The Pro335 --> Leu polymorphism of type 3 inositol 1,4,5-trisphosphate receptor found in mouse inbred lines results in functional change.

Inositol 1,4,5-trisphosphate receptor (IP3R) is an intracellular Ca2+ channel involved in various cellular signaling. Type 3 IP3R (IP3R3) retains ligand-gated Ca2+ channel properties differing from other subtypes in terms of IP3-binding affinity and regulation of its channel activity by effector molecules. In this study, we found the natural Pro335 --> Leu polymorphism of mouse IP3R3 between BALB/c and C57BL/6J. We investigated the functional differences between Pro335IP3R3 and Leu335IP3R3 with purified receptors reconstituted into proteoliposomes as well as with soluble ligand binding domains. Pro335IP3R3 exhibited significantly higher IP3-binding affinity and IP3-induced Ca2+ release than those of Leu335IP3R3 in both forms of the receptor. Moreover, the polymorphic change caused differences in the effect of external Ca2+ on IP3-induced Ca2+ release. The Pro335 --> Leu substitution alters the conformation of soluble ligand binding domain as revealed by intrinsic fluorescence and circular dichroism spectra with or without Ca2+. The results indicate that the polymorphism of IP3R3 causes changes in receptor function, presumably affecting intracellular Ca2+ signaling.

Inositol 1,4,5-trisphosphate (IP 3 ) 1 is a well known second messenger that is produced by hydrolysis of phosphoinositol 4,5-bisphosphate involving phospholipase (1). IP 3 plays a major role in mediating the release of Ca 2ϩ from the intracellular storage, thereby regulating numerous physiological processes in the cell (2,3). The inositol 1,4,5-trisphosphate receptor (IP 3 R) is an intracellular Ca 2ϩ channel that releases Ca 2ϩ from the endoplasmic reticulum, an internal calcium store, by various stimuli such as neurotransmitters, hormones, and growth factors (4). IP 3 Rs are encoded by three different genes and variably expressed in different cell types (5). IP 3 R forms a tetramer with each subunit being composed of an N-terminal ligand-binding domain (LBD), an intervening regulator, and a C-terminal channel (6,7).
The type 1 IP 3 R (IP 3 R1) is mainly expressed in the brain (8), whereas the type 3 IP 3 R (IP 3 R3) is the major isoform found in the intestine, kidney, and pancreatic islets (9,10). Two Ca 2ϩbinding sites were identified in the LBD of IP 3 R1 (11). Although Ca 2ϩ binding to the LBD of IP 3 R3 has not been reported, the inhibition of free Ca 2ϩ on the IP 3 binding of LBD of the type 3 receptor suggests that the domain also has an intrinsic Ca 2ϩ -binding property (12). IP 3 R3 shows distinct properties of Ca 2ϩ signaling upon stimulation in the cell as observed in other subtypes (13,14). For instance, DT 40 cells expressing only IP 3 R3 exhibit a single Ca 2ϩ peak upon stimulation of the B-cell receptor, whereas type 1 or type 2 generates Ca 2ϩ oscillation with the same signal (13). In IP 3 R1-knockdown, COS-7 cells, expressing primarily the type 3 receptor, also exhibited a single Ca 2ϩ peak, whose amplitude was reduced relative to that of the untreated cell (14).
Cytosolic Ca 2ϩ is an important modulator of the Ca 2ϩ channel activity of IP 3 Rs (15). The Ca 2ϩ dependence on the channel activity of IP 3 R3 is, however, still controversial. Study with vesicles from RIN-5F cells mainly expressing IP 3 R3 revealed that external Ca 2ϩ did not inhibit Ca 2ϩ release of IP 3 R3 (16), whereas the Ca 2ϩ channel activity of IP 3 R3 was shown in other studies to be biphasically modulated by Ca 2ϩ (17,18). A study using a microsome purified from COS-7 cells expressing IP 3 R3 (17) and a patch clamping with Xenopus oocyte demonstrated the biphasic dependence of the channel activity by Ca 2ϩ (18). The reason for the difference in Ca 2ϩ effect for IP 3 R3 is unclear, although it is likely that the sample preparations and membrane compositions are different.
The specificity of IP 3 R3 function is largely unknown. We observed that BALB/c (BC) mice exhibited a higher defecation than C57BL/6J (B6) mice under stress conditions (19). It has been known in Caenorhabditis elegans that IP 3 R regulates the defecation rhythm (20). Thus, it might be intriguing to speculate that the defecation behavior is associated with the functional polymorphism of IP 3 R3 found in BC and B6 mice. On the other hand, the major subtype of mammalian IP 3 R in the intestine is known to be the type 3 receptor (9,10). In this study, we discovered a polymorphism of IP 3 R3 causing a change in amino acids, and thus we investigated the functional consequences of receptor alteration in terms of IP 3 and Ca 2ϩ bindings. The results indicate that functional difference exists between receptor variants when tested with their ligand binding domains as well as the whole receptor.

