Expression and ultrastructural immunolocalization of a major 66 kDa phosphoprotein synthesized by chicken osteoblasts during mineralization in vitro.код для вставкиСкачать
THE ANATOMICAL RECORD 228:93-103 (1990) Expression and Ultrastructurat Immunolocalization of a Major 66 kDa Phosphoprotein Synthesized by Chicken Osteoblasts During Mineralization In Vitro L.C. GERSTENFELD, Y. GOTOH, M.D. McKEE, A. NANCI, W.J. LANDIS, AND M.J. GLIMCHER Department of Orthopedic Surgery, Laboratory for the Study of Skeletal Disorders and Rehabilitation, Harvard Medical School, The Children’s Hospital, Boston, Massachusetts (L.C.G., Y.G., M.D.M., W.J.L., M.J.G.); Departments of Anatomy and Stomatology, University of Montreal, Montreal, QC Canada (A.N.) ABSTRACT Embryonic chicken osteoblasts cultured over a 30 day period were used as a model system for studying the expression of bone phosphoproteins during cellular differentiation and the possible role of these proteins in extracellular matrix mineralization. Accumulation of total phosphoprotein in the cultures, as determined by 0-phosphoserine (Ser-P) and 0-phosphothreonine (Thr-P) amino acid analysis, revealed a > 10-fold increase over the 30 day period. Total phosphoprotein synthesis, a s assessed by (32P)-,(3H)-Ser-P, and (I4C)-Thr-P protein labeling, showed the highest levels concurrent with initial mineral deposition within the matrix. The major phosphoprotein present in chicken bones and synthesized by the cultured osteoblasts had a molecular weight of -66 kDa. This 66 kDa bone phosphoprotein (66 kDa BPP) was purified to homogeneity and was used for antibody production. Application of this antibody in Western blot analysis revealed that 66 kDa BPP was present only in protein extracts of mineralizing cultured osteoblasts and was absent in cultures of non-mineralizing chondrocytes, myoblasts, and tendon fibroblasts. The 66 kDa BPP in vitro accumulated continuously in the extracellular matrix in a manner that paralleled both phosphoprotein synthesis and total phospho-amino acid production. A comparison of the results obtained in vitro to those from developing embryonic tibiae in vivo demonstrated a similar qualitative and temporal expression of phosphoprotein and a continual accumulation of 66 kDa BPP in the matrix with advancing mineralization and developmental age. Ultrastructural immunocytochemistry using the 66 kDa BPP antibody and the protein A-gold technique revealed specific immunolabeling over electron-dense regions of mineralization in the cultures that appeared identical to the distribution of labeling observed in vivo (McKee et al.: Connect. Tissue Res., 21:21-29, 1989; Anat. Rec., 228:77-92, 1990). These results demonstrate that this major 66 kDa BPP was expressed concurrently with other differentiated osteoblast functions and suggests that it may play a role in the initiation or regulation of mineralization. Central to the processes of bone mineralization is the prerequisite synthesis and assembly of a n extracellular matrix which can facilitate mineral nucleation and also serve as the superstructure into which mineral can be deposited and accumulated (Glimcher, 1976). Extracellular matrix synthesis, assembly, and mineralization, which largely determine the unique structure and physiological role of bone in skeletogenesis and ion homeostasis, are ultimately regulated temporally and spatially by the osteoblasts of this tissue through cellmatrix interactions and secretion and interplay of noncollagenous matrix components with collagen. One class of non-collagenous extracellular matrix proteins found in all vertebrate mineralized tissues are 0 1990 WILEY-LISS, INC the phosphoproteins. Their ubiquitous nature and other considerations have led to the hypothesis that this class of protein may play a crucial role in initiating or controlling the spatial distribution of mineral in the Received October 3, 1989; accepted December 19, 1989. Address reprint requests to Dr. Louis C. Gerstenfeld, Department of Orthopedic Research G11, Children’s Hospital, 300 Longwood Avenue, Boston, MA 02115. Dr. M.D. McKee’s present address is Department of Stomatology, Facuity of Dental Medicine, University of Montreal, Montreal, Quebec, H3C 357, Canada. 94 L.C. GERSTENFELD ET AL. extracellular matrix of dentin and bone (Veis, 1978; Veis and Sabsay, 1983; Glimcher, 1976, 1989). For chicken bone, numerous phosphorylated polypeptides, ranging in molecular weight from 6 kDa to 150 kDa and containing both 0-phosphoserine (Ser-P) and 0phosphothreonine (Thr-P) residues, have been extracted and purified to homogeneity (Spector and Glimcher, 1972; Cohen-Solal et al., 1978; Lee and Glimcher, 1981; Uchiyama et al., 1986). The chicken bone phosphoproteins have been shown to be synthesized and secreted by osteoblasts in bone tissue in vivo and by cultured chicken bone organ cultures and cultured osteoblasts in vitro (Glimcher et al., 1982; Gotoh et al., 1983, 1989; Landis et al., 1984; Gerstenfeld e t al., 1989). Recently, i t has been demonstrated that many of the phosphoprotein species previously isolated from chicken bone are the result of proteolysis of a major phosphoprotein species having a n apparent molecular weight of -66 kDa (66 kDa BPP; Gotoh et al., 1990). The DEAE sephacel elution characteristics, the presence of sialic acid, the amino acid composition, and the presence of a n Arg-Gly-Asp amino acid sequence within 66 kDa BPP (Gotoh et al., 1990) would suggest that this protein is similar to a recently identified class of phosphorylated sialoproteins found in bovine, rat, and human bone (Fisher et al., 1983, 1987; Franzen and Heinegard, 1985; Oldberg et al., 1986; Prince et al., 1987). It is likely that the mammalian phosphorylated sialoproteins and the phosphoproteins present in chicken bone represent a similar class of acidic glycosylated phosphoproteins and play comparable functional roles with regard to bone formation. To date, these proteins have been identified in the bones of 20 vertebrate species and all contain Ser-P, Thr-P, and sialic acid and have high Glu and Asp contents (Glimcher, 1984). This laboratory has used cultured chicken osteoblasts a s a model system to study cellular differentiation and the mechanism(s) of bone matrix assembly and mineralization. The