DANGER, a Novel Regulatory Protein of Inositol 1,4,5-Trisphosphate-Receptor Activity*

We report the cloning and characterization of DANGER, a novel protein which physiologically binds to inositol 1,4,5-trisphosphate receptors (IP3R). DANGER is a membrane-associated protein predicted to contain a partial MAB-21 domain. It is expressed in a wide variety of neuronal cell lineages where it localizes to membranes in the cell periphery together with IP3R. DANGER interacts with IP3R in vitro and co-immunoprecipitates with IP3R from cellular preparations. DANGER robustly enhances Ca2+-mediated inhibition of IP3 RCa2+ release without affecting IP3 binding in microsomal assays and inhibits gating in single-channel recordings of IP3R. DANGER appears to allosterically modulate the sensitivity of IP3 RtoCa2+ inhibition, which likely alters IP3R-mediated Ca2+ dynamics in cells where DANGER and IP3R are co-expressed.

The inositol 1,4,5-trisphosphate receptor (IP 3 R) 3 is a large, endoplasmic reticulum (ER) resident protein, which is a key regulator of intracellular Ca 2ϩ signaling (1,2). Inositol 1,4,5trisphosphate (IP 3 ) is formed in response to the activation of G protein-coupled receptors or receptor tyrosine kinase receptors located in the plasma membrane (1), which elicit IP 3 Rmediated Ca 2ϩ release from ER stores.
The IP 3 recognition site of IP 3 R includes amino acids (aa) 225-578 in the N-terminal portion of the protein, while the Ca 2ϩ channel domain comprises ϳ300 aa in the extreme C terminus (2). The IP 3 binding site and the Ca 2ϩ channel are separated by ϳ2,000 aa, providing a large area for interactions with multiple regulatory proteins including calmodulin, chromogranins, glyceraldehyde-3-phosphate dehydrogenase, RACK1, and caldendrin (3). While these proteins regulate IP 3 R function in diverse ways, only two regulators have been shown to influence Ca 2ϩ sensitivity. Cytochrome c, which binds to the extreme C terminus of the IP 3 R, relieves the inhibitory actions of Ca 2ϩ upon the channel (4), and Bcl-XL binding to the C terminus also influences the Ca 2ϩ dependence (5).
We have identified a novel vertebrate protein, designated DANGER, which was isolated by yeast two-hybrid analysis with the regulatory region of the IP 3 R as bait. DANGER physiologically binds to IP 3 R and allosterically enhances the potency of Ca 2ϩ -inhibition of IP 3 R-mediated Ca 2ϩ release without affecting ligand binding.

EXPERIMENTAL PROCEDURES
Yeast Two-hybrid Analysis-The Matchmaker3 yeast-2-hybrid system from Clontech (Palo Alto, CA) was employed. AH109 ␤-galactosidase yeast was used. IP 3 R fragments were cloned into pGADT7 (␤-galactosidase acceptor domain) vector and were screened against a rat brain and human fetal kidney library (Clontech) as per the manufacturer's specifications. Expression of these fragments was determined by Western blotting using antibodies from Clontech, corresponding to the expression vector. Positive clones grew on minimal SD agar (Clontech) lacking adenine, histidine, leucine, and tryptophan and had ␤-galactosidase activity.
Calcium Release Measurements and Electrophysiology-Calcium release through recombinant type I (SIIϩ) IP 3 R was measured exactly as described previously (13,14). Calcium concentrations in all solutions were calibrated with a calcium-selective minielectrode, and confirmed with fluorescent dyes. When the effects of DANGER were examined, it was added simultaneously with IP 3 . In the absence of IP 3 , SERCA activity was not affected by the addition of DANGER. At the concentrations used in this study, DANGER did not change the calcium concentration of the assay solution as determined by a calcium minielectrode.
