The Neuronal Ca2+ Sensor Protein Visinin-like Protein-1 Is Expressed in Pancreatic Islets and Regulates Insulin Secretion*

Visinin-like protein-1 (VILIP-1) is a member of the neuronal Ca2+ sensor protein family that modulates Ca2+-dependent cell signaling events. VILIP-1, which is expressed primarily in the brain, increases cAMP formation in neural cells by modulating adenylyl cyclase, but its functional role in other tissues remains largely unknown. In this study, we demonstrate that VILIP-1 is expressed in murine pancreatic islets and β-cells. To gain insight into the functions of VILIP-1 in β-cells, we used both overexpression and small interfering RNA knockdown strategies. Overexpression of VILIP-1 in the MIN6β-cell line or isolated mouse islets had no effect on basal insulin secretion but significantly increased glucose-stimulated insulin secretion. cAMP accumulation was elevated in VILIP-1-overexpressing cells, and the protein kinase A inhibitor H-89 attenuated increased glucose-stimulated insulin secretion. Overexpression of VILIP-1 in isolated mouse β-cells increased cAMP content accompanied by increased cAMP-responsive element-binding protein gene expression and enhanced exocytosis as detected by cell capacitance measurements. Conversely, VILIP-1 knockdown by small interfering RNA caused a reduction in cAMP accumulation and produced a dramatic increase in preproinsulin mRNA, basal insulin secretion, and total cellular insulin content. The increase in preproinsulin mRNA in these cells was attributed to enhanced insulin gene transcription. Taken together, we have shown that VILIP-1 is expressed in pancreatic β-cells and modulates insulin secretion. Increased VILIP-1 enhanced insulin secretion in a cAMP-associated manner. Down-regulation of VILIP-1 was accompanied by decreased cAMP accumulation but increased insulin gene transcription.

VILIP-1 was first cloned from a rat brain cDNA library (9), and at the protein level, it was shown to have 100% homology to human visinin-like peptide-1 (10). To date, studies on VILIP-1 have focused primarily on its role in neurons. Some of these studies have demonstrated Ca 2ϩ -dependent association with plasma membranes (11,12) and Golgi membranes (12), which might be a mechanism for the coordinated regulation of signaling cascades. Furthermore, in a human embryonic kidney cell line (tsA201), VILIP-1 has been shown to increase cell-surface expression of ␣ 4 -and ␤ 2 -subunits of the neuronal nicotinic acetylcholine receptor, which belongs to a superfamily of ligand-gated ion channels (13). In addition, VILIP-1 has modulatory effects on signaling of cAMP and cGMP in neuronal and peripheral cells (11, 14 -18). VILIP-1 has also been shown to increase cGMP levels in transfected C6 and PC12 cells by directly acting on guanylyl cyclase (17). Recent studies indicate that VILIP-1 modulates the activity of the receptor, guanylyl cyclase B, through clathrin-dependent receptor cycling, supporting a general physiological role for VILIP-1 in membrane trafficking within the central nervous system (14). In addition, VILIP-1 has been shown to bind double-stranded RNA as a protein-RNA complex for regulating the stability and localiza-tion of specific mRNAs, implicating its role in the regulation of gene expression (19).
VILIP-1 is expressed primarily in the brain, but lower expression has also been found in some peripheral organs, including the heart, testes, ovaries, and colon (20). Although many studies have reported on the biological functions of VILIP-1 in neurons, very little information is available regarding physiological roles in other tissues. In MIN6 cells, however, expression profiling of fatty acid-responsive genes has shown that oleic acid can significantly induce expression of VILIP-1 (21). In this study, we identified VILIP-1 in mouse pancreatic islets, and using overexpression and small interfering RNA (siRNA) knockdown strategies, we determined that VILIP-1 can regulate insulin secretion and insulin gene transcription. This study provides the first evidence that VILIP-1 plays a functional role in pancreatic ␤-cells.

MATERIALS AND METHODS
Reagents-The polyclonal antibody to VILIP-1 (rabbit antiserum) and the vector harboring the full-length cDNA of VILIP-1 or the fusion protein VILIP-1-enhanced green fluorescent protein (EGFP) have been described previously (12,15). Vectors expressing SNAP25 (synaptosome-associated protein of 25 kDa) and syntaxin-1A in mammalian cells were described in previous studies (22,23). The constructs of firefly luciferase driven by the rat insulin I promoter and Renilla luciferase driven by the cytomegalovirus promoter were gifts from Dr. Donald Fleenor (Duke University). The vector expressing NCS-1 protein was kindly provided by Dr. John Roder (Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto). Bovine serum albumin, oleic acid, H-89, forskolin, ionomycin, collagenase V, dispase II, rabbit polyclonal antibody against ␤-actin, and mouse monoclonal antibodies against syntaxin-1A and SNAP25 were products of Sigma. Mouse monoclonal antibodies against cellular subfractionation marker proteins annexin II, caveolin-1, and lamin A/C were obtained from Clontech. Guinea pig anti-insulin and mouse anti-glucagon antibodies were from Dako. Fluorescein isothiocyanate-or Cy5-conjugated secondary antibody was from Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA). Lipofectamine 2000, TRIzol, SuperScript II RNase H reverse transcriptase, ROX reference dye, SYBR Green dye, and Platinum Taq were purchased from Invitrogen. The cAMP assay kit was obtained from Biomedical Technologies, Inc. (Stoughton, MA). The Dual-Luciferase reporter assay kit was from Promega Corp. (Madison, WI). The siRNA duplex was designed using the program provided by Integrated DNA Technologies, Inc. (Coralville, IA), and the synthesis and high pressure liquid chromatography purification of the siRNAs were performed by the same company. The fluorescein-conjugated scrambled siRNA was from Dharmacon, Inc. (Lafayette, CO). The Qproteome compartment kit was purchased from Qiagen Inc.
