THE JOURNAL OF B I O ~ I C A L CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 269, No. 21, Issue of May 27, pp. 14957-14961, 1994 Printed in U S A .

Interaction of Glutamine 165 in the Fourth Transmembrane Segment of the Human Neurokinin-1 Receptor with Quinuclidine Antagonists*

(Received for publication, December 8, 1993, and in revised form, February 8, 1994)

Tung Ming Fond, Hong Yu, Margaret A. Cascieri, Dennis Underwood$, Christopher J. Swain?, and Catherine D. Strader From the Department of Molecular Pharmacology and Biochemistry and the §Department of Molecular Systems, Merck Research Laboratories, Rahway, New Jersey 07065 and the Weuroscience Research Centre, Merck Sharp & Dohme Research Laboratories, Harlow, Essex C“20 2Pz United Kingdom

Substance P binds to and activates the neurokinin-1 receptor with high affinity, thereby modulating several neuronal pathways including pain transmission and neurogenic inflammation. Several high affinity non-pep- tide antagonists have recently been described. To eluci- date the molecular interactions specific for binding to the neurokinin-1 receptor, site-directed mutagenesis has been utilized to identify amino acid residues that interact directly with antagonists. Glutamine 165 in the fourth transmembrane segment was shown to be critical for the binding of CP-96,345 but not SR140333. Analysis of quinuclidine analogs suggests that glutamine 165 in- teracts with the C-3 heteroatom in this class of antago- nists, probably through a hydrogen bond. Glutamine 165 also plays a minor role in the binding of peptides and RP67580. In contrast, serine 169 was determined to be critical for the binding of RP67580. These data indicate that residues 165 and 169 in the fourth transmembrane segment, along with residues in the fifth, sixth, and sev- enth transmembrane segments as demonstrated previ- ously, form the non-peptide antagonist binding site in the neurokinin-1 receptor. Furthermore, the antagonist binding site overlaps with the binding site for peptide agonists in the fourth and seventh transmembrane segments.

Substance P (SP)’ is an undecapeptide neurotransmitter that has been implicated in various neuromodulatory functions (Helke et al . , 1990; Henry, 1993; Iversen et al., 1987; Regoli et al . , 1989) and binds preferentially to the neurokinin-1 receptor (NKlR). Molecular cloning and heterologous expression have confirmed that the NKlR is a member of the G protein-coupled receptor family and activates the phosphatidylinositol hydroly- sis pathway (Fong et al . , 1992; Hershey and Krause, 1990; Nakanishi, 1991). The deduced amino acid sequence of the NKlR reveals seven hydrophobic regions that would be pre- dicted to form transmembrane helices, analogous to rhodopsin (Schetler et al., 1993). In addition to SP, the related decapep- tides neurokinin A ( N U ) and neurokinin B (NKB) are also NKlR agonists, albeit with lower affinity for the NKlR than SP.

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Merck Research Laboratories, P. 0. Box 2000,126 E. Lincoln Ave., Rah- $ To whom all correspondence should be addressed: Bldg. 80M-213,

way, NJ 07065. “el.: 908-594-6711; Fax: 908-594-3337. The abbreviations used are: SP, substance P; NKlR, neurokinin-1

receptor; NKA, neurokinin A NKB, neurokinin B; Ph, phenyl; lZ5I-SP, [‘25I-TyrSISP.

Because SP has been implicated in neurogenic inflammation and pain modulation, SP antagonists might be expected to be useful in treating these pathological conditions (Snider et al . , 1991; Garret et al., 1991). However, some potent SP antagonists also possess undesirable properties such as calcium channel binding activity (Schmidt et al., 1992). Therefore, it is impor- tant to elucidate the molecular interactions involved in antago- nist-NK1R binding in order to develop more specific and selec- tive SP antagonists. Various approaches have been employed to identify the ligand binding site of the NKlR (Fong et al . , 1992; Gether et al., 1993; Sachais et al., 1993; Yokota et al., 1992). In particular, site-directed mutagenesis and analysis of various antagonist analogs have demonstrated that His-197, His-265, and Tyr-287 of the human NKlR interact with non-peptide antagonists such as CP-96,345 or RP67580 (Fong et al . , 1993, 1994; Huang et al . , 1994). Binding energy calculations and analysis of the effect of amino acid substitutions on binding affinity indicate that interactions with these three residues cannot fully account for the nanomolar binding affinity of CP- 96,345 and RP67580 (Andrews, 1986). Therefore, other as yet unidentified side chains and/or the protein backbone must also be involved in antagonist binding. To further map the antago- nist binding site in the NKlR, several residues in the trans- membrane domain were substituted with Ala. In the present report, Gln-165 has been identified as another critical residue specifically involved in the binding of CP-96,345 to the NKlR. Structure-activity analysis of several analogs is consistent with the hypothesis that Gln-165 forms a hydrogen bond with the C-3 heteroatom in quinuclidine antagonists. In addition, Gln- 165 plays a minor role in the binding of SP to the receptor. Combined with previous results, these experiments indicate that the non-peptide antagonist binding site in the NKlR is defined by at least 5 residues in the fourth to seventh trans- membrane segments, while the specific interactions with dif- ferent antagonists can vary slightly within this region. These data should be useful in constructing a three-dimensional model for the NKlR and related receptors.

