Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Cbfa1 Contributes to the Osteoblast-specific Expression of type I collagen Genes

Cbfa1 Contributes to the Osteoblast-specific Expression of type I collagen Genes THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 276, No. 10, Issue of March 9, pp. 7101–7107, 2001 © 2001 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Cbfa1 Contributes to the Osteoblast-specific Expression of type I collagen Genes* Received for publication, July 13, 2000, and in revised form, November 29, 2000 Published, JBC Papers in Press, December 5, 2000, DOI 10.1074/jbc.M006215200 Britt Kern, Jianhe Shen, Michael Starbuck, and Gerard Karsenty‡ From the Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, Texas 77030 Type I collagen is composed of two chains, a1(I) and throughout life (8), suggesting that the factor(s) controlling a2(I), encoded by two distinct genes, the a1(I) and a2(I) their expression in these cells could also control osteoblast collagen genes, that are highly expressed in osteoblasts. differentiation and function. Another possibility, which is not a1(I) and a2(I) collagen In most physiological situations, exclusive of the previous one, is that different transcription expression is coregulated, suggesting that identical factors may control their expression in osteoblasts at various transcription factors control their expression. Here, we stages of development and of postnatal life. Thus, the elucida- studied the role of Cbfa1, an osteoblast-specific tran- tion of the molecular mechanisms controlling a1(I) and a2(I) a1(I) and a2(I) collagen scription factor, in the control of collagen gene expression in osteoblasts is of critical importance expression in osteoblasts. A consensus Cbfa1-binding in understanding how osteoblast differentiation and, thereby, site, termed OSE2, is present at the same location in the bone matrix deposition by differentiated osteoblasts is regu- a1(I) collagen promoter at approximately 21347 base lated. Ultimately, these studies may shed light on the patho- pairs (bp) of the rat, mouse, and human genes. Cbfa1 can genesis of genetically acquired bone diseases and help design bind to this site, as demonstrated by electrophoretic appropriate therapies for some of these diseases. mobility shift assay (EMSA) and supershift experiments The critical role that Cbfa1, a Runt-related osteoblast-spe- a1(I) using an anti-Cbfa1 antibody. Mutagenesis of the a1(I) cific transcription factor, plays in osteoblast differentiation and collagen OSE2 at 21347 bp reduced the activity of a function has been demonstrated in mouse and in human using collagen promoter fragment 2- to 3-fold. Moreover, mul- timers of this OSE2 at 21347bp confer osteoblast-spe- both molecular and genetic approaches (9 –14). Cbfa1 was iden- a1(I) collagen promoter frag- cific activity to a minimum tified as a key regulator of osteoblast-specific gene expression ment in DNA transfection experiments as well as in through its binding to the OSE2 element of the mouse Osteo- transgenic mice. An additional Cbfa1-binding element is calcin genes 1 and 2 (OG1 and OG2) (9) and other genes a1(I) collagen promoter of mouse, rat, and present in the expressed in osteoblasts. The early and cell-specific expression human at approximately position 2372. This site binds of this gene together with its biological role in vivo as a factor Cbfa1 only weakly and does not act as a cis-acting acti- required for osteoblast differentiation (9 –11), indicate that vator of transcription when tested in DNA transfection Cbfa1 must control the expression of multiple target genes that a1(I) collagen, the mouse a2(I) experiments. Similar to are expressed earlier than Osteocalcin. Conceivably, these tar- collagen gene contains multiple OSE2 sites, of which one get genes could include the a1(I) and a2(I) collagen genes that is conserved across multiple species. In EMSA, Cbfa1 are expressed early during development. This hypothesis was a2(I) OSE2 ele- binds to this site and multimers of this confirmed indirectly by the observation that expression of a ment confer osteoblast-specific activity to the minimum dominant negative form of Cbfa1 in differentiated osteoblasts a1(I) collagen promoter in DNA transfection experi- leads to a decrease in expression of the type I collagen genes in ments. Thus, our results suggest that Cbfa1 is one of the vivo (14). To date, no osteoblast-specific cis-acting elements to positive regulators of the osteoblast-specific expression of both type I collagen genes. which Cbfa1 may bind have been identified in these genes. Two groups have extensively studied the regulation of ex- pression of the a1(I) collagen gene in osteoblasts and have Type I collagen is the most abundant protein of the bone identified a region in the promoter of the rat and mouse a1(I) extracellular matrix, accounting for 90% of the matrix protein collagen gene that plays an important role in this regulation of content (1). It is a heterotrimer made of two a1(I) chains and expression (15, 16). The sequence of this region bears no ho- one a2(I) chain (2). The a1(I) and a2(I) chains are encoded by mology to a Cbfa1-binding site, and several homeobox-contain- two distinct genes that are expressed most highly in two cell ing proteins can bind to this sequence and affect a1(I) collagen types: the fibroblast and the osteoblast. Moreover, the expres- expression. However, a cell-specific transcription factor bind- sion of these two genes is often regulated by identical transcrip- ing to this region has not yet been identified. Moreover, no tion factors (3–7). The type I collagen genes are expressed in osteoblast-specific cis-acting element has yet been identified in osteoblastic cells at all stages during development and the a2(I) collagen promoter. Given the large size of these genes and their expression at multiple stages of osteoblast differen- tiation, it is likely that several distinct osteoblast-specific cis- * This work was supported by National Institutes of Health Grants acting elements, besides those already described (15, 16), con- R01AR45548 and R01DE11290, March of Dimes Grant F-198-0082, and tribute to the expression of the type I collagen genes in a grant from Eli Lilly and Co. (to G. K.). The costs of publication of this article were defrayed in part by the payment of page charges. This osteoblast progenitors and/or in fully differentiated osteoblasts. article must therefore be hereby marked “advertisement” in accordance Consistent with this hypothesis, we noticed the existence of two with 18 U.S.C. Section 1734 solely to indicate this fact. Cbfa1-binding sites (OSE2s) in the mouse a1(I) collagen pro- ‡ To whom correspondence should be addressed: Dept. of Molecular moter and one OSE2 in the mouse a2(I) collagen gene that are and Human Genetics, Baylor College of Medicine, Houston, TX 77030. Tel.: 713-798-5489; Fax: 713-798-1465; E-mail [email protected]. conserved among multiple species. The functional importance This paper is available on line at http://www.jbc.org 7101 This is an Open Access article under the CC BY license. 7102 Cbfa1 Regulation of Type I Collagen of these sites has never been studied before. Here we present evidence suggesting that Cbfa1 is one of the factors controlling osteoblast-specific expression of both type I collagen genes. EXPERIMENTAL PROCEDURES DNA Constructs—For DNA transfection and generation of trans- genic mice, multimers of double-stranded oligonucleotides (see Table I) were cloned into the SmaI site upstream of a chimeric pK1-luc reporter plasmid containing the a1(I) collagen 286 minimal promoter (4) fused to a luciferase gene. The expression plasmid was pCMV-Osf2 (9). Site-specific mutations were created in the intact promoter by using polymerase chain reaction-directed mutagenesis (17) on a construct containing 2.4 kb of the a1(I) collagen promoter upstream of a lucifer- ase reporter gene. Sole presence of the desired mutations was verified by sequencing. Cell Culture and DNA Transfection—COS7, NIH 3T3, HeLa, C2C12, and 10T1/2 cells were cultured in Dulbecco’s minimal essential medium (Life Technologies, Inc.), 10% fetal bovine serum (Life Technologies, Inc.). ROS 17/2.8 cells were cultured in Dulbecco’s minimal essential medium-F12 medium (Life Technologies, Inc.), 10% fetal bovine serum (Life Technologies, Inc.). Twenty hours before transfection, cells were plated at a density of 5 3 10 cells/dish and allowed to grow under normal culture conditions. For cotransfection experiments, we used 5 mg of Cbfa1 expression vector or empty vector, 5 mg of reporter plasmid or empty vector, and 2 mgofpSVb-gal vector using the calcium phos- phate coprecipitation procedure (17). Transfection conditions were iden- tical to those used in the cotransfection experiments, except that 5 mgof reporter plasmid were used. Twenty hours following transfection, cells were washed in phosphate-buffered saline and incubated in medium an additional 24 h. C2C12 cells were changed to media containing 10% horse serum (Life Technologies, Inc.) and allowed to incubate for 48 h. Cells were collected by scraping into 0.25 M Tris-HCl, pH 7.8, and lysed FIG.1. Type I collagen OSE2 sites are well conserved across by three freeze-thaw cycles. b-Galactosidase and luciferase assays were species. A, three OSE2 sites are present in the mouse a1(I) collagen carried out as described previously (18). b-Galactosidase assay results promoter, but only the a1A site and the a1C site are conserved across were used to normalize the luciferase assay results for transfection species. B, diagram of the a1(I) and a2(I) collagen promoters and the efficiency. All DNA transfection experiments were repeated at least OSE2 sites of each. The arrow indicates the start site of transcription three times in triplicate. (11). The position indicated for each element is relative to the start site Electrophoretic Mobility Shift Assays—Nuclear extract from ROS of transcription. C, a2(I) collagen OSE2 sequence is conserved across 17/2.8 cells, primary osteoblasts, and other tissues were prepared as vertebrate species for which the sequence is available: h, human; m, described previously (19) from 4-day-old wild-type mice and stored at mouse; r, rat; c, chicken; C.f., Canis familiaris; B.t., Bos taurus; R.c., 280 °C until use. Glutathione S-transferase-Cbfa1 was purified from Rana catesbeiana. transformed Escherichia coli bacteria using glutathione beads as de- scribed previously (17). Double-stranded oligonucleotides (see Table I) expressing the p4a1AB-luc construct were analyzed as follows: Organs were end-labeled and purified as previously described (19). 