Mouse Islet Isolation and Treatment-Mouse islets were isolated from male FVB/N mice (ϳ2 months of age) as described previously (24). To obtain single islet cells, these intact mouse islets were dispersed in dispase II solution (0.6 -2.4 units/ml) for 5 min and loaded onto glass coverslips after being washed with culture medium. The intact islets and single islet cells were cultured in RPMI 1640 medium containing 11.1 mM glucose supplemented with 10% fetal bovine serum, 10 mM HEPES, 100 units/ml penicillin, and 100 g/ml streptomycin for 24 h before being assayed.
Construction and Expression of siRNAs, Viruses, and Plasmids-The sequences of the siRNA duplex targeted to VILIP-1 were 5Ј-GAACAAAGAUGACCAGAUUTT-3Ј (sense) and 5Ј-AAUCUGGUCAUCUUUGUUCTT-3Ј (antisense), corresponding to nucleotides 480 -498 of mouse VILIP-1 cDNA (GenBank TM accession number D21165). The oligonucleotide sequences were subjected to a BLAST search, and no significant identity to other sequences was detected. The scrambled nonsense siRNA duplex sequences used as a negative control in MIN6 cells were 5Ј-UCAGAGUCUCGCAAUCACGTT-3Ј (sense) and 5Ј-CGUGAUUGCGAGACUCUGATT-3Ј (antisense) as described (25). The control duplex showed no significant homology to any known protein as assessed using the BLAST and NCBI Databases. VILIP-1 down-regulation was obtained by transfecting cells with the siRNA duplex (referred to as V1-siRNA) using Lipofectamine 2000 according to the manufacturer's instructions. 100 pmol of siRNA complexed by 5 l of Lipofectamine 2000 was used in 1 ml of medium. Fluorescein-conjugated siRNA duplex was used for cotransfection with V1-siRNA or scrambled siRNA to identify cells incorporated with the siRNA duplex. The full-length cDNA of VILIP-1-EGFP was subcloned into a cytomegalovirus promoter-driven adenoviral vector (AdLox.HTM) via HindIII and XbaI, yielding Ad-VILIP-1-EGFP. The virus was created according to the protocol described previously (26). Recombinant adenovirus particles of Ad-VILIP-1-EGFP or the control virus (Ad-EGFP) produced after recombination were amplified by passage in CRE8 cells as described (26). Recombinant adenovirus particles (ϳ10 10 plaque-forming units/ml) were used to infect isolated mouse islets (multiplicity of infection of 10 4 , assuming 10 3 ␤-cells/islet). Overexpression of VILIP-1 in MIN6 or single islet cells was achieved by transfection of VILIP-1-EGFP using Lipofectamine 2000 (27), and that in intact mouse islets by adenoviral transduction. A pilot study with the control virus Ad-EGFP demonstrated that 70 -80% of the cells in islets were infected as detected by confocal fluorescence microscopy.
Evaluation of Insulin Secretion and Content in MIN6 Cells and Mouse Islets-MIN6 cells were seeded into 24-well plates and incubated for 48 -72 h before glucose-stimulated insulin secretion (GSIS) studies. Cells were preincubated for 1 h in glucose-free Krebs-Ringer HEPES buffer (125 mM NaCl, 5.9 mM KCl, 1.28 mM CaCl 2 , 5.0 mM NaCO 3 , 25 mM HEPES, and 0.1% (w/v) bovine serum albumin) in a 37°C humidified incubator. The cells were then incubated in the same buffer containing 2.8 mM (basal concentration) or 20 mM (stimulatory concentration) glucose for 1 h or 20 mM glucose with 20 mM arginine for 5 min. Islets (10/condition) were preincubated for two sequential 30-min periods in Krebs-Ringer HEPES buffer and for 2 h at the indicated glucose concentration (either 2.8 or 20 mM) (24). The supernatants of the cells were collected and centrifuged at 300 g ϫ for 10 min to remove cell debris for insulin secretion assessment, and the cell pellets were lysed with acid ethanol (75% ethanol containing 1.5% (v/v) HCl) for DNA quantification or insulin content assessment (21). The insulin concentration in the supernatant or cell lysate was measured by radioimmunoassay as described previously (28). The insulin secretion data were normalized by total DNA content.
