Regulation of the promoters for the human bone morphogenetic protein 2 and 4 genes

Gene 256 (2000) 123–138 www.elsevier.com/locate/gene

Regulation of the promoters for the human bone morphogenetic protein 2 and 4 genes Leah M. Helvering *, Robert L. Sharp, Xuemei Ou, Andrew G. Geiser Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, IN 46285, USA Received 13 March 2000; received in revised form 6 July 2000; accepted 8 August 2000 Received by A.J. van Wijnen

Abstract The bone morphogenetic proteins 2 and 4 are known to be important in bone formation and are expressed in both the developing and adult mammalian bone. Understanding the regulation of these genes in osteoblasts may yield methods by which we can control expression to induce bone formation. We have isolated and characterized the human BMP-2 and BMP-4 promoters and report substantially more upstream sequence information than that which has been published. Human osteoblasts were found to have a single transcript initiation site that is conserved across species, rather than multiple start sites, as has previously been reported (Feng, J.Q., Harris, M.A., Ghosh-Choudhury, N., Feng, M., Mundy, G.R., Harris, S.E., 1994. Structure and sequence of mouse morphogenetic protein-2 gene (BMP-2): comparison of the structures and promoter regions of BMP-2 and BMP-4 genes. Biochim. Biophys. Acta 1218, 221–224; Heller, L.C., Li, Y., Abrams, K.L., Rogers, M.B., 1999. Transcriptional regulation of the Bmp2 gene. J. Biol. Chem. 274, 1394–1400; Sugiura, T., 1999. Cloning and functional characterization of the 5∞-flanking region of the human bone morphogenetic protein-2 gene. Biochem. J. 338, 433–440). A series of promoter deletions for both human BMP-2 and BMP-4 fused to the luciferase reporter gene were analyzed thoroughly in human and murine osteoblastic cell lines. Several compounds and growth factors that stimulate general or osteogenic pathways were used to treat cells transfected with the promoter constructs. Retinoic acid compounds and the phorbol ester, PMA were found to stimulate BMP-2 and, to a lesser degree, BMP-4. The combination of all trans-RA and PMA caused a synergistic increase in BMP-2 promoter activity and endogenous mRNA. The RA stimulation appears to be an indirect effect on the BMP-2 promoter, as the most highly conserved RRE in the BMP-2 promoter was unable to functionally bind or compete for protein binding. Potential binding sites in both promoters for the bone-specific transcription factor, Cbfa-1, were found to specifically bind Cbfa-1 protein in osteoblast nuclear extracts; however, deletion of these sites did not significantly affect transcriptional activity of the promoters in osteoblasts. These data thus present new sequence and regulatory information for the human BMP-2 and BMP-4 promoters and clarify the human BMP-2 gene transcriptional start site in osteoblasts. © 2000 Elsevier Science B.V. All rights reserved. Keywords: BMP-2; BMP-4; CBFA-1; Promoter sequence; Retinoic acid; Transcription

1. Introduction Bone morphogenetic proteins were originally extracted from bone and identified by their ability to Abbreviations: at-RA, all-trans retinoic acid; bgal, b-galactosidase; BMP, bone morphogenetic protein; bp, base pair(s); cAMP, 8-bromo cyclic AMP; CBE, Cbfa-1 binding element; Cbfa-1, core-binding factor alpha-1; CRE, cyclic AMP response element; DTT, dithiothreitol; EMSA, electrophoretic mobility shift assay; GAPDH, glyceraldehyde3-phosphate-dehydrogenase; hBMP, human bone morphogenetic protein; oligo, oligonucleotide; PMA, phorbol 12-myristate 13-acetate; RA, retinoic acid; RAR, retinoic acid receptor; RRE, retinoic acidresponsive element; RXR, retinoic X receptor; SSC, 0.15 M sodium chloride/0.015 M sodium citrate pH 7.6; UTR, untranslated region. * Corresponding author. Tel.: +1-317-277-4401; Fax: +1-317-276-1414. E-mail address: [email protected] (L.M. Helvering)

form normal, endochondral bone when implanted into nonskeletal sites in rodents ( Urist, 1965; Reddi and Huggins, 1972; Reddi, 1981; Urist et al., 1982). This de-novo bone-forming activity was found to contain several structurally related proteins from the transforming growth factor-b superfamily ( Wozney et al., 1988). Several of the BMP isoforms, including BMP-2 and BMP-4, have been shown to form ectopic bone when implanted as single recombinant proteins ( Wang et al., 1990; Hammonds et al., 1991) and have also demonstrated effective therapeutic potential in the clinical setting for the repair of non-union fractures and in fracture healing (Riley et al., 1996). In addition to their ability to induce bone formation, these proteins show diverse expression patterns throughout embryonic devel-

0378-1119/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S0 3 7 8 -1 1 1 9 ( 0 0 ) 0 0 36 4 - 4

124

L.M. Helvering et al. / Gene 256 (2000) 123–138

opment and are thought to regulate proliferation, apoptosis, differentiation, cell-fate determination, patterning, and morphogenesis in multiple organ systems (Lyons et al., 1990; Hogan, 1996). While BMP-2 and BMP-4 are similar in that they bind to common receptors (Massague and Weis-Garcia, 1996) and can induce de-novo bone formation, their promoter sequences are different, suggesting that their regulation at the transcriptional level may be different. Understanding the regulation of the human BMP-2 and BMP-4 genes could be useful in identifying elements necessary for bone-specific expression of these transcripts and aid in the recognition of other molecular targets involved in bone formation. In order to understand possible transcriptional controls of these genes, it is essential to isolate and characterize the promoter of each in osteoblastic cells. Furthermore, an extensive side-by-side analysis of promoter activity in bone cells in response to treatments with osteogenic agents and various stimulators or inhibitors of signaling pathways has not been done. Here, we provide new insight into the transcriptional control of the largest fragments of human BMP-2 and BMP-4 promoter reported to date and provide evidence that human osteoblasts use a single, conserved distal promoter for BMP-2 gene transcription rather than multiple sites, as has been reported previously (Heller et al., 1999; Sugiura, 1999). We present data that identify the involvement of the protein kinase A and C pathways in the transcriptional regulation of BMP-2 and BMP-4, Cbfa-1 binding sites in both promoters, and we show that the stimulation of BMP2 mRNA by retinoids (Rogers et al., 1992; Rogers, 1996; Heller et al., 1999) is likely an indirect effect rather than a direct effect on the promoter through a defined retinoid-responsive element.

