Modulation of Osteopontin Post-translational State by 1,25-(OH)2-Vitamin D3

In osteoblastic ROS 17/2.8 cells, 1,25-(OH)2-vitamin D3 stimulates transcription of the extracellular matrix phosphoprotein osteopontin (OPN). We now show post-translational regulation of OPN production by 1,25-(OH)2D3. Prior to transcriptional up-regulation of OPN, 1,25-(OH)2D3 induces a shift in OPN isoelectric point (pI) from 4.6 to 5.1. Loading equal amounts of OPN recovered from ROS 17/2.8 cells exposed to 1,25-(OH)2D3 or carrier for 3 h reveals that the pI shift represents reduced phosphorylation. Trypsin cleavage patterns of OPN produced after 1,25-(OH)2D3treatment indicates phosphorylation changes in the resulting peptides. Using structural analogs to 1,25-(OH)2D3, we found that analog AT (25-(OH)-16-ene-23-yne-D3), which triggers Ca2+ influx but does not bind to the vitamin D receptor, mimicked the OPN pI shift, whereas analog BT (1,25-(OH)2-22-ene-24-cyclopropyl-D3), which binds to the vitamin D receptor without triggering Ca2+influx, did not. Likewise, inclusion of the Ca2+ channel blocker nifedipine blocks the charge conversion of OPN. Isolation of OPN from rat femurs and tibiae demonstrates the existence of two OPN charge forms in vivo. We conclude that 1,25-(OH)2D3 regulates OPN not only at the transcriptional level, but also modulates OPN phosphorylation state. The latter involves a short term (<3 h) treatment and is associated with membrane-initiated Ca2+ influx.

Osteopontin (OPN) 1 is a non-collagenous, glycosylated phosphoprotein originally found in bone matrix (1,2), but now known to be expressed in many tissues including kidney, hypertrophic chondrocytes, placenta, T-lymphocytes, macrophages, secretory epithelia and ganglia of the inner ear, and smooth muscle of the vascular system (3,4). OPN is also found in biological fluids such as milk (5), urine (6), and plasma (7), and it displays elevated expression in many transformed cells (8). OPN is highly acidic with approximately 25% of the amino acid composition aspartate and glutamate and significant numbers of phosphoserine and phosphothreonine (1,2). The amino acid sequence contains a conserved Gly-Arg-Gly-Asp-Ser (GRGDS) sequence (2,3). As a result, OPN binds effectively to the ␣ V ␤ 3 integrin (9), as well as to the ␣ V ␤ 5 and ␣ V ␤ 1 integrins (10). OPN also contains a thrombin cleavage site located near the middle of the molecule. Both peptides produced by thrombin cleavage are capable of supporting integrin-mediated cell attachment (11). Potential roles for OPN function through the ␣ V ␤ 3 integrin were proposed (12). OPN can promote attachment of various cell types and can initiate signal transduction through integrin-associated kinases. Other proposed roles for OPN include chemotaxis (13,14), inhibition of nitric oxide synthase expression (15), activation of pp60 c-src (16), hydroxyapatite binding (17), and Ca 2ϩ binding (18).
Hormone regulation plays an important role in OPN production. Our laboratory has examined the effect of the seco-steroid hormone 1,25-dihydroxyvitamin D 3 (1,25-(OH) 2 D 3 ) on OPN expression and secretion. After uptake of 1,25-(OH) 2 D 3 into a cell, the hormone binds to the vitamin D receptor (VDR), translocates to the nucleus, dimerizes preferentially with the retinoid X receptor, then binds to the vitamin D response element (VDRE) located in the promoter region of 1,25-(OH) 2 D 3 -responsive genes (19). In the case of OPN, this results in increased transcription as seen by higher OPN mRNA steady state levels at 24 -48 h (20) and, eventually, higher secreted protein levels (21). The OPN gene in rat is regulated by the additive action of two VDREs (3), which respond to both 1,25-(OH) 2 D 3 and to bioactive analogs of 1,25-(OH) 2 D 3 that bind to nuclear receptors (20).
