membrane domain charged amino acids have been found to be important for function in several G-linked receptors. Mutagen- esis experiments with these other ...
THEJOURNAL OF BIOLOGICAL CHEMISTRY
Vol. 269, No. 32, Issue of August 12, pp. 20548-20553, 1994 Printed in U.S.A.
-1-1 Opiate Receptor CHARGED TRANSMEMBRANE DOMAIN AMINOACIDS ARE CRITICAL FOR AGONIST RECOGNITION AND INTRINSIC ACTMTY* (Received for publication, May 11, 1994,and in revised form, June 6, 1994)
Christopher K. Surratt, Peter S . Johnson$, Akiyoshi Moriwaki, Brian K. Seidleck, Carrie J. Blaschak, J i a Bei Wang, and George R. Uhlg From the Molecular Neurobiology Branch and Office of the Director, Intramural Research Program, National Institute on Drug Abuse and the Departments of Neurology and Neuroscience, Johns Hopkins University School of Medicine, Baltimore, Maryland 21224
The p opiate receptor is a principal brain site for ac- in circuits important for behavioral reward (3,4).Activation of tivities of morphine, other opiate drugs, and opioid pep- the p opiate receptor can reduce stimulated activities of adentides in modulating pain and altering mood. Recent ylate cyclase, open inwardly-rectifyingpotassium channels, cloning of cDNAs encoding rat and human p receptors and cause closure of calcium channels (1, 5-10). The binding reveals charged amino acid residues within putative properties of this receptor are alteredby GTP analogs, suggesttransmembrane domains (TMs) 11,111, and VI, a substan- ing G protein linkage (11-15). Opiate actions at thep receptor tial N-terminal extracellular domain, and a C-terminal also trigger tolerance and dependence on opiate drugs in fashintracellular domain. Deletion of 64 N-terminal amino ions that are currentlypoorly understood. Understanding the acids producedlittle effect on receptor function (Wang, molecular mechanisms that underlie activity of opiate drugs J. B., Imai, Y.,Eppler, C. M., Gregor, P., Spivak, C. E.,and and opioid peptides at the p opiate receptor is, thus, of subUhl, G. R. (1993)Proc. Natl. Acad. Sci. U. S.A. 90,10230- stantial interest. 10234). Further deletion of 33 C-terminal amino acids Elucidation by cDNA cloning of the sequences of rodent and yielded a receptor at which morphine,but not the substituted enkephalin DAMGO ([~-Ala~,MePhe",Gly-human p opiate receptors, as well as the related6 and K receptors, reveals that each is a member of the seven-transmem~l~lenkephalin), inhibited adenylate cyclase.Alanine substitution for each charged TM residue in the N-ter- brane domain G protein-linked receptor family(1, 16-21). The p receptors contain domainsof polar amino acid residues likely minallydeleted receptor reducedaffinitiesformorphine, DAMGO, and the opiate antagonist naloxone. Re- to represent nontransmembrane regions. Such nontransmemplacement of TM I1 Asp"" with asparagine or glutamic brane regions might participate inligand recognition, Na+-meacid increased p receptor affinity for naloxone. TM I1 diated effects on agonist binding, or G protein coupling as has and TM I11 glutamic acid substitutions for Asp"" and been inferred from studies of other G protein-coupled receptors Asp'47reduced agonist binding affinities but allowed full (22-24). The p receptors also contain charged amino acid resiinhibition of adenylate cyclase at high agonist concen- dues within putative transmembrane domains. These transtrations. TM VI histidine substitution with alanine membrane domain charged amino acids havebeen found to be yielded a receptor that produced almost twice the cy- important for function in several G-linked receptors. Mutagenclase inhibition displayed by the wild type receptor in esis experiments with these other receptors suggest that the parallel transient expression assays. These findings un- charged residues create hydrophilic pockets within the transderscore the importance of charged residues in TM 11, membrane portion of the protein that contribute ligand to bind111, and VI for different receptor functions and the mod- ing sites (25-27). est involvement of extensive portions of N- and C-termiThe p opiate receptor includes charged amino acidsin three nal receptor domains in these processes. positions likely to represent transmembranedomains (Fig. 1). A highly conserved aspartic acid residue (Asp114)in TM I1 lies in a position at which the analogous residue has demonstrated The p opiate receptor (pOR)I is the receptor subtype most importance for agonist binding by the qadrenergic receptor identified with the analgesic properties of opiate drugs; p re- and by the enkephalin-preferring but relatively morphine-inceptor agonists canmodify pain in virtually every test of spinal sensitive 6 opiate receptor (25, 28). This residue has also been and supraspinal analgesia (2). p opiate receptor agonists are postulated t o play arole in theNa+-mediated decreasein opiate prominently implicated in mechanisms of opiate-induced re- receptor agonist binding affinity (25, 28-30). The p receptor of dopamine possesses a TM I11 aspartic acid (Asp'47),a feature common to ward andreinforcement and can stimulate release * This work was supported by the Intramural Research Program of several G-linked receptors that bind amine-containing ligands the National Institute on Drug Abuse and the PRAT program of the (25-27). Mutagenesis data and the selective presence of this NIGMS, National Institutes of Health. The costs of publication of this amino acid in receptors for amine-containing ligands havesugarticle were defrayed in part by the payment of page charges. This gested that the P-adrenergic receptor TM I11 residue Asp"3 article must thereforebe hereby marked "advertisement"in accordance forms an ion pair with thepositively charged primary amineof with 18 U.S.C.Section 1734 solely to indicate this fact. norepinephrine (27). Finally, the p receptor contains a pre$ Pharmacology Research Associate, NIGMS, National Institutes of dicted positive charge in putative TM VI, histidine 297. This Health. 5 To whom correspondence shouldbe addressed: Molecular Neurobi- histidine is conserved in several membersof the opiate receptor ology, Box 5180, Baltimore,MD 21224. subfamily, but is only intermittently conserved in other mem'The abbreviationsused are: pOR, p opiatereceptor;PBS, phos- bers of the G protein-linked seven transmembrane domain rephate-buffered saline; GMP-PNP, guanosine 5'-(p,y-imido)triphosphate; WT, shortened wild type AN64 receptor; TM, transmembrane ceptor family (31). Determination of the roles that each of these residues play domain; DAMGO, [~-Ala~,MePhe~,Gly-ol~Ienkephalin.
