Muscarinic acetylcholine receptors: Structural basis of ligand binding and G protein coupling

Life Sciences, Vol. 56, No.s 11/l& pp. 915-922, 1995 1995 Elswier Science Ltd Printed in the USA. All rights reserved 0024-3205/95 $950 + .OO

Pergamon

0024-3205(95)00028-3

MUSCARINIC ACETYLCHOLINE RECEPTORS: STRUCTURAL OF LIGAND BINDING AND G PROTEIN COUPLING

BASIS

Jtirgen Wess, Nathalie Blin, Ernst Mutschler* and Klaus Bltiml Laboratory of Bioorganic Chemistry, National Institute of Diabetes and Digestive and Kidney Diseases, NIH, Bldg. 8A, Room BlA-09, Bethesda, MD 20892, U.S.A., and *Department of Pharmacology, Biocenter Niederursel, University of Frankfurt, D-60053 Frankfurt, Germany

Summarv Muscarinic acetylcholine receptors (ml -m5) were studied by a combined molecular genetic/pharmacologic approach to elucidate the molecular characteristics of the ligand binding site and of the receptor domains involved in G protein coupling. Sitedirected mutagenesis studies of the rat m3 muscarinic receptor suggest that the acetylcholine binding domain is formed by a series of hydrophilic amino acids located in the “upper” half of transmembrane domains (TM) III, V, VI, and VII. Moreover, we showed that mutational modification of a TM VI Asn residue (Asn507 in the rat m3 receptor sequence) which is characteristic for the muscarinic receptor family has little effect on high-affinity acetylcholine binding and receptor activation, but results in dramatic reductions in binding affinities for certain subclasses of muscarinic antagonists. The N-terminal portion of the third intracellular loop (i3) of muscarinic and other G protein-coupled receptors has been shown to play a central role in determining the G protein coupling profile of a given receptor subtype. Insertion mutagenesis studies with the rat m3 muscarinic receptor suggest that this region forms an amphiphilic o-helix and that the hydrophobic side of this helix represents an important G protein recognition surface. Further mutational analysis of this receptor segment showed that Tyr254 located at the N-terminus of the i3 loop of the m3 muscarinic receptor plays a key role in muscarinic receptor-induced Gq activation. The studies described here, complemented by biochemical and biophysical approaches, should eventually lead to a detailed structural model of the ligandreceptor-G protein complex. Key Words:

site-directed

mutagenesis,

ligand binding

site, receptor/G

protein

coupling,

G protein

coupling

The muscarinic acetylcholine (ACh) receptors (ml-m5) are members of a huge superfamily of plasma membrane receptors which regulate cellular activity via coupling to heterotrimeric guanine nucleotide-binding proteins (G proteins) (1,2). All these receptors are predicted to share a similar three-dimensional structure consisting of a tightly packed bundle of seven transmembrane helices (TM I-VII) linked by three extracellular and three intracellular loops. Over the past couple of years, we have used a combined molecular genetic/pharmacologic approach to study how muscarinic receptors function at a molecular level. In this paper, we review some of our recent mutagenesis studies aimed at elucidating the molecular structure of the ACh binding site and the receptor/G

