Pituitary Adenylate Cyclase-activating Polypeptides Directly Stimulate Sympathetic Neuron Neuropeptide Y Release through PAC1 Receptor Isoform Activation of Specific Intracellular Signaling Pathways*

Pituitary adenylate cyclase-activating polypeptides (PACAP) have potent regulatory and neurotrophic activities on superior cervical ganglion (SCG) sympathetic neurons with pharmacological profiles consistent for the PACAP-selective PAC1 receptor. Multiple PAC1 receptor isoforms are suggested to determine differential peptide potency and receptor coupling to multiple intracellular signaling pathways. The current studies examined rat SCG PAC1 receptor splice variant expression and coupling to intracellular signaling pathways mediating PACAP-stimulated peptide release. PAC1 receptor mRNA was localized in over 90% of SCG neurons, which correlated with the cells expressing receptor protein. The neurons expressed the PAC1(short)HOP1 receptor but not VIP/PACAP-nonselective VPAC1 receptors; low VPAC2 receptor mRNA levels were restricted to ganglionic nonneuronal cells. PACAP27 and PACAP38 potently and efficaciously stimulated both cAMP and inositol phosphate production; inhibition of phospholipase C augmented PACAP-stimulated cAMP production, but inhibition of adenylyl cyclase did not alter stimulated inositol phosphate production. Phospholipase C inhibition blunted neuron peptide release, suggesting that the phosphatidylinositol pathway was a prominent component of the secretory response. These studies demonstrate preferential sympathetic neuron expression of PACAP-selective receptor variants contributing to regulation of autonomic function.

Pituitary adenylate cyclase-activating polypeptides (PACAP) have potent regulatory and neurotrophic activities on superior cervical ganglion (SCG) sympathetic neurons with pharmacological profiles consistent for the PACAP-selective PAC 1 receptor. Multiple PAC 1 receptor isoforms are suggested to determine differential peptide potency and receptor coupling to multiple intracellular signaling pathways. The current studies examined rat SCG PAC 1 receptor splice variant expression and coupling to intracellular signaling pathways mediating PACAP-stimulated peptide release. PAC 1 receptor mRNA was localized in over 90% of SCG neurons, which correlated with the cells expressing receptor protein.
The neurons expressed the PAC 1 (short)HOP1 receptor but not VIP/PACAP-nonselective VPAC 1 receptors; low VPAC 2 receptor mRNA levels were restricted to ganglionic nonneuronal cells. PACAP27 and PACAP38 potently and efficaciously stimulated both cAMP and inositol phosphate production; inhibition of phospholipase C augmented PACAP-stimulated cAMP production, but inhibition of adenylyl cyclase did not alter stimulated inositol phosphate production. Phospholipase C inhibition blunted neuron peptide release, suggesting that the phosphatidylinositol pathway was a prominent component of the secretory response. These studies demonstrate preferential sympathetic neuron expression of PACAP-selective receptor variants contributing to regulation of autonomic function.
The cloning of cDNAs for three putative seven-transmembrane G-protein-coupled receptors for VIP and PACAP demonstrated receptor subtype diversity and functional heterogeneity (17)(18)(19)(20)(21)(22)(23)(24)(25). The PACAP-selective PAC 1 receptor demonstrates high affinity for only PACAP38 and PACAP27 and is coupled to multiple intracellular signaling cascades. The VPAC 1 and VPAC 2 receptors, in contrast, exhibit approximate equal high affinity for the PACAP38, PACAP27, and VIP peptides and are coupled to adenylyl cyclase. Multiple PAC 1 receptor isoforms result from the alternative splicing of two exons in the aminoterminal extracellular domain and/or two (HIP and HOP) exons in the third cytoplasmic loop (19,23,24,26). Cell-specific expression of PAC 1 receptor splice variants determines differential peptide potency and distinct patterns of adenylyl cyclase and phospholipase C stimulation by PACAP38 and PACAP27 (23).