2. Methods 2.1. Isolation of human BMP-2 and BMP-4 promoters Oligos homologous to portions of the UTR of the published mouse BMP-2 (Accession No. L25602) and BMP-4 (Accession No. M22490) genes were used with the Human Promoter Finder Kit (Clontech, Palo Alto, CA) to obtain upstream genomic sequences according to the manufacturer’s directions. Fragments obtained from the Human Promoter Finder Kit were gel-purified with GeneClean (BIO101, Vista, CA), cloned into pCR2.1 with the TA cloning kit (Invitrogen, Carlsbad, CA) and sequenced. Clones containing upstream sequence of human BMP-2 and BMP-4 were digested with EcoRI (Gibco BRL, Gaithersburg, MD) and their inserts were gel-purified as before. These human BMP2 and BMP-4 inserts were used as probes for hybridization screens of human genomic libraries by Genome

Systems (St. Louis, MO). Positive BAC clones from Genome Systems were used to subclone an approximately 7.8 kb EcoRI fragment of BMP-2 and a 6.8 kb XhoI (Gibco BRL) fragment of BMP-4 into pUC18 (Pharmacia, Piscataway, NJ ). 2.2. Promoter/reporter constructs Different lengths of the human BMP-2 and BMP-4 promoters were cloned into the luciferase reporter plasmid, pGL3-Basic (Promega, Madison, WI ). The BMP2 promoter serial 5∞ deletion construct inserts were generated by PCR amplification utilizing primers with engineered restriction endonuclease recognition sites and cloned into pGL3-Basic. The various BMP-4 promoter constructs were made by cloning 2443 bp of the promoter into the XhoI site of pGL3-Basic and subsequently generating smaller constructs by using specific restriction endonucleases to remove 5∞ portions of the promoter. Additional BMP-4 promoter constructs (the −2443, −1917, −733) and BMP-2 promoter constructs (−2124 bp and −1819) were made by subcloning the promoter fragments from the pGL3-Basic promoter constructs into the pbgal Basic Vector (Clontech) ( Figs. 3 and 5). 2.3. Transient transfection assays U-2OS cells were plated in a 24-well dish at 1×104 cells per well in DMEM complete media (Gibco BRL) with 5% fetal calf serum (Gibco BRL). The following day, 500 ng of each luciferase deletion construct and 40 ng of a pCMV-b plasmid (Clontech) containing a b-galactosidase gene were transfected with 2 ml of Fugene reagent according to the manufacturer’s specifications ( Roche, Indianapolis, IN ). Cells were lysed after 16–24 h with 100 ml of lysis buffer (100 mM potassium phosphate, 0.2% Triton X, 1 mM DTT, pH 7.8). Half of the lysate was used to measure b-gal activity using a Tropix kit ( Tropix, Bedford, MA), and the rest was used to measure luciferase activity with a Promega luciferase kit and a Labsystems Luminoskan (Needham Heights, MA). Each deletion construct was run in quadruplicate, and luciferase activity was normalized to b-gal prior to analysis. Co-transfection assays were performed in six-well dishes with 40% confluent UMR106 cells, 0–80 ng of a mouse Cbfa-1 expression plasmid, 400 ng of the BMP/bgal constructs, and 2 ml of Fugene. Lysates were prepared and assayed as above. The CMV-driven fulllength mouse Cbfa-1 cDNA was kindly provided by Gerard Karsenty. 2.4. Determination of the transcriptional start sites S1 nuclease analysis was performed as previously described (Geiser et al., 1991). Briefly, various double-

L.M. Helvering et al. / Gene 256 (2000) 123–138

stranded DNA fragments of the human BMP-2 and BMP-4 genomic constructs spanning the promoter and 5∞- UTR regions were first digested with a restriction enzyme that cut within the 5∞ UTR region. This end was labeled with [33P] ATP (Amersham Pharmacia Biotech, Piscataway, NJ ), and a second restriction enzyme was used within the putative promoter region to release the labeled probe fragment and gel-purified. Probe (2–5×104 cpm) was then added to 80 mg of total RNA isolated from human osteoblast cells and denatured at 75–85°C (the temperature depended on the individual probe) in 71% formamide/buffer for 15 min, then allowed to drop to hybridization temperature (60–70°C ) overnight. S1 nuclease (150 units, Roche) was used to digest unhybridized mRNA for 1 h at 37°C, then samples were extracted with phenol/chlorophorm, ethanol-precipitated, resuspended in loading buffer, heated to 80°C for 5 min and loaded onto 6% denaturing acrylamide gels. The annealing temperatures for the BMP-2 probes were 65°C for the PstI(+352)/SgrAI(+911) and BamHI(−105)/SgrAI(+911) fragments and 70°C for the BamHI(−105)/ApaLI(+567) and BamHI(−105)/ EcoNI(+348) fragments. The NcoI (−296)/XhoI (+208) fragment of BMP-4 was annealed at 64.5°C. Restriction enzymes were purchased from Gibco BRL and New England Biolabs (Beverly, MA). 2.5. Cell culture and transfections Stable cell lines, C2C12 (a myoblastic precursor cell with osteoblastic potential ), two human osteosarcoma cell lines U-2OS and SaOS-2, the rat osteosarcoma line UMR106, the mouse preosteoblast cell line MC3T3, and the human kidney line 293 were maintained in Alpha MEM media supplemented with 5% fetal calf serum, penicillin–streptomycin, non-essential amino acids, and glutamine (Gibco BRL). Selection of stable neomycin cell clones was in 500 mg/ml of G418 (Gibco BRL), and maintained in 250 mg/ml of G418. 2.6. Compounds used in treatment All-trans retinoic acid, 9-cis RA, 13-cis RA, forskolin (2 and 10 mM ), 8-bromo cAMP (1 mM ), ascorbate (50 mg/ml ), b-glycerol phosphate (10 mM ), 17 b-estradiol (0.1–10 nm), dexamethasone (0.1 5.0mM ), bovine insulin (10 mg/ml ), osteonectin 6.25 mg /ml, and phorbol 12-myristate 13-acetate (PMA) were purchased from Sigma (St. Louis, MO). Targretin and insulin-like growth factor-I (25 ng/ml ), and insulin-like growth factor-II (150 ng/ml ), were synthesized at Lilly Research Labs (Indianapolis, IN ). Bombesin (10 nM ), suramin (100 mM ), mastoparam-7 (15 mM ), osteonectin, interleukin-6 (100 and 200 ng/ml ), interleukin-6 soluble receptor, and prostoglandin E (1 and 5 mM ) were 2 purchased from Calbiochem (San Diego, CA), epider-

125

mal growth factor-1 (50 ng/ml ) was purchsed from Gibco BRL, and 1,25-dihydroxy vitamin D3 (100 and 10 nM ) was obtained from Biomol (Plymouth Meeting, PA). Transforming growth factor-b 1 (2 ng/ml ), osteocalcin (4.4 mg/ml ), acidic (1 ng/ml ) and basic fibroblast growth factor (0.3 ng/ml ), platelet-derived growth factor aa (10 ng/ml ), platelet-derived growth factor bb (10 ng/ml ) were purchased from R & D Systems (Minneapolis, MN ). 2.7. Northern blot analysis RNA was isolated by the guanidium/CsCl procedure as described previously (Maniatis et al., 1982). Thirty micrograms of RNA were loaded onto a 1% formaldehyde/agarose gel and transferred to a Nytran (Schleicher and Schuell, Keene, NH ) filter following electrophoretic separation. The RNA was then crosslinked to the membrane in a UV Stratalinker 1800 (Stratagene, La Jolla, CA) and hybridized as described previously (Church and Gilbert, 1984). Probes used in RNA analysis were generated by random-priming an approximately 200 bp cDNA fragment of human BMP2 and GAPDH with a Boehringer Mannheim HighPrime DNA Labeling Kit (Roche) and [33P] dCTP (3000 ci/mMol, Amersham). Washed membranes were exposed to phosphorimaging plates, and relative mRNA levels were quantified on a Molecular Dynamics PhosphorImager (Sunnyvale, CA). 2.8. Dot blot analysis Twenty micrograms of RNA in 4 M formaldehyde and 6.5× SSC were heated to 65°C for 15 min and then chilled on ice. Samples were then dot-blotted onto a Nytran membrane (Schleicher & Schuell ) under vacuum as suggested by the manufacturer (Stratagene). The RNA was cross-linked to the Nytran, and the membranes were probed and quantitated as described above. In addition to dot blotting, all samples were subjected to Northern analysis (as described above) to verify the integrity of the RNA and the size of the transcripts following hybridization. 2.9. Electrophoretic mobility shift assay (DNA/protein binding) Nuclear extracts were isolated from cultured cell lines as described previously ( Krishnan et al., 1997). Doublestranded DNA probes (putative binding elements in bold ) were prepared by annealing the following oligos: osteocalcin CBE 5∞GCTGCAATCACCAACCACAGCA and 5∞GGATGCTGTGGTTGGTGATTG; BMP-2 CBE (−2054) 5∞ CGAACTCATTTCCACCACAAGAGG and 5∞CCTCTTGTGGTGGAAATGAGT; BMP-4 CBE (−865) 5∞GGTTACTGCTTCTGTGGTTATC-