916

Mechanisms

of Muscarinic

Receptor

Function

Vol. 56, No.s llj12,

1995

protein interface. Amino acids involved in ligand binding Characterization of the ACh binding site Accumulating evidence suggests that ACh as well as other biogenic amine neurotransmitters such as norepinephrine, dopamine, or 5-HT bind to their target receptors in a narrow cleft enclosed by the ring-like arrangement of the seven transmembrane helices, about lo-15 A away from the membrane surface. Affinity labelling (3,4) and site-directed mutagenesis studies (5) have shown that the positively charged ammonium (amino) head group of ACh and other muscarinic ligands is likely to form an ionic bond with a negatively charged TM III Asp residue which is highly conserved among all biogenic amine receptors (Asp147 in Fig. 1). Genetic analysis of adrenergic, dopamine, 5HT, and histamine receptors (for a review, see ref. 6) suggests that this Asp residue plays a similar role (as suggested for the muscarinic receptors) in all receptors which bind biogenic amine ligands. Mutational analysis of the rat m3 muscarinic receptor (78) has recently shown that high-affinity binding of ACh to muscarinic receptors critically depends on the presence of four Tyr and two Thr residues which are conserved among all muscarinic receptors which have been cloned so far (Fig. 1). All of these residues are located in the “upper” portion of TM III, V, VI, and VII, at a similar level as the conserved TM III Asp residue, suggesting that these amino acids define the plane in which ACh’interacts with the receptor protein. Fig. 2 shows where these residues are predicted to be located in a model of the transmembrane core of G protein-coupled receptors recently proposed by Baldwin (9,lO). This model does not rely on the use of bacteriorhodopsin as a template and is compatible with a recently published low-resolution map of rhodopsin (11). The ion pair formed between the conserved TM III Asp residue (Asp147) and the ACh ammonium head group is shown in the center of Fig. 2. One may speculate that this ion pair is stabilized by charge transfer or hydrogen bond interactions with Tyr506 (TM VI), TyrS29 (TM VII), and Tyr533 (TM VII). On the other hand, the possibility exists that Tyr148 (TM III), Thr23 I (TM V), and Thr234 (TM V) can interact with the ACh ester moiety by means of hydrogen bonding. Consistent with these findings, mutagenesis studies on other G protein-coupled receptors which bind biogenic amines and other small molecules also suggest that residues on TM III, V, VI, and VII are critically involved in ligand binding (6,9,10). Interestingly, many of these studies have shown that residues located at positions analogous to those which we have identified as being important for high-affinity ACh binding in the muscarinic receptors also play key roles in ligand binding in other receptor subclasses (10). Functional role of a conserved TM VI Am residue Based on the assumption that G protein-coupled receptors fold in a fashion similar to the structurally well-characterized bacterial membrane protein, bacteriorhodopsin, several threedimensional models of muscarinic receptors have been proposed (12-14). All these models predict that a TM VI Asn residue which is characteristic for the muscarinic receptor family is ideally positioned to interact with the ACh ester moiety by means of hydrogen bonding. To test the correctness of this hypothesis we have replaced this residue (Asn507) in the rat m3 muscarinic receptor with Ala or Asp which should disrupt potential hydrogen bond interactions with the partially negatively charged oxygen atoms of the ACh ester moiety. Pharmacological analysis of these mutant receptors in COS-7 cells showed that they were able to bind ACh as well as the ACh derivative, carbachol, with relatively high affinities which differed from the corresponding wild type values by less than 5fold (15). In addition, both mutant receptors were able to stimulate carbachol-induced phosphatidyl inositol (PI) hydrolysis in a fashion remarkably similar to that of the wild type receptor (15). We therefore conclude that the conserved TM VI Asn residue does not play a critical role in high-affinity ACh binding and agonist-induced receptor activation. These findings also suggest that the recently published three-dimensional muscarinic receptor models need

Vol. 56, No.s 11112, 1995

Mechanisms of Muscarinic Receptor Function

917

further refinement to allow more reliable predictions regarding the molecular structure of the AChreceptor complex. One possible reason for the observed discrepancy between our experimental data and the predictions made based on molecular modelling studies is that the arrangement of the TM helices in G protein-coupled receptors differs from that found in bacteriorhodopsin (11). In this context, it should also be noted that bacteriorhodopsin is not coupled to G proteins and shares virtually no sequence homology with G protein-coupled receptors. While having little effect on ACh binding and agonist-induced receptor activation, mutational modification of Asn507 resulted in mutant m3 muscarinic receptors which displayed drastically reduced affinities for most antagonists tested (15). Most notably, the binding affinities for atropinelike agents (atropine, N-methylatropine, scopolamine, and N-methylscopolamine) and the tricyclic compound, pirenzepine, were reduced by factors ranging from 235 to 28,000. Interestingly, the binding affinities of some other muscarinic antagonists including (-)-quinuclidinyl benzilate (which was used as tritiated radioligand in this study) and trihexyphenidyl were much less affected by the various Asn507 mutations. These data suggest that the conserved TM VI Asn residue plays a key role in the binding of certain subclasses of muscarinic antagonists. It remains unclear at present

