ADP-Ribosylation Factor-Dependent Phospholipase D Activation by VPAC Receptors and a PAC₁ Receptor Splice Variant

Experimental Procedures

Materials

All tissue culture media, including animal serum, geneticin, penicillin, and streptomycin were obtained from Life Technologies, Irvine, UK. Radiochemicals; [¹²⁵NaI],¹²⁵I-PACAP-27, myo-[2-³H]inositol, and [9,10-³H]palmitic acid were obtained from PerkinElmer Life Science Products, Dreiech, Germany. All peptides were from Novabiochem, Nottingham, UK, and all biochemicals were from Sigma, Poole, UK, unless otherwise stated. Any reagents with relatively lower aqueous solubility (such as brefeldin A or U 73122) were added to cell signaling assays from solutions in dimethylformamide. Corresponding additions of vehicle were made to control and test wells and were limited to a concentration of 0.3% (at which level no effects could be detected on any of the responses).

Generation of Stable Chinese Hamster Ovary (CHO) Cell Lines

cDNAs encoding the rat VPAC₁, VPAC₂, PAC_1-null, and PAC_1-hop1 receptors were introduced into the expression vector pcDNA1, containing the neomycin resistance gene (Invitrogen BV, Groningen, The Netherlands). CHO cells were transfected with the receptor plasmids using lipofectamine (Life Technologies). Forty-eight hours after transfection, geneticin (500 μg/ml) was added to cells grown in Ham's F-12 nutrient media with 10% NCS, 100 U/ml penicillin, and 100 μg/ml streptomycin to select for cells expressing constructs. Clonal cells were picked and growth was continued for 1 month in geneticin-containing media. Clonal lines expressing the different receptors were selected for their ability to stimulate cAMP production in response to VIP, PACAP, and other VIP-like peptides. The expression of the mRNA for the various receptors was also confirmed by Northern analysis (data not shown).

Construction of Chimeric Receptors

Chimeric receptors were made by replacement of the i3 domain of the rat VPAC₂ receptor with the i3 domains from either the short form or the hop-1 form of the rat PAC₁ receptor. Exchange sites were within transmembrane domain (tm) 5 and tm7 (Fig. 4). This was achieved using cDNAs encoding the rat VPAC₂ receptor (R4, pBluescript) and the null (R7b, pBluescript) and hop1 (R7/9.1, pBluescript) splice variants of the rat PAC₁receptor. The first domain exchange was made by using a conserved restriction (HincII) site in the region of the cDNAs encoding the fifth transmembrane region of the VPAC₂ and PAC₁ receptors. After digestion with HincII, the appropriate cDNA fragments were gel purified and then ligated with T4 DNA ligase (Promega, Southampton, UK). These were inserted into pBluescript for selection of appropriate clones by sequence analysis of the domain exchange region. The second domain exchange within tm7 was made by overlap extension polymerase chain reaction (PCR) mutagenesis (Huang et al., 1995). The reaction was heated to 95°C for 5 min and then maintained at 80°C while adding 2.5 U of Pfu (Pyrococcus furiosus) polymerase (Stratagene, Amsterdam, The Netherlands), after which the reaction was put through 30 cycles with denaturing at 94°C (1 min), annealing at 57°C (1 min), and extension at 72°C (3 min). After the first round of PCR, 10-μl samples were analyzed by electrophoresis. The remaining PCR reactions were purified by extracting with the Wizard cDNA purification system (Promega), and then in the second round of PCR amplification 1-μl volumes of each appropriate extract were mixed and amplified using the flanking pBluescript primers under the same conditions as the first round of amplification. These were ligated into pBluescript for selection of appropriate clones by sequence analysis and then inserted into the expression vector pcDNA1 for functional expression in COS 7 cells.

Cell Culture and Transient Transfection of Receptor cDNAs

CHO cell lines stably expressing the VPAC₁, VPAC₂, PAC_1-null, and PAC_1-hop1receptors were grown in 80-cm² flasks in Ham's modified F-12 medium containing 10% NCS, 300 μg/ml geneticin, 100 U/ml penicillin, and 100 μg/ml streptomycin, in a humidified atmosphere of 5% CO₂ and 95% O₂ at 37°C. GH₃ and αT3-1 cells were cultured as described previously (Johnson et al., 2000). Confluent cultures were trypsinized and seeded into 12-well cell culture plates for assay of PLC or PLD activity, or 24-well plates for assay of cAMP production. COS 7 cells were grown in 175-cm² flasks in Dulbecco's modified Eagle's medium containing 10% NCS, 100 U/ml penicillin, and 100 μg/ml streptomycin, in a humidified atmosphere of 5% CO₂ and 95% O₂ at 37°C. The cDNAs for the chimeric VPAC₂/PAC_1-null and VPAC₂/PAC_1-hop1(VP/4/7b/2.1c and VP/4/7/2.1c, pcDNA1) receptors were transfected into COS 7 cells using 30 μg of cDNA/6 × 10⁶cells and DEAE dextran (Promega) or FuGENE 6 (Roche Diagnostics Ltd., Lewes, UK) as described previously (Morrow et al., 1993;MacKenzie et al., 2001).

