Regulation of Type VI Adenylyl Cyclase by Snapin, a SNAP25-binding Protein* □ S

In the present study, we used the N terminus (amino acids 1 (cid:1) 160) of type VI adenylyl cyclase (ACVI) as bait to screen a mouse brain cDNA library and identified Snapin as a novel ACVI-interacting molecule. Snapin is a binding protein of SNAP25, a component of the SNARE complex. Co-immunoprecipitation analyses confirmed the interaction between Snapin and full-length ACVI. Mutational analysis revealed that the interaction domains of ACVI and Snapin were located within amino acids 1 (cid:1) 86 of ACVI and 33–51 of Snapin, respectively. Co-localization of ACVI and Snapin was observed in primary hippocampal neurons. Moreover, expression of Snapin specifically eliminated protein kinase C (PKC)-mediated suppression of ACVI, but not that of cAMP-de-pendent protein kinase (PKA) or calcium. Mutation of the potential PKC and PKA phosphorylation sites of Snapin did not affect the ability of Snapin to reverse the PKC inhibitory effect on ACVI. Phosphorylation of Snapin by PKC or PKA therefore might not be crucial for Snapin action on ACVI. In contrast, Snapin (cid:1) 33–51 , which harbors an internal deletion of amino acids 33–51 did not affect PKC-mediated inhibition

lation. To date, at least 9 membrane-bound ACs have been isolated and characterized (1). These enzymes are capable of integrating positive and negative signals that act directly through stimulation of G protein-coupled receptors (GPCRs) or indirectly via intracellular signaling molecules in isozyme-specific patterns. In addition, the regulatory properties and expression patterns of different AC isoforms greatly diverge and may account for the distinctive cell-and tissue-specific responsiveness of ACs. Recently, several different proteins, including RGS2 and the protein associated with Myc (PAM), have been shown to interact and modulate activity of different AC isozymes (2,3), adding additional dimensions to the isozymespecific regulation of the AC superfamily.
Except for the newly identified soluble AC, all membranebound AC members share a primary structure consisting of 12 transmembrane regions and 3 large cytoplasmic domains (N, C1a/b, and C2). The C1a and C2 domains, which form the catalytic core complex, are highly conserved and are homologous to each other. The N-terminal domains of ACs, in contrast, are variable among ACs, and have been demonstrated to play mostly regulatory roles (4,5). Among the AC isozymes, ACVI is of particular interest, because this enzyme can be suppressed by several intracellular signaling molecules including Ca 2ϩ , cAMP-dependent kinase (PKA), and protein kinase C (PKC) (6 -8). Although ACVI has been found in many tissues examined (9), its expression in the brain is mainly detected in neurons (10). ACVI therefore might play a critical role in integrating multiple signal inputs in neurons. We previously demonstrated that the N-terminal domain of ACVI (designated N 1-160 ) is critical for PKC-mediated suppression of ACVI (5). Using N  as the bait to screen a mouse brain cDNA library, we report herein that Snapin is a novel ACVI-interacting protein. Snapin has been found on synaptic vesicle membranes and was originally identified as a novel binding protein of SNAP-25, a plasmalemmal component of SNAREs (soluble Nethylmaleimide-sensitive factor attached protein receptors) (11,12). Assembly of the SNAREs complex is the key step for docking and fusion between vesicles and target membranes. Snapin was found to modulate neurotransmitter release by enhancing the interaction between the SNAREs complex and synaptotagmin, a putative calcium sensor, through direct interaction with SNAP-25 (12). In addition, Snapin is a physiological PKA substrate and can modulate neurotransmitter secretion by the cAMP-dependent pathway (13). Similar to the expression profile of Snapin in the brain, we previously demonstrated that ACVI is widely expressed in many brain regions (10). Double immunostaining of ACVI and Snapin shown in the present study reveals the colocalization of ACVI and Snapin in the brain and in primary hippocampal neurons, supporting the physiological relevance of this interaction. Most importantly, * This work was supported by Grants NSC89-2320-B001-011 and NSC90 -2320-B001-009 from the National Science Council and Grants NHRI-EX92-9203NI and NHRI-EX93-9203NI from the National Health Research Institutes and Academia Sinica, Taipei, Taiwan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. PKC loses its ability to suppress ACVI activity in the presence of Snapin. These findings demonstrate, for the first time, a fine-tuning mechanism for the cAMP production system by the vesicle-transporting machinery in the brain.