Recent studies have suggested that PACAP peptides may be physiological regulators of sympathetic neuron function. PACAP has been identified in pericellular fiber networks enveloping SCG postganglionic neurons, and sympathetic preganglionic neurons in the spinal cord projecting to the SCG express PACAP (27)(28)(29). We demonstrated previously that PACAP peptides are potent and efficacious regulators of SCG neuron transmitter and peptide production, secretion, and mRNA with pharmacological profiles congruous for the PACAP-selective PAC 1 receptor (30,31). Similarly, PACAP peptides selectively and potently stimulate sympathetic neuroblast mitogenesis, neuritogenesis, and survival and sympathetic neuron depolarization and membrane excitability (10,30,32). 2 Accordingly, PAC 1 receptor mRNA has been demonstrated in the rat SCG (30,31,33). The cellular targets of PACAP and the mechanisms underlying PACAP sympathetic actions, however, have remained unclear. To further evaluate the mechanisms mediating sympathetic neuron responses to PACAP, the present studies investigated SCG neuron expression of specific PAC 1 receptor molecular forms and determined whether PACAP-stimulated receptor activation of specific second messenger pathways modulates peptide secretion. The analyses demonstrate the preferential expression of specific PACAP-selective receptor isoforms coupled to multiple intracellular signaling pathways in sympathetic neurons; PAC 1 receptor stimulation of phospholipase C is a prominent component of SCG neuron neuropeptide Y (NPY) release.

EXPERIMENTAL PROCEDURES
Animals-Adult male (225-250 g) and pregnant female Harlan Sprague-Dawley rats were obtained from Charles River Canada; all animal procedures were approved by the University of Vermont IACUC. Animals were decapitated, and SCG were removed and frozen. Neonatal (postnatal day 1) rat ganglia were obtained from mixed sex litters.
Cell Culture-Primary neuronal SCG cultures were prepared as described previously (30,34). Neonatal rat SCG enzymatically dispersed to produce a pooled population of cells were plated at a density of 1.5 ϫ 10 4 neurons/cm 2 , treated with cytosine ␤-D-arabinofuranoside to eliminate nonneuronal cells, and maintained in defined complete serum-free medium containing 50 ng/ml nerve growth factor (34). For specific experiments, to obtain neuronal and nonneuronal cell populations, the SCG cells were preplated for 3 h; the neuronal cells were dislodged by agitation, recovered, and cultured in serum-free defined medium with mitotic inhibitors and nerve growth factor as described above. The remaining adherent cells were cultured as nonneuronal cells.
In Situ Hybridization Histochemistry-Adult male rat SCG cyrosections (20 m) were processed for in situ hybridization histochemistry as described previously (36 -38), using a riboprobe to the carboxyl-terminal intracellular domain of the PAC 1 receptor, which did not discriminate among the transcript splice variants. A 449-base pair fragment of the PAC 1 receptor cDNA synthesized from rat brain hypothalamus was amplified using the primer templates PACAPR5 and PACAPR6 (Table  I). The product was gel-purified and ligated into the EcoRV site of pBluescript II KS Ϫ cloning vector (Stratagene, La Jolla, CA). The nucleotide sequence of the insert was verified by automated fluorescent dideoxy dye terminator sequencing. Radiolabeled antisense and sense riboprobes were synthesized using [ 35 S]dCTP and T7 or T3 DNA polymerase, respectively; hybridization with sense riboprobes failed to produce detectable signals (data not shown).
Immunocytochemistry-Cultured sympathetic neurons on Aclar plastic (39) and cryosections of adult male rat SCG were immunocytochemically stained to localize the PAC 1 receptor protein using a modification of the protocol described previously (10,40). The neurons and sections were incubated for 24 h at 4°C with 1:2000 affinity-purified rabbit anti-PAC 1 receptor antiserum ERIQ, produced against amino acid residues Glu 35 to Cys 53 of the amino-terminal extracellular domain. PAC 1 receptor immunoreactivity was localized by incubation of the tissue with 1:500 indocarbocyanine (Cy3)-labeled donkey anti-rabbit IgG.