MATERIALS AND METHODS
Mice-Inbred strains of mice, BC and B6, were purchased from Daehan Biolink Co. Ltd. (Korea). All mice were raised in housing with five per cage in a specific pathogen-free and temperature-controlled facility at 22°C under a 12-h light-dark cycle with light on at 07:00. Humidity was maintained at 55% with food and water freely available. Male mice at the age of 5-10 weeks were used for this study. All experimental procedures with animals followed the National Institutes of Health Guideline for the Care and Use of Laboratory Animals.
Full-length Nucleotide Sequences of Mouse IP 3 R3-The full-length sequences of mouse IP 3 R3 cDNA were obtained from plasmid, G431004L16 (RIKEN). Based on the nucleotide sequences obtained, we designed specific primers for mouse IP 3 R3. The entire coding sequences of IP 3 R3 were compared between BC and B6 mice by RT-PCR. Briefly, total RNA was extracted mouse brain followed by reverse transcription. After PCR amplification of an ϳ2-3 kb DNA fragment with specific primers, the DNA fragments were sequenced on an ABI 3700 sequencer (PerkinElmer Life Sciences).
Ratio of Semi-quantitative RT-PCR-The ratio of semi-quantitative RT-PCR was conducted as described previously with minor modifications (21). After 2 g of total RNA were purified from whole brains and intestine, first strand cDNAs were synthesized by using the total RNA as template. PCR amplification was performed with the first cDNAs as templates. PCR conditions were as follows: 1 cycle at 94°C for 3 min, followed by 27 cycles at 94°C for 30 s, 55°C for 1 min (for subtypes of I/III), 52°C for 1 min (for subtypes of II/III), or 72°C for 1 min. The oligonucleotides used for PCR were as follows: for type I/III, 5Ј-GT(C/-G)ATCATCGACACCTTTGC-3Ј and 5Ј-GCT(G/C)ACCAGGGACATGG-C-3Ј; and for type II/III, 5Ј-T(C/T)ATCGTCATCATCATCG-3Ј and 5Ј-C-(G/T)TGATCATCTGAGCCAC-3Ј. The PCR products were electrophoresed on a 1.5% agarose gel and then transferred to nylon membranes. Blots were hybridized with 32 P-labeled type-specific oligonucleotide probes (type I, 5Ј-AGGGAAAGAAACCTTGAT-3Ј, type II, 5Ј-GAA-GGTGAAGGACCCGAC-3Ј, or type III, 5Ј-CCGAGTCAAGAACAAGAC-3Ј). To test the specificity of the oligonucleotide probe, plasmids containing the PCR product of each subtype were used as positive and negative controls. The control plasmids were constructed by ligating PCR products amplified with primer pairs, I/III or II/III, into the pGEM-T Easy vector. For normalizing the IP 3 R signals, glyceraldehyde-3phosphate dehydrogenase was amplified with specific primers (forward primer, 5Ј-ACCACAGTCCATGCCATCAC-3Ј, and reverse primer, 5Ј-T-CCACCACCCTGTTGCTGTA-3Ј), and then glyceraldehyde-3phosphate dehydrogenase signals detected by Southern blot analysis were used as standard. The radioactivity of the membrane was analyzed by a PhosphorImager (Fuji).
Construction of pGEX-LBD and pET-LBDHis Vector-The cDNA encoding the N-terminal 605 amino acids of the mouse Pro 335 IP 3 R3 was amplified by PCR with the forward primer ATCCCGGGAATGAAT-GAAATGTCCAGC and the reverse primer GCGAATTCTTAAAGCTTC-CGGTTGTTGCAG, by using plasmid G431004L16 (from RIKEN) as a template, and subcloned into the pGEM-T Easy vector. As there is an insertion mutation in the 1588th nucleotide position of the clone, G431004L16, a DNA fragment containing this region (nucleotide position, 1189 -1815) was changed with a DNA fragment that was amplified by RT-PCR using total RNA extracted from brains of B6 mice. For construction of pGEX-P 335 LBD, the SmaI and EcoRI fragment from pGEM-T-P 335 LBD was cloned into SmaI/EcoRI sites of pGEX-2T. For construction of pGEX-L 335 LBD, DNA fragment (nucleotide position, 505-1815) was amplified from RT-PCR by using first strand cDNA synthesized from total RNA extracted from brain of B6 mice. The HindIII fragment isolated from amplified DNA was cloned into pGEX-P 335 LBD. To construct pET-LBDHis, PCRs were performed with the forward primer GCATATGAATGAAATGTCCAGC and the reverse primer GCTCGAGAAGCTTCCGGTTGTTGT by using plasmid pGEX-P 335 LBD or pGEX-L 335 LBD as a template, and then amplified DNA fragments were subcloned into pET-21b plasmid after NdeI/XhoI digestion. All PCR products and the polymorphic site were confirmed by DNA sequencing.