approach, in using these cultures, has been to quantitate systematically the temporal expression of known matrix components secreted by osteoblasts and determine the extent of their accumulation within the extracellular matrix as i t relates to the ongoing extensive mineralization of these cultures in vitro. Previous studies using this system have demonstrated a n induction of several differentiated osteoblast functions (Gerstenfeld et al., 1987) that are identical to the induction seen for these functions in vivo (Yoon et al., 1988).Extensive ultrastructural examination of the cultures has also demonstrated that a n intrinsic property of the cultured osteoblasts is a n ability t o direct the assembly of their synthesized collagen into both individual fibrils and a suprafibrillar arrangement that is similar to that observed in vivo. Once preliminary matrix assembly has occurred, apatitic mineral is spatially disposed throughout the matrix in a manner resembling that found in bone (Gerstenfeld et al., 1988). These initial studies, which show numerous similarities between the processes of bone formation observed in vivo and in vitro, demonstrate the usefulness of such a culture system to aid in the understanding of bone development and mineralization. The results of the present study, using biochemi- cal and immunocytochemical techniques, extend these initial observations and relate the expression of a specific 66 kDa chicken bone phosphoprotein (BPP) to mineralization in vitro. MATERIALS AND METHODS Cell Culture Osteoblasts were isolated by sequential trypsinicollagenase treatment of 17-day embryonic chicken calvariae and cells capable of expressing a n osteoblastic phenotype were selected (Gerstenfeld et al., 1987). Cultures were initially grown for 3 wk in minimum medium and then subcultivated and grown for up to 36 days in BJG, Fitton-Jackson medium supplemented with 50 pg/ml ascorbate and 10 mM P-glycerophosphate. Protein Pulse Labeling Analysis All radiolabeled compounds were obtained from the Amersham Co., Inc. (Arlington Heights, IL). Cells were separately labeled, without serum, in a total of 5 ml per 100-mm dish with either 250 pCi of (3H)-leucine (120 mCi/mmol) for 24 h r or with 250 pCi of (32P)-orthophosphate (3,000 mciimmol) for 6 hr. In experiments where total phosphoprotein synthesis was to be monitored, cultures were double-labeled with 250 pCi (3H)serine (20-40 mCi/mmol) and 50 pCi (14C)-threonine (250 mCi/mmol) for 24 hr. Phosphoprotein Extraction and Purification From Tissue Metatarsals, tibiae, and femurs from 14-wk-old postnatal chickens were used for the extraction of bone phosphoproteins. The tissues were rapidly frozen in liquid N,immediately after dissection and powdered with a mortar and pestle. Ten grams of powdered tissue were suspended in 300 ml of deionized H,O containing 1mM p-hydroxymercurobenzoic acid, 1 mM phenylmethylsulfonyl fluoride, and 10 mM levamisole as protease and phosphatase inhibitors and homogenized at 0°C with a polytron apparatus (Tekman Co., Cincinnati, OH). The homogenate was brought to 0.3 N HC1 by addition of 1.0 N HC1 at 4°C. The tissue homogenate was then extracted in the HC1-acidified solution at 4°C for 48 h r while the pH was maintained a t -1.0. Insoluble material was removed by centrifugation a t 15,000 rpm in a n SS34 Sorvall rotor. The acid-extracted proteins were dialyzed against deionized H,O until a neutral pH was obtained, following which they were lyophilized. Preliminary chromatography on DEAE Sephacel and molecular sieving on Sephacryl S-300 (Schema 11, Uchiyama et al., 1986; Gotoh et al., 1990) were carried out for protein purification. Protein purity was determined by amino terminal sequence analysis and was carried out either directly on samples obtained from filtration through Sephacryl S-300 or after blotting onto PVDF membranes followed by the sequencing method a s described by Hunkapillar and Lujan (1986). Both analyses yielded identical amino terminal sequences of good quality, high repetitive yield, and good recovery (70-100 pM), and no underlying secondary sequences were detected. These analyses were carried out on a protein sequencer (Model 470A; Applied Biosystems, Inc., CA) equipped with a n on-line HPLC (Model 120A; Applied Biosystems, Inc.) a t the Mi- IMMUNOLOCALIZATION OF SYNTHESIZED PHOSPHOPROTEIN crochemistry Laboratory, Harvard University, Cambridge, MA. 95 point dried. Following drying, the specimens were coated by gold evaporation. Some cultures were viewed directly along their top surfaces while others were Phosphoprotein Analysis viewed from a n angle that revealed the full cross-secTotal amino acid analysis and Ser-P and Thr-P con- tional depth of the tissue. The latter cultures were pretents were determined from 6 N and 4 N HCl hydroly- pared by manually fracturing the flexible plastic cirsates, respectively, on a Model 121M automatic amino cular coverslips on which tissue had been grown and acid analyzer [Beckman Instruments, Inc., Palo Alto, then gold-coating the exposed surfaces of interest. CA (Cohen-Solal et al., 197811. Radiolabeled ( 14C)-Thr- Samples were observed with a JEOL JSM 840 having P and (,H)-Ser-P were quantitated by scintillation a n LaB, electron gun and operating a t 15 kV. counting after collecting the Ser-P and Thr-P peaks Immunocytochemistry was performed on 30-day from the amino acid analysis. Protein profiles were an- mineralizing osteoblast cultures which were fixed overalyzed by SDS-polyacrylamide gel electrophoresis on night a t 4°C with 1% glutaraldehyde in 0.1 M sodium 5-15% continuous linear gradient gels (Laemmli, cacodylate buffer, pH 7.3, followed by post-fixation 1970) followed by staining with rhodamine B, a stain with reduced osmium tetroxide at 4°C. After washing specific for phosphorylated bone proteins (Uchiyama et with 0.1 M sodium cacodylate buffer containing 4% sual., 1986), or with AgNO, by using a Silver Staining crose, pH 7.3, the cultures were dehydrated through a Kit (Bio-Rad Laboratories, Richmond, CA). Fluorogra- graded ethanol series and embedded in Epon 812 (E.F. phy of (,H)-labeled proteins was carried out by the Fullam, Inc., Latham, NY). The resin was polymerized methods of Bonner and Laskey (1974) and fluoro- for 2 days at 60°C. Thin