The Hop Cassette of the PAC1 Receptor Confers Coupling to Ca2+ Elevation Required for Pituitary Adenylate Cyclase-activating Polypeptide-evoked Neurosecretion*

We have identified the single PAC1 receptor variant responsible for Ca2+ mobilization from intracellular stores and influx through voltage-gated Ca2+ channels in bovine chromaffin cells and the domain of this receptor variant that confers coupling to [Ca2+]i elevation. This receptor (bPAC1hop) contains a 28-amino acid “hop” insertion in the third intracellular loop, with a full-length 171-amino acid N terminus. Expression of the bPAC1hop receptor in NG108-15 cells, which lack endogenous PAC1 receptors, reconstituted high affinity PACAP binding and PACAP-dependent elevation of both cAMP and intracellular Ca2+ concentrations ([Ca2+]i). Removal of the hop domain and expression of this receptor (bPAC1null) in NG108-15 cells reconstituted high affinity PACAP binding and PACAP-dependent cAMP generation but without a corresponding [Ca2+]i elevation. PC12-G cells express sufficient levels of PAC1 receptors to provide PACAP-saturable coupling to adenylate cyclase and to drive PACAP-dependent differentiation but do not express PAC1 receptors at levels found in postmitotic neuronal and endocrine cells and do not support PACAP-mediated neurosecretion. Expression of bPAC1hop, but not bPAC1null, at levels comparable with those of bPAC1hop in bovine chromaffin cells resulted in acquisition by PC12-G cells of PACAP-dependent [Ca2+]i increase and extracellular Ca2+ influx. In addition, PC12-G cells expressing bPAC1hop acquired the ability to release [3H]norepinephrine in a Ca2+ influx-dependent manner in response to PACAP. Expression of PACAP receptors in neuroendocrine rather than nonneuroendocrine cells reveals key differences between PAC1hop and PAC1null coupling, indicating an important and previously unrecognized role of the hop cassette in PAC1-mediated Ca2+ signaling in neuroendocrine cells.

The actions of PACAP in the sympathoadrenal axis are mediated mainly through its G-protein-coupled receptor, PAC1. Several PAC1 variants exist in neurons and endocrine cells, which differ at the N terminus and in the third intracellular loop (IC3) of this G-protein-coupled receptor (GPCR). N-terminal splice variants show slightly different affinities for PACAP binding but similar G-protein-coupling properties (20,21). IC3 variants originate from the alternative splicing of neither, one, or both of two exons, hip or hop, within this region (22,23). All variants, including PAC1 null , which has neither hop nor hip insertions in IC3, bind PACAP-38 and PACAP-27 with high affinity and couple to G s to activate adenylate cyclase (AC) (22,23). Insertion of the hip cassette in the PAC1 receptor either abolishes or impairs coupling to phospholipase C (PLC) when these receptors are expressed in the nonneuroendocrine LLC-PK1 (LLC-PK1, pig kidney epithelial) or NIH3T3 (mouse embryonic fibroblast) cell lines (22,23). Human PAC1 hop and PAC1 null receptors couple to both AC and PLC in the Chinese hamster ovary cell line (CHO) commonly used for receptor reconstitution; however, PAC1 hop uniquely binds to other signaling molecules, at least in these cells, that could explain differential signaling between the two receptor variants (24,25). Finally, a rat PAC1 variant consisting of an amino acid substitution in transmembrane region two and a two-amino acid deletion with a two-amino acid substitution in transmembrane region four has been reported to mediate Ca 2ϩ influx in CHO cells, without coupling to AC (26), although the existence of this variant in other species is in doubt (14). Thus, the structural basis for PAC1 receptor coupling to cytosolic Ca 2ϩ mobilization and extracellular influx in neuronal and endocrine cells remains unresolved, significantly impeding progress in understanding this crucial aspect of class II GPCR-Ca 2ϩ coupling.
Determining receptor variant specificity unambiguously with respect to second messenger coupling is key to understanding the mechanisms of PACAP-evoked exocytosis, since stimulation of exocytosis by PACAP and other neurotransmitters may be modulated by cAMP and interactions between cAMP and Ca 2ϩ in addition to the required rise in cytosolic Ca 2ϩ concentration ([Ca 2ϩ ] i ) that triggers the final step of exocytosis (27). PACAP-mediated secretion from bovine chromaffin cells (BCCs) and rat pheochromocytoma (PC12) cells is reported to require Ca 2ϩ influx through voltage-gated Ca 2ϩ channels (VGCCs), with predominant contributions from either L-type or non-L-type channels depending on the time frame of secretion and the type of cell (1,2,5,8,28,29). Since all PAC1 receptor variants elevate cAMP, various laboratories have investigated cAMP involvement in potentiating PACAPinduced [Ca 2ϩ ] i elevation or otherwise contributing to PACAP-induced secretion from chromaffin cells (30 -33). PACAP also causes depolarization of pituitary cells, pinealocytes, and sympathetic neurons through Na ϩ channels or Ca 2ϩ influx through nonspecific cation or transient receptor potential channels (17, 34 -36), pathways that may also contribute to Ca 2ϩ influx required for exocytosis. PAC1 receptor activation could therefore stimulate secretion both directly and indirectly through Ca 2ϩ mobilization from intracellular stores, Ca 2ϩ influx, and cAMP generation. Evidence for PACAP coupling to these second messengers through PAC1 comes from a variety of endocrine and nonendocrine cells, however, and may be specific to both the cell type and receptor variant. Therefore, identifying the mechanism of secretion triggered by PACAP at the adrenomedullary synapse, at which PACAP is the major physiologically relevant slow transmitter, requires (a) identification of the PAC1 variant or variants present in chromaffin cells, (b) reconstitution of PACAP-dependent calcium signaling and secretory competence in a neuroendocrine cell background, and (c) determination of the structural motifs responsible for reconstituted signaling leading to secretion. We demonstrate here that BCCs express a single PACAP receptor, bPAC1 hop , and reveal the structural requirements for coupling to intracellular Ca 2ϩ mobilization, Ca 2ϩ influx, and neurosecretion through reconstitution in PC12-G cells.

