Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Chinese hamster ovary (CHO) and mouse fibroblast (B82) cells are frequently used as host cells for the expression of G protein-coupled receptor cDNAs and the characterization of signal transduction pathways mediated by the expressed receptor proteins. The use of recombinant mammalian cell lines for screening compounds with potential as agonists or antagonists has many advantages over animal tissues. In addition to possible ethical concerns limiting the use of animal tissue, the recombinant cell lines provide an unlimited tissue source with a homogenous, well-defined receptor population. Also cell lines with different levels of spare receptors can easily be constructed. There are a number of examples, however, when the presence or absence of a component of the signal transduction cascade in the host cell produces unexpected results (Kenakin, 1996). For example, physiologically irrelevant signaling might be detected or the relevant signal transduction might be missing if the G protein pool present in the host cells is different from the G protein pool in the physiological environment of the receptor.

The activation of G proteins is the first step in the signal transduction cascade mediated by G protein-coupled receptors. The affinity of a drug in receptor binding assays and its potency and intrinsic activity in guanosine-5'-O-(3-[³⁵S]thio)triphosphate ([³⁵S]GTPS) binding assays can be used for calculations of drug efficacy (Burkey et al., 1998). CHO and B82 cells have been used in our laboratory for the expression of a number of receptors coupled to pertussis toxin (Ptx)-sensitive G proteins, including the M2 and M4 muscarinic (Kashihara et al., 1992; Kovacs et al., 1998) human - (Malatynska et al., 1995) and µ- (Hosohata et al., 1998) opioid, and CB1 cannabinoid (Landsman et al., 1998) receptors. The transfected cell lines were used to characterize signal transduction cascades mediated by these receptors and in some cases to calculate agonist efficacy values by measuring agonist stimulated [³⁵S]GTPS binding. The measured [³⁵S]GTPS binding, however, reflects the sum of activation of different G proteins. Numerous studies have demonstrated that different receptors can activate different G proteins depending on the G protein pool present in the examined tissue (for review, see Kenakin, 1996). To calculate a physiologically relevant efficacy for a potential drug it would be of interest to know if, and under what cellular conditions, the results obtained from a recombinant expression system are relevant for the pharmacological target tissues. It is very important, therefore, to characterize the cellular G protein pool in both the host cells and native tissues.

The G protein pool in B82 cells to our knowledge has not been examined. Agonist activation of the G protein -subunits has been studied by immunodetection methods in CHO cells (Dell'Acqua et al., 1993; Prather et al., 1994, 1995; Chakrabarti et al., 1995; Reisine et al., 1996). The immunological methods have, however, an inherent limitation, e.g., novel or unexpected proteins may not be detected because of the choice of specific antibodies.

The present study was designed to characterize the complete Ptx-sensitive G protein -subunit pool in CHO and B82 cells with a reverse transcription-polymerase chain reaction (RT-PCR) method with degenerate primers designed to highly conserved regions in the G_i/o family. Unexpectedly, we have isolated several clones with a sequence highly homologous to the rodent retinal rod transducin (G_t1) in CHO cells, and a cDNA fragment 100% homologous to the appropriate fragment of the mouse retinal cone transducin (G_t2) in the B82 cells. The presence of the transducin -subunit proteins in CHO and B82 cells was confirmed by immunoprecipitation with G_t1- and G_t2-specific antibodies, respectively. A Ptx-insensitive rod transducin -subunit mutant, where cysteine 347 was mutated to serine (t1C347S) was transfected into CHO cells expressing the human -opioid receptor (hDOR/CHO) to test if transducin can functionally couple to heterologously expressed G protein-coupled receptors in these cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

mRNA Isolation. mRNA was isolated from CHO and B82 cells (10⁶ cells) with the PolyATract mRNA isolation kit (Clontech, Palo Alto, CA) according to the manufacturer's instructions. First strand cDNA was synthesized with random primers and Superscript reverse transcriptase from the cDNA cycle kit (Life Technologies, Inc., Gaithersburg, MD).

Degenerate Primers. Degenerate primers to highly conserved regions of the -subunits of the G_i/o family were designed and synthesized (Integrated DNA Technologies, Coralville, IA). The sequence of the primers was upper primer (to the amino acid sequence STIVKQM): 5' >AG(C/T)AC(C/T/A)AT(C/T)GT(G/C/A)AA(G/A)CAGAT > 3'; and lower primer to the amino acid sequence KKWIHCF: 5' > (G/A)AA(G/A)CAGTG(G/A)ATCCA(C/T)TT(C/T)TT > 3'.