20548
p Opiate Receptor q ,,,,,,..,Q ,,.,.... Y. ...............................
,/'
U.!l ,...,,,,.,,,NH2
20549
2ml
n
1'233
\
HOOC
AC33
FIG.1. Schematic representation of predicted topology of pOR1. Seven transmembrane domains are indicated (gray cylinders), containing the threeloci for site-directed point mutations (circles). The AN64 mutant (WTJ lacks the wild type N terminus (broken line) andits five potential N-linked glycosylation sites (forks); AC33 lacks an additional 33 amino acid residues from the wild type C terminus. The region corresponding to an epitope employed for polyclonal antibody production is indicated by the stippled bar. 0147, Asp147;H297, Hiszg7;0114, Asp"'.
D114A
D114N
1
D114E
D147A
0147N
H297A D147E
FIG.2. Screening of pORl constructions for ligand binding and G c o u p l e d intrinsic activity. Point mutants and deletions were initially assessed for ability to bind 1and 5 nM ["HIDAMGO (black bars) and 1 and 5 nM [3H]naloxone(white bars); activation of inhibition of adenylate cyclase was measured as a function of DAMGO (striped bars) and morphine (checkered bars). All values are as compared with the WT, construction.
unlabeled compounds in 0.5 ml of Tris buffer and incubated for 90 min a t 25 "C. Binding was terminated by filtration through GFB filters (Whatman) followed by 3 washes with 4 ml of Tris buffer at 4 "C using a Brandel filtration device. Radioactivity was assessed in a Beckman liquid scintillation counter a t 40% efficiency. Forskolin-stimulated cAMPAccumulation-35-mm COS cellcultures could provideinsights into themodes whereby opiate drugs and preincubated for 15 min with 1mM isobutylmethylxanthine were incuopioid peptides interact with p receptors. Defining the contri- bated for 10 min a t 37 "C with 10 p~ forskolin and saline, morphine, or . medium was aspirated, 0.1 M HCI was added, butions of N- and C-terminal domains in receptor function al- DAMGO (1-10 p ~ ) The cells were sonicated for 1 min, and extracts were neutralized with lows comparisons of these domains' properties to those of trans- NaOH. CAMPlevels were determined by a scintillation proximity ramembrane regions. Deleted receptors retaining near wild type dioimmunoassay as described (1). levels of function also potentially provide simpler model sysPORI-like Zmmunostaining-A rabbit polyclonal antiserum raised tems for mutagenesis. We now report the resultof initial char- against a glutaraldehyde conjugate of the C-terminal 18 amino acids of receptor and keyhole limpet hemocyanin recognized pOR. acterization of the structurelfunction relationships of the p opi- the rat pORl The peptide, termed pOR-A, was synthesized by Multiple Peptide Sysate receptor through deletions and site-directed mutagenesis. tems (San Diego,CA), purified to -go%, and conjugated to keyhole limpet hemocyanin using glutaraldehyde as described previously (33). EXPERIMENTALPROCEDURES Three rabbits were immunized with subdermal and intramuscular inN Terminus and N- and C-terminal Deletions-pcDNApOR1 cDNA jections of 100 pg of conjugate in complete Freund's adjuvant and encodes the entire1194-base pair open reading frameof the pOR in the boosted a t 2-4-week intervals with 50 pg of the same conjugate in expression vector pcDNAl, as described previously (1). AN64, a deletion incomplete adjuvant (Cocalico Biologicals,Reamstown, PA). Sera were of the 64 N-terminal amino acids of pOR1, was obtained from a partial collected and tested for immunoreactivity using enzyme-linked immucDNAsubcloned into the same vector, as described previously (1).AC33, nosorbent assay and immunoblotting, as described (34). a deletion of 33 C-terminal amino acids from the AN64 construction, For immunocytochemistry, COS cells were transfected with wild type was prepared via deletion of a I-kilobase XhoI fragment that results in or mutant pORlexpressed in pcDNAl, grown for3 days on 22-mm glass the pOR sequence ending at arginine 365, followed byalanine, cysteine, coverslips, fixed with 4% paraformaldehyde in phosphate-buffered saisoleucine, and a chain-terminating codon as a consequence of the line (PBS)at 4 "C for l hand washed three times with PBS a t 25 "C. For