EXPERIMENTAL PROCEDURES
Materials-PACAP-38 and PACAP-27 were purchased from Phoenix Pharmaceuticals (Mountain View, CA). The radioligand, [ 125 I]PACAP-27 (specific activity 2200 Ci/mmol) was obtained from PerkinElmer Life Sciences. Tetrodotoxin (TTX), veratridine, and forskolin were purchased from Calbiochem-EMD Biosciences. All tissue culture reagents were purchased from Invitrogen unless otherwise specified. All other reagents used were purchased from Sigma-Aldrich.
Measurement of Met-Enkephalin Release-Release studies were performed on chromaffin cells 24 h after differential plating. Briefly, cells were pretreated for 30 min with 500 l of CCCM containing 30 M D600, 1 M TTX, 10 M nimodipine, vehicle, or Ca 2ϩ -free CCCM (Ca 2ϩ concentration from serum alone, ϳ100 M). A further 500 l of CCCM or Ca 2ϩ -free CCCM, either with or without 200 nM PACAP-38 or 20 M veratridine, was then added to each well containing each of the above inhibitors, resulting in a final concentration of 100 nM PACAP or 10 M veratridine/1000-l final well volume. Following 1 h of treatment, Met-enkephalin was assayed from medium aliquots or 0.1 N HCl-lyophilized cell extracts by radioimmunoassay (37). The statistical difference between control and treatment groups was determined using one-way analysis of variance by Tukey's multiple comparison test using GRAPHPAD PRISM, version 4 (GraphPad Software).
Preparation of cDNA-Bovine chromaffin, NG108-15, or PC12-G cells grown in Costar 162-cm 2 flasks (Corning) were harvested, washed once with phosphate-buffered saline, rapidly resuspended in Trizol (Invitrogen), snap-frozen, and stored at Ϫ80°C. Total RNA was then isolated using the RNeasy Mini total RNA isolation kit (Qiagen) according to the manufacturer's instructions with the following modifications. Samples were thawed on ice and extracted with a final concentration of 20% chloroform. Following centrifugation at 10,000 rpm for 10 min at 4°C, an equal volume of 70% ethanol was added to the aqueous phase, and the sample was applied to the RNeasy Mini spin column and processed according to Ref. 69. First strand cDNA was synthesized from 5 g of total RNA using the Superscript TM First-Strand Synthesis System according to the manufacturer's instructions (Invitrogen). Reverse transcription-PCR reactions were carried out using 1 l (ϳ250 ng) of the first-strand cDNA synthesis reaction as described below.
Identification and Cloning of the BCC PAC1 Receptor-The previously published sequence of the cloned bovine brain PAC1 receptor (accession number D17290) (39) was used to design a series of primers to amplify the receptor in overlapping segments and test for splice variants as described in Fig. 3. Primers (5Ј-ATGAGAGGCGGGCGGCACTG-3Ј and 5Ј-CCAACTG-TGTACAGGGCCTTCACTG-3Ј) were used to isolate the N-terminal region of the PAC1 receptor. The entire transmembrane 1-7 region, including the third intracellular loop known to contain inserts, was amplified using primers (5Ј-GCCCTG-TACACAGTTGGCTACA-3Ј and 5Ј-TTGATCTCCGCCTG-CACCTC-3Ј). The C terminus of the receptor was finally ampli-fied using primers (5Ј-AGAGGTGCAGGCGGAGATCA-3Ј and  5Ј-TCAGGTGGCCAGGTTGTCGG-3Ј), resulting in the generation of three overlapping amplicons representing the fulllength PAC1 receptor. To further investigate the presence of  third intracellular loop variants containing either hip, hop, or  hiphop cassette insertions, transmembrane region 5-6 was  amplified with primers (5Ј-GGCCCTGTAGTTGGCTC-CATAATGG-3Ј and 5Ј-GGGAGAAAGCAAAGACAGTG-3Ј). To determine which cassette was present in the IC3 region located between TM regions 5 and 6, the TM5 forward primer (5Ј-GGCCCTGTAGTTGGCTCCATAATGG-3Ј) was used in combination with reverse primers designed to anneal within the bovine hop (5Ј-AGAGTAATGGTGGACAGTTCTGACA-3Ј) or rat hip (5Ј-AGGAAAGGGGGTGAATGCTG-3Ј) cassettes. All PCRs contained a 0.25 M concentration of each primer, 1 mM dNTPs, 5 mM MgCl 2 , 2.5 units of high fidelity Pfu Turbo DNA polymerase (Stratagene), and ϳ250 ng of cDNA template. Amplification consisted of an initial denaturation step of 95°C for 5 min followed by 36 cycles of 95°C for 30 s, 65°C for 1 min, 72°C for 1 min with a final extension step of 72°C for 10 min.
The entire open reading frame of the bovine PAC1 receptor was obtained using primers (5Ј-ATGAGAGGCGGGCG-GCACTG-3Ј and 5Ј-TCAGGTGGCCAGGTTGTCGG-3Ј) under the PCR conditions specified above. The amplified bovine chromaffin PAC1 receptor cDNA was then directly cloned into the pCRII-TOPO plasmid (Invitrogen) according to the manufacturer's instructions. Recombinant plasmids containing the entire open reading frame (ORF) or various segments of the receptor were sequenced using M13 reverse and M13 forward primers by the dideoxy chain termination method (DNA Sequencing Facility, NINDS, National Institutes of Health). Following multiple sequence verification, the bovine chromaffin cell PAC1 receptor cDNA was subcloned into a mammalian expression vector, pcDNA3, and used for all subsequent studies.
