Opioid Peptide Receptor Studies, 11:Involvement of Tyr148, Trp318 and His319of the Rat m-Opioid Receptor in Binding

of m-Selective LigandsHENG XU,1 YI-FENG LU,1 JOHN S. PARTILLA,1 QIAO-XI ZHENG,1 JIA-BEI WANG,2

GEORGE A. BRINE,3 F. IVY. CARROLL,3 KENNER C. RICE,4 KAI-XIAN CHEN,5ZHI-QIANG CHI,5 AND RICHARD B. ROTHMAN1*

1Clinical Psychopharmacology Section, Division of Intramural Research, NIDA, NIH, PO, Baltimore, Maryland 212242Department of Pharmaceutical Sciences, School of Pharmacy, UMAB, Baltimore, Maryland 21201

3Chemistry and Life Sciences, Research Triangle Institute, Research Triangle Park, North Carolina 277094Laboratory of Medicinal Chemistry, NIDDK, NIH, Bethesda, Maryland 20892

5Shanghai Institute of Materia Medica, Chinese Academy of Sciences, Shanghai 200031, China

KEY WORDS opoid receptors; ligand binding; fentanyl; molecular modeling; muta-genesis

ABSTRACT Previous data obtained with the cloned rat µ opioid receptor demon-strated that the ‘‘super-potent’’ opiates, ohmefentanyl (RTI-4614–4) and its four enantio-mers, differ in binding affinity, potency, efficacy, and intrinsic efficacy. Molecularmodeling (Tang et al., 1996) of fentanyl derivatives binding to the µ receptor suggeststhat Asp147, Tyr148, Trp318, and His319 are important residues for binding. Accordingto this model, Asp147 interacts with the positively charged opiate agonist to form potentelectrostatic and hydrogen-bonding interactions. In this study, the role of weak electro-static and hydrogen-bonding ‘‘p-p’’ interactions of the O atom of the carbonyl group andthe phenyl ring structures of RTI-4614–4 and its four enantiomers with residues Tyr148,Trp318, and His319 were explored via site-directed mutagenesis. Tyr148 (in trans-membrane helix 3 5TMH36), Trp318 (TMH7), and His319 (TMH7) were individuallyreplaced with phenylalanine or alanine. Receptors transiently expressed in COS-7 cellswere labeled with [125I]IOXY according to published procedures. Mutation of Tyr148 tophenylalanine reduced the binding affinities of some µ-selective agonists (2–7 fold) butdid not alter the affinities of DAMGO, naloxone, and the non-selective opiates etorphineand buprenorphine. In contrast, this mutation significantly increased the bindingaffinities (decreased the Kd values) of [D-Ala2,D-Leu5]enkephalin, IOXY, and dermor-phin. Mutation of Trp318 decreased opioid receptor binding to almost undetectablelevels. Substitution of alanine for His319 significantly reduced binding affinities for theopioid ligands tested (1.3- to 48-fold), but did not alter the affinities of naloxone andbremazocine. These results indicate the importance of Tyr148 and His319 for the bindingof fentanyl derivatives to the µ receptor. Functional studies using the mutant receptorswill provide additional insight into the mechanism of action of RTI-4614–4 and its fourenantiomers. Synapse 32:23–28, 1999. Published 1999 Wiley-Liss, Inc.†

INTRODUCTION

Fentanyl and its congeners (Van Bever et al., 1974)are of interest not only because of their clinical use asanalgesics, but also because certain members of thisseries exhibit unique properties. For example, (1)-3-methylfentanyl and congeners such as ohmefentanyl(RTI-4614–4, (6)-cis-N-[1-(2-hydroxy-2-phenylethyl)-3-methyl-4-piperidyl]-N-phenylpropanamide HCl), lofen-tanil, and sufentanil are much more potent in tests of

antinociception than their Kd values would predict(Casy and Parfitt, 1986; Van Bever et al.,1974) despitethe general finding that the ED50 of an agent forproducing antinociception is generally correlated with

*Correspondence to: Richard B. Rothman M.D., Ph.D., CPS, DIR, NIDA, POBox 5180, 4940 Eastern Ave., Baltimore, MD 21224. E-mail: RROTHMAN@IRP.NIDA.NIH.GOV

Received 21 July 1998; Accepted 19 August 1998

SYNAPSE 32:23–28 (1999)

