A novel growth-promoting factor derived from fetal bovine cartilage, chondromodulin II. Purification and amino acid sequence.

During endochondral bone formation, cartilage cells show increased matrix synthesis and rapid proliferation. We found that cartilage matrix contains at least two types of heparin binding growth-promoting components. One, with a higher affinity to heparin, was identified as chondromodulin I (Hiraki, Y., Tanaka, H., Inoue, H. , Kondo, J., Kamizono, A., and Suzuki, F. (1991) Biochem. Biophys. Res. Commun. 175, 871-977). In this study, we isolated a novel growth-promoting component, chondromodulin II, which has a lower heparin affinity, from the dissociative extracts of fetal bovine epiphyseal cartilage. Chondromodulin II stimulated the proteoglycan synthesis in rabbit cultured growth plate chondrocytes, an expression of the differentiated phenotype of chondrocytes. It also stimulated DNA synthesis in chondrocytes in both the absence and the presence of fibroblast growth factor-2. The apparent molecular mass of chondromodulin II on SDS-polyacrylamide gel electrophoresis was 16 kDa. Its complete amino acid sequence was determined by overlapping sequences of the peptides released by endopeptidase digestion and CNBr cleavage. Chondromodulin II consists of 133 amino acids (calculated Mr = 14,548). The sequence was unique but homologous to the repeats 1 and 2 of the deduced amino acid sequence of the chicken mim-1 gene, which is specifically transactivated by the v-Myb oncogene product in promyelocytes. We also found a minor component with a higher heparin affinity, chondromodulin III, in cartilage extracts. Chondromodulin III stimulated DNA synthesis in chondrocytes in vitro, and its N-terminal sequence was identical with ribosomal protein L31 lacking the N-terminal three amino acids. These findings suggest that the growth and differentiation of chondrocytes are regulated by multiple components in the cartilage matrix.

During endochondral bone formation, cartilage cells show increased matrix synthesis and rapid proliferation. We found that cartilage matrix contains at least two types of heparin binding growth-promoting components. One, with a higher affinity to heparin, was identified as chondromodulin I (Hiraki, Y., Tanaka, H., Inoue, H., Kondo, J., Kamizono, A., and Suzuki, F. (1991) Biochem. Biophys. Res. Commun. 175, 871-977). In this study, we isolated a novel growth-promoting component, chondromodulin II, which has a lower heparin affinity, from the dissociative extracts of fetal bovine epiphyseal cartilage. Chondromodulin II stimulated the proteoglycan synthesis in rabbit cultured growth plate chondrocytes, an expression of the differentiated phenotype of chondrocytes. It also stimulated DNA synthesis in chondrocytes in both the absence and the presence of fibroblast growth factor-2. The apparent molecular mass of chondromodulin II on SDS-polyacrylamide gel electrophoresis was 16 kDa. Its complete amino acid sequence was determined by overlapping sequences of the peptides released by endopeptidase digestion and CNBr cleavage. Chondromodulin II consists of 133 amino acids (calculated M r ‫؍‬ 14,548). The sequence was unique but homologous to the repeats 1 and 2 of the deduced amino acid sequence of the chicken mim-1 gene, which is specifically transactivated by the v-Myb oncogene product in promyelocytes. We also found a minor component with a higher heparin affinity, chondromodulin III, in cartilage extracts. Chondromodulin III stimulated DNA synthesis in chondrocytes in vitro, and its N-terminal sequence was identical with ribosomal protein L31 lacking the N-terminal three amino acids. These findings suggest that the growth and differentiation of chondrocytes are regulated by multiple components in the cartilage matrix.
The growth of cartilage plays a key role in endochondral bone formation during embryonic development and during the lon-gitudinal growth of bone. Fibroblast growth factor (FGF) 1 exhibits pleiotropic effects depending on the target tissue. Cartilage is a major source of FGF (1), although growth factors of the FGF family are also widely distributed in the body (2). Recent DNA analysis has revealed point mutations in the FGF receptor 3 gene in achondroplasia (3,4). In individuals with this disorder, the growth cartilage of the long bones undergoes minimal proliferation. Thus, FGF signaling is considered to be important for the support of cartilage growth. FGF-2 is the most potent mitogen for chondrocytes (5) and stabilizes the phenotypic expression of these cells (6). As we have reported previously, some proteinaceous components in cartilage synergistically stimulated DNA synthesis and the growth of cultured chondrocytes in vitro as well as stimulating proteoglycan synthesis in chondrocytes (6,7). These findings suggest that FGF, in combination with some unique growth-promoting component(s) in cartilage, may act on chondrocytes.