5 fmol of from 4-week-old F1 animals were dissected and homogenized on ice in labeled oligonucleotide was incubated with 7 mg of ROS 17/2.8 nuclear a buffer containing 100 mM potassium phosphate (pH 7.8) and 1 mM extract or 0.1 mg of recombinant Cbfa1 protein. a2(I) collagen EMSA dithiothreitol (DTT). Protein homogenates were centrifuged, and super- experiments used twice the amount of extract or protein. The incuba- natants were assayed for luciferase activity according to standard pro- tion mix for nuclear extract binding assays consisted of binding buffer cedures (19). Protein levels were measured using the Bio-Rad protein (100 mM Tris-HCl, pH 7.5, 200 mM NaCl, 4 mM EDTA, 2 mM DTT, 0.2% assay. Relative luciferase activities were expressed as luciferase light Nonidet P-40, 10% glycerol, 5 mg/ml leupeptin, 5 mg/ml pepstatin) (20), units per 100 mg of protein expressed as a percentage of the activity in 2 mg of poly(dI-dC), and 0.5 fmol of single-stranded bottom strand bone. oligonucleotide. Incubation took place at room temperature for 5 min. RESULTS Supershift experiments were carried out as described above, except that the ROS17/2.8 nuclear extract was preincubated for 10 min at room The a1(I) collagen Promoter Contains Two Conserved OSE2 temperature with an antibody against Cbfa1 in binding buffer prior to Sites—DNA sequence inspection identified three potential their incubation with the labeled oligonucleotide for 10 min at room OSE2 sites in the promoter of the murine a1(I) collagen gene. temperature. Two are located side-by-side at positions 21347 and 21338 in For recombinant protein binding assays, the incubation mix con- the mouse gene (Fig. 1A), and a third site is found at 2372 in sisted of binding buffer (20 mM Tris-HCl, pH 8.0, 10 mM NaCl, 3 mM EGTA, 5 mM DTT, 0.05% Nonidet P-40) and 1 mg of bovine serum the mouse gene (Fig. 1A). These potential OSE2 sites were albumin. Incubation took place at room temperature for 5 min, followed termed a1A, a1B, and a1C, respectively (Fig. 1B). The presence by the addition of 1 ml of loading buffer (2 mM Tris-HCl, pH 8.0, 5% of a1A and a1C, but not of a1B, at approximately the same glycerol, 0.025% xylene cyanol, 0.025% bromphenol blue). location in the a1(I) collagen promoter sequences of rat and The reactions were run on 5% polyacrylamide gel, 0.253 TBE (89 mM human (Fig. 1A) suggested a biological role for these sites and Tris base, 89 mM boric acid, 2 mM EDTA, pH 8.0) for 90 min at 160 V. led us to study these two regions. To determine whether Cbfa1 The gels were then dried and exposed to film at 280 °C. Generation and Analysis of Transgenic Mice—The plasmids de- could bind to the OSE2-like sequences in the mouse a1(I) scribed above containing multimers of the a1AB or a1mutAmutB oli- collagen promoter, we generated double-stranded oligonucleo- gonucleotides were digested, and the insert was purified by two rounds tides to be used in DNA binding assays (Table I). One of them, of agarose gel electrophoresis. Linear DNA inserts were injected into called a1AB, contains the a1A and a1B sites and their sur- the pronuclei of fertilized B6D2F1 (Charles River Laboratory) mouse rounding sequences. A second one, termed a1wtAmutB, con- eggs, which were reimplanted in the oviduct of pseudo-pregnant CD1 tains the wild-type a1A site and a mutated a1B site. A third foster mothers (Jackson Laboratories). Transgenic animals were iden- tified by Southern blots of tail genomic DNA. The transgenic mice oligonucleotide, called a1mutAwtB, carries a mutated a1A se- quence and a wild-type a1B sequence. A fourth oligonucleotide, a1mutAmutB, contains mutations in both the a1A and the a1B The abbreviations used are: kb, kilobase(s); DTT, dithiothreitol; EMSA, electrophoretic mobility shift assay; bp, base pair(s). sites. Two other oligonucleotides, termed a1C and a1mutC, Cbfa1 Regulation of Type I Collagen 7103 TABLE I The a1A Site Acts as an Osteoblast-specific Activator of Tran- Oligonucleotides used in this study scription in Tissue Culture Experiments and in Vivo—We fur- Boldface letters indicate mutated nucleotides. ther addressed the functional relevance of the a1A and a1B Oligonucleotide sites using two additional approaches. First, to examine the Sequence name effect of the a1A site on activity of a 2.4-kb promoter fragment, a1AB 5 9-GATCCCTTCCCACACCACCCACACAGA-3 9 a site-specific mutation was generated in the a1A site via a1wtAmutB 5 9-GATCCCTTC GA ACACCACCCACACAGA-3 9 polymerase chain reaction and introduced into a 2.4-kb a1(I) a1mutAwtB 5 9-GATCCCTTCCCACACCAC GA ACACAGA-3 9 collagen promoter-luc chimeric gene. These 2-bp mutations a1mutAmutB 5 9-GATCCCTTC GA ACACCAC GA ACACAGA-3 9 resulted in a 54% decrease in promoter activity when tested in a1C 5 9-GATCTACTGAGGGCCCAGCCACACTCCA-3 9 a1mutC 5 9-GATCTACTGAGGGCCCAG GA ACACTCCA-3 9 DNA transfection experiments in ROS17/2.8 cells (Fig. 3A). a2A 5 9-GATCCCTTTGTGGATACGCGGACTTTGA-3 9 The same mutations did not reduce the activity of this 2.4-kb a2mutA 5 9-GATCCCTTTGT TC ATACGCGGACTTTGA-3 9 a1(I) collagen promoter-luc chimeric gene in other cell lines of nonosteoblastic nature of mesenchymal and nonmesenchymal origin (Fig. 3A). This result further suggests that this cis-acting were generated to test the binding activity of the a1C site. The element is active only in osteoblasts. Mutations in the a1B site mutations introduced into all the oligonucleotides mentioned did not affect the activity of the 2.4-kb fragment of the a1(I) above have previously been shown to abolish binding of nuclear collagen promoter used in this study (Fig. 3A). extract or recombinant Cbfa1 to the OSE2 sequence present in Second, in DNA cotransfection assays performed in COS7 the Osteocalcin (OG2) promoter (OSE2 ) (19). These double- OG2 cells, a cell line that does not express Cbfa1 (21), exogenous stranded oligonucleotides were then used as probes in electro- Cbfa1 transactivated a construct containing four copies of the phoretic mobility shift assays (EMSA) using either ROS 17/2.8 a1AB oligonucleotide fused to a minimal a1(I) collagen promot- nuclear extract or recombinant Cbfa1 as a source of protein. er-luc chimeric gene, p4a1AB-luc (Fig. 3B). Indeed, cotransfec- Cbfa1 Binds to an OSE2 Site in the Mouse a1(I) collagen tion of p4a1AB-luc with a recombinant Cbfa1-expressing vector Promoter—The complex formed upon incubation of ROS 17/2.8 resulted in a 180-fold activation, whereas an empty expression nuclear extract with a1AB migrated at the same location as the vector had no effect. This level of transactivation is similar to complex formed upon incubation of ROS 17/2.8 nuclear extract what we observed when using a vector containing six copies of with the OSE2 oligonucleotide (Fig. 2A, lanes 1 and 2), OG2 the OSE2 as a reporter (Fig. 3B). The minimal a1(I) colla- OG2 although it was of weaker intensity. In contrast, no protein- gen promoter fragment has virtually no transactivation ability DNA complex was observed when using a1mutAmutB as a on its own (4). When using a vector containing multimers of probe (Fig. 2A, lane 5). a1wtAmutB oligonucleotides cloned upstream of the minimal To determine which of the two upstream OSE2 sites was a1(I) collagen promoter fragment, we observed a 120-fold in- binding to this factor, we used a1wtAmutB oligonucleotide or crease in luciferase activity, indicating that the a1A site is the a1mutAwtB oligonucleotide as probes in EMSA. Incubation of main contributor to the transactivating function of this region labeled a1wtAmutB with ROS 17/2.8 nuclear extract generated of the a1(I) collagen promoter. This is consistent with the a protein-DNA complex that had the same mobility as the one observation that only the a1A site is able to bind Cbfa1 in vitro. observed when using a1AB oligonucleotide as a probe and was The slight decrease in activity seen with the loss of the a1B site of stronger intensity (Fig. 2A, lane 3). In contrast, when using may indicate a synergistic effect of these two sites in this type a1mutAwtB oligonucleotide as a probe, we observed only a of experiment. The activity of constructs containing four copies weak binding of ROS 17/2.8 nuclear extract to the DNA (Fig. of a1mutAmutB, or of a1mutAwtB upstream of the minimal 2A, lane 4). a1(I) collagen promoter-luc chimeric gene (Fig. 3B) could not be To determine whether this osteoblast-specific factor binding increased upon cotransfection with the Cbfa1-expressing vec- to the a1A OSE2-like element was indeed Cbfa1, we performed tor, thus demonstrating the specificity of the effect observed. In three types of experiments. First, we asked whether the factor this set of experiments, we used multimers of the a1(I) collagen present in ROS 17/2.8 nuclear extract and binding to the a1A OSE2 sites, because Cbfa1 does not transactivate the a1(I) site was expressed only in osteoblasts. For that purpose, we collagen promoter fragment in this type of assay. This is a prepared nuclear extract from primary osteoblasts and several consistent feature of Cbfa1 biology, indeed, we observed weak other tissues and used them in EMSA. As shown in Fig. 2B, the transactivation of the OG2 promoter when