1,25-(OH) 2 D 3 regulation also occurs through rapid plasma membrane-initiated responses, which have been well studied in osteoblasts. We previously reported the 1,25-(OH) 2 D 3 regulation of L-type voltage-sensitive calcium channels (VSCC) in ROS 17/2.8 osteosarcoma cells (22,23). Using patch-clamp techniques, nanomolar concentrations of 1,25-(OH) 2 D 3 were shown to increase open times of VSCCs and shift the threshold of activation toward the resting potential of the plasma membrane (22). Ca 2ϩ influx assays show this increase in VSCC open time leads to elevated Ca 2ϩ influx into the cell (23). 1,25-(OH) 2 D 3 activates other osteoblast signaling pathways that are independent of transcription such as a rapid increase in phospholipase C activity (24), activation of protein kinase C (25), and regulation of whole cell chloride currents (26).
In this publication, we report the effect of 1,25-(OH) 2 D 3 on osteoblast-like ROS 17/2.8 cells at a time period between the classic genomic response and the rapid membrane-associated responses. Examination of OPN secreted by ROS 17/2.8 cells after a 3-h exposure to 1,25-(OH) 2 D 3 reveals the production of a form of OPN with a higher isoelectric point (pI) than the original. Structural analogs of 1,25-(OH) 2 D 3 and Ca 2ϩ channel blockers were used to determine if a nuclear receptor-mediated response or a membrane-initiated response was responsible for the OPN pI shift. Our results provide evidence for short term regulation by 1,25-(OH) 2 D 3 of post-translational modification of OPN, a phenomenon that may modify its functional properties in bone matrix.

EXPERIMENTAL PROCEDURES
Cell Culture-ROS 17/2.8 cells were grown in Dulbecco's modified Eagle's medium/F12 medium (Atlanta Biologicals, Norcross, GA) containing 10% fetal bovine serum (Life Technologies, Inc.). ROS 17/2.8 osteoblast-like osteosarcoma cells were chosen because of their responsiveness to 1,25-(OH) 2 D 3 and because of their ability to synthesize and secrete high levels of OPN mRNA and protein. Cells were seeded at low density (30,000 cells/ml) and allowed to grow until they reached 60 -70% confluence. For protein isolation, cells were transferred to serumfree medium overnight, then treated with 2.5 nM 1,25-(OH) 2  OPN Purification from ROS 17/2.8 Cell Medium by Barium Citrate Precipitation-Medium was collected from ROS 17/2.8 cells that had been exposed to 1,25-(OH) 2 D 3 with or without Ca 2ϩ channel blockers, analog AT, analog BT, or carrier (ethanol) alone. OPN was isolated using the barium citrate procedure (5). Briefly, 3.8% sodium citrate and 15% BaCl 2 were added to the medium at 1/10 of the total volume. After shaking for 10 min at 4°C, the mixture was centrifuged for 10 min at 3800 rpm. The supernatant was discarded, and the pellet was then washed for 10 min with 15% BaCl 2 at 4°C. After centrifugation for 10 min at 3800 rpm, the supernatant was again discarded, and the pellet was washed with H 2 O for 10 min at 4°C. After another round of centrifugation, the supernatant was removed, and OPN was eluted from the pellet by dissolution in 0.2 M sodium citrate, pH 6.8. Excess salts were removed with a desalting column (Pierce). Further purification was accomplished by passing the sample through an affinity column (described below).
Immunoaffinity Chromatography-IgG was immobilized with cyanogen bromide-activated Sepharose 4B (Sigma) and coupled as described (27). Briefly, 1 g of cyanogen bromide-activated Sepharose 4B with 5 mg of ␣-rat OPN IgG (purified using ImmunoPure (A/G) IgG purification kit as per manufacturer's instructions (Pierce)) in 0.1 M Bicine-HCl, pH 8.5, 0.5 M NaCl was agitated continuously for 2 h at room temperature (coupling buffer). Remaining coupling sites were blocked by incubation in 1 M Tris-HCl, pH 8.0, with constant agitation for 2 h at room temperature. The gel was then washed extensively with coupling buffer and then with 0.1 M citric acid to remove nonspecifically bound protein.
Medium from ROS 17/2.8 cells was dialyzed through three buffer changes in either 10 mM Na 3 PO 4 , pH 7.5, or 0.25 mM ammonium bicarbonate (if performed following barium citrate precipitation). Samples were passed over the immunoaffinity column and OPN was eluted from the column with washes of Tween 20, PBS, 0.5 M NaCl, 0.025% SDS in PBS, and PBS.
