Identification of a Homeodomain Binding Element in the BoneSialoprotein Gene Promoter That Is Required for ItsOsteoblast-selective Expression*

Received for publication, June 21, 1999, and in revised form, February 14, 2000

M. Douglas Benson‡, Jeffrey L. Bargeon§, Guozhi Xiao§, Peedikayil E. Thomas§, Ahn Kim§,Yingqi Cui§, and Renny T. Franceschi‡§¶

From the ‡Department of Biological Chemistry, School of Medicine and the §Department of Periodontics, Prevention, andGeriatrics, School of Dentistry, University of Michigan, Ann Arbor, Michigan 48109-1078

Bone sialoprotein is a 70-kDa extracellular matrixcomponent that is intimately associated with biominer-alization, yet the cis-acting elements of the Bsp genethat restrict its expression to mineralizing cells remainuncharacterized. To identify such elements, we ana-lyzed a 2472-base pair fragment of the murine promoterthat directs osteoblast-selective expression of a lucifer-ase reporter gene and found that the region between–338 and –178 relative to the transcriptional start iscrucial for its osteoblast-selective activity. We identifiedan element within this region that binds a protein com-plex in the nuclear extracts of osteoblastic cells and isrequired for its transcriptional activity. Introduction ofa mutation that disrupts a homeodomain binding sitewithin this sequence eliminates both its in vitro bindingand nearly all of the osteoblastic-selective activity of the2472-base pair promoter. We further found that the Dlx5homeoprotein, which is able to regulate the osteoblast-specific osteocalcin promoter, can bind this element andstimulate its enhancer activity when overexpressed inCOS7 cells. These data represent the first description ofan osteoblast-specific element within the bone sialopro-tein promoter and demonstrate its regulation by a mem-ber of a family of factors known to be involved inskeletogenesis.

Bone sialoprotein (BSP)1 is a 70-kDa extracellular matrixcomponent that is selectively produced by mineralizing celltypes in a pattern that correlates with the onset of mineralformation in vivo. This expression pattern, combined with ev-idence that BSP binds collagen (1) and can nucleate hydroxy-apatite formation in vitro (2), has led to the widely held theorythat BSP plays a central role in biomineralization. In osteo-blasts, which produce the vast majority of BSP, expression isfurther restricted to those cells that have secreted, and areactively mineralizing, a type I collagen matrix (3). Thus, BSP isone of the primary markers of the terminally differentiated

osteoblast, and, as such, a study of its expression is importantto our understanding of the transcriptional mechanisms thatmediate osteoblast differentiation.

Extensive studies of the osteocalcin gene 2, which encodesanother osteoblast-specific protein, have revealed that its os-teoblast-selective expression is dependent on binding sites forthe Cbfa1 transcription factor in its promoter (named osteo-blast-specific element 2 (OSE2)) (4). Furthermore, Cbfa1,which itself exhibits bone-restricted expression, has since beenshown to be required for osteoblast development in general, asCbfa1 –/– mice lack functional osteoblasts (5). This findingprompted speculation that this factor may control the tran-scription of other osteoblast-related genes by binding similarOSE2 sites in their promoters. However, we recently investi-gated the contribution of two putative Cbfa1 binding sites tothe activity of a 2.5-kb fragment of the murine Bsp promoterthat exhibits osteoblast-selective expression. We found thatneither site exhibited significant enhancer activity nor contrib-uted to the activity of the 2.5-kb Bsp promoter, suggesting thatother cis-acting elements must be responsible for its osteoblast-selective expression (6). Although at first this result seemssurprising, it is, perhaps, less so when we consider that BSPand osteocalcin are associated with different biochemical func-tions in vivo. That is, whereas BSP is associated with mineralformation, osteocalcin appears to play a role in bone resorptionand turnover (7, 8). Thus, it is reasonable to expect that thesetwo genes, while both markers of the differentiated osteoblast,may be differently regulated.

In addition to Cbf/runt sites, a number of other known en-hancer sequences have been described within the Bsp pro-moter, but little progress has been made toward identificationof the elements necessary for its osteoblastic-selective expres-sion. Sodek and co-workers (9–11) have characterized elementswithin the rat promoter that mediate the actions of vitamin-D,glucocorticoids, and transforming growth factor-b, and Yangand Gerstenfeld (12) reported that parathyroid hormone-re-sponsive elements in the avian Bsp promoter are also tran-scriptionally active. However, none of these elements has beenshown to confer tissue specificity. Kerr et al. (13) described abinding site for Ying Yang 1, a factor implicated in the tran-scriptional initiation of some TATA-less promoters, within in-tron 1 of the rat gene, which displayed higher levels of en-hancer activity in UMR 106-01 osteosarcoma cells than infibroblasts. However, since this sequence is not conserved inthe corresponding region of either the human or mouse genes,and no intronic sequence is found in the –2472/141 murinepromoter, this element is clearly not essential for osteoblast-specific expression.

Because efforts to link the tissue specificity of the Bsp pro-moter to a variety of known transcription factor binding se-

* This work was supported by National Institutes of Health GrantsDE12211 and DE 11723, General Clinical Research Center Grant M01-RR00042, and Michigan Multipurpose Arthritis Center Grant AR20557. The costs of publication of this article were defrayed in part bythe payment of page charges. This article must therefore be herebymarked “advertisement” in accordance with 18 U.S.C. Section 1734solely to indicate this fact.

¶ To whom correspondence should be addressed: Dept. of Periodon-tics, Prevention, and Geriatrics, School of Dentistry, University ofMichigan, 1011 N. University Ave., Ann Arbor, MI 48109-1078. Tel.:734-763-7381; Fax: 734-763-5503; E-mail: rennyf@umich.edu.

1 The abbreviations used are: BSP, bone sialoprotein; bp, base pair(s);kb, kilobase(s); OSE, osteoblast-specific element; PCR, polymerasechain reaction; b-gal, b-galactosidase.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 275, No. 18, Issue of May 5, pp. 13907–13917, 2000© 2000 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www.jbc.org 13907

quences have so far been unsuccessful, we took the alternateapproach of conducting a systematic study of the 2472-bp pro-moter to identify osteoblast-specific elements. As we reporthere, this study has revealed that a homeodomain binding sitein the proximal promoter is required for the tissue-specificexpression of the murine Bsp gene. Furthermore, this site ispositively regulated by the Dlx5 homeoprotein, which is knownto play an important role in skeletogenesis.