To identify the population of sympathetic neurons containing NPY that coexpressed the PACAP-selective receptor, neurons were incubated with 1:2000 rabbit anti-PAC 1 receptor and 1:400 guinea pig anti-NPY (41); sites of antibody binding were localized using 1:500 Cy3-labeled donkey anti-rabbit IgG and 1:500 fluorescein isothiocyanate-labeled goat anti-guinea pig IgG, and the subpopulation of neurons containing NPY that coexpressed the PACAP-selective receptor was determined (34,35,38). No staining was observed with omission of primary or secondary antisera, incubation with preimmune serum, or absorption of the primary antiserum with immunogen or antigen (data not shown). Immunocytochemically stained samples were viewed by fluorescence microscopy using a Leica DMRB microscope equipped with a Cy3 filter set (Chroma, Brattleboro, VT) or imaged using a Bio-Rad MRC 1000 confocal scanning laser system.
cAMP Production-Primary cultured SCG neurons were incubated with PACAP38, PACAP27, or VIP in defined serum-free medium containing 50 M RO20-1724 (Calbiochem). Cultures were extracted in absolute ethanol containing 100 M RO20-1724 and processed for cAMP a Positions are given as the nucleotide sequences for the rat cDNA with the following GenBank™ accession numbers: rat PAC 1 (short) receptor, Z23272; rat VPAC 1 receptor, M86835; rat VPAC 2 receptor, Z25885.
b Expected amplified product sizes for the sets of oligonucleotide primer templates for the common rat PAC 1 receptor (PACAPR5 and PACAPR6), rat VPAC 1 receptor, and rat VPAC 2 receptor are predicted from the cDNA sequences given in footnote a. The potential rat PAC 1 receptor third cytoplasmic loop alternative splice variant product sizes and GenBank™ accession numbers are as follows: PAC 1 (short) (neither HIP nor HOP), 303 bp, Z23279; PAC 1 (short) HIPHOP1, 471 bp, Z23272; PAC 1 (short)HIP, 387 bp, Z23273; PAC 1 (short)HOP1, 387 bp, Z23274; PAC 1 (short)HOP2, 384 bp, Z23275. The PAC 1 receptor amino-terminal extracellular domain very short transcript without exons 4 and 5 and the short variant containing these alternatively spliced exons are predicted to produce 350-and 413-nucleotide amplified products, respectively (44). radioimmunoassay using the Biotrak nonacetylation protocol with 125 I-cAMP and Amerlex-M magnetic separation (Amersham Pharmacia Biotech). Assay midpoints were approximately 10 fmol.
Inositol Phosphate Production-Total inositol phosphates were quantitated as described previously (42). SCG neurons were cultured in defined medium containing 0.32 M myo-[ 3 H]inositol (19 Ci/mmol, NEN Life Science Products). After 48 h, the cultures were treated with 10 mM LiCl and PACAP38, PACAP27, or VIP. Following phase separation of the methanol/chloroform extract, inositol phosphates in the aqueous phase were separated from free inositol by ion exchange chromatography (AG 1-X8 resin, formate form; Bio-Rad).
Messenger RNA Analysis-Total RNA prepared using RNA STAT-60 total RNA/mRNA isolation reagent (Tel-Test "B"; Friendswood, TX), was used to synthesize first strand cDNA using SuperScript II reverse transcriptase and oligo(dT) primers with the SuperScript Preamplification System (Life Technologies, Inc.) as described previously (29,30,35,37,38). Amplification with AmpliTaq DNA polymerase (Perkin-Elmer) using the AmpliWax PCR gem-facilitated hot start (30) was conducted using oligonucleotide primers specific for the identification of PAC 1 receptor splice variants in the third cytoplasmic loop or aminoterminal domain or for VPAC 1 or VPAC 2 receptors (Table I). Complementary DNA synthesis in the absence of either RNA or reverse transcriptase or amplification without template, primers, or DNA polymerase failed to yield products (data not shown).
The identities of the amplified products were established by diagnostic restriction endonuclease analysis (43). Verification of HIP or HOP alternative splice variant expression was performed using sequencespecific hybridization as reported previously (38) with the synthetic antisense internal HIP-specific (5Ј-GTCTGAGGGCACAGGCAGGGGG-TCCTCTCGGGTTTTCTT-3Ј) or HOP-specific (5Ј-TGACATCTTGCAA-GAGTGCTGCTGAGCCCGCTGTGGCTT-3Ј) probes end-labeled with [ 32 P]ATP using T4 polynucleotide kinase to equal specific activities. Blots were apposed to autoradiographic film or analyzed by storage phosphor imaging. 1 Receptor mRNA and Protein-Neuroanatomical, pharmacological, and electrophysiological studies have implicated PACAP peptides among the physiological regulators of sympathetic neuron function (10,(27)(28)(29)(30). 2 The presence of PACAP-selective PAC 1 receptor expression in the cellular elements of the rat SCG was investigated morphologically. Using a riboprobe that did not discriminate the PAC 1 receptor transcript splice variants, hybridization was restricted to SCG principal neurons (Fig. 1A); almost all of the neurons were labeled. PAC 1 receptor immunoreactivity directly paralleled the morphological localization of the receptor transcripts; nearly all of the sympathetic neurons exhibited PAC 1 receptor staining (Fig. 1B). PC12 cells (Fig. 1F), which express PACAP-selective receptor mRNA and respond potently to PACAP, but not VIP (11), also stained, whereas no staining was observed in AtT-20/D16v cells (data not shown), which are regulated by both PACAP and VIP but express VPAC 2 receptor transcripts (35).