OPN Purification from Rat Long Bone-Purification of OPN from rat long bone was performed as described (1). Briefly, extraction of proteins from rat femurs and tibiae was accomplished by demineralization using 4 M guanidinium HCl in 50 mM Tris-HCl, pH 7.2, for 24 h at 4°C followed by decalcification in 4 M guanidinium HCl, 0.5 M EDTA in 50 mM Tris-HCl, pH 7.2, for 48 h at 4°C. All extractions contained the following protease and alkaline phosphatase inhibitors: 10 mM EDTA, 0.1 M 6-aminohexanoic acid, 5 mM benzamidine hydrochloride, 1 mM sodium iodoacetate, 10 mM phenylmethylsulfonyl fluoride, 5 mg/liter pepstatin, 1 mg/liter soybean trypsin inhibitor, and 5 mM levamisole. The extracts were pooled, centrifuged at 30,000 ϫ g for 15 min, and concentrated by ultrafiltration.
Concentrated samples were subjected to gel filtration on a 2.5 ϫ 30-cm Sephacryl S-200 column and eluted with 4 M guanidinium HCl in 50 mM Tris-HCl, pH 7.2. The high molecular weight fraction was concentrated to one-sixth volume, passed over a 2 ϫ 30-cm column of Bio-Gel P30, and eluted with 6 M urea in 50 mM Tris-HCl, pH 7.2. The elution was then run on a 2.5 ϫ 20-cm DEAE-Sephacel column. The column was eluted with a linear gradient formed from 750 ml each of 6 M urea in 50 mM Tris-HCl, pH 7.2, and 6 M urea in 50 mM Tris-HCl, pH 7.2 containing 0.7 M NaCl. The fractions eluting at 0.28 M NaCl (designated peak D4a) and 0.36 M NaCl (designated peak D4b) were rechromatographed on the 2.5 ϫ 20-cm DEAE-Sephacel column, but with a shallow NaCl gradient (0.14 -0.4 M).
Precipitation of Total Protein from ROS 17/2.8 Cells-Extraction buffer (1% Chaps, 4 M guanidinium HCl, PIC I and II (28), and 20 mM Tris-HCl, pH 7.0) was added to the plates for 10 min. The extraction buffer fraction was added to the medium fraction. Proteins were precipitated using 10% trichloroacetic acid, 3% phosphotungstic acid for 2 h at 4°C. The precipitate was centrifuged, and the pellets were solubilized in 2% cholate and sonicated. The soluble fraction was collected and dialyzed against 10 mM Tris, 0.1% cholate.
Trypsin Cleavage of OPN-OPN isolated from ROS 17/2.8 cells exposed to 1,25-(OH) 2 D 3 or carrier (ethanol) for times indicated in figures was lyophilized. The lyophilized protein was resuspended at 37°C for 3 h in 2.5 mg/ml trypsin in 0.125 M Tris-HCl, pH 6.8. After incubation, the peptide fragments were lyophilized, placed in SDS sample buffer, and separated by SDS-PAGE as described below.
SDS-Polyacrylamide Gel Electrophoresis-SDS-PAGE was performed using a modification of procedures previously described (29). Acrylamide and bisacrylamide concentrations were altered for a 12.5% resolving gel. Gels were stained with 20 mg of Stains All (Bio-Rad) in 10 ml of formamide, 50 ml of isopropyl alcohol, 1.0 ml of 3 M Tris-HCl, pH 8.8, 139 ml of H 2 O overnight in a dark room.
Isoelectric gel electrophoresis was performed using IsoGel agarose IEF plates (FMC Bioproducts, Rockland, ME). Proteins were lyophilized and resuspended in deionized H 2 O. Proteins were focused on pH range 3-10 gels following the manufacturer's instructions. Gels were then dried and exposed to autoradiographic film or transferred to nitrocellulose using a press blot procedure as described by the manufacturer.
