J Bone Miner Metab (2001) 19:345351

Original articles

Springer-Verlag 2001

Expression of mouse osteocalcin transcripts, OG1 and OG2, is differently regulated in bone tissues and osteoblast cultures
Takeshi Yanai1,2, Takenobu Katagiri1, Shuichi Akiyama1, Mana Imada1, Takeyoshi Yamashita3, Hiroshige Chiba2, Naoyuki Takahashi1, and Tatsuo Suda1
1 2

Department of Biochemistry, School of Dentistry, Showa University, 1-5-8 Hatanodai, Shinagawa-ku, Tokyo 142-8555, Japan Department of Oral and Maxillofacial Surgery, Tokyo Medical University, Tokyo, Japan 3 Pharmaceutical Research Laboratory, Kirin Brewery, Gunma, Japan

Abstract Osteocalcin is a noncollagenous protein that is abundant in mineralized bone matrix. Mice have a gene cluster of osteocalcin that consists of OG1, OG2, and ORG. We established a new method to directly analyze the expression levels of OG1, OG2, and ORG mRNAs relative to total osteocalcin mRNA. They were amplied as 371-bp fragments by reverse transcription-polymerase chain reaction (RT-PCR) at the same time using common primers, digested with ApaLI, and separated in a polyacrylamide gel. ApaLI digestion did not affect the mobility of the OG1-derived 371-bp fragment, whereas both 371-bp fragments, derived from OG2 and ORG, were digested into 350 bp. Total RNA prepared from mouse bone was then subjected to RT-PCR followed by ApaLI digestion. OG1 and OG2 mRNAs were found to be expressed at ratios of 80%86% and 14%20%, respectively, to the total osteocalcin mRNA in mouse bone. The ratios were almost constant in various bones in vivo, independent of the animals genetic background, age, or gender, or different parts of bone. RT-PCR using specic primers revealed that mouse bone tissues strongly expressed osteocalcin mRNA derived from OG1 and OG2, but not ORG. In contrast, cells cultured in vitro showed different expression ratios of osteocalcin mRNA: 53%65% for OG1 and 35%47% for OG2 to the total osteocalcin mRNA in the osteoblast cell line and primary osteoblasts in culture even though they formed many mineralized bone nodules. Similar results were obtained in both KS483 osteoblasts and C2C12 myoblasts, when they were cultured with bone morphogenetic protein-2 (BMP-2) to induce osteocalcin mRNA. Taken together, these ndings indicate that OG1 is the predominant transcript among the three osteocalcin genes in mouse bone in vivo. It is also suggested that the expression of OG1 and OG2 is regulated differently in bone tissues and osteoblast cultures. Key words osteocalcin gene expression osteoblasts mice

Introduction Bone essentially consists of hydroxyapatite crystals and matrix proteins secreted by osteoblasts. Type I collagen and osteocalcin are the most abundant collagen and noncollagenous protein, respectively, in bone matrix proteins [1,2]. Osteocalcin is synthesized and secreted by mature osteoblasts and odontoblasts. Three glutamic acid residues of the secreted osteocalcin have been found to be -carboxylated by vitamin K-dependent -carboxylase. These -carboxyglutamic acid residues (Gla) function in binding to hydroxyapatite crystals with a high afnity [3]. Osteoid, which is a nonmineralized bone matrix, gradually mineralizes with hydroxyapatite crystals in developing animals. Circulating osteocalcin is found in both fetal and adult animals, but it accumulates in bone only after matrix mineralization [4]. It is known that the administration of warfarin, an antagonist of vitamin K, inhibits the activity of vitamin K-dependent -carboxylase, leading to a loss of Gla residues in osteocalcin in vivo and in vitro. In warfarinadministered rats, osteocalcin is continuously circulated in serum, but not accumulated in bone matrix [5]. Due to its binding activity to hydroxyapatite crystals, osteocalcin has been thought for a long time to be a regulator of mineralization of bone matrix. Recently, three copies of osteocalcin genes, called OG1/OC-A, OG2/OC-B, and ORG/OC-X, were identied as a gene cluster within a 23-kb span of the mouse genome [6,7], and their genomic and deduced aminoacid sequences showed a high similarity. Recently, mice decient for both the OG1 and OG2 alleles were generated by gene targeting in embryonic stem cells [8]. These OG1/OG2-decient mice showed increased bone formation without abnormalities in bone mineralization. The phenotype of OG1/OG2-decient mice suggests that osteocalcin functions as an inhibitor of bone formation, rather than as a regulator of bone mineralization in vivo.