RT-PCR. The reactions were performed with 10 ng of first strand cDNA mixture as a template. The optimal Mg²⁺ concentration was 2 mM. The reaction mixture was denatured (5 min at 94°C). After addition of 2.5 U Taq DNA polymerase, 35 amplification cycles were performed with the following conditions: 95°C, 1 min (denaturation), 55°C, 1 min (annealing), and 72°C, 1 min (extension). The PCR products of the expected size (0.5 kilobase) were isolated from 2% agarose gels. After electroelution and precipitation, the fragment mixtures were ligated into the pCR 2.1 vector (TA cloning kit; Invitrogen, San Diego, CA) and transformed into OneShot competent cells. Randomly selected white clones were sequenced with the dideoxy chain termination method (Sequenase version 2.0 sequencing kit; Amersham, Arlington Heights, IL) with the promoter regions of the vector. The sequences were identified with the BLAST search program (GeneBank, National Center for Biotechnology Information, Rockville Pike, MD) and also aligned with the published G protein -subunit sequences with the DNAsis program.

Immunoprecipitation. The cells (10⁶ CHO or B82 cells/plate) were washed with PBS, and scraped from the culture plate into 1 ml of homogenization buffer [50 mM Tris, 250 mM sucrose, 1 mM EDTA, 5 mM MgCl₂, 1 mM dithiothreitol, 50 mM NaF, and 10 mM Na-pyrophosphate supplemented with 10 µl/ml protease inhibitor cocktail (Sigma Chemical Co., St Louis, MO) immediately before use]. The plates were washed with 2× 0.5 ml of homogenization buffer. Rat whole-brain membranes were isolated as previously described (Yamamura et al., 1991) and the final pellet taken up in 1 ml of homogenization buffer. Aliquots of the homogenate were used in the further steps to give the same amount (40 µg) of total protein as the CHO and B82 lysates. The homogenates from rat brain and from CHO or B82 cells were centrifuged at 14,000 rpm for 20 min and the pellets resuspended in 1 ml of RIPA buffer [50 mM Tris-HCl, 150 mM NaCl, 0.1% Igepal, 0.5% Triton X-100, 0.2% digitonin, 5 mM EDTA, 10 mM NaF, 10 mM -glycerol-phosphate with 10 µl/ml protease inhibitor cocktail (Sigma Chemical Co.) added immediately before use]. The solution was incubated on ice for 3 h and centrifuged at 14,000 rpm for 20 min. The lysate was precleared by incubation in the presence of 1 µg of preimmune rabbit IgG and 10 µl of protein A-agarose. The proteins that nonspecifically bound to the preimmune rabbit IgG/protein A were removed by cenrifugation (3000 rpm; 5 min). The precleared lysates were incubated overnight with 10 µl of the G_t1 or the G_t2 antibodies (Santa Cruz Biotechnologies, Santa Cruz, CA). Protein A-agarose (10 µl) bead slurry (Santa Cruz Biotechnologies) was added and the mixtures incubated on ice with gentle rocking for 3 h, centrifuged (3,000 rpm, 5 min) and washed three times with 10-min incubations in RIPA wash buffer (same as solubilization buffer except that the detergent concentrations were reduced to 0.075% Triton X-100, 0.05% Igepal, and 0.1% digitonin) in the presence of protease inhibitors. The antibody-transducin complexes were eluted from the final pellet by incubating with 10 µl of glycine-Cl buffer, pH = 2.3. The mixture was neutralized with 5 µl of neutralization buffer (0.5 M phosphate buffer; pH = 7.7) and boiled with 15 µl of 2× SDS-polyacrylamide gel electrophoresis (PAGE) sample buffer for 5 min. The immunoprecipitate was resolved on 10% SDS-PAGE and the protein bands were detected by silver staining with the Silver Stain Plus kit according to the manufacturer (Bio-Rad, Hercules, CA) instructions. The gel was dried in a GelAir dryer (Bio-Rad) and scanned with an Arcus II scanning densitometer with the Documax OneDScan software.