Measurement of Intracellular cAMP Content-Intracellular cAMP was measured as described previously (29). Briefly, the cells were washed with cold Krebs-Ringer HEPES buffer after removal of the supernatant. Intracellular cAMP was extracted with 80% ethanol and resuspended in cAMP assay buffer (0.05 mM sodium acetate (pH 6.2) and 0.01% sodium azide) and measured by radioimmunoassay using an intracellular cAMP assay kit.
Quantitative Real-time PCR-Total RNA was extracted from MIN6 cells using TRIzol reagent and further purified with RNeasy kits according to the manufacturer's instructions. Purified RNA was converted to cDNA using SuperScript II reverse transcriptase, and real-time PCR was performed using an ABI PRISM 7900HT sequence detection system (Applied Biosystems, Foster City, CA) according to the same protocol described previously (21). Primer Express Version 2.0 (Applied Biosystems) was used to design primers for real-time PCR. Platinum Taq DNA polymerase reagents were used to amplify target cDNA sequences, and SYBR Green was used to quantify PCR products. A standard curve was generated using mouse genomic DNA for quantification purposes. The measurements of gene expression were normalized by ␤-actin transcripts in the same sample. The sequences of the primers are shown in Table 1.
Luciferase Reporter Assay-24 h after siRNA duplex transfection, MIN6 cells were cotransfected with rat insulin I promoter-Renilla luciferase and cytomegalovirus promoter-Renilla luciferase (9:1 ratio); the latter was used for constitutive expression of Renilla luciferase as an internal control. 2 days after cotransfection, the cells were washed with cold phosphate-buffered saline and lysed with reporter buffer (Promega Corp.). The assay was performed using the Dual-Luciferase reporter assay system. The ratio of the activity of firefly to Renilla luciferase is presented as the relative luciferase activity. Luciferase expression without an upstream promoter was taken as the basal activity.
Fractionation, Immunoblotting, and Immunoprecipitation-Tissues and MIN6 cells were homogenized in lysis buffer (20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM Na 2 EDTA, 1 mM EGTA, 1% Triton, 2.5 mM sodium pyrophosphate, 1 mM ␤-glycerophosphate, 1 mM Na 3 VO4, 1 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride). The subcellular fractionation of MIN6 cells was performed according to the protocol of the Qproteome compartment kit. Differential speed centrifugation of the lysate was implemented to obtain subcellular fractions. The supernatant from the 1000 ϫ g centrifugation contained cytosolic proteins, that from the 6000 ϫ g centrifugation contained membrane proteins, and that from the 6800 ϫ g centrifugation contained nuclear proteins. The fractions or whole cell lysates were separated by 12% SDS-PAGE and transferred to polyvinylidene difluoride membranes. The rabbit polyclonal antibody against VILIP-1 used for immunoblotting has been described previously (30). The intensity of the appropriate immunoreactive bands was measured by densitometry, and the data were normalized by ␤-actin content as described (21) and analyzed using image analysis software (Scion Image Version 4.02, Scion Corp., Frederick, MD). Immunoprecipitation was performed as described previously (31). Briefly, HEK293 cells transfected with the cDNA of VILIP-1, syntaxin-1, or SNAP25 were harvested in binding buffer (25 mM HEPES (pH 7.4), 100 mM KCl, 2 mM EDTA, 1% MAPK/ERK growth and differentiation
Immunohistochemistry-Pancreases were fixed in 4% paraformaldehyde, dehydrated, and processed for paraffin embedding using routine procedures. Sections (5 m) were adhered to glass slides, rehydrated, and blocked with 3% H 2 O 2 for 10 min. After being washed with phosphate-buffered saline, sections were incubated in appropriate non-immune serum (5%) for 30 min (goat for VILIP and rabbit for insulin). Sections were blotted to remove excess serum prior to overnight application of anti-insulin (1:1000) or anti-VILIP-1 (1:800) primary antibody (32) at 4°C. Positive immunostaining was detected by incubating sections with goat antirabbit or rabbit anti-guinea pig peroxidase (1:100) for VILIP-1 and insulin, respectively, for 60 min at room temperature, followed by application of diaminobenzidene/ H 2 O 2 for 8 min.
Dispersed mouse islet cells on glass coverslips were fixed in 3.7% formaldehyde and permeabilized with 0.1% Triton X-100. After being washed with phosphatebuffered saline, the coverslips were blocked in 5% bovine serum albumin for 2 h before overnight application of primary antibody against VILIP-1, insulin, or glucagon at 4°C. Cy5-conjugated donkey anti-rabbit secondary antibody was used to detect VILIP-1; fluorescein isothiocyanate (FITC)conjugated rat anti-mouse and FITC donkey anti-guinea pig secondary antibodies were used to detect glucagon and insulin, respectively. The images were captured by confocal fluorescence microscopy.