126

L.M. Helvering et al. / Gene 256 (2000) 123–138

L.M. Helvering et al. / Gene 256 (2000) 123–138

127

Fig. 1. DNA sequence of the human BMP-2 and BMP-4 promoter regions. (A) DNA sequence of the human BMP-2 promoter, with potential transcription factor binding elements and the TATA box in bold print and hexamer sites underlined. The numbers refer to the position relative to the transcriptional start site (+1). The shaded portions indicate regions of 90% or greater homology to the mouse sequence. (B) DNA sequence of the human BMP-4 promoter, with potential transcription factor binding elements of interest in bold print. The numbers refer to the position relative to the transcriptional start site (+1).

128

L.M. Helvering et al. / Gene 256 (2000) 123–138

TCC and 5∞GGAGATAACCACAGAAGCAGTAACC; BMP-4 CBE (−2442) 5∞GATCACTAGGCTGTGGTTGATTTAG and 5∞GCTAAATCAACCACAGCCTAGTGAT; RARb RRE 5∞GGAGGGTTCACCGAAAGTTCACT and 5∞GGAGTGAACTTTCGGTGAACCCA; BMP-2 RRE (−1819) 5∞GCTGCAAGACATCCTTGAGGTCATCACAGGATA and 5∞GGTATCCTGTGATGACCTCAAGGATGTCTT. Annealed oligos (100 pM ) were labeled by filling in the 5∞ overlaps with 50 mCi [33P]-dCTP (Amersham) and other nucleotides with Klenow (Gibco BRL). Nuclear extracts (5 mg) were incubated with approximately 1.0 pM probe (25 000 cpm) for 15 min in a binding buffer consisting of 100 mM NaCl, 50 mM Tris, pH 7.5, 1 mM DTT, 1 mM EDTA, 5 mg poly (dI–dC ) and 1 mg poly (dG– dC ) (Pharmacia). Samples were separated on a 5% nondenaturing polyacrylamide gel, dried, exposed to phosphorimager plates, and imaged as described previously. Antibody specific to Cbfa1 was generated in rabbits (Genosys Biotechnologies, The Woodlands, TX ) with a peptide corresponding to the N-terminal 26 amino acids of Cbfa1: MASNSLFSAVTPCQQSFFWDPSTSRR. Serum antibody was IgG-purified using the BioRad (Hercules, CA) Econo-Pac serum IgG DEAE columns, as per the manufacturer’s instructions.

3. Results 3.1. Isolation and analysis of the human BMP-2 and BMP-4 promoters BAC clones positive for the 5∞ genomic fragments of the human BMP-2 and BMP-4 genes were used to subclone an approximately 7.8 kb EcoRI fragment of BMP-2 and a 6.8 kb XhoI fragment of BMP-4. Sequence analysis and comparison to the known human cDNA sequence (accession M22489) indicated that the BMP-2 subclone comprised a promoter region, the first two exons, the first intron and a portion of the second intron. The 5∞ end of the BMP-4 subclone contained an Alu sequence in addition to approximately 4.0 kb of promoter sequence and ended in the first (non-coding) exon. The BMP-2 gene has a potential TATA box within 30 bases of the start site, while the BMP-4 gene has no such proximal TATA box (Fig. 1). Although there were no obvious sequence similarities between the two promoters (only 40% homologous in the first 500 bp), various potential regulatory elements were identified in both promoters by their sequence homology to known transcription factor binding sequences, including Cbfa-1, RRE, SOX, CRE, and Sp1 (Mangelsdorf and Evans, 1995; Ducy et al., 1997; Lania et al., 1997; Montminy, 1997; Ng et al., 1997). The BMP-4 promoter had five potential binding sites for the osteoblast specific transcription factor, Cbfa-1, while BMP-2 had a single CBE.

The BMP-2 promoter had a cluster of four potential Sp1 sites within 200 bases of the start of transcription, while the BMP-4 promoter had a single potential Sp1 site within 110 bases from the start site. In addition, the BMP-2 promoter has potential sites reported to bind EGR-1, MZF-1, and NFkB, as well as a direct repeat (CGCCGCCGCCGCCG) of undetermined significance at the start site of transcription. The 4.75 kb human BMP-2 promoter was compared to the published 1.96 kb mouse promoter (Harris et al., 1996) and was found to have a homology of 73%. Within 300 bases of the transcription start site, the homology is nearly 90% with two of the Sp1 sites, the TATA box, and the direct repeat absolutely conserved. The homology of the region between −1100 and −300 bp from the start site falls to 69% when compared to the mouse sequence but resumes a greater degree of homology further upstream (75–80%). In this distal region, (from −1400 bp to the end of the mouse sequence), there are several stretches of 40 or more bases that are 90–100% conserved between species, including the Cbfa-1 binding site (−2036) and a potential RRE flanked by imperfect hexamer sites (−1437 to −1413 bp) that are absolutely conserved. An additional RRE flanked by two imperfect hexamer sites (−1814 to −1787 bp), and a SOX site are nearly conserved with only 1–2 bp mismatches between species. A recent report of the 5∞-flanking region of the human BMP-2 gene (Sugiura, 1999) reports 1030 bp of sequence upstream of the start site, while this study provides an additional 4745 bp of sequence. The human BMP-4 promoter was compared with the published 2.3 kb mouse promoter (Feng et al., 1994, 1997) and nearly half of the mouse sequence was found to be greater than 84% homologous to that of the human sequence. However, the most distal mouse CBE site and the 2 proximal CBE sites were not conserved between human and mouse ( Fig. 1B). This study reports an additional 1100 bp of upstream sequence information for human BMP-4 from that previously reported (Shore et al., 1998). 3.2. Determination of transcription initiation sites for human BMP-2 and BMP-4 In order to identify the transcription start site, S1 nuclease protection assays were performed using RNA from U-2OS cells. Initial probes for the BMP-2 gene were constructed to overlap with the published 5∞ end of the cDNA for human BMP-2 ( labeled the SgrAI site at +50 of the published cDNA or +911 of the start of transcription noted in Fig. 1). When hybridized to U-2OS RNA, the entire probe (PstI/SgrAI 567 bp fragment) was protected, indicating that the start site of transcription was further upstream of the probe ( Fig. 2A). Extending the probe upstream to the BamHI