MATERIALS AND METHODS The cDNA encoding the human NKlR was described previously

(Fong et al., 1992). All single residue mutations were constructed using the uracil selection method of site-directed mutagenesis (Bio-Rad). Binding experiments were carried out using [‘251-TyralSP (lZ5I-SP) and intact COS cells expressing the cloned NKlR or mutant receptors (Fong et al., 1992). All peptides were purchased from Peninsula Laboratories, Inc., and non-peptide antagonists were synthesized as described previ- ously (Lowe et al., 1992; Peyronel et al., 1992; Swain et al., 1993).

The binding affinities of various ligands for the wild type and mutant receptors were determined in the presence of varying concentrations of unlabeled ligands (Fong et al., 1992). Briefly, the binding reaction mix- ture contained 0.2 n M lz5I-SP, unlabeled ligands, and intact COS cells

14957

14958 Antagonist Binding Site of the NKlR

FIG. 1. Schematic model of the human NKlR. The relative posi- tions of the mutated residues are indicated. Residues shown in black circles interact directly with one or more NK1 antagonists.

expressing the NKlR or mutant receptor in a final volume of 0.24 ml of 50 mM Tris, pH 7.4, 5 mM MnCl,, 150 mM NaCl, 0.04 mg/ml bacitracin, 0.004 mg/ml leupeptin, 0.2 mg/ml bovine serum albumin, and 0.01 mM phosphoramidon. Non-peptide antagonists were diluted in dimethyl sulfoxide, and a 2.4-pl aliquot was added to each tube. The cell number was adjusted so that the final receptor concentration was less than 0.04 nM. The binding assay was performed at 4 "C and stopped by filtration over GFIC filters. The data were fit to the equation (cpm(L) - cpm(1 PM SP))/(cpm(O) - cpm(1 p~ SP)) = IC,&. + IC5,), in which cpm(L) and cpm(0) represent bound lZ5I-SP in the presence or the absence of unla- beled ligand, respectively; L represents the concentration of unlabeled ligand; and IC,, represents the concentration of unlabeled ligand that causes 50% inhibition of the specifically bound 1251-SP. All mutant re- ceptors were analyzed along with the wild type receptor control.

RESULTS

To further map the ligand binding site in the human NKlR, residues expected to face the interior of the receptor were se- lected on the basis of Fourier transform analysis of the amino acid sequence of the transmembrane domains (Rees et al., 1989). Some of these residues were then substituted with Ala (Fig. l), and the ligand binding affinities of mutant receptors bearing these substitutions were measured. As shown in Table I, substitution of Met-81 in helix 2, Asn-109 or Pro-112 in helix 3, or Pro-164 or Tyr-168 in helix 4 with Ala did not markedly change the binding affinity of SP or CP-96,345. However, sub- stitution of Gln-165 in helix 4 with Ala (the Q165A mutant in Table I) resulted in a modest reduction in the binding affinities of SP and RP67580, whereas the binding affinities of CP-96,345, NKA, and NKl3 were substantially reduced (Tables I and I1 and Fig. 2). In contrast, the affinity of another NK1 antagonist, SR140333 (Lefevre et al., 19931, was not signifi- cantly affected by the Q165A substitution (IC5o = 0.2 f 0.1 nM for the wild type NKlR and 0.5 2 0.2 nM for the Q165Amutant).

To further characterize the role of residue 165 in ligand bind- ing, other amino acid substitutions at this position were tested. Substitution of Gln-165 with Ser or Asn did not markedly affect the affinity of SP (Table I). However, both of these mutant receptors had reduced affinity for NKA and NKB (Table 11). In addition, neither Ser nor Asn could replace Gln-165 in the high affinity binding of CP-96,345. In contrast to CP-96,345, the binding affinity of RP67580 was slightly increased by the Q165N substitution.