Capacitance Measurements-Cells expressing EGFP or exhibiting fluorescein fluorescence were selected and patchclamped in the whole cell configuration at 33°C. The capacitance measurements were performed using an EPC-9 amplifier and PULSE software from HEKA Elektronik (Lambrecht, Germany). Patch pipettes were pulled from 1.5-mm thinwalled borosilicate glass tubes using a two-stage Narishige micropipette puller and had typical resistances of 3-6 megaohms when fire polished and filled with an intracellular solution containing 120 mM CsCl, 20 mM tetraethylammonium chloride, 1 mM MgCl 2 , 0.05 mM EGTA, 10 mM HEPES, 0.1 mM cAMP, and 5 mM MgATP, pH 7.3, with CsOH. Extracellular solutions contained 118 mM NaCl, 20 mM triethylammonium Confocal Fluorescence Microscopy-Islets overexpressing VILIP-1-EGFP were analyzed using a Zeiss LSM-510 laser scanning confocal imaging system with a ϫ40 water immersion Plan Apochromat objective. Samples were subjected to optical sectioning by moving the focal plane along the vertical (z) axis. HEK293 cells transiently transfected with VILIP-1-EGFP were treated with 1 mM ionomycin in Krebs-Ringer HEPES buffer. Serial images of these cells were taken every 15 s with the same apparatus.
Statistics-All data are presented as the means Ϯ S.E. Statistical comparisons were performed using Student's t test. Patch clamp data were analyzed with IGOR Pro Version 3.12 software (WaveMetrics, Lake Oswego, OR). Differences were considered to be significant when p Ͻ 0.05.

Expression of VILIP-1 in Cells and
Tissues-Immunostaining of dispersed mouse islets showed that VILIP-1 localized to insulin-expressing ␤-cells and glucagon-expressing ␣-cells (Fig. 1A). Western blotting showed VILIP-1 expression (ϳ22 kDa) in MIN6 ␤-cells, ␣-TC6 cells, and intact isolated mouse islets (Fig. 1B). VILIP-1 was also highly expressed in mouse brain but was not detected in mouse pancreatic extracts, suggesting little or no expression in acinar tissue (Fig. 1B). Immunohistochemical analysis of serial mouse pancreatic sections showed that VILIP-1 immunoreactivity was confined to islets and overlapped with insulin, further confirming VILIP-1 deficiency in acinar tissue (Fig. 1C). Immunoblot analysis of subcellular compartment fractionation in MIN6 cells revealed that VILIP-1 could be detected in cytosolic and membrane fractions but not in the nuclear fraction using annexin II, caveolin-1, and lamin A/C as the marker proteins for the cytosolic, membrane, and nuclear fractions, respectively (Fig. 1D).
Assessment of VILIP-1 Overexpression in MIN6 Cells and Isolated Mouse Islets-Overexpression of VILIP-1 in MIN6 cells and single islet cells was accomplished by transient transfection of the pEGFP-N1 vector expressing a fusion protein, yielding pVILIP-1-EGFP, in which EGFP was fused to the C terminus of VILIP-1 (12). Based on EGFP fluorescence, the transfection efficiencies routinely ranged from 80 to 90% in MIN6 cells and from 5 to 10% in single islet cells (data not shown). To accomplish VILIP-1 overexpression in mouse intact islets, an adenovirus harboring the VILIP-1-EGFP cDNA was constructed (Ad-VILIP-1-EGFP). Western blot analysis confirmed overexpression of VILIP-1-EGFP in MIN6 cells and mouse intact islets (Fig. 2, A and B) using cells overexpressing EGFP driven by the same promoter as the control. A transduction efficiency of 70 -80% with Ad-VILIP-1-EGFP was routinely achieved 48 h post-infection as observed by confocal fluorescence microscopy (Fig.  2C). All of the experiments were performed 48 h after plasmid transfection or viral infection.