Principal Postganglionic Sympathetic Neurons Express PAC
The cellular sites of the PAC 1 receptor protein in the post- ganglionic neurons of the SCG were evaluated using purified primary cultured sympathetic neurons, which exhibited punctate PAC 1 receptor immunoreactivity localized predominantly to the soma (Fig. 1C). The cellular distribution of the PACAPselective receptor protein, examined by fluorescent confocal microscopy, was over the plasma membrane often in patches that may represent receptor clusters (Fig. 1D).
Postganglionic sympathetic SCG neurons are catecholaminergic but are heterogenous with respect to neuropeptide expression. To demonstrate that the subpopulation of SCG neurons synthesizing NPY has the capability to selectively respond to PACAP, the neurons were dually labeled. Approximately 60% of SCG neurons expressed NPY immunoreactivity; all of the NPY-positive neurons coexpressed PAC 1 receptor immunoreactivity (Fig. 1E), implicating PACAP among prominent physiological regulators of sympathetic postganglionic neuron NPY expression.
SCG Neurons Predominantly Express the HOP Isoform of the PAC 1 Receptor-To define the signaling mechanisms underlying PACAP-elicited responses in sympathetic neurons, the molecular forms of PAC 1 receptor mRNA were characterized. Multiple PAC 1 receptor transcript isoforms that differ in either the amino-terminal extracellular region and/or the third intracellular cytoplasmic loop domain are synthesized in a tissuespecific manner by alternative splicing events (Fig. 2). Alternative HIP and/or HOP exon usage in the PAC 1 receptor mRNA encoding the third cytoplasmic loop was examined by reverse transcription-PCR, and amplification of cDNA from adult and neonatal rat SCG identified predominant expression of the one-exon isoform of the PAC 1 receptor mRNA (Fig. 3). Products corresponding to PAC 1 receptor transcripts with neither cassette or with both cassettes represented relatively minor receptor species. Similar to the intact ganglia, purified principal sympathetic neurons maintained in vitro without extrinsic signals expressed predominantly the PAC 1 receptor transcript with one cassette.
Since HIP versus HOP isoforms of the PAC 1 receptor could not be distinguished solely by PCR product size, the identity of the major SCG PAC 1 receptor splice variant resulting from alternative exon usage in the region encoding the third cytoplasmic loop was determined by amplified cDNA sequence-specific hybridization with exon-specific oligonucleotide probes and diagnostic restriction endonuclease analysis. Hybridizations with the HOP-specific oligonucleotide identified one major band, corresponding to the 384/387-nucleotide-amplified product with the HOP cassette insert (Fig. 4A); longer exposures revealed the expression of the much less abundant PAC 1 receptor mRNA containing both HIP and HOP exons. Hybridization with the HIP-specific oligonucleotide probe was revealed only after long autoradiogram exposures. Since the two probes were comparable in size, G/C content, and specific activity, this suggested that the mRNA expression for the onecassette PAC 1 -HIP receptor isoform was much lower than that for the HOP receptor; storage phosphor imaging analyses of the sequence-specific hybridization blots confirmed this relative receptor mRNA expression. Consistent with this finding, the 384/387-nucleotide products were not digested by the HIP exon-specific AvaII restriction enzyme; cleavage with BlpI at a recognition sequence common to HOP1 and HOP2 digested the PCR product to apparent completion (Fig. 4B), suggesting that virtually all of the PAC 1 receptors in the SCG neurons represented the HOP isoform. The HOP1 and HOP2 receptor variants result from alternative usage of two consecutive consensus splice acceptor sites at the 5Ј-end of the HOP exon, and the presence of the three upstream nucleotides in the HOP1 isoform generates a recognition site for the restriction enzyme PvuII. Digestion with PvuII yielded a predominant 328-nucleotide fragment (Fig. 4B), indicating primary expression of the HOP1 receptor transcript variant; the residual uncleaved material most likely represented the HOP2 isoform. Parallel reverse transcription-PCR of SCG cDNA using primers specific for either HIP or HOP cassettes generated amplified products congruous with the predominant expression of PAC 1 -HOP receptor transcripts (data not shown).