Purification of Recombinant GST-LBD and LBD-His Protein-Bacterial cultures of pGEX-LBD were grown at 37°C in a 500-ml volume of LB/ampicillin (100 g/ml) until an A 600 reached 0.5. Induction of the GST fusion protein was carried out by the addition of isopropyl ␤-Dthiogalactopyranoside (0.5 mM) and further incubation for 3 h. The bacteria were pelleted by centrifugation and resuspended in lysis buffer (50 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1 mM PMSF, 5 g/ml benzamidine, leupeptin, and pepstatin). The cells were lysed by passing through an Amicon French pressure cell. Cell lysates were applied to a column of glutathione-agarose pre-equilibrated with lysis buffer. The column was washed with lysis buffer and then eluted with elution buffer (50 mM Tris-HCl, pH 8.0, 10 mM reduced glutathione). The expression of pET-LBDHis was performed as described previously (22). To purify recombinant LBD-His proteins, harvested bacterial pellets were resuspended in the binding buffer (20 mM Tris-HCl, pH 8.0, 500 mM NaCl, 5 mM imidazole, 1 mM PMSF, 5 g/ml of benzamidine, leupeptin, and pepstatin). The cells were lysed by passing through an Amicon French pressure cell. Cell lysates were applied to a nickel-nitrilotriacetic acidagarose column pre-equilibrated with the binding buffer. The column was washed with binding buffer and washing buffer (20 mM Tris-HCl, pH 8.0, 500 mM NaCl, and 60 mM imidazole). The LBD-His proteins were eluted in the binding buffer containing 250 mM imidazole. The concentrations of recombinant proteins were quantified using bicinchoninic acid according to the manufacture's instruction (Pierce).
Partial Purification of IP 3 R3 from Intestine-Mouse intestines were mixed with 3 volumes of buffer A (50 mM Tris-HCl, pH 7.4, 0.32 M sucrose, 1 mM EDTA, 1 mM ␤-mercaptoethanol, 0.1 mM PMSF, 10 M leupeptin, and 10 M pepstatin) and were broken in a blender, followed by homogenization in a glass-Teflon homogenizer. The homogenates were then centrifuged at 100,000 ϫ g for 1 h to precipitate the membrane pellet. The pellet were resuspended in buffer B (50 mM Tris-HCl, pH 8.0, 1 mM EDTA, 1 mM ␤-mercaptoethanol, 0.1 mM PMSF, 10 M leupeptin, and 10 M pepstatin) containing 1% (w/v) Triton X-100 to give the membrane protein concentration of ϳ2 mg/ml. The membrane solution was stirred for 1 h and then centrifuged at 32,000 ϫ g for 1 h at 4°C. The supernatant obtained was mixed with an equal volume of buffer C (20 mM Tris-HCl, pH 7.5, 0.1 M NaCl, 1% Triton X-100, 1 mM ␤-mercaptoethanol, 0.1 mM PMSF, 10 M leupeptin, and 10 M pepstatin) and applied to a type 3 IP 3 R antibody-coupled immunoaffinity column equilibrated with buffer C as described above. The proteinloaded column was washed with 20 bed volumes of buffer C to remove unbound proteins, and the IP 3 R was eluted by 10 ml of elution buffer D (0.1 M glycine, pH 2.8, 0.2% (w/v) Triton X-100, 0.5 M NaCl, 1 mM ␤-mercaptoethanol, 0.1 mM PMSF, 10 M leupeptin, and 10 M pepstatin). The eluate was immediately neutralized by adding 1 M Tris-HCl, pH 9.5, mixed with an equal volume of buffer E (50 mM Tris-HCl, pH 8.0, 0.2% Triton X-100, 0.5 M NaCl, and 1 mM ␤-mercaptoethanol), and then applied to a benzamidine-Sepharose column equilibrated with buffer E to remove any residual proteases from the IP 3 R sample. The IP 3 R containing flow-through was collected and stored at Ϫ70°C until use.