Immunoblot Analysis-SDS-polyacrylamide gels were transferred to nitrocellulose at 15 V in transfer buffer (0.025 M Tris, 0.2 M glycine, 20% methanol) overnight. Blots were dried and then blocked in PBS containing 3% BSA and 0.15% Tween 20 at room temperature for 2 h. Blots were then incubated in PBS containing 3% BSA and 0.15% Tween 20 with a 1:10,000 dilution of primary antibody (goat ␣-rat OPN) at room temperature for 1 h. Blots were then washed five times for 10 min each in PBS containing 1.5% BSA and 0.075% Tween 20. Next, blots were incubated in PBS containing 3% BSA and 0.15% Tween 20 with a 1:100,000 dilution of peroxidase-conjugated donkey ␣-goat IgG secondary antibody (Jackson Immunoresearch, West Grove, PA) or a 1:50,000 dilution of alkaline phosphatase-conjugated swine ␣-goat IgG secondary antibody (Boehringer Mannheim). Blots were again washed 5 times for 10 min each in PBS containing 1.5% BSA and 0.075% Tween 20. For alkaline phosphatase detection, blots were incubated in carbonate buffer with detection substrate (0.35 mM nitro blue tetrazolium, 0.35 mM 5-bromo-4-choloro-3-indolyl phosphate, 0.1 M NaH 2 CO 3 , 1.0 mM MgCl 2 , pH 9.8) until appearance of coloration. Peroxidase was detected using a chemiluminescent procedure. Blots were incubated in 50% Luminol/enhancer solution, 50% stable peroxide solution (Pierce) for 5 min and exposed to autoradiographic film for 1 min.

RESULTS
Effect of 1,25-(OH) 2 D 3 on OPN pI-We examined the effect of 1,25-(OH) 2 D 3 on OPN production during the first 3 h of treatment, prior to nuclear receptor-mediated transcriptional upregulation. Proteins from the cell fraction were isolated as described under "Experimental Procedures" and visualized by two-dimensional gel electrophoresis. OPN from ROS 17/2.8 cells treated with vehicle was found to focus in two discrete spots at pI 4.6 and 5.1 (Fig. 1A). The identity of the spots focusing at these locations as OPN was confirmed by Western blotting (data not shown) using goat ␣-rat OPN. The ratio of protein found at pI 5.1 compared with pI 4.6 was 0.26:1 assessed by densitometry. In Fig. 1B, OPN from ROS 17/2.8 cells exposed to 2.5 nM 1,25-(OH) 2 D 3 for 3 h focuses almost completely at pI 5.1, with only a small fraction (12:1 ratio) at pI 4.6. Western blotting confirmed the focused spots as OPN (data not shown). These experiments were repeated with cells labeled with [ 35 S]methionine to visualize total protein from the cell fraction. Most spots were unchanged in location, but the spot corresponding to OPN shifted to a more basic pI after 3 h 1,25-(OH) 2 D 3 treatment (data not shown). The identity of the spot which disappears altogether (small arrow in Fig. 1, A and  B) is presently unknown.
For further verification of the pI shift, OPN was isolated from ROS 17/2.8 cell medium by immunoprecipitation. ROS 17/2.8 cells were cultured as described above, with the exception that the time of exposure to carrier or 1,25-(OH) 2 D 3 was increased to 18 h. This was done to permit the secretion of measurable levels of OPN protein prior to immunoprecipitation. Visualization of equal counts/min of immunoprecipitated OPN protein was accomplished by two-dimensional gel electrophoresis and autoradiography. Immunoprecipitated OPN from ROS 17/2.8 cells exposed to vehicle alone focused at pI 4.6 ( Fig.  1C). OPN immunoprecipitated from ROS 17/2.8 cells treated with 2.5 nM 1,25-(OH) 2 D 3 focused at pI 5.1 and, to a much lesser extent, pI 4.6 (Fig. 1D). Medium from ROS 17/2.8 cells treated with vehicle or 1,25-(OH) 2 D 3 and passaged over a preimmune IgG column produced no spots when analyzed by twodimensional gel electrophoresis (data not shown). These results show that OPN secreted into the medium undergoes a charge shift similar to that of OPN in the cell fraction (Fig. 1, A and B).