SCG Neuronal PAC 1 Receptors Represent the Short Variant-Alternative splicing of both exon 4 (21 nucleotides) and exon 5 (42 nucleotides) produces PAC 1 receptor variants with the presence or absence of a 21-amino acid insert in the aminoterminal extracellular domain (26,44). The short and very short variants were suggested to modulate PACAP27 and PACAP38 PAC 1 receptor binding and potency in stimulating phospho-  lipase C activity. Accordingly, SCG neuron expression of PAC 1 receptor mRNA amino-terminal extracellular domain splice variants was investigated. Reverse transcription-PCR suggested that the sympathetic neurons expressed predominantly the PAC 1 (short) receptor transcript variant containing both exons 4 and 5 encoding the 21 amino acid insert in the amino terminus (Fig. 5). Restriction analyses of the 413-nucleotide material using the exon-specific enzymes yielded cleavage patterns expected for the region of the PAC 1 receptor cDNA containing the 63-nucleotide insert, whereas the transcript variant in guinea pig SCG contains only exon 5. 3 Cleavage with endonucleases with unique recognition sequences in the other exons verified that all of the DNA segments in the fragment were represented (data not shown).

PAC 1 Receptor Transcripts Are Preferentially Expressed in SCG Postganglionic
Neurons-Pharmacologically, SCG neurons appeared to express predominantly the PACAP-selective receptor rather than either of the VIP/PACAP nonselective receptor subtypes (10,30). 2 In contrast to the PAC 1 receptor, amplification of SCG cDNA templates with primers specific for either the VPAC 1 or VPAC 2 receptors did not reveal significant expression of either receptor mRNA in the intact SCG (Fig. 6A). Amplification with VPAC 1 receptor primers, which yielded the predicted product from liver cDNA (data not shown), did not produce amplified fragments from SCG cDNA templates. SCG expression of VPAC 2 receptor mRNA was very low compared with that of the PAC 1 receptor. To establish the cell type expressing the VPAC 2 receptor, purified SCG neuron and nonneuronal cell cultures were prepared by differential plating. The pure postganglionic neurons demonstrated only PAC 1 receptor transcript expression, which correlated well with in situ hybridization and immunocytochemical studies (28,45); the nonneuronal cell cultures, in contrast, appeared to express only VPAC 2 receptor mRNA (Fig. 6B).
PACAP27 and PACAP38 Are Potent and Efficacious Activators of SCG Neuronal cAMP and Inositol Phosphate Production-Distinct PAC 1 receptor isoforms have been postulated to be differentially coupled to adenylyl cyclase and/or phospholipase C (23,26). Based on previous transfection studies, activation of PAC 1 (short)HOP1 receptors in SCG neurons by both molecular forms of PACAP were anticipated to elicit nearly equal high potencies in cAMP production, whereas PACAP38 was expected to be orders of magnitude more potent than PACAP27 in stimulating inositol phosphate production. VIP was not expected to potently stimulate intracellular signaling pathways. Unexpectedly, the patterns of SCG neuron 3 K. M. Braas, unpublished data.

FIG. 4. The alternatively spliced third cytoplasmic loop PAC 1 receptor isoform containing one cassette is primarily the HOP1 variant.
A, the PAC 1 receptor products amplified from adult SCG cDNA using the oligonucleotide primers PACAPR1 and PACAPR2 were hybridized with either HIP or HOP exon-specific internal radiolabeled oligonucleotide probes; the blots were apposed to film for 2 h (short exposure, upper autoradiograms) or 44 h (long exposure, lower autoradiograms). B, the 384/387-amplified nucleotide products were isolated for diagnostic restriction endonuclease digestion using isoform-specific enzymes. Control represents the undigested amplified 384/387-base pair product. The receptor mRNA schematic diagram is described in the legend to Fig. 3.  Table I) and can produce fragments of the indicated sizes with both (short) or neither (very short) exons 4 and 5 (left). The amplified product was recovered from the agarose gels for diagnostic restriction analyses with exon 4-and 5-specific enzymes, FokI and Tsp509I, respectively. The PAC 1 receptor mRNA schematic diagram is described in the legend to Fig. 3. The numbers in the diagrams for the restriction enzyme cleavage sites refer to the exons represented in the amplified product (44). Thick line, region amplified using PACAPR3 and PACAPR4.