In the course of the initial screening of these active components in the extracts of fetal bovine cartilage, we found at least two distinct factors in terms of their affinity to heparin (8): one was eluted from a heparin-Sepharose column with buffer containing 0.5 M NaCl, and the other was eluted with buffer containing 1.2 M NaCl. In a previous study, we purified the active component with a higher affinity to heparin from the heparin-bound fraction of cartilage extracts and named it chondromodulin I (ChM-I) (9); ChM-I is a novel 25-kDa glycoprotein that is expressed specifically in cartilage. In the present study, we purified the other component with a lower affinity to heparin to homogeneity and named it chondromodulin II (ChM-II). Chondromodulin II has an apparent molecular mass of 16 kDa on SDS-PAGE, and it stimulated proteoglycan synthesis in the cells. It also stimulated DNA synthesis in rabbit growth plate chondrocytes in vitro in both the absence and the presence of FGF-2. The complete amino acid sequence of ChM-II is reported here.

MATERIALS AND METHODS
Cell Culture and Bioassay-Chondrocytes were isolated from the growth plate cartilage of the ribs of young male New Zealand rabbits as described previously (10). The isolated cells were suspended in a mixture (1/1, v/v) of Ham's F-12 medium and Dulbecco's modified Eagle's medium (FAD medium) containing 10% fetal bovine serum (FBS). Inocula of 0.1 ml of cell suspension (1 ϫ 10 5 cells/ml) were plated in 96-microwell plates. The cells were grown to confluence in the same medium at 37°C under 5% CO 2 in air, and the medium was renewed every other day. Proteoglycan synthesis and DNA synthesis in chondrocytes were assayed as described previously (11). In brief, chondrocytes in the confluent culture were preincubated in FAD medium containing 0.3% FBS for 24 h. The medium was then replaced with FAD medium (0.1 ml) containing 0.3% FBS and test samples. After 3 h of incubation, the culture was labeled with [ 35 S]sulfate (5 Ci/ml) for another 17 h. The culture medium was collected and the cell layer was rinsed with ice-cold phosphate-buffered saline. A portion of the culture medium and phosphate-buffered saline rinse were combined. Proteoglycans were then precipitated by 1% cetylpyridinium chloride, and the radioactivity in the precipitates was measured. Another portion of the culture was labeled with [ 3 H]thymidine (13 Ci/ml) from 22 h after the addition of test samples for the following 4 h. Radioactivity incorporated into DNA was measured in a scintillation spectrometer. Recombinant human FGF-2 (rhFGF-2) was a gift from Takeda Chemical Industries (Osaka, Japan). Bovine bone-derived transforming growth factor-␤ (TGF-␤) and neutralizing antisera against TGF-␤ were generously provided by the Collagen Corporation (Palo Alto, CA).
Immunoblotting-Purified ChM-II and rhFGF-2 were subjected to SDS-PAGE analysis in 15% gel and then transferred electrophoretically to a nitrocellulose membrane (12). The membrane was incubated first in 5% dried milk in 0.1% Tween 20, 20 mM Tris-HCl, and 0.14 M NaCl, pH 7.6 (TTBS) at 37°C for 1 h, and it was then incubated with monoclonal antibody against FGF-2 (bFM-2, 100 ng/ml) (13) dissolved in TTBS containing 1% bovine serum albumin at 4°C for 16 h. The membrane was washed four times for 10 min each time with TTBS and then incubated with horseradish peroxidase-conjugated sheep antimouse IgG (Amersham Corp.) diluted with 1% bovine serum albumin in TTBS (100,000-fold dilution) at room temperature for 1 h. The membrane was then washed extensively with TTBS. Peroxidase activity was visualized by chemiluminescence using the ECL Western blotting detection system (Amersham Corp.), and the membrane was then exposed to Hyperfilm-ECL (Amersham Corp.) for 3 min.
