T K P P R S P N - - - - - - - - - - - - - - - .-. - - - - - - - - - - - - - - - - HM4 401 LW RK DKL AQK SVDDGGSFPKSFSKLPI QLESAVDTAKTSDVNSSVGKS HM3 0012 350 VI HM1 325 R TKKGRDRAGKGQKPRGKEOL T F KEKKAARTLSAILLAF HM4 451 TALPL sF E3EAT K L KFRSQ K KR S L EKKA TLSAILLAF HM3 362 - KSI RRNOQVKKQ RRTRI ]EIKVT IFAI L L A F PA[MRPALAJNVAR 0M2 350 -S GQN[DEE3QNIVARK K MTPKL- PAKKK P P S REKKVTRTILAI L A F VIl 16-70 375 |IL TWTPYNI MVLV KEE[L EILCGY WLCYLNSTINP|M|CYALCN|KB HM1 HM4 501 IITWTPYNIMVLVNTF DSCIDCIPKTF!NLGYWL CYI NSTIN P CY A L CNT HM3 410 I| TWTPYNVMVLVNTFCQSCIPDITVUASI G Y WL CYFINSTINPACYAL C N A T 397 IIT AACPYN A IPNT V TII GYWLCYINSTINPACYAL C N A T HM2 -COOH HM1 425 RDTR DKRRw I PKRPGSV HRTPS OC HM4 551 FRTTFK LLCQCDKKKRR QY QYQORQSVIFHK APEQAL HM3 460 F K K T FH L LLCOYRNI TAIR 447 F K K T F K H L LMCHYKNI GAT HM2 Fig. 1. Primary sequence homology between the four human muscarinic receptor subtypes. Since maximum homology occurs between HM1 and HM4 or HM3 and HM2, the sequence alignment is presented in this order. Gaps were introduced to maximize significant sequence identity and to compensate for the large size of the cytoplasmic loop 5-6 of HM4. Identities are indicated by the boxed amino acid residues. Asterisks denote potential N-linked glycosylation sites in the amino-terminal domains. Regions I-VII are putative transmembrane domains (Peralta et al., 1987) and are indicated by lines over the hydrophobic sequences. The hydrophilic loops are numbered according to the transmembrane regions which they connect and their orientations are indicated with respect to the extracellular (outer, o) or cytoplasmic (inner, i) side of the plasma membrane. The amino-terminal region is indicated by H2N- and is postulated to be extracellular, while the carboxy-terminal region is indicated by -COOH and is predicted to reside on the cytoplasmic side of the plasma membrane. Numbers at the left denote the position within each protein sequence of the first residue in each line. 3924 Human muscarinic acetycholine receptors In contrast to the virtual identity between the human and por- clones with this highly specific local identity, candidates were cine homologues of the Ml or M2 mAChRs, which suggests then characterized for their ability to hybridize with two sets of severe functional constraints on the structures of these pairs of oligonucleotide probes based upon the corresponding porcine M2 subtypes, all four human subtypes diverge dramatically in the mAChR sequences (amino acids 55-62 and 64-70, detailed in large loop connecting transmembrane domains 5 and 6 (Figure Materials and methods) (Peralta et al., 1987). Eighteen of 23 1). The length of this region differs significantly between the four clones hybridizing with the M2 mAChR restriction fragment were mAChR subtypes, consisting of 156, 181, 184 and 241 residues also found positive with both sets of oligonucleotides, and could in HM 1, HM2, HM3 and HM4 respectively. By analogy with be divided into four classes based on their ability to hybridize rhodopsin, the 5-6 loop region should reside on the cytoplasmic preferentially with individual members of each pool of side of the plasma membrane, suggesting that the exceptional oligonucleotides. Since the porcine Ml and M2 mAChRs are degree of divergence between mAChR subtypes in this region strikingly dissimilar in their large third cytoplasmic loop (Peralta may be related to their capacity for differential coupling to distinct et al., 1987), probes consisting of these subtype-unique sequences biochemical effectors or ion channels (Nathanson, 1987). Limited were used to identify one class of the human clones as the hom- ologue of the porcine cerebral Ml mAChR, and a second class homology occurs in the 5-6 loop region adjacent to the trans- membrane domains, as well as the serine, threonine and acidic as the homologue of the porcine atrial M2 mAChR. Nucleotide sequence analysis of the MI homologue predicts a polypeptide residue-rich middle region of the loop, between HM1 and HM4 of 460 amino acids sharing 98.9% identity with the porcine Ml and between HM2 and HM3 (21%), suggesting greater (22%) mAChR (Kubo et al., 1986a), while the M2 homologue sequence functional similarity between these pairs of mAChR subtypes reveals a protein of 466 amino acids with 97.4 % identity to the 1). Southern blot analysis of genomic DNA utilizing (Figure derived from the 5-6 loop of each subtype indicates that porcine M2 mAChR subtype (Peralta et al., 1987) (Figure 1). probes the human does not encode other sequences highly hom- The virtual sequence identity between each pair of human and genome mAChR subtypes, together with their ligand bind- to these unique mAChR domains (data not shown). porcine ologous ing characteristics (see below), suggests that these clones