Published 1999 WILEY-LISS, INC. †This article is a US governmentwork and, as such, is in the public domain in the United States of America.

its Kd for the µ binding site (Leysen et al.,1977; Wilsonet al.,1975). The fact that there is no correlationbetween the lipid solubility of compounds and theirantinociceptive activity suggests that pharmacokineticfactors can not fully explain the extraordinary potencyof certain (1)-cis-3-methylfentanyl analogs. Other fac-tors, such as an ability to achieve full efficacy at lowfractional occupation of the µ receptor, may explain thehigh potency of (1)-cis-3-methylfentanyl analogs (Rosen-baum et al.,1984).

The availability of four stereoisomers of RTI-4614–4allowed us to determine if enantiomeric ligands candistinguish among the four parameters of the ligand-receptor interaction: potency (ED50), efficacy (maximalstimulation), intrinsic efficacy (effect as a function ofreceptor occupation), and binding affinity. The use ofstereoisomers focuses the analysis on asymmetric struc-tural factors while avoiding confounding changes inphysiochemical characteristics. Our data, obtained withcloned rat µ receptors, demonstrated that the fourisomers of RTI-4614–4 [(2S,3R,4S)-1a, (2R,3R,4S)-1b,(2R,3S,4R)-1c, (2S,3S,4R)-1d] differed in binding affin-ity, potency, efficacy, and intrinsic efficacy as measuredby stimulation of [35S]-GTP-g-S binding (Xu et al.,1997).Isomer (2R,3S,4R)-1c had the lowest binding affinity,highest intrinsic efficacy, and was a partial agonist. Incontrast, (2S,3R,4S)-1a had the highest binding affinity,moderate intrinsic efficacy, and was a full agonist.(2S,3S,4R)-1d, which had higher affinity for the µreceptor than (2R,3S,4R)-1c, was an antagonist in thissystem. With the stereochemistry fixed to (3S,4R), the2-OH group has a substantial effect on the ability of thedrug to activate the µ receptor: in the 2S configuration,(2S,3S,4R)-1d is an antagonist, and in the 2R configura-tion, (2R,3S,4R)-1c has very high intrinsic efficacy. Incontrast, the (3R,4S) configuration promotes high-affinity binding, whereas the 2S configuration of the2-OH provides additional enhancement of binding affin-ity and potency. We speculate that these differencesresult from ligands binding to different domains of the µopioid receptor. The overall goal of our work in this areais to determine the molecular basis of the four param-eters associated with dose response curves: bindingaffinity (KI value), potency for receptor activation (ED50),efficacy (maximal response), and intrinsic efficacy (rela-tionship between receptor occupation and response).

Previous data indicated that the region from the Nterminal to the start of the TMH3 of the µ opioidreceptor determines the binding affinity of RTI-4614–4and its isomers and that the third extracellular (e3)loop also plays a major role in determining the bindingaffinity of µ agonist peptides (Lu et al.,1996,1998;Minami et al.,1995; Xue et al.,1995). Tang et al. (1996),based on computer modeling of the µ receptor usingbacteriorhodopsin as a template (Trumpp-Kallmeyer etal.,1992), identified residues Asp147, Tyr148, Trp318,

and His319 as being particularly important for thebinding of ohmefentanyl. Since Surratt et al. (1994)already showed the importance of ASP147 for agonistbinding, the present study determined the bindingaffinities of RTI-4614–4 and its isomers and otherselected agonists and antagonists for the cloned µreceptors mutated at residues Trp318, Tyr148, andHis319.

EXPERIMENTAL PROCEDURESSite-Directed Mutagenesis

A 1.5 kb HindIII fragment of the rat mu opioidreceptor gene (RMOR) containing the entire codingregion was subcloned into the phagemid pTZ18U (Wanget al.,1993). All mutations were prepared as describedby the manufacturer (Bio-Rad, Richmond, CA; Version2, Muta-Gene phagemid kit), and inserted into theexpression vector pcDNA3. The subcloned fragmentwas resequenced to verify the absence of inadvertentmutations.

Expression of Mutant Receptors

COS cells were transfected by electroporation with 20µg/107 cells of plasmid RMOR or mutant plasmid DNAand then grown for 2–3 days as described (Rothman etal.,1995; Wang et al.,1993). Cells were harvested 48–60hours following transfection. Cell pellets were homog-enized in 50 mM Tris-HCl (pH 7.4) using a polytron andcentrifuged at 37,000g for 10 minutes at 4°C. Thepellets were resuspended in 50 mM Tris-HCl (pH 7.4)and used in radioligand binding studies.