TABLE I Binding affinities of agonists and antagonists

IC,, values were determined from data such as those in Fig. 2. Mean -c S.E. and the number of independent experiments are listed.

Receptor IC50 SP CP-96,345 Rp67580

nM

hNK1R 0.6 f 0.2 (7) 0.5 f 0.1 (4) 18 .c 3 (4) M8 1A 0.3 2 0.1 (2) 0.8 5 0.2 (2) 84 f 2 (2) N109A 1.7 2 0.1 (2) 1.3 -c 0.7 (2) 44 2 17 (2) P112A 0.8 -c 0.1 (2) 1.5 f 0.2 (2) P164A

40 2 2 (2) 0.5 0.1 (3) 0.4 f 0.2 (3)

Q165A 32 f 3 (3)

Q165S 4.4 f 0.8 (6) 22.0 f 5 (2) 87 f 2 (2) 1.5 f 0.6 (4) 30.0 f 7 (2) 82 2 30 (2)

Q165N Y168A

1.1 0.3 (4) 61.0 f 30 (2) 5 5 0.8 (5) 0.7 f 0.1 (2)

S169A 0.7 f 0.1 (2) 55 f 5 (2)

0.3 0.2 (3) 2.6 f 0.3 (2) 339 91 (2)

TABLE I1 Binding affinity of peptides

Mean f S.E. and the number of independent experiments are listed.

Receptor IC,,

SP NKA NKB

40

20

-11 -10 -9 -8 -7 -6 -5

log [CP-96,345], M

I ' I 40 20 -11 -10 -9 -8 -7 -6 -5

log [SR140,333] , M

circles) and its Ql66A mutant (closed circbs) by SP, CP-96,345, FIG. 2. Inhibition of '261-SP binding to the human NKlR (open

and SR140333. Data shown here are representative of three to five independent experiments.

To identify which part of CP-96,345 may interact with Gln- 165, the interactions of analogs of CP-96,345 with the Q165A mutant receptor were analyzed. When the quinuclidine nitro- gen of CP-96,345 was replaced with carbon, the new compound (L-706,164) bound to the wild type receptor with a reduced affinity compared with CP-96,345. This compound showed a

Antagonist Binding Site of the NKlR 14959

TABLE I11 Binding affinities of quinuclidine analogs for the human N K l R and its Q165A mutant

R1

Compound a X Y R1 R2 R3 hNKl R Q165A Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

CP-96,345 N -NH- Ph Ph 2-0cH3 0.5 f 0.1 (4) 22 f 5 (2) 44

L-706,164 C -NH- Ph Ph 2-OCH3 22 f 6 (15) 297 f 66 (4) 14

L-703,605 N -NH- Ph Ph H

L-703,766 N -NH- Ph H 2-OCH3 232 f 31 (12) 1450 f 500 (2) 6

L-705,084 N -N(CH3) - Ph Ph 2-OCH3 8 f 3 (5) 21 f 10 ( 3 ) 3

L-706,158 N -0- Ph Ph 3-CF3 53 f 3 (2) 1760 f 500 (3) 33

L-706,220 N ( t r a n s ) =CH- Ph Ph 3-CF3 577 f 150 (8) 3300 f 2000 ( 3 ) 6

L-706,220 is 2 s and 2R racemates, whle all other compounds are 2

Ratio = IC, (&165A)/IC,, (hNK1R). * Mean S.E. from independent experiments are listed.

further 14-fold reduction in affinity for the Q165A mutant re- ceptor compared with the wild type (Table 111). Similarly, when the methoxy substituent of CP-96,345 was replaced with hy- drogen, the new compound (L-703,605) bound to the wild type receptor with a reduced affinity compared with CP-96,345. Its aEnity for the Q165A mutant was 75-fold lower than that for the wild type. Because the binding of L-706,164 or L-703,605 is sensitive to the Q165A substitution, it is inferred that neither the quinuclidine nitrogen nor the methoxy substituent inter- acts with Gln-165. When one phenyl ring was removed from the benzhydryl moiety of CP-96,345, the new compound (L-703,766) bound to the wild type receptor with a substan- tially reduced affinity compared with CP-96,345. Its affinity for the Q165A mutant was 6 times lower than that for the wild type. Therefore, L-703,766 is less sensitive to the chemical nature of the side chain at position 165 than is CP-96,345.