Measurement of Cell Capacitance in Single ␤-Cells
Overexpressing VILIP-1-cAMP can potentiate insulin secretion by various means, including acting directly on exocytosis (35). In pancreatic ␤-cells, exocytosis is regulated at several levels both temporally and spatially (36). The insulin granule pool consists of a low percentage of readily releasable pool (RRP) immediately accessible for release triggering and the reserve pool (RP) that requires mobilization to RRP before docking (37). It was shown very recently by cell capacitance measurements that overexpression of NCS-1, a member of NCS family Group II, facilitates glucosedependent exocytosis by increasing RRP via stimulation of phosphatidylinositol 4-kinase-␤ in pancreatic ␤-cells (Fig. 4, E and F) (27). Given these data, we performed cell capacitance measurements on single mouse ␤-cells to verify that VILIP-1 potentiates insulin secretion by directly stimulating exocytosis. At 20 mM glucose, the total increase in cell capacitance in control cells was 172 Ϯ 19 femtofarads (fF) (Fig. 4, A  and B) in response to a train of 10 depolarizations. Under the same conditions, cells overexpressing VILIP-1 demonstrated a much stronger exocytotic response evoked by the same depolarization (376 Ϯ 38 fF) (Fig. 4, A and B). Compared with control cells, VILIP-1 overexpression augmented cell capacitance by 2.2 Ϯ 0.3-fold (n ϭ 6; p Ͻ 0.001) (Fig. 4B). In general, exocytosis during the first two or three depolarizations is believed to represent the capacity of RRP, whereas exocytosis evoked by the subsequent depolarizations demonstrates the refilling of the RRP by mobilization of new granules from the RP (38,39). Close inspection of the exocytotic events in this study revealed that the capacitance increase elicited by the first three depolarizations in the cells transfected with VILIP-1 was 152 Ϯ 26 fF, markedly larger than that in control cells (69 Ϯ 8 fF, n ϭ 6; p Ͻ 0.01) (Fig. 4C). In addition, a consistent increase in cell capacitance was observed in response to the following seven additional depolarizations from 103 Ϯ 13 fF in control cells to 224 Ϯ 27 fF in VILIP-1-overexpressing cells (n ϭ 6; p Ͻ 0.01) (Fig. 4C), showing a more profound enhancement of the RRP refilling in VILIP-1-overexpressing cells. These results suggested that VILIP-1 overexpression in the pancreatic ␤-cells stimulated exocytosis by increasing RRP size and enhancing the refilling process of RRP through promoting granule priming from RP to RRP. It was noteworthy that, during the train of depolarizations, the Ca 2ϩ currents in both VILIP-1-overexpressing and control cells declined because of inactivation, and there was no significant difference . VILIP-1 overexpression enhances exocytosis in mouse ␤-cells. A train of ten 500-ms depolarizations from Ϫ70 to 0 mV was applied to mouse ␤-cells transfected with pEGFP (Control) or pVILIP-1-EGFP. Changes in cell membrane capacitance (⌬Cm) were measured at 20 mM glucose 2 days after transfection. A, representative ⌬Cm traces recorded from a control ␤-cell (pEGFP) and a VILIP-1-EGFP-overexpressing ␤-cell. The electrical depolarization protocol is indicated below the corresponding traces. B, total increase in ⌬Cm evoked by a train of depolarizations in cells overexpressing VILIP-1 and control EGFP (n ϭ 6). C, summary data of cumulative increases in cell capacitance for each depolarizing pulse. pC, picocoulomb. D, representative Ca 2ϩ currents and summary of integrated Ca 2ϩ currents (Q Ca ) evoked by the first and tenth pulse during a train of 10 depolarizations. Currents were recorded associated with the capacitance measurements. E and F, representative traces and summary of ⌬Cm, respectively, in cells transfected with EGFP (Control; n ϭ 8) or NCS-1 (n ϭ 11) used as a positive control for the effect of VILIP-1 on exocytosis according to Gromada et al. (27). Data are presented as the means Ϯ S.E. †, p Ͻ 0.05; ‡, p Ͻ 0.01; §, p Ͻ 0.001.
in the integrated Ca 2ϩ current between these two cells (Fig.  4D). This indicated that the effect of VILIP-1 on exocytosis was not caused by an increase in Ca 2ϩ influx.
Assessment of MIN6 Cells Transfected by the siRNA Duplex against VILIP-1-To further investigate the role of VILIP-1 in the pancreatic ␤-cell, we reduced VILIP-1 expression in the MIN6 ␤-cells by transiently transfecting MIN6 cells with siRNA duplexes against VILIP-1 mRNA (V1-siRNA). The control siRNA duplex was previously shown not to have any observable effects in MIN6 cells (25) and in the present study had no effect on a variety of parameters, including VILIP-1 and ␤-actin expression (Fig. 5A), insulin secretion, and cAMP production (data not shown), compared with nontransfected MIN6 cells. At 72 h post-transfection with V1-siRNA, expression of VILIP-1 was decreased by 55 Ϯ 14% ( p Ͻ 0.05) compared with the control using 50 nM siRNA and by 85 Ϯ 2% ( p Ͻ 0.05) using 100 nM siRNA (Fig. 5B). The suppression of VILIP-1 protein was observed for at least ϳ96 h post-transfection, with endogenous levels returning to the control level within 144 h (Fig. 5C). The VILIP-1 and control siRNA duplexes described above were used in all subsequent studies, with experiments being performed 72 h post-transfection (100 nM siRNA was used).
