Identification of a novel response element in the rat bone sialoprotein (BSP) gene promoter that mediates constitutive and fibroblast growth factor 2-induced expression of BSP.

Bone sialoprotein (BSP) is a sulfated and phosphorylated glycoprotein, found almost exclusively in mineralized connective tissues, that may function in the nucleation of hydroxyapatite crystals. We have found that expression of BSP in osteoblastic ROS 17/2.8 cells is stimulated by fibroblast growth factor 2 (FGF2), a potent mitogen for mesenchymal cells. Stimulation of BSP mRNA with 10 ng/ml FGF2 was first evident at 3 h ( approximately 2.6-fold) and reached maximal levels at 6 h ( approximately 4-fold). From transient transfection assays, a FGF response element (FRE) was identified (nucleotides -92 to -85, "GGTGAGAA") as a target of transcriptional activation by FGF2. Ligation of two copies of the FRE 5' to an SV40 promoter was sufficient to confer FGF-responsive transcription. A sequence-specific protein-DNA complex, formed with a double-stranded oligonucleotide encompassing the FRE and nuclear extracts from ROS 17/2.8 cells, but not from fibroblasts, was increased following FGF2 stimulation. Several point mutations within the critical FRE sequence abrogated the formation of this complex and suppressed both basal and FGF2-mediated promoter activity. These studies, therefore, have identified a novel FRE in the proximal promoter of the BSP gene that mediates both constitutive and FGF2-induced BSP transcription.

ated, and glycosylated protein that is characterized by its ability to bind to hydroxyapatite, through polyglutamic acid sequences, and to mediate cell attachment, through an RGD sequence (1)(2)(3)(4). The expression of BSP is essentially restricted in the mineralized connective tissues. Studies on the developmental expression of BSP have shown that BSP mRNA is produced at high levels at the onset of bone, dentin, and cementum formation (5,6). Furthermore, the temporo-spatial deposition of BSP into the extracellular matrix (6 -8) and the ability of BSP to nucleate hydroxyapatite crystal formation (9) indicate a role for this protein in the initial mineralization of bone, dentin, and cementum (4). Recent studies have shown that BSP is also expressed in osteotropic cancers, suggesting BSP might play a role in the pathogenesis of bone metastases (10). Thus, regulation of the BSP gene is important in the differentiation of osteoblasts, in bone matrix mineralization and in tumor metastasis. The human (11,12), mouse (13), and rat (14) BSP genes have been cloned and partially characterized. These promoters include a highly conserved region (BSP box) that extends upstream from the transcription start site to nt Ϫ370 (15). This region includes a functional, inverted TATA element (nts Ϫ24 to Ϫ19) (16), which overlaps a vitamin D response element (17). In addition, putative sites of regulation through an inverted CCAAT box (Ϫ50 to Ϫ46) (18) and an AP-2 site (Ϫ447 to Ϫ440), which overlaps a transforming growth factor-␤ activation element, have been identified in the proximal promoter (19). Further upstream, a glucocorticoid response element (GRE) overlapping an AP-1 site has been characterized (20,21). Recently, we have identified a pituitary-specific transcription factor-1 (Pit-1) motif through which the stimulatory effects of parathyroid hormone on BSP transcription are mediated (22), while studies by Benson et al. (23) have indicated that a homeodomain binding element in the BSP promoter is required for osteoblast-specific expression.
Fibroblast growth factor-2 (FGF2 or basic FGF), a member of the heparin-binding growth factor family of mitogens, has been implicated in a range of normal physiological processes from embryonic mesoderm induction and pattern formation to angiogenesis and wound repair (24,25). FGF2 is also synthesized by osteoblasts and is stored in a bioactive form in the extracellular matrix (26 -29), where it acts as a local regulator of bone formation. The FGF family of molecules transduce signals to the cytoplasm via a family of transmembrane receptors with , and by Research for the Frontier Science (The Ministry of Education, Science, Sports and Culture). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
tyrosine kinase activity (30 -34). Four distinct gene products encode highly homologous FGF receptors (FGFRs; FGFR1-4), which share 56 -71% amino acid sequence identity. The FGF receptors contain three extracellular immunoglobulin-like domains, a single transmembrane domain, and an intracellular tyrosine kinase domain. Immunoglobulin-like domains II and III are sufficient for FGF binding and determine affinity (30 -34). Mutations in the FGFR1 gene are associated with Pfeiffer syndrome, which is one of the classic autosomal dominant craniosynostosis syndromes (31), while mutations in FGFR2 and FGFR3 produce genetic disorders involving bone development. Jackson-Weiss and Crouzon syndromes are allelic with mutations in FGFR2 (32,33). Thanatophoric dysplasia, the most common neonatal lethal skeletal dysplasia, and achondroplasia are caused by mutations in FGFR3 (34). Analysis of FGFR3-deficient mice has revealed prolonged bone growth, showing that FGFR3 is a negative regulator of bone growth (35). Collectively these studies, and the observation that intravenous FGF2 stimulates bone formation and mineralization (28,36), indicate that FGF is an important regulator of bone formation.