EXPERIMENTAL PROCEDURES

Cell Culture and Transfections—The following cell lines were usedfor these studies: The isolation of preosteoblastic clone 4 cells from theparent MC3T3-E1 line (14) was described by Xiao et al. (15). UMR106-01 cells (16) were a gift from Dr. Ronald Midura (Cleveland Clinic,Cleveland, OH). C2C12 mouse myoblasts (17) were a gift from Dr.Daniel Goldman (University of Michigan, Ann Arbor, MI). ROS 17/2.8osteosarcoma cells (18), were provided by Dr. Laurie McCauley (Uni-versity of Michigan School of Dentistry). The 3T3-L1 preadipocyte (19),F9 teratocarcinoma (20), and S194 myeloma cell lines were purchasedfrom the American Type Culture Collection (Manassas, VA). Mainte-nance conditions for these cell lines were previously described (6). COS7monkey kidney cells (21) were the gift of Dr. Fred Askari (University ofMichigan) and were maintained in Dulbecco’s modified Eagle’s mediumwith 10% fetal bovine serum and 1% penicillin/streptomycin. C2C12cells were induced to differentiate into myotubes by switching to 2%horse serum for 6 days. An adipose phenotype of lipid accumulation wasinduced in 3T3-L1 cells by growth in 30% fetal bovine serum for 6 days(19). F9 cells were induced to differentiate into endoderm by the addi-tion of 0.1 mM retinoic acid and 1.0 mM dibutyryl cyclic AMP for 6 days(20). Clone 4 differentiation was accomplished by the addition of 50 mg/mlascorbic acid to the medium for 6 days to induce matrix synthesis andsubsequent expression of osteoblast-related genes, including osteocalcinand BSP (22). Media were changed every 2 days during treatment.

All adherent cell lines were plated in 35-mm dishes at a density of25 3 104 cells/cm2 and transfected with the appropriate firefly lucifer-ase reporter construct plus pRL-SV40 Renilla luciferase vector (Pro-mega Corp., Madison, WI) for normalization of transfection, using Li-pofectAMINE (Life Technologies, Inc.) as described previously (6). Cellswere switched to the appropriate differentiation media and grown for 6days, except for COS7 cells, which were additionally transfected with100 ng of either pcDNA3-Dlx5 or pCMV5-b-gal expression plasmid andharvested 48 after transfection. S194 myeloma cells were transfected byelectroporation as described elsewhere (6). They were then transferredto growth medium for 48 h before harvesting. Luciferase expression inthe cell lysates was assayed using the Dual Luciferase Reporter kit(Promega) and a Monolight 2010 luminometer (PharMingen, San Di-ego, CA). All transfection experiments were performed in triplicate atleast three times. Values shown are the means 6 standard deviations oftriplicate samples from representative experiments.

To produce Dlx5-containing or control COS7 nuclear extracts for gelshift experiments, cells were plated at the same density as above, but in150-mm dishes, and transfected with either pcDNA3-Dlx5, or pCMV5-b-gal plasmid (see below). The cells were harvested after 48 h, andnuclear extracts were prepared as described below.

DNA Constructs—Construction of pmBSP2472 was described previ-ously (6). This vector, renamed p2472, contains the region of the murineBsp gene from –2472 to 141 ligated upstream of a firefly luciferasereporter gene in pGL3Basic (Promega). The 59-deletion constructs p705,p338, p178, and p49 were created by ExoIII digestion of p2472 linear-ized with KpnI and MluI using the Erase-a-Base kit (Promega) accord-ing to the manufacturer’s instructions. The p49BSP luciferase reportervector was constructed by ligating a double-stranded, synthetic oligo-nucleotide containing the sequence of the basal murine Bsp promoterfrom –49 to 141 into the BglII and HindIII sites of pGL3Basic. Thep34OG2 vector containing a fragment of the osteocalcin gene 2 pro-moter from –34 to 11 was created in a similar fashion, as describedpreviously (6). This fragment has previously been shown to direct highlevels of matrix-induced transcription in MC3T3-E1 cells when posi-tioned downstream of a bone-specific enhancer. Double-stranded oligo-nucleotides containing wild type and mutant A, B, and C elements fromthe Bsp promoter (see Table I) were synthesized by the University ofMichigan Core Facilities. They were then self-ligated and inserted intothe BglII site of both p49BSP and p34OG2 in order to compare theenhancer activities of the Bsp sites with those of the minimal promotersalone. A, B, or C element sequences were deleted from the 705-base pairpromoter to create p705DA, p705DB, and p705DC as follows: primer