L.M. Helvering et al. / Gene 256 (2000) 123–138

A

B

129

site (BamHI/SgrAI 1028 bp fragment) resulted in a protected band smaller than the probe, indicating that this probe included a portion of the untranscribed region of the gene. Using shorter fragments of DNA with the same BamHI site at the 5∞ end as probes against U-2OS RNA, we mapped the transcriptional start site to approximately 870 bases upstream of the published cDNA end (Fig. 2B). This site appears to be the site of an upstream transcriptional start site reported for the mouse and deer BMP-2 gene ( Feng et al., 1994, 1997). As additional confirmation that the start site was indeed this far upstream, we performed RT-PCR on U-2OS mRNA using oligonucleotides that mapped between the new start site and the previously published 5∞ end of the cDNA, and a single fragment of the appropriate size was amplified (data not shown). It was necessary to add DMSO to our PCR reactions, however, to aid the polymerase in reading through the GC-rich 5∞ untranslated region of the BMP-2 gene. While no shorter transcripts were observed in S1 analysis with probes containing the entire 5∞ UTR, evidence of a larger protected transcript was apparent in RNA from U-2OS cells treated with at-RA and PMA (Fig. 2B). S1 analysis of the BMP-4 mRNA confirmed the approximate transcriptional start site identified by Shore et al. (1998) and was identical when either U-2OS or SaOS-2 cell mRNA was used. This site represents the start of transcription for the 1A exon splice variant, which is thought to represent the major BMP-4 transcript expressed in osteoblasts (van den Wijngaard et al., 1996; Metz et al., 1998; Shore et al., 1998). 3.3. Basal activity of the hBMP-2 and hBMP-4 promoters Initial BMP-2/luciferase plasmids contained a large portion of the 5∞ UTR because it was assumed that the transcriptional start site was similar to that of the proposed (downstream) mouse osteoblast BMP-2 start site. After it was determined that the human BMP-2 initiation of transcription in bone cells was actually 870 base pairs upstream of the previously reported site, we tested the entire BMP-2 promoter without the 5∞ UTR for activity in the human osteosarcoma cell line U-2OS

Fig. 2. S1 analysis of the human BMP-2 transcriptional initiation site. (A) The S1 probe extended from the PstI to the SgrAI sites (567 bp) within the 5∞ untranslated region of the transcript. A labeled 100 bp ladder was run in the first lane. A control for complete digestion of the probe is shown in the lane marked ‘tRNA’, and RNA from U-2OS cells (20 or 80 mg) demonstrates full-length protection of the probe. (B) The S1 probe extended further 5∞ to the BamHI site for a 453 bp fragment ending at the EcoNI site. U-2OS gave a protected band of approximately 350 bp, and the U-2OS stimulated with at-RA and PMA showed the same band of greater abundance in addition to a near full-length protected band.

130

L.M. Helvering et al. / Gene 256 (2000) 123–138

Fig. 3. BMP-2 promoter/reporter deletion constructs and their basal activity in U-2OS cells. A schematic representation of fragments of the BMP2 promoter cloned into pGL3-Basic is shown. The start of transcription is indicated at +1 and potential promoter elements of interest are noted. The solid line indicates various lengths of the BMP-2 promoter sequence and the LUC box represents the luciferase gene. Darkened bars to the right indicate the fold increase of luciferase activity of each promoter fragment (after normalization to b-Gal activity) relative to the −3073 bp construct. Error bars indicate the standard error of fold change over four separate experiments. Deletion constructs of the BMP-2 promoter fused to a luciferase reporter were transiently cotransfected with a CMV-b- Gal plasmid into U-2OS cells. After 16–24 h cell lysates were measured for luciferase and b-Gal activity.

cells. Subsequent experiments showed that this 5∞ UTR made little difference in the basal or induced level of expression (data not shown). Therefore, 5∞ serial deletion constructs of the BMP-2 promoter were made containing the long 5∞ UTR in addition to adjacent promoter sequence ranging from 320 bp to 4.75 kb upstream of the transcription start site. Basal activity in the U-2OS cell line varied within a 2.5-fold range in the constructs extending to 3073 bp upstream of the start site ( Fig. 3). However, the largest construct (4.75 kb of promoter) had far less basal luciferase activity, reflecting potential transcriptional repressor function. Other cell lines such as C2C12 and MC3T3 were also transiently transfected with the −3073 bp BMP-2 luciferase construct. A high basal activity for the −3073 bp BMP-2 promoter was found in the U-2OS and MC3T3 cell lines in contrast to a much lower activity in the C2C12 line ( Fig. 4). Serial 5∞ deletion constructs were made in the BMP4 promoter creating constructs ranging from 173 bp to 2.5 kb of upstream promoter sequence. Transient transfections of the BMP-4 promoter constructs done in U-2OS cells showed that basal activity varied only within a 2.5-fold range across these deletion constructs

Fig. 4. Relative luciferase activity of the −3073 bp BMP-2 construct in various cell lines. The −3073 bp BMP-2 luciferase construct was used to transiently transfect U-2OS, MC3T3, and C2C12 cell lines. The mean relative light units representing at least four samples is shown. Error bars=S.E.

L.M. Helvering et al. / Gene 256 (2000) 123–138

and was very similar to those described by Shore et al. (1998) (Fig. 5). The relative luciferase activity of all the BMP-4 promoter constructs was five to 10-fold higher than that of the BMP-2 constructs. 3.4. Retinoid activation of the human BMP-2 and 4 promoters Previous reports showed that RA can stimulate the expression of BMP-2 in embryonal carcinoma cells, adenocarcinoma cells and in the developing chick limb (Hatakeyama et al., 1996; Hogan, 1996; Rogers, 1996; Rogers et al., 1992). Initially, we tested our human BMP-2 promoter deletion constructs for inducibility by transient transfection in the presence of 1mM at-RA and saw nearly a twofold increase in luciferase activity with the −2124 bp construct in U-2OS cells. Further testing with the more stable retinoid, 13-cis RA, was done to evaluate the deletion constructs in U-2OS cells. We found that the constructs with at least 468 bp of BMP2 promoter sequence were consistently induced by 13cis RA, while targretin and at-RA had no effect on promoter stimulation in transient transfections (data not shown). Unlike the BMP-2 reporter plasmid, transient transfection of the full-length BMP-4 luciferase plasmid in the presence of 13-cis RA for 40 h showed no effect on luciferase activity in U-2OS cells. Regulation of the BMP-4 luciferase plasmid in the presence of targretin also differed from BMP-2 transients where luciferase activity for the BMP-4 reporter was sup-

131

pressed but had no effect on the BMP-2 construct (data not shown). Stable cell lines of the 3.0 kb BMP-2 promoter/ luciferase construct were established in C2C12, U-2OS, SaOS-2, MC3T3, and 293 cells to help determine cell specificity of the promoter in response to retinoids and other compounds. Additionally, a stable U-2OS cell line transfected with the 2.5 kb BMP-4 promoter/luciferase construct was established to evaluate retinoid activity on the promoter. Individual clones were characterized for basal and induced promoter activity. The majority of clones selected from each cell line could be induced within 24 h by treatment with at-RA. RNA analysis of U-2OS cells ( Fig. 6A) confirmed that the endogenous BMP-2 transcript is upregulated twofold after 24 h of treatment with 1 mM at-RA, and this induction can be observed as early as 8 h ( Fig. 8). The retinoids 9-cis, 13-cis, and targretin were tested on C2C12, MC3T3, and U-2OS stable BMP-2 and BMP-4 promoter/ luciferase clones. All-trans and 13-cis RA, both of which bind only RAR, and 9-cis, which binds both RAR and RXR, consistently stimulated the BMP-2 promoter after 48 h of treatment. Although the response was minimal for 13-cis RA in the MC3T3 and U-2OS BMP-2 and BMP-4 stable clones, it consistently showed a low level of activation in numerous experiments ( Fig. 6B). Targretin, an RXR-specific ligand, had no effect on the BMP-2 promoter at 1 and 10 mM (data not shown), and when treated in combination with at-RA, no further increase in stimulation over that of RA alone was observed ( Fig. 6B).