Effect of 1,25-(OH) 2 D 3 on OPN Phosphorylation Levels-Because the 1,25-(OH) 2 D 3 -induced pI change was over a short range (0.5 units), we hypothesized that this shift was the result of a reduced phosphorylation of OPN. An altered level of phosphorylation was a likely possibility because of the number of potential OPN phosphorylation sites. Radiolabeled medium was collected and proteins were precipitated using sodium citrate and BaCl 2 as described under "Experimental Procedures." OPN was isolated from the group of precipitated proteins by immunoaffinity chromatography as described above. (OH) 2 D 3 -treated cells were visualized after separation by 12.5% SDS-PAGE ( Fig. 2A). OPN from cells treated with 1,25-(OH) 2 D 3 (lane 2) contained lower levels of radioactive phosphate incorporation compared with OPN from control cells (lane 1). Measurement by densitometry and scintillation counting showed a 70% decrease in phosphorylation of OPN from 1,25-(OH) 2 D 3 -treated cells compared with that from control cells. Fig. 2B shows Western blot analysis of OPN (50 g) from vehicle-treated and 1,25-(OH) 2 D 3 -treated cells. We consistently observed increased immunostaining of the antibody to OPN from 1,25-(OH) 2 D 3 -treated cells, which may reflect increased accessibility to antigenic sites on the core protein.
To determine whether 1,25-(OH) 2 D 3 treatment, and consequent reduced phosphorylation, could alter sensitivity to proteolytic cleavage, OPN from vehicle-treated and 1,25-(OH) 2 D 3treated cultures was subjected to trypsin digestion (Fig. 3). Effect of Structural Analogs to 1,25-(OH) 2 D 3 on OPN pI-To assess the relative involvement of the VDR and Ca 2ϩ influx in these observed changes, we first utilized two structural analogs to 1,25-(OH) 2 D 3 that vary in their ability to stimulate pathways of 1,25-(OH) 2 D 3 action on ROS 17/2.8 cells (Fig. 4A). ROS 17/2.8 cells were labeled with [ 32 P]orthophosphate and treated with 25 nM analog BT or 2.5 nM analog AT in serum-free medium without Na 3 PO 4 . OPN was isolated using the barium citrate precipitation followed by immunoprecipitation as explained under "Experimental Procedures." Equal counts/min of OPN were visualized using isoelectric gel electrophoresis (Fig.  4B). In lane 1, OPN from cells exposed to analog BT focused at pI 4.6. However, OPN from cells exposed to analog AT duplicated the pI shift that occurred with the addition of 1,25-(OH) 2 D 3 (lane 2). Western blotting confirmed that the focused spots were OPN (data not shown).

Nifedipine Inhibits the 1,25-(OH) 2 D 3 -induced Phosphorylation Change in OPN-
Since analog AT was found to induce the OPN charge shift, we hypothesized that Ca 2ϩ influx through VSCCs was involved in the pathway leading to a shift in the OPN charge state. To examine this possibility, we inhibited the VSCCs with the L-type Ca 2ϩ channel blocker nifedipine. ROS 17/2.8 cells were labeled with [ 32 P]orthophosphate and treated with 2.5 nM 1,25-(OH) 2 D 3 or carrier with or without 50 nM nifedipine in serum-free medium without Na 3 PO 4 . OPN was isolated by barium citrate precipitation followed by immunoprecipitation as explained under "Experimental Procedures." 20 g per lane of OPN were visualized by SDS-PAGE (Fig. 5). Consistent with observations previously seen in Fig. 2, OPN from cells treated with 1,25-(OH) 2 D 3 contained lower levels of incorporated radioactive phosphate compared with OPN from control cells (compare lanes 1 and 2). In lane 3, cells treated with 1,25-(OH) 2 D 3 and nifedipine showed levels of radioactive phosphate incorporation comparable to control cells.
pI of OPN Isolated from Rat Long Bone-To determine the existence of similar multiple charge forms in vivo, OPN was isolated from rat femurs and tibiae. Analysis of OPN from peaks D4a and D4b of the DEAE-Sephacel column was accomplished by SDS-PAGE and by isoelectric gel electrophoresis (Fig. 6). OPN isolated from peak D4a migrated at 73.4 kDa (Fig. 6A, lane D4a; Fig. 6B, lane D4a). The streaking found in Fig. 6B (lane D4a) is likely the result of microheterogeneity associated with the aggregation of OPN or partial degradation during the bone extraction procedure. A minor band of OPN was found at 28.8 kDa, possibly the product of thrombin cleavage. OPN from peak D4b migrated at 91.3 kDa (Fig. 6A, lane D4b; Fig. 6B, lane D4b). After isoelectric focusing, the gel was transferred to nitrocellulose via a press blot procedure. Western analysis and detection by alkaline phosphatase (Fig. 6C) was performed as described under "Experimental Procedures." OPN1 from peak D4b (lane OPN1) was found to focus at pI 4.6,

. Effect of structural analogs of 1,25-(OH) 2 D 3 on OPN pI shift.