encode Limited amino acid conservation is also evident in tue amino- terminal of the four subtypes, which vary in length from the human Ml and M2 mAChR subtypes, thus designated HM1 regions and HM2 respectively. HMl and HM2, like the porcine Ml and 22 to 67 residues. However, a conserved feature of the amino terminus of all four subtypes is the presence of two to five poten- M2 mAChR subtypes, share 39% overall amino acid identity. tial sites for N-linked glycosylation (Figure 1), a feature common Sequence analysis of the remaining two classes of human clones to rhodopsins and the 3-adrenergic receptor (Zuker et al., 1985; with homology to the porcine mAChRs revealed that each class Dixon et al., 1986; Nathans et al., 1986; Yarden et al., 1986). encodes a polypeptide highly homologous to, yet quite distinct The high degree of native mAChR glycosylation (Nathanson, from the HM1 and HM2 subtypes (Figure 1). The first of these, 1987), as well as an absence of such sites in the other predicted designated HM3, predicts a protein of 479 amino acids and Mr extracellular regions of each subtype, suggest the functional 53 049 sharing 40 and 55% overall amino acid identity with HM1 of these otherwise unrelated amino-terminal sequences. similarity and HM2 respectively. The other novel class of clones, designated encodes a significantly larger protein of 590 amino acids HM4, The four human muscarinic receptor subtypes exhibit distinct and 66 127 displaying 43, 35 and 37% overall identity with M, binding properties antagonist HM2 and HM3 respectively (Figure 1). Analysis ofcDNA HM1, that the molecular clones isolated in this study encode clones corresponding to each class of mAChR indicates that all To confirm and to determine whether different human four human mAChR subtypes lack introns in their respective functional mAChRs, exhibit distinct pharmacologic characteristics coding regions (unpublished data). Accuracy of the initiation and mAChR subtypes in a similar cellular context, a transient mam- termination codon assignments shown in Figure 1 is further when presented et al., 1986) was utilized to indicated by the ability to express muscarinic ligand-binding malian expression system (Eaton of each to bind muscarinic compounds. utilizing these predicted signals (see below). study the ability subtype activity carried out with intact trans- Saturation binding experiments human muscarinic receptor contains a large Each or cell using the muscarinic antagonist fected cells homogenates domain unique to each subtype cytoplasmic benzilate revealed that all four human [3H]quinuclidinyl (QNB) the method of Kyte and Doolittle analysis by Hydropathicity encode polypeptides which specifically bind this mAChR genes each of the four human mAChRs contains indicates that (1982) with affinity (Table I). The apparent dissoci- radioligand high seven hydrophobic, potential transmembrane domains, suggesting constants determined by Scatchard analysis of the ation (KD) that all four subtypes adopt a structure within the plasma mem- from 22.8 to 112.2 for intact saturation data ranged pM binding to that for rhodopsin and (3-adrenergic brane similar proposed and 16.6 to 173.5 for cell homogenates (Table I). The cells pM extracellular amino terminus and intracellular with an receptors, for was lower in the HM4 notably of subtype [3H]QNB affinity carboxy terminus (Figure 1) (Zuker et al., 1985; Dixon et al., than that of the other human both whole cells and homogenates et Yarden et 1986; Peralta et al., Nathans al., 1986; al., 1986; whereas less variation is typically mAChR (Table I), subtypes The high degree of identity between all four subtypes 1987). of this receptors expressed in seen in the binding compound by segments and the short connect- within these membrane-spanning of animal tissues 1987). Non-transfected a (Nathanson, variety our previous proposal that (65%) (Figure 1) supports ing loops while trans- <200 mAChRs/cell, cells endogenous expressed well occur in the cleft created upon of acetylcholine may binding 7000 -90 000 depending upon fected cells receptors, expressed of these domains into the lipid bilayers insertion hydrophobic and transfection. the specific subtype particular et Two acid residues are conserved al., 1987). aspartic (Peralta experiments as the dis- [3H]QNB binding using Competition all four membrane in the otherwise highly hydro- among subtypes were carried out to determine the affinity of placeable ligand and third transmembrane domains suggesting that second phobic for each of the recombinant drugs additional anti-muscarinic residues directly in the these charged may participate negatively 2; Table I). The classic muscarinic miAChR subtypes (Figure of the neurotransmitter acetylcholine, a endogenous binding bound with to all four human affinity