Determination of Amino Acid Sequences-The purified ChM-II was reduced and S-carboxymethylated. The complete amino acid sequence was determined by subfragmentation by CNBr cleavage and enzyme digestion with Achromobacter protease I and trypsin. In the trypsin digestion, reduced S-carboxymethylated ChM-II was acetylated with acetic anhydride and digested with trypsin that had not been treated with N-tosyl-L-phenylalanyl chloromethyl ketone. The peptides obtained were separated by reverse-phase HPLC, and the sequence of each peptide was determined with a gas phase sequencer (Applied Biosystems 470A, ABI). The carboxyl-terminal sequence was determined as follows: reduced S-carboxymethylated ChM-II was digested with Achromobacter protease I in 50 mM N-ethylmorpholine-acetic acid buffer (pH 8.7) containing 4 M urea. The C-terminal peptide was isolated from the digests using p-phenylene diisothiocyanate-coated polymer (14). The C-terminal sequence of the isolated peptide was then determined by the thiohydantoin formation method, with trimethylsilylisothiocyanate as a coupling reagent (15).

RESULTS AND DISCUSSION
First, we attempted to confirm the bioactivity in the guanidinium extracts of cartilage from which contaminating FGF had been separated by heparin affinity chromatography. Fetal bovine epiphyseal cartilage was homogenized and extracted with 1 M guanidinium chloride. The resultant extract was fractionated with 45-65% acetone. Precipitates were solubilized in 4 M guanidinium chloride and fractionated by ultrafiltration. Materials with a molecular mass of 10 -50 kDa were concentrated (9). A portion of this 10 -50-kDa extract was fractionated into four on a heparin affinity column by batch elution. As shown in Table I, the 10 -50-kDa extract strongly stimulated [ 3 H]thymidine incorporation in the growth plate chondrocytes, largely due to contaminating FGF-like activity that was tightly bound to heparin (1). FGF-2 is the most potent mitogen for chondrocytes (7,16). FGF-2 stimulated DNA synthesis in quiescent primary chondrocytes in a dose-dependent manner, and treatment with an optimal dose of FGF-2 (0.4 ng/ml) resulted in a 5-15-fold increase of [ 3 H]thymidine incorporation in the cells, depending on the batch of isolated cells (Table I and Fig. 1B,  inset). Interestingly, the extract enhanced the effect of an optimal dose of FGF-2 that had produced maximal stimulation of [ 3 H]thymidine incorporation. This synergistic activity was recovered in the 0.5-1.2 M NaCl eluate (heparin 1.2 M) from the heparin affinity column, and this fraction alone stimulated [ 3 H]thymidine incorporation in the cells to 2 times the basal level. The 10 -50-kDa extract strongly stimulated proteoglycan synthesis in chondrocytes, an expression of the differentiated phenotype (Table I and Fig. 1A). The half-maximal dose was about 40 g/ml. This activity was also recovered in the heparin The following fractionated cartilage extract materials were added to confluent culture of rabbit growth plate chondrocytes in the presence or absence of rhFGF-2 (0.4 ng/ml): 1 M guanidine extract containing materials of 10 -50 kDa (10 -50 kDa extract, 200 g/ml) and the fractions obtained by heparin affinity chromatography of 10 -50 kDa extract: unbound-fraction (heparin pass, 100 g/ml), the first protein peak eluted with 0.5 M NaCl (heparin 0.5 M I, 6.2 g/ml), the second protein peak eluted with 0.5 M NaCl (heparin 0.5 M II, 2.4 g/ml), and the protein peak eluted with 1.2 M NaCl (heparin 1.2 M, 6.2 g/ml). Values are means Ϯ S.D. for triplicate samples.  (Table I and Fig. 1A). In contrast, FGF-2 had no effect on proteoglycan synthesis in the cells. As we have reported previously, ChM-I accounts for these bioactivities of the heparin 1.2 M fraction (9). However, residual proteoglycan synthesis-stimulating activ-ity was also recovered in the fraction corresponding to the second protein peak eluted by 0.5 M NaCl (heparin 0.5 M II) as shown in Table I. The presence of TGF-␤ in cartilage has been documented (17,18), although there is no definitive evidence that it is activated in cartilage under physiological conditions. Proteoglycan synthesis in chondrocytes was markedly stimulated by TGF-␤ (19). Neutralizing antisera against bovine bone-derived TGF-␤ 1 and -␤ 2 failed to inhibit the proteoglycan synthesis-stimulating activity of heparin 0.5 M II, while these sera completely blocked the actions of the corresponding TGF-␤ subtypes (Table II). These findings suggested that a novel growth factor was recovered in this fraction. Therefore, we decided to purify the novel factor on the basis of its proteoglycan synthesis-stimulating activity in the 10 -50-kDa extract.