EXPERIMENTAL PROCEDURES
Materials-Forskolin, cAMP, ATP, prostaglandin E1, thapsigargin, and essential amino acids were obtained from Sigma. Purified PKA, thapsigargin, and PKC were purchased from Calbiochem (La Jolla, CA). CellTracker TM CM-DiI was obtained from Molecular Probes (Eugene, OR). MgCl 2 and other chemicals were obtained from Merck (Darmstadt, Germany) unless stated otherwise.
Yeast Two-hybrid Experiments and Library Screening-The cDNA of rat ACVI (9) was kindly provided by Dr. R. Iyengar (Department of Pharmacology, Mount Sinai School of Medicine). The DNA fragment encoding the N-terminal domain of ACVI (amino acids 1ϳ160, N 1-160 ) was amplified by the polymerase chain reaction (PCR; see Supplemental Materials, Table S-1), subcloned into the pAS2-1 vector, and served as the bait to screen a mouse brain cDNA library using the Matchmaker Two-hybrid System 2 (Clontech, Palo Alto, CA) following the manufacturer's instructions. The identified cDNA clones were analyzed by nucleotide sequencing and functional annotation using standard bioinformatic programs.
In Vitro Binding-Recombinant N 1-160 , N 1-123 , N 1-86 , and Snapin variants were produced in the presence of [ 35 S]methionine (Amersham Biosciences) using the TNT® quick-coupled transcription/translation systems (Promega, Madison, WI) following the manufacturer's protocol. An in vitro binding assay was performed by incubating the reaction mixture containing [ 35 S]methionine-labeled recombinant proteins (25ϳ60 fmol) at 30°C for 1 h, followed by addition of 30 l of the AC6N antibody (10)-protein A complex (Sigma) and incubation at 4°C for 2 h with gentle agitation. The immunocomplexes were purified, extensively washed, and resolved on 15% SDS-polyacrylamide gels (15). Detection of 35 S-labeled proteins was enhanced by incubating the fixed gels in EN 3 HANCE (NEN TM Life Science Products, Zaventem, Belgium) for 30 min. Gels were dried and autoradiographed. Recombinant MBP and MBP-Snapin proteins were expressed in Escherichia coli and were purified using Amylose resin (New England BioLabs, Beverly, MA) following the manufacturer's protocol. For the interaction between MBP-Snapin and N 1-160 , recombinant MBP Snapin (2.5 g) or MBP (10 g) was first bound to amylose resin (30 l) and then incubated with the TNT lysates containing [ 35 S]methionine-labeled N 1-160 proteins (0.003 pmol) in 400 l of binding buffer (20 mM Tris, 200 mM NaCl, 1 mM EDTA, 1 mM sodium azide, and 1 mM ␤-mercaptoethanol) for 1 h at 30°C. The complexes bound on the amylose resins were extensively washed with binding buffer and analyzed using 15% SDS-PAGE gels. The halves of gels that contained proteins larger than 35 kDa were transferred to polyvinylidene difluoride membranes, and the levels of MBP-Snapin and MBP were analyzed by the Western blot analysis using anti-MBP antiserum (New England BioLabs; 1: 10,000 dilution) and anti-SNA-c (1:5,000 dilution), respectively. The other halves of the gels, which contained proteins smaller than 35 kDa were fixed and autoradiographed as described above. Note that an excess amount of MBP was used as a negative control to ensure the specificity of the binding between MBP-Snapin and N  .