Offprint requests to: T. Katagiri Received: February 2, 2001 / Accepted: June 4, 2001

346

T. Yanai et al.: Expression of osteocalcin transcripts in mice

Because osteocalcin is secreted specically by boneforming mature osteoblasts, the expression of osteocalcin is thought to be a marker for bone formation and osteoblast differentiation. Of the three osteocalcin genes in mice, OG1 and OG2 transcripts were specically expressed in bone [7]. In contrast, the ORG transcript was mainly expressed in gonadal tissues, and at lower levels in several other tissues, in mice [7,9]. To date, however, the expression levels of the three osteocalcin genes, especially OG1 and OG2, in bone are not clear. In the present study, we established a new method, using a reverse transcription-polymerase chain reaction (RT-PCR), followed by restriction enzyme digestion, to directly analyze the amounts of OG1, OG2, and ORG transcripts relative to the total amount of osteocalcin mRNA. Using this method, we found that OG1 was the major transcript, with expression levels higher than 80% of the total osteocalcin mRNA in mouse bone tissues in vivo. However, OG1 and OG2 were expressed at roughly equal levels in osteoblastic cells tested in culture. These ndings suggest that the regulation of the expression of mouse osteocalcin mRNA is different in bone tissues in vivo and in osteoblast cultures in vitro. Materials and methods RNA preparation from animals, and cultured cells Mice were obtained from Sankyo Laboratories Animal Center (Tokyo, Japan). Tibia were obtained from 7week-old male mice of different genetic backgrounds (ddY, C57BL/6, BALB/c, and C3H/He). Sevenweek-old female ddY mice were also used as a source of tibia. Calvaria were obtained from newborn ddY mice. Kidney, muscle, and liver were obtained from 7week-old male ddY mice. The tissues obtained were cleaned, washed and minced in ice-cold phosphatebuffered saline, and then homogenized in Trizol (GIBCO BRL, Grand Island, NY, USA) to extract total RNA. A mouse osteoblastic cell line, KS483 [10], and a mouse myoblast cell line, C2C12, were cultured in minimal essential medium (-MEM) containing 10% fetal bovine serum (FBS; Iwaki, Chiba, Japan) and Dulbeccos modied Eagles medium (DMEM) containing 15% FBS, respectively. To examine the effect of bone morphogenetic protein-2 (BMP-2) on the expression of osteocalcin mRNA, these cell lines were cultured with or without 300 ng/ml of recombinant human BMP-2 (kindly supplied by Yamanouchi Pharmaceuticals, Tokyo, Japan). Primary osteoblasts were prepared from newborn mouse calvaria by sequential collagenase-dispase digestion, as described previously [11]. The cells released as fractions IIIV were pooled and

inoculated with -MEM containing 10% FBS without subcultures. Total cellular RNA was extracted from these cultured cells, using Trizol (GIBCO BRL).