cotransfected with factor binding to the a1A oligonucleotide was present only in Cbfa1 (9), compared with the strong transactivating effect of primary osteoblast nuclear extract and not in nuclear extract of Cbfa1 observed when using 6OSE2-luc (9). other tissues. Next we performed supershift experiments using Third, we asked whether these OSE2 sites could confer bone- anti-Cbfa1 antibody or a nonspecific antiserum. Incubation of specific expression to a reporter gene in vivo. For that purpose, we the nuclear extract with an antibody against Cbfa1 prior to generated transgenic mice containing two of the constructs used in addition of labeled a1AB oligonucleotide led to the formation of the above transfections, p4a1AB-luc and p4a1mutAmutB-luc. In a second protein-DNA complex of slower mobility (Fig. 2C, lane transgenic mice harboring p4a1AB-luc, luciferase activity could be 2), whereas a nonspecific serum had no effect (Fig. 2C, lane 1), detected in bone but neither in other tissues expressing type I demonstrating that the protein-DNA complex formed upon in- collagen, nor in tissues which do not express type I collagen (Fig. cubation of the labeled a1AB with ROS 17/2.8 nuclear extract 3C). As expected, given their respective sizes, the expression of contains Cbfa1, because this antibody is specific for Cbfa1 (14). p4a1AB-luc was considerably lower than that of a1(I) collagen Third, we asked whether recombinant Cbfa1 could bind to a1AB but not to a1mutAmutB (Fig. 2D, lanes 1 and 4). Incu- (data not shown). The p4a1mutAmutB-luc construct was not ex- pressed in bone or any other tissue (Fig. 3C). Taken together with bation of labeled a1wtAmutB oligonucleotide with recombinant Cbfa1 resulted in the formation of a protein-DNA complex (Fig. the results of the mutagenesis of the a1A site and of the a1B site, these results indicate that Cbfa1 contributes to the expression of 2D, lane 2), whereas incubations using labeled a1mutAwtB oligonucleotide did not (Fig. 2D, lane 3). Taken together, these a1(I) collagen in osteoblasts through the a1A site. results indicate that the a1A site, a site conserved in multiple Cbfa1 Binds to the OSE2 Site Located at 2372 bp in the species, is the major binding site for Cbfa1 in this region of the Mouse a1(I) collagen Promoter—As mentioned at the beginning a1(I) collagen promoter. of “Results,” there is a third OSE2 sequence in the a1(I) colla- 7104 Cbfa1 Regulation of Type I Collagen FIG.2. Cbfa1 binds to the a1A site in the a1(I) collagen promoter. DNA binding was analyzed by EMSA. A, la- beled oligonucleotides OSE2 (lane 1), OG2 a1AB (lane 2), a1wtAmutB (lane 3), a1mutAwtB (lane 4), and a1mutAmutB (lane 5) were incubated with ROS 17/2.8 nuclear extract. The arrow indicates the complex of interest. B, labeled a1AB was used as a probe and incubated with nu- clear extract from primary osteoblasts (lane 1), brain (lane 2), kidney (lane 3), lung (lane 4), muscle (lane 5), and spleen (lane 6). The arrow indicates the complex containing Cbfa1. C, supershift EMSA was performed using an antiserum against Cbfa1 (lane 2), or nonspecific an- tiserum (lane 1) using a1AB as a probe. The arrow indicates the complex of lower mobility observed after incubation with the anti-Cbfa1 antibody. D, EMSA using recombinant Cbfa1 as a source of protein and a1AB (lane 1), a1wtAmutB (lane 2), a1mutAwtB (lane 3), and a1mutAmutB (lane 4) oligonucleotides as probes. gen promoter, located at 2372bp in mouse (site a1C, Fig. 1A). weakly, Cbfa1, because Cbfa1 was able to bind to a1C oligonu- This site is also conserved across species, and we first asked cleotide but not to a1mutC oligonucleotide (Fig. 4B, lanes 1 and whether this OSE2-like sequence could be bound by Cbfa1 in 2). These results indicate that Cbfa1 is able to bind only weakly EMSA. Labeled a1C oligonucleotides were incubated with ROS to the a1C site in the mouse a1(I) collagen promoter, suggest- 17/2.8 nuclear extract as described above, leading to the for- ing that this OSE2 site may not play a critical role. To test this mation of a protein-DNA complex that migrated at the same hypothesis, we cloned four copies of wild-type or mutated a1C location as that formed upon incubation of ROS 17/2.8 nuclear oligonucleotides upstream of the minimal a1(I) collagen pro- extract with labeled OSE2 (Fig. 4A, lanes 1 and 3). How- moter fragment-luciferase chimeric gene used in Fig. 3A.As OG2 ever, the protein-DNA complex was of weak intensity compared seen in Fig. 4C, in DNA transfection experiments in COS7 with that we observed when using OG2 or even a1wtAmutB cells, neither wild-type nor mutant a1C constructs could in- OG2 oligonucleotides as probes (Fig. 4A, lanes 1 and 2). No complex crease the activity of this reporter gene upon cotransfection of this size was observed after incubation of labeled a1mutC with a Cbfa1 expression vector. This lack of an overt role for the with ROS 17/2.8 extract (Fig. 4A, lane 3). a1C site is consistent with the rather poor binding of Cbfa1 to To show that Cbfa1 was part of this protein-DNA complex, this site. Taken together, these results indicate that the a1C supershift experiments were performed using an anti-Cbfa1 site is not a critical cis-acting element in controlling the osteo- antibody, a labeled a1C oligonucleotide, and ROS17/2.8 nu- blast-specific expression of the mouse a1(I) collagen gene. clear extract as a source of protein. The incubation of ROS Cbfa1 Binds to an OSE2 Site in the Mouse a2(I) collagen 17/2.8 nuclear extract with an antibody against Cbfa1 prior to Gene—Because a1(I) and a2(I) collagen genes are often coregu- the addition of labeled oligonucleotide led to the formation of a lated, we next asked whether Cbfa1 was also regulating the slower mobility complex. This complex was specific, because it expression of a2(I) collagen. Sequence analysis of the a2(I) was not observed when using a nonspecific antiserum (Fig. 4A, collagen promoter uncovered the existence of several potential lanes 5 and 6). EMSA experiments using recombinant Cbfa1 OSE2 sites. Only one of these, located in the first exon of the provided further evidence that this site could bind, albeit gene, is present at the same location in the a2(I) collagen gene Cbfa1 Regulation of Type I Collagen 7105 FIG.4. The a1C site is bound by Cbfa1 but is unable to activate a minimal a1(I) collagen promoter. A, EMSA was performed using FIG.3. The a1A site can activate transcription and is required labeled OSE2 (lane 1), a1C (lane 3), or a1mutC (lane 4) oligonucleo- OG2 for the full activity of the a1(I) collagen promoter. A, constructs tides and ROS 17/2.8 nuclear extract as a source of protein. The arrow containing 2.4 kb of the a1(I) collagen promoter with or without muta- indicates complex of interest. Supershift EMSA was performed using an tions were transfected into several cell types. a1AB (black bars), a1mu- antiserum against Cbfa1 (lane 6) or nonspecific antiserum (lane 5). The tAwtB (gray bars), and a1wtAmutB (white bars) promoters were cloned arrow indicates the complex of lower mobility formed after incubation upstream of a luciferase reporter gene and used in DNA transfection with the anti-Cbfa1 antibody. B, EMSA using recombinant Cbfa1 as a experiments in ROS 17/2.8 cells, C2C12 cells, 10T1/2 cells, NIH 3T3 source of protein and a1C (lane 1)or a1mutC (lane 2) oligonucleotides cells, and HeLa cells. Values represent percentage activity compared as a probe. C, multimers of a1C and a1mutC oligonucleotides were with the wild-type promoter. a1mutAwtB displays lower transcrip- placed upstream of a minimal a1(I) collagen promoter. These constructs tional activity only in ROS 17/2.8 cells. B, multimers of the oligonucleo- were used in cotransfection assays in COS7 cells with a Cbfa1-express- tides used for EMSA were placed upstream of a minimal a1(I) collagen ing vector (dark bars) or an empty vector (open bars). Values represent promoter fused to a luciferase reporter gene. These constructs were -fold activation in relation to an empty reporter vector and are an transfected into COS7 cells in the presence of a recombinant Cbfa1 average of at least three experiments done in triplicate. expression construct (dark bars), or an empty vector (open bars). Values represent -fold activation in relation to an empty reporter vector and are the average of at least three experiments done in triplicate. C, wild-type (dark bars) and double-mutant (open bars) constructs from DNA transfection experiments were used to generate transgenic mice. nuclear extract, we also used recombinant Cbfa1 protein in Luciferase activity per 100 mg of protein was determined for several EMSA. The incubation of recombinant Cbfa1 with a2A oligo- tissues, and the data are expressed in terms of percentage activity compared with that of bone. nucleotide, again using a 2-fold higher amount of Cbfa1 com- pared with that used to see binding of Cbfa1 to a1wtAmutB, of multiple vertebrate species (Fig. 1C). For this reason, this resulted in the formation of a protein-DNA complex (Fig. 5B, site was studied further. First, DNA binding was studied. lane 1). This complex did not form upon incubation of Cbfa1 EMSA was performed using a labeled double-stranded oligonu- with a2mutA oligonucleotide (Fig. 5B, lane 2). These data show cleotide containing the OSE2 element (a2A) as a probe and that the conserved OSE2 sequence present in the a2(I) collagen ROS 17/2.8 nuclear extract as a source of protein. Incubation of gene can bind Cbfa1, albeit more weakly than the a1A site. labeled a2A oligonucleotide with ROS 17/2.8 nuclear extract Cbfa1 Can Activate Transcription through the a2A Site—To resulted in the generation of a protein-DNA complex migrating study the function of the a2A site, cotransfection assays were at the same location as the protein-DNA complex formed upon performed in COS7 cells. In this assay, exogenous Cbfa1 trans- incubation of ROS 17/2.8 nuclear extract with