When the secondary amine in CP-96,345 was converted to a tertiary amine, the new compound (L-705,084) bound to the wild type receptor with a reduced affinity. However, the affinity of L-705,084 for the Q165A mutant (IC5o = 21 nM) was only marginally lower than that for the wild type receptor (IC5o = 8 nM), suggesting that the binding of L-705,084 is insensitive to the chemical nature of the side chain at position 165. Similarly, the Q165A mutant was insensitive to the addition of the N- methyl group to CP-96,345, as evidenced by the equal affinities of CP-96,345 and L-705,084 for the Q165A mutant receptor (Table 111).

The above data appeared to suggest a specific interaction between Gln-165 and the C-3 amine. To test this hypothesis from a different perspective, we compared the binding proper- ties of another pair of compounds. L-706,158 contains an ether linkage at the C-3 position, whereas L-706,220 is similar to L-706,158 except for the ethylene group at the C-3 position. As

,3-cis racemates.

shown in Table 111, the binding affinity of L-706,220 for the wild type receptor was lower than that of L-706,158, suggesting that the oxygen atom is required for high affinity binding of quinu- clidine ethers to the NKlR. The Gln-165 of the NKlR is also required for the binding of L"706,158, as evidenced by the 33- fold reduction in its affinity for Q165A. In contrast, the Q165A substitution caused only a 6-fold reduction in the affinity of

Although Gln-165 is critical for the binding of CP-96,345, this residue plays a relatively minor role in the binding of the perhydroisoindole antagonist RP67580 (Table I). To further map the binding site for RP67580, we have investigated the role of nearby residues. Substitution of Ser-169, which is pre- sumably one helical turn above Gln-165, had no effect on the binding of SP (Table I). In contrast, the S169A substitution resulted in a 19-fold reduction in the affinity of RP67590, with a concomitant 5-fold reduction in the affinity of CP-96,345. These data further support the hypothesis that the fourth transmembrane segment forms part of the binding site for these small molecule antagonists.

L-706,220.

DISCUSSION

The present study represents a continuing effort to map the antagonist binding site in the NKlR. Previous studies have implicated His-197 in helix 5, His-265 in helix 6, and Tyr-287 in helix 7 in direct interactions with non-peptide antagonists (Fong et al., 1993, 1994; Huang et al., 1994). We have now identified another region in the fourth transmembrane seg- ment, encompassing Gln-165 and Ser-169, that plays a critical role in the binding of the non-peptide antagonists CP-96,345 and RP67580. The effects of substituting Gln-165 or Ser-169 on antagonist binding affinity do not appear to result from gross conformational changes because not all ligands have signifi-

14960 Antagonist Binding Site of the NKlR cantly reduced affinity in these substitution mutants. However, the possibility of subtle local conformational changes cannot be ruled out without a three-dimensional structure of the receptor.

Comparison of CP-96,345 and L705,084 showed that the binding of L-705,084 is not sensitive to the Q165A mutation (Table 111). Because the only structural difference between the two compounds is the hydrogen versus methyl of the benzylic amine, it is postulated that Gln-165 interacts with the benzylic amino group. This hypothesis is further tested with L-706,158 and L-706,220, where the only structural difference between the two compounds is the ether versus ethylene a t C-3 position. Again, the binding of L-706,220 is relatively insensitive to the Q165Amutation compared with L-706,158. These data suggest that Gln-165 is involved in the binding of quinuclidines having an 0 or NH group at C-3, possibly through a hydrogen bond between these groups and Gln-165. I t should be pointed out that the affinities of L-705,084 and L-706,220 for the wild type receptor are slightly higher than those for the Q165A mutant receptor. This small difference in affinities may result from a better van der Waals' fit between these antagonists and the wild type receptor or a subtle conformational change in the receptor due to the Q165A substitution. Alternatively, Gln-165 may also interact weakly with a second functional group in these quinuclidine antagonists. One potential interaction sug- gested by modeling studies may involve the benzhydryl moiety, because the benzhydryl group is in proximity to Gln-165 when the benzhydryl interacts with His-197 (Fig. 3B) (Fong et al., 1993). This is consistent with the observation that the binding affinity of L-703,766 is less sensitive to the Q165A substitution than is CP-96,345.

Ser-169 is more important for the binding of RP67580 than is Gln-165. The specific intermolecular interactions with this residue have not been determined for RP67580, although Ser- 169 is probably involved in a hydrogen bond. The identification of Gln-165 and Ser-169 as part of the antagonist binding site indicates that different antagonists can bind to the same gen- eral region within the receptor, although the specific intermo- lecular interactions utilized by different antagonists vary within this region. For example, CP-96,345 appears to interact strongly with Gln-165 but weakly with Ser-169, while the re- verse is true for RP67580 (Table I).