FGF2 inhibits alkaline phosphatase activity in ROS 17/2.8 cells (27) and the calcification of hypertrophic chondrocytes (37). FGF2 also inhibits type I collagen and osteocalcin transcription in ROS 17/2.8 cells and MC3T3-E1 cells (26,27,38). However, the combination of FGF2 and forskolin markedly up-regulates osteocalcin mRNA accumulation in MC3T3-E1 cells (38). An osteocalcin FGF2 response element (GCAGTCA motif) has been identified in the proximal promoter of the rat osteocalcin gene as a target of FGF2 and cAMP stimulation (38). The induction of human osteocalcin transcription by FGF2 requires the interaction of CCAAT motif that overlaps with three tandem repeats of a nuclear factor-1 half-site (TTGGC) (39).
To determine the molecular mechanism of FGF2 regulation of the BSP gene, we have analyzed the effects of the FGF2 on the expression of BSP in ROS 17/2.8 cells. These studies have revealed a novel FRE that mediates both the constitutive and FGF2-induced expression of BSP in osteoblastic cells.
Cell Culture-The rat clonal cell lines, ROS 17/2.8 (generously provided by Dr. G. A. Rodan, Merck-Sharpe and Frosst, West Point, PA) was used in these studies as an osteoblastic cell line that synthesizes BSP (20). Cells grown to confluence in 60-mm tissue culture dishes in ␣-MEM medium containing 10% fetal bovine serum were changed to ␣-MEM without serum and incubated with or without 10 ng/ml FGF2 for time periods extending over 3-24 h. To determine the effect of FGF2 on the stability of BSP mRNA, cells were first incubated for 6 h in the presence or absence of 10 ng/ml FGF2 and the incubation continued for up to 24 h in the presence of 60 M concentration of the transcription inhibitor, 5,6-dichloro-1-␤-D-ribofuranosyl benzimidazole. RNA was isolated from triplicate cultures at various time intervals and analyzed for the expression of BSP mRNA by Northern hybridization as described below.
Northern Hybridization-Total RNA from the culture cells was extracted with guanidium thiocyanate and, following purification, 20-g aliquots of RNA were fractionated on a 1.2% agarose gel and transferred onto a Hybond N membrane, as described previously (20). Hybridizations were performed at 42°C with either a 32 P-labeled rat BSP or rat glyceraldehyde-3-phosphate dehydrogenase (GAPDH), DNA probe. Following hybridization, membranes were washed four times for 5 min each at 21°C in 300 mM sodium chloride, 30 mM trisodium citrate, pH 7.0, containing 0.1% SDS. This was followed by two, 20-min washes at 55°C in 15 mM sodium chloride, 1.5 mM trisodium citrate, pH 7.0, 0.1% SDS. The hybridized bands, representing the two polyadenylated forms (1.6 and 2.0 kilobases) of rat BSP mRNA, were scanned in a Bio-imaging analyzer (Fuji BAS 2000, Tokyo, Japan) and normalized to the expression of GAPDH.
Transient Transfection Assays-Exponentially growing ROS 17/2.8 cells were used for transfection assays. Twenty-four hours after plating, cells at 50-70% confluence were transfected using a LipofectACE reagent. The transfection mixture included 1 g of a luciferase (LUC) construct (20) and 2 g of pSV-␤-galactosidase vector as an internal control. Two days post-transfection, cells were deprived of serum for 12 h, and 10 ng/ml FGF2 was added for 6 h prior to harvesting. The luciferase assay was performed according to the supplier's protocol (picaGene, Toyo Inki, Tokyo, Japan) using a Luminescence reader BLR20 (Aloka) to measure the luciferase activity. The protein kinase inhibitor H89 (5 M) and H7 (5 M) were used to inhibit protein kinases A and C. Herbimycin A (1 M) and PP1 (10 M) were used for tyrosine kinase and Src tyrosine kinase inhibition, respectively (40). U0126 (5 M) was used for MAP kinase kinase (MEK) inhibitor (41). The FRE sequence, identified from the transient transfection studies, was cloned in the BglII site of pGL2-promoter vector, immediately upstream of the enhancerless SV40 promoter. This construct was used to identify the sequence within the BSP promoter that is required for transcriptional induction by FGF2. Oligonucleotide-directed mutagenesis by PCR was utilized to introduce the dinucleotide substitutions using the Quikchange Site-directed Mutagenesis Kit (Stratagene, La Jolla, CA). All constructs were sequenced as described previously (20) to verify the fidelity of the mutagenesis. The pCG-ATF2 and pCG-ATF3 expression plasmids were kindly provided by Dr. T. Hai (42).