Site-Directed Mutagenesis and Stable Transfection. The QuickChange Site-Directed mutagenesis kit (Stratagene, Inc., La Jolla, CA) was used to introduce a Cys-to-Ser point mutation at position 347 (C347S) into the bovine G_t1 cDNA (American Type Culture Collection, Rockville, MD) according to the manufacturer's instructions. The sequence of the mutagenic primers was 5' > CTC AAA GAC AGC GGG CTC TTC > 3' (sense) and 5' >GAA GAG CCC GCT GTC TTT GAG > 3' (antisense), the mutant nucleotide is in bold. The mutation was verified by sequencing with the dideoxy chain termination method (Sequenase version 2.0 sequencing kit; Amersham). The mutant cDNA was ligated into the HindIII and SalI sites of the LK 444 (pHAPr-Neo, a gift from L. Kedes, Stanford, CA) mammalian expression vector. The plasmid (5 µg) was transfected into a previously described hygromycin-resistant hDOR/CHO cell line (Malatynska et al., 1995) with the DOTAP mammalian transfection kit (Boehringer Mannheim, Indianapolis, IN). Double transfectant clones (hDOR/t1C347S/CHO) were selected in Ham's medium containing 400 µg/ml hygromycin and 400 µg/ml G418. The clones were screened for G_t1 overexpression by Western blot with the G_t1 antibody (Santa Cruz Biotechnologies) according to the manufacturer's instructions. The immunocomplexes were detected with the Immun-Star chemiluminescent detection kit (Bio-Rad) and quantitated by scanning densitometry.

[³⁵S]GTPS Binding Assay. A method described by Wieland et al. (1995) was used with minor modifications to measure -opioid agonist (SNC 80)-stimulated [³⁵S]GTPS binding in permeabilized hDOR/CHO and hDOR/t1C347S/CHO#18 cells after Ptx pretreatment. hDOR/CHO or hDOR/t1C347S/CHO#18 cells were grown in Ham's physiological medium complemented with 10% fetal bovine serum in the presence of 400 µg/ml G418 and 400 µg/ml hygromycin at 37°C in humidified CO₂ atmosphere. Forty-eight hours before the assay, cells were plated in 24-well culture plates to give a cell density of ~200,000 cells/well on the day of the assay. The growth medium was aspirated, the cells washed twice with Iscove's modified Dulbecco's medium (IMDM) and incubated in IMDM medium containing 200 ng/ml Ptx for 4 h, where indicated. After the pretreatment the cells were washed twice with IMDM and incubated with fresh IMDM at 37°C for 10 min. The medium was replaced with 1 ml permeabilization buffer (25 mM Tris-HCl, 150 mM NaCl, 2.5 mM MgCl₂, 1 mM EDTA, and 5 µM digitonin). After 15 min of permeabilization at 37°C, the buffer was replaced with assay buffer (25 mM Tris-HCl, 150 mM NaCl, 2.5 mM MgCl₂, 1 mM EDTA, 50 µM GDP, and 5 µM digitonin) containing 0.5 nM [³⁵S]GTPS and (+)-4-[(R)--((2S,2R)-4-allyl-2,5-dimethyl-1-piperazinyl)-3-methoxybenzyl]-N,N-diethylbenzamide (SNC 80) (0.3-10,000 nM) with or without 1 µM naltrindole (NTI), in a 1-ml sample volume. After a 30-min incubation at 37°C the solution was removed and the cells washed with 1 ml of ice-cold wash buffer (25 mM Tris-HCl, 120 mM NaCl). The cells were solubilized by incubation in 0.5 ml 10% SDS overnight and transferred into EcoLite liquid scintillation cocktail. The radioactivity was measured in a Beckman LS 6000SE liquid scintillation spectrophotometer.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

A low stringency (2 mM Mg²⁺; annealing temperature 55°C) RT-PCR was used to identify Ptx-sensitive G proteins in CHO and B82 cells. First strand cDNA was synthesized from CHO and B82 cell mRNA and used as PCR template. The degenerate primers were derived from highly conserved regions in the -subunits of the G_i/o family. The primers were not expected to amplify -subunits of the G _s and G₁₂ families. The PCR products (0.5 kilobase) were subcloned, randomly selected, and sequenced. As Table 1 shows, we have detected the presence of mRNA for G_i2, G_i3, and G_o in both cell lines. The sequences of G_oA and G_oB are identical in the amplified region; therefore, in the present experiments, we could not discriminate between the two variants. The degenerate primers could possibly have amplified members of the G_q family, but no G_q clones were obtained. No G_i1 and G_z mRNAs were detected in our experiments. The identified sequences show high (>85%) homology to the appropriate rodent G-subunits, the differences are presumably due to species (Chinese hamster versus rat or mouse) variation. No novel G-subunit mRNAs were found.