Whole-Cell Ligand Binding

The density and affinity of ligand binding sites in the VPAC₁ VPAC₂, PAC_1-null, and PAC_1-hop1receptor-expressing cell lines were determined by nonlinear curve-fitting analysis of homologous displacement curves (Swillens, 1992). This method allows calculation of the number and affinity of binding sites in circumstances (such as with ligands iodinated in-house) where the precise ligand specific activity is not known and therefore Scatchard type analysis is not possible. Experiments were carried out at 37°C using intact cells in 12-well plates. This enabled assessment of both cell-surface association and internalization of ligand, reflecting the cellular disposition of the receptors under near-physiological conditions. Cells were incubated in 0.5 ml of Medium 199 with 0.2% BSA, 30 μg/ml bacitracin, and 10 μg/ml aprotinin, plus ¹²⁵I-helodermin (for VPAC₁/VPAC₂ receptors) or¹²⁵I-PACAP-27 (for PAC₁receptors), using 20,000 to 50,000 cpm/well.¹²⁵I-Helodermin (approximately 770 Ci/mmol) was prepared by iodination using chloramine-T and purified by high-performance liquid chromatography according to methods described previously (Ogier et al., 1987). Increasing concentrations of unlabeled helodermin/PACAP-27 were also present as required. Nonspecific binding was defined with 300 nM unlabeled helodermin or PACAP-27, respectively. The plates were incubated at 37°C for 20 min unless otherwise indicated. The assay was terminated by aspiration of the medium and the cells were washed three times with 0.5 ml of ice-cold EBSS containing 0.1% BSA. Externally bound ligand was dissociated by a 5-min wash with 0.5 ml of an ice-cold acid strip solution (0.2 M acetic acid/0.5 M NaCl) (Slice et al., 1994). Internalized ligand was determined by solubilization of the cells after the acid strip wash using 1% Triton X-100 in 0.1 M NaOH. Protein content was determined using the bicinchoninic acid system (Pierce, Rockford, IL). After incubation of¹²⁵I-PACAP-27 with PAC_1-null receptor-containing CHO cells, the integrity of the ligand was assessed by reverse-phase chromatography on C18-silica (Waters Ltd., Watford, UK) using an H₂O/methanol gradient, containing 0.2% trifluoroacetic acid. After 30-min incubation at 37°C, 80 to 83% of the ligand in the extracellular medium and that released from the cells by hypotonic lysis in 10 mM formic acid still eluted as authentic¹²⁵I-PACAP-27.

Determination of cAMP Production

CHO/COS 7 cells expressing the VPAC₁, VPAC₂, PAC_1-null, and PAC_1-hop1 receptors were preincubated for 10 min with the phosphodiesterase inhibitor, 3-isobutyl-1-methylxanthine (500 μM) and then stimulated with agonist (for 15 min). Intracellular cAMP levels were measured using a radioimmunoassay technique described previously (Morrow et al., 1993; Lutz et al., 1999).

Determination of [³H]Inositol Phosphate Production

After 16-h labeling with 1 μCi/ml myo-[2-³H]inositol (20 Ci/mmol), PLC activity in response to 30-min stimulation with agonist was monitored as formation of [³H]inositol phosphate ([³H]InsP) in the presence of 10 mM LiCl, as described previously (MacKenzie et al., 1996, 2001).