Cell Culture and Transfection-Primary neuronal cultures were prepared from Sprague-Dawley rat brains at embryonic day 18ϳ19 as described elsewhere (16) and grown in Neurobasal medium supplemented with B27 (Invitrogen Life Technologies). At 8ϳ9 days in vitro (DIV), cultures were washed twice with KRH buffer (125 mM NaCl, 4.8 mM KCl, 1.3 mM CaCl 2 , 1.2 mM MgSO 4 , 1.2 mM KH 2 PO 4 , 5.6 mM glucose, and 10 mM HEPES, pH 7.2), and then treated with KRH buffer or KRH buffer containing 56 mM KCl (Hi-K ϩ ) for 5 min at room temperature. At the end of stimulation, cells were immediately fixed in 4% paraformaldehyde/4% sucrose for 20 min, and then were ready for double or triple label immunostaining. The CHOP cell line (17) was a generous gift from Dr. J. W. Dennis (Samuel Lunenfeld Research Institute, Toronto, Ontario, Canada) and was maintained in Dulbecco's modified Eagle's medium (HyClone, Logan, UT) supplemented with 10% fetal bovine serum (HyClone) in 10% CO 2 and 90% air. For transient transfection experiments, CHOP cells were transfected with the desired construct using the DEAE-dextran method (18) or LipofectAMINE 2000 TM (Invitrogen) following the manufacturer's protocol. Cells were harvested 72 h post-transfection for analysis. HEK293T cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) fetal bovine serum and 2 mM glutamine in an incubation chamber supplied with 5% CO 2 and 95% air at 37°C. HEK239T cells were transfected with the desired construct(s) at 3 ϫ 10 6 cells in 100-mm culture plates using LipofectAMINE 2000 TM as described above.
Production and Biotinylation of the Polyclonal Anti-Snapin Antibody-The oligopeptide (S 118 -136 , CARRRAMLDSGVYPPGSPSK, corresponding to amino acids 118ϳ136 of mouse Snapin) was purchased from Genosys (The Woodlands, TX) and conjugated to bovine serum albumin (BSA, Sigma) using m-maleimidobenzoyl-N-hydroxysuccinimide ester (MBS) as described by Harlow et al. (19). The antiserum was generated by injecting male New Zealand White rabbits with the BSAconjugated peptide using standard procedures (19). The resultant antibody was designated SNA-c. To remove the potentially existing anti-BSA IgG, SNA-c antiserum was preabsorbed with 1% BSA in phosphate-buffered saline at 4°C overnight. To characterize the specificity of SNA-c, the antigen peptide was conjugated to an irrelevant protein, ovalbumin, using MBS as described above and used for competition experiments. The addition of an excess amount of the peptide antigen (S 118 -136 ) conjugated to an irrelevant protein (ovalbumin; 1 mg/ml) completely abolished the immunoreactivity of Snapin in the brain and caused the disappearance of an immunoreactive band of 15 kDa, the predicted molecular mass of Snapin, in the plasma membrane fractions of the rat hippocampus ( Fig. S-1; Supplemental Materials). For double immunohistochemical staining, the SNA-c antibody was biotinylated. Briefly, the SNA-c antibody was first purified by a protein A-Sephadex column, and then incubated with NHS-LC-Biotin (3 mg/ml; Pierce) at a 1:15 molar ratio. Immunohistochemical staining of the rat brain using the biotinylated SNA-c antibody exhibited an identical expression profile of Snapin as that by the original SNA-c antiserum in rat brain (data not shown).
Immunoprecipitation and Western Blotting-Expressions of wildtype and mutant ACVI were carried out in a recombinant baculovirusdriven Sf21 cell system following the manufacturer's protocol (BD PharMingen, San Diego, CA) as described elsewhere (8). ACVI proteins were purified by immunoprecipitation with an anti-ACVI antibody (AC6D) as described previously (8). Briefly, membrane fractions (0.3 mg) collected from Sf21 cells infected with the desired ACVI virus were solubilized with 0.25 ml of ice-cold radioimmune precipitation assay buffer (50 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris, pH 8, 40 M phenylmethylsulfonyl fluoride, and 100 M leupeptin) at 4°C for 2 h, and then incubated with 50 l of the AC6D/protein A complex or the antigen-pre-absorbed AC6D/protein A complex at 4°C for 90 min with gentle agitation. After extensive washing, enriched ACVI proteins were mixed with [ 35 S]Snapin (0.005 pmol) in 50 l of binding buffer (150 mM NaCl, 50 mM Tris, pH 7.5, 20 mM MgCl 2 , 40 M phenylmethylsulfonyl fluoride, and 100 M leupeptin) at 30°C for 1 h to allow complex formation. Immunoprecipitates were analyzed on 8% separating gels. The half of gels that contained proteins larger than 50 kDa were transferred to polyvinylidene difluoride membranes and utilized to determine the expression of ACVI using AC6D (1:5000 dilution; 8). The other half of the gels, which contained proteins smaller than 50 kDa, were fixed, enhanced using En 3 HANCE, and autoradiographed as described above.