Statistical Analysis-Triplicate samples were analyzed for each experiment and experiments replicated to ensure consistency of the responses to FGF2. Significant differences between control and FGF2 treatment were determined using Student's t test.

Stimulation of BSP mRNA Expression in ROS 17/2.8
Cells-To study the regulation of BSP expression by FGF2, we used ROS 17/2.8 cells, which have been shown to have osteoblastic characteristics (44,45) and to express BSP mRNA con-

FGF2 Regulation of BSP Gene Expression
stitutively (20). First, a dose-response relation for FGF2 induction of BSP was established by treating the ROS 17/2.8 cells with different concentrations of FGF2 for 6 h and measuring the BSP mRNA levels by Northern blot analysis. At 1-50 ng/ml, FGF2 increased BSP mRNA with a maximal effect at 10 ng/ml (Fig. 1A). This optimal level of FGF2 (10 ng/ml) was used to determine a time course of BSP mRNA expression (Fig. 1B). FGF2 up-regulated BSP mRNA accumulation markedly in ROS 17/2.8 cells. A stimulation of 2.6-fold was evident 3 h after the addition of FGF2, with maximal levels (4.0-fold) of BSP mRNA obtained at 6 h. In comparison, osteopontin mRNA, which has been shown previously to be stimulated by FGF2 (27), was increased at 3 h and returned to base line at 12 h, whereas no effect on GAPDH mRNA was observed.
Analysis of BSP mRNA Stability-To determine whether the increase in BSP mRNA was due to an increased stability of the BSP mRNA in response to FGF2 treatment, ROS 17/2.8 cells were incubated in the presence of the transcription inhibitor 5,6-dichloro-1-␤-d-ribofuranosyl benzimidazole and the BSP mRNA levels determined over a 24-h period. From regression analysis a t1 ⁄2 of ϳ16 h was determined for BSP mRNA in the ROS 17/2.8 cells with no significant change observed in the presence of FGF2, indicating that the increase in mRNA was due to increased gene transcription (data not shown).
Transient Transfection Analysis of Rat BSP Promoter Constructs-To determine the site of FGF2-regulated transcription in the 5Ј-flanking region of the BSP gene, various sized promoter constructs ligated to a luciferase reporter gene were transiently transfected into ROS 17/2.8 cells and their transcriptional activity determined in the presence of FGF2. The constructs used, pLUC1-pLUC5, encompassing nucleotides Ϫ116 to ϩ60, gave a 2.9-fold increase in transcription after 6 h treatment with 10 ng/ml FGF2 (Fig. 2). FGF2 also increased transcription of pLUC4 (Ϫ425 to ϩ60) and pLUC5 (Ϫ801 to ϩ60). Within the DNA sequence that is unique to pLUC3 (between nts Ϫ116 to Ϫ43), an inverted CCAAT box (ATTGG; between nts Ϫ50 and Ϫ46), a possible cAMP response element (CRE; between nts Ϫ75 and Ϫ68), and a pituitary-specific transcription factor-1 (Pit-1) motif (between nts Ϫ111 and Ϫ105), which is the target of parathyroid hormone stimulation, are present (Fig. 3A).