Removal of the Hop Cassette-The entire hop cassette was removed from the bovine PAC1 receptor using PCR-mediated deletion mutagenesis based on a previously described method (40,41). Briefly, first round PCR was carried out using one chimeric and one nonchimeric primer to amplify the upstream and downstream regions immediately surrounding the hop cassette to be deleted. The PCR amplicons were then diluted 1:1000, and 1 l of the upstream and downstream amplicons were combined and used as template for a second round of ligation PCR using the outermost primer pairs (forward primer, 5Ј-TGAAT-GAGAGGCGGGCGGCA-3Ј; reverse primer, 5-GGCTCAG-GTGGCCAGGTTGT-3Ј). PCRs were carried out according to the conditions described above. The final single amplicon was purified and cloned into pcDNA3, and deletion of the hop cassette was confirmed through multiple sequencing reactions.
Stable and Transient Transfection Studies-NG108-15 cells were transfected (24-well plate, 2-cm 2 surface area) with a mixture of 0.8 g of DNA (pcDNA3 alone; pcDNA3 containing the bovine PAC1 receptor (bPAC1 hop ) cDNA; pcDNA3 containing the bovine PAC1 receptor deletion mutant (bPAC1 null ) cDNA; or pRT43.2, a cytomegalovirus promoter-driven GFP reporter plasmid, and 2 l of Lipofectamine TM (Invitrogen) in the absence of serum according to the manufacturer's instructions). For creation of bPAC1 hop stable cell lines, NG108-15 cells were transfected with PvuI (New England Biolabs)-linearized pcDNA3 containing bPAC1 hop receptor cDNA as described above. Cells were then selected using 500 g/ml G418 and continuously maintained in 300 g/ml G418 thereafter. Stable cell lines were screened through receptor binding studies and selected based on their ability to bind [ 125 I]PACAP-27 compared with mock-transfected NG108-15 cells. PC12-G cells were transfected (24-well plate, 2-cm 2 surface area) with 1 g of DNA as specified above and 2 l of Lipofectamine2000 TM (Invitrogen) in the presence of serum according to the manufacturer's instructions. For creation of stable cell lines, PC12-G cells were transfected with linearized plasmid as described above and selected and continuously maintained using 500 g/ml G418. PC12-G cells stably expressing the bPAC1 hop receptor were selected based on gain of [ 125 I]PACAP-27 binding and Ca 2ϩ mobilization compared with mock-transfected naive PC12-G cells. All transiently transfected cells were assayed 48 h post-transfection, whereas stable cell lines were plated 24 h prior to experimentation.
Preparation of Crude Cell Membranes-Crude membranes were prepared as previously described (43). Briefly, subconfluent cells were harvested in hypotonic buffer (50 mM Tris-HCl and 5 mM EDTA, pH 7.4), homogenized with a hand-held tissue homogenizer (Tissuemizer; Fisher Scientific) for 10 s, and centrifuged at 600 ϫ g for 5 min to remove cellular debris. The supernatant was centrifuged at 25,000 rpm for 20 min, and the membrane pellet was resuspended in PACAP wash buffer (50 mM Tris-HCl, 5 mM MgCl 2 , pH 7.4) supplemented with Roche cocktail inhibitor tablets (Roche Applied Science). Total protein concentration was determined using the Bio-Rad DC protein colorimetric assay according to the manufacturer's instructions (Bio-Rad).
Receptor Binding Studies-NG108-15 or PC12-G cell membranes (30 g) were incubated in the presence of 150 pM [ 125 I]PACAP-27 and increasing concentrations of unlabeled PACAP-38, PACAP-27, and VIP in PACAP binding buffer (50 mM Tris-HCl, 5 mM MgCl 2 , pH 7.4) supplemented with 0.5 mg/ml bacitracin, 200 M phenylmethylsulfonyl fluoride, 100 M aprotinin, and 1% bovine serum albumin) for 1.5 h at 37°C. All assays were set up and incubated in Multiscreen TM glass fiber plates (Millipore), precoated for 2 h at 4°C with 0.1% polyethyleneimine prepared in PACAP wash buffer as described above. Free and bound radioligand were then separated by filtration using the multiscreen vacuum manifold according to the manufacturer's instructions. Plates were washed four times with ice-cold PACAP wash buffer, filters were punched, and the retained radioactivity was measured on a ␥-counter. The radioligand binding data were analyzed using version 4 of the GRAPHPAD PRISM program.
Intracellular cAMP Measurements-Stably transfected NG108-15 cells were plated onto 24-well plates (Costar-Corning) coated with poly-L-lysine (0.1 mg/ml in phosphate-buffered saline) at a density of 70,000 cells/well 24 h prior to being serum-deprived for 24 h. For transient transfection studies, cells were plated at a density of 70,000 cells/well 24 h posttransfection according to the conditions described above and serum-deprived 24 h prior to treatments. The NG108-15 cells were then treated with varying concentrations (10 Ϫ7 to 10 Ϫ11 M) of PACAP-38, PACAP-27, or VIP for 20 min at 37°C in DMEM containing 100 units/ml penicillin/streptomycin, 2 mM glutamine, 1ϫ HAT supplement, 1 mM isobutylmethylxanthine, and 0.1% bovine serum albumin. Intracellular cAMP was then measured using the Biotrak nonacetylation enzyme immunoassay system with the supplied lysis reagent according to manufacturer's instructions (Amersham Biosciences). 24 h posttransfection, PC12-G cells were plated into 24-well plates coated with 0.1 mg/ml poly-L-lysine at a density of 250,000 cells/well, allowed to adhere for 24 h, and then treated with varying concentrations (10 Ϫ7 to 10 Ϫ11 M) of PACAP-38, PACAP-27, or VIP for 20 min at 37°C. Intracellular cAMP content was determined as described above.