For triple immunostaining with ACVI, Snapin, and MAP2 in primary hippocampal neurons, cells were blocked with 3% bovine serum albumin for 1 h, stained with a monoclonal anti-MAP2 antibody (1: 1000, Chemicon, Temecula CA) at 4°C overnight, and then incubated in a goat anti-mouse IgG conjugated to Cy5 (1:200, Jackson Immu-noResearch, West Grove, PA) for 2 h. After extensive rinsing, staining of ACVI, and Snapin was carried out as described above. For the triple immunostaining with ACVI, Snapin, and DiI, staining with ACVI and Snapin was carried out as described above. Cells were then stained with DiI (10 ng/l) in phosphate-buffered saline for 1 h at room temperature, followed by extensive washing with phosphatebuffered saline to remove excess DiI. Patterns of double or triple immunostaining were analyzed with the aid of a laser confocal microscope (Bio-Rad, MRC-1000). Controls for the specificity of immunofluorescence were determined by omission of the primary antibody from the staining process.
AC Assay and cAMP Measurement-AC activity was assayed as previously described (21). The enzyme activity was linear for up to 30 min with up to 40 g of membrane protein. The ACVI activity was determined as the difference between cyclase activities measured in the membrane fractions collected from CHOP cells transfected with the indicated ACVI expression construct and those from cells transfected with the empty vector alone. For CHOP cells expressing wild-type ACVI, endogenous cyclase activities represented ϳ40% of the total activity. The absolute values of ACVI activity observed in different transient transfection experiments might have slightly varied due to the number of passages and condition of cells. Nevertheless, the overall pattern of AC regulation was consistent among experiments.
Intracellular cAMP content was determined as described before (21,22) with slight modification. Cells were washed twice with the Ca 2ϩfree Locke's solution (150 mM NaCl, 5.6 mM KCl, 5 mM glucose, 1 mM MgCl 2 , and 10 mM HEPES adjusted to pH 7.4) containing 0.5 mM IBMX and resuspended in the same solution at 1 ϫ 10 5 cells/0.45 ml in each 1.5-ml tube. The effect of calcium on ACVI activities was determined as reported elsewhere (23). In brief, cells were washed once with calciumfree Locke's solution and pretreated with thapsigargin (1 M) for 3 min, followed by EGTA (2.5 mM) for 7 min, then stimulated with calcium-free Locke's solution containing forskolin (10 M) plus prostaglandin E1 (10 M) for 10 min at room temperature in the presence or absence of CaCl 2 (10 mM). Cellular cAMP was extracted by adding 0.3 ml of 0.1 N HCl to each tube with gentle mixing for 10 min on ice. The cAMP content was assayed using the 125 I-cAMP assay system (Amersham Biosciences). cAMP accumulation resulting from activation of ACVI was determined as the difference between cAMP levels measured in HEK293T cells expressing wild-type ACVI versus an empty vector (pCEP4).

FIG. 1. Snapin interacts with ACVI in vitro.