By using a series of 5Ј deletion constructs between nts Ϫ116 and Ϫ43, we found that the FGF2 response was mediated by a region between nts Ϫ108 and Ϫ84 of the promoter sequence (Fig. 4). A series of 2-base pair mutations were made between nts Ϫ92 and Ϫ85 within pLUC3 construct (Fig. 5). All four constructs (mutations 1-4; pLUC3M1-4) had lower basal activities than pLUC3 and resulted in near abolition of the FGF2 effects on the promoter. In particular, the mutation in pLUC3M2 drastically reduced basal expression and completely abolished the FGF2 effect (Fig. 5). Thus, the GGTGAGAA motif (FRE, FGF response element) in the region nts Ϫ92 to Ϫ85 is important for basal expression as well as being necessary for the FGF2 induction of BSP promoter activity. To examine whether the FRE of the rat BSP promoter confers FGF2-dependent inducibility in the context of a promoter that is not stimulated by FGF2, the DNA segment between nts Ϫ98 and Ϫ79 in the rat BSP gene was inserted 5Ј to the SV40 in the BglII site of the pGL2-promoter. While the insertion of a single FRE sequence in the same orientation as in the BSP promoter did not influence basal activity of the SV40 promoter, and only a modest increase in FGF-mediated transcription was found, two copies of FRE increased basal activity and significantly increased FGF-mediated transcription (Fig. 6). Notably, the

FGF2 Regulation of BSP Gene Expression
FRE sequences identified in the rat BSP promoter are conserved in the mouse and human BSP promoters ( Fig. 3B;  underlined).
Since protein kinases mediate FGF2 signaling activities, we also investigated the effects of the PKC inhibitor H7, the PKA inhibitor H89, the tyrosine kinase inhibitor herbimycin A, the Src kinase inhibitor PP1, and the MEK inhibitor U0126 on FGF-mediated transcription. Whereas FGF2-induced pLUC3 promoter activation was inhibited by herbimycin A, PP1, and U0126, no effects were observed for either H7 or H89 (Fig. 7), indicating an involvement of Src and MAP kinase in the signaling pathway.
Gel Mobility Shift Assays-To identify nuclear proteins that bind to the FRE and mediate the FGF2 effects on transcription, double-stranded oligonucleotides were end-labeled and incubated with equal amounts (3 g) of nuclear proteins extracted from confluent ROS 17/2.8 cells that were either not treated (control) or treated with 10 ng/ml FGF2 for 6 and 12 h. With nuclear extracts from confluent, control cultures of ROS 17/2.8 cells, a shift of a single FRE DNA-protein complex was evident (Fig. 8, lane 1). After stimulation by FGF2 (10 ng/ml) for 6 and 12 h, DNA binding activity was increased (Fig. 8, lanes 2 and  3). That the DNA-protein complex represents a specific inter- FIG. 4. Fine 5 deletion mapping of the nts ؊116 to ؊43 element in the BSP promoter. A series of rat BSP promoter 5Ј deletion constructs were analyzed for relative promoter activity after transfection into ROS17/2.8 cells and examined for induction in the presence of FGF2 (10 ng/ml). The results of transcriptional activity obtained from three separate transfections with constructs Ϫ43 BSPLUC (Ϫ43 to ϩ60), Ϫ60 BSPLUC (Ϫ60 to ϩ60), Ϫ84 BSPLUC (Ϫ84 to ϩ60), and Ϫ116 BSPLUC (Ϫ116 to ϩ60) have been combined and the values expressed with S.E. Significant differences from control: *, p Ͻ 0.1; #, p Ͻ 0.2.
To verify that the FGF2 was operating through a unique FRE, we also used gel mobility shift analyses to evaluate the potential effects of FGF2 on the nearby inverted CCAAT, NFB, and AP1 sites. When we used the inverted CCAAT sequence as a probe, the DNA-NF-Y protein complex (18) did not change after FGF2 stimulation (Fig. 11A, lanes 1-3). Also, whereas a 20 -100-fold molar excess of unlabeled homologous probe attenuated the signal of the complex (Fig. 11A, lanes  4 -6), the FRE did not compete the shifted band (Fig. 11A, lanes  7-9). In comparison, AP1 binding was increased by FGF2 (Fig.  11B, lanes 3-5). Although a BSP-NFB DNA-protein complex did not change after FGF2 stimulation (Fig. 11B, lanes 6 -8), an Ig-NFB DNA-protein complex was increased after 6 h and returned to control levels at 12 h (Fig. 11B, lanes 9 -11). As the BSP-FRE and -NFB response elements are close to each other in the BSP promoter (Fig. 3A), and their DNA-protein complexes have similar mobility, competition experiments were performed. A 20 -100-fold molar excess of unlabeled FRE-3Ј, which extends further 3Ј from FRE, competed the FRE DNAprotein complex, whereas BSP-NFB and Ig-NFB (20 -100fold excess) did not compete the complex formation (data not shown), showing that the FRE-binding protein is distinct from NFB family proteins. This was confirmed by the inability of NFB p50 or p65 antibodies to either supershift or disrupt the FRE-nuclear protein complex, which was also unaffected by antibodies to CREB, Pit-1, Oct1, c-Jun, and c-Fos (data not shown).