labeled OSE2 activated a construct containing a multimer of four a2A oligo- OG2 and a1wtAmutB (Fig. 5A, lanes 1 and 2). This protein-DNA nucleotides fused to a minimal a1(I) collagen promoter-luc complex was specific, because it did not form upon incubation of chimeric gene, p4a2A-luc (Fig. 5C), producing an ;15-fold in- ROS 17/2.8 nuclear extract with an oligonucleotide containing crease in luciferase activity. This effect was specific, because a mutation in this OSE2 sequence (a2mutA) (Fig. 5A, lane 3). the construct containing a multimer of six a2mutA oligonucleo- Because the binding of nuclear extract to a2A oligonucleotide tides (Fig. 5C) produced no activity. The relatively weak in- was weak, despite using a 2-fold higher amount of ROS 17/2.8 crease in luciferase activity compared with the effect observed 7106 Cbfa1 Regulation of Type I Collagen a1A element is the most potent activator of expression of all the OSE2 elements we studied in this promoter. These findings do not exclude the possibility that OSE2 sites present further upstream in the promoter and/or elsewhere in the gene may also contribute to the osteoblast-specific expression of the a1(I) collagen gene. Conceivably, one of these as of yet uncharacter- ized OSE2 sites may bind Cbfa1 with a higher affinity and act as a more powerful osteoblast-specific cis-acting element. We did not identify any conserved consensus OSE2 sites by exam- ining the DNA sequence of the region located between 21540 and 21656 that has been previously shown to be required for osteoblast expression (25). This reinforces the hypothesis that other cell-specific transcription factors must contribute to os- teoblastic expression of the type I collagen genes. Although Cbfa1 can bind to a site present in the a2(I) colla- gen gene and the expression of this gene is decreased in trans- genic mice expressing a dominant negative form of Cbfa1, the level of activation observed in cotransfection experiments with the a2A construct was lower than that seen when using the a1AB constructs based on the a1(I) collagen promoter. At least two explanations could account for this observation. First, and most importantly, the binding of ROS 17/2.8 nuclear extract to the a2A site was weaker than its binding to the a1A site of a1(I) collagen, indicating that this site has a lower affinity to Cbfa1. Second, considering the numerous OSE2 sites present in the a2(I) collagen promoter, it is likely that, for this gene and for the a1(I) collagen gene as well, some of the other OSE2 sites act alone or in concert with the conserved OSE2 site to control its expression in osteoblasts in vivo. If Cbfa1 is one positive regulator of type I collagen expression in osteoblasts, it is clear from the above data that it is not the only one. Indeed, Cbfa1 expression is initiated in osteoblast progenitors after type I collagen expression can be noticed in mesenchymal cells. Moreover, at least one other cis-acting el- FIG.5. Cbfa1 binds an OSE2 site in a2(I) collagen, and this site ement has been shown to be implicated in osteoblast-specific can activate transcription. A, EMSA was performed using labeled expression of the a1(I) collagen gene in mouse and rat (15, 16). OSE2 (lane 1), a2A (lane 2), or a2mutA (lane 3) oligonucleotides and OG2 Members of the DLX family of homeobox proteins are able to ROS 17/2.8 nuclear extract. B, EMSA using recombinant Cbfa1 as a source of protein and a2A (lane 1)or a2mutA (lane 2) oligonucleotides bind to this sequence and to activate transcription (22). An- as a probe. C, multimers of a2A and a2mutA oligonucleotides were other homeobox-related protein, MSX2, can bind to this se- placed upstream of the minimal promoter constructs used for a1(I) quence and repress expression of the a1(I) collagen gene (23), collagen transfections. These constructs were used in cotransfection and recent genetic evidence has demonstrated that MSX2 is assays in COS7 cells with a Cbfa1-expressing vector (dark bars) or with an empty vector (open bars). Values represent -fold activation in rela- upstream of Cbfa1 (24). These homeobox proteins are likely to tion to an empty reporter vector and are an average of at least three be expressed earlier than Cbfa1 and may even control its ex- experiments done in triplicate. pression, directly or indirectly. It is tempting to speculate that these homeobox proteins, possibly with other regulatory pro- with multimers of the a1wtAmutB oligonucleotide is consistent teins, act early during the specification of mesenchymal pro- with the weaker binding of Cbfa1 to the a2A sequence. genitor cells to the osteoblast lineage and that Cbfa1 is re- quired for osteoblast differentiation and for the maintenance of DISCUSSION the osteoblast phenotype. This hypothesis will be more easily Taken together, our data provide evidence indicating that testable when mice deficient for several DLX proteins are avail- Cbfa1 is one of the transcription factors contributing to the able. Regardless, the observation that Cbfa1 binds to and reg- expression of the two Type I collagen genes in osteoblasts in ulates the activity of both type I collagen genes in osteoblasts vivo. Moreover, along with CBF (3, 4, 18) and Sp1 (5, 7), Cbfa1 further illustrates how important Cbfa1 is in osteoblast belongs to a growing group of transcription factors accounting physiology. for the coordinated regulation of both genes. The finding that Acknowledgments—We thank Dr. P. Ducy and Dr. T. Schinke for Cbfa1 favors type I collagen expression in osteoblasts is con- critically reading the manuscript. sistent with the absence of bone extracellular matrix in Cbfa1- deficient mice and with the marked decrease of Type I collagen REFERENCES expression in transgenic mice overexpressing a dominant neg- 1. Gehron-Robey, P. (1996) Principles of Bone Biology, pp. 155–166, Academic Press, San Diego, CA ative form of Cbfa1 in osteoblasts (10, 11, 14). These results 2. van der Rest, M., and Garrone, R. (1991) FASEB. J. 5, 2814 –2823 broaden the spectrum of transcription factors able to regulate 3. Maity, S. N., Golumbek, P. T., Karsenty, G., and de Crombrugghe, B. (1988) type I collagen gene expression at various stages of osteoblast Science 241, 582–585 4. Karsenty, G., and de Crombrugghe, B. (1990) J. Biol. Chem. 265, 9934 –9942 differentiation and overall increase our understanding of type I 5. Nehls, M. C., Rippe, R. A., Veloz, L., and Brenner, D. A. (1991) Mol. Cell. Biol. collagen genes’ regulation. 11, 4065– 4073 6. Karsenty, G., and de Crombrugghe, B. (1991) Biochem. Biophys. Res. Com- The data presented here indicate that there is a clear func- mun. 177, 538 –544 tional hierarchy between the different Cbfa1-binding sites, or 7. Greenwel, P., Inagaki, Y., Hu, W., Walsh, M., and Ramirez, F. (1997) J. Biol. OSE2 sites, present in the a1(I) collagen promoter. Clearly, the Chem. 272, 19738 –19745 Cbfa1 Regulation of Type I Collagen 7107 8. Sandberg, M., Autio-Harmainen, H., and Vuorio, E. (1988) Dev. Biol. 130, 1421–1432 324 –334 17. Ausubel, F. M., Brent, R., Kingson, R. E., Moore, D. D., Seidman, J. G., Smith, 9. Ducy, P., Zhang, R., Geoffroy, V., Ridall, A., and Karsenty, G. (1997) Cell. 89, J. A., and Struhl, K. (1995) Current Protocols in Molecular Biology, pp. 747–754 8.5.5, 9.1.7, and 10.7.1, John Wiley & Sons, New York, NY 10. Komori, T., Yagi, H., Nomura, S., Yamaguchi, A., Sasaki, K., Deguchi, K., 18. Zhang, R., Ducy, P., and Karsenty, G. (1997) J. Biol. Chem. 272, 110 –116 Shimizu, Y., Bronson, R. T., Gao, Y. H., Inada, M., Sato, M., Okamoto, R., 19. Ducy, P., and Karsenty, G. (1995) Mol. Cell. Biol. 15, 1858 –1869 Kitamura, Y., Yoshiki, S., and Kishimoto, T. (1997) Cell. 89, 755–764 20. Pepling, M. E., and Gergen, J. P. (1995) Proc. Natl. Acad. Sci. U. S. A. 92, 11. Otto, F., Thornell, A. P., Crompton, T., Denzel, A., Gilmour, K. C., Rosewell, 9087–9091 I. R., Stamp, G. W., Beddington, R. S., Mundlos, S., Olsen, B. R., Selby, 21. Kurikawa, M., Tanaka, T., Tanaka, K., Hirano, N., Ogawa, S., Mitani, K., P. B., and Owen, M. J. (1997) Cell. 89, 765–771 Yazaki, Y., and Hirai, H. (1996) J. Biol. Chem. 271, 16870 –16876 12. Mundlos, S., Otto, F., Mundlos, C., Mulliken, J. B., Aylsworth, A. S., Albright, 22. Tadic, T., Dodig, M., Kronenberg, M. S., Kaliterna, M., Kosher, R. A., Ferrari, S., Lindhout, D., Cole, W. G., Henn, W., Knoll, J. H., Owen, M. J., Mertel- D. A., and Lichtler, A. C. (1998) Bone 23, S539 smann, R., Zabel, B. U., and Olsen, B. R. (1997) Cell. 89, 773–779 23. Dodig, M., Kronenberg, M. S., Bedalov, A., Kream, B. E, Gronowicz, G., Clark, 13. Lee, B., Thirunavukkarasu, K., Zhou, L., Pastore, L., Baldini, A., Hecht, J., S. H., Mack, K., Liu, Y., Maxon, R., Pan, Z. Z., Upholt, W. B., Rowe, D. W., Geoffroy, V., Ducy, P., and Karsenty, G. (1997) Nat. Genet. 16, 307–310 and Lichtler, A. C. (1996) J. Biol. Chem. 271, 16422–16429 14. Ducy, P., Starbuck, M., Priemel, M., Shen, J., Pinero, G., Geoffroy, V., Amling, 24. Satokata, I., Ma, L., Ohshima, H., Bei, M., Woo, I., Nishizawa, K., Maeda, T., M., and Karsenty, G. (1999) Genes Dev. 13, 1025–1036 Yoshiro, T., Uchiyama, M., Heany, S., Peters, H., Tang, Z., Maxson, R., and 15. Bogdanovic, Z., Bedalov, A., Krebsbach, P. H., Pavlin, D., Woody, C. O., Clark, Maas, R. (2000) Nat. Genet. 24, 391–395 S. H., Thomas, H. F., Rowe, D. W., Kream, B. E., and Lichtler, A. C. (1994) J. Bone Miner. Res. 9, 285–292 25. Rossert, J. A., Chen, S. S., Eberspaecher, H., Smith, C. N., and de Crombrug- 16. Rossert, J., Eberspaecher, H., and de Crombrugghe, B. (1995) J. Cell Biol. 129, ghe, B. (1996) Proc. Natl. Acad. Aci. U. S. A. 93, 1027–1031 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Biological Chemistry Unpaywall

Cbfa1 Contributes to the Osteoblast-specific Expression of type I collagen Genes

Kern, Britt; Shen, Jianhe; Starbuck, Michael; Karsenty, Gerard
Journal of Biological Chemistry – Mar 1, 2001
Download PDF
7 pages

Loading next page...