A, interaction between the N terminus of ACVI (N 1-160 ) and Snapin in the yeast two-hybrid system. Y187 yeast transformed with the indicated plasmids in selection plates (-Trp, -Leu) is shown. The black color represents the expression of the reporter (␤-galactosidase). Co-transformation of pVA3-1 and pDT1-1 provided by the manufacturer was included as a positive control. Note that the co-transformation of neither Snapin and pVA3-1, N 1-160 and pDT1-1, nor N 1-160 and an empty vector (pACT2) caused the expression of ␤-galactosidase. B, co-immunoprecipitation of N 1-160 with Snapin. Recombinant N 1-160 , Snapin, and uORF5 (an irrelevant protein) were produced using the in vitro TNT system in the presence of [ 35 S]methionine. Production of these proteins was visualized by loading 1 l of each TNT reaction mixture (0.001ϳ0.002 pmol) into each lane as shown on the 3 right-most lanes. For co-immunoprecipitation, recombinant N 1-160 (0.007 pmol) and the indicated proteins (0.002ϳ0.004 pmol) were incubated for 60 min at 30°C to allow complex formation. Immunoprecipitation was performed using anti-N 1-160 antiserum (IP with AC6N, 10) or normal rabbit serum (IP with NRS) as indicated. The bound proteins were separated by 15% SDS-PAGE, dried, and autoradiographed. C, co-immunoprecipitation of full-length ACVI with Snapin. ACVI was produced in Sf21 cells and purified immunologically with an anti-ACVI antibody (AC6D, 8) or AC6D pre-absorbed with the antigen (recombinant C2 domain containing amino acids 987ϳ1187 of ACVI; designated CON). For co-immunoprecipitation, purified ACVI and [ 35 S]Snapin (0.005 pmol) were incubated for 60 min at 30°C to allow complex formation and were analyzed by Western blot analysis to determine the levels of ACVI (top panel), and by autoradiography to visualize [ 35 S]Snapin (bottom panel). Note that the 2 ACVI-immunoreactive bands in the top panel respectively represent glycosylated and unglycosylated ACVI proteins in Sf21 cells (41).

Snapin Interaction with the N-terminal Domain of ACVI-
Although the N terminus is variable among different AC members, it is highly conserved between species. The N terminus therefore might mediate developmentally conserved regulatory modes of ACVI. Compared with other members in the AC superfamily, the N terminus of ACVI with 160 amino acids is relatively large and might mediate its functions through interacting with other proteins. To search for proteins that interact with the N-terminal domain of rat ACVI (designated N 1-160 ) in the brain, we used N 1-160 as the bait to screen a mouse brain cDNA library using the yeast two-hybrid system. This library was chosen because of its availability and because the N termini of rat and mouse ACVI are highly homologous (96% identify in amino acids). Out of 2 ϫ 10 5 clones, 2 independent clones encoding a full-length and a partial fragment (amino acids 33ϳ136, designated SNP  ) of Snapin were identified. Note that the amino acid sequences of mouse and rat Snapin are identical. Such an interaction was verified by co-transforming the identified pACT2-Snapin construct (full-length or the truncated Snapin 33-136 ) with the original N 1-160 bait into yeast Y187 cells. As shown in Fig. 1A, co-transformation of the Snapin variant and N 1-160 led to expression of the reporter gene (␤-galactosidase), confirming the interaction between Snapin and N 1-160 in yeast. In vitro binding of the recombinant N 1-160 and Snapin was demonstrated by co-immunoprecipitation using an anti-ACVI antibody (AC6N). As expected, Snapin, but not an unrelated UORF5 protein, was co-immunoprecipitated with N 1-160 (Fig. 1B). This interaction between N 1-160 and Snapin was specific because in the absence of N 1-160 , no Snapin was detected in the immunoprecipitated complex (Fig.  1B). Moreover, full-length ACVI prepared from Sf21 cells also bound Snapin (Fig. 1C). Collectively, these data demonstrate that ACVI interacts with Snapin through direct interaction at its N-terminal domain.
To define the Snapin-interacting site of N 1-160 , we performed in vitro binding analyses of the recombinant Snapin and various lengths of the N-terminal domain of ACVI ( Fig. 2A). Recombinant proteins were produced using the in vitro TNT system in the presence of [ 35 S]methionine. Complex formation was identified by immunoprecipitation using antiserum (AC6N) against N 1-160 . As demonstrated in Fig. 2A, the fragment comprising amino acids 1-86 (N 1-86 ) was effectively co-immunoprecipitated with the recombinant Snapin as were the fulllength N terminus (N 1-160 ) and N 1-123 . The Snapin-interacting region of N 1-160 therefore might reside in the region containing amino acids 1ϳ86.