Measurement of [³H]Phosphatidylbutanol ([³H]PtdBut) Production

Cells in 12-well plates that had reached 80 to 100% confluence were placed in serum-free medium and labeled by incubation with 10 μCi/ml/well [9,10-³H]palmitate (40 Ci/mmol) for 18 h before assay. PLD activity was monitored as the production of [³H]PtdBut when cells were stimulated in the presence of 30 mM butan-1-ol (Mitchell et al., 1998). Before stimulation, cells were washed twice with MEM containing fatty acid-free BSA (1%), before replacement with minimal essential medium containing 0.5% BSA. The assay (30 min) was started with addition of agonist and terminated by aspiration of the medium and addition of 0.5 ml of ice-cold methanol. Cells were homogenized and samples transferred to 2-ml glass vials, before chloroform and H₂O were added to give a ratio of methanol/chloroform/H₂O of 1:1:0.8. Samples were vortexed and left for 15 min and then spun for 8 min in a low-speed centrifuge to separate the aqueous and organic layers. The upper aqueous layer was removed and an aliquot of the lower organic phase was evaporated under vacuum at 30°C in a centrifugal evaporator (Jouan, Nottingham, UK). Lipids were redissolved in 50 μl of chloroform/methanol (19:1) and separated by thin-layer chromatography on LK5D silica gel plates (Whatman, Maidstone, UK) using the upper phase of a mixture of 110 ml of ethyl acetate, 50 ml of 2,2,4-trimethylpentane, 20 ml of acetic acid, and 100 ml of water. The region of the thin-layer chromatography plate corresponding to [³H]PtdBut, as determined by authentic standards, was scraped into vials and the radioactivity was quantified by liquid scintillation counting.

Receptor:Small G Protein Coimmunoprecipitation Studies

Native ARF.

Experiments to assess association of native ARF with the VPAC₂ receptor were carried out using the VPAC₂ receptor-expressing CHO cell line. Cells in Dulbecco's modified Eagle's medium were incubated for 10 min with/without 10 nM VIP before washing in cold EBSS and solubilization (30 min at 4°C) in 20 mM sodium HEPES (pH 7.5) with 1 mM sodium orthovanadate, 1 mM NaF, 1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride, 2 μg/ml aprotinin, 4 μg/ml leupeptin, 2 μg/ml pepstatin A, 50 μg/ml soybean trypsin inhibitor plus 5 mM CHAPS, 0.1% sodium cholate, and 1 M NaCl. Extracts were diluted 1:1 with the same buffer lacking salt but including 20% glycerol. After preclearing with protein G-Sepharose, samples were centrifuged at 12,000g for 30 min. Aliquots of supernatant were retained for ligand binding by polyethylene glycol 8000 (PEG) precipitation (see below), whereas others were incubated (16 h rolling at 4°C) with sheep anti-ARF1_98-112 immunoglobulin at 10 μl/ml, with/without blocking peptide at 10 μg/ml (Mitchell et al., 1998) or nonimmune sheep IgG (3 μg/ml; Sigma) as control. Excess protein G-Sepharose was added to each tube and incubated with rolling for 3 h at 4°C before centrifugation (12,000g for 5 min). The pellet was washed twice with equivalent buffer before resuspension into similar buffer with the addition of (sonicated) phosphatidylcholine (Sigma) to 3 mg/ml. The suspension was aliquoted (100 μl) into tubes with 400 μl of 50 mM Tris-HCl (pH 7.4) containing 6.25 mM MgCl₂, 1% BSA, 0.1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride, 0.5 mg/ml bacitracin, 7% glycerol, and 2 mg/ml phosphatidylcholine. Approximately 80,000 cpm of¹²⁵I-helodermin was added per tube, with/without 1 μM unlabeled helodermin to define nonspecific binding. Assays were incubated for 1 h on ice before the addition of 0.6 ml of 30% PEG and 0.1 ml of bovine γ-globulin (0.1%). After mixing and standing on ice for 15 min, the samples were centrifuged at 12,000g for 20 min, the supernatant was aspirated, and the tube tips were removed for gamma-counting of the pellets.

Epitope-Tagged ARF.

COS 7 cells were cotransfected with expression plasmids encoding the PAC_1-null and PAC_1-hop1 receptors and ARF1 with a carboxyl-terminal HA epitope tag (in pcDNA3 and pXS, respectively). Transfections were carried out with 8 μg of receptor plasmid DNA, 2 μg of ARF plasmid DNA, and 30 μl of FuGENE 6/175-cm² flask. Seventy-two hours later, quiescent cells were washed in cold EBSS before solubilization (1 h on ice) in standard phosphate-buffered saline (PBS) with the same protease- and phosphatase-inhibitors used for experiments with native ARF, plus 1% CHAPS. After preclearing with protein G-Sepharose, extracts were centrifuged at 12,000g for 30 min. Aliquots of supernatant were retained for binding and PEG precipitation, whereas others were incubated (16 h rolling at 4°C) with mouse monoclonal anti-HA IgG (clone 12CA5; Roche Diagnostics Ltd.) and/or control nonimmune mouse IgG to a total of 2 μg/ml. Excess protein G-Sepharose was added and incubated with rolling for 3 h at 4°C before centrifugation. The pellet was washed with solubilization buffer and then with PBS before resuspension into PBS. Samples were aliquoted (100 μl) into tubes with 400 μl of PBS and final concentrations of 10 mM MgCl₂ and 0.2% BSA. Approximately 11,000 cpm of¹²⁵I-PACAP-27 was added per tube, with/without 100 nM PACAP-38 to define nonspecific binding. After incubation for 1 h on ice, assays were terminated, and PEG precipitated and harvested as described above.

