Phosphoproteins of Chicken Bone Matrix PROOF OF SYNTHESIS IN BONE TISSUE*

Twelve-day-old embryonic chick mandibles cultured vitro that O-phosphoserine-and O-phosphothreonine-containing phosphoproteins which are similar to those present in embryonic and postnatal chicken bone The synthesis of the phosphoproteins identified in and in the EDTA-extractable phosphoproteins after mandibles in with radioactively labeled serine and threonine,

Twelve-day-old embryonic chick mandibles were cultured in vitro for 6 days. Measurements of the weights of the explants, their mineral and protein components, and the EDTA-extractable proteins established that bone tissue synthesizes O-phosphoserineand O-phosphothreonine-containing phosphoproteins which are similar to those present in embryonic and postnatal chicken bone matrix.
The synthesis of the phosphoproteins was further confirmed by the demonstration that radioactively labeled O-phosphoserine and O-phosphothreonine were identified in bone and in the EDTA-extractable phosphoproteins after pulse-labeling chick mandibles in vitro with radioactively labeled serine and threonine, respectively.
Phosphoproteins, present in all normally and pathologically calcified vertebrate tissues (1-4), have been postulated to play a significant role in the deposition of the solid phase of calcium-inorganic phosphorus during mineralization (1-7). While tissue culture experiments (8) and indirect 33P radioautographic evidence (9) strongly suggest that the phosphoproteins of dentin are synthesized in the tissue, no similar biochemical evidence has been presented for the bone phosphoproteins, which are chemically quite distinct from those of dentin and enamel (10). Indeed, two major noncollagenous proteins in bone matrix, albumin and a2HS-glycoprotein, are synthesized in the liver and then transported to bone where they are adsorbed and concentrated (11)(12)(13)(14).
From experiments utilizing bone cultured in uitro, we present data which establish for the first time that phosphoproteins containing O-phosphoserine and O-phosphothreonine, which behave chromatographically similar to well characterized bone matrix phosphoproteins from both embryonic and postnatal chickens, are synthesized by bone tissue. A preliminary report of these experiments has already been presented (14).

Tissue Culture Techniques
No Radioactively Labeled Acids Utilized-Portions of 12-day-old embryonic chick mandibles, 8.1 mm in length, were used as tissue * These studies were supported by National Institutes of Health Grant AM-15671 and by a grant from the New England Peabody Home for Crippled Children, lnc., Boston, MA. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ T o whom reprint requests should be addressed at, Children's Hospital Medical Center, 300 Longwood Avenue, Boston, MA 02115. culture explants in a completely defined synthetic medium according to methods previously described (15)(16)(17)(18). In this group of experiments carried out without the use of radioactively labeled amino acids, the tissue culture medium was changed every 2 days for a total incubation period of 6 days. Twelve-day-old chick mandibles which were not incubated and those incubated for 6 days and chick mandibles from 18-day-old chick embryos grown in oua were weighed and dried, and their dry weights, ash weights, and calcium, magnesium, inorganic orthophosphate, total protein, collagen, Ser(P),' Thr(P), and Gla contents were determined as described elsewhere (19)(20)(21)(22). From these analytic data it was possible to compute the composition of the mandibles before and after tissue culture and after growth in oua, as well as to detect any change in the net amounts of the major inorganic and organic constituents of the tissue which occurred during growth in tissue culture and in ouo.
Samples of 12-day-old embryonic chick mandibles before and after culture and mandibles from 18-day-old chick embryos grown in ouo were freeze-dried, powdered at liquid nitrogen temperature, and extracted for 10-14 days in 0.4 M EDTA, pH 7.5, at 4 "C in the presence of protease inhibitors (19). The EDTA extracts were dialyzed free of EDTA and salts at 4 "C by ultrafiltration in an Amicon Dialyzer Concentrator with a DC2 HIP5 cartridge which retains >5000 daltons (Amicon Corp., Lexington, MA). Ammonium bicarbonate, 0.05 M, pH 7.9, was used with the same inhibitors in the final reservoir. The samples were then lyophilized.