Binding Assays

[125I]IOXY (6b-iodo-3,14-dihydroxy-17-cyclopropyl-methyl-4,5a-epoxymorphinan) binding assays were con-ducted as described previously (Ni et al.,1993; Xu etal.,1997). Assays took place in 50 mM Tris-HCl, pH 7.4,along with a protease inhibitor cocktail (bacitracin [100µg/ml], bestatin [10 µg/ml], leupeptin [4 µg/ml], andchymostatin [2 µg/ml]), in a final assay volume of 0.5ml. Duplicate samples were filtered with Brandel CellHarvesters over Whatman GF/B filters after a 2–3 hourincubation at 25°C and washed twice with 5 ml ice-cold10 mM Tris-HCl, pH 7.4. The nonspecific binding wasdetermined with 10 µM naloxone (or 20 µM levallor-phan). All drug dilutions were made with 10 mMTris-HCl, pH 7.4, containing 1 mg/ml bovine serumalbumin. The filters were punched into 12 x 75 mmglass test tubes and counted in a Micromedic gammacounter at 80% efficiency. Membrane protein concentra-tions were determined using the Lowry method (Lowryet al.,1951).

24 H. XU ET AL.

Data Analysis and Statistics

The Kd and Bmax of receptors labeled by [125I]IOXYwere determined by displacing two concentrations of[125I]IOXY (0.014 nM and 0.114 nM) each by 8 concentra-tions of IOXY (0.03815 to 2,500 nM) as described indetail elsewhere (McLean et al.,1987). The data ob-tained from 3 independent experiments were pooledand fit to the one-site binding model using the nonlin-ear least squares curve fitting program MLAB-PC(Knott and Reece, 1972) (Civilized Software, Bethesda,MD) for the best-fit estimates of the Kd and Bmax.Inhibition curves were fit to the two parameter logisticequation for the best-fit estimates of the IC50 and slopefactor (N) using MLAB-PC. KI values were calculatedfrom the IC50 values according to the formula KI 5IC50/11[L]/Kd where L and Kd are the radioligandconcentration and its dissociation constant, respec-tively. Statistical significance between binding param-eters was assessed using the F-test as described else-

where (Munson and Rodbard, 1980; Rothman etal.,1991).

Chemicals

[D-Ala2-MePhe4,Gly-ol5]enkephalin (DAMGO) and[D-Ala2,D-Leu5]enkephalin were provided by RobertWalsh of the Research Technology Branch, NIDA. IOXY-AGO (6b-iodo-3,14-dihydroxy-17-methyl-4,5a-epoxy-morphinan) and IOXY were synthesized as described(Ni et al.,1993). [125I]IOXY (SA 5 2,200 Ci/mmol) wasprepared as described (Ni et al.,1993). RTI-4614–4 andits four enantiomers were prepared as described (Brineet al.,1995). Protease inhibitors were purchased fromPeptides International (Louisville, KY). Other opioidpeptides were purchased from Peninsula Laboratories(Belmont, CA). Various opiate compounds were ob-tained from Dr. Rice’s Laboratory (LMC, NIDDK,Bethesda, MD). The rat µ opioid receptor cDNA in

Fig. 1. Chemical structures of the four stereoisomers of RTI-4614–4.

25OPIOID PEPTIDE RECEPTOR STUDIES

expression vector pcDNA1/Neo was provided by Dr.J.-B. Wang (UMAB, Baltimore, MD).

RESULTS AND DISCUSSION

Figure 1 shows the structures of the four stereoiso-mers of RTI-4614–4. Figure 2 shows the detailed viewof the interaction of the protonated form of ohmefen-tanyl with the proposed binding sites of the µ opioid

receptor. In this model, His319 was not only included inTM regions, but was also regarded as a critical residue.

As reported in Table I, replacement of Tyr148 withphenylalanine failed to decrease [125I]IOXY (0.02 nM)binding. Replacement of H319 with alanine decreased[125I]IOXY (0.02 nM) binding by over 50%. Mutationof Trp318 eliminated the ability of the receptor tobind [125I]IOXY, [3H]][D-Ala2,D-Leu5]enkephalin, and

Fig. 2. A detailed view of the protonatedform of ohmefentanyl with the proposed bind-ing sites of the µ opioid receptor. Reproducedfrom Tang et al. (1996) with permission of thepublisher.