pGL3(–), which anneals to the antisense strand of the vector immedi-ately 39 of the pGL3Basic multiple cloning site, was used in conjunctionwith BglII linker primers –253(1), –226(1), and –180(1) to amplifythree different length fragments by the polymerase chain reaction(PCR) using Taq polymerase (Applied Biosystems, Foster City, CA).These fragments were ligated into the BglII and HindIII sites ofpGL3Basic. Primers –279(–), A(–), and B(–) were then used with primerpGL3(1), a KpnI linker primer that anneals to the antisense strand ofthe plasmid upstream of the multiple cloning site, to create PCR frag-ments representing the upstream segments of the final constructs.These fragments were blunted with Klenow, digested with KpnI, andligated into the KpnI and SmaI sites of the plasmids, which alreadyharbored the downstream PCR fragments described. The resulting Bspreporter constructs are missing 26, 30, and 50 bp of Bsp sequence for A,B, and C deletions, respectively, which is replaced with cloning sitesequence for net total deletions of 8, 13, and 33 bp. These constructs aregraphically depicted in Fig. 5A. The sequences of the primers used intheir construction are given in Table I. These internal deletion con-structs were transfected into clone 4 and UMR 106-01 cells under theconditions described above in order to assess the effect on promoteractivity of deletion of these elements in the context of the 705 base pairBsp promoter. To create the probe plasmid pBSPfp1, harboring theinsert used as a probe in our footprinting studies, we used linkerprimers –369(1) and –165(–) (see Table I) to PCR amplify a Bsp pro-moter fragment, which we then ligated into the BamHI and EcoRI sitesof pBluescript SK(1) (Stratagene, La Jolla, CA). Creation ofp2472mut3, containing the 2472-bp Bsp promoter with a 2-bp mutationin the C element sequence, was accomplished as follows: primersCmut3SD(1) and (–) (Table I) were used with Stratagene’s Quick-Change site-directed mutagenesis kit according to the manufacturer’sinstructions to introduce the desired mutation into p2472. An EcoRV/XhoI fragment containing Bsp promoter sequence from –1016 to 141was then isolated from the mutagenized plasmid and ligated into thecorresponding position in the unmutagenized p2472 plasmid. Thisstrategy was used so that the resultant construct contained a muchshorter length of in vitro polymerized sequence that would have toundergo confirmatory sequencing. The sequences of all in vitro synthe-sized constructs were confirmed by dideoxy sequencing using a Seque-nase 2.0 kit (Amersham Pharmacia Biotech). The pcDNA3-Dlx5 expres-sion plasmid was the generous gift of Dr. Dwight Towler (Merck, Inc.,West Point, PA) and contains a sequence that encodes a FLAG epitopetag 59 to codon 2 of a murine Dlx5 cDNA, resulting in the production ofan N-terminal “FLAG-tagged” protein (23). pCMV5-b-gal was createdby ligating a bacterial lacZ cDNA from pRSVb-gal (a gift from Dr.Daniel Wechsler, University of Michigan) into the EcoRI site of pCMV5(Sigma).

Northern Analysis—Total RNA from cells subjected to differentia-tion treatments as described above was isolated by the method ofChomczynski and Sacchi (24). 10 mg aliquots were fractionated on 1%agarose-formaldehyde gels and blotted to nylon membranes as de-scribed previously (25). The mouse BSP cDNA insert used as probe wasisolated from a plasmid obtained from Dr. Marion Young (NIDCR,National Institutes of Health) (26). The Dlx5 cDNA was obtained fromDr. Steven Harris (University of Texas Health Center at San Antonio,San Antonio, TX). Both were labeled with [a-32P]dCTP using a randomprimer kit (Roche Molecular Biochemicals), and incubated with themembranes in a Bellco Autoblot hybridization oven as described previ-ously (27). The washed membranes were then exposed to Kodak X-omatAR film at –70 °C.

DNase I Footprinting Assays—Nuclear extracts were prepared afterthe method of Dignam et al. (28) from cells grown under differentiationconditions. Additionally, nuclei from ascorbic acid-treated clone 4 cellswere isolated from the collagenous matrix by treatment with 2 mg/mlcollagenase A (Sigma) and 0.25% trypsin (Life Technologies, Inc.) for 1 hfollowed by exhaustive rinsing in cold phosphate-buffered saline beforecontinuing with the extraction of nuclear proteins. Footprinting assayswere performed as follows: 5 mg of probe plasmid was digested witheither BamHI (to label the positive strand) or EcoRI (to label thenegative strand). The cut linearized plasmid was then treated with calfintestinal alkaline phosphatase, labeled with [g-32P]ATP using T4polynucleotide kinase, and digested with either EcoRI or BamHI. Thereleased insert was then purified from a 5% acrylamide gel, and itsactivity was quantitated with a scintillation counter. 15,000 cpm ofthese probes and 0, 25, 50, or 75 mg of nuclear extract were incubatedfor 20 min at room temperature in a 100-ml reaction volume containing25 mM Tris-HCl, pH 8.0, 50 mM KCl, 6.25 mM MgCl2, 0.5 mM EDTA, 0.5mM dithiothreitol, and 10% glycerol, with 5 mg of poly(dIzdC). An equalvolume of a 2.5 mM CaCl2, 5 mM MgCl2 solution was added, and the

Osteoblast Specificity of the Mouse Bsp Promoter13908

reactions were digested with 1.2 units of RNase free DNase I (Promega)for 60 s at room temperature. Digestion was stopped by the addition of180 ml of stop solution (200 mM NaCl, 30 mM EDTA, 1% SDS, and 100mg/ml yeast tRNA), followed by extraction with phenol/chloroform/isoamyl alcohol (25:24:1) equilibrated with TEB (10 mM Tris, pH 8.0, 1mM EDTA, and 15 mM b-mercaptoethanol). The samples were precipi-tated with 1 ml of ethanol, rinsed twice with 70% ethanol, dried, andresuspended in 5 ml of loading buffer (95% formamide, 20 mM EDTA,0.05% bromphenol blue, 0.05% xylene cyanol). After heating for 5 min at95 °C, the samples were then electrophoresed on a sequencing gel of 8%acrylamide (19:1 with bis-acrylamide) in 13 TBE at 60 W. The gelswere dried and exposed to Kodak BioMax film at 270 °C. In order todetermine positions of the protected nucleotides in these assays, thesamples were run alongside aliquots of the same probes subjected tochemical sequencing reactions after the method of Maxam and Gilbert(29).

Gel Mobility Shift Assays—Gel shift probe was made by labeling 1.5pmol of double-stranded C oligonucleotide (shown in Table I) with[g-32P]ATP and T4 kinase; filling in with Klenow, cold G, A, and Tnucleotides, and [a-32P]dCTP; and purifying the labeled oligo on a 5%acrylamide gel. Gel mobility shift assays were performed by incubationat 4 °C for 30 min of 5 mg of nuclear extract with approximately 5 fmolof probe in a 15-ml reaction containing 2 ml of poly(dIzdC) in the samebinding buffer that was used in the footprinting assays described above.For experiments including antibody, samples were then incubated foran additional 30 min with either 4 mg of anti-FLAG M2 monoclonalantibody (Stratagene) or IgG as control (Life Technologies) and, in somecases, 100 ng of FLAG peptide (Sigma). The shifted complexes werethen electrophoresed on 5% acrylamide gels in 13 TGE (50 mM Tris-HCl, pH 8.5, 380 mM glycine, 2 mM EDTA, 0.2 mM b-mercaptoethanol)for 120 min at 170 V in a 4 °C cold room. Gels were dried and exposedto BioMax film.