sections (80 nm) were cut and graphs were prepared for quantitation by the method of immunocytochemical labeling, using a rabbit 66 kDa Laskey and Mills (1975). Scanning densitometry was BPP polyclonal antibody and the protein A-gold techperformed by using a n LKB Ultrascan laser densitom- nique, was carried out as detailed by McKee et al. eter (LKB, Broma, Sweden). (1990). The protein A-gold complex was as described by Bendayan (1984), using gold particles (14 nm diamAntibody Production and Characterization eter) made according to Frens (1973). The specificity of Purified 66 kDa BPP (1 mg/ml) was injected subcu- the immunocytochemical reactions was assessed by intaneously into New Zealand white rabbits by using a n cubating control sections with either pre-immune seinjection schedule and protocol previously published rum or antigen-adsorbed antibody prior to incubation for the generation of antibodies to osteocalcin (Gund- with the protein A-gold complex. Following the immuberg et al., 1987). The titer of the rabbit anti-serum nocytochemical procedures, tissue sections were conwas evaluated by serial dilution “dot blotting.” Con- ventionally stained with uranyl acetate for 5 min and trol, pre-immune serum prepared from the rabbit on lead citrate for 2 min and examined at 80 kV in a Philthe initial day of its innoculation with the 66 kDa BPP ips 410 transmission electron microscope. was simultaneously examined. Animals were initially RESULTS bled a t 0 wk, 4 wk, and 10 wk. After 10 wk, the titer Protein Purification and Antibody Characterization against BPP reached a maximal level with a detectable colorimetric response on a dot blot assay a t 50 pg. UsThe major 66 kDa BPP was extracted from 14-wk ing 11-wk rabbit anti-serum, immunoblotting was car- post-natal bone as described earlier. A representative ried out as described by Towbin and Gordon (1984). purification scheme by DEAE ion exchange chromatogProteins were electroblotted onto nitrocellulose by us- raphy and S-300 molecular sieving is shown in Figure ing a Trans Blot SD blotter (Bio-Rad Laboratories, 1A while the SDS polyacrylamide gel analysis of each Richmond, CA). Initial tests demonstrated that only a purification step is shown in Figure 1B. As Figure 1B short time of transfer (30 min a t 15 V) was needed for illustrates by silver staining, a single protein of -66 66 kDa BPP. Subsequent to blotting, the membranes kDa is obtained after the second S-300 Sephacryl chrowere stored and air dried at 4°C. The transferred pro- matographic procedure. Prior to antibody preparation, tein was visualized immunochemically with alkaline- protein purity was assessed by amino terminal sephosphatase-conjugated goat anti-rabbit IgG followed quence analysis of the second S-300 chromatographic by staining with nitro blue tetrazolium (Sigma Chem- peak or by electro-blotting onto a PVDF membrane. An ical Co., St. Louis, MO). Non-specific antibody binding identical sequence of high quality and good yield was was blocked by using 5% non-fat dry milk (Carnation obtained for both analyses at recoveries of 60-150 Co., Los Angeles, CA) in tris-HC1, pH 8.0, containing pmoles and no secondary sequence was observed 150 mM NaCl and 0.5% Tween 20 (Johnson et al., (Gotoh et al., 1990). 1984). Antibody specificity to the 66 kDa BPP was assessed by Western blot analysis of total protein extracts from Electron Microscopy and lmmunocytochemistry both embryonic bone and cultured osteoblasts. The anFor ultrastructural analysis by scanning electron tibody reacted strongly with a protein a t -66 kDa and microscopy, 30-day mineralizing osteoblast cultures less strongly with a protein a t -45 kDa (Fig. 2A). The were fixed overnight at 4°C with 3% glutaraldehyde in 45-kDa species is a proteolytic product of the 66-kDa 0.1 M sodium cacodylate buffer, pH 7.3, and post-fixed protein which accumulates in the extracellular matrix with potassium ferrocyanide-reduced osmium tetrox- of the bone (Gotoh et al., 1989). Indirect immunopreide (Karnovsky, 1971) for 1 h r a t 4°C. The cultures cipitation of (,H)-leucine-labeled protein (Fig. 2A) were then washed extensively with 0.1 M sodium ca- demonstrated that the -66 kDa phosphoprotein was a codylate buffer containing 4% sucrose, pH 7.3, dehy- major secretory protein synthesized by the osteoblast drated through a graded ethanol series, and critical- cultures. Two additional (3H) bands were observed in O O B 1.5 B T IP 1.o 0.5 kD6645- 0 1.5 29- a 10- s .- 1.0 rn c 4- A Q, Stained Immuno- (d V .. I a 0 100 50 I A S-300. A 40 50 60 2nd 70 3(H) Leucine Blot 0 0.5 I m 0 80 Fraction Number 1 2 3 B 66KD- Fig. 1. Purification of the major 66-kDa chicken bone phosphoprotein (66-kDa BPP). A: Representative chromatography profiles of each step of the 66-kDa BPP phosphoprotein purification. The DEAESephacel gradient was from 0-500 mM NaCl and the gradient was initiated at fraction 20. Subsequent purification was performed on two sequential S-300 Sephacryl columns. Bars represent pooled fractions used for each subsequent purification procedure. B: SDS polyacrylamide gel electrophoresis profiles of each purification step. Gels were 5-10% linear gradients and proteins were visualized by silver staining. 1: Total protein applied to DEAE Sephacel. 2 Fraction D of DEAE Sephacel (top panel, A). 3: Second S-300 fraction. Fig. 2. Characterization of the 66-kDa BPP polyclonal antibody. A Analysis of antibody specificity for BPP in total protein extracts from 30-day cultured osteoblasts (0)or total HCl protein extracts from 14-wk bone (B). Total newly synthesized media protein from 30-day cultures (TI was indirectly precipitated UP). Samples were resolved on linear 5-10% gradient SDS polyacrylamide gels. The method by which individual lanes were visualized is denoted a t the bottom of the figure. B Immunoblot analysis of antibody specificity for 66-kDa BPP in total protein extracts from 30-day osteoblasts (OB), tendon fibroblasts (Fibro), myoblasts (Myo), and 4-day sternal chondrocytes (Chon). the fluorograph, one a t -50 kDa, which has previously been shown to be a result of non-specific (3H)-labeled protein binding to the large subunit of the IgG used in indirect immunoprecipitations (Gerstenfeld et al., 1984),and a band a t -45 kDa, which presumably