Tissue Culture in Presence of Radioactively Labeled Amino Acids-
Experimentsutilizing [3H]proline to determine the conversion of [3H] proline to [3H]hydroxyproline and the proportion of the [3H]hydroxyproline which is incorporated into the tissue explant were carried out as described previously (15). In the first group of experiments utilizing [3H]serine and [3H]threonine, the radioactively labeled amino acids were added to the tissue culture in separate independent experiments and the tissue explants were exposed for 4-8 h to the radioactively labeled amino acid. The tissue culture media used during this puke time period were made up free of serine and threonine, respectively. Media for all of the subsequent changes contained serine and threonine.
[3H]Serine and ['Hlthreonine at 1.5 &/ml were used. A total of 1.5 ml was used for each tissue culture medium change in a tissue culture dish containing four mandibles. After the initial 4-8-h pulse labeling, the explants were washed with cold tissue culture medium and the washings were added to the tissue culture medium which had been removed. The explants were recultured and the procedure was repeated for the time periods shown under "Results." The extracts removed at the end of the total tissue culture period were washed and hydrolyzed appropriately (7), as were the individual tissue culture medium samples at each of the time periods, and then counted in a Beckman Model 250 scintillation counter (18).
Aliquot8 samples of the explants were extracted with EDTA as described above, and the salt-free extracts and calcified bone explants were analyzed independently for radioactivity. In a second group of experiments, ['*Cjserine (0.5 C/ml, 1.5 ml/tissue culture dish containing four explants) and [3H]threonine (1.5 &/ml) were added together to the explants and simultaneously used to pulse label the explants for 4-8 h. The tissue culture medium was changed at the

290
Synthesis of Bone Phosphoproteins 291 end of the 4-8 h, and the tissue was recultured for 6 days with changes every 2 days, At the end of the tissue culture period, the explants were washed and aliquots were analyzed for the presence of ["C] Ser(P) and [3H]Thr(P) (19)(20). The major portion of the explants were extracted in 0.4 M EDTA, pH 7.5, at 4 "C, and the nondiffusible components were utilized for molecular sieving and ion exchange experiments (vide infra).

Identification of 0-Phosphoserine and 0-Phosphothreonine
The identification of Ser(P), Thr(P), f'H]Ser(P), [14C]Ser(P), and [3H]Thr(P) was carried out by a slight modification of the method previously described (19,20,23). It was found that equilibration and elution of the analytical amino acid analyzer column with 0.2 N citric acid, pH 7.4, greatly improved the resolution and separation of Ser(P) and Thr(P). In brief, the identification of Ser(P) and Thr(P) is carried out as follows. Partial acid hydrolysates (19,20) and chromatographed on a preparative amino acid analyzer column (19,20), and in the case of unlabeled samples, the aliquots eluted in the positions of standard Ser(P) and Thr(P) rechromatographed on a Beckman 121" automatic amino acid analyzer (19,20) to confirm their chromatographic elution positions as Ser(P) and Thr(P) in another chromatographic system. The majority of the sample eluted from the analytical amino acid analyzer in the positions corresponding to authentic Ser(P) and Thr(P) is then hydrolyzed on 6 N HCI at 106 "C for 22 h and rechromatographed on a preparative amino acid analyzer, and the quantitative release of serine and threonine is determined. In the case of the radioactively labeled samples, the fractions eluting from the preparative amino acid analyzer column in the positions of authentic Ser(P) and Thr(P) are not further chromatographed on the 121" amino acid analyzer, but hydrolyzed in 6 N HCI at 106 "C for 22 h and rechromatographed on a preparative amino acid analyzer, and the recovery of [3H]serine, ["Clserine,and [3H]threonine is determined quantitatively. In view of the very close elution positions of many of the phosphorylated amino acids and other phosphorylated components such as phosphoethanolamine, we feel that these additional procedures are necessary for the absolute identification of Ser(P) and Thr(P). This is especially important in those instances where one is attempting to identify for the first time the nature of the phosphorylated components in whole connective tissues or in macromolecules derived from connective tissues, which we have found to always contain small amounts of phosphorylated components other than amino acids, viz. DNA, RNA, phospholipids, etc., which are tightly bound to the protein components (7, 19).