Quantication of the expression levels of OG1, OG2, and ORG relative to the total amount of osteocalcin mRNA Twenty micrograms of total RNA was treated with RNase-free DNase I (Takara Biochemicals, Shiga, Japan), then extracted with Trizol again. First-strand cDNA was synthesized from 10 g of DNase I-treated RNA, using Superscript II (GIBCO BRL) and random primers in a nal volume of 20 l. One microliter of the reaction mixture was subjected to PCR amplication with EX Taq DNA polymerase (Takara Biochemicals) in the presence of [-33P]dCTP, in a nal volume of 50 l. Primer sets used in PCR for total osteocalcin, ORG, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were: total osteocalcin, 5-CAAGTCCC ACACAGCAGCTT-3 (primer A) and 5AAAGCCGAGCTGCCAGAGTT-3 (primer B) [7]; ORG, 5-AGCATTCTGGCCTTCTGGTG-3 (primer C) and primer B; and GAPDH, 5-TGAAGGTCGG TGTGAACGGATTTGGC-3 and 5-CATGTAGG CCATGAGGTCCACCAC-3, respectively. PCR conditions were 1 min at 94 C, 1 min at 60 C, and 1 min at 72 C for 25 cycles (total osteocalcin and GAPDH), and the same conditions for 30 cycles for ORG. The products amplied in PCR were resolved by electrophoresis in a 5% polyacrylamide gel. The PCR products for total osteocalcin were precipitated in ethanol with Ethachinmate (Nippon Gene, Tokyo, Japan), and dissolved in a digestion buffer for ApaLI. The mixture was separated into two tubes and then incubated for at least 2 h at 37 C with and without ApaLI (Takara Biochemicals) before electrophoresis. The gels were dried and exposed to X-ray lms at 80 C. The radioactivity of each band on the membrane was quantied with a Bioimage Analyzer BAS2000 (Fuji Film, Tokyo, Japan). To analyze the DNA sequence of the PCR products, the 371-bp fragment was amplied from bone and kidney cDNAs of 7-week-old ddY mice by RT-PCR, as described above, and subcloned in pGEM-T Easy (Promega, Madison, WI, USA). The sequence of the insert DNA was analyzed with a GeneRapid DNA sequencer (Amersham Pharmacia Biotech, Buckinghamshire, UK). Plasmids carrying OG1, OG2, and ORG cDNAs (pGEM-T/OG1, pGEM-T/OG2, and pGEM-T/ ORG, respectively) were used as positive controls for the total osteocalcin in PCR, as described above.

T. Yanai et al.: Expression of osteocalcin transcripts in mice

347

Northern blot analysis Northern blot analysis was performed as described previously [12]. In brief, 20 g of total RNA was resolved by electrophoresis in a 1.2% agarose-formaldehyde gel, and transferred onto a Hybond-N membrane (Amersham International, Amersham, UK). The membrane was sequentially hybridized with [32P]-labeled cDNA probes for mouse OG1 as osteocalcin and GAPDH. In some experiments, additional OG2 and ORG probes were used. Probes for OG1, OG2, and ORG were obtained by EcoRI digestion from pG EM-T/OG1, pGEM-T/OG2, and pGEM-T/ORG, respectively, as described above.

Establishment of a method to analyze the expression levels of OG1, OG2, and ORG in the total amount of osteocalcin mRNA Next, we examined the expression levels of OG1, OG2, and ORG in mouse tissues by Northern blotting. As shown in Fig. 2A, all of the probes for OG1, OG2, and ORG hybridized with ~0.6 kb mRNA in bone, but no signals were detected in the kidney, skeletal muscle, and liver. We then attempted PCR for total osteocalcin, using common primers A and B, which amplify OG1, OG2, and ORG equally, followed by ApaLI digestion (Fig. 2B). Plasmids carrying OG1, OG2 and ORG cDNAs were used as the templates. We obtained a single 371-bp band from each of the plasmids (Fig. 2C). No fragments were observed from an empty vector (data not shown). When the PCR products were treated with ApaLI, the 371-bp fragments derived from OG2 and ORG were completely digested into a 350-bp length (Fig. 2C). In contrast, ApaLI digestion did not alter the 371-bp fragment amplied from OG1. These ndings indicated that an OG1-derived product can be separated from the mixture of OG1, OG2, and ORG-derived products by detecting the 371-bp fragment after ApaLI digestion. This method was employed for measuring the expression level of OG1 relative to total osteocalcin mRNA in mouse tissues. As shown in Fig. 2D, the 371-bp fragment was amplied from bone RNA on RT-PCR, using common primers A and B. A portion of this fragment was digested into a 350-bp fragment by ApaLI digestion. Quantication of the radioactivity in these two fragments showed that the ratios of the 371-bp and 350bp fragments to the total were approximately 80% and 20%, respectively. A longer period of ApaLI digestion (4 h) did not affect the relative amounts of the two

Results Cloning of mouse OG1, OG2 and ORG cDNAs First, we obtained cDNA clones encoding OG1, OG2, and ORG by RT-PCR from tibia and kidney of ddY mice, then we subcloned and sequenced them in plasmid vectors (Fig. 1). The DNA sequence showed that, of 35 clones obtained from tibia, 27 were for OG1, 8 were for OG2, and none for ORG. In contrast, 5 of 5 clones obtained from the kidney were for ORG. There were one or two nucleotide substitutions in osteocalcin cDNAs: OG1 (314C to T), OG2 (266C to G and 297G to A), and ORG (118C to G) from the sequences derived from 129Sv mice, which have been deposited on a database. The substitutions in OG2 and ORG caused amino-acid changes, from Asp75 to Glu and Ala86 to Thr, and Thr26 to Ser, respectively. The nucleotide substitution in ORG destroyed a Cfr10 I restriction site.

