Selective Interactions between Helix VIII of the Human μ-Opioid Receptors and the C Terminus of Periplakin Disrupt G Protein Activation*

Analysis of interactions between the C-terminal tail of the MOP-1 and MOP-1A variants of the human μ-opioid receptor with proteins derived from a human brain cDNA library resulted in identification of the actin and intermediate filament-binding protein periplakin. Mapping of this interaction indicated that the predicted fourth intracellular loop/helix VIII of the receptor interacts with the C-terminal rod and linker region of periplakin. Periplakin is widely expressed in the central nervous system of both man and rat and demonstrated an overlapping but not identical distribution with μ-opioid (MOP) receptors. Co-expression of periplakin with MOP-1 or a MOP-1-eYFP fusion construct in HEK293 cells did not interfere with agonist-mediated internalization of the receptor. When co-expressed with a MOP-1-Gi1α fusion protein periplakin significantly reduced the capacity of the agonist to stimulate binding of 35S-labeled guanosine 5′-3-O-(thio)triphosphate ([35S]GTPγS) to the receptor-associated G protein. By contrast, periplakin did not interfere with agonist-stimulation of [35S]GTPγS binding to either an α2A-adrenoreceptor-Gi1α fusion protein or a β2-adrenoreceptor-Gsα fusion protein, indicating its selectivity of function. This represents the first example of an opioid receptor-interacting protein that functions to disrupt agonist-mediated G protein activation.

The basic module of G protein-mediated signal transduction has long been considered to be a G protein-coupled receptor (GPCR), 1 a G protein, and an effector. However, in recent times a wide range of proteins has been identified that interacts with either GPCRs (1)(2)(3) or G proteins (4 -6), and these can modulate signal transduction efficiency, cellular localization, or the regulation of these polypeptides. Although certain protein-protein interactions can be anticipated based on the presence of well characterized protein interaction motifs in the primary sequence of the GPCR (1), many of the reported interactions do not involve previously characterized motifs. A widely used strategy to identify such interactions is the yeast two hybrid system (7).
The core opioid GPCR family comprises the MOP, KOP, and DOP receptors (8 -9). These have been studied extensively in the search for non-addictive analgesics and particular interest centers on the MOP receptor because it mediates most of the actions of morphine and other clinically relevant analgesic agents as well as drugs of abuse such as heroin. Mice in which the genes for the various opioid receptors have been knocked out have contributed significantly to understanding (10). A number of MOP receptor subtypes have been defined pharmacologically (11)(12), and it is possible that these represent hetero-dimers containing both the MOP and either DOP, KOP (12)(13)(14)(15), or other related GPCRs (16). In rodents a number of distinct MOP splice variants have been described (17)(18)(19) that vary in distribution (20). In man, two variants, MOP-1 and MOP-1A, were described initially (21), but further potential variation has recently been indicated (22).
Recent studies have demonstrated a number of opioid receptor-interacting proteins. These include the ezrin-radixin-moesin binding phosphoprotein-50/Na ϩ /H ϩ exchanger regulatory protein that has been shown to interact with KOP and prevent agonist-induced down-regulation of the receptor by enhancing its recycling rate (23) and a GPCR-associated sorting protein shown to interact with DOP and alter the recycling characteristics of this receptor (24). Further studies have recently indicated an interaction between the rat MOP and phospholipase D2 (25). This also appears to be involved in the regulation of agonist-induced internalization of the receptor (25). These interactions either did not produce significant effects on G protein activation by the receptors or this issue was not examined.