pcDNAl multiple cloning region sequence. some experiments, cells were incubated with 0.1% H,Oz for 30 min at Mutagenesis-A single-stranded template for mutagenesis was pre- 25 "C to suppress endogenous peroxidase. Fixed cellswere preincubated pared from a AN64 Bluescript I1 SK (Stratagene) subclone. Oligode- with PBS, 0.5% Triton X-100, 3% normal goat serum for 1 h a t 25 "C; oxynucleotides containing the mutantcodon were annealed to the tem- washed three times with PBS, 0.1% Triton X-100, 1%normal goat plate followed by in vitro synthesis and ligation of the mutant strand, serum for 5 min; and then incubated with primary antiserum 5051, or NciI-mediated nicking and exonuclease I11 digestion of the nonmutant other control sera diluted 1:5000 for 3 days at 4 "C. Coverslips were strand, and repolymerization and ligation of the gapped DNA as de- washed with PBS, and bound IgG was detected by biotinylated goat scribed previously (32) (Amersham kit, version 2.1). Mutations were anti-rabbit antisera and avidin-conjugated peroxidase as described confirmed by DNA sequencing. Fragments of 550 base pairs (BamHI- (Vectastain ABC kit). Reaction product was enhanced in some sections BamHI;Asp"4 and Asp147)or 250 base pairs (FspI-PuuI; Hiszg7)contain- by use of 0.8%nickel chloride in the diaminobenzidine solution. Control ing each mutation were isolated by agarose gel electrophoresis and experiments included staining of paraformaldehyde-fixed brain secsubcloned into AN64 previously digested with the appropriate restric- tions in the samefashion, use of preimmune sera, omission of secondary tion enzymes, and the subcloned fragments were resequenced to verify antibodies, preincubation in which the C-terminal peptide or an irrelthe absence of inadvertant mutations. evant peptide of the same length was added to primary sera a t 500 Expression ofMutant Receptors-COScells were transfected by elec- pg/ml overnight a t 4 "C before application to cells, and experiments in troporation with 20 pg/107cells of pcDNApOR1 as described previously which the C-terminally deleted receptor was expressed. (1).Transfected cells were plated in Dulbecco's modifiedminimal essenPharmacological Characterization-Results from screening experitial medium containing 10% fetal bovine serum and cultured for 3 days. ments that indicated detectable binding were followed with more deCells expressing each construction were screened for functional recep- tailed saturation and displacement experiments. Saturation analyses tor expression by radioligand binding, forskolin-stimulated CAMP accu- employed 0.1-15 n~ [3HlDAMG0 or [3Hlnaloxone. In some assays, 50 mulation, and immunostaining. mM GMP-PNP was added to incubations. Data analysesused the EBDA Radioligand Binding-Levorphanol-displaceable binding of 5 n~ and LIGAND programs (34). ['HIDAMGO ([~-Ala~,MePhe~,Gly-ol~lenkephalin) or nonradioactive naloxone-displaceable binding of 5 nM [3Hlnaloxone was assessed in RESULTS intact COS cells. Briefly, medium was removed from 150 mm plates Both the wild type p opiate receptor, pOR1, and the N-tercontaining 5 x lo6 COS cells; plates were rinsed with 50 mM Tris buffer, pH 7.4, and cells were harvested by scraping. Binding densities were minal deletion mutant AN64 (termed WT,; see below)exnormalized to total protein concentration, determined by the Bradford pressed robust binding ofDAMGO and naloxone as well as method (Bio-Rad). COS cells were combined with various labeled and morphine- and DAMGO-mediated inhibition of forskolin-
20550
JL
Opiate Receptor
60 -
40
~
0.1
106 -
1
Morphine (nM)
FII;.3. Morphine displacement of ['HIDAMGO hinding to WT, and AC.33 receptors. I('.,s o f 7.7 antl (5.6n\r and K, v a l n r s of 2.5 and 1.6 n\r wrrr o1,t:Iinrd for IVT" and AC9.7. rrsprrtivrly
80 49.5 -