FIG. 6. SCG postganglionic neurons express PAC 1 receptors but neither of the nonselective VIP/PACAP receptors. A, total RNA from individual adult SCG was reversed transcribed, and the cDNA templates were amplified using oligonucleotide primers specific for the PAC 1 , VPAC 1 , or VPAC 2 receptors (Table I). B, total RNA (2 g) from highly enriched SCG neuronal and nonneuronal cells produced by differential plating were amplified using primers for PAC 1 and VPAC 2 to identify the cellular sources of these receptor types. Representative of data from three separate experiments. PAC 1 (short)HOP receptor activation of second messengers differed from those previously observed for this particular receptor variant. Both PACAP27 and PACAP38 potently and efficaciously stimulated sympathetic neuron cAMP production. The potency of PACAP38 was approximately 10-fold greater than PACAP27; half-maximal stimulation of cAMP production with PACAP38 was observed at 0.3 nM, and the half-maximal effect with PACAP27 was 3 nM (Fig. 7A). However, PACAP27 appeared more efficacious than PACAP38 in stimulating cAMP production; the maximal level of PACAP27-stimulated SCG cAMP production was approximately 1.4-fold greater than that for PACAP38. As predicted by expression of the PACAP-selective receptor type, VIP was over 1000-fold less potent than the PACAP peptides in adenylyl cyclase activation.
While previous reports suggested that only PACAP38 potently stimulated phospholipase C, both PACAP27 and PACAP38 potently stimulated inositol phosphate production in sympathetic neurons with identical half-maximal effects of 0.5 nM peptide (Fig. 7B). Similar to the stimulation of SCG cAMP production, PACAP27 appeared more efficacious than PACAP38 in phospholipase C activation. As expected for PACAP-selective receptor activation of intracellular signaling pathways, micromolar concentrations of VIP were also required to stimulate inositol phosphate production.
In heterologous inhibitor studies, treatment of the neurons with the adenylyl cyclase inhibitor SQ22536 did not alter either basal or PACAP-stimulated inositol phosphate production (Fig.  8B, black bars); SQ22536, furthermore, did not alter the ability of the phospholipase C inhibitor U73122 to diminish inositol phosphate levels (light gray bars). In contrast, U73122 inhibition of phospholipase C not only failed to attenuate, but potentiated, PACAP peptide-stimulated cAMP production; SCG cAMP levels elicited by PACAP38 and PACAP27 in the presence of U73122 were increased 150% compared with either peptide alone (p Ͻ 0.001; Fig. 8A, dark gray bars). These results suggest that SCG PAC 1 (short)HOP1 receptor activation of these two signaling pathways resulted from distinct G-protein interactions, but furthermore, activation of the phosphatidylinositol pathway may be a means of modulating the intracellular rise in SCG cAMP levels following PACAP activation.
Sympathetic Neuron PACAP-stimulated NPY Release Is Reduced by Phospholipase C Inhibition-The contributions of the cAMP and phosphatidylinositol pathways to SCG neuropeptide release were examined under identical conditions to those used for the analysis of PAC 1 receptor coupling to intracellular signaling cascades. Incubation of primary cultured SCG neurons with 100 nM PACAP27 or PACAP38 potently and efficaciously stimulated NPY release approximately 10-fold compared with basal levels (p Ͻ 0.001; Fig. 9, open bars). NPY release was increased from 24 fmol of NPY/10 4 cells in control neurons to 220 fmol of NPY/10 4 cells and 260 fmol of NPY/10 4 cells with PACAP27 and PACAP38 treatment, respectively. Basal NPY secretion was not affected by cAMP or phosphatidylinositol pathway inhibitors. The pattern of PACAP-stimulated NPY release under the inhibitor paradigm, however, paralleled the inositol phosphate production profile. Inhibition of adenylyl cyclase activity with SQ22536 failed to diminish the PACAPinduced increase in neuron NPY release (Fig. 9, black bars). In contrast, U73122 attenuated PACAP-stimulated NPY secretion commensurate with decreased inositol phosphate production (p Ͻ 0.001; dark gray bars). The decreased receptor-mediated peptide release produced by U73122 was not altered by SQ22536 (light gray bars), suggesting that the PACAP-stimulated NPY secretion was regulated predominantly by the phosphatidylinositol signaling pathway.