N-Glycosylation \

TL-

I

II

-LG

IV

V

VI

VII

M

0

(NtL,

III NRWA

H T

ii

I I

CDS c I PKT Y

F

9

R TK

i2

E

\ ---

i3

/

K SL’

Extrace”u’ar

Y

’

(COOH)

Intracellular

Fig. 1 Predicted transmembrane topology of the rat m3 muscarinic receptor (modified according to ref. 35). Amino acids highlighted in black are critical for high-affinity ACh binding (4,7,8). Residues marked by open circles are conserved among most G protein-coupled receptors. Tyr254 which plays a key role in muscarinic receptor-mediated stimulation of PI hydrolysis (3 1,32) is shown boxed. Amino acids discussed in the text are numbered. Only the membrane-proximal portions of the N-terminal receptor domain (total length: 66 aa), the i3 loop (total length: 240 aa), and the C-terminal tail (total length: 43 aa) are shown. In total, the rat m3 muscarinic receptor is composed of 589 amino acids (34).

918

Mechanisms

of Muscarinic

Receptor

Function

Vol. 56, No.s 11/12, 1995

Fig. 2 Arrangement of TM I-VII in the rat m3 muscarinic receptor according to the “Baldwin model” (9,lO). The view is from the intracellular side of the membrane. Only the first I8 amino acids of the upper portion of TM I-VII (facing the extracellular surface of the membrane) are shown. ACh was “docked” manually into the binding cleft formed by TM III-VII. Residues which have been implicated in high-affinity ACh binding (4,7,8) are highlighted by black bars. In contrast to predictions made based on molecular modelling studies (12-14), a recent mutagenesis study (15) suggests that Asn507 (TM VI; shown circled) does not play a major role in ACh binding. The positions of the a-C atoms are marked by filled circles whose size reflects their relative depth in the membrane.

whether the Asn side chain directly interacts with distinct functional groups on the antagonist molecules or whether it only stabilizes a receptor conformation required for high-affinity antagonist binding. To distinguish between these two possibilities, the binding affinities of a greater number of antagonists containing systematic structural modifications need to be studied.

Receptor reeions involved in G protein coupling By a mechanism that is as yet unknown, ACh binding to the transmembrane core of the muscarinic receptors triggers conformational changes in the receptor proteins which enable distinct intracellular receptor domains to recognize and activate specific sets of G proteins. Mutagenesis studies