stimulated cyclic AMP accumulation in COS cells (Fig. 2) as 32.5 previously reported ( 1j. The AN64 deletion, therefore, simplified the receptor. A further deletion of 3.7 C-terminal amino 27.5 acids was constructed to assess whether an even simpler p receptor could servr as the hasis for mutagenesis experiments. 18.5 This N- and C-terminally deleted mutant receptor fAC.73) recognized DAMGO, naloxone. and morphine, and mediated morphine-induced inhihition of forskolin-stimulated cyclic AMP accumulation in a fashion similar to that displayed hy the wild Preimmune type receptor (Figs. 2 and 3).The GTP analog GMP-PNP re(1 :2000) duced DAMGO binding affinity for both AN64 and LC33 by 3and 4-fold, respectively (Table I), declines similar to those disImmune played hy t h e wild typepOR1receptorandotherG-linked (1 :5000) receptors. The ability of DAMGO to inhibit adenylate cyclase activitywasdecreasedto 15'i of AN64 values in t h e AC33 MuOR Peptide mutant. Morphine-mediated inhihition remained intact (Fig. 0.1 pgirnl 2). a result reproduced in 10 experiments. The AN64 receptor is was thus used as the hasis for further mutagenesis and referred to as the shortened wild type (WT,) for convenience. Polyclonal anti-p receptor antihodies were employed to add to verification of expression of mutant receptors. Antisera directed against the peptide FOR-A displayed reactivity in enzyme-linked immunosorhent assay that produced titers rangingfrom15775to 1:40,000. Theserumwiththehighest enzyme-linked immunosorhent assay titer, 5051, detected immunoreactivity in a discrete band in Western immunohlots of protein purified from rat brain (Ref. 35: Fig. 4). Further evidence for specificityof immunostaining was adduced in experiments in which rat brain sections yielded positive staining in was 30'6 of WT, values in the alanine mutant and grratrr than regions rich in p opiate receptors.? Preadsorption with the ap- WT, values in the glutamic acid and asparaginc mutants. Thc propriate peptide eliminated COS cell and brain section stain- preservation of naloxone binding (Tahlc 111prrmittrd ~ S S P S Y ing. No staining was observed when the primary antihody was ment of agonist potencies in displacing I 'il Inaloxnnr 1 Figs. 6 eliminated or when COS cells were transfected with either the and 7 ) . DAMGO andmorphineaffinities for t h r aspar:tCinr pcDNAl vector aloneor t h e AC.73 construction, which lacks the mutantreceptorwere5277-and 74-fold lrss th:ln thosr fhr C-terminalepitopeagainstwhichtheantibodywasraised. WTh, respectively. Reductions in DAlIGO and morphinc affniEach deletionor mutant except the AC.78 construction revealed ties of 133.7- and 3.7-fold were sustninrd, rrsprctivrly. whm positive immunostaining in transiently expressing COS cells. glutamic acid was suhstitutrd (Figs. 6 and 7 1. No differences in the cellular distribution of immunoreactivity The ability to inhibit stimulated adenylate cyclasr \vas also could he consistently ohserved with any of the mutants exam- reduced h.v asparagine substitution for Asp"' (Fig. 21. (:yclasr ined; each appeared to equally distribute immunoreactive ma- inhihition at less than2 S f ; of wild typr values \vas o h s r n r d by terial to different cellular regions as identifiedhv light micro- agonist concentrations that wrrt morc than 1 0 timcs thr h': scopic observation (Fig. 5).Some nuclear staining was notedin values inferred from analysis ofdisplaccmc~nt of nnloxonr hindeach case (Fig. 5 ) . ing, suggesting influences on intrinsic activity a.5 tvrll as nn Replacement of t h e TM I1 Asp".' by either alanine, aspara- affinity. Glutamic acid substitution mutants, hourvrr. rrtninrd gine, or glutamic acid reduced bindingof the p-preferring ago- full intrinsic activitv for DAMGO and morphinc.. nist [?HIDAMGO to less than 5'5 of WT, values in screening Substitution of the TM 111 rrsidur Asp!" \vith nlaninr or experiments (Fig. 2 ) . Binding of 5 nu ["Hlnaloxone, however, asparagine dramatically rrtlucrd binding of [ 'H I D t \ l l ~ ; Oantl I"Hlnaloxone in scrreningexprrimrnts'FiC. 21. Rindint: of .. . .- . . ["Hlnaloxone \vas 30"; of Ii'T. values for thcs t:lut:~micacid muA. Moriwaki and G. R. LJhI, manuscript in prrparation.
-
+
-
~
"
+
+
-
+
+
,u Opiate
a
O0
0 ]*D114E 0.1 100010300 100 10 1
b
DAMGO inMi
3
FIG.6. DAMGO displacement of I'HInaloxonc. hinding to mutant p receptors. IC,,,, valurs o f 17 nv. H . 4 . and 2.0 p v a n d h' w l u r s of 7.2 n\c. .?X. a n d 9.6 p v w'rrrnhtaincdfnr respectivrly.
LVT., 1)114S. a n d l ) l l 4 E ,
n
0 0 V"Ts
40.