We next performed experiments to determine what portion on Snapin N 1-160 was bound using the in vitro binding assay as described above. Five recombinant proteins comprising different portions of mouse Snapin (amino acids 26ϳ136, 33ϳ136, 51ϳ136, 1ϳ95, and 1ϳ115, designated SNP 26 -136 , SNP 33-136 , SNP 51-136 , SNP 1-95 , and SNP 1-115 , respectively; Fig. 2B) were prepared in vitro. As shown in Fig. 2, B and C, N 1-160 immunoprecipitated all of the Snapin variants examined, except for SNP 51-136 . Since SNP  retained the ability to interact with N 1-160 , amino acids 33ϳ51 of Snapin appeared to be crucial for the interaction with N 1-160 . Furthermore, addition of a synthetic peptide comprising amino acids 33ϳ51 of Snapin (designated S 33-51 ) abated the interaction between Snapin and N 1-160 (Fig. 2, D and E), confirming that the interactive domain of Snapin with N 1-160 resides in the region containing amino acids 33ϳ51 of Snapin. The peptide containing amino acids 118ϳ136 of Snapin (designated S 118 -136 ) did not exert a significant effect on the binding of Snapin to N 1-160 and thus served as a control peptide in this experiment (Fig. 2, D and E). Collectively, the ACVI-binding region of Snapin is located within amino acids 33ϳ51, a region different from that which was used to interact with SNAP25 (12) or SNAP23 (24).
Colocalization of ACVI and Snapin-We previously demonstrated that in the brain, ACVI is expressed in most areas examined and is present in cells with a neuronal phenotype (10). Since the regional expression of Snapin in the central nervous system (CNS) has not yet been reported, we first set out to examine the expression of Snapin in the brain. Detailed analysis revealed that similar to the expression of ACVI in the brain, Snapin was widely expressed in many brain areas with different intensities (Supplemental Materials, Fig. S-1B-G). Double immunohistochemical staining further demonstrated that ACVI was colocalized with Snapin in the adult rat brain (Supplemental Materials, Fig. S-1H-M), and in primary hippocampal neurons (Fig. 3). Importantly, depolarization of these hippocampal neurons by external high K ϩ (56 mM, 5 min) did  not alter the colocalization of ACVI and Snapin (Fig. 3, D-N), which could readily be observed in somas and proximal dendrites. Consistently, in vitro binding of ACVI and Snapin determined by co-immunoprecipitation could be observed both in the absence (Fig. 1C), and presence of calcium (data not shown). Interactions between ACVI and Snapin therefore appeared to be independent of the neuron's activity. The ACVI and Snapin complexes were also colocalized with MAP2 (Fig.  3I), a somatodendritic marker, indicating their somatodendritic localization. In addition to the plasma membrane of somas (Fig. 3N), co-localization of Snapin and ACVI on the tips of growth cones could also readily be observed (Fig. 3, J-M).
Such an interaction between these two molecules therefore might play an important role in regulating neuronal activity. Although adenylyl cyclases in general are supposed to exist in plasma membranes, we consistently found that ACVI could also be detected in intracellular compartments (Ref. 10 and this study). Similarly, in cultured hippocampal neurons, immunoreactivities of ACI and ACVIII were found in various subcellular locations, in addition to plasma membranes (25,26). Transport of ACs therefore might be an important regulatory role for the AC family. This is of particular importance since the ACVI-interacting Snapin is known to associate with the SNARE complex, which plays an important role in intracellular membrane cycling. It is therefore possible that Snapin might mediate the cycling of ACVI through interactions with the SNARE complex.
Many neuromodulators, including those that elevate cAMP contents, have been shown to rapidly modulate synaptic strength. We previously showed that ACVI protein is widely expressed in the brain and may participate in regulation of the classical neurotransmitter systems (10). Interactions between ACVI and Snapin as reported herein further support the potential contribution of ACVI in synaptic transmission. Since the cAMP-dependent pathway has been implicated in neurotransmitter release (27), and because phosphorylation of Snapin by PKA enhances its interaction with SNAP25 and synaptotagmin of the assembled SNARE complex (13), direct interactions between Snapin and a cAMP-synthesizing enzyme (ACVI) might set the stage for enhancement of neurotransmitter release by cAMP. Moreover, a somatodendritic expression pattern of Snapin was found in several brain areas and in primary hippocampal neurons (Supplemental Materials, Fig. S-1C-G; Fig. 3G-I). Thus, Snapin/ACVI complexes might also possess certain postsynaptic functions. For example, SNAREs complexes are critical for membrane trafficking (28), and therefore might mediate the incorporation of postsynaptic proteins (such as AMPA receptors) into synapses to facilitate synaptic plasticity (29). More importantly, ample evidence suggests that the cAMP system plays a key role in neuronal plasticity. Specifically, phosphorylation of AMPA receptors by PKA has been shown to regulate the insertion of AMPA receptors into synapses and to contribute to plasticity (30,31). An association of cAMP pathway components (e.g. AKAP and PKA) with receptors (e.g. AMPA receptors) important for synaptic strength has also been reported at postsynaptic sites (32). Interactions between ACVI and Snapin therefore might provide a previously uncharacterized mechanism to enhance the neuronal plasticity regulated by the cAMP pathway.