TABLE I. Comparison of [125I]OXY Binding to Rat m (RMOR)and Mutant Opioid m-Receptors Expressed in COS Cells†

Specific binding(50 mM Tris-HCl,

pH 7.4)

Specific binding(20 mM HEPES/100 mM NaCl/10 mM MgCl2/

1 mM EDTA/pH 7.4)

CPM% of

RMOR CPM% of

RMOR

RMOR 1296 6 12 100 2475 6 18 100Y148F (Replacement of tyrosine with phenylalanine) 1256 6 30 96.9 2452 6 30 99.1W318A (Replacement of tryptophan with alanine) 45 6 3* 3.5 118 6 45* 4.8H319A (Replacement of histidine with alanine) 510 6 45* 39.4 656 6 20* 26.5

†[125I]OXY (0.02 nM) binding assays to the RMOR and the Y148F, W318A and H319A mutants proceeded as described inMethods. Assays took place in 50 mM Tris-HCl, pH 7.4 (or in HEPES buffer, pH 7.4), along with a protease inhibitorcocktail, in a final assay volume of 0.5 ml. The values are the mean 6 SD of three experiments (each in triplicate).*P , 0.001 compared with RMOR control values (ANOVA).

26 H. XU ET AL.

[3H]bremazocine (some data not shown). These resultsindicated the importance of Trp318 and His319 for thebinding of opioids.

As shown in Table II, mutations Y148F and H319Adecreased the Bmax of the receptor. The Bmax valuesranged from 564 fmol/mg of protein for RMOR to 384

fmol/mg of protein for Y148F and 234 fmol/mg ofprotein for H319A. The Y148F mutation significantlydecreased the Kd of [125I]IOXY binding, while theH319A mutation increased the Kd of [125I]IOXY bind-ing.

The ligand-selectivity study of the mutant receptorswas conducted with opioid drugs representative of moststructural classes of opioid ligands and included bothagonist and antagonist ligands. After initial concentra-tion-ranging experiments, a parallel experimental de-sign, in which assays using the RMOR and two mutantreceptors were executed at the same time using thesame dilution curves, was used to determine the Kivalues of test agents. As reported in Table III, theY148F mutation decreased the binding affinities ofµ-selective agonists (2–7 fold) but did not alter theaffinities of DAMGO, naloxone, etorphine, and buprenor-phine. Interestingly, the Y148F mutation significantlyincreased the binding affinities of DADL, IOXY, anddermorphin. The H319A mutation significantly re-

TABLE II. Kd and Bmax of [125I]IOXY Binding to Rat m (RMOR)and Mutant Opioid m-Receptors Expressed in COS Cells†

RMOR Y148F H319A

Kd (nM) 2.36 6 0.18 1.23 6 0.01* 3.83 6 0.23*Bmax (fmol/mg of protein) 564 6 48 384.8 6 6.8* 233.7 6 17.5*

†Two concentrations of [125I]IOXY (0.014 and 0.114 nM) were each displaced by 8concentrations of IOXY (0.03815 to 2500 nM). The data of three independentexperiments were combined (n 5 60 data points) and then fit to a one-site bindingmodel for the best-fit parameter estimates (6SD) reported above. Statisticalsignificance between the binding parameters of sets of data was determined bysimultaneously fitting the two sets of data to the one site binding model first withthe parameters unconstrained and then a second time with the parametersconstrained to be the same values, as described in Methods. The F-test was usedto determine the corresponding P value for the increase in the sum-of-squareswhich resulted from the constraint.*P , 0.01 compared with wild-type control values.