Spodoptera frugiperda (Sf9) cells (Invitrogen) were grown and maintained in SF-900II serum-free media (Invitrogen) as described previously (9). An ice-cold nuclear isolation solution, containing (in mM): 140 KCl, 250 sucrose, 1.5 ␤-mercaptoethanol, 10 (15). Single channel traces were used for dwell time analyses by QuB software (16). The number of active channels in nuclear membrane patches (N A ) and the mean channel activity duration T a were determined as described (9). All data are averages of at least six experiments performed at the same [InsP 3 ] and [Ca 2ϩ ] i . Error bars indicate S.E. Cell Lines and Expression Protocols-Human HEK cells were obtained from ATCC and cultured in Dulbecco's minimal essential medium supplemented with 10% fetal bovine serum, 2 mM L-glutamine, and 1% penicillin-streptomycin. PC12 cells were cultured in Dulbecco's minimal essential medium supplemented with 10% horse serum, 5% fetal bovine serum, 2 mM L-glutamine, and 1% penicillin-streptomycin.
Transfection of PC12 and HEK with either 100 nM small interfering RNA (siRNA) duplex and 1 g YFP, or 1 g of DAN-GER and 1 g of YFP, was performed using Lipofectamine 2000 (Invitrogen), according to the manufacturer's specifications.
Expression Constructs-The type 1 IP 3 R SIIϩ splice variant and SERCA-2b in pcDNA3.1 have been described elsewhere (17) and were kindly provided by Dr. Suresh K. Joseph (Thomas Jefferson University, Philadelphia, PA). DANGER was cloned from human, rat, and mouse cDNA libraries generated from brain using the Roche reverse transcriptase. PCR was performed using 5Ј and 3Ј 21-mer primers corresponding to the following sequences human (GI: 29789287), Mouse (GI: 39645725), and Rat (GI:30466774), which have 9-mer overhangs on each end to engineer the appropriate restriction sites.
Antibody Generation-Rabbit polyclonal antiserum against DANGER was generated by injecting New Zealand White male rabbits with peptide RAEPLNLFRPFVLQR SLYR coupled to keyhole limpet hemocyanin (Pierce). Initial injection was with Complete Freund's adjuvant, and boost injections at days 14, 21, and 49 were with Incomplete Freund's Adjuvant. Rabbit injections, bleeds, and housing were performed by Cocalico Biologicals (Reamstown, PA).
Monoclonal type 1 IP 3 R antibody m18A10 was a kind gift form K. Mikoshiba. Anti-Myc antibody was from Sigma.
Subcellular Fractionation-After inducing differentiation by NGF, cells were harvested by gently scraping plates with a cell scraper, and were washed once with cold PBS. The washed pellet was subjected to one freeze-thaw cycle in liquid nitrogen. Pellets were resuspended in 1 ml of Buffer A (250 mM sucrose, 10 mM Tris-HCl, pH 7.5, 1 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, one protease inhibitor pill). Cells were then homogenized on ice using a 1-ml glass Dounce homogenizer with a tight fitting pestle until ϳ95% of cells were disrupted as judged by trypan blue staining. Crude lysates were centrifuged at 1,000 ϫ g for 15 min at 4°C to remove nuclei and unbroken cells. The supernatant was collected and the pellet (P1) discarded. The low speed supernatant was then subjected to 10,000 ϫ g centrifugation for 15 min, which yielded the 10,000 ϫ g pellet (P2). The supernatant from the P2 pellet was centrifuged at 15,000 ϫ g to completely rid the supernatant of any remaining mitochondria. Finally the 15,000 ϫ g supernatant was separated into cytosol (S3) and light membrane (P3) fractions by centrifugation at 100,000 ϫ g for 1 h. The 100,000 ϫ g supernatant was collected as the S3 fraction, and the pellet was resuspended in 40 -70 l of Buffer A. P2 was washed twice by resuspending cells in 100 l of Buffer A and pelleting (10,000 ϫ g for 15 min). After the final wash, P2 was resuspended in 50 -100 l of Buffer A. All fractionations were repeated a minimum of four times with essentially identical results. Subcellular fractions were characterized using nucleoporin (P1, nuclear), cytochrome c oxidase (P2, mitochondria, heavy ER), heme oxygenase 2 (P3, light ER), and lactate dehydrogenase (S3, cytosol).