By analysis of proteins identified to interact with the Cterminal tail of the human MOP-1 and MOP-1A isoforms herein we demonstrate the interactions of these receptors with periplakin (PPL). Periplakin does not significantly alter agonist-induced internalization of MOP-1, but by interacting with the postulated helix VIII of the receptor that likely runs parallel to the plasma membrane (26 -29) it interferes with agonist-mediated activation of G protein. This region of rhodopsin has been demonstrated to play an integral role in G protein activation (30). This interaction is selective because the presence of periplakin did not interfere with the ability of agonists at the ␣ 2A -adrenoreceptor or the ␤ 2 -adrenoreceptor to activate their cognate G proteins. Generation of DNA Constructs, Yeast Two-hybrid Analysis, TAQ-MAN Analysis, and Immunoblotting Studies-PPLC was inserted into PQE30 (Qiagen) to generate His-tagged protein. MOP-1-eYFP was made by inserting MOP-1 into plasmid eYFP-Ni (Clontech). All truncated forms of MOP-1 were generated by PCR using human MOP-1 cDNA as the template. For GST fusions these truncations were cloned into PGEX-4T-1 (Amersham Biosciences), whereas for His fusions they were cloned into PQE30 (Qiagen). Fusion proteins between MOP-1 and Cys 351 -Ile-G i1 ␣ (31), the ␣ 2A -adrenoreceptor and Cys 351 -Ile-G i1 ␣ (32), and the ␤ 2 -adrenoreceptor and G s ␣ (33) have been described previously. Yeast two-hybrid analysis was conducted as described previously (34). RNA purification and TAQMAN reverse transcription-PCR analysis of human tissue were performed as described previously (35). Immunoblotting studies were performed on homogenates of dissected rat brain regions or on pre-prepared gel-ready Medley samples of human brain obtained from BD Biosciences.
Cell Culture and Transient Transfection-HEK293 cells were maintained in DMEM containing 10% newborn calf serum and 2 mM glutamine. The day before transfection, cells were seeded either in 10-cm dishes or on coverslips in 6-well plates at 50 -70% confluency. Transfection was performed using LipofectAMINE reagent (Invitrogen). 48 h later cells in the dishes were washed twice with ice-cold PBS in situ, harvested in 5 ml of PBS, and pelleted by centrifugation at 1,600 rpm at 4°C. These pellets were kept at Ϫ80°C until membrane preparation. For the cells in 6-well plates, 24 h after transfection they were fixed or subjected to immunostaining as follows.
Immunofluorescence Staining and Confocal Microscopy-Cells on coverslips were fixed in 4% paraformaldehyde in PBS containing 5% sucrose for 10 min at room temperature. Cells were then permeabilized for 10 min in TM buffer (0.15% Triton X-100 and 3% nonfat milk in PBS). Coverslips were incubated for 1 h at room temperature with mouse anti-HA antibody (2.5 g/ml, Roche Applied Science), washed in TM buffer and PBS, and then incubated for a further 1 h with Alexaconjugated anti-mouse 594. After washing, coverslips were mounted onto glass slides and examined using a laser-scanning Zeiss LSM510 confocal microscope.
Preparation of Membranes-Cell pellets were resuspended in TE buffer (10 mM Tris HCl/0.1 mM EDTA, pH 7.5) and homogenized with 30 -50 strokes of a Teflon-on-glass homogenizer. Unbroken cells and nuclei were removed by centrifugation at 1000 rpm for 10 min. The supernatant was then centrifuged at 50,000 rpm for 30 min. The pellets were resuspended in TE buffer at Ϫ80°C until use. 3 H Ligand Binding Assays-The level of expression of MOP-1 and the MOP-1-Cys 351 -Ile-G i1 ␣ fusion protein was determined by the binding of [ 3 H]diprenorphine (2 nM) in TEM buffer (75 mM Tris⅐HCl, pH 7.4, 1 mM EDTA, 12.5 mM MgCl 2 ). Nonspecific binding was defined with 50 M naloxone. Samples were incubated at 25°C for 1 h and stopped by adding 5 ml of cold TE buffer followed by immediate filtration through GF/C filters and washing. Binding assays to measure levels of expression of the ␣ 2A -adrenoreceptor-Cys 351 -Ile-G i1 ␣ and the ␤ 2 -adrenoreceptor-G s ␣ fusion proteins have been described previously (32)(33).