Fig. 5. Human BMP-4 promoter/reporter deletion constructs and their basal activity in U-2OS cells. A schematic representation of fragments of the BMP-4 promoter cloned into pGL-3 Basic is shown. The start of transcription is indicated at +1, and potential promoter elements of interest are noted. The solid line indicates various lengths of the BMP-2 promoter sequence and the LUC box represents the luciferase gene. Darkened bars to the right indicate the fold increase of luciferase activity of each promoter fragment (after normalization to b-Gal activity) relative to the −2443 bp construct. Error bars indicate the standard error of fold change over seven separate experiments. Deletion constructs of the BMP-4 promoter fused to a luciferase reporter were transiently cotransfected with a CMV-b-Gal plasmid into U-2OS cells. After 16–24 h, cell lysates were measured for luciferase and b-Gal activity.

132

L.M. Helvering et al. / Gene 256 (2000) 123–138

Fig. 6. Retinoid effects on the –3073 BMP-2 and –2447 BMP-4 promoters in various stable cell lines. (A) U-2OS cells were treated with 5 mM at-RA or 0.01% DMSO for 24 or 48 h. Thirty micrograms of total RNA were gel-electrophoresed and then blotted onto filters that were subsequently hybridized with a human BMP-2 and GAPDH (not shown) cDNA radiolabeled probes. Membranes were washed and exposed to phosphorimaging plates. (B) Each cell line was treated for 48 h with 1 mM of each retinoid indicated or 0.01% DMSO. Luciferase activity was measured, and the values are expressed as the fold increase over the DMSO luciferase activity of the corresponding cell line. N=3 and error bars=S.E.

3.5. Stimulation of the promoters by PMA and synergism with retinoids Individual stable clones of the BMP-2 promoter construct in C2C12 and U-2OS cells and a stable U-2OS BMP-4 clone were examined for potential activity under several different treatment paradigms that have precedence for stimulating osteoblast activity. These included growth factors, nuclear receptor hormones, stimulators or inhibitors of signaling pathways, and bone matrix proteins (see Section 2). We found that the majority of these treatments had no effect on BMP-2 or BMP-4 promoter activity. However, the protein kinase A and protein kinase C pathways both regulated expression of the BMP-2 and BMP-4 promoters. The phorbol ester, PMA, which is known to activate the protein kinase C pathway, stimulated the BMP-2 promoter at nM concentrations ( Fig. 7). A consistent low level of stimulation (150% of control ) of the BMP-4 promoter was also observed when treated with PMA (data not shown). The stimulation by a combination of at-RA and PMA of the BMP-2 promoter was greater than the sum of either alone suggesting a synergistic effect of the two compounds (Fig. 7). RNA analysis confirmed the synergistic effect on the endogenous BMP-2 transcript in U-2OS within 8 h of treatment (Fig. 8). Activators of the kinase A pathway, however, consistently downregulated expression of BMP-2 in stable clones (Fig. 7) and

Fig. 7. BMP-2 promoter response to at-RA, PMA, and cAMP in a U-2OS Stable Clone. The 3073 BMP-2/luciferase stable U-2OS clones were treated with 0.01% DMSO, 5 mM AT-RA, 10 nM PMA, 5 mM at-RA and 10 nM PMA, and 1 mM cAMP. Luciferase activity was measure in each cell lysate after 24 h of treatment. Values are expressed as the fold change in luciferase activity after each treatment compared to the DMSO luciferase activity. N=4 and error bars=S.E.

in stable BMP-4 clones (data not shown). When cAMP was given in combination with at-RA/PMA, the activity of the BMP-2 promoter was similar to the activity of at-RA alone (data not shown).

L.M. Helvering et al. / Gene 256 (2000) 123–138

133

3.6. Protein/DNA-binding potential of an RRE in the BMP-2 promoter A potential RRE in the BMP-2 promoter (a consensus hexamer site flanked by imperfect hexamer sequences) identified between −1814 and −1787 and a consensus RRE found in the RARb promoter were tested for their ability to compete for binding to a consensus RRE using UMR106 rat osteosarcoma cell line nuclear extracts by EMSA. As shown in Fig. 9, a 100× molar excess of unlabeled RARb-RRE completely blocked binding to the labeled RARb-RRE, while 10× excess blocked the majority of binding. Competition with unlabeled 100× molar excess of the putative BMP2 RRE sequence only partially blocked binding and was nearly equivalent to competition with 1× unlabeled RARb-RRE. Additionally, when the BMP-2 RRE sequence was labeled and used as a probe with the same nuclear extracts, it was difficult to discern a distinct DNA/protein band (data not shown). Fig. 8. Northern analysis of U-2OS BMP-2 mRNA following treatment with at-RA. Untransfected U-2OS cells were treated with 0.01% DMSO, 5 mM at-RA, 10 nM PMA, or 5mM at-RA and 10 nM PMA for 8 h. Twenty micrograms of total RNA were denatured and dotblotted onto filters that were subsequently hybridized with a human BMP-2 or GAPDH cDNA radiolabeled probe. Membranes were washed and exposed to phosphorimaging plates. Images were quantitated on a phosphorimager, and the fold increase of BMP-2 message relative to the DMSO treated sample was calculated after normalizing for the expression of GAPDH in each sample.

3.7. Protein/DNA binding of Cbfa-1 in the BMP-2 and BMP-4 promoters An oligo corresponding to the putative CBE located between −2041 and −2035 in the BMP-2 promoter was found to effectively compete for binding to a consensus radiolabeled CBE from the osteocalcin promoter using UMR106 nuclear extracts (Fig. 10A). The

Fig. 9. Retinoic acid receptor element (RRE) competition with an element in the human BMP-2 promoter. Gel shift of nuclear extract from UMR106 cells was used as a source for binding to a labeled consensus RARb-RRE. Unlabeled competitor double-stranded oligos were used to compete for binding as indicated with 1×, 10×, or 100× the molar amount of the probe. Lane marked ‘probe’ did not contain any nuclear extract, while the ‘NC’ lane had no competitor DNA. Note the inability of Sp1 oligos to compete for binding.