Fractionation and Purification of EDTA-soluble Nondiffusible
Phosphoproteins of the Bone Matrix of Chick Mandibles EDTA-soluble nondiffusible proteins (7, 19, 24) were extracted from 12-day-old chick mandibles, 12-day-old chick mandibles after 6 days of tissue culture, 18-day-old chick embryos in ovo, and 12-dayold chick embryo mandibles labeled with radioactively labeled serine and/or threonine and cultured over a 6-day period. The EDTA-soluble nondiffusible proteins were fractionated as described by Lee and Climcher (7).

RESULTS
A typical experiment demonstrating the mineral and organic constituents of 12-day-old embryonic chick mandible and chick mandible after 6 days of growth in tissue culture and in ouo is shown in Table I. Changes in the major mineral components (Ca, P, Mg) and in the total protein and collagen contents of the bone were also very similar to those of mandibles grown over a similar period of time in ouo and to the changes accompanying mineralization in uiuo in postnatal animals (25). The data demonstrate that active synthesis of bone matrix protein as well as calcification of bone matrix have occurred in the bone cultured in uiuo although not quite to the extent that they have in the mandibles grown in ovo. Freeze-thawed specimens of mandible incubated in tissue culture medium for similar periods of time showed a maximum of 5% increase in weight of the mineral component and organic constituents.
From 96-99% of the Ser(P) and Thr(P) in mandibular bone before and after growth in tissue culture and in ouo was  (19). This establishes that almost all of the Ser(P) and Thr(P) synthesized in organ culture and in ouo is present in EDTAextractable nondiffusible proteins (7-19).
Molecular Sieve and Chromatographic Behavior of EDTAextracted Proteins Containing Ser(P) and Thr(P)-The molecular sieve and ion exchange chromatographic behavior of the EDTA extracts from 12-day-old chick mandibles and from 12-day-old chick mandibles after 6 days of growth in vitro and in ouo was similar and had the same characteristic behavior previously described for the EDTA-extractable phosphoproteins of 10-14-week-old chicken bone (7, 24). For example, the concentration of Ser(P) in the second peak from the Sephacryl S-200 column (Fig. 3) corresponding to the approximately M, = 12,000 purified phosphoprotein of postnatal chicken bone matrix (7) was 12.0 nmol/mg and that of Thr(P) was 2.7 nmol/mg, an approximate 6-fold increase from the Ser(P) concentration in the crude EDTA extract. (No corrections were made for destruction and maximum yield for the phosphoamino acids.) Peak b (Fig. 3) was not purified to homogeneity. The position of this peak corresponded to the M, = 28,000-30,000 fraction isolated from postnatal chick bone (7). The Ser(P) concentration in peak b varied somewhat in different preparations, but in general was from 1.5 to 2.0 times that of peak a (7,26).
The last peak eluted from the Sephadex G-100 column contains the majority of the Gla applied to the column (27). This confirms earlier work that the Gla-containing protein, osteocalcin, is also synthesized in bone (27,28). Further purification of this fraction on DEAE-cellulose ion exchange chromatography resulted in the elution of a peak which behaved as a single band by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Ser(P) was detected in this fraction and in a single component eluted by high precision liquid chromatography from a DEAE-cellulose-purified fraction whose amino acid composition and concentration of Gla are identical with those of authentic osteocalcin (27). This subject is being further investigated at this time.
Incorporation of Radioactively Labeled Serine and Threonine as Radioactively Labeled Ser(P) and Thr(P), Respectiuely, into Bone Matrix Proteins during Tissue Culture of 12day-old Chick Mandibles-In the first group of experiments, [3H]Ser(P) and [3H]Thr(P) were identified in the partial acid hydrolysates of whole chick mandible bone and in the EDTAextractable nondiffusible proteins after the mandibles had been cultured in vitro with [3H]serine and [3H]threonine for 4-8 h. 95-99% of the [3H]Ser(P) and [3H]Thr(P) was extracted in the EDTA solutions as nondialyzable components.