Fig. 1. Alignment of OG1, OG2, and ORG cDNAs. Total RNA was prepared from tibia and kidney of 7-week-old male ddY mice, and subjected to reverse transcription-polymerase chain reaction (RT-PCR), using primers A and B. A 371-bp fragment was subcloned then sequenced in plasmid vectors.

OG1 and OG2 cDNAs were obtained from tibia and ORG cDNA was obtained from the kidney. Identical bases in OG1, OG2, and ORG are boxed. Primers A and B are shown by arrows. ApaLI sites (gtgcac) in OG2 and ORG are underlined

348

T. Yanai et al.: Expression of osteocalcin transcripts in mice

Fig. 2. Establishment of a method to directly compare the expression ratios of OG1, OG2, and ORG transcripts to the total amount of osteocalcin mRNA. A Northern blot analysis for OG1, OG2, and ORG. Total RNA was prepared from tibia (T), kidney (K), skeletal muscle (M), and liver (L) of 7week-old male ddY mice. Twenty micrograms of total RNA was resolved by electrophoresis in a 1.2% agarose gel, and transferred onto a membrane. The membrane was hybridized with cDNA probes for OG1, OG2, ORG, and GAPDH. B Schematic structure of OG1, OG2, and ORG cDNAs, 371bp in length, amplied on RT-PCR. Coding regions are shown by solid boxes. Primers A, B, and C used for RT-PCR are shown by arrows. Positions of the ApaLI site in OG2 and ORG are shown by open triangles. C Characterization of a 371-bp fragment amplied separately from OG1, OG2, and

ORG cDNAs. PCR was performed with primers A and B, using pGEM-T/OG1, pGEM-T/OG2, and pGEM-T/ORG as templates. The PCR products were digested with (w/) and without (w/o) ApaLI before electrophoresis in a 5% polyacrylamide gel. Note that both OG2 and ORG were digested into 350-bp fragments by ApaLI, but OG1 was not. D Analysis of the expression levels of osteocalcin transcripts in total osteocalcin mRNA. Total RNA was prepared from tibia (T), kidney (K), skeletal muscle (M), and liver (L) of 7-week-old male ddY mice. DNase I-treated total RNA was transcribed into cDNA, using random primers, and subjected to PCR, using primers A and B for osteocalcin, primers C and B for ORG, and specic primers for GAPDH. The PCR products for osteocalcin were digested with (w/) and without (w/o) ApaLI before electrophoresis in a 5% polyacrylamide gel

fragments (data not shown). We specically amplied ORG on RT-PCR, using a specic primer C, and a common primer B (Fig. 2B). A very faint band, 546 bp in length, was observed in bone and kidney, but not in skeletal muscle and liver under the present PCR conditions (Fig. 2D). In agreement with the ndings in a previous study [7], the expression level of the 546-bp fragment in bone was lower than that in the kidney. OG1 is the major transcript in total osteocalcin mRNA in mouse bone in vivo We further examined the expression levels of OG1, OG2, and ORG in tibia prepared from 7-week-old male

mice of various genetic backgrounds, including C57BL/ 6, BALB/c, C3H/He, and ddY. Osteocalcin mRNA was detected in all strains examined by Northern blotting and the amplication of the 371-bp fragment on RTPCR (Fig. 3A). The 350-bp fragment appeared after ApaLI digestion in all samples. In these mice, the ratios of the 371-bp, and 350-bp fragments to the total amount of osteocalcin mRNA were 80.0%86.2% and 13.8% 20.0%, respectively (Fig. 3B). Similar ndings were observed not only in tibia prepared from 7-week-old female ddY mice but also in calvaria prepared from newborn ddY mice (Fig. 3A,B). ORG-specic PCR amplied an extremely faint 546-bp fragment in all of the samples.