Receptor Internalization Assay by Biotin Labeling of MOP-1-24 h after transfection with or without HA-PPL cells were transferred into 6-well plates and cultured for further 24 h. Cells were incubated with the MOP-selective agonist DAMGO (10 M) for varying times and washed immediately 2ϫ with PBS and 2ϫ with PBS-CM (PBS containing 1 mM MgCl 2 , 0.1 mM CaCl 2 ). Biotin labeling was performed in a dark room under dim light. Cells were treated with ice-cold 10 mM sodium periodate in PBS-CM and then with ice-cold 1 mM Biotin-LC-hydrazide in acetate buffer (0.1 M sodium acetate, 1 mM MgCl 2 , 0.1 mM CaCl 2 ) for 30 min. After 3 washes with PBS, cells were lysed in radioimmune precipitation assay buffer (50 mM HEPES pH 7.5, 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 5% ethylene glycol, and a mixture of protease inhibitors). Solubilized proteins were cleared by a 15-min centrifugation at 13,000 rpm at 4°C. Equal amounts of protein from each sample were used for immunoprecipitation using an antibody against the C terminus of human MOP-1 (anti-MT43). Immuno-complexes were immobilized by protein A-Sepharose 4B (Sigma), resolved in 4 -12% Tris-Bis NuPAGE (Invitrogen), and transferred to nitrocellulose. Biotin-labeled MOP-1 was detected by horseradish peroxide-conjugated streptavidin and visualized by ECL.
Purification of His-tagged Proteins-Plasmids bearing the desired His fusion inserts were transformed into competent Escherichia coli DH5a cells. From an overnight culture, 10 ml was used to inoculate 500 ml of LB media containing 100 g/ml ampicillin, and cells were allowed to grow at 37°C until the culture reached an A 600 of 0.4 -0.6. 1 mM isopropyl-␤,D-thiogalactopyranoside was added for 4 h before harvesting by centrifugation at 8,000 rpm for 15 min at 4°C. The pellet was resuspended in 10 ml of lysis buffer, and His-tagged proteins were purified according to the manufacturer (Qiagen). Eluted proteins were dialyzed against at least three changes of PBS containing 5% glycerol at 4°C over a period of 2 days before storage at Ϫ80°C.
Purification of GST Fusion Proteins and GST Pull-down Assays-Bacterial cultures as above were harvested at 8,000 rpm for 15 min. The cell pellets were then lysed in 10 ml of BugBuster containing 10 l of benzonase (Novagen) and a protease inhibitor mixture, incubated for 1 h at room temperature with rotation, and cleared by centrifugation at 16,000 rpm for 30 min at 4°C after 2 ϫ 1 min of sonication. The supernatants were either stored at Ϫ80°C until used for GST pulldown assays or directly purified using glutathione-Sepharose 4B beads (Amersham Biosciences).
For GST pull-down assays, 1-5 ml of the soluble lysates were incubated with 100 l of 50% (v/v) slurry of glutathione-Sepharose beads for 2 h at 4°C. After a brief centrifugation (1000 rpm, 2min), the beads were washed 3ϫ with PBS containing 1% Triton X-100 and resuspended in 1 ml of PBS/Triton X-100 containing 50 g of required His fusions. The mix was then incubated for an additional 2 h before collection of the beads. These were washed 5ϫ with PBS/Triton X-100, washed again with PBS, and then eluted in 50 l of 10 mM glutathione in a Tris⅐HCl buffer, pH 8.0. The eluates were resolved, and the Histagged protein was detected by immunoblotting. Samples were precleared with Pansorbin (Calbiochem) and immunoprecipitated with an antiserum that identifies the C-terminal decapeptide of G i1 ␣. Finally, the immunocomplexes were washed twice with solubilization buffer, and bound [ 35 S]GTP␥S was measured by liquid scintillation spectrometry. Equivalent experiments were performed after transfection of fusion proteins between the ␣ 2A -adrenoreceptor and Cys 351 -Ile-G i1 ␣ (32) or between the ␤ 2 -adrenoreceptor and G s ␣ (33) in the presence or absence of HA-PPL. End of assay immunocapture of the ␤ 2 -adrenoreceptor-G s ␣ fusion protein utilized an antiserum that identifies the C-terminal decapeptide of G s ␣ (33).