A, structure of 1,25-(OH) 2 D 3 , analog AT, and analog BT. B, OPN was analyzed on 3-10 isoelectric focusing gels as described under "Experimental Procedures." Lane 1, OPN precipitated by the barium citrate procedure from ROS 17/2.8 cells exposed to 25 nM analog BT for 3 h. pI was determined as in Fig. 1A. Arrow indicates isoelectric focusing at pI 4.6. Lane 2, OPN precipitated by the barium citrate procedure from ROS 17/2.8 cells exposed to 2.5 nM analog AT for 3 h. Arrow indicates isoelectric focusing at pI 5.1. are also able to stimulate separate pathways, with structural confirmation and side chain additions or subtractions affecting the ability of the analog to bind to VDR or stimulate Ca 2ϩ influx (33)(34)(35)(36).
To the growing list of non-genomic or membrane-initiated responses to 1,25-(OH) 2 D 3 (37)(38)(39)(40)(41)(42)(43)(44)(45), we now report that addition of 1,25-(OH) 2 D 3 to ROS 17/2.8 osteoblast-like cells results in the production of a less phosphorylated form of the secreted extracellular matrix protein OPN. Ca 2ϩ influx through VSCCs is necessary for production of the less phosphorylated form of OPN as inhibition of VSCCs by the Ca 2ϩ channel blocker nifedipine prevents its production. Although changes in phosphorylation state stimulated by 1,25-(OH) 2 D 3 have been found in cytosolic and plasma membrane fractions (46,47), this is the first report of a change in post-translational processing initiated by 1,25-(OH) 2 D 3 .
OPN has been found in multiple charge forms, previously. Nemir et al. (48) reported the existence of a phosphorylated form and a non-phosphorylated form of OPN produced by normal rat kidney cells. Chang and Prince (49) showed that 1,25-(OH) 2 D 3 could stimulate the synthesis of a non-phosphorylated form in mouse JB6 epidermal cells, which normally produce low amounts of a phosphorylated form. The stimulation, however, appears to be genomic, as protein levels increase after a 24-h exposure to 1,25-(OH) 2 D 3 .
Data from the in vitro experiments indicate the existence of multiple charge forms of OPN secreted by osteoblast-like cells, but does not answer the question of the existence of these charge forms in bone extracellular matrix. Looking for equivalent charge forms in vivo, we examined OPN purified from rat tibiae and femurs employing a well-developed isolation procedure (1). On the final purification column (DEAE-Sephacel), OPN can be isolated from two peaks (designated D4a and D4b). Since OPN from these peaks elutes at two different concentrations of NaCl, we hypothesized that these peaks might correspond to OPN with two distinct isoelectric points. The functional significance of these charge forms must be determined.
With the addition of this new phosphorylation shift in OPN from ROS 17/2.8 cells stimulated by 1,25-(OH) 2 D 3 and in OPN from rat long bone, we have the following model representing 1,25-(OH) 2 D 3 action on osteoblasts (Fig. 7). Osteoblasts, when not exposed to 1,25-(OH) 2 D 3 , produce a higher phosphorylated OPN, designated OPN1. This form is found in rat long bone from peak D4b. 1,25-(OH) 2 D 3 produces, after 3 h, a shift to a lower phosphorylation form of OPN, designated OPN2. This form is found in rat long bone from peak D4a. Since the shift in OPN phosphorylation state can be mimicked by analog AT or inhibited by Ca 2ϩ channel blockers in ROS 17/2.8 cells, this pathway apparently does not involve the VDR, and appears to be a result of a Ca 2ϩ -influx-dependent signaling response.