 
/lp/unpaywall/cbfa1-contributes-to-the-osteoblast-specific-expression-of-type-i-tAXM5ODX4l

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Unpaywall
ISSN
0021-9258
DOI
10.1074/jbc.m006215200
Publisher site
See Article on Publisher Site

Abstract

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 276, No. 10, Issue of March 9, pp. 7101–7107, 2001 © 2001 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Cbfa1 Contributes to the Osteoblast-specific Expression of type I collagen Genes* Received for publication, July 13, 2000, and in revised form, November 29, 2000 Published, JBC Papers in Press, December 5, 2000, DOI 10.1074/jbc.M006215200 Britt Kern, Jianhe Shen, Michael Starbuck, and Gerard Karsenty‡ From the Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, Texas 77030 Type I collagen is composed of two chains, a1(I) and throughout life (8), suggesting that the factor(s) controlling a2(I), encoded by two distinct genes, the a1(I) and a2(I) their expression in these cells could also control osteoblast collagen genes, that are highly expressed in osteoblasts. differentiation and function. Another possibility, which is not a1(I) and a2(I) collagen In most physiological situations, exclusive of the previous one, is that different transcription expression is coregulated, suggesting that identical factors may control their expression in osteoblasts at various transcription factors control their expression. Here, we stages of development and of postnatal life. Thus, the elucida- studied the role of Cbfa1, an osteoblast-specific tran- tion of the molecular mechanisms controlling a1(I) and a2(I) a1(I) and a2(I) collagen scription factor, in the control of collagen gene expression in osteoblasts is of critical importance expression in osteoblasts. A consensus Cbfa1-binding in understanding how osteoblast differentiation and, thereby, site, termed OSE2, is present at the same location in the bone matrix deposition by differentiated osteoblasts is regu- a1(I) collagen promoter at approximately 21347 base lated. Ultimately, these studies may shed light on the patho- pairs (bp) of the rat, mouse, and human genes. Cbfa1 can genesis of genetically acquired bone diseases and help design bind to this site, as demonstrated by electrophoretic appropriate therapies for some of these diseases. mobility shift assay (EMSA) and supershift experiments The critical role that Cbfa1, a Runt-related osteoblast-spe- a1(I) using an anti-Cbfa1 antibody. Mutagenesis of the a1(I) cific transcription factor, plays in osteoblast differentiation and collagen OSE2 at 21347 bp reduced the activity of a function has been demonstrated in mouse and in human using collagen promoter fragment 2- to 3-fold. Moreover, mul- timers of this OSE2 at 21347bp confer osteoblast-spe- both molecular and genetic approaches (9 –14). Cbfa1 was iden- a1(I) collagen promoter frag- cific activity to a minimum tified as a key regulator of osteoblast-specific gene expression ment in DNA transfection experiments as well as in through its binding to the OSE2 element of the mouse Osteo- transgenic mice. An additional Cbfa1-binding element is calcin genes 1 and 2 (OG1 and OG2) (9) and other genes a1(I) collagen promoter of mouse, rat, and present in the expressed in osteoblasts. The early and cell-specific expression human at approximately position 2372. This site binds of this gene together with its biological role in vivo as a factor Cbfa1 only weakly and does not act as a cis-acting acti- required for osteoblast differentiation (9 –11), indicate that vator of transcription when tested in DNA transfection Cbfa1 must control the expression of multiple target genes that a1(I) collagen, the mouse a2(I) experiments. Similar to are expressed earlier than Osteocalcin. Conceivably, these tar- collagen gene contains multiple OSE2 sites, of which one get genes could include the a1(I) and a2(I) collagen genes that is conserved across multiple species. In EMSA, Cbfa1 are expressed early during development. This hypothesis was a2(I) OSE2 ele- binds to this site and multimers of this confirmed indirectly by the observation that expression of a ment confer osteoblast-specific activity to the minimum dominant negative form of Cbfa1 in differentiated osteoblasts a1(I) collagen promoter in DNA transfection experi- leads to a decrease in expression of the type I collagen genes in ments. Thus, our results suggest that Cbfa1 is one of the vivo (14). To date, no osteoblast-specific cis-acting elements to positive regulators of the osteoblast-specific expression of both type I collagen genes. which Cbfa1 may bind have been identified in these genes. Two groups have extensively studied the regulation of ex- pression of the a1(I) collagen gene in osteoblasts and have Type I collagen is the most abundant protein of the bone identified a region in the promoter of the rat and mouse a1(I) extracellular matrix, accounting for 90% of the matrix protein collagen gene that plays an important role in this regulation of content (1). It is a heterotrimer made of two a1(I) chains and expression (15, 16). The sequence of this region bears no ho- one a2(I) chain (2). The a1(I) and a2(I) chains are encoded by mology to a Cbfa1-binding site, and several homeobox-contain- two distinct genes that are expressed most highly in two cell ing proteins can bind to this sequence and affect a1(I) collagen types: the fibroblast and the osteoblast. Moreover, the expres- expression. However, a cell-specific transcription factor bind- sion of these two genes is often regulated by identical transcrip- ing to this region has not yet been identified. Moreover, no tion factors (3–7). The type I collagen genes are expressed in osteoblast-specific cis-acting element has yet been identified in osteoblastic cells at all stages during development and the a2(I) collagen promoter. Given the large size of these genes and their expression at multiple stages of osteoblast differen- tiation, it is likely that several distinct osteoblast-specific cis- * This work was supported by National Institutes of Health Grants acting elements, besides those already described (15, 16), con- R01AR45548 and R01DE11290, March of Dimes Grant F-198-0082, and tribute to the expression of the type I collagen genes in a grant from Eli Lilly and Co. (to G. K.). The costs of publication of this article were defrayed in part by the payment of page charges. This osteoblast progenitors and/or in fully differentiated osteoblasts. article must therefore be hereby marked “advertisement” in accordance Consistent with this hypothesis, we noticed the existence of two with 18 U.S.C. Section 1734 solely to indicate this fact. Cbfa1-binding sites (OSE2s) in the mouse a1(I) collagen pro- ‡ To whom correspondence should be addressed: Dept. of Molecular moter and one OSE2 in the mouse a2(I) collagen gene that are and Human Genetics, Baylor College of Medicine, Houston, TX 77030. Tel.: 713-798-5489; Fax: 713-798-1465; E-mail [email protected]. conserved among multiple species. The functional importance This paper is available on line at http://www.jbc.org 7101 This is an Open Access article under the CC BY license. 7102 Cbfa1 Regulation of Type I Collagen of these sites has never been studied before. Here we present evidence suggesting that Cbfa1 is one of the factors controlling osteoblast-specific expression of both type I collagen genes. EXPERIMENTAL PROCEDURES DNA Constructs—For DNA transfection and generation of trans- genic mice, multimers of double-stranded oligonucleotides (see Table I) were cloned into the SmaI site upstream of a chimeric pK1-luc reporter plasmid containing the a1(I) collagen 286 minimal promoter (4) fused to a luciferase gene. The expression plasmid was pCMV-Osf2 (9). Site-specific mutations were created in the intact promoter by using polymerase chain reaction-directed mutagenesis (17) on a construct containing 2.4 kb of the a1(I) collagen promoter upstream of a lucifer- ase reporter gene. Sole presence of the desired mutations was verified by sequencing. Cell Culture and DNA Transfection—COS7, NIH 3T3, HeLa, C2C12, and 10T1/2 cells were cultured in Dulbecco’s minimal essential medium (Life Technologies, Inc.), 10% fetal bovine serum (Life Technologies, Inc.). ROS 17/2.8 cells were cultured in Dulbecco’s minimal essential medium-F12 medium (Life Technologies, Inc.), 10% fetal bovine serum (Life Technologies, Inc.). Twenty hours before transfection, cells were plated at a density of 5 3 10 cells/dish and allowed to grow under normal culture conditions. For cotransfection experiments, we used 5 mg of Cbfa1 expression vector or empty vector, 5 mg of reporter plasmid or empty vector, and 2 mgofpSVb-gal vector using the calcium phos- phate coprecipitation procedure (17). Transfection conditions were iden- tical to those used in the cotransfection experiments, except that 5 mgof reporter plasmid were used. Twenty hours following transfection, cells were washed in phosphate-buffered saline and incubated in medium an additional 24 h. C2C12 cells were changed to media containing 10% horse serum (Life Technologies, Inc.) and allowed to incubate for 48 h. Cells were collected by scraping into 0.25 M Tris-HCl, pH 7.8, and lysed FIG.1. Type I collagen OSE2 sites are well conserved across by three freeze-thaw cycles. b-Galactosidase and luciferase assays were species. A, three OSE2 sites are present in the mouse a1(I) collagen carried out as described previously (18). b-Galactosidase assay results promoter, but only the a1A site and the a1C site are conserved across were used to normalize the luciferase assay results for transfection species. B, diagram of the a1(I) and a2(I) collagen promoters and the efficiency. All DNA transfection experiments were repeated at least OSE2 sites of each. The arrow indicates the start site of transcription three times in triplicate. (11). The position indicated for each element is relative to the start site Electrophoretic Mobility Shift Assays—Nuclear extract from ROS of transcription. C, a2(I) collagen OSE2 sequence is conserved across 17/2.8 cells, primary osteoblasts, and other tissues were prepared as vertebrate species for which the sequence is available: h, human; m, described previously (19) from 4-day-old wild-type mice and stored at mouse; r, rat; c, chicken; C.f., Canis familiaris; B.t., Bos taurus; R.c., 280 °C until use. Glutathione S-transferase-Cbfa1 was purified from Rana catesbeiana. transformed Escherichia coli bacteria using glutathione beads as de- scribed previously (17). Double-stranded oligonucleotides (see Table I) expressing the p4a1AB-luc construct were analyzed as follows: Organs were end-labeled and purified as previously described (19). 