T. Yanai et al.: Expression of osteocalcin transcripts in mice

349

Fig. 3. OG1 is the major transcript in the total amount of osteocalcin mRNA in mouse bone in vivo. A Tibias (T) were obtained from 7-week-old male (M) and female (F) mice of various genetic backgrounds, including C57BL/6, BALB/c, C3H/He, and ddY. Newborn ddY mice (NB) were examined for the preparation calvaria (C). Twenty micrograms of total RNA extracted from bones was studied in Northern blot analysis (two upper panels). DNase I-treated total RNA was subjected to RT-PCR for total osteocalcin, ORG, and

GAPDH (four lower panels). The PCR products for total osteocalcin were digested with (w/) and without (w/o) ApaLI before electrophoresis in a 5% polyacrylamide gel. B Quantication of two fragments, of 371 bp and 350 bp, in A. The radioactivity of the two fragments, of 371 bp (open bars) and 350 bp (closed bars) after ApaLI digestion was determined with a Bioimage Analyzer (BAS2000; Fuji Film, Tokyo, Japan)

OG1 and OG2 are expressed equally in osteoblast cultures in vitro Finally, we examined the expression levels of OG1, OG2, and ORG in osteoblast cultures. Primary osteoblasts were prepared from the calvaria of newborn ddY mice, which expressed predominantly OG1 mRNA in vivo (Fig. 3). The expression of osteocalcin mRNA by primary osteoblasts began before day 4, reached a peak at day 6, and gradually decreased thereafter (Fig. 4A). RT-PCR followed by ApaLI digestion showed that the ratios of the 371-bp and 350-bp fragments to the total amount osteocalcin of mRNA were 54.2%66.5% and 34.5%45.8%, respectively (Fig. 4A,B). We also used the osteoblastic cell line, KS483 cells, which expresses osteocalcin mRNA and forms many mineralized bone nodules in vitro [10]. In KS483 cells, the ratio of the 371bp fragment to the total osteocalcin mRNA was 53.2% 62.4% and that of the 350-bp fragment was 37.6% 46.8% (Fig. 4A,B). Treatment with BMP-2 for 3 days

induced the expression of amount of osteocalcin mRNA in both KS483 osteoblasts and C2C12 myoblasts (Fig. 4A). In the presence of BMP-2, the ratios of the 371-bp fragment to the total amount of osteocalcin mRNA were 46.5% in KS483 cells and 57.5% in C2C12 cells (Fig. 4A,B).

Discussion In the present study, we established a new method to directly examine the expression levels of OG1, OG2, and ORG mRNAs in mice. We were able to amplify these cDNAs as 371-bp fragments on RT-PCR, using common primers A and B [7]. When we treated the 371-bp fragments with ApaLI, OG2 and ORG were digested into 350-bp fragments, but OG1 was not. This method allowed us to separate OG1 and OG2/ORG in the 371-bp fragments mixture. The 350-bp fragment

350

T. Yanai et al.: Expression of osteocalcin transcripts in mice

Fig. 4. OG1 and OG2 are expressed equally in osteoblastic cell cultures in vitro. A Primary osteoblasts (POB), KS483 osteoblasts, and C2C12 myoblasts were cultured with or without 300 ng/ml of bone morphogenetic protein (BMP)-2 for the indicated periods. Total RNA was extracted from these cells at each time point. Twenty micrograms of total RNA was examined in Northern blot analysis (two upper panels). DNase I-treated total RNA was subjected to RT-PCR for total

osteocalcin, ORG, and GAPDH (four lower panels). The PCR products for total osteocalcin were digested with (w/) and without (w/o) ApaLI before electrophoresis in a 5% polyacrylamide gel. B Quantication of two fragments, of 371 bp and 350 bp, in A. Radioactivity of the two fragments, of 371 bp (open bars) and 350 bp (closed bars) after ApaLI digestion was determined using the Bioimage Analyzer (BAS2000; Fuji Film). N.D., Not determined