To facilitate purification of the activity, the 10 -50-kDa extract (141 mg) was subfractionated on Sephacryl S-200 in the presence of 4 M guanidinium chloride and 1 M NaCl. The first protein peak was collected (17.3 mg), since all bioactivity of the extract was recovered in this fraction (9). The heparin-bound materials eluted by 0.5 M NaCl (heparin 0.5 M; 11.4 mg) and then the heparin-bound materials eluted by 1.2 M NaCl (heparin 1.2 M; 2.66 mg) were collected as described under "Materials and Methods." The heparin 0.5 M fraction showed a dose-response curve with a slope shallower than that of the 10 -50-kDa extract, while the heparin 1.2 M fraction stimulated proteoglycan synthesis with a dose-response profile parallel to that of the 10 -50-kDa extract (Fig. 1A). The half-maximal dose of the heparin 0.5 M fraction was about 4 g/ml. Finally, the heparin 0.5 M fraction was subjected to reverse-phase HPLC. The elution profile is shown in Fig. 2A. Bioassay defined the presence of a novel component (ChM-II; 60 g), which had a longer retention time than ChM-I. Chondromodulin II purified from the heparin 0.5 M fraction had an apparent molecular mass of 16 kDa (Fig. 3A). The purified ChM-II stimulated proteoglycan synthesis in the growth plate chondrocytes in culture in a dose-dependent manner (Fig. 1A). The half-maximal dose of ChM-II was about 75 ng/ml. Insulin-like growth factor-I or -II also stimulates proteoglycan synthesis in rabbit growth plate chondrocytes, the half-maximal dose being about 40 ng/ml. 2 However, SDS-PAGE analysis (Fig. 3A) indicated that the purified ChM-II preparation did not contain insulin-like growth factor-I or -II to the extent that could account for its proteoglycan synthesisstimulating activity. ChM-II also stimulated [ 3 H]thymidine incorporation in chondrocytes (Fig. 1B). As shown in the inset of  contamination of FGF-2 by immunoblotting with a monoclonal antibody (bFM-2) against bovine brain-derived FGF-2 (12,13). As shown in Fig. 3B, immunoblotting clearly visualized the 10 -50-ng rhFGF-2 used as the reference substance. In contrast, no immunoreactive material was detected in the ChM-II preparation (2.5 g), suggesting that there was not more than a nanogram order of contamination of FGF-2 in the preparation. Further, the purified ChM-II enhanced [ 3 H]thymidine incorporation in the presence of an optimal dose (0.4 ng/ml) of FGF-2 (Fig. 1B). This enhancing effect cannot be accounted for by contamination with FGF-2, since additional FGF-2 did not stimulate [ 3 H]thymidine incorporation any further. The purified ChM-II preparation had no effect on [ 3 H]thymidine incorporation in cultured fibroblasts (data not shown). These results suggest that weak growth-promoting activity is also associated with ChM-II, although this agent plays a role primarily in the phenotypic expression of chondrocytes. As previously reported (9), ChM-I was the major protein in the heparin 1.2 M fraction associated with growth-promoting activity for chondrocytes (Fig. 2B). N-terminal amino acid sequencing indicated that ChM-II was also present in the heparin 1.2 M fraction as one of the two minor proteins associated with bioactivity. The other minor active component (17 kDa on SDS-PAGE) was found between ChM-I and ChM-II, and this was termed chondromodulin III (ChM-III) (Fig. 2B). The Nterminal amino acid sequence of ChM-III was determined to be Ala-Lys-Lys-Gly-Gly-Glu-Lys-Lys-Lys-Gly-Arg, which is identical to the N-terminal sequence of ribosomal protein L31 (20). Since ribosomal protein L31 has an additional three amino acids (Met-Ala-Pro), we assumed ChM-III to be a ribosomal protein with an N-terminal truncation. Thus, ChM-III seems irrelevant as a growth factor. It is possible that the extraordinarily high pI of the ribosomal protein may lead to erroneous stimulation of chondrocytes in culture. No ChM-III was found in the heparin 0.5 M fraction ( Fig. 2A). Interestingly, Chester and co-workers (21) reported overexpression of human ribosomal protein L31 mRNA in familial adenomatous polyposis and colorectal tumors. The gene is also expressed at abnormally high levels in various hematopoietic malignant tumor cells, but its expression was markedly down-regulated upon terminal differentiation of immature leukemic cell lines such as HL-60 promyelocytic leukemia cells and K562 erythroleukemia cells in vitro (22). These findings indicate that the ribosomal protein L31 is associated with growth regulation and differentiation. Recently, Fujita and co-workers (23) reported that the heparinbinding ribosomal protein L22 was copurified with FGF-1 from bovine brain. Thus, it is possible that these heparin-binding ribosomal proteins may have functional relevance in growth regulation through a potential interaction with FGF signaling.