TABLE III. Ki Values of Selective Opioid Ligands at Cloned Rat m and Mutant Opioid m ReceptorsLabeled by [125I]IOXY†

Test drug RMOR Y148F H319A Y148F/RMOR H319A/RMOR

RTI4614-4 0.59 6 0.10 2.71 6 0.30*** 20.5 6 2.1*** 4.59 34.7(1.15 6 0.21) (0.80 6 0.06) (0.67 6 0.04)

(2S,3R,4S)-1a 0.41 6 0.09 0.73 6 0.06* 6.83 6 1.01*** 1.78 16.7(0.72 6 0.11) (0.87 6 0.06) (0.78 6 0.08)

(2R,3R,4S)-1b 0.78 6 0.10 5.25 6 0.82*** 21.3 6 4.3*** 6.73 27.3(0.96 6 0.11) (0.78 6 0.08) (0.79 6 0.11)

(2R,3S,4R)-1c 205 6 50 980 6 106*** 1404 6 215*** 4.78 6.85(0.73 6 0.11) (0.95 6 0.09) (0.90 6 0.12)

(2S,3S,4R)-1d 96.6 6 15.5 574 6 74*** 960 6 141*** 5.94 9.94(0.97 6 0.13) (0.93 6 0.10) (1.1 6 0.1)

DAMGO 12.2 6 1.4 10.4 6 1.5 284 6 38*** 0.85 23.3(0.63 6 0.04) (0.66 6 0.06) (0.75 6 0.07)

[D-Ala2,D-Leu5]-enkephalin

137 6 24(0.90 6 0.13)

62.4 6 9.8**(0.62 6 0.05)

1084 6 183***(0.70 6 0.08)

0.46 7.91

IOXY 3.63 6 0.19 1.73 6 0.15*** 4.87 6 0.39** 0.48 1.34(1.07 6 0.05) (1.02 6 0.08) (1.11 6 0.09)

IOXY-AGO 10.0 6 1.4 8.02 6 0.76 34.2 6 2.9*** 0.80 3.42(0.86 6 0.09) (0.84 6 0.06) (1.00 6 0.07)

(2)Etorphine 0.96 6 0.15 1.01 6 0.15 1.91 6 0.22** 1.05 1.99(0.93 6 0.12) (1.0 6 0.1) (0.88 6 0.08)

Naloxone 8.65 6 1.10 9.36 6 1.14 12.3 6 1.9 1.08 1.42(1.02 6 0.11) (0.85 6 0.08) (0.91 6 0.12)

(2)Pentazocine 26.3 6 3.1 130 6 15*** 49.1 6 6.6** 4.94 1.87(0.93 6 0.09) (0.88 6 0.08) (0.90 6 0.09)

(2Buprenorphine 0.41 6 0.05(1.06 6 0.13)

0.43 6 0.08(0.90 6 0.14)

0.68 6 0.09**(1.01 6 0.12)

1.05 1.66

Etonitazine 0.26 6 0.04 1.09 6 0.19*** 12.4 6 0.9*** 4.19 47.7(1.14 6 0.18) (0.68 6 0.07) (0.70 6 0.03)

Morphine 9.62 6 3.19 29.3 6 9.0* 100 6 10*** 3.05 10.4(0.69 6 0.14) (0.64 6 0.11) (0.93 6 0.08)

Dermorphin 3.47 6 0.80 0.92 6 0.16*** 135 6 13*** 0.27 38.9(0.68 6 0.09) (0.81 6 0.10) (0.66 6 0.04)

Bremazocine 2.55 6 0.27 5.56 6 0.33*** 3.2 6 0.3 2.18 1.25(0.88 6 0.08) (1.0 6 0.1) (1.03 6 0.08)

†COS cells were transfected by electroporation with 20 µg/107 cells of plasmid DNA and then grown for 2–3 days asdescribed in Methods. Cells were harvested 48–60 hr after transfection. Cell pellets were homogenized in 50 mM Tris-HClbuffer (pH 7.4) using a polytron (setting #5, 30 sec) and centrifuged at 37,000 3 g for 10 min at 4°C. The pellets wereresuspended in 50 mM Tris-HCl buffer (pH 7.4) and used in radioligand binding studies. The data of three [125I]IOXYbinding experiments were pooled (n 5 60 data points) and fit to the two parameter logistic equation for the best-fitestimates of the IC50 and Slope Factor. Each value is 6 the SD. Statistical significance for the parameter ‘‘Ki’’ wasdetermined by simultaneously fitting two sets of data to the one site binding model first with the parametersunconstrained and then a second time with the parameters constrained to be the same value. As described in Methods, theF-test was used to determine the corresponding P value (compared with wild-type control values) for the increase in thesun-of-squares which resulted from the constraint.*P , 0.05.**P , 0.01.***P , 0.001.