During the preparation of this article, several studies that call for the reinvestigation of the role and expression of Snapin were published (24,33,34). In contrast to a brain-specific expression pattern (12), Snapin was reported to be a ubiquitously expressed soluble protein, which can be detected in both cytosol and peripheral membrane-associated fractions (24,34). With additional Snapin-interacting proteins being identified also been expanded into a broader range of intracellular membrane-fusion events, not just limited to neurotransmitter release as proposed earlier (12,13). Since ACVI, like Snapin, is also widely expressed in many tissues (9), it will be of great interest in the future to characterize whether the association of ACVI and Snapin might also occur in tissues other than the CNS and whether other functions involving intracellular membrane cycling might be regulated by such a close interaction between the cAMP and SNAREs systems.
Snapin Selectively Modulates the Inhibition of ACVI by PKC-Among AC isozymes, ACVI is unique because it exhibits relatively low catalytic activity (37), and most regulatory modes of this enzyme reported to date are negative. Specifically, ACVI can be inhibited by calcium, PKA, PKC, and G i ␣ protein (1,7,8,23,24,38). We first determined whether Snapin exerted a significant effect on the enzymatic properties of ACVI. As shown in Table I, expression of Snapin did not affect the basal, Gs␣-, or forskolin-evoked activities of ACVI in CHOP cells. We next examined whether interaction with Snapin affected the regulation of ACVI by PKA (7). As shown in Fig. 4A, PKA treatment suppressed the activity of ACVI evoked by Gs␣ plus forskolin expressed in CHOP cells. However, coexpression of Snapin did not modulate PKA-mediated inhibition of ACVI. To examine if Snapin exerted an important role in calciumevoked inhibition of ACVI, we followed the experimental designs of Cooper et al. (23). In brief, expression constructs of Snapin and ACVI were cotransfected into HEK293T cells. The intracellular calcium stores were emptied using thapsigargin and EGTA. As shown in Fig. 4B, calcium influx through capacitative Ca 2ϩ entry (CCE) initiated by addition of extracellular calcium caused a reduction in ACVI activity evoked by prostaglandin E1 plus forskolin. Importantly, coexpression of Snapin did not affect the inhibition of ACVI by calcium either.
We previously reported that PKC suppresses the catalytic activity of ACVI and that the N-terminal domain of ACVI significantly contributes to this inhibition. To investigate whether interaction with Snapin affects its susceptibility to PKC-mediated inhibition, expression constructs of ACVI and the indicated Snapin variant were cotransfected into CHOP cells. Membrane fractions prepared from CHOP cells transfected with the indicated expression constructs were incubated with purified PKCs which contained at least 10 different PKC isoforms including the ␦ and the ⑀ isoforms that inhibit ACVI (8), and then these were assayed for enzymatic properties. Intriguingly, the presence of wild-type Snapin abolished the PKC-mediated inhibition of forskolin-induced ACVI activity (Fig. 5A). Since expression of Snapin did not affect PKC activities (Supplemental Materials, Fig. S-2), Snapin did not exert its action through inhibiting PKC. Moreover, mutation of the only potential PKC phosphorylation site on Snapin (Thr 117 ) into alanine, although it reduced its expression level, did not disable the resultant mutant ability to reverse the inhibitory effect of PKC on ACVI. Since Snapin is a substrate of PKA (13), we next mutated the prominent PKA phosphorylation site (Ser 50 ) of Snapin into alanine. Similarly, the resultant Snapin mutant (Snapin-S50A) was as effective as the wildtype Snapin in reversing PKC-mediated inhibition of ACVI. Phosphorylation of Snapin by PKC or PKA therefore is not crucial for Snapin's action on ACVI. Conversely, Snapin ⌬33-51 which harbors an internal deletion of amino acids 33-51 did not affect PKC-mediated inhibition of ACVI, supporting that amino acids 33-51 of Snapin comprises the ACVI-interacting region as suggested in Fig. 2. Moreover, expression of Snapin did not affect PKC-mediated inhibition of ACVI-⌬A87, an ACVI variant, which lacked the Snapin-interacting region (amino acids 1-86, Fig. 5B and Ref. 5). Direct interaction between Snapin and ACVI therefore is crucial for Snapin action.