RESULTS

Osteoblast-selective Expression of the 2.5-kb Bsp Promot-er—In order to identify tissue-specific elements in the BSPpromoter, we employed a construct containing bases –2472 to141 of the murine gene driving expression of a luciferasereporter gene. Our decision to focus on this region was based onprevious work which demonstrated that this construct was ableto direct 5–10-fold higher levels of expression in osteoblasticcells than in nonbone cell lines (6), indicating that this frag-ment of the promoter likely contains sequence elements neces-sary to direct tissue-specific BSP expression. This finding is inagreement with data reported by Chen et al. (30), whichshowed mineralized tissue-restricted expression of a similarregion of the rat BSP promoter in transgenic mice. As a preludeto a detailed analysis of the sequence elements responsible forthis selective expression, a more thorough characterization ofthe expression of this reporter construct was performed invarious osteoblastic cell lines to determine whether its regula-tion pattern parallels that of the endogenous BSP message.

The murine MC3T3-E1 preosteoblastic cell line produceshigh levels of bone-related proteins, including BSP, in responseto induction of collagen matrix synthesis by ascorbic acid (31).For our studies, we employed a subclone, clone 4, which giveshigher levels of bone marker transcription than the more het-erogeneous parent cell line (15). Upon transfection with thep2472 construct, and treatment with ascorbic acid over a 7-daytime course, clone 4 cells showed a significant induction inreporter activity—up to 40-fold over that in nontreated cells byday 7—that paralleled the increase in endogenous BSP mes-sage (Fig. 1, A and B). By way of comparison, ROS 17/2.8osteosarcoma cells, which express high levels of osteocalcin butnot of BSP (Fig. 1C), exhibited only 20% of the luciferaseactivity seen in clone 4 cells when transfected with the p2472construct, comparable to those seen in C2C12 myoblasts and3T3-L1 preadipocytes, two mesenchymal cell lines that do notexpress BSP (Fig. 1D). However, UMR 106-01 osteosarcomacells, which constitutively express high levels of BSP (16),showed levels of expression of the transfected construct that

are comparable to those in clone 4 cells (Fig. 2). Thus, the2472-bp murine Bsp promoter is expressed at high levels onlyin cells that produce BSP and in a pattern that mirrors that ofthe endogenous message.

Isolation of a Region Necessary for Osteoblast Selectivity by59-Deletion Analysis—We created a series of 59-deletions of the2472-bp promoter construct in order to isolate regions of thefragment responsible for its osteoblast selective expression.These deletions were transfected into osteoblastic cell lines(MC3T3-E1 and UMR 106-01) and nonbone cells (C2C12 myo-blasts, 3T3-L1 preadipocytes, F9 teratocarcinoma, and S194lymphocytes) and grown under conditions that induced differ-entiation appropriate for each cell type (osteoblast, muscle,adipose, endoderm, and B lymphocyte, respectively; see under“Experimental Procedures”). The object of these treatmentswas to approximate a panel of different tissue types in cellculture so as to examine the tissue selectivity of the transfectedBsp promoter constructs. The results of these transfections,shown in Fig. 2, revealed three major features. (i) Deletion ofthe promoter sequence from –2472 to –178 caused no change inactivity in the nonbone cell lines, whereas further deletion to–49 resulted in a drastic loss of activity in all cell lines. Thislikely means that the region from –49 to –178 contains se-quences required for basal, non-tissue-specific transcription.(ii) Deletion from –2472 to –338 reduced promoter activity inclone 4 cells by approximately 60% but caused only a slightdrop in UMR 106-01 cells. This segment may contain regula-tory sequences that are required for high level transcription inthe MC3T3-E1 cell line but not in the tumor-derived UMR line.(iii) Deletion of sequences from –338 to –178 resulted in a lossof over half of the remaining activity in both clone 4 and UMRlines. This represents the most severe drop in activity to thatpoint and brings the activity in these cells down to that seen inthe nonosteoblastic cell lines, indicating that the region from–338 to –178 most likely contains elements crucial to the bone-selective expression of the promoter. Interestingly, we notedthe presence of a sequence similar to the OSE1 site, describedby Ducy and Karsenty (4) and Schinke and Karsenty (32) asimportant to the tissue-specific expression of osteocalcin, at–566 (TTACATCA). However, because deletion of the regioncontaining this sequence resulted in only a relatively smalldecrease in activity in clone 4 cells and none in UMR cells, andbecause this deletion did not abolish tissue specificity, it wasnot considered further.

DNase I footprint analysis was employed to determine whichbases in the –338 to –178 region contacted protein complexes inthe nuclear extracts of osteoblastic cells. Preincubation of aprobe containing promoter sequence from –390 to –165 withincreasing amounts of UMR nuclear extract resulted in theprotection of three regions (Fig. 3A) on both the sense strand(lanes 2–4) and the corresponding antisense strand (lanes6–8). The same pattern was seen when nuclear extracts fromascorbic acid-treated clone 4 cells were used (not shown). To-gether, these three regions, named A, B, and C, were foundbetween –278 and –187 and are represented graphically in Fig.3B. When compared with their corresponding sequences in therat and human Bsp promoters (Fig. 3C), the A, B, and Celements show a high degree of sequence conservation, consist-ent with a potentially crucial role in regulating transcription.Both the A and B regions are AT rich and do not contain anyrecognizable transcription factor binding sites. The C element,however, contains a consensus binding site for engrailed-1 (33,34), marking it as a potential binding site for a wide range ofhomeodomain-containing factors.