Results

VPAC₁ and VPAC₂ Receptors.

The cell-surface expression of ¹²⁵I-helodermin binding sites and the internalization of ligand were measured in selected CHO cell clones expressing VPAC₁ and VPAC₂ receptors (Fig.1A; Table1). Cell surface VPAC₁ and VPAC₂ receptors showed high affinity for helodermin (approximately 1–3 nM), as did the recognition sites from which ligand had subsequently become internalized (presumably receptors that had been present at the cell surface). The amount of accumulated ligand in intracellular stores was consistently greater than that remaining on the cell surface after incubations of more than 5 to 10 min. The time course of¹²⁵I-helodermin association with cell-surface VPAC₂ receptors and its subsequent internalization are illustrated in Fig. 1A, showing rapid equilibration of cell-surface binding to a steady state by 10 min and continuing extensive internalization reaching a maximum by 20 min. The time course of ligand binding in the VPAC₁ receptor CHO clone was not studied in detail.

The ability of the VPAC₁ and VPAC₂ receptor CHO clones to show activation of AC, PLC, and PLD in response to agonists was assessed by cAMP, [³H]InsP, and [³H]PtdBut production assays (Fig. 1, B and C; Table 2). VIP elicited robust, concentration-dependent increases in cAMP production at VPAC₁ and VPAC₂ receptors (Fig. 1B). The potency of the responses was similarly high at both receptors (subnanomolar) and the maximal size of response was less in the VPAC₁ receptor CHO cells; probably reflecting the lower level of receptor expression. Neither the VPAC₁ nor VPAC₂ receptor CHO cell clones showed produced any detectable rise in [³H]InsP production in response to VIP (0.3 nM–3 μM). Both VPAC₁ and VPAC₂ receptors did however mediate small but consistent [³H]PtdBut responses when stimulated by VIP (Fig. 1C). The potency of these PLD responses was considerably lower than that for cAMP responses of the receptors, but still in the low nanomolar range (Table 2). As with the cAMP responses, the maximal PLD response of the VPAC₁ receptor was less than that of the VPAC₂ receptor.

To assess whether the PLD responses might be downstream of the activation of G_s or adenylate cyclase, [³H]PtdBut production was measured in cells stimulated with 1 and 5 μg/ml cholera toxin (16 h) or 1 and 5 μM forskolin (during the 30-min assay). None of these stimuli caused any detectable increase in [³H]PtdBut production, despite large control PLD responses to 1 μM phorbol 12,13-dibutyrate being recorded in each assay (data not shown).

Receptors can activate PLD through many routes (Exton, 1997), including pathways downstream of PLC activation or involving the small G proteins ARF and Rho (Mitchell et al., 1998). Because PLC activation was undetectable in response to VIP in these cell lines, it seems unlikely that this is important in bringing about the PLD responses observed. In PLD responses of rhodopsin family GPCRs, sensitivity to BFA, an inhibitor of some ARF GTP-exchange factors, correlates closely with the involvement of ARF 1/3 in these responses, apparently through direct association with relevant receptors (Mitchell et al., 1998). Figure 1D shows that BFA inhibited VPAC₁ and VPAC₂ receptor PLD responses with relatively high potency, yet was without effect on cAMP responses.

Coimmunoprecipitation experiments were carried out on solubilized extracts of the VPAC₂ receptor-containing CHO cells to assess any ARF:receptor interaction. ARF 1/3 was immunoprecipitated broadly as described previously (Mitchell et al., 1998) and the VPAC₂ receptor was detected by specific binding of ¹²⁵I-helodermin to the precipitated proteins because suitable high-affinity antibodies for the VPAC₂ receptor are not available. No specific binding was detected in wild-type CHO cells. Approximately 28% of the initial specific ¹²⁵I-helodermin binding could be solubilized and recovered as viable specific ligand binding sites from VPAC₂ receptor-CHO cells by the procedure used. Figure 1E shows that specific binding of¹²⁵I-helodermin was associated with ARF immunoprecipitates and this was considerably reduced when the ARF antibody was blocked with excess of the antigenic peptide or when nonimmune IgG was substituted. Brief preincubation with the agonist VIP seemed to facilitate ARF:receptor association. The affinity of solubilized, and presumably also immunoprecipitated receptors for ligand was reduced so it is not possible to reliably calculate the proportion of the receptors engaging with ARF.