is the proteolytic product described above. Several different cell types of mesenchymal origin were examined for the presence of 66 kDa BPP. A comparable quantity of extracted protein from cultured sternal chondrocytes, myoblasts, and tendon fibroblasts was blotted and reacted with the 66 kDa BPP antibody. As Figure 2B shows, the 66-kDa BPP was only detected in proteins extracted from the mineralizing osteoblast culture. IMMUNOLOCALIZATION OF SYNTHESIZED PHOSPHOPROTEIN 97 phosphorylated protein accretion within the culture matrix, the nature of the specific phosphorylated proteins that were synthesized was assessed by polyacrylamide gel electrophoresis of (3H)-leucine and (32P) labeling of proteins, and specific 66 kDa BPP accumulation was determined by Western blot analysis. The results of hosphoprotein synthesis analyses, by labeling with (8P) are shown in Figure 5A. Relatively little synthesis of (”P)-labeled 66 kDa protein was seen before day 18, whereas thereafter, synthesis remained elevated and labeled protein was found associated with both the cell layer and the media. It is important to note the predominance of low molecular weight phosphoprotein species in the cell layer and the general increase in their quantity a t later time points. Since analysis of (32P)-labeled protein only examines posttranslational phosphorylation, and i t may not reflect actual protein synthesis, cultures were additionally labeled with (3H)-leucine to generate data shown in Figure 5B. The time frame of induction seen for this analysis was identical to that observed for the (32P) labeling, a result indicating that protein phosphorylation was coordinately regulated with the induction of protein synthesis. Days in Culture The accumulation of 66 kDa BPP was monitored by Western blot analysis of the culture cell layers a t varFig. 3. Mineral accumulation as a function of culture time. Total Ca was determined at 6-day increments over a 36-day experimental peious times throughout the 30-day experimental period. riod. Measurements represent the average of triplicate samples each Figure 5C shows that the 66-kDa BPP accumulation determined from three separate 100-mm dishes. Error bars denote parallels the total phosphoprotein content as detertotal range of variation. mined by amino acid analysis. As seen for the (32P)labeled protein, a progressive increase in the lower molecular weight species was found at later time points for the accumulated protein in the cell layer. This rePhosphoprotein Expression During Extracellular Matrix sult further suggests progressive proteolytic breakMineralization In Vitro down of the 66-kDa BPP as it accumulates with time in The temporal relationships between extracellular the extracellular matrix. matrix mineralization and phosphoprotein synthesis and accumulation were examined in vitro. The total lmmunocytochemical Localization of 66-kDa BPP in the accumulation of calcium (Ca) as a measure of mineral Extracellular Matrix of Cultured Osteoblasts in the cell layer is depicted in Figure 3. An apparent Immunocytochemistry using the 66-kDa BPP antibiphasic increase in total calcium was observed in the 36-day period. From day 6 to day 30, there was a n body was carried out to determine the microscopic loinitial very slow, 3-4-fold increase observed in min- calization of 66-kDa BPP and to compare its distribueral, subsequently followed by a very rapid 3.5-fold in- tion to that seen in vivo. Ultrastructural examination crease between days 30 and 36. In order to assess the of the surface of a 30-day culture by scanning electron relationship of total phosphoprotein synthesis by osteo- microscopy revealed numerous flattened osteoblasts blasts in culture to the accumulation of mineral, cul- having long cell processes extending between cells and tures were labeled with (3H)-Ser and ( 14C)-Thrfollowed throughout the layer of the extracellular collagen maby determination of the incorporation of these labels trix (Fig. 6A). Additional cell processes were observed into phosphoprotein. Total phosphoprotein accretion in about the cells in regions not associated with matrix. the matrix was simultaneously determined by a n anal- The extracellular matrix was comprised of a n extenysis of Ser-P and Thr-P accumulation. As shown in sive layer of interwoven and well-developed collagen. Figure 4, total phosphoprotein synthesis rapidly in- In cross-sectioned “on edge” views of these cultures, creased 3-4-fold after day 12 with (3H)-Ser-P levels osteoblasts were positioned between layers of extracelpeaking a t day 18 and (I4C)-Thr-Plevels reaching a lular matrix such that as many a s 4-6 alternating maximum a t day 24. Thereafter, protein synthesis uti- layers of cells and matrix were present by 30 days of lizing both phosphorylated amino acids decreased, but cell culture (Fig. 6B). Successive layers of collagenous it remained 1.5 times greater than the initial measured matrix were approximately equal in thickness. levels. Total accumulated phosphoprotein, as assessed Whereas a continuous monolayer of cells was generally by phospho-amino acid incorporation, showed a slow present a t the upper surface of the culture, this was not and continual 4-fold increase over the same 30-day pe- a consistent feature of the lower surface a s collagen riod and subsequently remained constant between days fibrils could often be observed immediately adjacent to the plastic Petri dish. 30 and 36 (Fig. 4). immunocytochemistry of 30-day osteoblast cultures While the accumulation of phosphorylated amino acids and incorporation of label into Ser-P and Thr-P showed a distribution of gold immunoreactivity, idenmeasured total phospho-amino acid synthesis and tifying the sites of 66-kDa BPP antigenicity, practi- 98 L.C. GERSTENFELD ET AL. CULTURED CHICK OSTEOBLAST PHOSPHOPROTEINS I S y n t h e s i s (Media) PSer PThr -a ' O0 Y xZ I Accumulation (cell L a y e r ) *OOt 150- EI 9 /. 