0
0 D114N
O V 0.
201 0 D114E
n
1
V
o v
D147E ,
1 0.1
0
10
100
n 0
V-O.".. 1033 10000
Morphlne fnMl
FIG.7. Morphine displacement of ['Hlnaloxone hinding to mutant p receptors. ICFfivnlurs o f 1.1. 660. 3211. :!nd 41; n v a n d K v a l r ~ v s o f 3.9,290. 130. a n d 1 9 n v w r r r o h t a i n r d for \YT.,I)114S. 1)lIJE. and
D147E,rrsprctivrly.
C
FIG.5. Immunostaining of transfected COS cells. Three selected constructions arr shown.n, pcDNAl vrctor; h, W T a : c, D114A.I n no case did prrimmunr sera exhihit immunostaining (data not shown).
tant, allowing its5-fold affinity loss for morphinr to brasscsscd (Fig. 7 ) . Cyclase inhihition was also significantly rrtlucrtl by either alanine or asparagine substitution for Asp"' (Fig. 21. Glutamic acid suhstitution retained nearly full IIA.MGO and morphine intrinsic activities, as cyclase inhihition nchirvrd by high agonist concentrations was similar to thatfound for \\.'T.. Cotransfection of t h e DllllA and D147A mutants yicldrd thr same poor [''HIDAMGO and ['Hlnaloxonc~ binding found with each mutant receptor alone (data not shown). Thr mutant rrceptors thus reveal no evidence for complcmrntatinn through heterodimeric or other mechanisms. A third charged residue, the positively charged histidinr 297, is predicted to lie within TM V I fig. 1). Rrplacrmcnt of this amino acid with alanine resulted in substantial rrtluctions in ['HIDAMGO and["Hlnaloxonebinding in scrrrning rxpcriments (Fig. 2). The reduction in binding rrndcrrd attrmpts to accurately estimate the precise afiinity loss unsuccessful. This mutant was still fully functional in inhibiting ndrnylntc cyclase; inhibition achieved a t high concentrations of DAXIGO and morphine was even greater than that achirvrdby the \!T. receptor.
20552
p Opiate Receptor DISCUSSION
This initial exploration of structurdfunction relationships of the p opiate receptor reveals strikingly differentfunctional consequences of changes in specific receptor regions. The present findings add to an emerging picture of p receptor structure/ function relationships derived from mutagenesis, gene family relationships, and analogy with results obtained in mutagenesis studies of other related G-linked receptors. Molecular modeling studies, in which structures are inferred by superimposing seven a-helical receptor TM domains onto the TM domain backbone of crystallized rhodopsin (361, have also provided possible scenarios for receptor interactions with ligands, ions, and G proteins. These results need to be interpreted in light of technical considerations arising in many mutagenesis studies. A significant concern in interpretation of results of many deletion and mutagenesis experiments relates to expression. None of the evidence developed in this report absolutely excludes the possibility that mutations and/or deletions produce modest effects on levels of cellular or membrane expression. Many mutants, however, display wild type function in at least some assays. Each, except AC33, produces FOR-like immunostaining indistinguishable from wild type patterns in expressing COS cells. The nuclear staining noted here could reflect differential compartmentalization of transiently expressed p receptor protein in COS cell nuclei,although less nuclear staining is observed in neurons.2 These immunohistochemical studies are moredetailed than those often accompanying reports of mutagenesis work, although they are limited to the light microscopic levelof resolution. Each of these features suggests that no major variation in the extent of cellular expression was exerted by the molecular alterations effected in these studies. Retention of full binding activity in receptors lacking virtually all of the N-terminal first extracellular domain was initially surprising (l),but the result was replicated in studies described here. This deletion removes each of the five consensus sites for N-linked glycosylation, as well as many of the amino acid residues unique to the p opiate receptor. The intactness of p agonist and antagonist binding in receptors displaying this deletion suggests that this region's function relates t o features other than agonist or antagonist recognition. Conceivably, specific sugars displayed on this receptor's N terminus could aid in cellular distribution of receptor protein; B,, values for the W T , receptor were somewhat lower than those of the native wild type receptor in preliminary experiments3 The importance of the C-terminal cytoplasmic domainof the p opiate receptor for ligand binding and G protein-couplingwas tested by deletion of the C-terminal33residues of the W T , construction. No pronounced ligand binding effects of the truncation were detected; DAMGO, morphine, and naloxone binding profiles were essentially indistinguishable from those of the wild type receptor (Figs. 2 and 3). Morphine remained an effective inhibitor of forskolin-stimulated cyclic AMP accumulation with these deletions, while DAMGO lost much of this capacity in this deletion construction (Fig. 2). Regulation of the high affinity state for agonist binding by GTP and analogs (11) was retained in theAC33 mutant (Table I). This outcome is similar to that noted after deletion of the 49 C-terminal residues of the human p-adrenergic receptor (37). The AC33 p receptor, lacking approximately one-third of the wild type protein sequence, thus retained activity in most, but not all, respects. Three charged amino acid residues predicted to lie within the transmembrane region of the p opiate receptor protein have been addressed as points of contact for incomingagonists and a C.