Co-immunoprecipitation-Cell lysis buffer (150 mM NaCl, 50 mM Tris, pH 7.8, 1% Triton, 1 mM EDTA) was added to 100 g of cell lysate to bring samples to a total volume of 500 l. Samples were cleared of insoluble debris by centrifugation at 10,000 ϫ g. Anti-IP 3 R-1, pre-immune sera, mouse IgG, or anti-DANGER antibodies, and protein A-Sepharose beads were added and incubated on a rotator for 30 min at 4°C. The protein A-Sepharose beads were washed three times with lysis buffer and quenched with 20 l of SDS sample buffer. Co-immunoprecipitates were resolved by SDS-PAGE and analyzed by Western blot analysis. Figure1. DANGER interacts directly with IP 3 R. A, top: schematic of the functional domains of the IP 3 R, including the ligand binding (Bait 2), modulatory and transmembrane regions, and trypsin digest (I-V) domains. The seven regions used as bait in the yeast two-hybrid screen are also indicated below the protein and the corresponding amino acids (rat type-1 sequence). When DANGER was screened against all IP 3 R baits, only bait 4 (red) supported growth on selective media. Bottom: schematic of the conserved domains of DANGER including the partial (ϳ25%) MAB-21 domain and an ER/plasma-membrane signaling peptide (N terminus). B, left: in vitro Coomassie stain of GST fusion protein binding of purified DANGER to glutathione-Sepharose. Right: GST pulldown assays using PC12 cell lysates (100 g) blotted with mouse monoclonal (M18A10) anti-type 1 IP 3 R antibody or polyclonal anti-type 3 IP 3 R antibody demonstrates binding to purified DANGER. Bottom: Western analysis of purified type 1 IP 3 R passed over GST-DANGER columns. C, left: Western analysis of (20 g) PC12 cell lysates Ϯ full-length Myc-tagged DANGER blotted with polyclonal antibody against DANGER or anti-Myc. Right: Western analysis of (20 g) PC12 cell lysates Ϯ siRNA duplexes against rat DANGER blotted with anti-DANGER and actin as loading control. Below: peptide sequence used for polyclonal antibody production. D, co-immunoprecipitation experiments from (100 g) PC12 cell lysates of endogenous DANGER and type 1 IP 3 R, visualized by Western analysis. E, Western analysis of DANGER from subcellular fractions of PC12 cells. P1 ϭ nucleus, P2 ϭ heavy ER/mitochondria, P3 ϭ light ER, S3 ϭ cytosol. DECEMBER 1, 2006 • VOLUME 281 • NUMBER 48

JOURNAL OF BIOLOGICAL CHEMISTRY 37113
Immunofluorescence and Immunohistochemistry-PC12 cells were grown on coverslips for 56 h then fixed with 0.4% paraformaldehyde in 2ϫ PBS for 30 min and washed three times in 2ϫ PBS for 15 min. Coverslips were quenched in 50 mM NH 4 Cl for 15 min and washed three times in 2ϫ PBS for 15 min. Cells were permeabilized in saponin solution (2ϫ PBS, 1% bovine serum albumin, 1% goat serum, and 0.075% w/v saponin) for 1 h. Coverslips were washed three times in 2ϫ PBS for 15 min and subsequently inverted on 250 l of anti-DANGER and IP 3 R in a wet chamber overnight at 4°C. Coverslips were washed three times in 2ϫ PBS for 15 min and inverted on 250 l of secondary antibody solution 1:300 for 1 h at room temperature. Finally, coverslips were washed three times in 2ϫ PBS for 15 min and mounted onto slides with Slo-fade. Fluorescence signals were detected with a Zeiss LSM410 confocal laser scanning microscope, using a 63ϫ/NA 1.4 objective.
Formalin-fixed, paraffin-embedded 5 M sections were dewaxed, rehydrated, and subjected to antigen retrieval by incubation in 10 mM sodium citrate, pH 6.0, at 98°C for 20 min. Sections were then immunostained using the Vectastain TM ABC system according to the manufacturer instructions using affinity purified anti-DANGER antibody described above at a dilution of 1/50.
Antibodies and Reagents-Plasmids were from the following sources: EYFP, Matchmaker©, and Myc-CMV vector cDNA from CLONETECH and human M5 muscarinic receptor cDNA from L. Birnbaumer (National Institutes of Health). Human Northern blot from Clontech (Palo Alto, CA). Carbachol, protein A-agarose, GST-agarose, Sigma. Fura-2/AM was from Molecular Probes (Eugene, OR). UTP was from Invitrogen. siRNA duplex was from Qiagen (Valencia, CA). Ca 45 was from Amersham Biosciences.