View this table: [in this window] [in a new window]   TABLE 1 Identification of Ptx-sensitive G-subunit cDNA clones obtained by RT-PCR from CHO and B82 cells The mRNA from CHO and B82 cells was amplified in an RT-PCR with degenerate primers derived from conserved regions of the Ptx-sensitive Gi/o family. The PCR fragments were subcloned, randomly selected, and sequenced. The table shows the number of clones identified by sequence analysis. The sequences of GoA and GoB are identical in the amplified region.

Unexpectedly, we have isolated three clones with a high (91%) sequence homology in the amplified region to the mouse retinal rod photoreceptor G protein G_t1 (Raport et al., 1989) in CHO cells. Figure 1 shows the alignment of the nucleotide sequence between the appropriate fragment of the mouse G_t1 and the transducin homolog isolated from the CHO cells. The clone showed lower homology (71%) to the mouse cone transducin G_t2 (Zigman et al., 1994). The deduced amino acid sequence of the CHO transducin homolog is 97.4% identical with that of the mouse G_t1, with three amino acid substitutions: ¹³⁷AspGlu; ¹⁵³SerLeu, and ¹⁷⁸ThrAla (the numbering corresponds to the deduced amino acid sequence of the mouse rod transducin). We also have isolated a clone in B82 cells showing 100% homology to the mouse cone transducin (Zigman et al., 1994) in the investigated region (amino acid 53 to 208 in mouse G_t2).

View larger version (67K): [in this window] [in a new window]   Fig. 1.   Nucleotide sequence alignment of the transducin cDNA fragment obtained from CHO cells with the appropriate fragment (from nucleotide 148 to nucleotide 465) of the mouse rod transducin cDNA. The  above the sequences marks the exon-intron boundaries in the mouse gene. (GeneBank accession no. AF157566).

The genes of the G_i/o family contain four introns in the amplified region (Itoh et al., 1988). No intron sequences were detected, however, in any of the amplified G sequences, showing that the isolated mRNA did not contain genomic DNA contamination and that the obtained transducin homologs are not of chromosomal origin. No PCR errors were detected in the B82 cell cDNA clones (cells of mouse origin) compared with the mouse G protein sequences deposited in GeneBank.

To verify the translation of the mRNA and the presence of the transducin -subunit proteins in the CHO and B82 cells, immunoprecipitation experiments were performed with antibodies (Santa Cruz Biotechnologies) raised against unique sequences in the human rod and cone transducin. The antibodies do not cross-react with any other G-subunits, but are expected to cross-react with the appropriate transducin homologs from different rodent species. The rod transducin (G_t1) antibody reacted with a 40-kDa protein in the CHO cell lysate, whereas the cone transducin (G_t2) antibody immunoprecipitated a 41-kDa protein from the B82 cell lysate (Fig. 2). A 41-kDa protein also was immunoprecipitated from the B82 cell lysate with the rod transducin antibody, presumably showing some cross-reactivity of the G_t1 antibody. The G_t2 antibody did not cross-react with the G_t1 present in the CHO cells. No immunoprecipitation was detected from rat brain membranes.

View larger version (91K): [in this window] [in a new window]   Fig. 2.   Immunoprecipitation analysis of transducin expression in CHO and B82 cells. CHO (lanes 1 and 4) and B82 (lanes 2 and 5) cell lysates were incubated with rabbit polyclonal antibodies against Gt1 (lanes 1-3) or Gt2 (lanes 3-6) and immunoprecipitated with protein A-agarose. The immunoprecipitates were resolved on a 10% SDS-PAGE and the gel was silver stained. The antibodies did not detect specific bands from the control rat brain membranes (lanes 3 and 6).

As CHO cells are widely used as host cells in numerous laboratories for the expression of G protein-coupled receptors, we tested whether heterologously expressed receptors can couple to transducin in these cells. We have previously (Malatynska et al., 1995) established a CHO cell line stably expressing hDOR. [³⁵S]GTPS binding to hDOR cell membranes stimulated by -opioid receptor agonists was used to calculate ligand efficacies at the hDOR (Quock et al., 1997). To test whether a fraction of agonist-stimulated [³⁵S]GTPS binding originates from the coupling of the hDOR to transducin in the CHO cells, we have stably transfected a Ptx-insensitive (C347S) mutant of G_t1 into the hDOR/CHO cells. SNC 80 stimulated [³⁵S]GTPS binding in digitonin (5 µM)-permeabilized CHO cells expressing the hDOR alone 231 ± 19% above basal levels with an EC₅₀ value of 18.6 ± 6.5 nM (Fig. 3). Ptx pretreatment (200 ng; 4 h) completely abolished agonist-stimulated [³⁵S]GTPS binding in the permeabilized hDOR/CHO cells (Fig. 3). However, in the permeabilized double-transfectant hDOR/t1C347S/CHO cells, SNC 80 stimulated [³⁵S]GTPS binding after Ptx (200 ng/ml; 4 h) pretreatment 193 ± 36% above basal levels (Fig. 4) with an EC₅₀ value of 54 ± 32 nM. The stimulation was antagonized by the -opioid receptor selective antagonist NTI (Fig. 4).