Reconstitution of IP 3 R into Proteoliposomes-Phospholipids were purchase from Avanti Polar Lipids. Chloroform solutions of lipids were stored in sealed ampules under argon gas at Ϫ20°C. Phosphatidylcholine (from bovine brain), phosphatidylethanolamine (from bovine brain), and phosphatidylserine (from bovine brain) dissolved in chloroform were mixed to give a molar ratio of 50, 30, and 20%, respectively. The final lipid concentration was 5 mM in a total volume of 1 ml. The solvent was evaporated under a stream of argon gas, and the residual chloroform was removed by speed vacuuming. The dry lipids were hydrated in buffer F (20 mM HEPES, pH 7.5, 100 mM NaCl, 1 mM CaCl 2 , 1% CHAPS) containing about 5 g/ml IP 3 R. The mixtures were dialyzed for 72 h against an excess volume of buffer G (buffer F without CHAPS). The resulting proteoliposomes were pelleted by centrifugation at 100,000 ϫ g for 30 min at 4°C, washed with buffer H (20 mM HEPES, pH 7.5, 100 mM NaCl), and then dialyzed against buffer E for 24 h at 4°C. The resulting proteoliposomes were passed through Chelex 100 to remove free Ca 2ϩ . The formation of proteoliposomes was monitored by measurement of light scattering during analysis with a spectrofluorometer (excitation and emission wavelength of 450 nm). The average diameter of the liposomes was about 360 when assayed as described previously (23). The amounts of reconstituted IP 3 R were determined using NanoOrange® protein quantitation kit (Molecular Probes). After the reconstitution, the ratio of phospholipid concentration (w/v) to that of protein (L/P) was determined as 890 Ϯ 27 for Pro 335 IP 3 R3 and 879 Ϯ 22 for Leu 335 IP 3 R 3 in reaction samples, respectively. IP 3 -induced Ca 2ϩ Release Measurement Using Indo-1 Fluorescence-Ca 2ϩ efflux from the proteoliposomes was observed by measuring the fluorescence changes of external indo-1. Fluorometric measurements were performed at 30°C by using a Shimadzu RF-5301 PC spectrofluorometer. The fluorescence intensity was measured at the emission wavelength of 393 nm under an excitation wavelength of 355 nm. The fluorescent intensities of 10 M indo-1 were calibrated to free Ca 2ϩ concentration using the Ca 2ϩ -EGTA buffering system (24). To quantify the amount of released Ca 2ϩ from proteoliposomes by IP 3 binding, the fluorescence intensity of indo-1 after addition of IP 3 was compared with the fluorescence intensity after addition of Triton X-100 (1% (w/v), final concentration) instead of IP 3 .
Effect of External Ca 2ϩ on the IP 3 -mediated Ca 2ϩ Release Measurement from the Proteoliposomes Using 45 Ca 2ϩ -The proteoliposomes were produced in the presence of 45 Ca 2ϩ to include ϳ20,000 cpm of 45 Ca 2ϩ in buffer H (20 mM HEPES, pH 7.5, 100 mM NaCl, 1 mM ␤-mercaptoethanol) according to the procedure described. To remove residual Ca 2ϩ bound to the vesicle surface, the sample was applied to Sephadex G-25 column equilibrated with buffer H, and the liposomes were pelleted by centrifugation (100,000 ϫ g, 30 min, 4°C). The pellet was then redissolved and dialyzed against excess volume of buffer H for 12 h at 4°C. The proteoliposomes were mixed with each indicated concentration of CaCl 2 and incubated for 10 min at 30°C. After further incubation of the sample for 10 min in the presence of 1 M of IP 3 in the reaction mixtures, the sample was diluted with buffer I (buffer H plus 1.5 M KCl). The liposomes were pelleted by centrifugation (100,000 ϫ g, 30 min, 30°C). The pellet was then dissolved with 1% (w/v) Triton X-100, and the radioactivity of each fraction (pellet and supernatant) was determined by scintillation counting.
Removal of Ca 2ϩ Contamination-Removal of Ca 2ϩ contamination was conducted according to the method described previously (25). Ca 2ϩ contamination during all experiments was checked using the fluorescence of the Ca 2ϩ indicator, indo-1, before measurements.
Preparation of IP 3 R3 Antibody-coupled Column-Type 3 IP 3 R antibody-coupled column was prepared by coupling 0.6 mg of affinitypurified anti-peptide type 3 IP 3 R antibody (Sigma) to 1.2 ml of the immobilized protein A resin from the ImmunoPure protein A IgG orientation kit (from Pierce) according to the instructions. The column was stored in 20 mM Tris-HCl, pH 7.5, containing 0.02% sodium azide until use.

H]IP 3 Binding to Recombinant GST-LBD Fusion
Protein or Reconstituted IP 3 R3-10 g/ml GST-LBD fusion protein or 2.8 g/ml IP 3 R3 in buffer J (buffer H plus 1 mM EGTA, 1 mM ␤-mercaptoethanol) was incubated with various concentrations of IP 3 containing 1:1000 as much [ 3 H]IP 3 . Binding experiments were carried out in the absence of free external Ca 2ϩ . After 20 min of incubation at 30°C, a 0.2 volume of each sample was filtered through a spun concentrator with molecular weight cut-off of 30,000. The radioactivity of each filtrate was determined by a liquid scintillation counter and compared with the control values without proteins. The same method was applied to measure Ca 2ϩ binding on the recombinant GST-LBDs.