5 a$ 100ZQ I n 0 50- 0 ./O 6 12 18 24 6 30 12 18 24 30 36 Days in Culture Fig. 4. Analysis of total phospho-amino acid synthesis and accumulation as a function of culture time. Phospho-amino acid synthesis was determined by amino acid analysis of pooled cell layers and media protein from each time point. Ninhydrin, Ser-P, and Thr-P peaks were collected and (3H)-Ser-P and (I4C)-Thr-P in each peak were determined. Each data point represents the average of two determinations. Total accumulations were determined from the average of triplicate amino acid analysis determinations. cally identical to that observed in vivo (McKee et al., 1990). Immunolabeling was found principally over extracellular matrix regions undergoing mineralization (Fig. 7A-C). Gold particles were most often associated with electron-dense patches of organic material in close association with collagen fibrils within the matrix. These immunolabeled matrix sites contained relatively high levels of Ca and P, a conclusion based on electron probe x-ray microanalysis of other sections of glutaraldehyde-fixed, unstained sections not subjected to the immunocytochemistry (data not shown). Following the immunocytochemical incubation procedures and staining of sections with uranyl acetate and lead citrate, Ca and P were lost from the tissue. These results demonstrate that 66-kDa BPP is associated with mineralization foci in the culture extracellular matrices. Since the tissue processing and immunocytochemical technique used here retain the phosphoproteins (McKee e t al., 1990) and the 66-kDa BPP is specifically localized to the dense patches as shown by immunolabeling, i t appears that the highest concentrations of 66 kDa BPP are to be found in such dense extracellular foci. expression of bone phosphoproteins and their relationship with other osteoblast functions and extracellular matrix proteins (such as, for example, osteocalcin, alkaline phosphatase, type I collagen, and fibronectin). The work provides a very well-defined in vitro model system of mineralization for correlation with the mineralization processes t h a t occur in vivo. A schematic summary of the temporal pattern of synthesis of 66-kDa BPP is presented in Figure 8 in relation to the other proteins previously examined in this model system. Several interesting correlations may be obtained from this comparison. Phosphoprotein expression increases concurrent with the expression of other proteins associated with advanced osteoblast differentiation, such as osteocalcin and alkaline phosphatase, a result suggesting that phosphoprotein synthesis is restricted to more differentiated or mature osteoblasts. In contrast, collagen and fibronectin, which have a more generalized expression in many connective tissues, are maximally expressed during the period of rapid cell proliferation associated with early culture times (Shalhoub e t al., 1989). These results would suggest that connective tissue proteins such a s collagen and fibronectin are necessary prerequisites for the growth of osteoblasts and for the initial assembly of the extracellular matrix. The subsequent expression, then, of high levels of 66-kDa BPP, osteocalcin, and alkaline phosphatase in bone are no doubt a reflection of the extracellular matrix specialization. It is also interesting to note that phosphoprotein begins to accumulate concurrent with, or slightly preceding, the first increases in mineral deposition. These results obtained in vitro are qualitatively identical to the sequence of expression seen in vivo for collagen type I (Moen et al., 1979), osteocalcin (Hauschka and Reid, 1978), and phosphoprotein (McKee e t al., 1990), in which collagen synthesis temporally precedes both osteocalcin and phosphoprotein synthesis during embryonic chicken DISCUSSION In a previous study from this laboratory of embryonic chicken bone development in vivo, analysis of the ultrastructural localization and extracellular matrix accumulation of the same major 66-kDa BPP examined here showed a n intimate association of this protein with areas of mineralization and a n unequivocal quantitative correlation between increased phosphoprotein accumulation and mineral deposition (McKee e t al., 1989). In the present work, phosphoprotein synthesis, extracellular matrix accumulation, and ultrastructural localization were examined in vitro in cultured embryonic chicken osteoblasts and compared with the results obtained in vivo. The investigation also was carried out in order to define more clearly the temporal 99 IMMUNOLOCALIZATION OF SYNTHESIZED PHOSPHOPROTEIN Days Day 6 18 30 6 3 9 21 24 30 18 30 MWx103 116946645- 29Osteocalcin A MEDIA CELL 12 18 24 30 66kDa- C MEDIA CELL bone development. Similar temporal profiles are found for the expression of these same collagenous and noncollagenous extracellular matrix proteins [including the rat phosphoprotein, osteopontin (Prince et al., 1987)] during mammalian embryonic bone development (Yoon et al., 1987). An identical sequence of expression for these respective genes was also seen in cultured primary rat osteoblasts (Stein e t al., 1989). This commonality indicates t h a t closely related mechanisms of bone matrix assembly and mineralization may be operative in all vertebrates. The chicken 66-kDa BPP shares many biochemical features with the rat phosphoprotein, osteopontin. Both proteins are glycosylated and contain sialic acid, have phosphorylated threonine and serine residues, have similar DEAE chromatography elution characteristics, and have approximately the same apparent sizes on gel electrophoresis (Franzen and Heinegard, 1985; Prince e t al., 1987). Analysis of recent sequence data for the chicken 66-kDa BPP (with -25% of the protein sequence completed and 30% of the cDNA sequence completed) shows that there is little homology with rat osteopontin. However, chicken 66-kDa BPP contains a nine-amino-acid sequence, with two conservative substitutions, that contains the Arg-Gly-Asp cell-binding sequence of osteopontin (Gotoh et al., 1990) and has a 6 I?