amine. high atropine antagonist quaternary 3925 Human muscarinic acetylcholine receptors 00- -// 0 X I. dissociation for the of Table Apparent constants (KD) binding antagonists 8--O and to the four human mAChR agonists subtypes at. pz. AF. 5 0- so- oxo. carb. HM1 HM2 HM3 HM4 HM1 HM1 Antagonist-binding I. o- if0 1 I 1 I1 [3H]QNB 10 6 7 6 9 5 4 3 8 7 6 5 4 3 2 homogenate KD (pM) 26.5 37.5 16.6 173 -[ I cells KD (pM) 22.8 83.3 39.1 112 3.3 16.6 Atropine KD (nM) 1.2 1.1 12.5 2.7 Pirenzepine KD (#M) 0.5 1.2 5 s 19] AF-DX1 16 KHa (nM) 50.0 20.0 - 300 ;0-~~ HM2 HM2 KLb (PM) 6.8 0.8 2.6 1.1 %Hc 13.5 40.7 19.5 0 I / m 10 9 6 7 6 5 4 AI 8 7 6 5 4 3 2 Agonist-binding Carbachol KHa 7.3 0.1 8.1 la (AM) a 130 200 KLb 470 560 (AM) - 6 %Hc 30.1 28.4 6.5 Oxotremorine KHa (nM) - 8.0 22.0 LHM3 HM3 KLb 4.1 5.3 4.1 4.0 (AM) %Hc - 20.0 5.1 0 - 0 II -Se, , , 10 9 6 7 6 5 4 3 6 766 54 3 2 The KD values for [3H]QNB were determined independently by Scatchard of analysis saturation binding curves obtained from the same population of intact cells (n = 4) or = cell homogenates (n 2) that were utilized for or antagonist agonist binding studies respectively. The statistical variation of each value and the mean receptor numbers determined from Scatchard analysis, expressed as [3H]QNB binding sites/cell and fmol/mg homogenate are described in Materials and methods. All protein competition displacement data were fitted non linear least with one by squares regression 5 3 9 8 7 6 5 4 3 8 7 6 4 or two site models and the best fit was selected of by analysis variance -log[agonlstJ -log[antagonlst] (Munson and Rodbard, 1980). aKD for the high affinity binding state. 2. of human mAChR bKD for the low state. Fig. Ligand binding properties affinity binding subtypes expressed by transfection of mammalian cells. cThe amount of high as a of Competition displacement of affinity receptors percentage the total number [3H]QNB muscarinic of receptors. binding by antagonists (left panel) was studied in intact cells transfection following with each of the four muscarinic clones; receptor was determined with cell agonist binding (right panel) homogenates with HM3 and HM4 similar subtypes, HM1, displaying KD as described in Materials and methods. The prepared studied antagonists values 1.2 and 1.1 nM In contrast to 0 (KD 3.3, respectively). were atropine ( ), pirenzepine (A) and AF-DX1 16 The (e). agonists were carbachol and oxotremorine studied The the small variations in for (A) (0). apparent regional atropine affinity reported constants dissociation of the different are summarized in Table (KD) ligands I. various tissues (Birdsall and Hulme, the HM2 1983), subtype bound atropine with significantly lower 16.6 affinity (KD nM) HM2 and HM3 = 0.8 and subtypes (KD 1.1 respectively), AM than the other three human mAChR as well as the natural subtypes whereas the low state of HM1 was lower affinity significantly porcine atrial M2 mAChR (KD 1.0 and nM) (Schimerlik = HM2 (KD 6.8 ,uM). the of displayed largest proportion high Searles, 1980) or a recombinant porcine M2 mAChR 4.1 (KD affinity AF-DX1 16 binding sites (40.7% of the total receptor nM) (Peralta et al., 1987). Substantial differences were also sites) as well as the highest 20 of the four affinity (KD nM) detected in the abilities of the four human mAChRs to bind the human for mAChRs this ligand (Table I). Thus, the HM2 subtype selective muscarinic antagonist pirenzepine, which distinguishes is most similar to M2 receptors of porcine and rat atria in display- between MI and M2 mAChR subtypes (Hammer et al., 1980), the ing highest overall affinity for AF-DX1 16 (Giraldo et al., and the cardioselective AF-DX1 16 antagonist (Giraldo et al., Korc et 1987; al., 1987). 1987). The affinity of pirenzepine for HM1 was 25-fold higher Multiple agonist binding states are a than for the HM2 a difference affinity property of subtype, comparable with that each muscarinic receptor subtype observed for MI receptors of rat cerebral cortex and hippocampus and M2 of rat receptors atria, confirming the assignment of the In contrast to the of single class binding sites recognized by most HMl and HM2 subtypes (Table I). Notably, the affinity ofHM3 muscarinic antagonists, agonist binding generally reveals the and HM4 for is pirenzepine much closer to that of HMl than existence of multiple affinity binding states (Nathanson, 1987). HM2 (Table I), suggesting that these mAChR subtypes may have The proportion of high and low agonist affinity sites can be altered been characterized as Ml receptors by previous investigators. by guanine nucleotides suggesting that heterogeneous agonist Of the four recombinant human mAChRs, only HM3 exhibited binding reflects conformation states of receptor which arise homogeneous binding to the antagonist 16 (Figure AF-DX1 2). through interactions