To determine whether these three elements were able tofunction as transcriptional enhancers, we synthesized A, B,

Osteoblast Specificity of the Mouse Bsp Promoter 13909

and C double-stranded oligonucleotides containing the se-quences shown in Fig. 3C (see also Table I) and subclonedmultimers of them, in both orientations, into a firefly luciferasereporter plasmid upstream of a minimal –49/141 Bsp promoterfragment. When transfected into both the clone 4 and UMR celllines, only the C element construct exhibited elevated reporteractivity relative to the minimal promoter alone (Fig. 4A, com-pare rows 1, 4, and 7), with levels approximately 3-fold higherin clone 4 cells and 2-fold higher in UMR. The same patternwas observed when the three elements were placed into asimilar vector containing a 34-bp minimal promoter from themurine osteocalcin gene 2, which has been shown to be en-hanced by multimers of the bone-specific OSE2 element (4, 6).In this case, however, the C element displayed a 2–3-fold stim-ulation in both osteoblastic cell lines when placed in the for-ward orientation (Fig. 4B, row 4) and a 10–13-fold stimulationwhen in the reverse orientation (row 7). This apparent orien-tation dependence may result from the presence of one morecopy of the C element in the reverse construct (five) than in theforward construct (four). To assess the transcriptional contri-butions of these elements in the context of the native promoter,we then created three reporter constructs from which eitherthe A, B, or C element had been deleted from the –705-bp Bsppromoter and compared their activities to that of the wild type–705-bp construct when transfected into both clone 4 and UMRcell lines. The construction of these internal deletion mutants(named p705DA, DB, and DC) is diagrammed in Fig. 5A. The705-bp promoter was chosen as the template for these deletionsbecause it contains enough promoter sequence to exhibit osteo-blast specificity but is short enough to make confirmatory se-quencing of the resultant PCR-produced fragments feasible. Asis shown in Fig. 5B, deletion of the A element did not effectpromoter activity in clone 4 cells, and although deletion of theB element reduced promoter activity by approximately 35% inUMR cells, it did not affect activity in clone 4 cells. Because theB element also failed to show enhancer activity when placed infront of a minimal promoter (Fig. 4), it was not considered

was conducted on RNA from MC3T3-E1 clone 4, ROS17/2.8, C2C12, and3T3-L1 cells grown for 6 days under differentiating conditions. D, cellline specificity of Bsp promoter activity. Clone 4, ROS17/2.8, C2C12,and 3T3-L1 cells were transfected with p2472 and grown for 6 daysunder the same conditions as in D. Firefly luciferase activities werenormalized to Renilla luciferase activities and are shown as a percent-age of the clone 4 ascorbate-treated sample.

FIG. 1. Osteoblast-selective activity of the 2472-bp Bsp pro-moter. A, time course of Bsp promoter induction. MC3T3-E1 clone 4cells were transfected with p2472 and grown for up to 6 days with (A) orwithout (C) ascorbic acid. Cells were harvested at the times indicatedand assayed for firefly and Renilla luciferase activities. B, induction ofendogenous BSP mRNA in MC3T3-E1 cells. Total RNA from clone 4cells, grown with or without ascorbic acid and harvested at the timesshown, was subjected to Northern blot analysis with a labeled BSPcDNA. C, cell line specificity of BSP mRNA expression. Northern analysis

FIG. 2. 5*-Deletion analysis of the Bsp promoter. ExoIII deletionof the p2472 construct yielded the set of 59-deletions of the murine Bsppromoter shown. These deletion constructs were transfected into clone4, UMR 106-01, C2C12, 3T3-L1, F9, and S194 cells and grown underdifferentiation conditions. Firefly luciferase activities were assayed andnormalized to Renilla luciferase and are shown as a percentage of theactivity of the full-length construct in clone 4 cells.

Osteoblast Specificity of the Mouse Bsp Promoter13910

further. In contrast to the A and B deletions, however, deletionof the C element from the 705-bp promoter caused a major lossof transcriptional activity in both cell types. And, in fact, thisdeletion accounted for the entire 60% drop in activity seen uponremoval of the entire region from –338 to –178 in these celllines (compare Figs. 2 and 5B). The results from these deletiontransfections are in agreement with the data from the multim-erized oligonucleotide experiments described above, and, takentogether, these two complementary studies indicate that the Celement is likely the only one of the three that is important forpromoter activity in osteoblastic cells. Thus, we chose to focusour further investigations on this sequence element.

Identification of a Homeodomain Binding Site in the C Ele-ment that is Necessary for Osteoblast-selective Bsp PromoterActivity—Gel mobility shift assays of nuclear extracts from

UMR and clone 4 cells with radiolabeled double-stranded Coligonucleotide as probe yielded a complex array of bands (Fig.6, lane 2). All but one of these species (bands 1–5) could becompeted for, to varying degrees, by the addition of 50–100-foldmolar excess of unlabeled C oligonucleotide (Fig. 6, lanes 3 and4) but not by the addition of a heterologous oligonucleotidecontaining the proximal OSE2 site from the Og2 promoter (4)(data not shown), indicating that this element is able to specif-ically bind a protein complex in the nuclear extracts of theseosteoblastic cells. We synthesized a series of mutated C oligo-nucleotides to determine which bases in the C element wereessential for protein binding. These contained the same se-quence as the wild type C oligonucleotide shown in Table I,with the exception of a 2-base pair change in each. The identityof the mutated bases in each, and their effects on the binding of

FIG. 3. DNase I footprint analysis ofthe –338 to –178 promoter region. A, aconstruct containing the murine Bsp pro-moter from –338 to –178 was linearizedwith EcoRI (to label the (1) strand) orBamHI (for the (–) strand), end-labeledwith 32P, and incubated with the indi-cated amount of UMR nuclear extract be-fore digestion with DNase I. Digestedproducts were run on a sequencing geland compared with digested probe with-out extract. The three regions that wereprotected on both strands are named A, B,and C and are indicated by brackets. Thebase pair numbering of the ends of eachprotected region is also shown. Asterisksmark bands that were intensified due tothe creation of DNase I hypersensitivitycaused by the adjacent bound proteins. B,shown is the region of the Bsp promoterfrom –290 to –181 with the bases pro-tected from DNase I digestion in Amarked by brackets (only the (1) strandsequence is shown for simplicity). Brack-ets above the base sequence denote theprotected bases on the (1) strand in A,whereas those below the sequence denotethe corresponding protected bases on the(–) strand. C, sequence conservation of A,B, and C sequences. Brackets mark thesequences of the A, B, and C oligonucleo-tides that contain the protected bases in-dicated in B. The base pair numbers inthe promoter sequence corresponding tothe ends of each oligonucleotide areshown above the brackets. Shown beloweach are the homologous regions of the ratand human Bsp promoters for sequencecomparison. A consensus binding site forthe engrailed-1 homeodomain protein wasfound in region C and is boxed.