- B Fig. 5. Characterization of 66-kDa BPP synthesis as a function of culture time. A Specific protein phosphorylation was examined by (32P)labeling for 6 hr. Both cell layer and media were separately analyzed and the days examined are denoted at the top of the figure. A constant 50 pg of protein was resolved on a linear 510% gradient SDS polyacrylamide gel and autoradiographic exposure was for 24 hr. B Specific protein synthesis was examined by (3H)-leucine labeling for 24 hr. The secreted proteins of the media were analyzed and the days examined are denoted in the figure. A constant 5 x lo4 cpm of (3H)-leucinewas applied for each sample and the proteins were resolved on a linear 5-18% gradient SDS polyacrylamide gel. Fluorographic exposure was for 3 days. The positions of osteocalcin, 66-kDa BPP, collagen a chains, and procollagen are denoted in the figure. An osteocalcin standard was included in the left lane as a reference. C : 66-kDa BPP accumulation detected by Western blot analysis. Both cell layer and media were separately analyzed and the days examined are denoted in the figure. A constant 50 pg of protein was resolved on a linear 5-10% gradient SDS polyacrylamide gel. nine-amino-acid sequence of continuous aspartatic acid residues (Moore, Gotoh, and Gerstenfeld, unpublished data). Indeed, whole embryo immunolocalization studies using the 66-kDa BPP antibody shows a n immunohistochemical distribution of staining in the intestine and kidney (Bruder et al., 1990) that is similar to the localization for osteopontin and osteopontin mRNA (Mark et al., 1987a,b, 1988; Nomura et al., 1988). Furthermore, 66-kDa BPP promotes cellular adherence of MG63 osteosarcoma cells (M. Pierschbacher, unpublished results), a property also observed for osteopontin and periodontal fibroblasts (Somerman et al., 1987). These results would suggest that 66-kDa BPP is either the chicken homologue to the rat protein, but has undergone extensive sequence divergence during evolution, or is a very similar functionally related protein. The presence of a n Arg-Gly-Asp sequence in the rat phosphoprotein, osteopontin, and the cell attachment studies of Oldberg e t al. (1986, 198813) and Somerman et al. (1987) led these researchers to hypothesize that this protein transmitted positional information directly from the osteoblast surface to the mineralized bone matrix. The synthesis of fibronectin determined in studies from this laboratory (Winnard et al., 1989) during the early period of culture growth indicates that osteoblasts synthesize multiple Arg-Gly-Asp-contain- 100 L.C. GERSTENFELD ET AL. Figs. 6-7. 101 IMMUNOLOCALIZATION OF SYNTHESIZED PHOSPHOPROTEIN ing proteins and that fibronectin serves to promote cell attachment and spreading during osteoblast proliferation, while the later expression of 66-kDa BPP probably is related to a different function. Recent data indicate that the class of integrin receptor that recognizes bone sialoprotein and osteopontin is more similar to thevitronectinreceptor (Oldberget al., 1988a,b),aresult thereby suggesting that osteoblasts may also have multiple types of integrin receptors which impart different types of positional information. In the present study in vitro and in the accompanying study in vivo (McKee et al., 1990), chicken 66-kDa BPP was observed predominantly in the extracellular matrix, was clearly associated with mineral, and was well removed from the cell surface. These results indicate that any cell surface association of 66-kDa BPP is probably only transient in nature. Several possible types of positional information might be imparted by such a transient cell surface association. Previous results indicated that cultured osteoblasts assemble their collagen fibrils in a tissue-specific orthogonal array (Gerstenfeld et al., 1988). Since 66-kDa BPP accumulates a t extracellular sites of mineralization in a n identical fashion both in vitro and in vivo, specific positional information may be needed during matrix assembly to direct specific interactions of matrix proteins during their translocation from the cell to the matrix. Other functions that may be transient in nature might involve the post-translational modification of 66-kDa BPP by specific cell-surface-associated kinases, phosphatases, or proteases. Alternatively, the localization of 66-kDa BPP in the kidney and intestine, Fig. 6. General ultrastructural characteristics of a 30-day osteoblast culture. A Scanning electron micrograph of the upper surface of a 30-day osteoblast culture. Osteoblasts (OB) generally cover the underlying collagenous matrix (Coll) and develop long cell processes (asterisks) extending both among the numerous collagen fibrils and into the space not associated with matrix just above the cells. B: Scanning electron micrograph of a 30-day osteoblast culture fractured open to reveal a n “on edge” view of the tissue. Layers of osteoblasts (OB) alternate with layers of the collagenous matrix (Coll) to form 3 relatively thick bone-like tissue having a lamellar structure. Fig. 7. Immunolocalization of 66-kDa BPP in a 30-day mineralizing osteoblast culture, fixed with glutaraldehyde and osmium tetroxide, embedded in Epon, and stained with uranyl acetate and lead citrate. A Transmission electron micrograph after incubation of the tissue section with the 66-kDa BPP antibody and the protein A-gold complex. The most intense specific immunolabeling is observed over areas of extracellular matrix undergoing mineralization (arrowheads) and is associated with electron-dense organic material in close proximity to collagen fibrils (Coll). Some gold particles are also observed nonspecifically over areas of collagenous matrix not associated with detectable mineral. OB, osteoblast. B: Higher-magnification electron micrograph of a site of mineralization in the extracellular matrix (MM). In this region some collagen is visible in partial longitudinal profiles and a specific immunocytochemical reaction is found over electron-dense organic matrix in close association with the fibrils. The electron-dense foci contain high levels of Ca and P detectable by x-ray microanalysis, but these elements are lost following antibody incubation and staining with uranyl acetate and lead citrate. The observed uptake of stain is indicative of an organic composition of the foci, contributed in part by 66-kDa BPP. C: Electron micrograph of the extracellular matrix of a 30-day mineralizing osteoblast culture in which the collagen fibrils (Coll) are predominantly sectioned transversely. Numerous gold particles are observed specifically over the electron-dense organic matrix associated with areas of mineralization among the fibrils (arrowheads). 