Previous work in ROS 17/2.8 cells has provided evidence for a 44-kDa form and a 55-kDa form of OPN (50). The 44-kDa form was found to be highly phosphorylated, while the 55-kDa form was found to be less phosphorylated. In ROS 17/2.8 cells, our data indicate the existence of two different charge states. The 4.6 pI form apparently is equivalent to this 44-kDa form, and the 5.1 pI form of OPN is equivalent to the 55-kDa form. Contrary to their conclusions, however, is our finding that both forms can be isolated and do exist in rat long bone. OPN from rat long bone migrates differently than these 44-and 55-kDa forms. We find that the more acidic form, the more phosphorylated form, migrates slower on SDS-PAGE (Fig. 6B), compared with faster migration of the highly phosphorylated, more acidic 44-kDa OPN. These results are similar to the findings of Nemir et al. (48), where the phosphorylated OPN migrated slower than the non-phosphorylated form. In fact, the discrepancy in migration on SDS-PAGE between OPN from peaks D4a and D4b might be attributed to decreased binding of SDS to OPN from peak D4b. This decreased binding to SDS could result in slower migration on SDS-PAGE.
Currently, our laboratory is investigating the phosphorylation sites found on the two OPN isoforms isolated from rat long bone (D4a and D4b). Analysis of phosphorylation from bovine OPN revealed the existence of 28 sites of phosphorylation (51). Phosphorylation motifs matched the recognition sequence utilized by CKI and CKII coinciding with the evidence that a CKII-like activity phosphorylates OPN in the Golgi (52). In rat long bone, OPN from peak D4a has been analyzed for posttranslational modifications (53). Phosphorylations were found on equivalent sites as those in bovine OPN, but some of the phosphorylations were partial (not found in every analysis). Partial phosphorylation of residues might play a role in the unusual streaking found when OPN is detected by Western blot procedures (Fig. 6). Although staining reveals the presence of distinct bands (Fig. 6A), Western blot analysis of OPN from rat long bone reveals not only the distinct bands, but also a broad range of forms (Fig. 6B). This is especially evident in the isoelectric gel (Fig. 6C), where OPN from D4a and D4b focused at major bands of pI 5.1 and 4.6, respectively, but was also detected in a broad band at higher isoelectric points. Notice that the D4a fraction does not appear to contain any of the lower pI fraction of OPN. Another interesting result with antibody detection is that the antibody to OPN utilized in this study apparently binds to the lower phosphorylated form of OPN more efficiently than to the higher phosphorylated form. This can be seen in OPN isolated from ROS 17/2.8 cells (Fig. 2) and in OPN from rat long bone (Fig. 6). A possible explanation is that this antibody was developed using OPN purified from peak D4a, and, therefore, might be more specific for the lower phosphorylated form.
OPN contains a GRGDS sequence that allows the protein bind to cells containing the ␣ V ␤ 3 integrin. As previously mentioned, OPN was found in both phosphorylated and non-phosphorylated forms in normal rat kidney cells (48). Phosphorylated OPN showed cell surface association, apparently GRGDSdependent, while the non-phosphorylated form was not found on the cell surface. Instead, the non-phosphorylated OPN was associated with fibronectin (via co-precipitation). Another study of OPN phosphorylation demonstrated that partially dephosphorylating OPN with tartrate-resistant acid phosphatase created a form that could no longer bind to osteoclasts (54). It may be that the OPN phosphorylation loss after 1,25-(OH) 2 D 3 treatment results in decreased binding to cells containing the ␣ V ␤ 3 integrin. Osteoblasts (ob) secrete OPN1 into newly synthesized osteoid unless signaled by 1,25-(OH) 2 D 3 to deposit the lower phosphorylated form, OPN2. OPN1 is designated as the 4.6 pI charge form isolated from ROS 17/2.8 cells or the product of peak D4b during purification of OPN from rat long bones. OPN2 is designated as the 5.1 pI charge form or the product of peak D4a during purification of OPN from rat long bones.
In conclusion, we have isolated two charge forms of the phosphoprotein OPN from ROS 17/2.8 cells exposed to vehicle or 1,25-(OH) 2 D 3 that have uniquely different isoelectric points. This pI shift appears to be the result of reduced phosphorylation in cells treated with 1,25-(OH) 2 D 3 . The loss in phosphorylation does not occur through a VDR-mediated genomic event, but is associated with Ca 2ϩ influx. Finally, we report that the two charge isoforms of differing pI can be isolated from rat long bone.