K. Surratt, unpublished observations.
antagonists. These residues have been postulated to contribute toward formation of a hydrophilic central pocket comprisinga ligand binding site (27). One of the interesting features of the opiate receptors relates to the fashion in which they bind small molecule ligands and larger neuropeptides. Many p receptor small molecule and neuropeptide ligands possess positively charged nitrogen atoms at physiological pH values. Molecular modeling has suggested that these positive charges might interact with negatively charged transmembrane aspartic acid residues conserved in several G-coupled receptors (27,31, 38-40). TM I1 contains an aspartic acid residue that is conserved among most G-coupled receptors. The dramatic reductions in binding activity observed for p receptors altered at this TM I1 residue are consistent with the important roles that homologous aspartic acids may play in adrenergic and neuropeptide receptors (26, 27, 29, 41, 42). These data also fit with the role that aspartic acid 95 of the 6 opiate receptor may play in agonist binding, as inferred from studies of an asparagine mutant at this position (28). Important roles ascribed to the TM I1 aspartic acid include direct effects in ligand recognition and/or interaction with sodium ions important for modifying agonist binding and participation in transduction of agonist binding signal to G protein activation or vice versa (25, 26, 28-30, 43). Direct effects on agonist binding could account for the substantial losses in agonist potency incurred by the three substitution mutations of Asp1I4, although the sparing of naloxone affinity with the asparagine and glutamic acid substitutions may point to differential involvement of Asp114in binding domains for agonist and antagonist ligands (Figs. 2,6, and 7).The effect of replacement of the TM I1 aspartic acid on ligand binding has been variable in studiesof other receptors. Asparagine substitution at the corresponding position did not diminish agonist binding for the somatostatin, angiotensin 11, muscarinic, and qadrenergic receptors (25, 26, 30, 43), while 101000-fold decreases were observed for P-adrenergic, 6 opiate, and lutenizing hormonelchoriogonadotropin receptors (27-29). Interestingly, antagonist binding was undiminished or even enhanced by mutations in several of these G-coupled receptors and in the current workon the p receptor. These data are consistent with postulates that agonist and antagonist recognition domains are not identical. Further, retention of antagonist affinity suggests that thedramatic loss of agonist recognition noted with the Asp"* mutations may not be due to severe generalized conformational changes in the receptorprecipitated by substitutions at this position. The negative charge in TMI1 has also been suggested to contribute to Na' binding. Allosteric regulation of ligand binding by Na+ has been reported to be mediated by homologous residues in the qadrenergic (441, 6 opiate (281, somatostatin type 2 (30), and lutenizing hormone/choriogonadotropin (29, 41) receptors. Conceivably, Na' recognitionat such a site could allosterically contribute to the reductions in agonist aflinity characteristic of Na+ effects first noted in binding studies of opiate receptors (45). DAMGO displacement of 13Hlnaloxone was not altered appreciably byNa' for the mutant receptor, suggesting that the D114N mutant is Na+-insensitive,in contrast t o WT, and D114E.4 The affinity of the D114E and D114N mutants for morphine was reduced to 130 and 290 m, respectively. Adenylate cyclase inhibition mediated by 10,000 n~ morphine in these mutant receptors differed more strikingly. Glutamic acid substitution retained wild type levels of cyclase inhibition, while asparagine substitution yielded less than 15%of wild type values (Fig. 2). These data strongly imply the necessity for a negative charge a t C. K. Surratt and G. R. Uhl, manuscript in preparation.