RESULTS AND DISCUSSION
We conducted a yeast two-hybrid analysis using multiple fragments of IP 3 R as described previously (6) (Fig. 1A). A novel interacting sequence, corresponding to a previously uncharacterized expressed sequence tag designated KIAA1754 (7), bound to aa 923-1581 fragment of IP 3 R. This expressed sequence tag, which we now denote as DANGER, maps to a human genomic sequence on chromosome 10q25.1 comprising a single coding exon (4.3 kb). The open reading frame codes for a protein of 547 aa and is predicted to contain several coiled-coil regions, an N-terminal signal peptide for ER/plasma-membrane retention, and a partial MAB-21 domain (Fig. 1A, schematic).
Direct binding of DANGER to full-length IP 3 R was detected in vitro by co-immunoprecipitation with both overexpressed and endogenous DANGER proteins. We observe robust interactions between a GST-DANGER fusion protein and IP 3 R (type 1 and type 3) from cell lysates as well as purified type 1 IP 3 R from rat cerebellum (Fig. 1B). Binding to type 2 IP 3 R could not be confirmed as the type2 IP 3 R antibodies we employed failed to demonstrate immunoreactivity to the samples assayed. To study endogenous DANGER, we developed a polyclonal antibody against C-terminal aa 496 -514 of DANGER (Fig. 1C). Using a Myc-tagged DANGER expression construct in PC12 cells, the polyclonal antibody recognizes both endogenous and over-expressed DANGER at the same molecular weight (Fig.   1C) as shown by anti-Myc staining of overexpressed DANGER. Western analysis of PC12 cell lysates blotted with polyclonal anti-DANGER reveals a single major band (59 kDa) corresponding to the full-length protein as well as two other bands of lower intensity (49 and 41 kDa). To ensure the specificity of the antibody, we employed DANGER-specific siRNA to deplete endogenous DANGER from PC12 cells (Fig. 1C). The siRNA abolishes DANGER immunoreactivity indicating that the three bands observed by our antibody arise from the same mRNA. Next, we tested the interaction of endogenous DANGER and IP 3 R in intact cells. We observed co-immunoprecipitation of DANGER with IP 3 R using either our polyclonal antibody to DANGER or a monoclonal antibody to type 1 IP 3 R (Fig. 1D). Subcellular fractionation of PC12 cells (Fig. 1E) reveals the greatest enrichment of DANGER in fractions P2 and P3, which contain ER membranes, with lower levels in the crude nuclear P1 fraction.
The DANGER clone isolated from the yeast two-hybrid screen comprises aa 259 -547, which contain a partial MAB-21 region. We attempted to map the site in DANGER that binds IP 3 R more extensively, using yeast two-hybrid and GST fusion proteins of DANGER (data not shown). Thus far we are unable to identify binding to IP 3 R with any fragment of DANGER smaller than the one isolated from the yeast two-hybrid screen suggesting that the tertiary structure of the MAB21-containing region is critical for binding. To more carefully characterize the intracellular localization of DANGER, we performed immunofluorescence using Alexa-488-conjugated DANGER antibody and Alexa-568-conjugated type 1 IP 3 R antibodies in PC12 cells treated with NGF for 2 days. We observe a membrane staining pattern for endogenous DANGER, which has a high degree of co-localization with type 1 IP 3 R (54% Ϯ 8% concurrence in 50 cells measured) (Fig. 2, A-C). We also observe a small amount of DAN-GER staining in the nucleus, consistent with the results from the subcellular fractionation. Higher magnification reveals substantial co-localization of DANGER with type 1 IP 3 R near the plasma membrane (Fig 2C1). DANGER and type 1 IP 3 R also co-localize in NGF-induced neurites to a similar extent as in the cell body (46% Ϯ 11%) (Fig.  2C2). DANGER is concentrated in cerebellar Purkinje cells, which contain the highest densities of IP 3 R in the brain (Fig. 2D). Whereas cerebellar immunoreactivity for IP 3 R is confined to Purkinje cells, DANGER is also concentrated in basket cells. DANGER is also expressed throughout the body in numerous terminally differentiated cell types including neurons, neuroendocrine cells, crypt cells, muscle cells, B-cells, and T-cells (data not shown).