3 6 12 18 24 TIME IN CULTURE 30 Fig. 8. Schematic representation of the temporal profiles of osteoblast functions during mineralization in vitro. A summary comparison of the percent maximal of each culture component is depicted for different osteoblast functions. Data for osteocalcin and alkaline phosphatase are from Gerstenfeld et al. (1987). Data for collagen are from Gerstenfeld et al. (1988), for fibronectin from Winnard et al. (1990), and for phosphoprotein and mineral from the present study. and by analogy to the known vitamin D and PTH regulation of osteopontin expression in the rat and phosphoprotein expression in the chicken (Prince and Butler, 1987; Noda and Rodan, 1988; Lian et al., 19821, might indicate that this class of proteins plays a generalized role in phosphate metabolism, functioning either in phosphate mobilization or phosphate transport. However, neither the exact phosphorylation state of this protein in these tissues nor its quantitative expression is currently known. In summary, the data presented here demonstrate that 66-kDa BPP is expressed only concurrently with other differentiated osteoblast functions. The ultrastructural localization of this phosphoprotein in vitro is identical to that seen in vivo, thereby indicating that osteoblasts in vitro retain their phenotypic ability to express proteins that appear to be important in regu- 102 L.C. GERSTENFELD ET AL. lating the assembly and mineralization of their extracellular matrix. The chicken 66-kDa BPP is primarily associated with the extracellular organic matrix and does not accumulate on the osteoblast cell surface. These data obtained in vitro and those of McKee et al. (1989, 1990) from work in vivo possibly reflect on the suggestion t h a t a n Arg-Gly-Asp-containing protein may be only transiently associated with the cell surface and may transmit positional information a t a distance well removed from the cell surface. its possible linkage to the organic matrix by protein-bound phosphate bonds. Philos. Trans. R. SOC. Lond. [Biol.], B304:479-508. Glimcher, M.J. 1989 Mechanisms of calcification in bone: Role of collagen fibrils and collagen-phosphoprotein complexes in uztro and in uiuo. Anat. Rec., 224t139-153. Glimcher, M.J., D. Brickley-Parsons, and D. Kossiva 1982 Proof that the phosphoproteins containing 0-phosphoserine and O-phosphothreonine are synthesized in bone. Trans. Orthop. Res. SOC.,7r36. Gotoh, Y., M. Sakamoto, S. Sakamoto, and M.J. Glimcher 1983 Biosynthesis of 0-phosphoserine-containingphosphoproteins by isolated bone cells of mouse calvaria. FEBS Lett., 154:116-120. Gotoh, Y., L.C. Gerstenfeld, and M.J. Glimcher 1990 Identification and characterization of the major bone specific phosphoprotein synthesized by cultured embryonic chick osteoblasts. Eur. J. BioACKNOWLEDGMENTS chem., in press. Gundberg, C.M., P.V. Hauschka, J.B. Lian, and P.M. Gallop 1987 This work was supported by grants from the NaOsteocalcin: Isolation, characterization, and detection. In: Methtional Institutes of Health (AR 34078, AR 34081, and ods of Enzymology. K. Moldave, ed. Academic Press, New York, Vol. 107, pp. 516-544. HD22400), the National Aeronautics and Space AdP.V., and M.L. Reid 1978 Timed appearance of a calcium ministration (NAG-2-5381, the Orthopaedic Research Hauschka, binding protein containing y-carboxyglutamic acid in developing and Education Foundation Bristol-MyersiZimmer Corchick bone. Dev. Biol., 65t426-434. poration Institutional Grant, the Peabody Foundation, Heinegard, D., K. Hultenby, A. Oldberg, F. Reinholt, and M. Wendle 1989 Macromolecules in bone matrix. Connect. Tissue Res., 21: the Medical Research Council of Canada (A.N.), and 3-14. CAFIR of the University of Montreal (A.N.). The au- Hunkapillar, M.W., and E. Lujan 1986 In: Methods of Protein Microthors are grateful to the Department of Anatomy, characterization. J.E. Shively, ed. Human Press, Clifton, New McGill University, for use of their x-ray microanalytJersey. ical equipment and to M. Fortin, G. Lambert, and S. Johnson, D.A., J.W. Gautsch, J.W. Sportsman, and J.H. Elder 1984 Improved method for utilizing non-fat dry milk for analysis of Zalzal (Montreal) and to K. Hodgens and K. Silva (Bosproteins and nucleic acids transferred to nitrocellulose. Gene ton) for expert technical assistance. Anal. Tech., 1:3-8. Karnovsky, M.J. 1971 Use of ferrocyanide-reduced osmium tetroxide in electron microscopy. Am. SOC. Cell Biol., Abstract 284, New LITERATURE CITED Orleans, 146. Bendayan, M. 1984 Protein A-gold electron microscopic immunocy- Laemmli, U.K. 1970 Cleavage of structural proteins during assembly tochemistry methods, applications and limitations. J. Electron of the head of bacteriophage T4. Nature 277t680-685. Microsc. Tech., lr243-270. Landis, W.J., C.F. Sanzone, D.-Brickley-Parsons, and M.J. Glimcher Bonner, W.M., and R.A. Laskey 1974 A film detection method for 1984 Radioautographic visualization and biochemical identificatritium labeled proteins and nucleic acids in polyacrylamide gels. tion of 0-phosphoserine and 0-phosphothreonine-containing Eur. J. Biochem., 46:83-88. phosphoproteins in mineralizing embryonic chick bone. J. Cell Bruder, S., A. Caplan, Y. Gotoh, L.C. Gerstenfeld, and M.J. Glimcher Biol., 981.986-990. 1990 Immunohistochemical localization of a -66 kDa glycosy- Laskey, R.A., and A. Mills 1975 Quantitative film detection of 3H and lated phosphoprotein during development of the embryonic chick ’“C in polyacrylamide gels by fluorography. Eur. J. Biochem., tibia. Calcif. Tissue Int., in press. 