with G proteins (Nathanson, 1987). How- Binding of AF-DX1 16 to the other three mAChR subtypes was since ever, individual subtypes may bind a given agonist with heterogeneous and best fit to computer models assuming two different and affinities, respond differently to guanine nucleotide classes of sites, indicating that this drug distinguishes between addition, the of presence multiple mAChR subtypes within a high and low affinity states for each of these receptors. The value single tissue could possibly give the appearance of heterogeneous calculated for the state of the HM3 agonist binding independent of the effects ofG proteins. To deter- single affinity subtype (KD = 2.6 was with the low mine the state of the agonist binding properties of individual human tM) mAChR comparable affinity 3926 E.G.Peralta et al. Z ID0 ZIO 0 Z I o C Tissue-specific expression of muscarinic receptors Binding studies with selective muscarinic antagonists have sug- gested that mAChR subtypes are differentially expressed in various mammalian tissues (Hammer et al., 1980; Birdsall and ,w r .Hulme, 1983; Giraldo et al., 1987). To the directly investigate 28S- U s expression of mAChRs in tissues in which mAChR subtypes have been pharmacologically defined, RNA isolated from rat heart, 18S- whole brain, cerebral cortex, pancreas and NG108-15 rat/mouse neuroblastoma x glioma cells was subjected to Northern blot analysis under stringent hybridization conditions utilizing probes derived from the subtype-specific sequences contained in the 5-6 loop region of each human mAChR. Each probe detected a dis- crete species of mRNA within a subset of the tissues or cells analyzed (Figure 3); the estimated sizes of the mRNAs corre- HM1 HM2 HM3 HM4 sponding to eachsubtypewere3.1 kb forHM1, 5.2 kb forHM2, Fig. 3. Tissue specificity of mAChR subtype gene expression. Polyadenylated 3.3 kb for HM3 and 4.5 kb for HM4. Notably, expression of mRN. A was isolated from the indicated rat tissues, denatured, separated by each rat mAChR subtype mRNA was detected in whole brain, electriophoresis through a 1.1% agarose-formaldehyde gel (Maniatis et al., providing definitive evidence for the transcription of the novel 1982) transferred to nitrocellulose and hybridized with uniformly 32p_ HM3 and HM4 genes, and suggesting a role for each subtype labeleA restriction fragments specific for each subtype. mRNA samples central nevous system function. In contrast, subtype-specific in includle: NG108 (NG), heart (H), cerebral cortex (C), pancreas (P) and whole brain (WB); each lane contained 5yg of each type of mRNA, except expression was observed in other tissues. Heart mRNA was found cerebiral cortex samples which contained 3 The positions of 28S to hybridize only with the HM2 probe, supporting the M2 sub- Ag/lane. and 1 8S ribosomal RNA are indicated. Hybridization of the filter with a type assignment of this human mAChR clone. With pancreatic 32p_er nd-labeled oligonucleotide encoding a portion of the mouse alpha-actin RNA was observed with the HM4- significant hybridization only codinl region (Alonso et al., 1986) was performed to confirm that specific probe, suggesting that this subtype corresponds to the equiv. alent amounts of each type of mRNA were present in the s corres sponding lanes (not shown). 'glandular' mAChR subtype described in earlier studies (Figure 3) (Giraldo et al., 1987; Korc et al., 1987). This observation (pes, [3H]QNB competition binding analysis was carried out is consistent with the significantly larger size previously detected subty the muscarinic agonists oxotremorine and carbachol, a close for the mAChR found in pancreatic acinar cell preparations, as with tural relative of the natural neurotransmitter acetylcholine. expected for HM4 (Hootman et al., 1985). Interestingly, ex- struc len in Figure 2, heterogeneous binding was readily apparent pression of mAChR RNA in NG108-15 cells is restricted to As se IMl,HM2 andHM4 with carbachol, and for HM2 and HM3 HM3, suggesting that biochemical events mediated by muscarinic for oxotremorine. In each case, computer analysis of the bind- agonists in this neuron-derived cell line may largely reflect bind- with dlata eave a statisticallv better fit with a two site than a one ing to this subtype. inz Biological role of muscarinic receptor subtype diversity site model (Table I). Oxotremorine exhibited substantially higher The studies presented here demonstrate that the human gen- affinity than carbachol for both classes of sites for a given human ome contains a family of at least four evolutionarily conserved mAChR subtype (Table I), consistent with binding studies per- mAChRs which differ strikingly in their primary structures, their formed with tissue preparations (Schimerlik and Searles, 1980; ability to bind various ligands, and