Single Cell [Ca 2ϩ ] i Measurements-[Ca 2ϩ ] i was measured as previously described (44). Briefly, 150,000 (NG108-15) or 400,000 (PC12-G) cells were seeded onto 1.5-cm diameter glass coverslips (Assistent) coated with 0.5 mg/ml poly-L-lysine in 12-well plates (Costar-Corning). Transiently transfected cells were plated 24 h post-transfection at the cell densities specified above. For transient transfection studies, cells were cotransfected at a ratio of 1:3 (bPAC1 hop or bPAC1 null in pcDNA3:pRT43.2GFP) in 162-cm 2 flasks and plated 24 h posttransfection under the specified conditions. 24 h postseeding, cells were washed once with Krebs-Ringer buffer (KRB) containing 125 mM NaCl, 5 mM KCl, 1 mM Na 2 HPO 4 , 1 mM MgSO 4 , 1 mM CaCl 2 , 5.5 mM glucose, and 20 mM HEPES, pH 7.3. The cells were then loaded with 4 M Fura-2/AM (Invitrogen) in KRB for 22 min at room temperature. Following loading, cells were washed once with KRB and incubated for a further 22 min at room temperature in Fura-2/AM-free KRB. The coverslips were then mounted onto a custom-built perfusion chamber and placed on an inverted Olympus microscope equipped with a 40ϫ high numerical aperture 340-nm optimized lens and perfused with KRB in the presence or absence of drugs. Ca 2ϩ -free studies were carried out as described above using Ca 2ϩ -free KRB supplemented with 100 M EGTA. The preparations were then excited at 340 and 380 nm, respectively, and fluorescence intensity was measured every 2 s at an emission wavelength of 510 nm. Images were digitally captured and analyzed with the software MetaFluor (Molecular Devices). Among transiently transfected cells, bPAC1 hop -or bPAC1 nullexpressing cells were identified by GFP fluorescence prior to [Ca 2ϩ ] i measurements, compared with GFP-nonexpressing, untransfected NG108-15 or PC12-G cells. Ratio values were then converted to calibrated [Ca 2ϩ ] i values. F max values were obtained by exposing the cells to 10 M ionomycin in 1 mM Ca 2ϩ -containing KRB, and F min values were obtained by exposing ionomycin-treated cells to Ca 2ϩ -free KRB supplemented with 100 M EGTA. The corresponding ratios, F min and F max were substituted into the equation described by Grynckiewicz et al. (45) to convert data point ratios to [Ca 2ϩ ] i . Individual calibrations were used for each cell type analyzed in this study.
Catecholamine Uptake and Release Assay-Naive PC12-G cells or PC12-G cells stably expressing the bPAC1 hop receptor were plated at a density of 200,000 cells/well in 0.1 mg/ml poly-L-lysine-coated 24-well plates and allowed to adhere for 24 h. Cells were labeled with 1 Ci/well levo-[7-3 H]norepinephrine (1 mCi/ml; PerkinElmer Life Sciences) in PC12 complete media for 4 h at 37°C. For treatments, cells were washed once with phosphate-buffered saline and preincubated for 30 min at 37°C in 400 l of KRB as described above. Cells were subsequently stimulated by the addition of 100 l of KRB with or without 5ϫ concentrated secretagogues, resulting in a final concentration of 100 nM PACAP-38 or 55 mM KCl. Following stimulation for 30 min at 37°C, the total medium (500 l) was collected, and cells were lysed with 500 l of KRB containing 1% Triton X-100. Ca 2ϩ -free studies were carried out using Ca 2ϩ -free KRB supplemented with 100 M EGTA as described above. The level of radioactivity in both medium and cell lysates was then determined by liquid scintillation counting. Levo-[7-3 H]norepinephrine secretion was calculated as the percentage of total radioactivity released, where total activity is the amount released plus the amount remaining in the cells. The statistical difference between control and treatment groups was determined using one-way analysis of variance by Tukey's multiple comparison tests using GRAPHPAD PRISM, version 4.

Characteristics of PACAP-stimulated Neuropeptide Secretion and [Ca 2ϩ ] i Elevation in BCCs-
We examined the Ca 2ϩ -and cAMP-dependent characteristics of PACAP-regulated acute (1-h) secretion from bovine chromaffin cells using endogenous Met-enkephalin as a marker of release from large dense core vesicles (46). The ability of PACAP to cause Met-enkephalin release from BCCs was not dependent on Na ϩ influx-evoked cellular depolarization, since TTX, which inhibits voltage-dependent Na ϩ influx leading to action potential generation and opening of VGCCs, failed to block PACAP-evoked secretion (Fig. 1). Treatment of cells with 25 M forskolin did not mimic the acute secretagogue effects of PACAP under conditions (Met-enkephalin secretion during 60 min of stimulation) in which 40 mM KCl or 100 nM PACAP gave a 3-5-fold stimulation over basal level (data not shown), indicating that the ability of PACAP to elevate intracellular cAMP by coupling to adenylate cyclase through G␣ s is insufficient on its own to stimulate acute secretion from large dense core vesicles in BCCs. PACAP-evoked secretion of Met-enkephalin was abolished in Ca 2ϩ -free medium (Fig. 1) and by exposure of BCCs to 30 M D600, a pan-specific inhibitor of VGCCs (Fig. 1), consistent with previous reports that PACAP-dependent Ca 2ϩ influx occurs through one or more VGCCs in BCCs (1,36). Blockade of PACAP-evoked acute secretion, unlike regulation of neu-ropeptide biosynthesis and sustained (24 -72-h) secretion (2,8), was also blocked by the L-type Ca 2ϩ channel inhibitors, nimodipine ( Fig. 1) (39) to design a set of primers to amplify the entire open reading frame (ORF) of the receptor in three overlapping segments: the N terminus, transmembrane regions 1-7, and the C terminus (Fig. 3). First, amplification of the N terminus region indicated the presence of a full-length transcript without the 21-or 57amino acid truncations reported in some PAC1 transcripts (20,21) or insertion of 24 amino acids reported in others (47). Generation of an 857-bp transcript following amplification of the TM1 to -7 region of the receptor could be accounted for only by the presence of a single hip or hop cassette, but not both, in the third intracellular loop of the receptor (22,48). The use of internal primers to the hop or hip cassettes confirmed that the bovine PAC1 receptor contained the hop cassette within its IC3 region.