In addition to playing a critical role in non-peptide antago- nist binding, Gln-165 appears to interact with peptide agonists as well. Although substitution of Gln-165 has a minor effect on SP binding, NKA and NKB binding are more sensitive to the substitution of Gln-165. Further studies are required to eluci- date the differential interactions of Gln-165 with the three peptides. The observation that both peptide agonists and non- peptide antagonists interact with Gln-165 is consistent with previous pharmacological studies indicating that these non- peptide antagonists are competitive antagonists of SP binding (Cascieri et al., 1992; Snider et al., 1991; Garret et al., 1991).

The identification of several residues in helices 4-7 of the NKlR as forming part of the non-peptide antagonist binding site allows us to further refine our model of the ligand binding domain of the NKlR. In the absence of a high resolution struc- ture, a model of the NKlR can be constructed based on the assumption of seven transmembrane a-helices. However, the relative placement of each transmembrane helix and its lateral orientation are more speculative. In light of the mutagenesis results demonstrating that Gln-165, Ser-189, His-197, His-265, and "287 form part of the antagonist binding site, it is pos- sible to refine the NKlR model by restricting the interresidue distances to the limits of flexibility of these non-peptide antago- nists. We have previously shown that His-197 interacts with the benzhydryl moiety, and His-265 is in proximity to the sub- stituted benzyl moiety of some quinuclidine antagonists (Fong

A Helix 7

I

" " _ H265

interactions between the human NKlR and CP-96,345 are indicated by FIG. 3. A, schematic model of antagonist binding site, in which the

arrows. B , top view of a three-dimensional model of the human N K l R with CP-96,345 bound. Only the transmembrane helices (coils) of the NKlR are shown. Gln-165. His-197, and His-265 are represented as Corey-Pauling-Koltun model in gray, and CP-96.345 is represented as Corey-Pauling-Koltun model in black.

et al., 1993, 1994). The interaction of Gln-165 with the C-3 heteroatom imposes another constraint on the three-dimen- sional model of the NKlR (Fig. 3). Furthermore, the observa- tions that retinal (Thomas and Stryer, 1982), adrenergic ago- nists and antagonists (Straderet al., 1989). and non-peptide SP antagonists bind to the transmembrane domain of their respec- tive receptors suggest that this region of G protein-coupled receptors may form a general binding site for small molecules.

The current results also indicate that two conserved Pro residues in the NKlR do not appear to play a role in specific interactions with SP or several antagonists, although Pro has the potential of producing a kink in helices (Williams and De- ber, 1991). Pro-112 is unique to all three subtypes of neurokinin receptors. Its location in helix 3, at a position analogous to the Asp-113 that binds the amino group of biogenic amines in ad- renergic receptors, led to the speculation that this region of the NKlR may be important for SP binding. However, the lack of effect of the P112Asubstitution suggests that this residue is not directly involved in SP binding, although chimeric receptor studies have suggested that helices 2 4 may still be important for subtype selectivity (Yokota et al., 1992). Pro-164 is next to Gln-165 in the primary amino acid sequence and is conserved

Antagonist Binding Site of the NKlR 14961

in most G protein-coupled receptors (Attwood et al., 1991). However, the P164A substitution does not affect the binding affinity of CP-96,345. Since the Q165N substitution substan- tially reduces the affinity of CP-96,345, any significant dislo- cation of the Gln-165 side chain would be expected to affect the binding of CP-96,345. The absence of effect of the P164A sub- stitution implies that the precise position of Gln-165 is not significantly affected upon Pro-164 substitution. One likely ex- planation is that structural constraints elsewhere in the recep- tor are sufficient to maintain the local conformation of Gln-165.

In summary, the present results establish that the non-pep- tide antagonist binding site in the NKlR is defined by at least 5 residues (Gln-165, Ser-169, His-197, His-265, and Tyr-287). While this region of the NKlR appears to accommodate many different antagonists, the specific interactions utilized by each small molecule vary and depend on the bound conformation of the antagonist. The binding site information derived from mu- tagenesis studies will be useful in further analysis of the mechanisms of receptor activation and antagonism.

Acknowledgments-We thank Drs. T. Ladduwahetty, V. Sabin, E. Seward, B. Williams, P. Finke, and M. MacCoss for synthesizing various non-peptide antagonists and R. R. C. Huang and L. L. Shiao for help with mutagenesis and binding assays.