Effect of VILIP-1 Down-regulation on Insulin Secretion, Insulin Content, and cAMP Content in MIN6 Cells-To examine the consequences of significantly reducing VILIP-1 expression in MIN6 cells, insulin secretion assays were performed. Interestingly, compared with control cells, down-regulation of VILIP-1 resulted in a significant increase (2.14 Ϯ 0.29, n ϭ 21; p Ͻ 0.001) in basal insulin secretion (0.51 Ϯ 0.05 versus 0.23 Ϯ 0.02 ng/g of DNA, respectively) (Fig. 6A). In the presence of a stimulatory glucose concentration (20 mM), the insulin secretion was higher in cells expressing V1-siRNA than in those expressing control siRNA (0.91 Ϯ 0.06 versus 0.61 Ϯ 0.03 ng/g of DNA, respectively, n ϭ 20; p Ͻ 0.001) (Fig. 6A). When treated with H-89, MIN6 cells expressing V1-siRNA demonstrated decreased GSIS compared with those without treatment (from 0.91 Ϯ 0.06 to 0.57 Ϯ 0.11 ng/g of DNA, n ϭ 9; p Ͻ 0.001). However, GSIS in V1-siRNA cells with H-89 treatment was still larger than that in control cells (0.57 Ϯ 0.11 versus 0.41 Ϯ 0.04 ng/g of DNA, n ϭ 9; p Ͻ 0.05) (Fig. 6A). Despite enhanced insulin secretion, the cAMP content in MIN6 cells transfected with V1-siRNA was significantly decreased by 26.3 Ϯ 4.1% (n ϭ 15; p Ͻ 0.001) compared with control cells at the basal glucose concentration (0.73 Ϯ 0.02 versus 0.99 Ϯ 0.04 pmol/g of DNA) and by 27.1 Ϯ 6.5% (n ϭ 12; p Ͻ 0.05) at 20 mM glucose (0.63 Ϯ 0.04 versus 0.87 Ϯ 0.06 pmol/g of DNA) (Fig. 6B). H-89 treatment had no effect on cAMP accumulation in either V1-siRNA-expressing cells or control cells because PKA is downstream of cAMP (Fig. 6B). Thus, the augmented insulin secretion was attributed to mechanisms other than altered cAMP accumulation. To examine the capacity of MIN6 cells to secrete insulin, 20 mM arginine in the presence of glucose stimulus was used and shown to increase secretion significantly in V1-siRNA cells compared with control cells (1.41 Ϯ 0.08 versus 1.00 Ϯ 0.09 ng/g of DNA, n ϭ 10; p Ͻ 0.001) (Fig.  6A). This increased insulin capacity was supported by total insulin content measurement in V1-siRNA-transfected MIN6 cells, which was increased (66 Ϯ 0.05%, n ϭ 15; p Ͻ 0.001) compared with control cells (42.8 Ϯ 1.2 versus 25.8 Ϯ 0.8 ng/g of DNA) (Fig. 6C). H-89 did not abolish this effect (33.3 Ϯ 2.3 ng/g of DNA in V1-siRNA cells versus 19.4 Ϯ 1.8 ng/g of DNA in control cells, n ϭ 9; p Ͻ 0.05), although H-89 did decrease insulin content in both groups (Fig. 6C). To examine insulin expression in V1-siRNA-transfected MIN6 cells, we performed real-time PCR on RNA from these cells. Our results showed that preproinsulin mRNA in V1-siRNA cells was significantly increased by 54 Ϯ 14.7% (n ϭ 12; p Ͻ 0.001) over that in control cells (Fig. 6D). These results suggested that downregulation of VILIP-1 likely induced insulin gene transcription independent of the cAMP/PKA pathway. To further verify induction of insulin gene transcription, we performed luciferase assays in which luciferase expression was driven by the rat insulin I promoter. The relative activity of lucifer- ase in V1-siRNA-expressing cells was significantly increased (41 Ϯ 3%, n ϭ 9; p Ͻ 0.05) compared with that in control cells (Fig. 6E).
Effect of VILIP-1 Down-regulation on Cell Capacitance-Cell capacitance measurements record changes in the size of the insulin granule pool in single cells. To verify that VILIP-1 down-regulation caused an increase in total insulin content, we measured cell capacitance in MIN6 cells expressing V1-siRNA. MIN6 cells were cotransfected with V1-siRNA or control siRNA and fluorescein-conjugated scrambled siRNA (4:1 ratio). The fluorescein-conjugated siRNA served as a nonsense siRNA for optimal negative control and demonstrated siRNA incorporation into cells. Single cells demonstrating fluorescein were selected for cell capacitance measurements. Throughout the depolarization, the increase in cell capacitance in MIN6 cells expressing V1-siRNA was almost 2-fold of that in control cells (201 Ϯ 35 versus 112 Ϯ 20 fF, n ϭ 7; p Ͻ 0.05) (Fig. 7). This finding suggested that the size of the insulin granule pool was increased in V1-siRNA-expressing cells, further confirming the previous finding that VILIP-1 knockdown caused an enlarged insulin secretion capacity.