Importantly, the hop variant was the only PAC1 receptor species present in BCCs (Fig. 3B). No C terminus receptor variants of the PAC1 receptor were identified in bovine chromaffin cells, as shown by the single amplicon of the expected length of 193 bp in Fig.  3B. Following multiple sequence verifications, our results indicated that BCCs express a single splice variant of the receptor with the hop cassette in the IC3 loop. Finally, multiple sequencing of amplicons encompassing the fourth transmembrane domain failed to identify the reported transmembrane variant responsible for PAC1 receptor coupling to Ca 2ϩ influx reported previously (26), consistent with the absence of this sequence in either the human or rodent PAC1 gene loci (14). Since it is the only form of the PAC1 receptor expressed in bovine chromaffin cells, bPAC1 hop can be identified unambiguously with PACAP coupling to both cAMP generation and Ca 2ϩ mobilization and influx in these cells. This receptor was next expressed in homologous (neural and neuroendocrine) cells to recapitulate PACAP-dependent Ca 2ϩ signaling required for secretion.  (Fig. 4A). Both PACAP-38 and PACAP-27 stimulated the accumulation of intracellular cAMP in a concentration-dependent manner, with EC 50 values of 0.14 and 0.05 nM, respectively, in NG108-15 cells stably transfected with bPAC1 hop , indicating that coupling to AC, presumably through Gs, was reconstituted in these cells (Fig. 4B).
Exposure of NG108-15 cells stably expressing the bPAC1 hop receptor to 100 nM PACAP resulted in a rapid rise in [Ca 2ϩ ] i  . Schematic representation of PCR-based identification and cloning of the BCC PAC1 receptor. A, the previously published bovine brain PAC1 receptor sequence was used to design a series of primers that could amplify the receptor in three major overlapping segments that results in the generation of the entire open reading frame of the receptor. B, primer set 1 (1f-1r) was used to amplify the segment of the receptor from the start of the ORF to the beginning of transmembrane region 1. This resulted in the generation of a single 530-bp amplicon, indicating a full-length N terminus. Primer set B (2f-2r) was used to amplify TM regions 1-7, that includes both extracellular and intracellular loop regions of the receptor. An amplicon of 857 bp was generated, indicating a receptor variant with either a hop or hip cassette in the IC3 region. This was confirmed by the generation of a 271-bp amplicon using primer set 4 (4f-4r) designed to amplify the region between TM5 and TM6. To discriminate between the two cassettes, a reverse primer was designed to bind within either the hop (5r) or hip cassettes (6r). This resulted in the generation of a 200 bp using the hop specific primer (4f-5r), whereas no amplification was detected using the hip-specific primer (4f-6r), demonstrating that the bovine PAC1 receptor contains the hop cassette within the IC3 region. The C terminus region of the receptor from the TM7 to the end of the ORF was amplified using primer set C (3f-3r) and resulted in the generation of a 193-bp fragment. Sequence verification of the overlapping segments allowed the construction of a full-length sequence identical to the previously identified bovine brain PAC1 hop receptor. The entire ORF was then amplified (1f-3r), and the full-length, 1542-bp cDNA was cloned into mammalian expression vector pcDNA3.
(resting, 48 Ϯ 1.9 nM; peak, 742.2 Ϯ 61 nM; plateau, 53.5 Ϯ 2.96 nM [Ca 2ϩ ] i ) that was not observed in control untransfected NG108-15 cells (47 Ϯ 1.09 nM [Ca 2ϩ ] i ), confirming this as a bPAC1 hop receptor-mediated event (Fig. 4C). Removal of extracellular Ca 2ϩ did not affect the rapid initial rise of [Ca 2ϩ ] i fol-lowing PACAP (resting, 61.9 Ϯ 8.6 nM; resting Ca 2ϩ -free KRB, 63.9 Ϯ 4.1 nM; peak, 598.7 Ϯ 58 nM; plateau, 68.6 Ϯ 2.3 nM [Ca 2ϩ ] i ), indicating that the bulk of the Ca 2ϩ release into the cytoplasm was contributed by intracellular stores (Fig. 4C) with no obvious Ca 2ϩ influx "plateau phase" as observed in the   (Fig. 4D). Exposure to PACAP-38 and PACAP-27 also stimulated intracellular cAMP generation in a concentration-dependent manner with IC 50 values of 0.3 and 0.17 nM, respectively (Fig. 4E). These results demonstrated that removal of the hop cassette from the IC3 region of bPAC1 did not impair the N terminus binding properties of the receptor, receptor trafficking to the plasma membrane compartment, or its ability to couple to AC through G s in NG108-15 cells. Both bPAC1 hopand bPAC1 null -expressing NG108-15 cells had identical receptor levels (B max ϭ 11-12 pM receptor/mg of protein) (see Table  1), determined using data generated from homologous competitive binding experiments (Fig. 4, A and D).
Expression of the bPAC1 null receptor in NG108-15 cells at a density equivalent to that of bPAC1 hop reconstituted both high affinity PACAP-27 binding and PACAP-dependent cAMP generation. However, PACAP did not evoke [Ca 2ϩ ] i elevation in bPAC1 null -expressing NG108-15 cells (56.3 Ϯ 3.6 nM [Ca 2ϩ ] i ) (Fig. 4F), suggesting that the hop cassette plays a critical role in transducing PACAP-dependent Ca 2ϩ mobilization signaling in neural cells.