56t335-341. Cohen-Solal, L., J.B. Lian, D. Kossiva, and M.J. Glimcher 1978 The Lee, S.L., and M.J. Glimcher 1981 The purification, composition and identification of 0-phosphothreonine in the soluble non-collage31Pspectroscopic properties of a non-collagenous phosphoprotein nous phosphoproteins of bone matrix. FEBS Lett., 89t107-110. isolated from chicken bone matrix. Calcif. Tissue Int., 33t385Fisher, L.W., S.W. Whitson, L.V. Avioli, and J.D. Termine 1983 Ma394. trix sialoprotein of developing bone. J. Biol. Chem., 258t12723Lian, J.B., L. Cohen-Solal, D. Kossiva, and M.J. Glimcher 1982 12727. Changes in chicken bone matrix in vitamin D deficient rickets. Fisher, L.W., G.R. Hawkins, N. Tuross, and J.D. Termine 1987 PuriFEBS Lett., 144:123-125. fication and partial characterization of small proteoglycans I and Mark, M.P., C.W. Prince, S. Gay, R.L. Austin, M. Bhown, R.D. Finkel11, bone sialoproteins I and 11, and osteonectin from the mineral man, and W.T. Butler 1987a A comparative immunocytochemical compartment of developing human bone. J. Biol. Chem., 262: study on the subcellular distributions of 44 kDa bone phosphopro9702-9708. tein and bone y-carboxyglutamic acid (Gla)-containingprotein in Franzen, A., and D. Heinegard 1985 Isolation and characterization of osteoblasts. J. Bone Miner. Res., 2t337-346. two sialoproteins present only in bone calcified matrix. Biochem. Mark, M.P., C.W. Prince, T. Oosawa, S. Gay, A.L.J.J. Bronckers, and J., 232t7 15-724. W.T. Butler 198713 Immunohistochemical demonstration of a 44 Frens, G. 1973 Controlled nucleation for the regulation of the particle kDa phosphoprotein in developing rat bones. J. Histochem. Cysize in monodisperse gold solutions. Nat. Phys. Sci., 241t20-22. tochem., 35t707-715. Gerstenfeld, L.C., J.C. Beldekas, G.E. Sonenshein, and C. Franzblau Mark, M.P., W.T. Butler, C.W. Prince, R.D. Finkelman, and J.-V. 1984 Processing of procollagen types I11 and I in cultured bovine Ruch 1988 Developmental expression of 44 kDa bone phosphosmooth muscle cells. J. Biol. Chem., 259t9158-9162. protein (osteopontin) and bone y-carboxyglutamic acid (GlahonGerstenfeld, L.C., S.D. Chipman, J. Glowacki, and J.B. Lian 1987 taining protein (osteocalcin) in calcifying tissues of the rat. DifExpression of differentiated function by mineralizing cultures of ferentiation, 37t123-136. chicken osteoblasts. Dev. Biol., 122:49-60. McKee, M.D., A. Nanci, W.J. Landis, L.C. Gerstenfeld, Y. Gotoh, and M.J. Glimcher 1989 Ultrastructural immunolocalization of a maGerstenfeld, L.C., S.D. Chipman, C.M. Kelly, K.J. Hodgens, D.D. Lee, and W.J. Landis 1988 Collagen expression, ultrastructural asjor phosphoprotein in embryonic chick bone. Connect. Tissue sembly and mineralization in cultures of chicken embryo osteoRes., 2 1 2 - 2 9 . McKee, M.D., A. Nanci, W.J. Landis, Y. Gotoh, L.C. Gerstenfeld, and blasts. J. Cell Biol., 106t979-989. M.J. Glimcher 1990 Developmental appearance and ultrastrucGerstenfeld, L.C., J.B. Lian, Y. Gotoh, D.D. Lee, W.J. Landis, M.D. tural immunolocalization of a major 66 kDa phosphoprotein in McKee, A. Nanci, and M.J. Glimcher 1989 Use of cultured emembryonic and post-natal chicken bone. Anat. Rec., submitted bryonic chicken osteoblasts as a model of cellular differentiation (this issue). and bone mineralization. Connect. Tissue Res., 21t215-225. Moen, R.C., D.W. Rowe, and R. Palmiter 1979 Regulation of procolGlimcher, M.J. 1976 Composition, structure, and organization of bone lagen synthesis during the development of chick embryo calvaria. and other mineralized tissues and the mechanisms of calcificaJ. Biol. Chem., 254:3526-3530. tion. In: Handbook of Physiology,” Vol. 7, Endocrinology. R.O. Greep and E.B. Astwood, eds. American Physiological Society, Noda, N., and G.A. Rodan 1988 Transcriptional regulation of osteopontin production in rat osteoblast like cells by parathyroid horWashington, D.C., pp. 25-116. mone. J. Cell Biol., 108t713-718. Glimcher, M.J. 1984 Recent studies of the mineral phase in bone and IMMUNOLOCALIZATION OF SYNTHESIZED PHOSPHOPROTEIN Nomura, S., A.J. Wills, D.R. Edwards, J.K. Heath, and B.L.M. Hogan 1988 Developmental expression of 2ar (osteopontin) and SPARC (osteonectin) RNA as revealed by in situ hybridization. J. Cell Biol., 106t441-450. Oldberg, A., A. Franzen, and D. Heinegard 1986 Cloning and sequence analysis of rat bone sialoprotein (osteopontin) cDNA reveals a n Arg-Gly-Asp cell binding sequence. Proc. Natl. Acad. Sci. USA, 83t8819-8823. Oldberg, A., A. Franzen, and D. Heinegard 1988a The primary structure of a cell binding bone sialoprotein. J. Biol. Chem., 263: 19430-19432. Oldberg, A., A. Franzen, D. Heinegard, M. Pierschbacher, and E. Ruoslahti 198813 Identification of a bone sialoprotein receptor in osteosarcoma cells. J. Biol. Chem., 263,19433-19436, Prince, C.W., and W.T. Butler 1987 1,25-dihydroxyvitamin D, regulates the biosynthesis of osteopontin, a bone derived cell attachment protein in clonal osteoblast-like osteosarcoma cells. Coll. Relat. Res., 7t305-313. Prince, C.W., T. Oosawa, W.T. Butler, M. Tomana, AS. Bhown, M. Rhown. and R.E. Schronhenloher 1987 Isolation. characterization and biosynthesis of a phosphorylated glycoprotein from rat bone J Biol Chem , 2 6 2 2900-2907 Shalhoub, V., L.C. Gerstenfeld, D. Collart, J.B. Lian, and G.S. Stein 1989 Down regulation of cell growth and cell cycle regulated genes during chick osteoblast differentiation with the reciprocal expression of histone gene variants. Biochemistry, 28t53185322. Somerman, M.J., C.W. Prince, J. Sauk, R.A. Foster, and W.T. Butler 1987 Mechanism of fibroblast attachment to bone extracellular 103 matrix: Role of a 44 kilodalton bone phosphoprotein. J. Bone Miner. Res., 2t259-265. Spector, A.R., and M.J. Glimcher 1972 The extraction and characterization of soluble anionic proteins from bone. Biochim, Biophys. Acta, 303t360-362. Stein, G.S., J.B. Lian, L.C. Gerstenfeld, V. Shalhoub, M. Aronow, T. Owen, and E. Markose 1989 The onset and progression of osteoblast differentiation is functionally related to cellular proliferation. Connect. Tissue Res., 20:3-13. Towbin, H., and J. Gordon 1984 Immunoblotting and dot immunobinding: Current status and outlook. J . Immunol. Methods, 72: 313-340. Uchiyama, A,, M. Suzuki, B. Lefteriou, and M.J. Glimcher 1986 Isolation and chemical characterization of the phosphoproteins of chicken bone matrix: Heterogeneity in molecular weight and composition. Biochemistry, 25t7572-7583. Veis, A. 1978 The role of acidic proteins in biological mineralization. In: Ions in Macromolecular and Biological Systems. D.H. Everett and B. Vincent, eds. Scientechnia, Bristol. Veis, A., and B. Sabsay 1983 Bone and tooth formation. Insights into mineralization strategies. In: Biomineralization and Biological Metal Accumulation. P. Westbroek and E.W. deJong, eds. D. Reidel Pub. Co., Dordrecht, Netherlands, pp. 273-284. Winnard, R.G., R.T. Francheschi, and L.G. Gerstenfeld 1990 Fibronectin gene expression, synthesis and accumulation in mineralizing cultures of chicken osteoblasts. Trans. Orthop. Res. SOC., 15t395. Yoon, K., R. Buenaga, and G. Rodan 1987 Tissue specificity and expression of rat osteopontin. Biochem. Biophys. Res. Commun., 148t1129-1136.