27OPIOID PEPTIDE RECEPTOR STUDIES

duced the binding affinities of most opioid ligandstested but did not alter the binding affinities of nalox-one and bremazocine. The largest effects were seenwith RTI-4614–4, (2S,3R,4S)-1a, (2R,3R,4S)-1b,DAMGO, dermorphin, and etonitazine.

According to the model proposed by Tang et al.(1996), Asp147, Tyr148, Trp318, and His319 are the keyamino acids contributing to the binding affinity ofohmefentanyl (Fig. 2). The N1 atom on the piperidinylring is postulated to form potent electrostatic andhydrogen-bonding interactions with the carboxy groupof residue Asp147 on TM3 of the µ receptor. The O16atom in the carbonyl group is postulated to form weakelectrostatic and hydrogen-bonding interactions withthe imidazolyl plane of residue His319 in the TM7. Onephenyl ring is inserted into 2 aryl ring planes ofresidues Trp318 and His319 forming p-p interactions.The other phenyl ring forms a p-p interaction with thering plane of Tyr148. The electrostatic, hydrogen-bonding and p-p interactions promote the high affinitybinding of ohmefentanyl to the mu opioid receptor andits potent analgesic activity.

As noted above, Tang et al. (1996) hypothesized that:(1) H319 contributes to the binding of ohmefentanyl viaweak electrostatic and hydrogen bonding with the C 5O moiety, (2) that Y148 contributes to the binding ofohmefentanyl via hydrogen bonding with the 2-OHgroup, with the 2S configuration preferred over the 2Rconfiguration, and (3) that Trp318 contributes to thebinding of ohmefentanyl. The data reported here con-firm that H319 is important for the binding of RTI-4614–4, especially for the (3R,4S) configuration andthat H319 is also very important for the binding of theµ-selective ligands DAMGO, dermorphin, and eton-itazine. These data confirm the importance of Trp318for the binding of opioid ligands to the µ receptor but donot show a strong effect of Y148 on the binding ofRTI-4614–4. In particular, mutation of this residueshowed a stronger effect on the 2R configuration thanthe 2S configuration. Future studies will address theeffect of these mutations on agonist-stimulated [35S]-GTP-g-S binding, why H319 is important for the bind-ing of DAMGO, dermorphin, and etonitazine, why theY148 mutation decreases the binding of pentazocine,etonitazine, dermorphin, bremazocine, and the effect ofmutating Y148 to Ala or other amino acids to furthertest the hydrogen-bond and p-p interaction hypothesis.

REFERENCES

Brine GA, Stark PA, Liu Y, Carroll FI, Singh P, Xu H, Rothman RB.1995. Enantiomers of diastereomeric cis-N-[1-(2-hydroxy-2-phenyl-ethyl)-3-methyl-4-piperidyl]-N-phenyl propanamide: synthesis, x-ray analysis, and biological activities. J Med Chem 38:1547–1557.

Casy AF, Parfitt RT. 1986. Fentanyl and the 4-anilinopiperidine groupof analgesics. In: Casy AF, Parfitt RT, editors. Opioid analgesics.Chemistry and receptors. New York: Plenum Press, p 287–301.

Knott GD, Reece DK. 1972. MLAB: a civilized curve fitting system.Proc Online 1972 Int Conf 1:497–526.

Leysen J, Tollenaere JP, Koch MH, Laduron P. 1977. Differentiation ofopiate and neuroleptic receptor binding in rat brain. Eur J Pharma-col 43:253–267.

Lowry OH, Rosenbrough NH, Farr AL, Randall PJ. 1951. Proteinmeasurement with the Folin phenol reagent. J Biol Chem 193:265–275.

Lu Y F, Xu H, Partilla JS, Brine GA, Carroll FI, Rice KC, Sadee W,Rothman RB. 1996. Opioid peptide receptor studies. 5. Furtherevidence that the methylfentanyl congener RTI-4614–4 and its fourenantiomers bind to different domains of the rat µ opioid receptor.Analgesia 2:291–297.

Lu Y F, Xu H, Liu-Chen L-Y, Chen C, Partilla JS, Brine GA, Carroll FI,Rice KC, Lai J, Porreca F, Sadee W, Rothman RB. 1998. Opioidpeptide receptor studies. 7. The methylfentanyl congener RTI-4614–4 and its four enantiomers bind to different domains of the ratm opioid receptor. Synapse 28:117–124.