Effect of External Ca 2ϩ on the IP 3 Binding to Recombinant GST-LBD-10 g/ml IP 3 R-GST fusion protein was incubated with various concentrations of Ca 2ϩ in the presence of a fixed concentration of IP 3 containing 1:1000 as much [ 3 H]IP 3 . After 20 min of incubation at 30°C, a 0.2 volume of each sample was filtered through a spun concentrator, and the radioactivity of filtrate was determined as described.
Fluorescence and CD Spectroscopy-To measure the intrinsic fluorescence of Trp residues in the recombinant LBD-His protein (1 M), the emission intensity in the range 310 -400 nm was recorded under 295 nm of excitation wavelength using a Shimadzu RF-5301 PC spectrofluorometer at 25°C. CD spectra in the far-UV and near-UV regions were monitored at room temperature with a Jasco J715 spectropolarimeter (Japan Spectroscopic, Tokyo); the optical path lengths were 0.1 and 1. SDS-PAGE and Immunoblotting-Immunoprecipitated IP 3 R samples were dissolved with a 1% SDS solution and then analyzed by 7% SDS-PAGE. This was followed by immunoblotting and detection with anti-IP 3 R antibody.

Polymorphism of IP 3 R3 Found in BC and B6 Mice-
The coding sequences of the IP 3 R3 of BC and B6 mice were compared with the search for polymorphism that may potentially be associated with functional variation. The open reading frame consists of 8013 bp and 2670 encoded amino acids, showing 95 and 92% identity with rat and human orthologs, respectively. A total of 11 nucleotide polymorphisms were founded in the open reading frame. One of the polymorphisms at the 335th position resulted in a change in amino acid from proline to leucine (Fig. 1A). Sequencing of genomic DNA obtained from the two mouse strains confirmed the polymorphism. The change is located in the N-terminal LBD (amino acids 1-605) with variations among species (Fig. 1B). On the basis of the crystal structure of the IP 3 R1 LBD core (amino acids 224 -604), we were able to position the 335th residue in a loop between the ␤6 and ␤7 strands (26). The LBDs seems to have similar structure regardless of the receptor subtypes, because the sequence homologies among them are fairly high, e.g. 68% identity between types 1 and 3 in the LBD core sequence.
In order to compare the relative expressions of three IP 3 Rs and the ones between two mouse strains, semi-quantitative RT-PCR (21) was conducted with specific primers designed to amplify two isoforms simultaneously. The expression levels of IP 3 Rs of BC mice were ϳ30% higher than those of B6 mice in the brain regardless of the receptor type ( Fig. 2A). In the intestine, type 1-and type 2-specific bands were not detected in both mice. On the other hand, IP 3 R3 mRNA was expressed more in the BC intestine than that of B6 by ϳ60%. A Western blotting analysis after immunoprecipitation using a specific antibody against IP 3 R3 showed that more IP 3 R3 proteins are expressed in the BC intestine than that of B6 (Fig. 2B), which is consistent with the results of semi-quantitative RT-PCR. The results confirmed that IP 3 R3 is the major isoform expressed in the mouse intestine, with a slight variation in their expressions among inbred mice.
IP 3  whether the polymorphism affects the intrinsic function of IP 3 R3, an IP 3 -binding assay was performed with purified Pro 335 IP 3 R3 and Leu 335 IP 3 R3 reconstituted into proteoliposomes. Considering that IP 3 R3 was highly expressed (10) and the expression of type 1 or 2 IP 3 R was not detected in intestines by the semi-quantitative RT-PCR ( Fig. 2A), the purified IP 3 R was expected to be mostly type 3. From the results, Pro 335 IP 3 R3 showed 2.4-fold higher IP 3 -binding affinity than that of Leu 335 IP 3 R3 (Fig. 3A). In particular, the difference was more significant at values less than 1 M IP 3 , which is the range of physiological concentration after stimulation (27,28). Above 1 M IP 3 concentration, both IP 3 R3s exhibited similar IP 3 -binding affinities. The IP 3 dissociation constants (K d ) for Pro 335 IP 3 R3 and for Leu 335 IP 3 R3 were estimated to be 301.5 Ϯ 28.9 and 722.4 Ϯ 27.0 nM, respectively. Hill coefficients of both Pro 335 IP 3 R3 and Leu 335 IP 3 R3 were 1.1 and 2.7, respectively. These values indicate that IP 3 binds to each subunit of Pro 335 IP 3 R3 independently, whereas IP 3 binding to Leu 335 IP 3 R3 exhibits a positive cooperativity.