Osteoblast Specificity of the Mouse Bsp Promoter 13911

the C probe when used as competitors in gel shift assays, areshown in Fig. 6. We chose a series of mutations that spannedthe consensus homeodomain site because this is the only rec-

ognizable transcription factor binding site in the C sequence.Addition of a 50–100-fold molar excess of unlabeled mutant 1or mutant 6 oligonucleotides (Fig. 6, lanes 5, 6, 15, and 16)

TABLE IOligonucleotides used in this study

Oligo name Sequence

A GATCCTTTGCTATTTATTTTATTTGAAGAAAACGATAAATAAAATAAACTTCTAG

B GATCCAGTATATTATTATATATATTCAGTCATATAATAATATATATAAGTCTAG

Ca GATCCTCCTCACCCTTCAATTAAATCCCACAAGAGGAGTGGGAAGTTAATTTAGGGTGTTCTAG

2369(1) TATGAATTCACTTTAGACCCCATGCAGTG2253(1) TATGGATCCGCAGTATATTATTATATATATTCAGAACTC2226(1) TATGGATCCCTCTAACTACCATCTTCTCC2180(1) TATGGATCCCAAGCCTCTTGGCAGAAGG2279(2) GATCGACTTAAGCATTTAAGCTTTC2165(2) TATGGATCCCAAGAGGCTTGCATTGTGGpGL3(1) TATGGTACCTGCCAGAACATTTCTCTATCGpGL3(2) ATGCCAAGCTTACTTAGATCGCmut3SDb CCATCTTCTCCTCACCCTTCCCTTAAATCCCACAATGCAAG

GGTAGAAGAGGAGTGGGAAGGGAATTTAGGGTGTTACGTTCa The sequences of the C mutant oligonucleotides used in this study are the same as those of the wild type shown here except for the two base

pair changes shown in Fig. 6A.b The two base pairs that differ from the wild type BSP sequence are shown in boldface.

FIG. 4. Enhancer activity of A, B,and C elements. A, A, B, and C double-stranded oligonucleotides were ligatedinto multimers and inserted immediatelyupstream of a 49-bp minimal Bsp pro-moter driving luciferase expression in thep49BSP vector. Constructs containingoligonucleotides oriented in either the for-ward (F) or reverse (R) direction weretransfected into clone 4 and UMR cells.Cells were harvested after 6 days and as-sayed for luciferase activity (clone 4 cellswere treated with ascorbic acid). Valuesshown are fold induction of each constructover that of the insertless p49BSP vector(row 1). Each arrow represents one copyof a double-stranded oligonucleotide. Theopen box represents the Bsp promoterfrom –49 to 141. B, multimers of A, B,and C oligonucleotides were also ligatedinto the p34OG2 vector, which is the sameas the p49BSP vector described above, ex-cept that the osteocalcin gene 2 promoterfrom –34 to 11 is substituted for the –49/141 Bsp fragment. The resulting con-structs were transfected into clone 4 andUMR cells as in A.

Osteoblast Specificity of the Mouse Bsp Promoter13912

showed that these mutants were able to compete for proteinbinding nearly as well as the wild type C probe. Mutants 2–5,however, were unable to compete with the wild type sequence.Taken together, these assays show that disruption of the ho-meodomain binding sequence, TCAATTAA, abolishes the abil-ity of the C element to bind protein complexes in the nuclei ofosteoblastic cells.

To test whether this sequence requirement for in vitro bind-ing also applied to the enhancer activity of the C element, theability of the mutant oligonucleotide 3 (Cmut3) to stimulatetranscription of the –49/141 Bsp minimal promoter was com-pared with that of the wild type C element described above. Asexpected, a construct containing three copies of the wild typesequence gave up to 3-fold higher activity in osteoblastic cellsthan the minimal promoter alone (Fig. 7A, row 2), but themutant construct, which differed from the wild type by only 2bp, showed no such stimulation. To assess the contribution ofthe homeodomain binding site to the osteoblast specificity ofthe Bsp promoter, we introduced the same 2 bp mutation, by

site-directed mutagenesis using primers Cmut3SD(1) and (–)(see Table I), into the full-length 2472-bp promoter construct.Results from the transfection of this mutated construct intobone and nonbone cell lines are shown in Fig. 7B. The relativenormalized activities of the wild type 2472-bp construct inthese lines were the same as those shown in Fig. 2. However, inFig. 7B, we have set the activity of the full-length construct at100% in each cell line and expressed the activities of the sitemutant and deleted constructs as a percentage of that in orderto clearly illustrate the drop caused by the 2-bp change. Al-though the mutation had no effect on the activity of the pro-moter in nonosteoblastic cells, it abolished 60% of the promoteractivity in the osteoblastic clone 4 and UMR cells, causing adrop nearly equivalent to the deletion of the entire sequencebetween –2472 and –178. Thus, the disruption of the homeodo-main binding site resulted in the ablation of almost all of theosteoblast selective activity of the 2472-bp Bsp promoter.

Dlx5 Regulation of the C Element Homeodomain Site—Wefurther characterized the activity of this element by examining

FIG. 5. Effect of A, B, or C internaldeletions on the activity of the p705construct. A, different combinations ofprimers (represented as arrows) wereused with p705 as a template to createPCR fragments containing sequences ei-ther upstream or downstream of the de-sired deletion, as described under “Exper-imental Procedures.” These fragmentswere ligated into pGL3Basic to create thep705DA, p705DB, and p705DC constructsrepresented here. The hatched bar repre-sents Bsp promoter sequence. The openbar is pGL3 sequence. Dashed lines indi-cate the sequences that are deleted fromthe final constructs. The end points of thedeleted promoter sequences are shownrelative to the start of transcription aboveeach construct. B, the deletion constructsshown in A, as well as the complete p705,were transfected into clone 4 and UMRcells. Their activities are shown as per-centages of p705 activity in each cell line.