McLean S, Rice KC, Lessor RA, Rothman R B. 1987. [3H]CycloFOXY, aligand suitable for positron emission tomography, labels mu andkappa opioid receptors. Neuropeptides 10:235–239.

Minami M, Onogi T, Nakagawa T, Katao Y, Aoki Y, Katsumata S, SatohM. 1995. DAMGO, a mu-opioid receptor selective ligand, distin-guishes between mu-and kappa-opioid receptors at a differentregion from that for the distinction between mu- and delta-opioidreceptors. FEBS Lett 364:23–27.

Munson PJ, Rodbard D. 1980. LIGAND: a versatile computerizedapproach for characterization of ligand-binding systems. Anal Bio-chem 107:220–239.

Ni Q, Xu H, Partilla JS, de Costa BR, Rice KC, Rothman RB. 1993.Selective labeling of k2 opioid receptors in rat brain by [125I]IOXY:interaction of opioid peptides and other drugs with multiple k2abinding sites. Peptides 14:1279–1294.

Rosenbaum JS, Holford NH, Sadee W. 1984. Opiate receptor binding-effect relationship: sufentanil and etorphine produce analgesia atthe mu-site with low fractional receptor occupancy. Brain Res291:317–324.

Rothman RB, Bykov V, Mahboubi A, Long JB, Jiang Q, Porreca, F, deCosta BR, Jacobson AE, Rice KC, Holaday JW. 1991. Interaction ofbeta-funaltrexamine with [3H]cycloFOXY binding in rat brain:further evidence that beta-FNA alkylates the opioid receptor com-plex. Synapse 8:86–99.

Rothman RB, Xu H, Wang JB, Partilla JS, Kayakiri H, Rice KC, UhlGR. 1995. Ligand-selectivity of cloned human and rat opioid mureceptors. Synapse 21:60–64.

Surratt CK, Johnson PS, Moriwaki A, Seidleck BK, Blaschak CJ,Wang J-B, Uhl GR. 1994. µ Opiate receptor: charged transmem-brane domain amino acids are critical for agonist recognition andintrinsic activity. J Biol Chem 269(32):20548–53.

Tang Y, Chen KX, Jiang HL, Wang ZX, Ji RY, Chi ZQ. 1996. Molecularmodeling of µ opioid receptor and its interaction with ohmefentanyl.Acta Pharmacol Sinica 17:156–160.

Trumpp-Kallmeyer S, Hoflack J, Bruinvels A, Hibert M. 1992. Model-ing of G-protein-coupled receptors: application to dopamine, adrena-line, serotonin, acetylcholine, and mammalian opsin receptors. JMed Chem 35:3448–3462.

Van Bever WFM, Niemegeers CJE, Janssen PAJ. 1974. Syntheticanalgesics. Synthesis and pharmacology of the diastereoisomers ofN-[3-Methyl-1-(2-phenylethyl)-4-piperidyl]-N-phenylpropanamideand N-[3-Methyl-1-(1-methyl-2-phenylethyl)-4-piperidyl]-N-phenyl-propanamide. J Med Chem 17:1047–1051.

Wang J-B, Imai Y, Eppler CM, Gregor P, Spivak CE, Uhl GR. 1993. µopiate receptor: cDNA cloning and expression. Proc Natl Acad SciUSA 90:10230–10234.

Wilson RS, Rogers ME, Pert CB, Snyder SH. 1975. HomologousN-alkylnorketobemidones. Correlation of receptor binding withanalgesic potency. J Med Chem 18:240–242.

Xu H, Lu YF, Partilla JS, Brine GA, Carroll FI, Rice KC, Lai J, PorrecaF, Rothman RB. 1997. Opioid peptide receptor studies. 6. The3-methylfentanyl congeners RTI-4614–4 and its enantiomers differin efficacy, potency and intrinsic efficacy as measured by stimulationof [35S]GTP-g-S binding using cloned mu opioid receptors. Analgesia3:35–42.

Xue JC, Chen C, Zhu J, Kunapuli SP, de Riel JK, Yu L, Liu-Chen LY.1995. The third extracellular loop of the mu opioid receptor isimportant for agonist selectivity. J Biol Chem 270:12977–12979.

28 H. XU ET AL.