DISCUSSION
The present studies demonstrate that SCG postganglionic neurons express predominantly the PAC 1 (short)HOP1 receptor splice variant coupled to both adenylyl cyclase and phospholipase C. Moreover, coupling of this receptor isoform to inositol phosphate production contributes to PACAP-stimulated sympathetic neuron NPY release. Unlike many tissues that express not only multiple PAC 1 receptor isoforms, but also VPAC 1 and/or VPAC 2 receptors (17,18,23,25,51), our studies demonstrated high SCG sympathetic postganglionic neuron expression of only the PAC 1 (short)HOP1 receptor splice variant. In contrast, we identified low levels of VPAC 2 receptor mRNA in SCG nonneuronal cells, whereas we and others reported that VPAC 1 receptor mRNA is not expressed in the SCG (45). Thus, PACAP and VIP have different cellular targets in the SCG, resulting in potentially important functional differences. PACAP27, PACAP38, and VIP not only modulate neuron functions but also nervous system nonneuronal cell cytokines and growth factors (52)(53)(54)(55)(56)(57). The high potencies of the PACAP peptides at both PAC 1 and VPAC 2 receptors suggest that these peptides may possess regulatory functions on both neuronal and nonneuronal cells in the SCG, while the roles of VIP are predicted to be restricted to nonneuronal cell VPAC 2 receptor activation. Under conditions that result in increased sympathetic neuron PACAP and/or VIP expression, such as neuronal injury (58,59), activation of specific receptor subtypes on SCG neuronal and nonneuronal cells may contribute to the neuronal repair response. PAC 1 receptor activation of both adenylyl cyclase and phospholipase C is characteristic of the group III family of Gprotein-coupled receptors (44,60). Alternative usage of the HIP and/or HOP exons in the region of the PAC 1 receptor transcript encoding the third cytoplasmic loop, a region important for receptor-G-protein interaction, has been suggested to deter-mine receptor signaling pathway activation selectivity (23,26,33). Furthermore, expression of variants differing in the presence (short) or absence (very short) of exons 4 and 5 in the amino-terminal extracellular domain has been suggested to modulate PACAP27 and PACAP38 binding and discriminate peptide potency in second messenger production (26). The corresponding amino-terminal region in the VPAC 1 receptor contains essential amino acids defining the VIP binding domain (61)(62)(63). Identification of the PAC 1 receptor isoforms expressed by SCG neurons and the physiological coupling of these receptors to intracellular signaling pathways was therefore fundamental to understanding the mechanisms of sympathetic neuron responses to PACAP.
In contrast to most nervous tissues, which express prevalently the PAC 1 (short) receptor transcript variant with neither HIP nor HOP exons (23,33), SCG neurons express predominantly the PAC 1 (short)HOP1 receptor isoform. PACAP27 and PACAP38 demonstrated comparable high potency in augmenting cAMP levels in transfected LLCPK1 cells expressing the PAC 1 (short)HOP receptor, but PACAP27 was at least 100-fold less potent than PACAP38 in stimulating inositol phosphate production; moreover, PACAP38 stimulated intracellular inositol phosphate ϳ40-fold less potently than cAMP (23,26). Similar patterns of second messenger activation were demonstrated in adrenal medullary chromaffin or PC12 pheochromocytoma cells (11,64,65). By contrast, PACAP-mediated second messenger responses in SCG neurons through the PAC 1 (short)HOP1 receptor were unique in that both PACAP27 and PACAP38 exhibited high potency in stimulating cAMP and inositol phosphate production. In accord with the pattern of activation predicted for the HOP variant (23), both peptides potently stimulated cellular adenylyl cyclase; unexpectedly, PACAP38 was ϳ10-fold more potent than PACAP27. Importantly, in departure from these reports, PACAP27 and PACAP38 demonstrated equal subnanomolar high potencies in augmenting inositol phosphate production in sympathetic neurons. In addition, the concentration dependence of PACAPinduced inositol phosphate production was characteristic of the PAC 1 (very short) receptor (26) but was unanticipated for the PAC 1 (short) variant expressed by SCG postganglionic neurons. These results implied that alternative splicing of the aminoterminal putative peptide binding domain may not represent the only or major determinant of PACAP27 potency in phospholipase C stimulation and suggested that other factors, such as receptor density or expression of specific cell signaling components, may ultimately dictate the cellular responses (60).