Receptor
p Opiate
position 114 for efficient G protein coupling with themorphinestimulated receptor. and glutamic acid substitutions at the homologous TM 11 aspartic acid residue in the angiotensin I1 (subtype 1) receptorrevealed similar effects and also yielded increased antagonist binding (Ref. 42, see also "
20553 2. Porreca, F., and Burks, T. F. (1993)Handb. EXP.Pharrnacol. 104,21-51 3. DiChiara, G., and Imperato,A. (1988)Proc. Natl.Acad. Sei. U. S. A. 85,5274C"", a& (0
4. Spanagel, R., Herz, A,, andShippenberg, T. S. (1990) J. Neurochem. 55,17341740 5. Johnson, P. s.,Wang, J. B., Wang, W. F., and Uhl, G. R. (1994)Neuroreport 5, 507-509 6. Carter, B. D., and Medzihradsky, F. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, Ref. 28). TM I11 Asp147substitutions produced changes in p receptor 40624066 7. Miyake, M., Christie, M. J., and North, R.A. (1989) Proc. Natl. Acad. Sci. binding and second messenger inhibition that closely reflect U. S. A . 86, 3419-3422 results obtained in studiesof the p-adrenergic receptor (Ref. 27; 8. Schroeder, J . E., Fischbach, P. S., Zheng, D., and McCleskey, E. W. (1991) Neuron 6, 13-20 Fig' 2)' These marked reduction in agonist 9. Kennedy, C., and Henderson, G. (1991) Mol. Pharmacol. 40, 1000-1005 and antagonist affinities and agonist intrinsicactivities VJhf!n 10. Seward, E., Hammond, C., and Henderson, G. (1991) Proc. R. Soc. Lond. B substitutions remove the negativecharge. The reduced funcB i d . Sci. 244, 129-135 11. Blume, A. J. (1978) Proc. Natl. Acad. Sci. U. S. A . 75, 1713-1717 tional activitiesof the alanine and asparagine mutants at 12. Koski, G . , and Klee, W. A. (1981) Proc. Natl. Acad.Sei. U. S. A . 78,41854189 Sition 147 could fit with roles of the wild tYPe aspartic acid in 13. Frances, B., Moisand, c., and Meunier, J.-c. (1985) EUCJ . Pharrnacol. 117, direct ligand recognition, as has been implicated instudies of 223-232 agonist binding to the p-adrenergic (27) and other receptors. A 14. Crain, S. M., Crain, B., and Makman, M. H. (1987) Brain Res. 400, 185-190 15. Demoliou-Mason, C. D., and Barnard, E. A. (1986) J. Neurochem. 46, 1118glutamic acid residue at the analogous position in rhodopsin 1128 ~~~. provides the counterion for the positively charged Schiff base of 16. Wang, J. B., Johnson, P. S., Persico, A. M., Hawkins, A. L., Griffin, C. A,, and Uhl, G. R. (1994) FEBS Lett. 338, 217-222 its ligand, 114s-retinal,while a TM I11lysine of the endothelin 17. Evans, C. J., Keith, D. E., Morrison, H., Magendzo, K., and Edwards, R. H. B receptor possesses a corresponding positively charged lysine (1992) Science 258, 1952-1955 residue whose substitution by aspartic acid virtually elimi- 18. Kieffer, B. L., Befort, K., Gaveriaux-Ruff, C., andHirth, C. G. (1992)Proc.Natl. Acad. Sei. U. S . A. 89, 12048-12052 nates bindingof negatively chargedETb-selective agonists (4619. Chen, Y., Mestak, A,, Liu, J., Hurley,J. A., and Yu, L. (1993) Mol. Pharmacol. 49). Although Asp147of the p receptor is more likely than Asp114 44, a 1 2 20. Fukuda, K., Kato,S., Mori, K., Nishi, M., and Takeshima, H. (1993)FEBS Lett. to form an ion pair with the positively charged agonist, both 327,311-314 transmembrane aspartic acid residues were required for effl- 21. Yasuda, K., Raynor, K., Kong, H., Breder, C., Takeda, J., Reisine, T.,and Bell, G. I. (1993)Proc. Natl. Acad. Sci. U. S. A. 90, 6736-6740 cientradiolabeled agonist binding. Moreover, the charged groups alone are insufficient for high agonist affinity;the extra 22. ODowd, B. F., Hnatowich, M., Regan, J. W., Leader, W. M., Caron, M. G., and Lefkowitz, R. J. (1988) J. Biol. Chem. 263, 15985-15992 methylene groupof the glutamicacid side chain disturbedbind- 23. Samama, P., Cotecchia, S., Costa, T., and Lefiowitz,R. J. (1993)J . Biol. Chem. 268,46254636 ing significantly in each case. Fraser, C. M. (1989) J. B i d . Chem. 264, 92669270 The results of TM I11 Asp147substitutions could also reflect 24. 25. Wang, C.-D., Buck, M. A,, and Fraser, C. M. (1991) Mol. Pharmacol. 40, 168alterations inreceptor structure through displacement of pre179 existing interactions between this amino acid's sidechain and 26. Fraser, C. M., Wang, C.-D., Robinson, D. A., Gocayne, J. D., and Venter, J. C. (1989) Mol. Pharrnacol. 