To determine the amino acid sequence, ChM-II was first reduced and protected by S-carboxymethylation. First, the Nterminal amino acid sequence of 37 residues was determined (Fig. 4). Then lysylendopeptidase digestion yielded several peptide fragments that were separated by reverse-phase HPLC on a Bakerbond C 8 column. The amino acid sequences of the peptides were determined, and the C-terminal amino acid sequence was determined from one of the peptides isolated by the thiohydantoin formation method (15). Cyanogen bromide cleavage yielded two internal amino acid sequences of ChM-II. Finally, ChM-II was acetylated to protect lysine residues. The resultant acetylated ChM-II was digested with trypsin. Overlapping sequences of these fragments defined the complete primary amino acid sequence of ChM-II, which had one tryptophan near the N terminus, had six cysteine residues, and was rich in lysine residues (Fig. 4).
In a previous study, we reported that cartilage-derived factor, which has a molecular mass of 16 kDa, stimulated proteoglycan synthesis in rabbit growth plate chondrocytes in culture (8). The N-terminal amino acid sequence determined at that time was Gly-Pro-Trp-Ala-Ile-Ile-Xaa-Ala-Gly-Lys-Ser-Ser-Asn-Glu-Ile-Arg-Thr. 2 The seventh residue was assumed to be cysteine that had been oxidized during purification. The Nterminal sequence of ChM-II completely matched the sequence above, suggesting that ChM-II was identical to cartilage-derived factor. In the study above (8), we presented the amino acid composition of cartilage-derived factor assuming its molecular mass to be 16 kDa. The amino acid composition of ChM-II was compatible with that of cartilage-derived factor.
Screening of data bases indicated that ChM-II had a novel amino acid sequence. We were unable to identify any previously characterized sequence motifs. However, ChM-II had a similarity to the repeats 1 and 2 of the mim-1 gene product (24) (Fig. 5). The similarity spread over the entire region of the repeats 1 and 2, each of which contains seven cysteine residues. The positions of the six cysteine residues in ChM-II were all conserved in these mim-1 repeats. The promyelocyte-specific mim-1 gene was first identified by differential hybridization as a v-myb regulated gene (24). The mim-1 gene is directly transcribed by c-Myb transcription factor in cooperation with other myeloid-specific transcription factors (25,26). The mim-1 mRNA encodes a secretable 35-kDa protein that is specifically expressed in normal bone marrow promyelocytes. The MIM-1 protein is stored in granules of promyelocytes in the 35-kDa form as well as in the cleaved 15-18-kDa forms. In situ hybridization and immunolocalization of mim-1 expression have indicated that MIM-1 protein may participate in the regulation of granulopoietic differentiation as a local growth factor (24,27). However, no biological function of MIM-1 protein has yet been defined.
Although ChM-II has a sequence showing 57% homology to the mim-1 repeats, we assume that ChM-II is a product of a gene distinct from mim-1. It seems unlikely that ChM-II is derived from a larger protein with internal repeats, such as MIM-1, since we were unable to find a peptide with a similar amino acid sequence in our preparation of cartilage extract. Our preliminary sequence data for ChM-II cDNA indicated that the CUA codon corresponding to the C-terminal leucine residue was directly followed by a TAG stop codon (data not shown), while the putative cleavage products of MIM-1 have an extension of 11 amino acids at their C terminus. However, it is possible that there is some evolutionary relationship between the mim-1 gene and the gene encoding ChM-II. Although ChM-II was purified on the basis of its action on the growth and phenotypic expression of chondrocytes, its sequence similarity to MIM-1 implies that ChM-II may act on the growth and differentiation of other cell types, such as the cells involved in hematopoietic differentiation (28,29). Further studies along these lines are in progress.  We suggest that subscribers photocopy these corrections and insert the photocopies at the appropriate places where the article to be corrected originally appeared. Authors are urged to introduce these corrections into any reprints they distribute. Secondary (abstract) services are urged to carry notice of these corrections as prominently as they carried the original abstracts.