PAC_1-null and PAC_1-hop1 Receptors.

Similar levels of cell-surface ¹²⁵I-PACAP-27 binding sites and amounts of subsequently internalized ligand were measured in CHO cell clones expressing PAC_1-nulland PAC_1-hop1 receptors (Fig.2, A and B; Table3). The accumulation of ligand in both compartments of PAC_1-null and PAC_1-hop1 receptor CHO cells was inhibited with moderately high affinity (9–21 nM) by unlabeled PACAP-27 with no clear evidence for multiple components (mean Hill coefficient of 0.99 ± 0.11 from all experiments). The time courses for cell surface association of ¹²⁵I-PACAP-27 and for the internalization of ligand bound to receptors were generally similar between PAC_1-null and PAC_1-hop1 receptors (Fig. 2, A and B) and to data from the VPAC₂ receptor (Fig. 1A). The steady-state accumulation of ¹²⁵I-PACAP-27 into the intracellular compartment seemed to be slightly greater for the PAC_1-null receptor than for the PAC_1-hop1 receptor but neither this, nor slight differences in rates between the splice variants, were investigated further.

The cAMP, [³H]InsP, and [³H]PtdBut production responses of the PAC_1-null and PAC_1-hop1receptor CHO cells are shown in Figs. 2, C and D; 3, A and B; Table4. Their receptors mediated robust cAMP responses of very high potency in response to PACAP-38. Both EC₅₀ values and maximal responses were almost identical (Fig. 2C). Consistent with previous reports (Spengler et al., 1993; van Rampelbergh et al., 1997), both PAC_1-null and PAC_1-hop1receptors also showed clear [³H]InsP responses to PACAP-38 (Fig. 2D). The EC₅₀ values were similar (4–9 nM) but of considerably lower potency than the cAMP responses. The maximal evoked [³H]InsP response of the PAC_1-null receptor was consistently 2.5- to 3-fold greater than that of the PAC_1-hop1receptor despite the very similar cAMP responses. Both PAC_1-null and PAC_1-hop1clones also produced consistent [³H]PtdBut responses of similar potency to the [³H]InsP responses (Fig. 3A). The maximal evoked PLD response of the PAC_1-hop1 receptor was consistently around 3-fold that of the PAC_1-nullreceptor, in contrast to the pattern of their PLC responses. Time courses of these PLD responses showed no evidence of desensitization for at least 15 min (Fig. 3B). The [³H]PtdBut response of the PAC_1-hop1 receptor was significantly inhibited by BFA (at concentrations of 50 μM and above), whereas that of the PAC_1-null receptor was little affected; consistent with a greater involvement in the former response of an ARF-dependent pathway. cAMP and [³H]InsP responses of the PAC_1-hop1 receptor were unaffected by BFA. Figure3D shows that the [³H]PtdBut response of the PAC_1-null receptor was inhibited to a greater extent than that of the PAC_1-hop1 receptor by the PLC inhibitor U 73122; the converse of their sensitivity to BFA and consistent with a greater contribution of a PLC-dependent pathway to the PAC_1-null receptor response.

PAC_1-null and PAC_1-hop1receptors were cotransfected with ARF1-HA into COS 7 cells and receptor:small G protein coimmunoprecipitation was investigated in HA-directed immunoprecipitates. These were probed for specific¹²⁵I-PACAP-27 binding sites because suitable high-affinity PAC₁ receptor antibodies are not available. The solubilization procedure extracted similar levels of specific ¹²⁵I-PACAP-27 binding sites from PAC_1-null and PAC_1-hop1receptor-expressing cells (i.e., approximately 100 fmol/ml extract), as assessed by PEG precipitation binding assays, but no detectable specific binding in extracts of untransfected cells. With extracts of PAC_1-hop1 but not PAC_1-nullreceptor-expressing cells, increasing levels of anti-HA IgG in the immunoprecipitation gave a concentration-dependent increase in the amount of specific ¹²⁵I-PACAP-27 binding recovered in the precipitate (Fig. 3E). Under the present experimental conditions, preincubation with agonist did not seem to be a prerequisite for this association. The overall levels of IgG were kept constant by substitution of nonimmune mouse IgG.