As we observed an alteration in IP 3 -binding affinity, we further examined whether the polymorphism influences IICR. In a previous study, purified bovine IP 3 R1 reconstituted into proteoliposomes showed Ca 2ϩ release in dose-dependent manner of IP 3 (29). The dose-dependent Ca 2ϩ release of IP 3 from reconstituted proteoliposomes in each IP 3 R3 was compared. As shown in Fig. 3B, Pro 335 IP 3 R3 released a significantly higher amount of Ca 2ϩ than that of Leu 335 IP 3 R for all IP 3 concentrations tested. In particular, the difference was remarkable in the range less than 3 M IP 3  Ca 2ϩ Effects on IICR of Pro 335 -and Leu 335 IP 3 R3-External Ca 2ϩ is a well known modulator of IP 3 R3 Ca 2ϩ channel activity (15). Two Ca 2ϩ -binding sites were identified previously in LBD of IP 3 R1 exposed to cytosol, and the residues were conserved in LBD between IP 3 R1 and IP 3 R3 (11). As the polymorphism is located near the putative Ca 2ϩ -binding sites, we investigated the effects of external Ca 2ϩ on IICR of both receptors. As shown in Fig. 4, external Ca 2ϩ differentially modulated the Ca 2ϩ release of each receptor in the presence of 1 M IP 3 , showing a maximum difference in IICR, as illustrated in Fig. 3B. In the case of Pro 335 IP 3 R3, IICR exhibited a characteristic biphasic pattern with a maximum release of Ca 2ϩ at 1 M of Ca 2ϩ . A further decrease or increase in external Ca 2ϩ concentration reduced the IICR of Pro 335 IP 3 R3 from the maximum value. In contrast, IICR of Leu 335 IP 3 R3 was monotonically inhibited up to 3 M Ca 2ϩ . It has been suggested that a biphasic mode of Ca 2ϩ on the channel activity of IP 3 R is important in generating characteristic Ca 2ϩ signaling within the cells (30,31). Therefore, the present biphasic regulation by Ca 2ϩ suggests that Pro 335 IP 3 R3 shows different Ca 2ϩ signaling from that of Leu 335 IP 3 R3 in vivo. IP 3 -binding Affinities and Ca 2ϩ Effects on IP 3 Binding of Recombinant Pro 335 and Leu 335 of IP 3 R3 LBD-In order to investigate further the different modes of IP 3 bindings between the receptors, we constructed bacterial expression vectors containing 605 amino acid sequences of each LBD fused to GST. In a previous study, LBD (amino acids 1-576) of rat IP 3 R3 fused to GST was expressed in bacteria and retained proper IP 3binding affinity (10). Fig. 5A shows that IP 3 -binding affinity on

FIG. 2. Expression of IP 3 R3 between B6 and BC mice.
A, ratio of semi-quantitative RT-PCR of total RNA extracted from brain and intestine. After first cDNA was synthesized with total RNA extracted from the indicated tissues, PCR was conducted with primer pairs (type I/III or type II/III) using the cDNA as a template. The subtype-specific signals were detected by Southern blotting with 32 P-labeled type-specific probes. RT and w/o RT indicate that RT-PCRs were performed with and without reverse transcriptase, respectively. EcoRI-digested plasmids containing the PCR fragment of each subtype were loaded on the subtype lanes. B, Western blot analysis of immunoprecipitated intestine extracts with anti-IP 3 R3 antibody. Lanes 2 and 4 are loaded double quantities of immunoprecipitated samples compared with lanes 1 and 3, respectively. A and B represent typical results of four and three experiments, respectively. Br, brain; Int, intestine; G3PDH, glyceraldehyde-3-phosphate dehydrogenase.
the Pro 335 GST-LBD (BC allele) was about 2.4-fold higher than that of Leu 335 GST-LBD (B6 allele) in the absence of free Ca 2ϩ . Again, the difference found in the submicromolar range was remarkable. The result is similar to that obtained from native IP 3 R3s. The K d values of Pro 335 GST-LBD and Leu 335 GST-LBD were 95.4 Ϯ 2.9 and 229.4 Ϯ 11.0 nM, respectively. The K d values of two LBDs were smaller than those of native IP 3 R3. However, the fold difference between the two recombinant proteins was consistent with that of the native receptor.
In the reconstitution study, external Ca 2ϩ induced a different IICR between Pro 335 -and Leu 335 IP 3 R3. Therefore, we examined the effect of polymorphism on IP 3 binding to GST-LBDs by external Ca 2ϩ fixed at 0.4 M IP 3 , which displayed the most significant difference in IP 3 binding, as described in Fig. 5A. The results indicated that IP 3 binding to Pro 335 IP 3 R3GST-LBD was slightly increased at 0.5 M Ca 2ϩ and then decreased up to 4 M Ca 2ϩ (Fig. 5B). In contrast, the binding to Leu 335 GST-LBD was inhibited progressively with increasing Ca 2ϩ concentrations. The external Ca 2ϩ -induced inhibition was 53% for Pro 335 GST-LBD and 64% for Leu 335 GST-LBD at a 4 M concentration. These results indicate that the single polymorphism causes different Ca 2ϩ effects on IP 3 binding to the LBD.
