Dermorphin Gene Sequence Peptide with High Affinity and Selectivity for δ-Opioid Receptors

Skin of the frog Phyllomedusa sauvagei contains a cDNA sequence that codes for the selective preceptor peptide demorphin and a new heptapeptide we have designated as dermorphin  gene-associated peptide (DGAP). Investigation of the opioid receptor binding characteristics of synthetic DGAP and [D-M~~~IDGAP revealed that the latter peptide had high affinity and selectivity for 6-type opioid receptors in rat brain synaptosomes. The IC50 values for DGAP on pand 6receptors were only 28 pM and 670 nM, respectively, while that for [D-M~~']DGAP was 0.80 nM for 6-receptors and >1 p~ for p-receptors yielding a very high 6 selectivity ratio (SR) of 1345. In comparison, the SR values for [~-Ala',D-Leu~]enkephalin, [~-Ser',Leu', Thr'lenkephalin, and [~-Pen'~~]enkephalin, ligands which are considered to be specific for &receptors, were 20, 42, and 301, respectively. Dermorphin, which contains a D-Ala' residue and is a selective preceptor ligand (Lazarus, L. H., Guglietta, A., Wilson, W. E., Irons, B. J., and de Castiglione, R. (1989) J. Biol. Chem. 264, 354-362), exhibits a SR of 0.0055 similar to that for the conventional p-agonist [DAla2,NMePhe4,Gly-ollenkephalin (0.0040). This finding that frog skin cDNA contains the information to code for dermorphin and DGAP, or the presumed [DMet'IDGAP molecule, which are among the most selective high affinity opioid ligands described for pand 6receptors, may permit new insight into the design of future opioid receptor agonists and antagonists.

To whom correspondence should be addressed: NIEHS, P. heptapeptide isolated from the skin of the South American frog Phyllomedusa sauuagei (1, 2). The cDNA was recently shown to code for a repetitive pattern of five extensively homologous sequences, each comprising about 35 amino acids, four of which code for a single copy of the amidated heptapeptide DM (3). A fifth heptapeptide sequence, dermorphin gene-associated peptide (DGAP, Table I), is immediately flanked by amino acid sequences which are nearly identical to those surrounding DM, including paired dibasic amino acids (3). Thus, both DM and DGAP are potential products of similar peptide processing enzymes described in yeast (4, 5) and mammalian tissue (6). Precedents for the occurrence of multiple peptides with varying bioactivity profiles and receptor specificities within precursor polypeptides are exemplified by pro-opiomelanocortin (7), @-prepro-tachykinin (8), and calcitonin/calcitonin gene-related peptide (9, 10); in each case, bioactive peptides are excised by selective endoproteases (4-6) whose activities are influenced by the secondary structure of the precursor (11, 12). Prepro-DM cDNA possesses the codon for Ala2 in DM, while the cDNA code for DGAP indicates a Met' residue (3). However, DM contains a D-Ala' residue (1, 2); the D-enantiomer is rarely found outside microbial sources (13), which also contain specific D-amino acid peptidases (14). In this study, we sought to determine ( a ) the extent to which DGAP interacts with p-and 6-type opioid receptors and ( b ) by analogy with DM, the influence of the configuration about the a-carbon of Met2 on receptor affinities and selectivities.

EXPERIMENTAL PROCEDURES
The peptides DM, DGAP, and [ D -M~~~I D G A P were synthesized a t Farmitalia; DM was prepared by the solution method according to de Castiglione et al. (15), while DGAP and [ D -M~~~I D G A P were synthesized by conventional solid phase methods. DAGO, DADLE, DSLET (16), and DPDPE (17) were purchased from Peninsula Laboratories, Inc. (Belmont, CA) and Bachem (Torrance, CA).
The preparation of rat whole brain (minus cerebellum) synaptosomal membranes followed a slightly modified method (18) of that described by Chang and Cuatrecasas (19); the P2 fraction was initially preincubated a t 22 "C for 60 min in a dissociation solution containing 50 mM HEPES, pH 7.5, 50 pg of soybean trypsin inhibitor/ml, 100 mM NaC1, and 0.1 mM GDP in order to remove endogenously bound opioid peptides (20). The membrane fraction was then washed extensively in cold buffer (50 mM HEPES, pH 7.5, on ice) and stored at -70 "C in buffer containing trypsin inhibitor (50 pg/ml) and 20% glycerol (v/v) (18).
Binding studies were performed by incubating an aliquot of synaptosomes (800 pg of protein) in an assay containing a final concentration of 50 mM HEPES, pH 7.5, 20 pg of trypsin inhibitor, 8% glycerol (v/v), 1 mg of bovine serum albumin, 16 pg of bacitracin, 1 p M bestatin, and 0.2 nM [3H]DAG0 (1-type receptor agonist) or 0.2 nM [3H]DADLE (&receptor agonist) plus 2.6 ~L M [NMe-Phe3,0-Pro4] morphiceptin to suppress binding to p-receptor sites (20) in a final volume of 100 pl. After 60 min a t 22 "C, the membranes were collected on glass fiber filters (GF/C), presoaked in buffer containing 0.1% bovine serum albumin, and washed thrice in the same ice-cold buffer that in the presence of excess (2 p~) DAGO or DADLE for p-and 6-solution. Specific binding is the difference between total binding and receptor assays, respectively. The IC50 values (nM), representing the midpoints of the competition binding curves, are expressed as the mean k S.E. for the indicated number ( n ) of assays performed in duplicate.
Binding assays conducted with [3H]DADLE to label 8receptors indicated that [D-M~~']DGAP had &type receptor affinities similar to those of DADLE and DSLET (Fig. 2). Although &affinity of DPDPE was 4.5-fold less than those of the other &agonists (Table I), our ICw value was one-fourth of that published (17). The apparent &receptor affinities for DAGO and DM were more than 2 orders of magnitude less than that for DADLE (Table I). DGAP also exhibited low affinity for the 6-receptor with an IC50 930 times higher than that of DADLE, suggestive that DGAP has an intrinsic affinity for this receptor type. In contrast, [D-M~~']DGAP exhibits dramatically high affinity for 6-receptors (Table I).
The double reciprocal plot of peptide affinities versus receptor type selectivity (SR) (21) provides a very convenient means for presenting relationships between each of the peptide ligands (Fig. 3). That figure shows that the SR of the @specific ligands DAGO and DM is typically cO.01, while the &specific ligands have SR values greater than 10. The respective SR values for DADLE and DSLET are approximately 20 and 40, twice the values reported by others (21)