134

L.M. Helvering et al. / Gene 256 (2000) 123–138

A

B

Fig. 10. Cbfa-1 binds to an element in the human BMP-2 promoter. (A) The gel shift shown used nuclear extract from UMR106 cells as a Cbfa-1 protein source for binding to the proximal Cbfa-1 element in the osteocalcin promoter. The top band is Cbfa-1 specific and binds to the proximal osteocalcin, human BMP-2 promoter (−2041), and human BMP-4 promoter (−848) equivalently as demonstrated by the ability of each unlabeled competitor (100× and 20× molar excess) to displace the labeled complex. ‘NC’ signifies no competitor DNA. Note that 100× molar concentration of oligos containing the consensus Sp1 binding site do not interfere with binding to the labeled CBE. (B) Nuclear extract from UMR106 cells binds to the labeled –2041 BMP-2 Cbfa-1 element. The first lane contains no nuclear extract. Note that the Cbfa-1 specific antibody decreases the complex and creates a supershift complex that is not seen by preimmune serum.

L.M. Helvering et al. / Gene 256 (2000) 123–138

putative BMP-2 CBE was also radiolabeled and shown to directly bind to a protein in UMR106 nuclear extracts with identical mobility to the labeled consensus CBE and was competed by 100× molar excess of unlabeled consensus CBE ( Fig. 10B). In addition, a peptide antibody specific for Cbfa-1 blocked some of this binding and created a supershift complex with the BMP-2 CBE (Fig. 10B). Further proof that the BMP-2 CBE complex found in UMR106 nuclear extracts included Cbfa-1 protein was demonstrated by the presence of a nuclear protein complex in COS cells of identical mobility expressing transfected Cbfa-1 (data not shown). Two of the putative BMP-4 CBEs (at −848 and −2426) were similarly analyzed and found to compete for binding Cbfa-1 protein with an affinity equivalent to that of the osteocalcin CBE (data shown in Fig. 10A for the −848 site).

135

A

3.8. Cbfa-1 activation of the BMP-2 and BMP-4 promoters A mouse Cbfa-1 expression plasmid co-transfected with the −2124 and −1819 bp BMP-2/b-gal and the −2443, −1917, −733, BMP-4/b-gal promoter constructs activated transcription of each promoter in a dose-dependent manner in UMR106 cells. These results were unexpected because a similar activation was achieved for each construct regardless of its length or inclusion of Cbfa-1 binding sites (Fig. 11).

4. Discussion This study has described the characterization of the human BMP-2 and BMP-4 promoters in several different cell lines. Despite some ambiguity in the literature regarding the start site of BMP-2 gene transcription, we have definitively determined by S1 analysis that the BMP-2 initiation of transcription lies approximately 870 bp upstream of the originally reported 5∞ end of the hBMP-2 cDNA ( Wozney et al., 1988) and 700 bp upstream of the most distal start site reported recently by Sugiura et al (Sugiura, 1999). Our data suggest that the BMP-2 gene does not have additional downstream sites of initiation in human osteoblasts, as recently described (Sugiura, 1999), or as reported in mouse osteoblasts ( Feng et al., 1994, 1997). The apparent discrepancy between our data and that of Sugiura et al. could be explained by the GC richness of the BMP-2 promoter. It has been well established that GC-rich templates can create secondary structures that are difficult for a polymerase to read through. Sugiura et al. solely utilized a PCR-based approach to determine the start of the BMP-2 gene and possibly obtained truncated products due to artifacts produced by inefficient readthrough of the GC-rich regions (Sugiura, 1999). We

B Fig. 11. Non-specific activation of the BMP-2 and BMP-4 deletion constructs by Cbfa-1. (A) The −2124 or −1819 bp BMP-2/bgal promoter construct was cotransfected with increasing amounts of a mouse Cbfa-1 expression plasmid into UMR106 cells. The mean relative light units representing at least four samples is shown. Error bars=S.E. (B) The −2443, −1919, or −733 bp BMP-4//bgal promoter construct was cotransfected with increasing amounts of a mouse Cbfa-1 expression plasmid into UMR106 cells. The mean relative light units representing at least four samples is shown. Error bars=S.E.

were able to fully protect the region where multiple start sites had been reported in two different osteoblast lines ( Fig. 2). Our conclusion that the human BMP-2 gene has a single start site of transcription is further supported by the fact that this site is also conserved across species in agreement with the distal start site identified in mouse and deer BMP-2 transcripts (Feng et al., 1994, 1997). Our analysis of the human BMP-4 start site was found to be consistent with a recent report by Shore et al. (Shore et al., 1998). Our study reports additional upstream promoter sequence for both the human BMP-2 and BMP-4 genes compared to sequences reported to date. Both promoters share potential transcription factor binding sites including those important in osteogenesis (Cbfa-1) and chondrogenesis (SOX-9) in addition to other common regulatory elements (Sp1, CRE, and RRE ). While basal activity of serial 5∞ deletions within each promoter did not vary more than 2.5-fold, the relative activity of the BMP-4 promoter constructs were five to 10-fold greater than those of BMP-2, suggesting greater gene expression

136

L.M. Helvering et al. / Gene 256 (2000) 123–138

of BMP-4 than BMP-2 in bone cell lines (Figs. 1, 3 and 4). The 5∞ untranslated region of the BMP-2 transcript was compared to the available sequences of the BMP-2 gene in mouse (Feng et al., 1994) and deer (Feng et al., 1997) and was found to be highly homologous (78– 86%). In addition, the sequence upstream of the start of transcription was compared to 1.96 kb of the published mouse sequence (Harris et al., 1996) and was found to be 73% homologous. Several potential regulatory elements (i.e. CBE, Sp1, TATA box, RRE, and a direct repeat) are absolutely conserved between the mouse and human sequences ( Fig. 1A). The human BMP-4 promoter region shares a high degree of homology to the mouse sequence and includes conservation of many possible regulatory elements (Shore et al., 1998). However, some of the CBEs are not conserved between the human and the mouse BMP-4 promoter. Retinoic acid had been reported to stimulate the expression of BMP-2 in various cell lines and in the developing chick limb (Rogers et al., 1992; Hatakeyama et al., 1996; Hogan, 1996; Rogers, 1996; Heller et al., 1999). Heller et al. recently showed modest stimulation of mouse BMP-2 promoter/reporter constructs with alltrans and 9-cis retinoic acid in mouse F9 cells and in yeast cells over-expressing the RARb, RARc, or RXRc receptors (Heller et al., 1999). In the yeast studies, it was demonstrated that individual agonists of RXR and RAR stimulated the BMP-2 promoter two-fold and a combination of agonists stimulated about 3.6-fold. To investigate the potential of retinoids to activate the BMP-2 and the BMP-4 promoters in bone cells, our study utilized stable and transiently transfected cell lines with promoter constructs that were then treated with various retinoids. Compounds known to bind RAR, such as at-RA and 13-cis RA, were able to stimulate the BMP-2 but not the BMP-4 promoter in transient transfections (data not shown). However, upon stable integration of the –2443 bp BMP-4 and −3073 bp BMP-2 constructs into U-2OS cells, at-RA, 13-cis RA and 9-cis RA were able to stimulate both promoters (Fig. 6B). Stimulation was not seen by an RXR-specific agonist, (targretin) in stable ( Fig. 6B) or transiently transfected cells (data not shown). Northern analysis confirmed that the endogenous BMP-2 transcript could also be induced by at-RA two-fold over controls in U-2OS cells ( Fig. 5A). However, the best consensus retinoid response element identified by sequence analysis in the BMP-2 promoter competed poorly with a consensus RRE in nuclear extracts (Fig. 9). These results suggest that the retinoic acid stimulation of BMP-2 transcription in osteoblastic and multipotential cells is not mediated by direct binding of RAR or RAR/RXR heterodimers but by an indirect mechanism yet to be identified. A potential RRE located at position −1802 flanked