5 fmol of from 4-week-old F1 animals were dissected and homogenized on ice in labeled oligonucleotide was incubated with 7 mg of ROS 17/2.8 nuclear a buffer containing 100 mM potassium phosphate (pH 7.8) and 1 mM extract or 0.1 mg of recombinant Cbfa1 protein. a2(I) collagen EMSA dithiothreitol (DTT). Protein homogenates were centrifuged, and super- experiments used twice the amount of extract or protein. The incuba- natants were assayed for luciferase activity according to standard pro- tion mix for nuclear extract binding assays consisted of binding buffer cedures (19). Protein levels were measured using the Bio-Rad protein (100 mM Tris-HCl, pH 7.5, 200 mM NaCl, 4 mM EDTA, 2 mM DTT, 0.2% assay. Relative luciferase activities were expressed as luciferase light Nonidet P-40, 10% glycerol, 5 mg/ml leupeptin, 5 mg/ml pepstatin) (20), units per 100 mg of protein expressed as a percentage of the activity in 2 mg of poly(dI-dC), and 0.5 fmol of single-stranded bottom strand bone. oligonucleotide. Incubation took place at room temperature for 5 min. RESULTS Supershift experiments were carried out as described above, except that the ROS17/2.8 nuclear extract was preincubated for 10 min at room The a1(I) collagen Promoter Contains Two Conserved OSE2 temperature with an antibody against Cbfa1 in binding buffer prior to Sites—DNA sequence inspection identified three potential their incubation with the labeled oligonucleotide for 10 min at room OSE2 sites in the promoter of the murine a1(I) collagen gene. temperature. Two are located side-by-side at positions 21347 and 21338 in For recombinant protein binding assays, the incubation mix con- the mouse gene (Fig. 1A), and a third site is found at 2372 in sisted of binding buffer (20 mM Tris-HCl, pH 8.0, 10 mM NaCl, 3 mM EGTA, 5 mM DTT, 0.05% Nonidet P-40) and 1 mg of bovine serum the mouse gene (Fig. 1A). These potential OSE2 sites were albumin. Incubation took place at room temperature for 5 min, followed termed a1A, a1B, and a1C, respectively (Fig. 1B). The presence by the addition of 1 ml of loading buffer (2 mM Tris-HCl, pH 8.0, 5% of a1A and a1C, but not of a1B, at approximately the same glycerol, 0.025% xylene cyanol, 0.025% bromphenol blue). location in the a1(I) collagen promoter sequences of rat and The reactions were run on 5% polyacrylamide gel, 0.253 TBE (89 mM human (Fig. 1A) suggested a biological role for these sites and Tris base, 89 mM boric acid, 2 mM EDTA, pH 8.0) for 90 min at 160 V. led us to study these two regions. To determine whether Cbfa1 The gels were then dried and exposed to film at 280 °C. Generation and Analysis of Transgenic Mice—The plasmids de- could bind to the OSE2-like sequences in the mouse a1(I) scribed above containing multimers of the a1AB or a1mutAmutB oli- collagen promoter, we generated double-stranded oligonucleo- gonucleotides were digested, and the insert was purified by two rounds tides to be used in DNA binding assays (Table I). One of them, of agarose gel electrophoresis. Linear DNA inserts were injected into called a1AB, contains the a1A and a1B sites and their sur- the pronuclei of fertilized B6D2F1 (Charles River Laboratory) mouse rounding sequences. A second one, termed a1wtAmutB, con- eggs, which were reimplanted in the oviduct of pseudo-pregnant CD1 tains the wild-type a1A site and a mutated a1B site. A third foster mothers (Jackson Laboratories). Transgenic animals were iden- tified by Southern blots of tail genomic DNA. The transgenic mice oligonucleotide, called a1mutAwtB, carries a mutated a1A se- quence and a wild-type a1B sequence. A fourth oligonucleotide, a1mutAmutB, contains mutations in both the a1A and the a1B The abbreviations used are: kb, kilobase(s); DTT, dithiothreitol; EMSA, electrophoretic mobility shift assay; bp, base pair(s). sites. Two other oligonucleotides, termed a1C and a1mutC, Cbfa1 Regulation of Type I Collagen 7103 TABLE I The a1A Site Acts as an Osteoblast-specific Activator of Tran- Oligonucleotides used in this study scription in Tissue Culture Experiments and in Vivo—We fur- Boldface letters indicate mutated nucleotides. ther addressed the functional relevance of the a1A and a1B Oligonucleotide sites using two additional approaches. First, to examine the Sequence name effect of the a1A site on activity of a 2.4-kb promoter fragment, a1AB 5 9-GATCCCTTCCCACACCACCCACACAGA-3 9 a site-specific mutation was generated in the a1A site via a1wtAmutB 5 9-GATCCCTTC GA ACACCACCCACACAGA-3 9 polymerase chain reaction and introduced into a 2.4-kb a1(I) a1mutAwtB 5 9-GATCCCTTCCCACACCAC GA ACACAGA-3 9 collagen promoter-luc chimeric gene. These 2-bp mutations a1mutAmutB 5 9-GATCCCTTC GA ACACCAC GA ACACAGA-3 9 resulted in a 54% decrease in promoter activity when tested in a1C 5 9-GATCTACTGAGGGCCCAGCCACACTCCA-3 9 a1mutC 5 9-GATCTACTGAGGGCCCAG GA ACACTCCA-3 9 DNA transfection experiments in ROS17/2.8 cells (Fig. 3A). a2A 5 9-GATCCCTTTGTGGATACGCGGACTTTGA-3 9 The same mutations did not reduce the activity of this 2.4-kb a2mutA 5 9-GATCCCTTTGT TC ATACGCGGACTTTGA-3 9 a1(I) collagen promoter-luc chimeric gene in other cell lines of nonosteoblastic nature of mesenchymal and nonmesenchymal origin (Fig. 3A). This result further suggests that this cis-acting were generated to test the binding activity of the a1C site. The element is active only in osteoblasts. Mutations in the a1B site mutations introduced into all the oligonucleotides mentioned did not affect the activity of the 2.4-kb fragment of the a1(I) above have previously been shown to abolish binding of nuclear collagen promoter used in this study (Fig. 3A). extract or recombinant Cbfa1 to the OSE2 sequence present in Second, in DNA cotransfection assays performed in COS7 the Osteocalcin (OG2) promoter (OSE2 ) (19). These double- OG2 cells, a cell line that does not express Cbfa1 (21), exogenous stranded oligonucleotides were then used as probes in electro- Cbfa1 transactivated a construct containing four copies of the phoretic mobility shift assays (EMSA) using either ROS 17/2.8 a1AB oligonucleotide fused to a minimal a1(I) collagen promot- nuclear extract or recombinant Cbfa1 as a source of protein. er-luc chimeric gene, p4a1AB-luc (Fig. 3B). Indeed, cotransfec- Cbfa1 Binds to an OSE2 Site in the Mouse a1(I) collagen tion of p4a1AB-luc with a recombinant Cbfa1-expressing vector Promoter—The complex formed upon incubation of ROS 17/2.8 resulted in a 180-fold activation, whereas an empty expression nuclear extract with a1AB migrated at the same location as the vector had no effect. This level of transactivation is similar to complex formed upon incubation of ROS 17/2.8 nuclear extract what we observed when using a vector containing six copies of with the OSE2 oligonucleotide (Fig. 2A, lanes 1 and 2), OG2 the OSE2 as a reporter (Fig. 3B). The minimal a1(I) colla- OG2 although it was of weaker intensity. In contrast, no protein- gen promoter fragment has virtually no transactivation ability DNA complex was observed when using a1mutAmutB as a on its own (4). When using a vector containing multimers of probe (Fig. 2A, lane 5). a1wtAmutB oligonucleotides cloned upstream of the minimal To determine which of the two upstream OSE2 sites was a1(I) collagen promoter fragment, we observed a 120-fold in- binding to this factor, we used a1wtAmutB oligonucleotide or crease in luciferase activity, indicating that the a1A site is the a1mutAwtB oligonucleotide as probes in EMSA. Incubation of main contributor to the transactivating function of this region labeled a1wtAmutB with ROS 17/2.8 nuclear extract generated of the a1(I) collagen promoter. This is consistent with the a protein-DNA complex that had the same mobility as the one observation that only the a1A site is able to bind Cbfa1 in vitro. observed when using a1AB oligonucleotide as a probe and was The slight decrease in activity seen with the loss of the a1B site of stronger intensity (Fig. 2A, lane 3). In contrast, when using may indicate a synergistic effect of these two sites in this type a1mutAwtB oligonucleotide as a probe, we observed only a of experiment. The activity of constructs containing four copies weak binding of ROS 17/2.8 nuclear extract to the DNA (Fig. of a1mutAmutB, or of a1mutAwtB upstream of the minimal 2A, lane 4). a1(I) collagen promoter-luc chimeric gene (Fig. 3B) could not be To determine whether this osteoblast-specific factor binding increased upon cotransfection with the Cbfa1-expressing vec- to the a1A OSE2-like element was indeed Cbfa1, we performed tor, thus demonstrating the specificity of the effect observed. In three types of experiments. First, we asked whether the factor this set of experiments, we used multimers of the a1(I) collagen present in ROS 17/2.8 nuclear extract and binding to the a1A OSE2 sites, because Cbfa1 does not transactivate the a1(I) site was expressed only in osteoblasts. For that purpose, we collagen promoter fragment in this type of assay. This is a prepared nuclear extract from primary osteoblasts and several consistent feature of Cbfa1 biology, indeed, we observed