DISCUSSION
The NCS family, a group of EF-hand-containing proteins, plays a variety of roles in cell signaling (44), including regulation of ion channel function (4), activation of multiple target proteins such as kinases (6), and interactions with nucleic acids and control of cyclic nucleotide levels (5,19). In this study, we investigated VILIP-1 expression and its biological function in pancreatic ␤-cells. Our results showed that VILIP-1 was indeed The value of preproinsulin mRNA obtained by real-time PCR was normalized by ␤-actin transcripts in the sample. Expression of luciferase is presented as the ratio of the activity of rat insulin I promoter-driven firefly luciferase (RIP-luciferase) to that of the constitutively expressed Renilla luciferase. Luciferase expression without an upstream promoter was taken as the basal condition. Results are the means Ϯ S.E. from three independent experiments carried out in three to four replicates. * and **, p Ͻ 0.05 and p Ͻ 0.001 (n Ն 9), respectively, compared with the control. expressed in mouse pancreatic islets (␤-and ␣-cells) but not in acinar tissue (Fig. 1). Although VILIP-1 could not be detected in the whole pancreas, islet cells, which occupy only a small fraction of the whole pancreas, showed strong VILIP-1 expression. Expression of VILIP-1 within the endocrine pancreas suggests a potential role in islet function.
It has been well demonstrated that VILIP-1 is involved in cAMP and cGMP signaling by modulating the cyclic nucleotide level in neurons (14,15,17,30). Early studies showed that VILIP-1 overexpression in glioma cells enhances ␤-adrenergic receptor-stimulated cAMP accumulation and that this effect is still seen after stimulation of adenylyl cyclase with forskolin (15), indicating that VILIP-1 acts downstream of the receptor. Subsequent studies provided more evidence implicating the role of VILIP-1 in adenylyl cyclase modulation (11,16). In a Ca 2ϩ -dependent manner, VILIP-1 associates with membrane structures to interact with proteins, including adenylyl cyclase (11,12,45). In our study, in agreement with the findings in neurons, we found that VILIP-1 overexpression resulted in elevated cAMP accumulation in pancreatic ␤-cells (Fig. 3E). In contrast, VILIP-1 knockdown caused decreased cAMP levels (Fig. 6B). As shown in a previous study (11), VILIP-1-EGFP expressed in HEK293 cells translocated from the cytosol to the plasma membrane (supplemental Fig. 1) upon stimulation with ionomycin (a Ca 2ϩ ionophore). This translocation is considered to facilitate adenylyl cyclase modulation. Thus, cAMP may possibly be a downstream molecule through which Ca 2ϩ and subsequently VILIP-1 exert their biological functions.
It is well known that cAMP, as a second messenger in the amplification pathway of stimulus coupling in ␤-cells, enhances insulin secretion only in the presence of stimulatory glucose concentrations. In MIN6 cells and mouse islets, VILIP-1 overexpression led to enhanced GSIS without significant effects on basal insulin secretion (Fig. 3, A and F). Enhanced GSIS can be mimicked by forskolin, which stimulates adenylyl cyclase to activate cAMP/PKA signaling cascades, leading to increased insulin secretion at stimulatory glucose concentrations (46). The increased GSIS observed in VILIP-1-overexpressing cells was blocked by the PKA inhibitor H-89, confirming that this effect is dependent upon the cAMP/PKA pathway. In our study, we used 2.8 mM glucose as the low glucose condition, which is likely insufficient for cAMP to potentiate insulin secretion. This could provide an explanation for our observation in VILIP-1-overexpressing ␤-cells that insulin secretion was enhanced at high but not basal glucose concentrations. Consistent with our findings, cAMP accumulation elicited by genistein, a soy isoflavone, stimulates insulin secretion in mouse islets at high but not low glucose concentrations (47).