Expression of bPAC1 Confers High Density PACAP Binding and Hop-dependent Coupling to Ca 2ϩ Mobilization from ICS
and Ca 2ϩ Influx in PC12-G Cells-Our observations in NG108-15 cells (above) suggested that, unlike results predicted from regulation of PLC coupling through both PAC1 hop and PAC1 null receptors reported in nonneuronal cells, regulation of Ca 2ϩ mobilization by PAC1 receptor stimulation in neural cell lines may be specific to the hop variant of the receptor. We wanted to confirm these results in another neural cell line and extend them to analysis of functional sequelae of Ca 2ϩ mobilization by PACAP leading to exocytosis. bPAC1 hop and bPAC1 null variants were therefore expressed in rat PC12-G cells, a well characterized neuroendocrine cell line in which Ca 2ϩ -dependent catecholamine secretion has been well studied.
Expression of the bPAC1 hop or bPAC1 null receptors in PC12-G cells reconstituted high density and high affinity [ 125 I]PACAP-27 binding that could be specifically competed for by unlabeled PACAP-38 and PACAP-27 with IC 50 values of 4.8 and 26 nM, respectively, for bPAC1 hop (Fig. 5A) and 7.5 and 38 nM, respectively, for bPAC1 null (Fig. 5B). VIP did not compete for [ 125 I]PACAP-27 binding to either the endogenous rat or either of the bovine receptors. Mock-transfected PC12-G cells expressed high affinity but very low density [ 125 I]PACAP-27 binding (Ͻ10% of total specific binding measured in PC12-G cells transfected with bPAC1 hop ; data not shown). Treatment of either naive, bPAC1 hop -transfected, or bPAC1 null -transfected PC12-G cells with varying concentrations of PACAP-38 resulted in a potent concentration-dependent elevation of intracellular cAMP levels at comparable levels with EC 50 values of 0.78 nM (PC12-G), 1.2 nM (bPAC1 hop ), and 3 nM (bPAC1 null ) (Fig. 5, B and D). These data indicate that PC12-G cells are capable of appropriate trafficking, expression, and G-protein coupling of PAC1 receptors. The endogenous PAC1 receptor appeared to be maximally efficiently coupled to AC through G␣ s in PC12-G cells, since cAMP production induced by 100 nM PACAP-38 in PC12-G cells was not further increased by the Ͼ10-fold increase in PACAP-27 binding after expression of bPAC1 hop or bPAC1 null and is equivalent to that generated by direct supramaximal stimulation of adenylate cyclase with 25 M forskolin (results not shown).
Ca 2ϩ signaling was next examined in bPAC1 hop -expressing, bPAC1 null -expressing, or naive PC12-G cells following transient transfection. A GFP expression vector served as a marker of transfection efficiency to discriminate between bPAC1-and non-bPAC1-expressing PC12-G cells for single cell Ca 2ϩ imaging (Fig. 6). Following Fura-2 loading, GFP-positive and -negative cells were identified (Fig. 6, inset A). Expression of GFP alone affected neither basal [Ca 2ϩ ] i nor elevation of [Ca 2ϩ ] i elicited by either 100 nM PACAP or 40 mM KCl (data not shown). bPAC1 hop -expressing (GFP-positive) cells responded to exposure to 100 nM PACAP with a large and rapid rise in [Ca 2ϩ ] i typical of Ca 2ϩ release from ICS (resting, 64.6 Ϯ 3.95 nM; peak, 731 Ϯ 58 nM; plateau, 213 Ϯ 26 nM [Ca 2ϩ ] i ). This phase was followed by a prolonged phase of [Ca 2ϩ ] i elevation, attributable to extracellular Ca 2ϩ influx (Fig. 6), as shown previously in BCCs ( Fig. 2A; see Ref. 8). Untransfected cells responded to PACAP treatment with very small increases in [Ca 2ϩ ] i that reflect the limited coupling of endogenous PAC1 receptors expressed in PC12-G cells with Ca 2ϩ mobilization and influx (Fig. 6)  We next investigated whether reconstitution of [Ca 2ϩ ] i elevation by bPAC1 was specifically dependent on the presence of the hop cassette in PC12-G cells, as observed in NG108-15 cells. Expression of bPAC1 null in PC12-G cells, with GFP as a marker for transfection, allowed Fura-2 visualization of [Ca 2ϩ ] i following exposure to PACAP-38, which was not elevated beyond the small rise observed in untransfected PC12-G cells (Fig. 6). Thus, removal of the hop cassette uncoupled the PAC1 receptor from its ability to promote Ca 2ϩ mobilization from ICS as well as Ca 2ϩ influx, implicating the hop cassette within IC3 as a critical determinant in PAC1 receptor coupling to both Ca 2ϩ mobili- PACAP stimulation of bPAC1 hop -expressing PC12-G cells in Ca 2ϩ -free medium prevented the plateau phase of [Ca 2ϩ ] i elevation, indicating its dependence on extracellular Ca 2ϩ influx following Ca 2ϩ mobilization (Fig. 7). These findings indicate that PC12-G cells express molecular components required for PAC1 hop coupling to [Ca 2ϩ ] i elevation through Ca 2ϩ mobilization from ICS and via extracellular Ca 2ϩ entry, provided that the receptor is expressed at sufficient density (Ca 2ϩ -free KRB: resting, 70.23 Ϯ 4.9 nM; peak, 259.33 Ϯ 21 nM; plateau, 67.27 Ϯ 12.17 nM; Ca 2ϩ -KRB: resting, 62.3 Ϯ 2.5 nM; peak, 424.31 Ϯ 23.6 nM; plateau, 233.1 Ϯ 4 nM [Ca 2ϩ ] i ). These results are further confirmed by expression of the rat PAC1 hop receptor under the control of the cytomegalovirus promoter, which increased rPAC1 hop expression in PC12-G cells around 10-fold and also reconstituted both PACAP-mediated Ca 2ϩ mobilization from ICS (Ca 2ϩ peak) and influx (Ca 2ϩ plateau) in PC12-G cells (data not shown).