in their pattern of tissue- Peterson and Schimerlik, 1984). The low affinity binding states specific expression, revealing a heretofore unappreciated degree calculated for either agonist were similar for each of the four of specialization in muscarinic cholinergic function. Despite the mAChR subtypes, with KD values ranging from 130 to 560yM substantial amino acid divergence between the four mAChR for carbachol, and 4.0 to 5.3 ,uM for oxotremorine. The high subtypes, 151 of 241 residues (65%) contained within the seven affinity carbachol binding state was notably higher for HM2 (KD predicted membrane-spanning domains and their short connecting = 0.1 M) than for HM1 and HM4 (KD = 7.3 and 8.1 AM loops are identical in each. The majority of amino acid differences respectively), and in the range of the high affinity carbachol sites among the four receptor subtypes within these regions correspond reported for rat cerebral Ml and porcine atrial M2 mAChRs to conservative substitutions, supporting the notion that these (KD = 1.41 and 0.83 respectively) (Fisher et al., 1983; IAM sequences constitute the receptor ligand-binding domain. The Peterson and Schimerlik, 1984; Collins and Crankshaw, 1986). most unique structural feature of each subtype is the large loop The high affinity binding states of HM2 and HM3 for oxotre- between the fifth and sixth transmembrane regions, which is morine were comparable (KD = 8.0 and 22.0 nM respective- predicted to reside on the cytoplasmic side of the plasma mem- ly) and similar to the porcine atrial M2 mAChR (KD = 2.3 nM) brane, and is thus likely to play an important role in specifying (Peterson and Schimerlik, 1984). Significantly, the appearance which effector system(s) will couple to a given subtype. This of multiple affinity states for either oxotremorine or carbachol possibility is supported by studies with (3-adrenergic receptor upon transfection of each human mAChR subtype cannot be deletion mutants showing that deletions of the corresponding attributed to endogenous mAChRs since their level (< 200 recep- domain of this receptor abolish its ability to regulate adenylyl tors/cell) cannot account for even the smallest population of high cyclase activity without impairing ligand binding (Dixon et al., agonist binding sites observed in these experiments. Fur- affinity 1987). A greater degree of similarity within the 5-6 loop region the variable proportion of high and low affinity (or thermore, is evident between the HM1 and HM4 subtypes, and the HM2 of high affinity) binding sites detected for each human absence and HM3 subtypes, suggesting that both of these pairs of sub- mAChR subtype suggests that the four subtypes may differ in types share greater functional similarity. their abilities to couple efficiently with the G proteins present The ligand binding properties of human mAChR subtypes ex- within these cells. 3927 Human muscarinic acetylcholine receptors mammalian as well as the thus suggest that the HM3 be pressed by transfected cells, pattern subtype may preferentially coupled to adenylyl cyclase. The pancreas-specific expression of HM4 of gene expression detected for the homologous rat mRNAs, clues to the relationship between the subtypes suggests that this mAChR be to the provide important subtype may coupled poly- identified by earlier phosphoinositide turnover and calcium mobilization associated characterized here and those investigators. analysis of mRNA from rat tissues revealed that with the carbachol stimulated secretory activities of pancreatic Northern blot are in the heart acinar cells et Sekar et only the HM2 and HM4 homologues expressed (Korc al., 1987; al., 1987). corre- In tissues in which different mAChR subtypes may be co- and pancreas, respectively, suggesting that these subtypes 'glandular' human mAChR the contribution of to the exist- spond to the 'cardiac' and subtypes. expressed, subtype heterogeneity subtype is also supported by ence of multiple agonist binding affinity states cannot easily be The assignment of the HM2 ligand HM2 discerned. The here for cells trans- binding analysis demonstrating that the recombinant poly- binding experiments presented a fected with individual subtypes conclusively demonstrate that exhibits low affinity for pirenzepine, and high propor- peptide sites for AF-DX1 16, two distinguishing multiple agonist affinity states are the property of all four mAChR tion of high affinity M2 receptors in mammalian atria (Hammer et polypeptides. High affinity agonist binding by mAChRs is be- characteristics of al., 1980; Giraldo et al., 1987). In addition, HM2 shares virtual lieved to reflect interactions with G proteins since the proportion with the M2 mAChR protein purified from of such high affinity sites can be substantially decreased by the amino acid identity