Osteoblast Specificity of the Mouse Bsp Promoter 13913

its ability to be directly regulated by the Dlx5 homeoprotein, amember of the Dlx family of Drosophila distalless homologues.We chose this factor as the most likely known candidate for theC element-binding protein because it has been shown to beexpressed on bone and to be able to regulate the osteoblast-specific osteocalcin promoter (23, 35). Additionally, homozy-gous knockout of the Dlx5 gene produces mice with severeskeletal deformities, implying a crucial role in skeletogenesis(36). As represented in Fig. 8A, Dlx5 mRNA is expressed onlyin those cell lines that also produce BSP, thus qualifying it asa possible regulator of Bsp transcription. Furthermore, overex-pression of Dlx5 in COS7 cells (a highly transfectable, Dlx5-and BSP-null background) stimulated the activity of the Celement-driven minimal Bsp promoter construct 6-fold overcells transfected with a b-galactosidase control vector, but onlyif the homeodomain site was left intact (Fig. 8B, compare C w.t.and Cmut3 constructs).

Finally, gel shift assays performed with nuclear extract pre-pared from COS cells overexpressing Dlx5 demonstrated thesequence-specific binding of Dlx5 to the C element. These ex-periments produced two major shifted species (Fig. 8C, lane 6),the upper one of which (asterisk) was competed by the additionof a 50-, 100-, or 200-fold excess of unlabeled wild type Coligonucleotide (lanes 7–9) but not by the addition of Cmut3

oligonucleotide (lanes 10–12). Thus, this complex is dependenton the intact homeodomain site for binding. The lower specieswas competed by both wild type and mutant C oligos and wasalso found in control nuclear extracts made from COS7 cellstransfected with a b-galactosidase expression vector (Fig. 8C,lanes 2–4). Therefore, it does not represent homeodomain site-specific binding. To confirm that the specifically shifted speciesobserved was the result of Dlx5 binding, we repeated the gelshift assay with a monoclonal antibody (M2) recognizing aFLAG epitope tag built into the overexpressed Dlx5. This an-tibody blocked the appearance of the specific species whilehaving no effect on the lower, nonspecific band (Fig. 8C, lane13). Furthermore, this blocking action was counteracted by theaddition of FLAG peptide to compete for antibody binding,indicating that the protein complex observed contains Dlx5(Fig. 8C, lane 14). Addition of M2 antibody had no effect inbinding reactions with b-gal control extract (Fig. 8C, lane 5).Together, these experiments show that the Dlx5 gene is ex-pressed in our osteoblastic cell lines, and that the Dlx5 homeo-protein is able to bind the Bsp C element and stimulate itstranscriptional activity.

DISCUSSION

Of the known bone extracellular matrix constituents, bonesialoprotein is, perhaps, the most intimately associated withthe primary function of the differentiated osteoblast, namelythe mineralization of a collagenous matrix, yet little is knownabout the cis-acting elements that restrict its expression to themineralizing osteoblast. We previously reported the isolation ofa 2472-bp fragment of the murine Bsp promoter that is able todirect osteoblastic-selective expression of a luciferase cDNAunder growth conditions which are known to support osteoblastdifferentiation in cell culture (6). Here we describe the resultsof a systematic study to identify sequences within this frag-ment that may direct this tissue-specific expression.

We found that the –2472/141 base pair Bsp promoter, inaddition to its preferential expression in osteoblastic versusnonosteoblastic cell lines, was induced by collagen matrix pro-duction in parallel with the endogenous BSP message. Thisup-regulation in response to matrix interaction is indicative ofgenes whose expression is associated with osteoblast differen-tiation and provides further evidence that the 2472-bp pro-moter contains the requisite information to appropriately reg-ulate Bsp transcription during the differentiation process (37).Deletion of sequence from –2472 to –705 caused a marked dropin promoter activity in clone 4 cells, but not in UMR 106-01cells, which constitutively produce BSP. One possible explana-tion for this observation is that this region contains sequencesthat are necessary for the normal matrix-stimulated inductionof BSP but that are inactive in osteosarcoma cells which arefree from the matrix requirement. It will be interesting toidentify the cis-acting elements in this region of the Bsp pro-moter that are responsible for the differences in reporter ex-pression between these two osteoblastic cell lines. Character-ization of such elements may yield additional information onthe matrix-induced signaling pathways that are active inosteoblasts.

This present study, however, focuses on the isolation of se-quences that are important for promoter activity in osteoblasticversus nonbone cells, as these are likely to be the elements thatconfer bone specificity. We identified one such element, a con-sensus homeodomain binding site, between –199 and –192(TCAATTAA). Disruption of this sequence caused a 60% dropin transcriptional activity only in the osteoblastic cells andaccounted for almost all of the difference in activity betweenthe bone and nonbone cell lines used in this study. To ourknowledge, this is the first description of an osteoblast-specific

FIG. 6. Interaction of the C element with osteoblast nuclearproteins requires an intact homeodomain binding motif. A seriesof six double-stranded mutant C oligonucleotides was synthesized tospan the entire homeodomain binding motif. Each oligonucleotide con-tains a 2-bp mutation (as denoted by the brackets) in, or adjacent to, thehomeodomain consensus binding sequence. Although the correspondingmutations were made in the (–) strand, only the (1) strand is shown.The sequences of all mutants are identical to that of the wild type Coligonucleotide shown in Table I except for these 2-bp changes. Gelmobility shift assays were conducted with either UMR or clone 4 nu-clear extract as described under “Experimental Procedures.” Lane 1contains no extract. Lane 2 contains extract but no cold competitor. Theremaining lanes contain binding reactions (with extract) incubatedwith a 50- or 100-fold molar excess (increasing amounts indicated by thetriangles) of unlabeled wild type (w.t.) C oligo or the mutant C oligoindicated. The six shifted species observed are numbered 1–6 in orderfrom lowest to highest mobility.

Osteoblast Specificity of the Mouse Bsp Promoter13914

transcriptional element within the bone sialoprotein genepromoter.

There is a large body of genetic evidence implicating home-odomain proteins in the development of mineralized tissues,particularly with respect to members of the Dlx and Msx fam-ilies, which are known to be expressed in tissues that give riseto skeletal structures, and mutations in which result in thedisplay of severe skeletal phenotypes. For example, mice con-taining mutations in Msx1 show craniofacial abnormalitiesthat include cleft palate and absence of specific teeth (38), whilemice expressing a mutated Msx2 transgene exhibit the symp-toms of craniosynostosis, a disease characterized by prematureclosure of the cranial sutures (39). Furthermore, both Dlx1 andDlx2 knockout mice display abnormal formation of the skullbones derived from the first and second branchial arches (40,41), while mice deficient in both factors show additional defectsin dentition (42). In addition, in a recent study by Acampora etal. (36), Dlx5 –/– mice were shown to have the most severephenotype of all the known Dlx knockouts, displaying not onlycranial and mandibular malformations, but defects in longbone development as well.