by imperfect hexamer sequences was identified by promoter sequence analysis. The basal activity of the BMP2 promoter was substantially increased upon deletion of the region (Fig. 3), but there was no substantial decrease in retinoid response (data not shown). Since the transcription factor COUP-TF is also known to bind to RRE and is known to be a potent repressor of transcriptional activity (Cooney et al., 1993; Tsai and O’Malley, 1994), it is possible that this site is occupied by COUP-TF. Furthermore, this site was a weak competitor to a consensus RARb-RRE in an EMSA using osteoblast nuclear extract. COUP-TFI has previously been shown to inhibit BMP-4 promoter activity (Feng et al., 1995) and may also act as a regulator of the BMP-2 promoter. This study evaluated the ability of various osteogenic agents and modulators of various signaling pathways (see Section 2) to affect BMP-2 and BMP-4 promoter activity in osteoblasts. In addition to retinoid activation, the phorbol ester, PMA, was also able to stimulate the BMP-2 promoter ( Fig. 7) and, to a lesser degree, the BMP-4 promoter (data not shown). When retinoids were added in combination with PMA, a synergistic stimulatory effect on the BMP-2 promoter activity ( Fig. 7) and endogenous message was observed (Fig. 8). It could be concluded from these data that activation of the protein kinase C pathway by treatment with low concentrations of PMA is sufficient to increase BMP-2 and 4 transcription. Furthermore, when the kinase C and retinoid pathways are activated simultaneously, they are able to synergistically activate transcription of the BMP-2 gene. Conversely, when the protein kinase A pathway is activated by treatment with cAMP, promoter activity of BMP-2 and BMP-4 is reduced. Several hormones and growth factors exerted no effect on either promoter, suggesting that their osteotropic mechanism of action does not include transcriptional regulation of BMP-2 and BMP-4 genes. Cbfa-1 may also contribute to the regulation of the BMP-2 and BMP-4 promoters. BMPs have been shown to regulate Cbfa-1 expression (Ducy et al., 1997; Ito et al., 1998; Lee et al., 1999), and the presence of CBEs in the BMP promoters suggests reciprocal regulation. We were able to demonstrate that the CBEs in the BMP-2 and BMP-4 promoters identified by sequence analysis could specifically bind Cbfa-1 protein in vitro ( Fig. 10). The EMSA data presented in this study show that the consensus sites identified in the BMP-2 and BMP-4 promoters were able to compete for binding to a consensus osteocalcin CBE site. The in vivo interaction of Cbfa-1 with the osteocalcin promoter has been well established (Ducy and Karsenty, 1995; Banerjee et al., 1996; Otto et al., 1997). However, there was no change in BMP-2 or BMP-4 promoter activity when the CBEs were deleted in co-transfections assays ( Fig. 11). One possible explanation for the inability of the functional

L.M. Helvering et al. / Gene 256 (2000) 123–138

binding sites to influence transcription may be that the CBEs identified in the BMP-2 and BMP-4 promoters lie distal to the start of transcription (~2 and 1.2 kb, respectively) and may be inaccessible to the transcriptional machinery. For instance, Frendo et al. identified the proximal CBE as the most relevant for activation of osteocalcin transcription ( Frendo et al., 1998). It could also be concluded that Cbfa-1 is not relevant to the in vivo regulation of BMP-2 and BMP-4 by observation of the Cbfa-1 knockout animals. These animals, which are void of osteoblasts, retain the normal number and shapes of cartilagenous structures (Otto et al., 1997). One would expect abnormal patterning or morphology to occur if Cbfa-1 was necessary for proper regulatory control of the BMP-2 and BMP-4 genes in vivo. These data present a significant amount of new sequence of the human BMP-2 and BMP-4 promoters so that a comparison of their regulatory elements and a thorough analysis of their regulation in multiple cell lines could be accomplished. Our study provides evidence that retinoids likely indirectly influence BMP-2 and BMP-4 promoter activity in bone cells. Furthermore, retinoids given in combination with a protein kinase C activator such as PMA, synergistically activate BMP-2 gene expression. We have also presented evidence that while Cbfa-1, a bone-specific transcription factor has functional binding sites in both the BMP-2 and BMP-4 promoters, they may not influence the promoter in vivo. These data taken together help to further the understanding of the regulation of the BMP2 and BMP-4 genes.

References Banerjee, C., Hiebert, S.W., Stein, J.L., Lian, J.B., Stein, G.S., 1996. An AML-1 consensus sequence binds an osteoblast-specific complex and transcriptionally activates the osteocalcin gene. Proc. Natl. Acad. Sci. USA 93, 4968–4973. Church, G.M., Gilbert, W., 1984. Genomic sequencing. Proc. Natl. Acad. Sci. USA 81, 1991–1995. Cooney, A., Leng, X., Tsai, S., O’Malley, B.W., Tsai, M.-J., 1993. Multiple mechanisms of chicken ovalbumin upstream promoter transcription factor-dependent repression of transactivation by the vitamin D, thyroid hormone and retinoic acid receptors. J. Biol. Chem. 268, 4152–4160. Ducy, P., Karsenty, G., 1995. Two distinct osteoblast-specific cis-acting elements control expression of a mouse osteocalcin gene. Mol. Cell. Biol. 15, 1858–1869. Ducy, P., Zhang, R., Geoffroy, V., Ridall, A., Karsenty, G., 1997. Osf2/Cbfa1: a transcriptional activator of osteoblast differentiation. Cell 89, 747–754. Feng, J.Q., Harris, M.A., Ghosh-Choudhury, N., Feng, M., Mundy, G.R., Harris, S.E., 1994. Structure and sequence of mouse morphogenetic protein-2 gene (BMP-2): comparison of the structures and promoter regions of BMP-2 and BMP-4 genes. Biochim. Biophys. Acta 1218, 221–224. Feng, J.Q., Chen, D., Cooney, A.J., Tsai, M.J., Harris, M.A., Tsai, S.Y., Feng, M., Mundy, G.R., Harris, S.E., 1995. The mouse bone morphogenetic protein-4 gene. J. Biol. Chem. 270, 28364–28373.