Chimeric VPAC₂/PAC₁ Receptors.

To address the importance of the i3 sequence of the PAC₁ receptor splice variants in determining the route and extent of coupling to PLD, we constructed chimeric VPAC₂/PAC₁ receptors (Fig.4). These contained the body of the VPAC₂ receptor with a segment from tm5 to tm7 (i.e., including the i3 domain) derived from either PAC_1-null or PAC_1-hop1receptor sequences. Thus, the only difference between these two constructs was the additional 28 amino acid hop-1 cassette in the VPAC₂/PAC_1-hop1 construct.

The chimeric receptors were transiently expressed in COS 7 cells for assessment of their signaling responses and transfected GLP-1 receptors were also studied for comparison. Both VPAC₂/PAC_1-null and VPAC₂/PAC_1-hop1 receptor constructs mediated robust cAMP responses to VIP (mean maximal responses of 149- and 143-fold of basal and EC₅₀values of 0.8 ± 0.1 and 1.5 ± 0.2 nM, respectively; means ± S.E.M., n = 6; Fig.5A). The potency of these responses was less than in CHO cells (as described previously for wild-type PAC₁ and VPAC₂ receptors in COS 7 cells (Lutz et al., 1999). This may relate to different complements of relevant cellular proteins in the two cell types. GLP-1 (7-36) amide potently activated cAMP production (in GLP-1 receptor-expressing cells only) with an EC₅₀value of 0.034 ± 0.014 nM (mean ± S.E.M., n= 6). [³H]InsP responses of the receptors were not investigated. Both chimeric constructs, however, mediated clear and similar [³H]PtdBut production responses to VIP (Fig. 5B). Although the EC₅₀ values were greater than those for cAMP production, they were in the low nanomolar range (2.5 ± 1.7 and 5.6 ± 3.9 nM for the VPAC₂/PAC_1-null and VPAC₂/PAC_1-hop1 constructs, respectively, means ± S.E.M., n = 6–10). The corresponding maximal response values were also similar; 3.2 ± 0.2- and 2.5 ± 0.4-fold of basal, respectively. The GLP-1 receptor mediated a [³H]PtdBut response only at very high concentrations of GLP-1 (7-36) amide, with an EC₅₀ value of 0.98 ± 0.22 μM (mean ± S.E.M., n = 6) (i.e., greater than 10,000-fold less potent than the receptor's cAMP response). When the [³H]PtdBut responses of the receptors were challenged with BFA, only that of the VPAC₂/PAC_1-hop1 construct was significantly inhibited at BFA concentrations of 50 μM and above (Fig. 5C). This is consistent with the idea that transfer of the PAC_1-null or PAC_1-hop1 i3 domains into the body of the VPAC₂ receptor had respectively conferred ARF-independent and ARF-dependent coupling to PLD activation.

Native VPAC₂ and PAC₁ Receptors.

To assess whether similar mechanisms for PLD activation might occur in native cells, we examined responses of the VPAC₂receptor in GH₃ cells and the PAC₁ receptor in αT3-1 cells. In each case these are the only members of the VPAC/PAC receptor family expressed, and the predominant form of the PAC₁ receptor in αT3-1 cells is known to be the hop-1 splice variant (Rawlings et al., 1995; MacKenzie et al., 2001). In GH₃ cells, VIP elicited a robust cAMP response with an EC₅₀value of 1.9 ± 0.4 nM and a maximal response 7.6 ± 0.4-fold of basal. A smaller PLD response was also seen, with an EC₅₀ value of 30.6 ±11.2 nM and maximal response of 2.6 ± 0.1-fold of basal (means ± S.E.M.,n = 6) (i.e., only around 12-fold lower potency than the cAMP response) (Fig. 6A). In αT3-1 cells, PACAP-38 elicited cAMP production with an EC₅₀ value of 0.19 ± 0.13 nM and maximal response 12.1 ± 0.5-fold of basal. PLD was also clearly activated with an EC₅₀ value of 6.7 ± 1.0 nM and a maximal response 3.2 ± 0.1-fold of basal (means ± S.E.M.,n = 6) (i.e., approximately 35-fold lower potency than the cAMP response but still within the low nanomolar range of ligand concentration) (Fig. 6B). Both PLD responses were sensitive to BFA with mean IC₅₀ values of 64 μM for the VPAC₂ receptor in GH₃ cells and 65 μM for the PAC₁ receptor in αT3-1 cells (Fig. 6C), concurring with the properties observed for VPAC₂ and PAC_1-hop1receptors in transfected cells.