atria (Peralta et al., 1987). The identity of HM1 is in- addition of guanine nucleotides or by pertussis porcine toxin-catalyzed affinity for pirenzepine and expression in ADP-ribosylation, treatments which uncouple receptors from G dicated by its high HM1 exhibits virtual identity with proteins (Stryer and Boume, 1986). The differences in the cerebral cortex. Furthermore, affinity MI clones corresponding to the major subtype porcine mAChR and proportion of the multiple agonist affinity states detected for porcine cerebral cortex (Kubo et al., 1986a). While individual mAChR subtypes expressed purified from in a single cell type may HM1 the highest pirenzepine affinity of the four human reflect the capacity of individual subtypes to interact differentially exhibits ability of the HM3 and HM4 subtypes to bind with G proteins, and thus to couple specifically with the various mAChRs, the with relatively high affinity, and their comparable biochemical effector systems present. pirenzepine to HM1 in brain, suggests that previous studies of In conclusion, our results demonstrate that the human mAChR abundance Ml displaying high pirenzepine affinity may have gene family is composed of at least four members; specific struc- brain mAChRs to discriminate between these three mAChR subtypes. tural features distinguish each receptor and are manifested in the failed In to the preferential expression of the HM2 and HM4 distinct ligand-binding properties displayed by individual contrast sub- respectively, all four mAChR sub- types expressed by transfection of mammalian cells. This strategy subtypes in heart and pancreas extents in whole rat brain. This will be further employed to study the biochemical properties of types are expressed to similar nature of synaptic signalling each mAChR and thereby continue the investigation of difference may reflect the complex subtype, in nevous system compared with the per- the physiological significance of receptor subtype multiplicity. that occurs the central indicate a physiological require- ipheral nervous system, or may Materials and methods receptor subtypes which may regulate different ment for multiple expression of mAChRs within effector systems. Subtype-specific Isolation of molecular clones central nervous is suggested by previous studies, which the system A human genomic library (Lawn et al., 1978) was screened under low stringency hybridization conditions with a 680-bp uniformly 32P-labeled restriction that M I and M2 mAChRs occupy postsynaptic and pre- fragment indicate encoding the first five transmembrane domains of the porcine M2 muscarinic within cholinergic nerve terminals synaptic locations respectively receptor described as (Peralta et al., 1987). Positive clones were screened with of the cerebral cortex (Mash et al., 1985). Moreover, the effects 32P-end-labeled oligonucleotide pools, each containing 64 degeneracies, encoding of acetylcholine on ion conductances within the central nervous either amino acids 55-62 (Gln-Thr-Val-Asn-Asn-Tyr-Phe-Leu) or 64-70 (Ser- system have been used to correlate mAChR subtypes with specific Leu-Ala-Cys-Ala-Asp-Leu) of the porcine atrial M2 muscarinic receptor (Peralta et al., 1987) and with oligonucleotides encoding unique regions of the porcine electrophysiological functions; M2 mAChRs appear to mediate M1 (amino acids 259-286) and M2 (amino acids 240-263) mAChRs. Restriction an increase in potassium conductance within brain cells, while fragments from each genomic clone which hybridized with the oligonucleotide MI mAChRs are associated with an inhibition of potassium per- probes described above were subcloned into M13 vectors and sequenced by the meability (Egan and North, 1986). The predominant expression chain termination method (Sanger et al., 1980; Messing et al., 1981). Nucleotide sequence data will be submitted to the EMBL/GenBank Data Libraries. of a single mAChR subtype, HM3, in NG108-15 neuronal cells and Nirenberg, 1975) suggests the interesting possi- (Matsuzawa Protein sequence homologies that of mAChR subtype expression may be highly bility regulation Sequence comparisons were first determined by separate alignment of each pair of receptors (Needleman and Wunsch, 1970; Fitch and Smith, 1983). This analysis differentiated at the level of the individual neuron as well as in revealed a high degree of identity within the predicted transmembrane domains gross regions of the brain. and short hydrophilic loops of each pair of mAChR and thus allowed the alignment Earlier studies have revealed correlations between the occu- of all four sequences by visual inspection. The large cytoplasmic domain joining pation of a given receptor subtype which has been defined by the putative fifth and sixth transmembrane domains was analyzed separately for ligand specificity, and