To ascertain a possible modulatory role of DANGER on IP 3 R function, we monitored IP 3 -induced release of Ca 2ϩ from microsomal preparations of COS7 cells transfected with type 1 IP 3 R and SERCA2b (Fig. 3, A-C) as described previously (4). Compared with GST alone (black), incubation with purified GST-tagged DANGER (500 nM) has little effect on IP 3 -induced Ca 2ϩ release in the presence of 300 nM Ca 2ϩ over a wide range of IP 3 concentrations (Fig. 3A). In addition, overexpression or depletion (siRNA) of DANGER does not alter IP 3 production in response to muscarinic and purinergic receptor stimulator (data not shown).
At low Ca 2ϩ (Ͻ150 -200 nM) concentrations, Ca 2ϩ enhances IP 3 R channel activity, while supraphysiologic concentrations of Ca 2ϩ inhibit IP 3 R activity, reflecting feedback regulation (8). DANGER (500 nM) has no effect on the stimulatory actions of Ca 2ϩ but markedly enhances the inhibitory effects of elevated Ca 2ϩ levels (Fig. 3C). Thus, in the absence of DANGER, moderately high Ca 2ϩ levels (0.8 and 3.8 M) do not influence IP 3 R responses to 100 nM IP 3 ; however, in the presence of DANGER (500 nM), IP 3 -mediated release of Ca 2ϩ is virtually abolished at these Ca 2ϩ levels. Similar but lesser effects are observed with 100 nM DANGER (data not shown). DANGER-elicited alterations on IP 3 R-mediated Ca 2ϩ release do not appear to result    DECEMBER 1, 2006 • VOLUME 281 • NUMBER 48 JOURNAL OF BIOLOGICAL CHEMISTRY 37115 from changes in binding affinity of IP 3 for IP 3 R as DANGER does not affect IP 3 binding affinity (Fig. 3B).

DANGER Modulates IP 3 R Ca 2؉ Dependence
We further examined the effects of DANGER on activity of IP 3 R by performing single IP 3 R channel recordings with patch clamp electrophysiology. We utilized endogenous Sf9 IP 3 R in the outer membrane of freshly isolated nuclei (5,9). The cytoplasmic face of the channel is exposed to the pipette solution containing 7 M Ca 2ϩ and 100 nM InsP 3 . In this preparation, the channel open probability (P o ) is ϳ0.7 (Fig. 4), as observed previously (9). Inclusion of 1 M DANGER in the pipette solution reduces channel P o by ϳ50%. Channel P o is diminished primarily as a consequence of an increase of the mean closed time (t c ), indicating that DANGER decreases the channel opening rate. In contrast, the rate of channel inactivation, monitored by the channel activity duration T a and the mean number of channels activated in each patch, N A , is unaffected by DANGER (Fig. 4).
In summary, DANGER is a novel membrane-associated IP 3 R binding protein that enhances the sensitivity of IP 3 R to inhibition by Ca 2ϩ . By contrast, cytochrome c and Bcl-XL, which also modulate Ca 2ϩ sensitivity of IP 3 R, decreases the inhibitory actions of high Ca 2ϩ (4,5). What might be the major physiologic role of DANGER? Its enhancement of Ca 2ϩ inhibition of IP 3 R channel function is striking. In microsomal membranes treated with DANGER, IP 3 R channel function decreases by 90% as [Ca 2ϩ ] increases from 300 to 800 nM, reflecting a highly cooperative process. This suggests that in DANGER-enriched cells, Ca 2ϩ is a major regulator of channel opening. DANGER occurs in numerous terminally differentiated tissues (data not shown), but its marked enrichment in cerebellar Purkinje cells is unique among neuronal lineages. IP 3 R in Purkinje cells are highly insensitive to IP 3 (10 -12). This decreased sensitivity fits well with the influences we have observed of DANGER on IP 3 R channel function, which may reflect allosteric modulation of the Ca 2ϩ inhibition site. Last, many of the proteins which have been demonstrated to bind to the IP 3 R have not been tested for their ability to alter IP 3 R activity over a large range of external Ca 2ϩ concentrations. Thus it may be that regulation of IP 3 R Ca 2ϩ dependence via IP 3 R binding partners may be a more common function than has been described previously.