Discussion

Like other GPCRs in the secretin/parathyroid hormone receptor family (Probst et al., 1992; Segre and Goldring, 1993; Harmar and Lutz, 1994; Donnelly, 1997), the PAC₁ and VPAC receptors increase cellular cAMP levels, presumably via coupling to Gs. VPAC₁, VPAC₂, and PAC₁ receptors expressed here in CHO cells mediate robust increases in cellular cAMP levels but in addition can lead to the activation of PLD (and in some cases, PLC). cAMP signals occur at higher potency than phospholipase activation in all the receptors studied here, but substantial phospholipase activation still occurs at low nanomolar concentrations that are potentially relevant in a physiological context.

Neither the potency nor maximal response of cAMP production was altered by the presence of the hop1 cassette in i3 of the PAC₁ receptor (compared with the PAC_1-null form). Similar numbers of receptors were expressed both at the cell surface and in pools internalized from the cell surface in the PAC_1-null and PAC_1-hop1 CHO cells, matching their similar cAMP responses. In the VPAC₁ receptor CHO cells, receptor expression was lower than that in VPAC₂and PAC_1-null or PAC_1-hop1cells and the maximum but not the potency of the cAMP response was diminished.

PAC₁ receptors activate PLC (through a pertussis toxin-insensitive route) in a variety of cell types (Deutsch and Sun, 1992; Spengler et al., 1993; Schomerus et al., 1994; Pantaloni et al., 1996; Pisegna and Wank, 1996; van Rampelbergh et al., 1997). Rat PAC_1-null and PAC_1-hop1 receptors expressed in LLC-PK1 cells activate PLC with similar potency (Spengler et al., 1993) although other splice variants, the hip form and an N-terminally deleted form show reduced, and facilitated, potency of PLC activation, respectively (Spengler et al., 1993; Pantaloni et al., 1996). Human PAC_1-null and PAC_1-SV-2 (equivalent to hop1) receptors expressed in NIH/3T3 cells display similar potencies of PLC activation but the PAC_1-SV-2 variant gave a greater maximal response when expressed at similar levels of total receptors per cell (Pisegna and Wank, 1996). In the present study, the PLC response of the rat PAC_1-null receptor in CHO cells was greater than that of the PAC_1-hop1 variant for similar levels of receptor expression. Both species differences and host cell differences may contribute to the disparity.

In contrast, VPAC₁ and VPAC₂ receptors are less well known to activate PLC. Both receptors can elicit a modest PLC response (which is partly sensitive to pertussis toxin) when expressed in COS 7 cells (MacKenzie et al., 1996, 2001). In addition, another group reported a 1.5-fold increase in phosphoinositide hydrolysis in CHO cells expressing the VPAC₁ receptor (van Rampelbergh et al., 1997). However, the receptor density in their stable clones was 20 pmol/mg protein compared with the expression of only of 0.06 ± 0.01 pmol VPAC₁ receptor/mg protein here. In the present experiments, neither VPAC₁ nor VPAC₂ receptor CHO cell clones demonstrated detectable [³H]InsP responses to agonists.

The activation of PLD by members of the secretin/parathyroid hormone receptor family has been little studied, although it has been reported in the case of glucagon, calcitonin, and parathyroid hormone receptors (Pittner and Fain, 1991; Friedman et al., 1999; Naro et al., 1998). This is the first report of PLD activation by the VPAC₁, VPAC₂, PAC_1-null, and PAC_1-hop1receptors. Unlike the PLD responses of many rhodopsin family GPCRs there was no evidence for rapid desensitization, but a variety of factors, including assay conditions, cellular context, and receptor type could contribute to this and the issue was not further investigated here. Both VPAC₁ and VPAC₂ receptors mediated modest PLD responses to VIP (with a lower maximal response for the VPAC₁receptor, matching its lower level of expression). Similar nanomolar potencies were seen at both receptors although these were much weaker than the effects on cAMP production. PLD activation could not be mimicked by activators of Gs or adenylate cyclase and occurred in the absence of any detectable PLC responses, suggesting that activation of PLD did not occur downstream of either of these pathways. A lack of concurrent PLC activation was also seen with the glucagon receptor (Pittner and Fain, 1991), whereas PLD activation by the parathyroid hormone receptor was unaffected after the inhibition of PLC activity by U 73122 (Friedman et al., 1999). Instead, both VPAC₁ and VPAC₂ receptor PLD responses were inhibited with relatively high potency by the ARF inhibitor BFA, whereas cAMP responses of the VPAC₂ receptor, for example, were unaffected. These data are consistent with the physical association between VPAC receptors and the small G protein ARF demonstrated in Fig. 1E (a link that could potentially provide a basis for facilitated ARF-dependent PLD activation). Both PAC_1-null and PAC_1-hop1 receptors also displayed PLD responses and although these were again of lower potency than cAMP responses, they were of similar (or greater) potency than the PLC responses of the receptors. The PLD response of the PAC_1-nullreceptor (but not the hop1 variant) was inhibited by low concentrations of the PLC inhibitor U 73122, suggesting that it may result substantially from PLC-dependent pathways. In contrast, the PAC_1-hop1 receptor displayed a much greater maximal PLD response that (unlike that of the null variant) was sensitive to BFA, whereas its cAMP and PLC responses were unaffected. This suggests that the presence of the hop1 cassette is critical in linking the PAC₁ receptor to an ARF-dependent route of PLD activation. Correspondingly, immunoprecipitation of epitope-tagged ARF1 resulted in coprecipitation of PAC_1-hop1 but not PAC_1-nullreceptors (Fig. 3E).