Vol. 56, No.s 11/12, 1995

Mechanisms of Muscarinic Receptor Function

919

with muscarinic and other G protein-coupled receptors have shown that several of the about twenty amino acids which are highly conserved among most G protein-coupled receptors (Fig. 1) are essential for ligand-induced receptor activation. Such residues include, for example, a TM II Asp residue (5), the members of the Asp-Arg-Tyr motif located at the C-terminus of TM III (5,16), a TM V Phe (Tyr) residue (6), and a TM VII Pro residue (17). One may speculate that these highly conserved amino acids (most of which are located in the “lower” half of the transmembrane receptor core facing the cytoplasm) may provide a common structural framework through which the conformational activation of most G protein-coupled receptors occurs. A great number of mutagenesis studies as well as experiments with peptides that can mimic or inhibit receptor interactions with G proteins have implicated four distinct intracellular receptor regions in G protein coupling (for reviews, see refs. 6,18,19). These include the second intracellular loop (i2), the N- and the C-terminal portions of the third intracellular loop (i3), and the membrane-proximal portion of the C-terminal tail (i4). These receptor domains are thought to form a three-dimensional surface which, upon agonist binding to the receptor, can functionally interact with specific G proteins. Functional and structural characteristics of the N-terminal segment of the i3 loop Based on their G protein-coupling profiles, the five muscarinic receptors can be classed into two major functional categories. The ml, m3, and m5 receptors are preferentially linked to G proteins of the Gq/Gll family (20-22) which mediate the activation of several isoforms of PLCP resulting in the breakdown of phosphatidyl inositol lipids (PI hydrolysis). The m2 and m4 receptors, on the other hand, are efficiently coupled to G proteins of the Gi/Go class (22,23) which, at a biochemical level, mediate the inhibition of adenylyl cyclase. Studies with chimeric m2/m3 muscarinic receptors have shown that the first 16-21 amino acids of the i3 loop are critically involved in determining G protein coupling specificity (24-26). It could be demonstrated that exchange of this receptor segment between the m2 and m3 receptors was able to confer on the resultant mutant receptors the ability to interact with the same G proteins as the wild type receptor from which this short sequence was derived. Sequence analysis of the N-terminal portion of the i3 loop of the muscarinic and many other G protein-coupled receptors shows that this region is particularly rich in charged residues. These charged amino acids alternate with non-charged residues with a certain periodicity suggesting that the N-terminal portion of the i3 loop forms an amphipathic cc-helical extension of TM V (27). To test the correctness of this hypothesis, we have functionally characterized a series of five mutant m3 muscarinic receptors (m3(+1A) to m3(+5A)) in which l-5 additional Ala residues were inserted at the N-terminus of the i3 loop (following Arg252; Fig. l)(ref. 28). If the N-terminal segment of the i3 loop is in fact o-helically arranged, one would expect that the insertion of extra Ala residues between TM V and the beginning of the i3 loop should lead to a rotation of the adjacent helical segment thus severely affecting receptor/G protein coupling. Functional analysis of the various mutant receptors in COS-7 cells showed that m3(+1A), m3(+3A), and m3(+4A) were still able to stimulate PI hydrolysis, although with somewhat reduced efficiencies as compared with the wild type receptor. In contrast, m3(+2A) and m3(+5A) proved to be functionally virtually inactive. Helical wheel models show that this pattern is fully consistent with the notion that the N-terminal segment of the i3 loop is in fact o-helically arranged and that one gide of this helix is involved in G protein interactions (28). Based on several recent mutagenesis studies which exclude the charged residues located in this receptor region from playing specific roles in G protein coupling (29-3 l), it is highly likely that the non-charged surface of this putative o-helix is involved in G protein recognition. Since the stepwise insertion of extra Ala residues at the beginning of the i3 loop is predicted to lead to a clockwise rotation of the adjacent a-helical segment in 1000 increments, the hydrophobic surface of this helix faces a direction approximately opposite to that found in the wild type

Mechanisms

920

of Muscarinic

Receptor

Vol. 56, No.s 11/12,

Function

1995

receptor in the two functionally inactive mutant receptors, (m3(+2A) and m3(+5A), and can thus no longer be recognized by G proteins. However, in the functionally active mutant receptors, m3(+1A), m3(+3A), and m3(+4A), the position of this surface differs by only 40-1000 from the wild type orientation and therefore still allows productive G protein coupling. The insertion mutagenesis approach described here should be useful in studying the potential a-helicity of other functionally important receptor domains. Amino acids criticalfor muscarinic receptor-mediated activation of Gq Given the pivotal functional importance of the N-terminal portion of the i3 loop, experiments were designed to identify specific amino acids located within this region which are of particular importance for muscarinic receptor-mediated activation of Gq (31,32). Initially, we replaced short segments of the N-terminal region of the i3 domain of the rat m3 muscarinic receptor with the corresponding m2 receptor sequences and studied the effect of these substitutions on m3 receptormediated PI hydrolysis. We found that substitutions comprising amino acids 5-16 (Glu256Leu267) of the i3 loop of the m3 receptor were functionally well tolerated (31). In contrast, all mutant receptor constructs in which the N-terminal four amino acids of the i3 loop contained m2 receptor sequence were severely impaired in their ability to mediate stimulation of PI hydrolysis. Functional analysis of this four amino acid-segment by site-directed mutagenesis showed that Tyr254 (located at position 3 of the i3 loop of the m3 receptor) plays a key role in muscarinic receptor-mediated Gq activation (3 1,32). This residue is conserved among all Gq-coupled muscarinic receptors (ml, m3, and m5), but is replaced with a Ser residue in the G&coupled receptor subtypes (m2 and m4). Whereas replacement of Tyr254 with Ser or Ala resulted in mutant m3

m

0

m3

m2

120m3(w) L E

CR1

80

% Q) t ,r 40

0 -9 CRl+Y

-8

I

I

I

I

I

I

I

-7

-6

-5

-4

-3

-2

-1

Carbachol,

log M

Fig. 3 Functional importance of Tyr254 (rat m3 receptor sequence) for muscarinic receptormediated stimulation of PI hydrolysis (31,32). In CRI, the first 16 amino acids (RIYKETEKRTKELAGL) of the i3 loop of the rat m3 muscarinic receptor are replaced with the corresponding human m2 receptor sequence (HISRASKSRIKKDKKE). In CRl+Y, the Ser residue present at position 3 of the i3 loop of CR1 (corresponding to position 254 in the rat m3 receptor sequence) is replaced with Tyr. Data are presented as % increase in inositol monophosphate (IPI) above basal levels (Emax of wild type m3 = 100%). A representative experiment is shown (n=3).