REFERENCES

Andrews, P. R. (1986) Dends Pharmacol. Sci. 1986, 148-151 Attwood, T. K., Eliopoulos, E. E., and Findlay, J. B. C. (1991) Gene (Amst. 98,

Cascieri, M. A,, Ber, E., Fong, T. M., Sadowski, S., Bansal, A,, Swain, C., Seward, 153-159

E., Frances, B., Burns, D., and Strader, C . D. (1992) Mol. Pharmacol. 42, 458-463

Fong, T. M., Yu, H., Huang, R. R. C., and Strader, C. D. (1992) Biochemistry 31, 1180611811

Fong, T. M., Cascieri, M. A., Yu, H., Bansal,A., Swain, C., and Strader, C. D. (1993)

Fong, T. M., Yu, H., Cascieri, M. A,, Underwood, D., Swain, C . J., and Strader, C. Nature 362,350-353

D. (1994) J. Biol. Chem. 269,2728-2732 Garret, C., Carruette, A,, Fardin, V., Moussaoui, S., Peyronel, J.-F., Blanchard,

Gether, U., Johansen, T. E., Snider, R. M., Lowe, J. A,, 111, Nakanishi, S., and J.-C., andLaduron, P. M. (1991)Proc. Natl. Acad. Sci. U. 5. A. 88,10208-10212

Helke, C. J., Krause, J. E., Mantyh, P. W., Couture, R., and Bannon, M. J. (1990) Schwartz, T. W. (1993) Nature 362, 345-348

Henry, J. (1993) Regul. Pept. 46, 138-143 FASEB J. 4,1606-1615

Hershey, A. D., and Krause, J. E. (1990) Science 247, 95&962 Huang, R. R. C., Yu, H., Strader, C. D., and Fong, T. M. (1994) Biochemistry 33,

Iversen, L. L., Watling, K J., McKnight, A. C., Williams, B. J., and Lee, C. M. 3007-3013

Lefevre, I. A., Oury-Donat, E , Bordey, A,, Thurneyssen, O., Emonds-Alt, X., Feltz, (1987) Top. Med. Chem. 65, 1-9

Lowe, J. A,, 111, Drozda, S. E., Snider, R. M., Longo, K. P., Zorn, S. H., Morrone, J., P., Lefur, G., and Soubrie, P. (1993) SOC. Neurosci. Abstr 19, 722

Jackson, E. R., McLean, S., Bryce, D. K., Bordner, J., Nagahisa, A., Kanai, Y., Suga, O., and Tsuchiya, M. (1992) J. Med. Chem. 35,2591-2600

Nakanishi, S. (1991) Annu. Reu. Neurosci. 14, 123-136 Peyronel, J.-F., Truchon, A,, Moutonnier, C., and Garret, C. (1992) Bioorg. & Med.

Regoli, D., Drapeau, G., Dion, S., and DOrleans-Juste, P. (1989) Pharmacology 38,

Rees, D. C., DeAntonio, L., and Eisenberg, D. (1989) Science 245,510-513 Sachais, B. S., Snider, R. M., Lowe, J. A,, 111, and Krause, J. E. (1993) J. Biol.

Schertler, G. F., Villa, C., and Henderson, R. (1993) Nature 362,770-772

Snider, R. M., Constantine, J. W., Lowe, J. A., 111, Longo, K. P., Lebel, W. S., Woody, Schmidt, A. W., McLean, S., and Heym, J. (1992) Eur J. Pharmacol. 215,351-352

H. A., Drozda, S. E., Desai, M. C., Vinick, F. J., Spencer, R. W., and Hess, H. J. (1991) Science 251, 435437

Chem. Lett. 2 , 3 7 4 0

1-15

Chem. 268,2319-2323

Strader, C. D., Sigal, I. S., and Dixon, R. A. F. (1989) FASEB J. 3, 1825-1832 Swain, C. J., Seward, E. M., Sabin, V., Owen, S., Baker, R., Cascieri, M. A.,

Sadowski, S., Strader, C . D., and Ball, R. G. (1993) Bioorg. & Med. Chem. Lett. 3, 1703-1706

Thomas, D. D., and Stryer, L. (1982) J. Mol. Biol. 154, 145-157 Yokota, Y., Akazawa, C., Ohkubo, H., and Nakanishi, S. (1992) EMBO J. 11,

Williams, K. A., and Deber, C. M. (1991) Biochemistry 30, 891943923 3585-3591