Chimeric VPAC₂/PAC₁receptors containing either PAC_1-null or PAC_1-hop1 i3 domains were constructed to address whether BFA-sensitivity/insensitivity could be conferred just by an i3 domain swap. The chimeric VPAC₂/PAC₁ receptors with i3 domains from either PAC_1-null or PAC_1-hop1 receptors showed no apparent difference in their cAMP responses. However, the VPAC₂/PAC_1-null construct (just like the wild-type PAC_1-null receptor) showed a PLD response insensitive to BFA, despite the main body of the construct, apart from i3, being of VPAC₂(BFA-sensitive) origin. The VPAC₂/PAC_1-hop1 chimera retained BFA sensitivity, indicating that the i3 sequence of PAC₁ receptors is a critical determinant of coupling to ARF-dependent PLD activation. Some analogy can be drawn with the dopamine D₂ receptor, where i3 splice variants couple differentially to G_αi2 (Guiramand et al., 1995) and the calcitonin receptor where i1 variants couple differentially to PLC but not to AC (Nussenzveig et al., 1994). Thus, the alternative splicing of receptors may allow a more subtle selection of signals and hence control of cellular activity to be achieved. Analogous behavior of natively expressed VPAC₂and PAC₁ receptors was demonstrated using GH₃ and αT3-1 cell lines, respectively (Fig.6). Modest but significant PLD responses to agonists were seen in each case with sensitivity to BFA (matching in the case of αT3-1 cells, their predominant expression of the hop-1 splice variant (Rawlings et al., 1995).

A number of studies have pointed to a role of amphipathic helical domains incorporating basic amino acids in the coupling of GPCRs to G proteins. Peptides derived from the α₂-adrenergic receptor activate G_i/o in vitro, providing they possess basic amino acids spaced throughout the peptide and end with a BBxxB or BBxB (where x is any residue, and B is basic residue or in the last position either a basic or aromatic residue) (Ikezu et al., 1992; Wade et al., 1996). Groupings of basic and hydrophobic amino acids, situated in i3 of the muscarinic and the α₂-adrenergic receptors, have been implicated in interactions with G proteins (Burstein et al., 1998; Okamoto and Nishimoto, 1992; Wade et al., 1996; Wess, 1997). The VPAC₂ receptor contains a classical BBxxB motif at the i3/tm6 junction, whereas both the VPAC₁and PAC_1-hop1 receptors contain a motif similar to that seen in the α₂-adrenergic receptor with spaced basic residues and a cluster of basic amino acids upstream of tm6. The presence of the hop-1 insert in the PAC₁receptor provides the cluster of basic amino acids that completes a spaced basic residue motif. The PAC_1-nullreceptor has no classical or spaced base motif present in its i3 and this may potentially underlie its minimal ARF-dependent coupling to PLD.

In summary, the VPAC and PAC₁ receptors can activate PLD and although this is at higher concentrations of agonist than those required to elicit cAMP production, they may still be physiologically relevant. There are marked differences in the mechanisms apparently used to bring about this activation, with both ARF-dependent routes and PLC-dependent routes being implicated in different cases. From data with i3 splice variants of the PAC₁ receptor and with chimeric receptor constructs incorporating i3 domain swaps it seems that the i3 structure of the PAC₁ receptor is a critical determinant of both its physical association with ARF1 and its ARF-dependent coupling to PLD.