Synthesis of Bone Phosphoproteins
In a typical experiment, approximately 2% of the total counts/ min incorporated into the mandibles was recovered as [3H] Ser(P) (11,500 cpm/mandible) and [3H]Thr(P) (3,208 cpm/ mandible) as determined by preparative amino acid chromatography. When these putative [3H]Ser(P) and [3H]Thr(P) fractions were further hydrolyzed in 6 N HCl at 106 "C for 22 h, 87.2 and 86.6% of the counts/min were recovered as 13H] serine and [3H]threonine, respectively. In comparison, approximately 90% of authentic Ser(P) and Thr(P) was recovered as serine and threonine, respectively, under similar conditions. Table I1 presents the results of a typical group of the second type of experiments carried out which demonstrate that very little of (3H]Ser(P) or [3H]Thr(P) synthesized in the bone is released as such into the tissue culture medium. In contrast and in keeping with previous findings (15), approximately 50% of the hydroxyproline synthesized during the same time period was recovered into the tissue culture medium.   [14C]Ser(P) and ['H]Thr(P) were identified in all of the radioactively labeled peaks, establishing that these fractions contained proteins having Ser(P) and Thr(P) as components. Peak D from the Sephadex G-100 molecular sieving was further chromatographed on DEAE-cellulose. [14C]Ser(P) was identified in the purified fraction, but insufficient amounts of this fraction were available to further pursue this avenue of research at this time.

DISCUSSION
Utilizing 12-day-old chick mandibles grown in vitro in tissue culture for 6 days, the data from the current experiments establish for the first time both by direct chemical analyses and by the incorporation of radioactively labeled serine and threonine as radioactively labeled Ser(P) and Thr(P), respectively, that phosphorylated proteins containing both Ser(P) and Thr(P) and behaving chromatographically similar to those extracted from embryonic and postnatal bone (7) are synthesized in bone tissue. One of the postulated roles of the phosphorylated bone matrix proteins is their participation in the initiation of the formation of a solid phase of calcium-inorganic phosphorus from solution (heterogeneous nucleation) (29, 30). For such components to play the critical role postulated, it is certainly almost mandatory that these proteins be synthesized in bone and be available at the sites in the extracellular organic matrix where calcification occurs. The present experiments, which have established that the characteristic phosphorylated proteins of bone matrix are indeed synthesized in bone, are therefore consistent with the concept that the phosphoproteins of bone matrix play a significant role in the initiation and formation of a solid mineral phase of calcium-inorganic phosphorus (2, 30, 31). The similarities in the basic physical chemical characteristics of the phosphoproteins of all the other vertebrate mineralized tissues thus far studied (5,6,32-35) suggest that the postulated role of phosphoproteins in the mineralization of bone may apply in general sense to all of the calcified vertebrate tissues (2-4).

Synthesis of Bone Phosphoproteins
From the data presented in this report, it is impossible to state which cell is responsible for the synthesis of the bone phosphoproteins. What is not well appreciated is the fact that the vast majority of cells present in bone as a tissue or organ are marrow and connective tissue cells and not the classical bone cells (osteoblast, osteocytes, and osteoclasts). This is especially true in the case of young growing animals. In a series of recent experiments, however, we have been able to obtain radioautographic and biochemical evidence that osteoblasts are the principal cell type synthesizing the phosphorylated extracellular matrix proteins of bone (36,377.