Vol. 56, No.s 11/12, 199.5

Mechanisms of Muscarinic Receptor Function

921

receptors with strongly reduced functional activity, replacement with other aromatic amino acids such as Phe or Trp yielded fully functional receptors (3 1,32). Consistent with these findings, we could demonstrate that insertion of Tyr254 into a mutant m3 muscarinic receptor in which the first 16 amino acids of the i3 loop were replaced with m2 receptor sequence was able to confer on this functionally inactive mutant receptor the ability to efficiently stimulate PI hydrolysis (Fig. 3; ref. 32). One may speculate that Tyr254 (or another aromatic residue instead) can interact with specific sites on the C-terminal portion of the ct subunit of Gq, a region known to play a key role in determining the specificity of receptor/G protein interactions (33). To test the hypothesis that substitution of Tyr254 into the homologous position of the human m2 muscarinic receptor is sufficient to confer on the resultant mutant receptor the ability to activate Gq, an m2(Ser210->Tyr) mutant receptor was constructed and functionally analyzed in COS-7 cells (32). However, we found that m2(Ser210->Tyr), similar to the wild type m2 receptor, was unable to mediate a significant stimulation of PI breakdown. Based on this finding, we speculated that other intracellular muscarinic receptor regions, besides the N-terminal segment of the i3 loop, also contribute to the specificity of receptor/G protein interactions. To test this hypothesis, we substituted the first (i 1) or the second intracellular loop (i2) as well as the C-terminal segment of i3 (Ci3) or the C-terminal tail (i4) of the m3 receptor into the m2(Ser210->Tyr) mutant receptor. Preliminary functional studies with these chimeric receptors suggest that the i2 and Ci3 regions, but not the i 1 and i4 domains, contain residues critical for muscarinic receptor-mediated activation of Gq (N. Blin and J. Wess, unpublished results). Site-directed mutagenesis studies are currently underway to identify functionally important amino acids contained within these two regions.

Conclusions The mutagenesis approach described here, complemented by biochemical, biophysical, and molecular modeling studies, should contribute to a better understanding of the molecular mechanisms governing the specificity of ligand/receptor/G protein interactions. Since G protein-coupled receptors are involved in numerous fundamental physiological and pathophysiological processes, this knowledge should eventually lead to the development of novel therapeutic agents with a more selective mechanism of action.

References 1. E.C. HULME, N.J.M. BIRDSALL and N.J. BUCKLEY, Annu. Rev. Pharmacol. Toxicol. 30 633-673 (1990). 2. J. WESS, Trends. Pharmacol. Sci. 14 308-3 13 (1993). 3. C.A.M. CURTIS, M. WHEATLEY, S. BANSAL, N.J.M. BIRDSALL, P. EVELEIGH, E.K. PEDDER, D. POYNER and E.C. HULME, J. Biol. Chem. 264 489-495 (1989). 4. T.A. SPALDING, N.J.M. BIRDSALL, C.A.M. CURTIS and E.C. HULME, J. Biol. Chem. 269 4092-4097 (1994). 5. C.M. FRASER, C.-D. WANG, D.A. ROBINSON, J.D. GOCAYNE and J.C. VENTER, Mol. Pharmacol. 36 840-847 (1989). 6. C.D. STRADER, T.M. FONG, M.R. TOTA, D. UNDERWOOD and R.A.F. DIXON, Annu. Rev. Biochem. 63 101-132 (1994). 7. J. WESS, D. GDULA and M.R. BRANN, EMBO J. M 3729-3734 (1991). 8. J. WESS, R. MAGGIO, J.R. PALMER and Z. VOGEL, J. Biol. Chem. 267 19313-19319 (1992). 9. J.M. BALDWIN, EMBO J. 12 1693-1703 (1993). 10. J.M. BALDWIN, Curr. Opin. Cell Biol. 6 180-190 (1994).