36,840-847 other receptor constituents. Rhodopsin-based molecular mod- 27. Strader, C. D., Sigal, I. S., Candelore, M. R., Rands, E., Hill, W. S., and Dixon, eling of the p opiate receptor indicates that thereceptor's TM R. A. F. (1988) J . Biol. Chem. 263, 10267-10271 28. Kong, H., Raynor, K., Yasuda, K., Moe, S. T., Portogbese, P. S., Bell, G. I., and I11 helix could lie close t o that of TM VI. In this orientation, the T. (1993) J. Bid. Chem. 268, 23055-23058 TM I11aspartic acid could participate inionic interactions with 29. Ji, Reisine, I., and Ji, T. H. (1991) J. Biol. Chem. 266, 14953-14957 histidine 297 of TM VI (data not shown). The TM VI histidine 30. Kong, H., Raynor, K., Yasuda, K., Bell, G. I., andReisine, T. (1993) Mol. Pharrnacol. 44, 380-384 residue is conserved among thep, S, and K opiate receptors (1, 31. Trumpp-Kallmeyer, S., Hoflack, J., Bruinvels, A,, and Hibert, M. (1992) J. 17-21) and 13 of the 42 other G-coupled receptors recently Med. Chem. 35, 3448-3462 reviewed by "rumpp-Kallmeyer et al. (31). 32. Kunkel, T. A,, Roberts, J. D., and Zakour, R. A. (1987) Methods Enzymol. 154, 367-382 Alanine substitution for Hiszg7 diddramatically reduce ago33. Vaughan, R. A,, Uhl, G. R., and Kuhar, M. J. (1993) Mol. Cell. Neurosci. 4, nist affinity while maintaining normal or increased intrinsic 209-215 activity of the mutant receptor. While our modeling fails to 34. Munson, P. J., and Rodbard, D. (1980)AnaL Biochem. 107,220-239 provide plausible scenarios for direct histidine involvement in 35. Eppler, C. M., Hulmes, J . D., Wang, J. B., Johnson, B., Corbett,M., Luthin, D., Uhl, G. R., and Linden, J. (1993) J. Biol. Chem. 268,26447-26451 morphine or DAMGO recognition, involvement of Hiszg7in 36. Schertler, G. F. X., Villa, C., and Henderson, R. (1993)Nature 362, 770-772 models of lofentanil binding to thep receptor has been postu- 37. ODowd, B. F., Hnatowich, M., Caron, M. G., Lefiowitz, R. J., and Bouvier, M. (1989) J . B i d . Chem. 264, 7564-7569 lated by Moreels and co-workers (cited in Ref. 50). Transmem- 38. Hibert, M. F., Trumpp-Kallmeyer, S., Bruinvels,A., andHoflack, J . (1991) Mol. brane histidine residues in two G-coupled receptors have rePharrnacol. 40, 8-15 cently been directly implicated in antagonist recognition (51, 39. Wess, J. (1993) Pends Pharrnacol. Sci. 14, 308-313 M. F., Trumpp-Kallmeyer, S., Hoflack, J., and Bruinvels, A. (1993) 52). Hiszg7could also potentially contribute to the pH sensitiv- 40. Hibert, Dends Pharmacol. Sci. 14, 7-12 ity noted for the receptor's ability to recognize ligands (53), as 41. Quintana, J., Wang, H., and Ascoli, M. (1993) Mol. Endocrinol. 7, 767-775 the pK, of its imidazole nitrogen is closer to the physiological 42. Savarese, T. M., and Fraser, C. M. (1992) Biochem. J. 283, 1-19 43. Bihoreau, C., Monnot, C., Davies, E., Teutsch, B., Bernstein, K. E., Conrol, P., range than thatof any other p receptor amino acids. and Clauser, E. (1993) Proc. Natl. Acad. Sci. U. S. A . 90, 5133-5137 Studies of structure/function relationships in the p opiate 44. Horstman, D. A., Brandon, S., Wilson, A. L., Guyer, C. A,, Cragoe, E.J.,Jr., and Limbird, L. E. (1990)J . Biol. Chem. 265,21590-21595 receptor, combined with structurelactivity correlates of opiate 45. Pert, C. B., and Snyder, S. H. (1974) Mol. Pharrnacol. 10, 868-879 agonists and antagonists,should provide increasingly detailed 46. Zhukovsky, E. A., and Oprian, D. D. (1989)Science 246, 928-930 pictures of the interactions between therapeutically important 47. Sakmar, T. P., Franke, R. R., and Khorana, H. G. (1989) Proc. Natl. Acad.Sci. U. S. A. 86,8309-8313 and abused drugs and p receptors that provide both medical 48. Nathans, J. (1990) Biochemistry 29, 97469752 benefit and addiction. 49. Zhu, G., Wu, L.-H., Mauzy, C., Egloff, A. M., Mirzadegan, T., and Chung, E-Z. (1992) J . Cell. Biochem. 50, 159-164 Acknowledgment-We thank Dr. Stacie Yuhasz for help with com- 50. Uhl, G. R., Childers, S., and Pasternak, G. (1994) Pends Neurosci. 17,89-93 51. Fong, T.M., Cascieri, M. A., Yu, H., Bansal, A,, Swain, C., and Strader, C. D. puter analyses and molecular modeling. (1993) Nature 362,350-353 52. Woodward, R., Daniell, S. J.,Strange, P. G., and Naylor, L. H. (1994) J. REFERENCES Neurochem. 62, 1664-1669 1. Wang, J. B., Imai, Y., Eppler, C. M., Gregor, P., Spivak, C. E., and Uhl, G. R. 53. Pert, C. B., and Synder, S. H. (1973)Proc. Natl. Acad. Sci. U. S. A. 70, 2243(1993) Proc. Natl. Acad. Sei. U. S. A . 90, 10230-10234 2247
-
"