Expression of bPAC1 hop Confers PACAP-evoked [ 3 H]Norepinephrine Release in PC12-G Cells-PC12-G cells express very
low levels of PAC1 and do not exhibit either Ca 2ϩ mobilization or a substantial Ca 2ϩ influx following exposure to PACAP in KRB (Fig. 6). We hypothesized that these cells might lack competence for PACAP-evoked secretion, unlike some PC12 cell lines that are competent for both PACAP-evoked Ca 2ϩ influx and catecholamine secretion (28), and would therefore represent an effective model for reconstitution of PAC1-dependent catecholamine secretion.
As demonstrated in Fig. 8A, PC12-G cells efficiently accumulate [ 3 H]norepinephrine yet fail to release it in response to PACAP-evoked secretion in KRB, suggesting that the low density of endogenously expressed PAC1 receptors in PC12-G cells do not generate a sufficient Ca 2ϩ signal for regulated exocytosis. Membrane depolarization of naive PC12-G with 55 mM KCl caused significant secretion of norepinephrine (30% of total content) (Fig. 8A), thus confirming sufficient uptake/loading of cells with [ 3 H]norepinephrine. In contrast, bPAC1 hop -express- ing PC12-G cells exhibit robust stimulation of catecholamine release in response to 30-min exposure to 100 nM PACAP (30% of total content) (Fig. 8B), which is abolished upon removal of Ca 2ϩ from the medium (Fig. 8C), demonstrating that expression of the bPAC1 hop receptor in PC12-G cells confers Ca 2ϩ influx-dependent secretion evoked by PACAP, as previously demonstrated in BCCs.

DISCUSSION
The structural basis for PAC1 receptor coupling to Ca 2ϩ mobilization and influx has not been previously addressed in a heterologous neuroendocrine cell culture system; nor has the structural basis for PAC1-dependent Ca 2ϩ mobilization from ICS been investigated, which along with elevation of intracellular cAMP, is the hallmark of combinatorial signaling by the slow transmitter PACAP. The bPAC1 hop receptor isolated from BCCs contains an intact full-length N terminus and a hop cassette in IC3 as previously identified in rat and human tissues and bovine brain. This receptor variant binds PACAP with high affinity and activates second messenger systems, such as AC and presumably PLC. This phenotype is common to most of the known PAC1 receptor splice variants, which have been examined upon expression in nonneuroendocrine heterologous cell lines (22,50). Expression of the PAC1 hop variant is sufficient for coupling PAC1 receptor occupancy to Ca 2ϩ signaling, leading to exocytosis. The ability to couple to Ca 2ϩ influx has been previously ascribed to a transmembrane domain 4 variant of PAC1 (26), but this variant has never been demonstrated to cause Ca 2ϩ influx in a neuroendocrine cell, and the possibility of its natural occurrence has been challenged on structural and genetic grounds (14). In contrast, transfer of the bPAC1 hop receptor to PC12-G cells simultaneously conferred PACAP-dependent Ca 2ϩ coupling and catecholamine secretion. The ability of PACAP to cause large and long lasting changes in [Ca 2ϩ ] i , required to trigger regulated exocytosis, provides a mechanism for the physiological role of PACAP as a slow transmitter subserving long lasting adrenal catecholamine secretion in response to metabolic and other stressors (9). We have further shown that transfer of PACAP-dependent secretory competence is specific for PAC1 receptor expression at levels found at or near physiological receptor density (i.e. that found in bovine chromaffin cells) and that Ca 2ϩ coupling, but not coupling to AC, is highly specific for the hop cassette in IC3 of the receptor.
The hop and null forms of the PAC1 receptor are the major variants found in rats and humans, whereas systematic comparison of variants present in bovine tissues has not been carried out except in the adrenal medulla (i.e. this study). With respect to studies in heterologous cells, no dif-   ferences between these two major variants have been demonstrated that could account for their differential involvement in neurosecretion. Thus, the human and rat hop and null receptors have both been shown to be efficacious in coupling PAC1 receptor occupancy to PLC-mediated IP 3 generation, presumably through G q , in nonneuroendocrine cell lines (22)(23)(24)(25). Our current findings, however, demonstrate that in neural cell lines, such as NG108-15 and PC12-G, loss of the hop cassette results in a PAC1 receptor that is clearly deficient in coupling to Ca 2ϩ mobilization and influx, although it is fully competent for PACAP-dependent coupling to activation of AC.
PC12 cells are a commonly employed cellular model for investigating the neurotrophic effects of PACAP (51). Unlike some other PC12 cell variants, binding studies demonstrated that the PC12-G cells used here expressed very low levels of PAC1 receptors (ϳ10 times less, as measured by [ 125 I]PACAP-27 binding, than in bPAC1 hop -transfected PC12-G cells and severalfold less than in BCCs) and displayed only small elevations in [Ca 2ϩ ] i following exposure to PACAP. Previous studies have demonstrated the ability of PACAP to induce relatively large Ca 2ϩ transients in some PC12 cell lines, perhaps due to higher levels of expression of PAC1 hop than in PC12-G (52)(53)(54) and similar commonly employed PC12 cell lines (55). Reverse transcription-PCR analysis of the naive PC12-G cells used in our studies revealed the presence of two PAC1 receptor variants, PAC1 hop and PAC1 hip , which are expressed at similar but very low levels (56). 4 We found that increasing the expression of the endogenous rat PAC1 hop receptor to levels similar to those of bPAC1 hop in BCCs or in bPAC1 hop -PC12-G cells also reconstituted coupling to both intracellular Ca 2ϩ mobilization and Ca 2ϩ influx. Endogenous PAC1 expression in neurons in vivo can vary significantly in response to physiological and pathophysiological stimuli (55,57). The ability of PAC1 hop to couple to Ca 2ϩ mobilization and influx in combination with cAMP, rather than to cAMP alone, upon up-regulation of receptor expression in vivo, may correspondingly alter the postsynaptic slow neurotransmitter properties of PACAP, uncovering a new type of GPCR functional regulation through expression level-dependent coupling to cAMP alone (at low receptor density) or cAMP and Ca 2ϩ mobilization and influx (at higher receptor density).