cAMP can potentiate insulin secretion in several ways.  These include increasing voltage-sensitive Ca 2ϩ channel opening (48) and Ca 2ϩ influx (49), activating ryanodine receptors in the endoplasmic reticulum (50), stimulating lipolysis (51), and directly acting on the exocytotic machinery (35). Most of these actions are mediated by the protein PKA pathway, but the direct effect on exocytosis is partly PKA-independent (34), being mediated by the cAMP-binding protein cAMP-regulated guanine nucleotide exchange factor II (53). In pancreatic ␤-cells, the insulin granule pool consists of a small RRP that is immediately accessible for release and a large RP that refills the RRP by granule mobilization (37). In general, a PKA-independent process increases the release probability of granules already in the RRP, whereas a PKA-dependent process promotes the refilling of the RRP by stimulation of mobilization of granules from the RP (34). VILIP-1 overexpression significantly increased exocytosis in response to the train of depolarizations in single ␤-cells compared with control cells (Fig. 4). This result is consistent with what has been observed with a related NCS protein, NCS-1. Gromada et al. (27) demonstrated that NCS-1 is expressed in pancreatic ␤-cells and that overexpression stimulates insulin exocytosis via phosphatidylinositol 4-kinase-␤. Our results in VILIP-1-overexpressing cells employing capacitance recordings suggested that VILIP-1 increased the size and promoted the refilling of the RRP by stimulating the mobilization of granules in the RP, thereby increasing the capacity for exocytosis. It is known that NCS-1 partially co-localizes with the synaptic protein SNAP25 at the plasma membrane in NG108-15 cells (a mouse neuroblastoma and rat glioma hybrid) and that this interaction may facilitate exocytosis (54). However, in our study, immunoprecipitation experiments showed no interaction between VILIP-1 and the SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptor) proteins syntaxin-1A and SNAP25 (supplemental Fig. 2) (31). Thus, our data suggested that VILIP-1 overexpression increased cAMP accumulation in pancreatic ␤-cells and potentiated insulin secretion via a cAMPdependent effect on insulin exocytosis. It is noteworthy that, in addition to stimulation of exocytosis in ␤-cells, activation of the cAMP/PKA pathway stimulates insulin gene transcription (55,56). More recently, it has been shown that GLP-1, through activation of the cAMP/PKA pathway, increases the levels of PDX-1 protein and its translocation from the cytosol to nuclei in insulin-secreting RIN 1046-38 cells (57). PDX-1 is involved in several aspects of pancreatic ␤-cells function, including the positive regulation of insulin gene transcription (58), and directly activates insulin transcription by forming complexes with coactivators on the proximal insulin promoter (43). In our study, however, despite elevated cAMP accumulation, no change in either preproinsulin mRNA or pdx-1 expression was observed with VILIP-1 overexpression. Conversely, in MIN6 cells transfected with V1-siRNA, despite decreased cAMP accumulation, insulin content and preproinsulin mRNA were actually increased. Therefore, insulin secretion at both basal and stimulatory glucose concentrations was elevated accordingly, whereas H-89 did not abolish this effect (Fig. 6). Of further interest, pdx-1 gene expression was actually up-regulated, and insulin gene transcription was accordingly stimulated (Figs. 6   and 8). It appears that VILIP-1, via an alternative pathway, may exert inhibitory effects on insulin gene transcription, possibly through the suppression of pdx-1, and that this effect is independent of the cAMP/PKA pathway. These two pathways thus may allow dual effects of VILIP-1 in ␤-cells, stimulating insulin exocytosis and inhibiting insulin gene transcription. Although we have observed increased insulin gene transcription and content in VILIP-1 knockdown ␤-cells, we have not yet conclusively shown a direct effect on proinsulin biosynthesis.
Our data suggest that VILIP-1 may act in part as an insulin gene repressor, either directly or indirectly. Interestingly, VILIP-1 has been shown to bind to double-stranded RNA (19), implying a possible role of this protein in the regulation of gene expression. In addition, another NCS protein, KChIP3/ DREAM (Kv channel-interacting protein-3/downstream regulatory element antagonistic modulator), a member of NCS protein family, has been reported to repress dynorphin gene transcription and production of dynorphin A peptides, which are involved in memory acquisition and pain. This effect is exerted directly on the prodynorphin promoter (5). Thus, previous studies and our present findings suggest a possible role of VILIP-1 in gene regulation. As for the mechanism, more studies are required to elucidate whether VILIP-1 itself acts as a transcription regulator or exerts its effect via other molecules to regulate genes like pdx-1 and insulin.
VILIP-1 expression also influenced expression of genes involved in cell proliferation and cell cycle regulation (cyclin D 2 , Gsk3b, and CREB) (Fig. 8). Cyclin D 2 associates with the cyclin-dependent kinase CDK4 or CDK6 to promote cell cycle progression (59) and in rodents is required for postnatal islet growth and glucose homeostasis throughout life, as revealed by cyclin D 2 (Ϫ/Ϫ) mice (60). Gsk3b is linked with phosphorylation and proteolytic turnover of cyclin D 1 during the cell division cycle and thus has effects on cell cycle regulation (42). CREB is a direct downstream transactivator of the cAMP/PKA pathway, and stimulation of the cAMP pathway leads to PKAmediated phosphorylation of CREB, which is found to occupy up to 4000 promoter sites in vivo and therefore can have profound effects on a variety of cellular processes (52). Further direct evidence is needed to demonstrate specific effects of VILIP-1 on cell proliferation and cell replication.
In summary, this is the first study demonstrating that VILIP-1, an NCS protein, is expressed in pancreatic ␣and ␤-cells and is involved in the regulation of insulin secretion and insulin gene expression. Increased VILIP-1 enhanced GSIS in a cAMP-associated manner, whereas down-regulation of VILIP-1 was accompanied by decreased cAMP accumulation but increased insulin content. This study further supports a role for NCS proteins and in particular VILIP-1 in islet function.