To further support the effect of Snapin on PKC-mediated inhibition of ACVI, recombinant Snapin fused to maltose-binding protein (designated MBP-Snapin, Fig. 6A) was prepared to test its ability to reverse the PKC-evoked inhibition of ACVI. As shown in Fig. 6B, recombinant MBP-Snapin, but not MBP alone, effectively pulled down recombinant N 1-160 in the in vitro binding assays, suggesting that recombinant MBP- Snapin likely exhibits the proper conformation for their interaction as that of Snapin. Note that a smaller protein existed in the purified MBP-Snapin preparation (Fig. 6A). This protein appeared to be a degradation product of MBP-Snapin because it was recognized by anti-MBP antiserum but not by anti-Snapin antiserum (SNA-c) which was raised against the most C terminus 18 amino acids of Snapin. Since the interaction between MBP-Snapin and N 1-160 could be established (Fig.  6B), the existence of this degradation product was unlikely to interfere with the action of Snapin under the experimental conditions tested. Most importantly, recombinant MBP-Snapin, but not MBP alone, dose-dependently relieved the inhibition of ACVI by PKC (Fig. 6C). This observation further strengthens our hypothesis that Snapin selectively modulates the inhibition of ACVI by PKC.
We previously reported that phosphorylation of ACVI by PKC caused suppression of its activity and subsequently led to reduced cAMP signaling during prolonged stimulation of the A 2A adenosine receptor in PC12 cells (8,21,39). Further biochemical analyses indicated that the regulatory domain (N 1-160 ) and the catalytic core complex (C1/C2) of ACVI are phosphorylated by PKC, and thus contribute to PKC-mediated inhibition (5,40). These studies suggest that the 3 cytosolic domains of ACVI might form a regulatory complex (40), and that the N terminus might contribute to its overall conformation. Regulation of ACVI by PKC appears to be sensitive to its conformation since defects in glycosylation at extracellular loops 5 and 6 caused ACVI to be less sensitive to PKC-mediated inhibition (41). Herein, we report that the interaction with Snapin caused ACVI to become insensitive to PKC-mediated suppression (Figs. 5 and 6). It is possible that Snapin, by interacting with the N terminus of ACVI, promotes a conformation of ACVI, which is resistant to PKC-mediated regulation.
Phosphorylation of Snapin itself by either PKC or PKA did not appear to be important for the ability of Snapin to regulate the effect of PKC on ACVI (Fig. 5A). We were unable to determine whether Snapin intervened in the action of PKC by interfering with PKC-mediated phosphorylation of ACVI, because the interaction between membrane-bound ACVI and Snapin in CHOP cells was not strong enough to survive membrane solubilization using detergents, a required step for coimmunoprecipitation of ACVI and Snapin from transfected CHOP cells (data not shown). Dissociation of assembled protein complexes during detergent solubilization and protein purification is a common problem during preparation of protein complexes (42,43).
Taken together, these findings demonstrate that the action of Snapin on PKC-mediated regulation of ACVI is selective (Figs. 4 -6). It is also of great interest to note that only part, but not all, of the ACVI detected in the brain and primary hippocampal neurons was colocalized with Snapin (Supplemental Materials, Fig. S-1 and Fig. 3). Association with Snapin therefore endows ACVI with a highly selective, finely tuned regulatory mode for PKC-mediated inhibition.
Conclusions-Data presented herein demonstrate that Snapin is a physiologically relevant interacting protein of ACVI, and functions as a modulator of ACVI on PKC-mediated inhibition via interacting with its N-terminal domain. Tight association of the cAMP pathway and SNAREs might be centrally involved in modulating channel activities or the rapid movement/organization of synaptic proteins, which underlie cAMP-mediated plasticity.