weak other tissues and used them in EMSA. As shown in Fig. 2B, the transactivation of the OG2 promoter when cotransfected with factor binding to the a1A oligonucleotide was present only in Cbfa1 (9), compared with the strong transactivating effect of primary osteoblast nuclear extract and not in nuclear extract of Cbfa1 observed when using 6OSE2-luc (9). other tissues. Next we performed supershift experiments using Third, we asked whether these OSE2 sites could confer bone- anti-Cbfa1 antibody or a nonspecific antiserum. Incubation of specific expression to a reporter gene in vivo. For that purpose, we the nuclear extract with an antibody against Cbfa1 prior to generated transgenic mice containing two of the constructs used in addition of labeled a1AB oligonucleotide led to the formation of the above transfections, p4a1AB-luc and p4a1mutAmutB-luc. In a second protein-DNA complex of slower mobility (Fig. 2C, lane transgenic mice harboring p4a1AB-luc, luciferase activity could be 2), whereas a nonspecific serum had no effect (Fig. 2C, lane 1), detected in bone but neither in other tissues expressing type I demonstrating that the protein-DNA complex formed upon in- collagen, nor in tissues which do not express type I collagen (Fig. cubation of the labeled a1AB with ROS 17/2.8 nuclear extract 3C). As expected, given their respective sizes, the expression of contains Cbfa1, because this antibody is specific for Cbfa1 (14). p4a1AB-luc was considerably lower than that of a1(I) collagen Third, we asked whether recombinant Cbfa1 could bind to a1AB but not to a1mutAmutB (Fig. 2D, lanes 1 and 4). Incu- (data not shown). The p4a1mutAmutB-luc construct was not ex- pressed in bone or any other tissue (Fig. 3C). Taken together with bation of labeled a1wtAmutB oligonucleotide with recombinant Cbfa1 resulted in the formation of a protein-DNA complex (Fig. the results of the mutagenesis of the a1A site and of the a1B site, these results indicate that Cbfa1 contributes to the expression of 2D, lane 2), whereas incubations using labeled a1mutAwtB oligonucleotide did not (Fig. 2D, lane 3). Taken together, these a1(I) collagen in osteoblasts through the a1A site. results indicate that the a1A site, a site conserved in multiple Cbfa1 Binds to the OSE2 Site Located at 2372 bp in the species, is the major binding site for Cbfa1 in this region of the Mouse a1(I) collagen Promoter—As mentioned at the beginning a1(I) collagen promoter. of “Results,” there is a third OSE2 sequence in the a1(I) colla- 7104 Cbfa1 Regulation of Type I Collagen FIG.2. Cbfa1 binds to the a1A site in the a1(I) collagen promoter. DNA binding was analyzed by EMSA. A, la- beled oligonucleotides OSE2 (lane 1), OG2 a1AB (lane 2), a1wtAmutB (lane 3), a1mutAwtB (lane 4), and a1mutAmutB (lane 5) were incubated with ROS 17/2.8 nuclear extract. The arrow indicates the complex of interest. B, labeled a1AB was used as a probe and incubated with nu- clear extract from primary osteoblasts (lane 1), brain (lane 2), kidney (lane 3), lung (lane 4), muscle (lane 5), and spleen (lane 6). The arrow indicates the complex containing Cbfa1. C, supershift EMSA was performed using an antiserum against Cbfa1 (lane 2), or nonspecific an- tiserum (lane 1) using a1AB as a probe. The arrow indicates the complex of lower mobility observed after incubation with the anti-Cbfa1 antibody. D, EMSA using recombinant Cbfa1 as a source of protein and a1AB (lane 1), a1wtAmutB (lane 2), a1mutAwtB (lane 3), and a1mutAmutB (lane 4) oligonucleotides as probes. gen promoter, located at 2372bp in mouse (site a1C, Fig. 1A). weakly, Cbfa1, because Cbfa1 was able to bind to a1C oligonu- This site is also conserved across species, and we first asked cleotide but not to a1mutC oligonucleotide (Fig. 4B, lanes 1 and whether this OSE2-like sequence could be bound by Cbfa1 in 2). These results indicate that Cbfa1 is able to bind only weakly EMSA. Labeled a1C oligonucleotides were incubated with ROS to the a1C site in the mouse a1(I) collagen promoter, suggest- 17/2.8 nuclear extract as described above, leading to the for- ing that this OSE2 site may not play a critical role. To test this mation of a protein-DNA complex that migrated at the same hypothesis, we cloned four copies of wild-type or mutated a1C location as that formed upon incubation of ROS 17/2.8 nuclear oligonucleotides upstream of the minimal a1(I) collagen pro- extract with labeled OSE2 (Fig. 4A, lanes 1 and 3). How- moter fragment-luciferase chimeric gene used in Fig. 3A.As OG2 ever, the protein-DNA complex was of weak intensity compared seen in Fig. 4C, in DNA transfection experiments in COS7 with that we observed when using OG2 or even a1wtAmutB cells, neither wild-type nor mutant a1C constructs could in- OG2 oligonucleotides as probes (Fig. 4A, lanes 1 and 2). No complex crease the activity of this reporter gene upon cotransfection of this size was observed after incubation of labeled a1mutC with a Cbfa1 expression vector. This lack of an overt role for the with ROS 17/2.8 extract (Fig. 4A, lane 3). a1C site is consistent with the rather poor binding of Cbfa1 to To show that Cbfa1 was part of this protein-DNA complex, this site. Taken together, these results indicate that the a1C supershift experiments were performed using an anti-Cbfa1 site is not a critical cis-acting element in controlling the osteo- antibody, a labeled a1C oligonucleotide, and ROS17/2.8 nu- blast-specific expression of the mouse a1(I) collagen gene. clear extract as a source of protein. The incubation of ROS Cbfa1 Binds to an OSE2 Site in the Mouse a2(I) collagen 17/2.8 nuclear extract with an antibody against Cbfa1 prior to Gene—Because a1(I) and a2(I) collagen genes are often coregu- the addition of labeled oligonucleotide led to the formation of a lated, we next asked whether Cbfa1 was also regulating the slower mobility complex. This complex was specific, because it expression of a2(I) collagen. Sequence analysis of the a2(I) was not observed when using a nonspecific antiserum (Fig. 4A, collagen promoter uncovered the existence of several potential lanes 5 and 6). EMSA experiments using recombinant Cbfa1 OSE2 sites. Only one of these, located in the first exon of the provided further evidence that this site could bind, albeit gene, is present at the same location in the a2(I) collagen gene Cbfa1 Regulation of Type I Collagen 7105 FIG.4. The a1C site is bound by Cbfa1 but is unable to activate a minimal a1(I) collagen promoter. A, EMSA was performed using FIG.3. The a1A site can activate transcription and is required labeled OSE2 (lane 1), a1C (lane 3), or a1mutC (lane 4) oligonucleo- OG2 for the full activity of the a1(I) collagen promoter. A, constructs tides and ROS 17/2.8 nuclear extract as a source of protein. The arrow containing 2.4 kb of the a1(I) collagen promoter with or without muta- indicates complex of interest. Supershift EMSA was performed using an tions were transfected into several cell types. a1AB (black bars), a1mu- antiserum against Cbfa1 (lane 6) or nonspecific antiserum (lane 5). The tAwtB (gray bars), and a1wtAmutB (white bars) promoters were cloned arrow indicates the complex of lower mobility formed after incubation upstream of a luciferase reporter gene and used in DNA transfection with the anti-Cbfa1 antibody. B, EMSA using recombinant Cbfa1 as a experiments in ROS 17/2.8 cells, C2C12 cells, 10T1/2 cells, NIH 3T3 source of protein and a1C (lane 1)or a1mutC (lane 2) oligonucleotides cells, and HeLa cells. Values represent percentage activity compared as a probe. C, multimers of a1C and a1mutC oligonucleotides were with the wild-type promoter. a1mutAwtB displays lower transcrip- placed upstream of a minimal a1(I) collagen promoter. These constructs tional activity only in ROS 17/2.8 cells. B, multimers of the oligonucleo- were used in cotransfection assays in COS7 cells with a Cbfa1-express- tides used for EMSA were placed upstream of a minimal a1(I) collagen ing vector (dark bars) or an empty vector (open bars). Values represent promoter fused to a luciferase reporter gene. These constructs were -fold activation in relation to an empty reporter vector and are an transfected into COS7 cells in the presence of a recombinant Cbfa1 average of at least three experiments done in triplicate. expression construct (dark bars), or an empty vector (open bars). Values represent -fold activation in relation to an empty reporter vector and are the average of at least three experiments done in triplicate. C, wild-type (dark bars) and double-mutant (open bars) constructs from DNA transfection experiments were used to generate transgenic mice. nuclear extract, we also used recombinant Cbfa1 protein in Luciferase activity per 100 mg of protein was determined for several EMSA. The incubation of recombinant Cbfa1 with a2A oligo- tissues, and the data are expressed in terms of percentage activity compared with that of bone. nucleotide, again using a 2-fold higher amount of Cbfa1 com- pared with that used to see binding of Cbfa1 to a1wtAmutB, of multiple vertebrate species (Fig. 1C). For this reason, this resulted in the formation of a protein-DNA complex (Fig. 5B, site was studied further. First, DNA binding was studied. lane 1). This complex did not form upon incubation of Cbfa1 EMSA was performed using a labeled double-stranded oligonu- with a2mutA oligonucleotide (Fig. 5B, lane 2). These data show cleotide containing the OSE2 element (a2A) as a probe and that the conserved OSE2 sequence present in the a2(I) collagen ROS 17/2.8 nuclear extract as a source of protein. Incubation of gene can bind Cbfa1, albeit more weakly than the a1A site. labeled a2A oligonucleotide with ROS 17/2.8 nuclear extract Cbfa1 Can Activate Transcription through the a2A Site—To resulted in the generation of a protein-DNA complex migrating study the function of the a2A site, cotransfection assays were at the same