RESULTS
Protein Interaction Studies-Bio-informatic analysis of expressed sequence tags suggests only two forms of MOP are expressed in man. These differ only in the extreme C-terminal tail, resulting in MOP-1A being eight amino acids shorter than MOP-1. Based on the presence of a potential, weak coiled-coil domain in the C terminus of human MOP-1 we sought proteininteracting partners for the C-terminal tails of MOP-1 and MOP-1A. The C-terminal 75 amino acids of MOP-1 and 67 amino acids of MOP-1A were employed as bait for yeast twohybrid screens using a human brain-derived cDNA library. Of 147 hits for the MOP-1 tail from 1.2 ϫ 10 7 -transformed cDNAs and 58 hits for the MOP-1A tail from 6 ϫ 10 6 -transformed cDNAs, multiple clones corresponded to PPL (36 -38) (Fig. 1a).
The interaction of PPL with both MOP-1 and MOP-1A eliminated the possibility that binding was to the extreme C terminus of MOP-1. The interaction with the C-terminal tails of the MOP-1 and MOP-1A receptors required the extreme C-terminal linker domain of PPL and/or part of the C-terminal region of its rod domain as all the characterized hits from the yeast two hybrid assays contained these regions. A series of fragments of the C-terminal tail of MOP-1 was generated (Fig. 1b) and used in further yeast two hybrid studies. These studies indicated the site of interaction must be in a region proximal to the plasma membrane because truncation to Ser 358 , just beyond the putative palmitoylated cysteine residues (Cys 348 and Cys 353 ), did not compromise the interaction. Given the similarity of sequence of the MOP and DOP receptors in this region, it was interesting to note that equivalent screens using the Cterminal tail of the DOP receptor also resulted in positive identification of PPL as an interacting protein (data not shown).
To confirm the results of the yeast two-hybrid analyses, a GST fusion protein containing the C-terminal 80 amino acids of MOP-1 was generated and linked to glutathione-Sepharose 4B beads. This was able to capture a His-tagged form of the Cterminal 208 amino acids of PPL (PPLC) (Fig. 2a). When GST fusion proteins of both the C-terminal 43 amino acids and the remaining 37 amino acids of the 80 amino acid MOP-1 fragment were generated, interaction with PPLC was preserved only for the membrane proximal 37-amino acid section (Fig.  2a). Elimination of 10 or 20 amino acids from the N terminus of the 80-amino acid fragment did not prevent interactions with PPLC (Fig. 2a). However, removal of a further 10 amino acids abolished the interaction (Fig. 2a). The combination of these data (Fig. 2b) delimited the site of interaction of the C-terminal 208 amino acids of PPL with the region between amino acids 341-351 (LDENFKRCFRE) of MOP-1. Based on the structure of bovine rhodopsin (26) and comparisons of other related GPCRs (29), this sequence is likely to represent helix VIII of the MOP-1 and MOP-1A receptors. Further GST fusion proteins, including a 23-amino acid segment corresponding to the third intracellular loop of the human MOP receptors, failed to indicate interactions between PPLC and other linear, intracellular regions of the MOPs (data not shown). Importantly for subsequent studies, the addition of enhanced green fluorescent protein to the C-terminal end of the MOP-1 tail did not prevent interaction between the C-terminal tail of the receptor and PPLC (Fig. 2c). Reciprocal experiments demonstrated that a GST fusion protein containing the C-terminal 208 amino acids of PPL (amino acids 1548 -1756) was able to capture a Histagged form of the MOP-1 C-terminal tail (Fig. 2d). However, we were unable to define the region of PPLC responsible for this interaction in detail as two similar-sized sections derived from this C-terminal region (periplakin 1548 -1666 and periplakin1667-1756) were both unable to capture the Histagged form of the MOP-1 C-terminal tail (data not shown).