the activation of a specific cellular effector. both amino acid and nucleotide homology. The scoring parameters are those of Dayhoffet al. (1983) with a deletion penalty identical to that of Lipman and Pearson For example, preferential coupling of the MI subtype to poly- (1985). phosphoinositide turnover in rat brain and mouse anterior pituitary Expression of human mAChRs cells, and of the M2 subtype to inhibition of adenylyl cyclase Each mAChR coding sequence was inserted into a derivative of pF8CIS9080 in rat heart or mouse anterior pituitary cells and activation of containing EcoRI, SmaI and BamHI linker sites replacing the Factor VIII coding the inward potassium channel of myocardium have been reported region (Eaton et al., 1986). The following coordinates indicate the nucleotide (Gil and Wolfe, 1985; Pfaffinger et al., 1985; Watson et al., sequences of each receptor that were inserted into the expression vector: positions 1986). The pattern of expression observed for the HM3 and HM4 -66 to + 1418, HM1; bp -42 to + 1467, HM2; bp -81 to + 1598, HM3; bp -7 to + 1795, HM4 (Figure 2). For transient expression of mAChRs, 60-mm mAChRs suggests specific roles in effector coupling for these plates of human embryonic kidney cells (50% confluent) were transfected as subtypes as well. Carbachol stimulation of NG108-15 cells results described et method (Wigler previously (Eaton al., 1986) by the calcium phosphate in with no effect on inhibition, adenylyl cyclase et hours after PBS contain- apparent phos- al., 1979). Sixty transfection cells were harvested in and our 5 mM EDTA and tested in phoinositide hydrolysis Harden, ligand-binding assays. (Hughes 1987); findings ing 3928 E.G.Peralta et al. Ligand-binding analysis and 118- Pharm. 7her., 240, Korc,M., Ackerman,M. Roeske,W. (1987) Exp. cells or with cell as Binding studies were carried out with intact homogenates, intact Mishina,M., earlier (Peralta et al., 1987). Incubations were 75 min at 37°C for Kubo,T., Fukuda,K., Mikami,A., Maeda,A., Takahashi,H., Haga, described Hirose,T. and Numa.S. cells and 60 min at 20-22°C for cell homogenates. Non-specific binding was K., Ichiyama,A., Kangawa,K., Kojima,M., Matsuo,H., 411-416. determined in the presence of 10 atropine and was always < 10% of the (1986a) Nature, 323, /sM total [3H]QNB binding. The mean KD values for [3H]QNB binding and receptor Takahashi,H., Kubo,T., Maeda,A., Sugimoto,K., Akida,I., Mikami,A., Haga,K., Hirose,T. and Fed. numbers expressed per transfected cell were determined by Scatchard analysis Numa.S. (1986b) Ichiyama,A., Kangawa,K., Matsuo,H., Biochem. Soc. 367-373. of [3H]QNB saturation binding by intact cells (n = 4) utilizing the LIGAND Eur. Lett., 209, J. Mol. 105-132. program (Munson and Rodbard, 1980) and are as follows: HM1, 22 A 8 pM, and Kyte,J. Doolittle,R. (1982) Biol., 57, and Maniatis,T. 1157- 9000 ± 2100 sites/cell; HM2, 83 + 41 pM, 48 600 + 1400; 39 6 Cell, 15, HM3, + pM, Lawn,R., Fritsch,E., Parker,R., Blake,G. (1978) 27 400 ± 7600; HM4, 112 i 55 pM, 89 600 21 700. Similar results for and 1435-1441. [3H]QNB-binding were obtained for cell homogenates (n = 2) with the range Lipman,D. Pearson,W. (1985) Science, 227, and Sambrook,J. Molecular A Labora- of values as follows: 17 pM, 13 fmol/mg homogenate protein (n = 1); HMI, Maniatis,T., Fritsch,E.F. (1982) Cloning: Manual. Cold Harbor New York. HM2, 26 10 pM, 155 ± 31 fmol/mg; HM3, 37 29 pM, 44 A20 fmol/mg; L tory Spring Laboratory Press, and 115-117. HM4, 172 i 73 pM, 332 241 fmol/mg. Antagonist competition displacement Mash,D., Flynn,D. Science, Potter,L. (1985) 22?, and Proc. Acad. Sci. experiments were performed with a subset of the transfections described above Matsuzawa,H. Natl. USA, 72, Nirenberg,M. (1975) 3472-3476. and the mean KD values for [3H]QNB binding and receptor numbers are: HMl, Nucleic Acids 309-321. and 19.5 ± 1 pM, 7100 ± 3300 sites/cell (n = 2); HM2, 95 i 60 pM, 58 100 Res., 9, Messing,J., Crea,R. Seeburg,P. (1981) and Anal. Biochem., 220-239. + 13 800 (n = 3); HM3, 45 ± 9 pM, 16 400 + 3200 (n = 3); 112 Munson,P. 107, HM4, Rodbard,D. (1980) and 193-200. + 83 pM, 82 900 28 800 (n = 3). The agonist competition Science, displacement Nathans,J., Thomas,D. Hogness,D. (1986) 232, Annu. Rev. 195-236. experiments were performed with cells from three transfections and the values Neurosci., KD Nathanson,N. (1987) 10, and J. Mol. 443-453. for [3H]QNB binding and receptor numbers for one Biol., representative experiment Needleman,S. Wunsch,C. (1970) 48, are as follows: HM1, 42 pM, 13 250 sites/cell; 57 78 Ramachandran,J., HM2, pM, 000; HM3, Peralta,E., Winslow,J., Peterson,G., Smith,D., Ashkenazi,A., and 600-605. 