137

Feng, J.Q., Chen, D., Ghoshchoudhury, N., Esparza, J., Mundy, G.R., Harris, S.E., 1997. Bone morphogenetic protein 2 transcripts in rapidly developing deer antler tissue contain an extended 5∞ noncoding region arising from a distal promoter. Biochim. Biophys. Acta — Gene Struct. Express. 1350, 47–52. Frendo, J.L., Xiao, G., Fuchs, S., Franceschi, R.T., Karsenty, G., Ducy, P., 1998. Functional hierarchy between two OSE2 elements in the control of osteocalcin gene expression in vivo. J. Biol. Chem. 273, 30509–30516. Geiser, A.G., Kim, S.J., Roberts, A.B., Sporn, M.B., 1991. Characterization of the mouse transforming growth factor-b1 promoter and activation by the Ha-ras oncogene. Mol. Cell. Biol. 11, 84–92. Hammonds, R.G., Schwall, R., Dudley, A., Berkemeier, L., Lai, C., Lee, J., Cunningham, N., Reddi, A.H., Wood, W.I., Mason, A.J., 1991. Bone-inducing activity of mature BMP-2b produced from a hybrid BMP-2a/2b precursor. Mol. Endocrinol. 5, 149–155. Harris, S.E., Mundy, G.R., Ghosh-Choudhury, N., Feng, J.Q., 1996. Methods and compositions for identifying osteogenic agents. World Intellectual Property Organization, International Publication number 96/38590. Geneva, Switzerland. Hatakeyama, S., Oharanemoto, Y., Kyakumoto, S., Satoh, M., 1996. Retinoic acid enhances expression of bone morphogenetic protein-2 in human adenocarcinoma cell-line (Hsg-S8). Biochem. Mol. Biol. Int. 38, 01235–01243. Heller, L.C., Li, Y., Abrams, K.L., Rogers, M.B., 1999. Transcriptional regulation of the Bmp2 gene. J. Biol. Chem. 274, 1394–1400. Hogan, B.L.M., 1996. Bone morphogenetic proteins in development. Curr. Opin. Genet. Dev. 6, 432–438. Ito, Y., Noda, M., Tsuji, K., 1998. Expression of the PEBP2 alpha A/ AML3/CBFA1 gene is regulated by BMP4/7 heterodimer and its overexpression suppresses type I collagen and osteocalcin gene expression in osteoblastic and nonosteoblastic mesenchymal cells. Bone 22, 87–92. Krishnan, V., Elberg, G., Tsai, M.J., Tsai, S.Y., 1997. Identification of a novel sonic hedgehog response element in the chicken ovalbumin upstream promoter-transcription factor II promoter. Mol. Endocrinol. 11, 1458–1466. Lania, L., Majello, B., de, Luca P., 1997. Transcriptional regulation by the Sp family proteins. Int. J. Biochem. Cell Biol. 29, 1313–1323. Lee, M.H., Javed, A., Kim, H.J., Shin, H.I., Gutierrez, S., Choi, J.Y., Rosen, V., Stein, J.L., Wijnen, A.J., Stein, G., Lian, J., Ryoo, H.M., 1999. Transient upregulation of Cbfa1 in response to bone morphogenetic protein-2 and transforming growth factor B1 in C2C12 myogenic cells coincides with suppression of the myogenic phenotype but is not sufficient for osteoblast differentiation. J. Cell. Biochem. 73, 114–125. Lyons, K.M., Pelton, R.W., Hogan, B.L.M., 1990. Organogenesis and pattern formation in the mouse: RNA distribution patterns suggest a role for Bone Morphogenetic Protein-2A (BMP2A). Development 109, 833–844. Mangelsdorf, D.J., Evans, R.M., 1995. The RXR heterodimers and orphan receptors. Cell 83, 841–850. Maniatis, T., Fritsch, E.F., Sambrook, J., 1982. Molecular Cloning. A Laboratory Manual. first ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Massague, J., Weis-Garcia, F., 1996. Serine/threonine kinase receptors: Mediators of transforming growth factor beta family signals. Cancer Surv. 27, 41–64. Metz, A., Knochel, S., Buchler, P., Koster, M., Knochel, W., 1998. Structural and functional analysis of the BMP-4 promoter in early embryos of Xenopus laevis. Mech. Dev. 74, 29–39. Montminy, M., 1997. Transcriptional regulation by cyclic AMP. Ann. Rev. Biochem. 66, 807–822. Ng, L.J., Wheatly, S., Muscat, G.E., Conway-Campbell, J., Bowles, J., Wright, E., Bell, D.M., Tam, P.P., Cheah, K.S., Koopman, P., 1997. SOX9 binds DNA, activates transcription, and coexpresses

138

L.M. Helvering et al. / Gene 256 (2000) 123–138

with type II collagen during chondrogenesis in the mouse. Dev. Biol. 183, 108–121. Otto, F., Thornell, A.P., Crompton, T., Denzel, A., Gilmour, K.C., Rosewell, I.R., Stamp, G.W.H., Beddington, R.S.P., Mundlos, S., Olsen, B.R., Selby, P.B., Owen, M.J., 1997. Cbfa1, a candidate gene for Cleidocranial Dysplasia Syndrome, is essential for osteoblast differentiation and bone development. Cell 89, 765–771. Reddi, A.H., Huggins, C., 1972. Biochemical sequences in the transformation of normal fibroblasts in adolescent rats. Proc. Natl. Acad. Sci. USA 69, 1601–1605. Reddi, A.H., 1981. Cell biology and biochemistry of endochondral bone formation. Collagen Rel. Res. 1, 209–220. Riley, E.H., Lane, J.M., Urist, M.R., Lyons, K.M., Lieberman, J.R., 1996. Bone morphogenetic protein-2: Biology and applications. Clin. Ortop. Rel. Res. 324, 39–46. Rogers, M.B., Rosen, V., Wozney, J.M., Gudas, L.J., 1992. Bone morphogenetic proteins-2 and -4 are involved in the retinoic acidinduced differentiation of embryonal carcinoma cells. Mol. Biol. Cell 3, 189–196. Rogers, M.B., 1996. Receptor-selective retinoids implicate retinoic acid receptor-alpha and receptor-gamma in the regulation of Bmp-2 and Bmp-4 in F9 embryonal carcinoma-cells. Cell Growth Diff. 7, 115–122. Shore, E.M., Xu, M.Q., Shah, P.B., Janoff, H.B., Hahn, G.V., Deardorff, M.A., Sovinsky, L., Spinner, N.B., Zasloff, M.A., Wozney, J.M., Kaplan, F.S., 1998. The human bone morphogenetic protein

4 (BMP-4) gene: Molecular structure and transcriptional regulation. Calc. Tiss. Int. 63, 221–229. Sugiura, T., 1999. Cloning and functional characterization of the 5∞-flanking region of the human bone morphogenetic protein-2 gene. Biochem. J. 338, 433–440. Tsai, M.J., O’Malley, B.W., 1994. Molecular mechanisms of steroid/ thyroid receptor superfamily members. Annu. Rev. Biochem. 63, 451–486. Urist, M.R., 1965. Bone: Formation by autoinduction. Science 150, 893–899. Urist, M.R., Lietze, A., Mizutani, H., Takagi, K., Triffitt, J.T., Amstutz, J., DeLange, R., Termine, J., Finerman, G.A.M., 1982. A bovine low molecular mass bone morphogenetic protein (BMP) fraction. Clin. Orthop. 162, 219–232. van den Wijngaard, A., van Kraay, M., van Zoelen, J.J., Olijve, W., Boersma, C.J.C., 1996. Genomic organization of the human bone morphogenetic protein-4 gene: Molecular basis for multiple transcripts. Biochem. Biophys. Res. Commun. 219, 789–794. Wang, E.A., Rosen, V.D, Alessandro, J.S., Bauduy, M., Cordes, P., Harada, T., Israel, D.I., Hewick, R.M., Kerns, K.M., LaPan, P., Luxenberg, D.H., McQuid, D., Moutsatsos, I.K., Nove, J., Wozney, J.M., 1990. Recombinant human bone morphogenetic protein induces bone formation. Proc. Natl. Acad. Sci. USA 87, 2220–2224. Wozney, J.M., Rosen, V., Celeste, A.J., Mitsock, L.M., Whitters, M.J., Kriz, R.W., Hewick, R.M., Wang, E.A., 1988. Novel regulators of bone formation: Molecular clones and activities. Science 242, 1528–1534.