Expression and Distribution of Periplakin-Low levels of MOP-1 transcripts were detected in a number of distinct regions of human brain using TAQMAN quantitative reverse transcription-PCR (Fig. 3a). Although PPL has been most actively studied as a 195-kDa protein of the keratinocyte cytoskeleton and desmosomes (36 -37), as noted previously (38), its transcript was also highly and widely expressed in the central nervous system. High levels of mRNA were detected in regions of human brain including the frontal and temporal TAQMAN quantitative reverse transcription-PCR was used to detect the expression of MOP-1 (a) and PPL (b) mRNA in discrete regions of human brain. Homogenates (50 g of protein) of pre-prepared, gel-ready samples of human brain regions (c) or of dissected regions of rat brain (d) were resolved by SDS-PAGE and immunoblotted with an antiserum directed against the C-terminal region of PPL. e, cell lysates were generated from HEK293, SHSY-5Y, NG108 -15 cells and from human keratinocytes (K.cytes). After resolution by SDS-PAGE of 10 g of each cell lysate, samples were immunoblotted with antiserum CR3 (37) directed toward the C terminus of PPL. lobes, amygdala, thalamus, hippocampus, and cerebellum (Fig.  3b). Immunoblotting studies with antibodies directed toward the C-terminal region of PPL identified a polypeptide of some 195 kDa in lysates of all available regions of human (Fig. 3c) and rat (Fig. 3d) brain. Immunodetected levels of PPL were relatively similar in individual, gross regions of human brain and were especially high in the pituitary and olfactory bulb from rat. Antibodies directed toward the N- (Fig. 3e) or Cterminal (data not shown) of PPL also easily detected a single 195-kDa protein in human keratinocytes. However, PPL expression in the neuron-derived cell lines, NG108 -15 and SHSY-5Y, was below immunodetectable levels (Fig. 3e). Anti-PPL antibodies were also unable to detect protein expression in HEK293 cells (Fig. 3e).
Periplakin Does Not Interfere with Internalization of MOP-1-Transient expression of full-length MOP-1-eYFP in HEK293 cells resulted in a predominantly plasma membrane-delineated distribution when examined by confocal microscopy (Fig. 4a). Exposure of these cells to the highly MOP selective enkephalin analogue DAMGO (10 M) resulted in rapid redistribution of the eYFP signal to punctate intracellular vesicles that are likely to represent recycling endosomes (Fig. 4a). HA-tagged PPL was distributed more widely in cells but was excluded from the nucleus and showed a distinct corona of staining close to the plasma membrane (Fig. 4b).
With co-expression, there was a clear overlap of the signals corresponding to the two proteins at the cell surface (Fig. 4c). The presence of HA-PPL did not prevent DAMGO mediated internalization of MOP-1-eYFP in HEK293 cells (Fig. 4c), but there was no evidence that DAMGO treatment altered the cellular distribution of HA-PPL. Indeed, the signals corresponding to the two polypeptides separated during exposure to the agonist (Fig. 4c). Unlike the receptor, HA-PPL did not move into an endocytic compartment after agonist treatment as there was no overlap of the signals in the MOP-1-eYFP-positive intracellular vesicles (Fig. 4c). The internalization of MOP-1 in intact HEK293 cells in response to DAMGO was also assessed by the removal from the cell surface of receptors available to be biotinylated (Fig. 5a). Biotinylated MOP-1 migrated in SDS-PAGE predominantly as an ϳ80-kDa species. Higher molecular mass species may represent dimeric and aggregated forms of the receptor. Co-expression with HA-PPL did not prevent the agonistinduced removal of MOP-1 from the cell surface (Fig. 5). Over short time periods, removal of MOP-1 from the cell surface did not reflect a down-regulation of the total cellular levels of the receptor (Fig.  5b).