must be derived mainly from OG2 in bone, because the expression level of ORG was essentially zero when evaluated by the ORG-specic RT-PCR. This was further conrmed by the direct DNA sequencing of the 371-bp fragments. Taken together, these ndings suggest that the new method established in this study is useful for analyzing, simply and directly, the ratios of OG1 and OG2/ORG transcripts relative to the total amount of osteocalcin mRNA. Using this method, we examined the relative expression levels of OG1 and OG2 in mice in vivo and in vitro. In bone tissues, the OG1 transcript was always detected at a level of greater than 80% of the total osteocalcin mRNA. Similar ndings were observed in all mice examined regardless of the various genetic backgrounds, age, and gender, and regardless of different parts of bone. These ndings indicated that OG1 was the major transcript among the three osteocalcin genes in mouse bone tissues in vivo. There are three amino acid substitutions in the signal peptides of OG1 and OG2. Because the signal peptide is essential for protein secretion, it is possible that the difference affects the efciency of OG1 and OG2 secretion by osteoblasts. The development of antibodies that can recognize the difference between OG1 and OG2 proteins would help to clarify this question.

In contrast to the in vivo situation, osteoblastic cells expressed roughly equal amounts of OG1 and OG2 transcripts in culture. These ndings were observed not only in a clonal osteoblastic cell line (KS483 cells) but also in primary osteoblasts, which predominantly expressed the OG1 transcript in vivo. We have reported that BMP-2 induces osteoblast differentiation of the mouse myoblast cell line, C2C12 [12]. In C2C12, as well as KS483 cells, osteocalcin mRNA induced by BMP-2 consisted of roughly equal amounts of OG1 and OG2 mRNAs. The reason for the different expression levels of OG1 and OG2 in vivo and in vitro is not clear at present, but there are several possible explanations for this. There may be two types of osteocalcin-expressing cells in bone in vivo. One type of cell expresses almost equal amounts of OG1 and OG2 mRNA, as seen in cultured cells. Another type of cell predominantly expresses OG1 mRNA in vivo. The latter may have poor proliferating capacity in vitro. Thiede et al. [13] reported that megakaryocytes in bone marrow and platelets in peripheral blood expressed osteocalcin mRNA, in rats and humans. These cells present in mouse bone tissues may affect the expression levels of OG1 and OG2 in vivo. It is necessary to examine whether these cells also express osteocalcin mRNA in mice in vivo and in vitro. It is also possible that a factor(s) that stimulates