View larger version (13K): [in this window] [in a new window]   Fig. 3.   The effect of Ptx treatment on -opioid receptor agonist (SNC 80)-stimulated [35S]GTPS binding in permeabilized hDOR/CHO cells. hDOR/CHO cells (average cell count of 177,000 cells/well) were plated in 24-well culture plates and treated () with 200 ng of Ptx at 37°C for 4 h, or left untreated (). The cells were washed and permeabilized in buffer containing 5 µM digitonin and incubated with different concentrations of SNC 80 in the presence of 0.5 nM [35S]GTPS for 30 min. The average value of basal [35S]GTPS binding in the absence of agonist was 2480 dpm for the untreated and 2100 dpm for the Ptx-treated cells. SNC 80 stimulated [35S]GTPS binding 231 ± 19% above basal levels with an EC50 value of 18.6 ± 6.5 nM. The data are means ± S.E. from four experiments performed in triplicate.

View larger version (13K): [in this window] [in a new window]   Fig. 4.   -Opioid receptor agonist (SNC 80)-stimulated [35S]GTPS binding in permeabilized hDOR/t1C347S/CHO cells after Ptx pretreatment. CHO cells expressing the hDOR and a Ptx-insensitive mutant of the bovine rod transducin (hDOR/t1C347S/CHO) were plated in 24-well culture plates (average cell count of 295,000 cells/well) and treated with 200 ng of Ptx at 37°C for 4 h. The cells were washed and permeabilized in permeabilization buffer containing 5 µM digitonin and incubated with 0.5 nM [35S]GTPS for 30 min and different concentrations of SNC 80 in the absence () or presence () of 1 µM NTI. The average value of basal [35S]GTPS binding in the absence of agonist was 3350 dpm. SNC 80 stimulated [35S]GTPS binding 193 ± 36% above basal levels with an EC50 value of 54 ± 32 nM. The data are means ± S.E. from four experiments performed in triplicate.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study, we used a RT-PCR method with degenerate primers to identify the Ptx-sensitive G-pool in CHO and B82 cells. We have confirmed the presence of G_i2, G_i3, and G_o in the CHO cells. No clones with G_i1 or G_z sequences have been identified, although in one study (Reisine et al., 1996) the precoupling of the -opioid receptor (in the absence of an agonist) to G_i1 has been detected by an immunoprecipitation method in CHO cell membranes. Reasons for this discrepancy include the limited specificity of the antibody used in the previous study or an unexpected bias in the sequence of the degenerate primers used in our study. The degenerate primers were designed based on the cloned -subunit sequences from different mammalian species. No Chinese hamster -subunit sequences are, however, deposited in the GeneBank. It is possible therefore, that unexpected species differences in the Chinese hamster G_i1 sequence prevented the annealing of the degenerate primers to the cDNA of G_i1. The regions selected for the design of the degenerate primers, however, are very highly conserved among mammalian species, and also no other studies have detected the presence of G_i1 in CHO cells. We have detected a similar Ptx-sensitive G protein population (G_i2, G_i3, and G_o) in the B82 cells. As in the CHO cell line, no G_i1 or G_z clones were found.

Unexpectedly, in our study we also detected the presence of the mRNA for another Ptx-sensitive G protein -subunit, rod transducin, in CHO cells. Based on nucleotide and deduced amino acid sequence alignment to the published mouse retinal G_t1 sequence (Raport et al., 1989), our clone is presumably a species (Chinese hamster) homolog of the mouse retinal G_t1. However, B82 cells contained the mRNA for cone transducin G_t2. The sequence of the cloned PCR fragment was identical with the published mouse cone transducin (Zigman et al., 1994) sequence in the appropriate region. No intron sequences were found in any of the -subunit sequences, showing that the identified -subunit fragments were not of chromosomal origin. Our immunoprecipitation experiments confirm the presence of G_t1 or G_t2 protein in the CHO and B82 cells, respectively.