Periplakin Interferes Selectively with MOP Activation of G Protein-Given that helix VIII of the rhodopsin-family receptors is believed to be a key functional contact site for the N (26,29) and/or C terminus (39) of G protein ␣ subunits, we explored whether PPL would interfere with G protein activation by MOP-1. After transient expression in HEK293 cells of a fusion protein in which the N terminus of a pertussis toxin-resistant (Cys 351 -Ile) variant of G i1 ␣ was linked in-frame to the C-terminal tail of MOP-1 (31), binding of [ 35 S]GTP␥S to the fusion protein was stimulated by DAMGO (Fig. 6). Co-expression of HA-PPL, which migrated in SDS-PAGE as a 200-kDa polypeptide (Fig. 6a) substantially reduced DAMGO stimulation of [ 35 S]GTP␥S binding produced by equal amounts of the fusion protein (Fig. 6). Although the yeast two-hybrid and GST pulldown studies demonstrated interactions between helix VIII of the MOP receptors and PPLC, this does not appear to be the only region of PPL that contributes to its effect on agonistactivation of G protein. An HA-tagged version of PPL lacking the C-terminal 208 amino acids was constructed and expressed (Fig. 7a). This was also able to inhibit DAMGO-stimulated binding of [ 35 S]GTP␥S to the MOP-1-G i1 ␣ fusion protein (Fig. 7b).
The effects of PPL were selective. Yeast two-hybrid assays using the C-terminal tails of the ␣ 2A -adrenoreceptor, the 5-hydroxytryptamine 5-HT 1A receptor, or the ␤ 2 -adrenoreceptor did not identify PPL as a potential interacting protein (data not shown). The ability of adrenaline to stimulate the binding of cells. An HA-tagged form of full-length PPL (HA-PPL) was introduced into HEK293 cells. Periplakin was excluded from the nucleus but displayed both a corona of immunostaining at the plasma membrane and a distinctly punctate intracellular distribution pattern. c, periplakin does not co-internalize with MOP-1-eYFP. In the absence of ligand a distinct overlap (yellow) of the distribution of HA-PPL (red), and MOP-1-eYFP (green) was observed at the plasma membrane in cells expressing both polypeptides (i). Stimulation of the cells with DAMGO for 10 (ii) or 30 (iii) min resulted in the resolution of MOP-1-eYFP from PPL. ␤ 2 -adrenoreceptor-G s ␣ fusion protein (Fig. 8) was unaffected by the expression of PPL. DISCUSSION Opioid receptors signal predominantly via members of the G i family of heterotrimeric G proteins (8). They also internalize and recycle to the cell surface after challenge with efficacious peptide and alkaloid ligands as part of the processes of receptor desensitization and resensitization. Recently, interactions of each of the DOP (24), the KOP (23), and the MOP receptors (25) with proteins that alter their intracellular sorting and recycling rates have been reported. By using the C-terminal tails of the human MOP-1 and MOP-1A receptors as bait in yeast two-hybrid screens with proteins generated from a human brain cDNA library we identified an interaction with PPL. Such interactions were confirmed in a range of pull-down studies. PPL is a member of the plakin family of cytolinker proteins (36). It has been most fully studied in keratinocytes and produces the scaffold on which the cornified envelope is formed (37). It is well suited to such a role as it is a large, 195-kDa, multi-domain protein known to interact with actin and intermediate filament proteins (40 -42). This role reflects specific contributions from the N-terminal domain and the ability of the rod segment of the polypeptide to allow both homodimerization and heterodimerization with other related plakins such as envoplakin. The C terminus appears to play roles in interactions with intermediate filaments and has recently been shown to be the region involved in interaction with protein kinase B (43). However, although detailed studies on the function of PPL have been largely restricted to skin, early cloning and mapping studies indicated it to be expressed in the brain (38). Quantitative reverse transcription-PCR confirmed this and in combination with direct immunoblotting studies demonstrated that significant levels of PPL mRNA and protein expression could be detected in a wide range of human and rat brain regions, including those that express the MOP receptors.