T. Yanai et al.: Expression of osteocalcin transcripts in mice

351
2. Desbois C, Karsenty G (1995) Osteocalcin cluster: implications for functional studies. J Cell Biochem 57:379383 3. Poser JW, Price PA (1979) A method for decarboxylation of carboxyglutamic acid in proteins. J Biol Chem 254:431436 4. Price PA, Lothringer JW, Nishimoto SK (1980) Absence of the vitamin K-dependent bone protein in fetal rat mineral: evidence for another -carboxyglutamic acid containing component in bone. J Biol Chem 255:29382942 5. Price PA, Williamson MK (1981) Effects of warfarin on bone. J Biol Chem 256:1275412759 6. Rahman S, Oberdorf A, Montecino M, Tanhauser SM, Lian JB, Stein GS, Laipis PJ, Stein JL (1993) Multiple copies of the bone-specic osteocalcin gene in mouse and rat. Endocrinology 133:30503053 7. Desbois C, Hogue DA, Karsenty G (1994) The mouse osteocalcin gene cluster contains three genes with two separate spatial and temporal patterns of expression. J Biol Chem 269:11831190 8. Ducy P, Desbois C, Boyce B, Pinero G, Story B, Dunstan C, Smith E, Bonadio J, Goldstein S, Gundberg C, Bradley A, Karsenty G (1996) Increased bone formation in osteocalcindecient mice. Nature 382:448452 9. Sato M, Tada N (1995) Preferential expression of osteocalcinrelated protein mRNA in gonadal tissues of male mice. Biochem Biophys Res Commun 215:412421 10. Yamashita T, Ishii H, Shimoda K, Sampath TK, Katagiri T, Wada M, Osawa T, Suda T (1996) Subcloning of three osteoblastic cell lines with distinct differentiation phenotypes from the mouse osteoblastic cell line KS-4. Bone 19:429436 11. Komaki M, Katagiri T, Suda T (1996) Bone morphogenetic protein-2 does not alter the differentiation pathway of committed progenitors of osteoblasts and chondroblasts. Cell Tissue Res 284:917 12. Katagiri T, Yamaguchi A, Komaki M, Abe E, Takahashi N, Ikeda T, Rosen V, Wozney JM, Fujisawa-Sehara A, Suda T (1994) Bone morphogenetic protein-2 converts the differentiation pathway of C2C12 myoblasts into the osteoblast lineage. J Cell Biol 127:1755 1766 13. Thiede MA, Smock SL, Petersen DN, Grasser WA, Thompson DD, Nishimoto SK (1994) Presence of messenger ribonucleic acid encoding osteocalcin, a marker of bone turnover, in bone marrow megakaryocytes and peripheral blood platelets. Endocrinology 135:929937 14. Geoffroy V, Ducy P, Karsenty G (1995) A PEBP2/AML-1 related factor increases osteocalcin promoter activity through its binding to an osteoblast-specic cis-acting element. J Biol Chem 270:3097330979 15. Banerjee C, Hiebert SW, Stein JL, Lian JB, Stein GS (1996) An AML-1 consensus sequence binds an osteoblast-specic complex and transcriptionally activates the osteocalcin gene. Proc Natl Acad Sci USA 93:49684973 16. Ducy P, Zhang R, Geoffroy V, Ridall AL, Karsenty G (1997) Osf2/Cbfa1: a transcriptional activator of osteoblast differentiation. Cell 89:747754 17. Ducy P, Karsenty G (1995) Two distinct osteoblast-specic cisacting elements control expression of a mouse osteocalcin gene. Mol Cell Biol 15:18581869 18. Komori T, Yagi H, Nomura S, Yamaguchi A, Sasaki K, Deguchi K, Shimizu Y, Bronson RT, Gao YH, Inada M, Sato M, Okamoto R, Kitamura Y, Yoshiki S, Kishimoto T (1997) Targeted disruption of Cbfa1 results in a complete lack of bone formation owing to maturational arrest of osteoblasts. Cell 89:755764

the expression of OG1 in vivo may be decient in culture conditions. Conversely, the levels of an inhibitor(s) of OG1 expression may be different in in vivo and in vitro conditions. Further studies are necessary to clarify these issues. In bone cells, Runx 2/Cbfa-1/Osf-2 has been identied as a transcription factor required for the bonespecic expression of osteocalcin [1416]. The OSE2, which is a Runx 2 consensus sequence in the osteocalcin promoter, is present at the same regions in both OG1 and OG2 promoters [17]. Furthermore, these two promoters have very similar DNA sequences, at least up to 1.0 kb [17]. The present ndings, however, showed that the expression of OG1 was predominant in mouse bone in vivo. These ndings suggest that the OG1 promoter has additional regulatory elements, which are recognized by some unidentied transcriptional factors, rather than by Runx 2. All these factors may be required for the full expression of osteocalcin mRNA in mice in vivo. It was reported that BMP-2 induced osteocalcin expression in Runx 2 (/) cells in vitro [18]. It would, therefore, be interesting to examine the ratio of OG1 and OG2 mRNAs in Runx 2 (/) cells stimulated by BMP-2. Further studies are necessary to clarify the molecular mechanisms of osteocalcin mRNA expression. In conclusion, we have developed a new method to determine directly the expression levels of mouse OG1, OG2, and ORG transcripts relative to the total amount of osteocalcin mRNA. Using this method, we found that OG1 was the major transcript among the three osteocalcin mRNAs in mouse bone in vivo. However, OG1 and OG2 were almost equally expressed in osteoblasts in vitro. These ndings suggest that there is a novel mechanism regulating the full expression of osteocalcin mRNA in vivo. Acknowledgments. We thank Yamanouchi Pharmaceutical, Japan for their kind supply of BMP-2. This work was supported in part by a Grant-in-Aid from the Ministry of Science, Education, and Culture of Japan (11771146) and a Grant-in-Aid for young researchers from the Kitasato University Alumni Association (to T. K.).

References
1. Hauschka PV, Lian JB, Cole DEC, Gundberg CM (1989) Osteocalcin and matrix Gla protein: vitamin K-dependent proteins in bone. Physiol Rev 69:9901047