Transducin was previously thought to be expressed specifically in the retina and the pineal gland. Transducin homologs have recently been identified in peripheral and CNS cells (Zigman et al., 1994; Yamaguchi et al., 1997). These data, together with our results, indicate that the visual system G proteins are present in nonretinal cells more frequently than previously reported. The functional coupling of the cone transducin (G_t2) to a nonopsin receptor (dopamine D4 receptor) in a mouse mesenkephalic cell line (MN9D) has been recently shown (Yamaguchi et al., 1997).

The physiological role of transducin in nonphotoreceptor cells is presently unclear. In the vertebrate retina, transducin modulates the activity of cGMP phosphodiesterase, leading to decreased cGMP levels and the closure of cGMP-gated channels (Baylor, 1996). Activation of rhodopsin by light also leads to the reduction of cAMP formation in the retina (Weiss et al., 1995). A rod transducin homolog gustducin (G_,gust), however, modulates cAMP-phosphodiesterase and cAMP-gated channels in the gustatory signal transduction pathway (Koleshnikov and Margolskee, 1995). Several amino acid substitutions are present, however, in the vicinity of the putative loop II region of the CHO transducin fragment compared with the rodent retinal rod transducin. This domain has been involved in the effector recognition of the G protein -subunits (Medina et al., 1996). These substitutions may modify the effector coupling preference of the CHO transducin.

Agonist binding to the - (Prather et al., 1994), µ- (Chakrabarti et al., 1995), and  (Prather et al., 1995)-opioid receptors has been shown to activate G_i2, G_i3, G_oB, and an additional G protein -subunit designated G? in CHO cells. To test the hypothesis that the -opioid receptors can couple to transducin in CHO cells, we have transfected a Ptx-insensitive (C347S) mutant of G_t1 into CHO cells expressing the hDOR. In membrane preparations from the double-transfected cell line, no -opioid agonist-stimulated [³⁵S]GTPS binding was detected after Ptx treatment.

Transducin, however, is weakly anchored to the membrane and can easily be removed during the membrane preparation (Chabre and Deterre, 1990) because transducin is not S-acylated (palmitoylated) in the N-terminal domain (Duncan and Gilman, 1996). Membrane anchoring for transducin is provided only by myristoylation at the Gly² residue, common for the G_i/o family (Wedegaertner et al., 1995). We hypothesized that although agonist-induced activation of transducin is not observed in CHO cell membrane preparations, transducin might contribute to receptor-mediated signaling in intact cells. Indeed, in permeabilized whole-CHO cells, coexpressing the -opioid receptor and the C347S mutant rod transducin, SNC 80 stimulated [³⁵S]GTPS binding by 193 ± 36% above basal levels after Ptx pretreatment. However, Ptx treatment completely abolished agonist-stimulated [³⁵S]GTPS binding in permeabilized CHO cells expressing the hDOR alone. Based on these results, the presence of transducin in CHO cells is not likely to contribute to [³⁵S]GTPS binding with membrane preparations. Caution is needed, however, when interpreting functional signal transduction data obtained from intact CHO cells.

Additional studies are necessary to determine the physiological relevance of transducin activation by opioid receptors. The pupillary effects of opioids (miosis in humans, mydriasis in some other species) are well known. Although it is generally held that the pupillary effects of opioids are mediated via the CNS, mainly at the Edinger-Westphal nucleus, there is evidence for peripheral mechanisms as well (Murray et al., 1983). The density of opioid-binding sites in rat retina is comparable to that of rat brain. The presence of - (131 fmol/mg), µ- (13 fmol/mg), and  (88 fmol/mg)-receptors in chick retina has been shown (Slaughter et al., 1985). However, the modulation of cAMP phosphodiesterase activity by opioid receptors in NG108-15 cells (Law and Loh, 1993) has previously been shown. It would be interesting to show what role, if any, transducin might play in these signaling mechanisms.

In summary, we have shown the presence of mRNA for G_i2, G_i3, and G_o in both CHO and B82 cells. G_i1 and G_z mRNAs were not detected in either cell line. We also have identified a rod transducin (G_t1) homolog in CHO cells and a mouse retinal cone transducin (G_t2) mRNA in B82 cells. The presence of the transducin -subunit proteins in CHO and B82 cells was confirmed by immunoprecipitation with G_t1- and G_t2-specific antibodies, respectively. We also have shown that the hDORs expressed in CHO cells stimulate guanine nucleotide exchange in a Ptx-insensitive mutant of the rod transducin after Ptx treatment in permeabilized whole-cell preparations.