Coupling of one receptor to multiple intracellular signaling cascades poses complex issues related to signal integration, and studies have shown multiple potential points of intersection among different signaling pathways. For example, agonist-stimulated increased IP 3 production also has been shown to enhance cAMP formation by a calcium-calmodulindependent mechanism (66,67). Contrary to expectation, inhibition of phospholipase C augmented PACAP-mediated cAMP production. Several phosphodiesterase isoforms exhibit serine/threonine kinase-dependent activities, and decreased protein kinase C activity following phospholipase C inhibition may diminish phosphodiesterase activities, resulting in an apparent elevation in cellular cAMP levels (68,69). Alternatively, expression of specific adenylyl cyclase isoforms by SCG neurons may be important. AC9, prevalent in many neuronal tissues, is inhibited by intracellular calcium (70), and decreased calcium mobilization following phospholipase C inhibition could result in sustained AC9 activity and enhanced cAMP production. The phosphorylation state of adenylyl cyclase or PAC 1 receptors following phospholipase C FIG. 9. Sympathetic neuron PACAP-stimulated NPY secretion is regulated by the phospholipase C signaling pathway. Primary cultured SCG neurons were incubated as described for Fig. 7 with defined medium containing vehicle or 100 nM PACAP27 or PACAP38; neurons under these conditions were treated with vehicle (white bar) or a 10 M concentration of the inhibitor SQ22536 (black bar), U73122 (dark gray bar), or SQ22536 plus U73122 (light gray bar), and the conditioned medium was analyzed for NPY immunoreactivity. Data represent the mean NPY released Ϯ S.E. (n ϭ 4 assayed in triplicate). inhibition may also affect total cellular cAMP production. Whether any of these mechanisms underlie the observed interactions between PACAP-mediated second messenger products remains to be investigated.
We demonstrated previously that direct activation of either the cAMP or inositol phosphate signaling pathways stimulates sympathetic neuron transmitter and peptide secretion (34); consequently, PACAP stimulation of PAC 1 (short)HOP1 receptor-mediated neurosecretion (30) was predicted to represent a synergistic response of the two signaling cascades. Inhibition of adenylyl cyclase did not affect sympathetic neuron PACAPmediated NPY release; by contrast, phosphatidylinositol signaling, presumably through inositol 1,4,5-trisphosphate-mediated intracellular calcium release, appeared to represent the predominant means of PACAP-stimulated sympathetic neurosecretion. Similarly, U73122 attenuated tachykinin-stimulated adrenal gland corticosteroid and aldosterone secretion and endothelin-1-and GnRH-induced luteinizing hormone release from gonadotropes (49,50). PACAP, primarily through PAC 1 (very short) receptors, also stimulates release of acetylcholine from cardiac ganglia parasympathetic neurons amplifying cardiac ganglia parasympathetic inhibition, but the intracellular signaling mechanisms coupled to these receptormediated events are unknown (40,71). PACAP peptides also appear to be noncholinergic secretagogues of adrenal medullary cell peptides and catecholamines (65,72,73). Activation of chromaffin cell PAC 1 HOP1 receptors stimulates cAMP and inositol phosphate production; however, the contributions of these signaling pathways to release remains unclear (9, 74 -77).
In summary, molecular characterization has shown that SCG sympathetic postganglionic neurons express predominantly the PAC 1 (short)HOP1 receptor isoform. Atypically, PACAP27 and PACAP38 binding to this sympathetic neuron receptor variant results in potent and efficacious stimulation of both adenylyl cyclase and phospholipase C activities; activation of the latter appears to be a prominent component to the neuronal secretory response. In the VIP/PACAP family of peptides, these studies demonstrate the specific and preferential actions of PACAP and PAC 1 receptors in sympathetic autonomic physiology.