Ca 2ϩ Binding to the LBDs and Change in Intrinsic Fluorescence of LBDs by Ca 2ϩ -In an effort to investigate further the differential IP 3 binding by external Ca 2ϩ to both LBDs, we compared 45 Ca 2ϩ -binding affinity between two LBDs. From the results of the 45 Ca 2ϩ -binding assay, the K d values of Pro 335 GST-LBD and Leu 335 GST-LBD were found to be nearly identical at 260.4 Ϯ 25.4 and 287.7 Ϯ 51.4 nM, respectively (Fig.  6A). This indicates that the polymorphism does not affect the intrinsic Ca 2ϩ -binding property of the LBD.
Next, Ca 2ϩ -induced conformational changes of the two recombinant LBDs were examined using intrinsic emission fluorescence. The LBD of IP 3 R3 contains 7 tryptophan and 17 tyrosine residues. Pro 335 -and Leu 335 GST-LBD showed maximum emission peaks ( max ) at 355 and 356 nm, respectively, under an excitation at 280 nm (result not shown). Fig. 6B shows that the intensities at max gradually increased with increasing Ca 2ϩ concentration. The fluorescence values of Pro 335 -and Leu 335 GST-LBD were increased by 75 and 35%, respectively, compared with the Ca 2ϩ -free case. It suggests that both LBDs have different conformations upon Ca 2ϩ binding. To obtain the binding affinity for the Ca 2ϩ (K Ca ) and the Hill coefficient, the average fluorescence changes versus calcium concentration were plotted and fitted by using the Hill equation (Fig. 6B, inset). The K Ca values of Pro 335 GST-LBD and Leu 335 GST-LBD were estimated at 549.3 Ϯ 27.5 and 791.6 Ϯ 58.3 nM, respectively. The Hill coefficients of both LBDs were comparable (Pro 335 GST-LBD n ϭ 1.58; Leu 335 GST-LBD n ϭ 1.44). This suggests that more than one molecule of Ca 2ϩ binds to one LBD with a positive cooperativity, which is in agreement with the previous result that the LBD of IP 3 R1 has two Ca 2ϩ -binding sites (304 -381 and 378 -450) (11).
Structural Difference of Pro 335 -and Leu 335 LBD-His upon Ca 2ϩ Binding-The fluorescence results suggest that the two LBDs assume different conformations upon Ca 2ϩ bindings. In order to investigate further the structural differences between Any difference in the secondary structure of LBD-His proteins was not detected in the far-UV region without Ca 2ϩ (Fig.  7A). The ␣-helical content estimated from mean residue ellipticity ([] R ) at 222 nm (32) was 34.4 and 32.8% with Pro 335 -and Leu 335 LBD-His, respectively. In addition, the far-UV CD spectra of both proteins were not significantly different in the presence of 1 mM Ca 2ϩ (results not shown). This indicates that both proteins have a similar secondary structure regardless of the presence or absence of Ca 2ϩ . CD spectra in the near-UV region differed from those of far-UV regions (Fig. 7, B and C). The near-UV CD spectra of Pro 335 -and Leu 335 LBD-His showed a marginal difference in the absence of Ca 2ϩ (Fig. 7B). In the presence of 1 mM Ca 2ϩ , on the contrary, the CD spectra were significantly different from each other, with Pro 335 LBD-His having a more negative molar ellipticity in the near-UV region (especially in the region of 280 nm) than that of Leu 335 LBD (Fig. 7C). The near-UV CD spectra reflect the contributions of aromatic side chains of amino acids and disulfide bonds. Therefore, these results imply that Pro 335 LBD-His has a more ordered tertiary structure in the vicinities of aromatic residues than that of Leu 335 LBD-His in the presence of 1 mM Ca 2ϩ , even though both species have a similar secondary structure. DISCUSSION For the Pro 335 3 Leu polymorphism found in the mouse IP 3 R3, we investigated their functional differences with the whole receptors and their LBDs. The polymorphism affects IP 3 -binding affinity and IICR of the receptor, i.e. Pro 335 IP 3 R3 exhibited higher IP 3 -binding affinity and IICR than those of Leu 335 IP 3 R3. In particular, the difference was noted in the range of submicromolar IP 3 concentrations. It has been reported that IP 3 concentration reaches ϳ100 -300 nM during stimulation in lymphocytes (27,28). Therefore, Pro 335 IP 3 R3 may respond to stronger cellular Ca 2ϩ signals than Leu 335 IP 3 R3 upon stimulation. In addition, Pro 335 IP 3 R3 exerted a distinct biphasic inhibition on IICR by external Ca 2ϩ , whereas monotonic inhibition was shown in the case of Leu 335 IP 3 R3. Moreover, the receptors showed differences in Ca 2ϩ -induced conformational change when assayed with intrinsic fluorescence and CD. The results obtained from near-UV CD spectroscopy indicate that the Pro 335 LBD retains a more ordered tertiary structure than that of Leu 335 LBD upon Ca 2ϩ binding, although the two LBD proteins do not show a significant difference