Mapping of sites of interaction between PPL and the human MOP isoforms indicated that it was a region within the last 208 amino acids of the C-terminal domain of PPL that interacted with the C-terminal tail region of MOP receptors. Fine mapping of the region of the MOP receptors responsible for this interaction defined a region of some 11 amino acids that are proximal to the plasma membrane. The vast majority of rhodopsin-like GPCRs contain one or more cysteine residues that can be post-translationally acylated within 10 -15 amino acids of the end of transmembrane helix VII (44). Because direct studies on rhodopsin demonstrate that these acyl chains are able to insert into the plasma membrane to provide a point of anchorage (45), this region became known as the fourth intracellular loop. With crystallization it became apparent that this section forms an eighth helix that runs parallel to the plasma membrane (26). Structural similarity between the rhodopsinlike receptors suggests that this will be a common feature (29) and, thus, that the site of interaction of the C terminus of PPL with the MOP isoforms is at this helix. Models of the interaction of GPCRs with G proteins indicate a likely interaction between helix VIII and the N terminus of the G protein ␣ subunit (28). Furthermore, although a key site of interaction of GPCRs is provided by the extreme C terminus of the G protein ␣ subunit, an important role for the N terminus of the G protein ␣ subunit has long been appreciated. We thus considered that the presence of PPL might disrupt agonist activation of G protein rather than internalization of the MOP receptors, as regulation of internalization generally involves the distal elements of the C-terminal tail. Indeed, internalization of MOP-1 was not altered by the presence of PPL. To explore the capability of PPL to interfere with MOP-1 activation of G protein, we took advantage of a fusion strategy in which a pertussis toxin-resistant variant of the ␣ subunit of G i1 ␣ was linked in-frame to the C-terminal tail of MOP-1 (31). A key rationale for this approach was that co-expression of isolated MOP-1 with PPL altered the levels of expression of the receptor. This, therefore, resulted in an alteration in the stoichiometry of MOP-1 to G protein. The fusion protein strategy ensures that the receptor to G protein ratio is the same no matter the absolute levels of expression (46). We have previously shown that the highly selective MOP receptor agonist DAMGO is able to stimulate binding of [ 35 S]GTP␥S to the G protein element of this construct (31). This was confirmed on addition of DAMGO to membranes of pertussis-toxin treated HEK293 cells expressing MOP-1-Cys 351 -Ile-G i1 ␣. This stimulation was reduced markedly when the fusion construct was co-expressed with PPL. However, although the yeast two-hybrid studies and the GST pull-down experiments demonstrated interaction between the helix VIII of these receptors and the C-terminal region of PPL, it does not appear to be the only interaction between these FIG. 7. The effect of periplakin is not produced only by the C-terminal region. a, a form of HA-tagged PPL lacking the C-terminal 208 amino acids was generated. Full-length HA-PPL or ⌬-208 HA-PPL was expressed in HEK293 cells along with the MOP-1-G i1 ␣ fusion protein, and cell lysates were immunoblotted with an anti-HA antibody to detect expression. b, experiments akin to those of Fig. 6 were performed in membranes expressing MOP-1-G i1 ␣ alone (open bars) or co-expressing MOP-1-G i1 ␣ with either full-length (gray bars) or ⌬-208 (black bars) HA-PPL. *, significantly different (p Ͻ 0.05) from MOP-1-G i1 ␣ alone. proteins. When we generated and expressed a form of PPL lacking the C-terminal 208 amino acids, this was also able to interfere with agonist-stimulated binding of [ 35 S]GTP␥S to the MOP-1-Cys 351 -Ile-G i1 ␣ fusion protein. Although yeast two-hybrid analyses are extremely useful in demonstrating interactions between linear peptide fragments from two proteins, they are not appropriate to examine complex interactions requiring sequences from more than one segment of a protein. We thus wished to confirm that the ability of PPL to interfere with agonist-stimulated binding of [ 35 S]GTP␥S to MOP-1-Cys 351 -Ile-G i1 ␣ was selective. When co-expressed with two other receptor-G protein fusions, including one that also contained Cys 351 -Ile-G i1 ␣ as the G protein, PPL was unable to modify agonist activation of the G proteins. Parallel yeast two-hybrid assays failed to demonstrate interactions of periplakin with the C-terminal tails of these GPCRs.
These results demonstrate interactions between helix VIII of the human MOP receptors and PPL and indicate that selective interactions between these polypeptides limit agonist activation of G protein. PPL is the first opioid receptor-interacting protein described that alters the effectiveness of G protein activation rather than the intracellular sorting and recycling of the receptor. The distribution of MOP and PPL in rat brain overlapped but was not identical. The current data suggest that MOP receptor signaling may be less effective in neurons that co-express the receptor and PPL than in those that do not. Future studies will test this hypothesis.