location as the protein-DNA complex formed upon performed in COS7 cells. In this assay, exogenous Cbfa1 trans- incubation of ROS 17/2.8 nuclear extract with labeled OSE2 activated a construct containing a multimer of four a2A oligo- OG2 and a1wtAmutB (Fig. 5A, lanes 1 and 2). This protein-DNA nucleotides fused to a minimal a1(I) collagen promoter-luc complex was specific, because it did not form upon incubation of chimeric gene, p4a2A-luc (Fig. 5C), producing an ;15-fold in- ROS 17/2.8 nuclear extract with an oligonucleotide containing crease in luciferase activity. This effect was specific, because a mutation in this OSE2 sequence (a2mutA) (Fig. 5A, lane 3). the construct containing a multimer of six a2mutA oligonucleo- Because the binding of nuclear extract to a2A oligonucleotide tides (Fig. 5C) produced no activity. The relatively weak in- was weak, despite using a 2-fold higher amount of ROS 17/2.8 crease in luciferase activity compared with the effect observed 7106 Cbfa1 Regulation of Type I Collagen a1A element is the most potent activator of expression of all the OSE2 elements we studied in this promoter. These findings do not exclude the possibility that OSE2 sites present further upstream in the promoter and/or elsewhere in the gene may also contribute to the osteoblast-specific expression of the a1(I) collagen gene. Conceivably, one of these as of yet uncharacter- ized OSE2 sites may bind Cbfa1 with a higher affinity and act as a more powerful osteoblast-specific cis-acting element. We did not identify any conserved consensus OSE2 sites by exam- ining the DNA sequence of the region located between 21540 and 21656 that has been previously shown to be required for osteoblast expression (25). This reinforces the hypothesis that other cell-specific transcription factors must contribute to os- teoblastic expression of the type I collagen genes. Although Cbfa1 can bind to a site present in the a2(I) colla- gen gene and the expression of this gene is decreased in trans- genic mice expressing a dominant negative form of Cbfa1, the level of activation observed in cotransfection experiments with the a2A construct was lower than that seen when using the a1AB constructs based on the a1(I) collagen promoter. At least two explanations could account for this observation. First, and most importantly, the binding of ROS 17/2.8 nuclear extract to the a2A site was weaker than its binding to the a1A site of a1(I) collagen, indicating that this site has a lower affinity to Cbfa1. Second, considering the numerous OSE2 sites present in the a2(I) collagen promoter, it is likely that, for this gene and for the a1(I) collagen gene as well, some of the other OSE2 sites act alone or in concert with the conserved OSE2 site to control its expression in osteoblasts in vivo. If Cbfa1 is one positive regulator of type I collagen expression in osteoblasts, it is clear from the above data that it is not the only one. Indeed, Cbfa1 expression is initiated in osteoblast progenitors after type I collagen expression can be noticed in mesenchymal cells. Moreover, at least one other cis-acting el- FIG.5. Cbfa1 binds an OSE2 site in a2(I) collagen, and this site ement has been shown to be implicated in osteoblast-specific can activate transcription. A, EMSA was performed using labeled expression of the a1(I) collagen gene in mouse and rat (15, 16). OSE2 (lane 1), a2A (lane 2), or a2mutA (lane 3) oligonucleotides and OG2 Members of the DLX family of homeobox proteins are able to ROS 17/2.8 nuclear extract. B, EMSA using recombinant Cbfa1 as a source of protein and a2A (lane 1)or a2mutA (lane 2) oligonucleotides bind to this sequence and to activate transcription (22). An- as a probe. C, multimers of a2A and a2mutA oligonucleotides were other homeobox-related protein, MSX2, can bind to this se- placed upstream of the minimal promoter constructs used for a1(I) quence and repress expression of the a1(I) collagen gene (23), collagen transfections. These constructs were used in cotransfection and recent genetic evidence has demonstrated that MSX2 is assays in COS7 cells with a Cbfa1-expressing vector (dark bars) or with an empty vector (open bars). Values represent -fold activation in rela- upstream of Cbfa1 (24). These homeobox proteins are likely to tion to an empty reporter vector and are an average of at least three be expressed earlier than Cbfa1 and may even control its ex- experiments done in triplicate. pression, directly or indirectly. It is tempting to speculate that these homeobox proteins, possibly with other regulatory pro- with multimers of the a1wtAmutB oligonucleotide is consistent teins, act early during the specification of mesenchymal pro- with the weaker binding of Cbfa1 to the a2A sequence. genitor cells to the osteoblast lineage and that Cbfa1 is re- quired for osteoblast differentiation and for the maintenance of DISCUSSION the osteoblast phenotype. This hypothesis will be more easily Taken together, our data provide evidence indicating that testable when mice deficient for several DLX proteins are avail- Cbfa1 is one of the transcription factors contributing to the able. Regardless, the observation that Cbfa1 binds to and reg- expression of the two Type I collagen genes in osteoblasts in ulates the activity of both type I collagen genes in osteoblasts vivo. Moreover, along with CBF (3, 4, 18) and Sp1 (5, 7), Cbfa1 further illustrates how important Cbfa1 is in osteoblast belongs to a growing group of transcription factors accounting physiology. for the coordinated regulation of both genes. The finding that Acknowledgments—We thank Dr. P. Ducy and Dr. T. Schinke for Cbfa1 favors type I collagen expression in osteoblasts is con- critically reading the manuscript. sistent with the absence of bone extracellular matrix in Cbfa1- deficient mice and with the marked decrease of Type I collagen REFERENCES expression in transgenic mice overexpressing a dominant neg- 1. Gehron-Robey, P. (1996) Principles of Bone Biology, pp. 155–166, Academic Press, San Diego, CA ative form of Cbfa1 in osteoblasts (10, 11, 14). These results 2. van der Rest, M., and Garrone, R. (1991) FASEB. J. 5, 2814 –2823 broaden the spectrum of transcription factors able to regulate 3. Maity, S. N., Golumbek, P. T., Karsenty, G., and de Crombrugghe, B. (1988) type I collagen gene expression at various stages of osteoblast Science 241, 582–585 4. Karsenty, G., and de Crombrugghe, B. (1990) J. Biol. Chem. 265, 9934 –9942 differentiation and overall increase our understanding of type I 5. Nehls, M. C., Rippe, R. A., Veloz, L., and Brenner, D. A. (1991) Mol. Cell. Biol. collagen genes’ regulation. 11, 4065– 4073 6. Karsenty, G., and de Crombrugghe, B. (1991) Biochem. Biophys. Res. Com- The data presented here indicate that there is a clear func- mun. 177, 538 –544 tional hierarchy between the different Cbfa1-binding sites, or 7. Greenwel, P., Inagaki, Y., Hu, W., Walsh, M., and Ramirez, F. (1997) J. Biol. OSE2 sites, present in the a1(I) collagen promoter. Clearly, the Chem. 272, 19738 –19745 Cbfa1 Regulation of Type I Collagen 7107 8. Sandberg, M., Autio-Harmainen, H., and Vuorio, E. (1988) Dev. Biol. 130, 1421–1432 324 –334 17. Ausubel, F. M., Brent, R., Kingson, R. E., Moore, D. D., Seidman, J. G., Smith, 9. Ducy, P., Zhang, R., Geoffroy, V., Ridall, A., and Karsenty, G. (1997) Cell. 89, J. A., and Struhl, K. (1995) Current Protocols in Molecular Biology, pp. 747–754 8.5.5, 9.1.7, and 10.7.1, John Wiley & Sons, New York, NY 10. Komori, T., Yagi, H., Nomura, S., Yamaguchi, A., Sasaki, K., Deguchi, K., 18. Zhang, R., Ducy, P., and Karsenty, G. (1997) J. Biol. Chem. 272, 110 –116 Shimizu, Y., Bronson, R. T., Gao, Y. H., Inada, M., Sato, M., Okamoto, R., 19. Ducy, P., and Karsenty, G. (1995) Mol. Cell. Biol. 15, 1858 –1869 Kitamura, Y., Yoshiki, S., and Kishimoto, T. (1997) Cell. 89, 755–764 20. Pepling, M. E., and Gergen, J. P. (1995) Proc. Natl. Acad. Sci. U. S. A. 92, 11. Otto, F., Thornell, A. P., Crompton, T., Denzel, A., Gilmour, K. C., Rosewell, 9087–9091 I. R., Stamp, G. W., Beddington, R. S., Mundlos, S., Olsen, B. R., Selby, 21. Kurikawa, M., Tanaka, T., Tanaka, K., Hirano, N., Ogawa, S., Mitani, K., P. B., and Owen, M. J. (1997) Cell. 89, 765–771 Yazaki, Y., and Hirai, H. (1996) J. Biol. Chem. 271, 16870 –16876 12. Mundlos, S., Otto, F., Mundlos, C., Mulliken, J. B., Aylsworth, A. S., Albright, 22. Tadic, T., Dodig, M., Kronenberg, M. S., Kaliterna, M., Kosher, R. A., Ferrari, S., Lindhout, D., Cole, W. G., Henn, W., Knoll, J. H., Owen, M. J., Mertel- D. A., and Lichtler, A. C. (1998) Bone 23, S539 smann, R., Zabel, B. U., and Olsen, B. R. (1997) Cell. 89, 773–779 23. Dodig, M., Kronenberg, M. S., Bedalov, A., Kream, B. E, Gronowicz, G., Clark, 13. Lee, B., Thirunavukkarasu, K., Zhou, L., Pastore, L., Baldini, A., Hecht, J., S. H., Mack, K., Liu, Y., Maxon, R., Pan, Z. Z., Upholt, W. B., Rowe, D. W., Geoffroy, V., Ducy, P., and Karsenty, G. (1997) Nat. Genet. 16, 307–310 and Lichtler, A. C. (1996) J. Biol. Chem. 271, 16422–16429 14. Ducy, P., Starbuck, M., Priemel, M., Shen, J., Pinero, G., Geoffroy, V., Amling, 24. Satokata, I., Ma, L., Ohshima, H., Bei, M., Woo, I., Nishizawa, K., Maeda, T., M., and Karsenty, G. (1999) Genes Dev. 13, 1025–1036 Yoshiro, T., Uchiyama, M., Heany, S., Peters, H., Tang, Z., Maxson, R., and 15. Bogdanovic, Z., Bedalov, A., Krebsbach, P. H., Pavlin, D., Woody, C. O., Clark, Maas, R. (2000) Nat. Genet. 24, 391–395 S. H., Thomas, H. F., Rowe, D. W., Kream, B. E., and Lichtler, A. C. (1994) J. Bone Miner. Res. 9, 285–292 25. Rossert, J. A., Chen, S. S., Eberspaecher, H., Smith, C. N., and de Crombrug- 16. Rossert, J., Eberspaecher, H., and de Crombrugghe, B. (1995) J. Cell Biol. 129, ghe, B. (1996) Proc. Natl. Acad. Aci. U. S. A. 93, 1027–1031

Journal

Journal of Biological Chemistry – Unpaywall

Published: Mar 1, 2001

There are no references for this article.