in secondary structure as assayed with far-UV CD. Therefore, it seems plausible that the functional differences between the receptors are mainly attributable to their structural differences caused by single amino acid substitution in the LBD region that is exposed to cytosol. Based on these results, we speculate that the Pro 335 3 Leu polymorphism of IP 3 R3 causes a change in receptor functions, with distinct cellular Ca 2ϩ -signaling properties. However, implica- It has been shown that IP 3 R reconstituted into proteoliposomes displays similar functional properties with the native receptors including IP 3 -binding affinity, Ca 2ϩ selectivity, and cytosolic Ca 2ϩ dependence of channel activity (33)(34)(35). Therefore, our results from the reconstitution study may indicate actual differences between the two receptors. Recently, the same receptor, Leu 335 IP 3 R3, from the B6 mouse was cloned and characterized (36), in which the K d value of the purified receptor for IP 3 (340 nM) was ϳ2-fold higher than that of the receptor reconstituted into the proteoliposome. In this study, Leu 335 IP 3 R3 reconstituted into the proteoliposome showed positive cooperativity for IP 3 binding. However, Iwai et al. (36) indicated that IP 3 binds to the receptor with negative cooperativity based on the model of four IP 3 -binding sites. The reason for this contradiction is not clear, although we cannot rule out the involvement of various accessory factors. Here we conducted an IP 3 -binding assay with purified IP 3 R3 reconstituted into proteoliposomes, excluding the possibility of accessory factors. Iwai et al. (36) used membrane fractions obtained from Sf9 cells expressing mouse IP 3 R3, which may contain additional factors influencing receptor activity.
The K d values of LBDs for ligand binding were reported for the two mutant forms of mouse IP 3 R3 (10, 37). The IP 3 -binding affinity of GST-fused LBD of rat IP 3 R3 (K d ϭ 66.0 Ϯ 13.2 nM) is close to that of Pro 335 LBD obtained here (95.4 Ϯ 2.9 nM). In the case of human IP 3 R3 using the N-terminal 750 residues of LBD, the K d for IP 3 was 151 nM (37). Differences in the IP 3binding affinities of IP 3 R3 are in the order of rat LBD Ͼ mouse Pro 335 LBD Ͼ human LBD Ͼ mouse Leu 335 LBD. It is interesting that the 335th residue of rat, BC mouse, human, and B6 mouse are proline, proline, methionine, and leucine, respectively. Although we cannot rule out the possibility that the difference may lie in other variable residues among species or in different experimental conditions, the association of the 335th residues with IP 3 -binding affinities together with our results suggest that the residue is one of the major determinants for the ligand-binding affinity among species.
The putative calcium sensor regions of IP 3 R isoforms were localized in the regulatory domain (38). According to the results, Ca 2ϩ induced a conformational change of recombinant protein made of that region, and calcium sensitivity was interchanged by swapping the sensor regions of two isoforms (38). Also, mutations within the sensor region (D2100E and E2100Q) shifted the bell-shaped Ca 2ϩ dependence curve into the high Ca 2ϩ range, indicating a reduction of Ca 2ϩ -binding affinities (39). However, in our results, IP 3 binding to LBD was regulated by conformational changes in the presence of free Ca 2ϩ . In addition, the Pro 335 3 Leu polymorphism changed the Ca 2ϩ effect on IICR. Taken together, the results suggest that LBD of IP 3 R3 plays an important role in Ca 2ϩ sensing. The Ca 2ϩ sensor region of IP 3 R3 in the regulatory domain (Met 1835 -Arg 2199 ) is associated with Ca 2ϩ with a K Ca of 100 nM (38). Therefore, the LBD of IP 3 R3 may function as a low affinity Ca 2ϩ sensor as opposed to the regulatory domain.
Based on the crystal structure of the LBD core (amino acids 224 -436) of IP 3 R1, the 335th residue is localized in the undefined loop between the Ϫ6 and Ϫ7 strands (26). Although the loop structure is not precisely determined, the polymorphic residue is unlikely to affect directly the IP 3 coordinating region and Ca 2ϩ -binding site because of the distant configuration in the receptor. The crystallographic structure provides only the LBD core without an N-terminal inhibitory domain (residues 1-225). Many studies have suggested that N-terminal inhibitory domain plays an important role in the functional regulation of LBD by a possible domain-domain interaction (26,40,41). Therefore, we cannot exclude the possibility that the polymorphic residue (335th) affects the structure and function of LBD in association with the N-terminal inhibitory domain.