27 pM, 34 200; HM4, 16 pM, 133 000. The KD Science, values for competing antagon- Schimerlik,M. Capon,D. (1987) 236, and 33-74. Biochem., 14, ists and agonists were determined from three transfections and displacement exper- Peterson,G. Schimerlik,M. (1984) Prep. and Hille,B. Nature, iments performed in duplicate (with the exception of cholinergic binding to HM 1 Pfaffinger,P., Martin,J., Hunter,D., Nathanson,N. (1985) 536-538. 317, which was analyzed in duplicate in one experiment); equivalent results were and J. Mol. Roe,B. Biol., obtained and the values of one representative experiment are presented in Table Sanger,F., Coulson,A., Barrel,B., Smith,A. (1980) 161-178. I. Untransfected cells expressed <200 muscarinic receptors/cell (< 1.3 fmol/mg 143, and 3407-3413. 19, homogenate protein). Schimerlik,M. Searles,R. (1980) Biochemistry, Dixon,J. and J. Biol. 340-344. Chem., 262, Sekar,M., Hokin,L. (1987) Northern hybridization analysis and Annu. Rev. Cell 391-419. Biol., 2, Stryer,L. Bourne,H. (1986) Subtype-specific restriction fragment probes encoded the following unique regions The In and Taylor,P. (1985) Goodman,L. Gilman,A. (eds), Pharmacological of the large cytoplasmic domain amino acids 265- of each receptor: HM1, Basis New 100. of 7herapeutics. Macmillan, York, p. 363; HM2, amino acids 268-381; HM3, amino acids 265-394; HM4, amino and Watson,M., Roeske,W., Vickroy,T., Smith,T., Warmsley,J. Yamamura,H.I. acids 292-390. with Each fragment was uniformly radiolabeled [a-32P]dCTP Trends Pharmacol. 7 46-55. (1986) Sci., (Suppl.), (Maniatis et al., 1982). Stringent hybridization conditions for Southern and North- and Wigler,M., Pellicer,A., Axel,R., Urlaub,G. Chasin,L. (1979) Silverstein,S., ern filters are as previously described (Peralta et al., 1987). Proc. Acad. Sci. 1373-1376. Natl. USA, 76, Brandt,D., Burnier,J., Harkins,R., Yarden,Y., Rodriguez,H., Wong,S., May,D., and Proc. Natl. Acad. Ross,E. Chen,E., Ramachandran,J., Ullrich,A. (1986) Acknowledgements Sci. 6795-6799. USA, 83, for Dr P.Moore for We thank Drs J.Witkin and C.Gorman helpful discussions, 851-858. and Cell, Zuker,C., Cowman,A. Rubin,G. (1985) 40, advice on computer analysis of ligand-binding data, S.Marsters for tissue culture assistance, M.Vasser, P.Jhurani and P.Ng for oligonucleotide synthesis, Drs on 1987 Received September 14, for mRNA Drs and A.Rosenthal and C.Logsdon providing samples, P.Seeburg and C.Morita and S.Peters for P.Braun for communicating unpublished data, prep- added in Note proof aration and AF-DX1 16 < of figures. Pirenzepine 11-[[2-(diethylamino)methyl] of this Bonner et al. - the submission 1-piperidinyllacetyl > -5,1 1-dihydro-6H-pyrido[2,3,6][1,4]benzodiazepine- Following original (Science, 235, manuscript the and of three 6-one > Drs G.Trummlitz and W.Reuter. were kindly provided by R.Hammer, 527-532, 1987) reported cloning partial sequence analysis Rml Rm3 and Hm4 of their Inc. and CA16417 to J.R. mAChR This work was supported by Genentech, NIH grant subtypes. (rat), (human). Comparison (rat) here indicates that Hm4 is to with the equivalent subtypes mnAChRs reported Rml and Rm3 are rat of HM1 and HM4 and that HM3, suggests homologues References the rat warrants further since of respectively. Assignment subtypes investigation Southern that there be five additional Alonso,S., Minty,A., and J. Mol. Evol., 23, Bonner et al. showed Bourlet,Y. Buckingham,M. (1986) by analysis may rat 11-22. uncharacterized muscarinic genes. Birdsall,N. and Trends Pharmacol. Sci., 4, 459-463. Hulme,E. (1983) Brown,J.H., and Mol. Pharmacol., 27, 525-531. Goldstein,D. Masters,S. (1985) Cohen,S. and Trends Pharmacol. Sci., 8, 1-4. Sokolovsky,M. (1987) Collins,S. and Am. J. 251, 546-552. Crankshaw,D. (1986) Physiol, Dayhoff,M., Barker,W. and Methods 91, 524-545. Hunt,L. (1983) Enzymol., Dixon,R., Kobilka,B., Dolhman,H., Frielle,T., Bolan- Strader,D., Benovic,J., owski,M., Bennett,C., Mumford,R., Slater,E., Rands,E., Diehl,R., Sigal,I., Caron,M., Lefkowitz,R. and Nature, 321, 75-79. Strader,C. (1986) Dixon,R., Sigal,I., Candelore,M., Blake,A. and Strader,C. Rands,E., Register,R., (1987) Nature, 326, 73-77. Eaton,D.L., Wood,W., Eaton,D., Wion,K., Mather,J., Hass,P., Hollingshead,P., Lawn,R., Vehar,G. and Gorman,C. (1986) Biochemistry, 26, 8343-8347. Egan,T. and North,R. (1986) Nature, 319, 405-407. Fisher,S., Klinger,P. and Agranoff,B. (1983) J. Biol. Chem., 258, 7358-7363. Fitch,W. and Smith,T. (1983) Proc. Natl. Acad. Sci. USA, 80, 1382-1386. Gil,D. and Wolfe,B. (1985) J. Phann. Exp. Ther., 232, 608-616. Giraldo,E., Hammer,R. and Ladinsky,H. (1987) Life Sci., 40, 833-840. Hammer,R., Berrie,C., Birdsall,N., Burgen,A. and Hulme,E. (1980) Nature, 283, 90-92. Hootman,S., Picardo-Leonard,T. and Burnham,D. (1985) J. Biol. Chem., 260, 4186-4194. Harden,T. J. Pharm. Ther., 237, 173-178. Hughes,A. and (1987) Exp.

Journal

The EMBO Journal – Springer Journals

Published: Dec 1, 1987

There are no references for this article.