TABLE I
Peptide affinities for p-and 6-type opioid receptors on rat brain synaptosomes The sequences of the opioid peptides and their relative affinities for rat brain p-and &receptors are given by the ICso values, which represent the midpoint in the competitive binding curves illustrated in Figs. 1 and 2. The number of duplicate assay repetitions (n) is listed in parentheses after each ICbo value f S.E. The binding assays are described in the text and published elsewhere (18). The SR is defined as the ratio of the I C~/ I C W ' according to Shimohigashi et al. (21,22

Tyr-D-Ala-Gly-Phe-D-Leu-OH DSLET
Tyr-D-Ser-Gly-Phe-Leu-Thr-OH DPDPE -DGAP Two-dimensional analysis of receptor selectivity and the interrelationship between peptides. Plotted on the abscissa are the reciprocal IC50 values (nM-') for preceptors, with the exception of DGAP for which the IC60 value is estimated to be approximately 0.000036 nM", and on the ordinate, the ICs0 values for &receptors (nM"). Nonselective ligands (SR = 1) would fall within the stippled region of the graph, as reported (18,21). The diagonal lines indicate 10-fold differences in the SR; values >1 characterize pselective ligands, while values >1 indicate &selective ligands. The dushed lines connecting related structural sequence analogues emphasize the marked differences in opioid receptor selectivities: the DM prohormone-derived peptides (DM, DGAP, and [D-M~~*]DGAP) on the one hand, and the enkephalin-derived peptides (DAGO, DADLE, DSLET, and DPDPE) on the other. SR of 42 for DGAP reflects a preference for &receptors primarily based upon its very minimal interaction with preceptors (Table I); on the other hand, [D-M~~']DGAP exhibited a SR of over 1300. This SR value is 4.5-fold greater than that of DPDPE (SR = 301), which is considered to be one of the most effective synthetic &agonists; the SR for DPDPE (Table I) exceeds that value obtained under different assay conditions (17).

Tyr-Met-Phe-His-Leu-Met-Asp-NH2
Our findings indicate that even though the selectivity for the &receptor by [D-M~~']DGAP exceeds that for DADLE by over 1.5 orders of magnitude, the apparent receptor affinities for these peptides are identical ( Fig. 2 and Table I). By analogy, the presence of the D-isomer of Ala' in DM is absolutely essential for bioactivity (1,2,23,(25)(26)(27)(28) and binding to p-receptors (18,25,27); it also affects peptide conformation and stabilization in solution (29)(30)(31). The absence of Gly4 and Tyr' residues of DM in [D-M~~']DGAP (Table I) may prohibit this peptide from assuming the intramolecular H-bonded folded structure characteristic of DM (32) and DM N-terminal tetrapeptide (31). In the case of the enkephalin analogues, arguments have been presented that a more open molecular structure (33, 34) confers higher affinity to 6receptors (16,17,20,24); DAGO is the exception, being a pagonist (20,24). These data tend to support the membraneassisted opioid receptor selection model of Schwyzer (35) in that a C-terminal carboxyl function of the peptide is presumed to influence interaction with a positively charged group in the &type receptor site. The conformational properties imposed on DGAP by the inclusion of a D-Met' residue may permit the peptide to bind more tightly or selectively, or both, at a specific site in the &receptor.
At the present time, neither DGAP nor [D-M~~']DGAP has been isolated from extracts of frog skins. Although it would be of considerable interest to know whether the occurrence of D-amino acids in DM or DGAP reflects an unanticipated direct incorporation into the prohormone, or the more probable post-translational modification of the configuration around the a-carbon of L-Met' (3), our results indicate that a single frog skin gene codes for two distinct heptapeptides which, after appropriate processing, possess unique structural information to allow them to exhibit both high affinity and selectivity for p-and &type opioid receptors in rat brain. In addition to the primary sequence, secondary structural properties of the DM precursor may influence the processing which leads to an inversion of the configuration about the a-carbon of Ala' and Met' adjacent to the N-terminal tyrosine residues of the excised heptapeptides. In this regard, the amino acid sequence of the DM prohormone (3) resembles that of a de nouo-designed a-helical protein (36) that could suggest a model in which the C-terminal cleavage sites for DM and DGAP lie within a loop between a-helices. Similarly, the two @-turns in pro-opiomelanocortin enable site-specific cleavage while the a-helical content confers a degree of rigidity that prevents proteolysis (11). It is conceivable then that the composite structural features of the DM prohormone may dictate the sites for enzymic excision and for enantiomer inversion.
In conclusion, the recognition of differential receptor type selectivities of two opioid peptides, DM and [D-M~~*]DGAP, coded by a single gene, is unprecedented. These peptides may prove invaluable in opioid research in the facilitation of the design of new high affinity peptide agonists and antagonists (28,31,37,38), which may establish clinically relevant paradigms (39-43).