DISCUSSION
FGFs have a prominent role in bone development and growth. During endochondral ossification unregulated FGF signaling can produce premature suture closure (craniosynostosis) and other craniofacial anomalies. BSP has been characterized as a unique marker of osteogenic differentiation that can regulate the formation of mineral crystals (4). Our studies have identified a novel GGTGAGAA element in the proximal promoter of the BSP gene that mediates both basal and FGF2stimulated transcription of BSP promoter-reporter constructs. This BSP-FRE binds a nuclear protein, the presence of which in unstimulated ROS 17/2.8 cells indicates that it is expressed constitutively. The nuclear protein also appears to be necessary for basal transcription of BSP, since double mutations in the FRE abrogate nuclear protein binding and transcription. FGF2 stimulated the expression of this nuclear protein in association with the increased transcription of both the endogenous BSP gene and chimeric constructs containing the FRE.
Since osteoblastic cells derived from an osteosarcoma were in our studies, the regulation of BSP by FGF2 that we have observed may also be relevant to the expression of BSP in tumors. In addition to being produced by osteogenic tumors, BSP is also expressed in breast, lung, thyroid, and prostate cancers (46 -49), and the presence of BSP in human primary breast cancers has been associated with an increased risk for subsequent bone metastases and a poor survival rate (10,46). The ability of BSP to bind to hydroxyapatite crystals and to mediate cell attachment through cell-surface integrins may be involved in the osteotropism of the metastatic cells (4, 10, 50). FGF2 is significantly increased in prostate cancers, relative to normal prostate (51), and the concentrations of FGF2 in nipple fluid is significantly increased in breast cancer patients (52). Recent reports have also shown expression of FGF and FGF receptor genes in human breast cancer cell lines and tumor samples (53), indicating a close relation between FGF expression and osteotropic cancers. Thus, it is conceivable that the BSP regulation by FGF2 could also relate to cancer cells.
The biological effects of FGFs are initiated by the autophosphorylation of receptor tyrosines, which provide high affinity binding sites for Src Homology 2 domain-containing signaling molecules, such as PLC␥, Src, FRS2, Grb2, PI3-K, and SHP-2, that activate a number of intracellular signaling pathways (54). In our studies the tyrosine kinase inhibitor, herbimycin A, and a Src family-selective tyrosine kinase inhibitor, PP1, inhibited FGF2-induced promoter activity (Fig. 7), indicating . DNA-protein complexes were separated on 5% polyacrylamide gel in low ionic strength Tris borate buffer, dried under vacuum, and exposed to an imaging plate for quantitation using an imaging analyzer. CONT, nuclear extract from control confluent cells.
that FGF2 increases expression of the BSP gene in ROS 17/2.8 cells through a tyrosine kinase and, more specifically, through Src tyrosine kinase-dependent pathway. Subsequent signaling also appears to involve the extracellular signal-regulated kinase/MAP kinase pathway, since the MEK1 and MEK2 inhibitor U0126 also inhibited FGF2-induced promoter activity. Notably, the Ras/Raf/MAP kinase pathway is required for mesoderm induction by FGF in Xenopus (55,56), for elongation of neurite outgrowth in PC12 cells by nerve growth factor and FGF (57) and for endothelial cell differentiation (58), whereas the mitogenic effects of FGF2 on rat osteoblastic Py1a cells are transduced by the PKC pathway (59). In contrast, protein kinase B is activated by stimuli such as insulin, platelet-derived growth factor, epidermal growth factor, and bFGF (60), whereas FGF receptor 3 (FGFR3) has been shown to be asso-ciated with Stat1 (61).
Ϫ84 are a target sequence of FGF2 (Fig. 4). Within this region of the promoter, a putative FRE was identified at nts Ϫ92 to Ϫ85. Notably, this sequence is conserved in the rat, mouse, and human BSP promoters (Fig. 3B), indicative of a potentially important role in transcriptional regulation of the BSP gene. That the FRE mediates FGF2 transcriptional regulation was shown using pLUC3 constructs encompassing mutations in the putative FRE element (Fig. 5). These experiments further identified a centrally located TGAG sequence to be crucial for FGF2-mediated transcription. The loss of basal transcription with pLUC3M2 further suggests that the FRE is critical for constitutive BSP expression.