922

Mechanisms of Muscarinic Receptor Function

Vol. 56, No.s 11112, 199.5

11. G.F.X. SCHERTLER, C. VILLA and R. HENDERSON, Nature 362 770-772 (1993). 12. M.F. HIBERT, S. TRUMPP-KALLMEYER, A. BRUINVELS and J. HOFLACK, Mol. Pharmacol. 40 8-15 (1991). 13. S. TRUMPP-KALLMEYER, S.J. HOFLACK, A. BRUINVELS and M.J. HIBERT, Med. Chem. 35 3448-3462 (1992). 14. G. NORDVALL and U. J. HACKSELL, Med. Chem. 36 967-976 (1993). 15. K. BLUML, E. MUTSCHLER and J. WESS, J. Biol. Chem. 269 18870-18876 (1994). 16. S.Z. ZHU, S.Z. WANG, J. HU and E.E. EL-FAKAHANY, Mol. Pharmacol. 45 5 17-523 (1994). 17. J. WESS, S. NANAVATI, Z. VOGEL and R. MAGGIO, EMBO J. 12,331-338 (1993). 18. H.G. DOHLMAN, J. THORNER, M.G. CARON and R.J. LEFKOWITZ, Annu. Rev. Biochem. 60 653-688 (1991). 19. T.M. SAVARESE and C.M. FRASER, Biochem. J. 283 l-19 (1992). 20. A.V. SMRCKA, J.R. HEPLER, K.O. BROWN and P.C. STERNWEIS, Science =804807 (1991). 2 1. G. BERSTEIN, J.L. BLANK, A.V. SMRCKA, T. HIGASHIJIMA, P.C. STERNWEIS, J.H. EXTON and E.M. ROSS, J. Biol. Chem. 267 8081-8088 (1992). 22. S. OFFERMANNS, T. WIELAND, D. HOMANN, J. SANDMANN, E. BOMBIEN, K. SPICHER, G. SCHULTZ and K.H. JAKOBS, Mol. Pharmacol. f15 890-898 (1994). 23. E.M. PARKER, K. KAMEYAMA, T. HIGASHIJIMA and E.M. ROSS, J. Biol. Chem. 266 519-527 (1991). 24. J. WESS, M.R. BRANN and T.I. BONNER, FEBS Lett. 258 133-136 (1989). 25. J. WESS, T.I. BONNER, F. DGRJE and M.R. BRANN, Mol. Pharmacol. 38 517-523 (1990). 26. J. LECHLEITER, R. HELLMISS, K. DUERSON, D. ENNULAT, N. DAVID, D.CLAPHAM and E. PERALTA, EMBO J. 9 438 l-4390 (1990). 27. C.D. STRADER, I.S. SIGAL and R.A.F. DIXON, FASEB J. 3 1825-l 832 (1989). 28. K. BLtiML, E. MUTSCHLER and J. WESS, Proc. Natl. Acad. Sci. USA 917980-7984 (1994). 29. J.R. ARDEN, 0. NAGATA, M.S. SHOCKLEY, M. PHILIP, J. LAMEH and W. SADEE, Biochem. Biophys. Res. Commun. 188 11 I l-l 115 (1992). 30. A.H. CHEUNG, R.-R. C. HUANG and C.D. STRADER, Mol. Pharmacol. 41 1061-1065 (1992). 31. K. BLtiML, E. MUTSCHLER and J. WESS, J. Biol. Chem. 269 402-405 (1994). 32. K. BLtiML, E. MUTSCHLER and J. WESS, J. Biol. Chem. 269 11537-I 1541 (1994). 33. B.R. CONKLIN, Z. FARFEL, K.D. LUSTIG, D. JULIUS and H.R. BOURNE, Nature 363 274-276 (1993). 34. T.I. BONNER, N.J. BUCKLEY, A.C. YOUNG and M.R. BRANN, Science 237 527-532 (1987). 35. J. WESS, Life Sci. 53 1447-1463 (1993).