One mechanism suggested for PACAP action on secretion through PAC1 receptors in various cell types is via cAMP/protein kinase A-mediated protein phosphorylation (31,32,53,58), possibly of VGCCs themselves, with a resultant potentiation of ion channel opening probability and duration (59 -63). However, as reported here, cAMP elevation alone following exposure to forskolin failed to mimic PACAP-evoked secretion or Ca 2ϩ elevation in BCCs or PC12-G cells, suggesting that bPAC1 hop -mediated cAMP generation alone is insufficient for initiating PACAP-stimulated Ca 2ϩ elevation, influx, or acute secretion. However, a modulatory role of cAMP in conjunction with [Ca 2ϩ ] i to maintain long term PACAP-evoked secretion from chromaffin cells is plausible and likely, based on partial inhibition of PACAP-stimulated secretion from perfused adrenal medulla by inhibition of protein kinase A (30).
Two major hypotheses have been proposed for initiation of [Ca 2ϩ ] i elevation by PACAP and other GPCR ligands in excitable (neuroendocrine) cells. The first is that receptor occupancy mediates membrane depolarization and subsequent facilitation of VGCC opening. PACAP has been variously shown to influence membrane potentials in supraoptic neurons (64), pinealocytes (34), sympathetic neurons (17), suprachiasmatic nuclear neurons (65), and chromaffin cells (36), through 4 T. Mustafa, unpublished results. Transfection of PC12-G cells with the bPAC1 hop receptor reconstituted PACAP-evoked norepinephrine secretion that was significantly higher compared with naive PC12-G cells (B). Removal of extracellular Ca 2ϩ abolished PACAP-evoked norepinephrine secretion without affecting basal secretion (C). Values represent the mean Ϯ S.E. of triplicate wells obtained from 3-7 independent experiments. Differences between control and treatment groups were determined using one-way analysis of variance followed by post hoc Tukey's multiple comparison test. **, p Ͻ 0.001 compared with corresponding control; ##, p Ͻ 0.001 compared with corresponding treatment. mechanisms proposed to involve Na ϩ influx through Na ϩ channels, Na ϩ or Ca 2ϩ entry through nonselective cation channels, or inhibition of outwardly directed K ϩ channels leading to depolarization sufficient to voltage-gate the opening of VGCCs (17, 34 -36, 62, 63). The second hypothesis is that PACAP affects Ca 2ϩ entry via prior activation of PLC-mediated Ca 2ϩ release from IP 3 -sensitive ICS (17), perhaps also requiring coactivation of protein kinase C (65), with PACAP-mediated Ca 2ϩ influx dependent on PLC-IP 3 -mediated Ca 2ϩ mobilization from ICS. This mechanism is the one most strongly supported by our results with bPAC1 hop , due to the association of the hop cassette with Ca 2ϩ mobilization in NG108-15 cells and with both mobilization and influx of Ca 2ϩ in PC12-G cells. Our identification of the specific PAC1 variant and the relevant IC3 domain, which supports PACAP-mediated Ca 2ϩ mobilization and influx in chromaffin cells, provides an important mechanistic insight into PAC1-Ca 2ϩ coupling and an experimental avenue for further exploration of the steps underlying initiation of Ca 2ϩ influx via this important class II GPCR.
We have provided evidence that removal of the hop cassette from the IC3 region of the bPAC1 receptor eliminates its ability to mediate PACAP-evoked Ca 2ϩ mobilization from ICS in both PC12-G and NG108-15 cells despite retention of high affinity binding and cAMP generation by PACAP. The requirement for a discrete region of the IC3 loop of the PAC1 receptor to activate Ca 2ϩ mobilization, Ca 2ϩ influx, and exocytotic secretion suggests the possibility that several components required for regulated exocytosis may be brought into close proximity through the hop-dependent formation of a specific signaling complex to trigger PACAP-dependent neurosecretion. Ronaldson and co-workers (25) have demonstrated that the small GTP-binding protein ARF6 (ADP-ribosylation factor-6) specifically interacts with the hop cassette to activate phospholipase D (24,25), and the role of ARF6 in membrane trafficking, synaptic vesicle biogenesis (66), and Ca 2ϩ -dependent large dense core vesicle exocytosis in both PC12 and chromaffin cells has been well established (67,68). ARF6 or a related GTPase may have a role in PAC1 hop -mediated exocytosis through formation of such a complex.
In conclusion, we have demonstrated that a single PAC1 receptor variant, bPAC1 hop , exists in BCCs and couples PAC1 receptor occupancy by PACAP to Ca 2ϩ mobilization, influx, and neurotransmitter release from large dense core vesicles. This mode of signaling appears to be neuroendocrine cell-specific and probably depends on the presence of specific signaling molecules in neuroendocrine cells. The demonstration that Ca 2ϩ coupling for secretion absolutely requires the hop cassette should allow rapid elucidation of the molecular bases for the sequential processes of Ca 2ϩ mobilization and influx preceding exocytosis evoked by PACAP through the class II G-protein-coupled receptor, PAC1.