We identified a protein in nuclear extracts of ROS 17/2.8 cells that selectively bound to the FRE and which was upregulated by FGF2 (Fig. 8) in association with increased transcription (Fig. 2). That the protein binds specifically to the FRE was demonstrated by a combination of competition gel mobility shift assays (Fig. 8) and gel mobility shifts with oligonucleotides representing the mutated forms of the FRE (Fig. 9) used for transcriptional analysis. Moreover, the FRE DNA-protein complex was unaffected by an inverted CCAAT, and consensus AP1 and GRE (Fig. 8), BSP-NFB, and Ig-NFB doublestranded oligonucleotides (results not shown). Interestingly, the protein binding to the FRE was not detected in nuclear extracts from cells that do not express the BSP gene, such as human gingival fibroblasts and human periodontal ligament cells (62), indicating that it could be bone cell specific. 2 Transcriptional regulation of the BSP gene by FGF2 contrasts mechanisms described for other genes. Thus, FGF2 leads to its own synthesis through an autoregulated transcriptional response that requires the transcription factor Egr-1, which binds to a site overlapping an SP1 binding motif (63). In comparison, up-regulation of matrix metalloproteinase-1 gene expression by FGF2 in NIH3T3 fibroblasts is mediated through an AP1 consensus sequence (64). FGF2 signaling in Xenopus requires AP1/Jun for early development and mesoderm induction (65), while FGF2 and cAMP synergistically activate proenkephalin gene expression via ATF3 and c-Jun through a cAMP response element (66). Members of the ATF/CREB family utilize a leucine zipper region to form heterodimers that bind to a consensus DNA sequence (TGACGTCA), and both ATF1 and ATF2 have been demonstrated to stimulate transcription. In contrast, ATF3 represses rather than activates transcription (42). In the rat osteocalcin gene, expression is synergistically induced by FGF2 and cAMP through a GCAGTCA motif rather than an AP1 sequence (TCAGTCA) (38). Transient coexpression of ATF3, but not ATF2, selectively attenuates the FGF2/ forskolin induction of rat osteocalcin transcription (38). The induction of human osteocalcin promoter activity by FGF2 requires the interaction of a CCAAT motif that overlaps with three tandem repeats of nuclear factor-1 half-sites (39). However, the factors mediating the FGF responses have not been identified. Our studies show that FGF2 alone can stimulate BSP transcription, and neither ATF2 and ATF3 has any effect on FGF2 stimulation (results not shown). Moreover, while AP1 binding activities were up-regulated by FGF2 in gel shift assays (Fig. 11B), and the AP1 sequence (TGAGTCA) is quite similar to central portion of the FRE motif (GGTGAGAA), the AP1 and FRE-protein complexes did not comigrate (Fig. 11B), and AP1 could not compete FRE DNA-protein complex (Fig. 8), indicating that the FRE-binding protein is distinct from AP1. Furthermore, we have shown that BSP gene expression is suppressed by 12-O-tetradecanoylphorbol 13-acetate through an AP1 motif (21). Although the FRE and NFB sequences are closely spaced in the BSP promoter, we found that unlabeled BSP-NFB and Ig-NFB could not compete FRE DNA-protein complex (results not shown), showing that the FRE-binding protein is distinct from NFB binding proteins. Current studies are aimed at identifying and characterizing the BSP-FRE binding protein because of its potential role in the regulating basal 2 E. Shimizu-Sasaki and Y. Ogata, unpublished results.  7-9). B, radiolabeled double-stranded FRE, consensus AP1, BSP-NFB, and Ig-NFB were incubated for 20 min at 21°C with nuclear protein extracts obtained from cells incubated without (lanes 1, 3, 6, and 9) or with FGF2 (10 ng/ml) for 6 h (lanes 4, 7, and 10) and 12 h (lanes 2, 5, 8, and 11). DNA-protein complexes were separated on a 5% polyacrylamide gel in low ionic strength Tris borate buffer, dried under vacuum, and exposed to an imaging plate for quantitation using an image analyzer. CONT, nuclear extract from control confluent cells.
as well as FGF2-induced transcription of BSP in osteoblasts.
In summary, we have shown that a FRE motif (GGTGAGAA) exists in the rat BSP proximal promoter through which the stimulatory effects of FGF2 on BSP gene transcription are mediated. Since BSP is expressed by differentiated osteoblasts, and FGF2 is a crucial factor for bone metabolism, it is conceivable that a FRE-binding transcription factor may contribute to the cell-specific expression of the BSP gene during the formation of the mineralized extracellular matrix of bone.