At the molecular level, Towler and colleagues (43) reportedthat Msx2 is able to bind in vitro to a homeodomain binding sitebetween –84 and –92 in the rat osteocalcin promoter (ACTA-ATTGG) and that forced overexpression of Msx2 inhibited ex-

pression in MC3T3-E1 cells of a 199-base pair promoter frag-ment containing this site. More recently, Newberry et al. (23)reported that overexpression of Dlx5 was able to reverse thisinhibition through heterodimerization with Msx2, further sup-porting the theory that these two families of homeoproteins arecapable of influencing the transcription of an osteoblast-spe-cific gene. Our finding that Dlx5 can bind the Bsp C elementand stimulate its activity is not only consistent with this the-ory, but extends it in that it represents the first example ofdirect activation of an osteoblastic-specific cis-acting elementby a Dlx factor.

Two other studies which also describe positive regulationthrough an osteoblast-selective homeodomain binding elementin vivo were performed by two independent groups on themurine and rat a1(I) collagen gene promoters. Taken together,these studies found that a homeodomain binding site (TTA-ATTA), which is conserved in the human, rat, and murinepromoters, was required for osteoblastic-selective activity intransgenic mice and that this site bound in vitro to a complexfound only in osteoblastic cells. Furthermore, both in vitrobinding and transcriptional activity were lost if the homeodo-main consensus sequence was mutated (44, 45).

Our discovery of an osteoblastic-specific homeodomain bind-ing site in the Bsp promoter that is similar to the one shown todirect expression of a1(I) collagen may provide new insight into

FIG. 7. Comparison of enhancer ac-tivities of wild type and mutant C el-ement sequences. A, a trimer of theCmut3 double-stranded oligonucleotidewas inserted into the p49BSP vector inthe forward orientation (row 2), and itstranscriptional activity in clone 4 andUMR cells was compared with that of theconstruct containing the wild type C ele-ment (row 3). Values shown are fold in-duction of each construct over that of theinsertless p49BSP vector (row 1). B, the2-bp mutation corresponding to that inthe Cmut3 oligonucleotide was intro-duced into p2472 by site-directed mu-tagenesis using primers Cmut3SD(1) and(–) (see Table I). The resulting p2472mut3construct was transfected into clone 4,UMR, C2C12, 3T3-L1, and F9 cells, andits transcriptional activity was comparedwith that of the wild type p2472, as wellas the p178 (which lacks A, B, and Celements) and p49 (minimal promoter)deletion constructs. Normalized lucifer-ase activities are shown as a percentageof the wild type p2472 (p2472 w.t.) con-struct activity, with the p2472 w.t. activ-ity set at 100% in each cell line in order toemphasize the percentage of drop attrib-uted to the mutation in each cell type.Normalized luciferase activities of thewild type promoter in osteoblastic andnonosteoblastic cell lines were similar toresults shown in Fig. 2.

Osteoblast Specificity of the Mouse Bsp Promoter 13915

the mechanism of osteoblast gene expression. If it is true, as iswidely held, that BSP plays a crucial role in the nucleation ofmineral formation in the collagenous matrix of bone, then it isplausible to speculate that the col1a(I) and Bsp promoters arecoordinately regulated by a shared set of osteoblast-specific

transcription factors. Given the new data reported here, it ispossible that one such factor may be a homeodomain proteincomplex that binds to both collagen and Bsp promoters, ena-bling their transcription in osteoblasts, and thus coordinatingboth the deposition and mineralization of the collagenous ma-

FIG. 8. Transcriptional activationand in vitro binding of the C elementby Dlx5. A, correlation of Dlx5 and BSPexpression. RNA was isolated fromMC3T3-E1 clone 4, UMR 106-01, C2C12,F9, and COS7 cells grown for 6 days un-der differentiating conditions as describedunder “Experimental Procedures” (exceptCOS cells, which were grown for 48 h) andassayed for Dlx5 and BSP expression byNorthern analysis. Hybridization with aprobe for 18 S rRNA was performed to con-firm equal loading of all lanes. B, the vec-tors shown in Fig. 7A, containing either theC wild type (C wt) or Cmut3 sequencesupstream of the 49 Bsp minimal promoter,were each transfected into COS7 cells witheither the pcDNA3Dlx5 or pCMV5-b-galexpression vector and assayed for lucifer-ase reporter activity after 48 h. Valuesshown are fold stimulation of Renilla nor-malized firefly luciferase activities ofDlx5 transfected samples over those ofb-galactosidase-transfected samples foreach construct. C, binding of Dlx5 to the Celement. 5 mg per binding reaction of nu-clear extract from COS7 cells transfectedwith either pCMV5-b-gal expression plas-mid (lanes 2–5) or pcDNA3-Dlx5 plasmid(lanes 6–16) was subjected to gel mobilityshift analysis with the wild type C oligo-nucleotide. Lane 1 contains no extract.50-, 100-, and 200-fold molar excesses ofwild type C oligo (lanes 7–9) or Cmut3oligo (lanes 10–12) were used as unla-beled competitors. A 200-fold excess of ei-ther wild type or mutant C oligo was usedin lanes 3 and 4, respectively. The speciesthat represents homeodomain site-spe-cific binding to the probe in lanes 6–16 ismarked with an asterisk. A faster migrat-ing band that is not homeodomain site-specific appears in both b-gal and Dlx5reactions and is competed by both wildtype and mutant sequences. M2 mono-clonal anti-FLAG epitope antibody (lane13), but not control IgG (lane 15), added tothe binding reaction blocked formation ofthe specific species. Addition of an ap-proximately 100-fold molar excess ofFLAG peptide reversed this blockage(lanes 14 and 16). Antibody addition to ab-gal control extract binding reaction(lane 5) had no effect.

Osteoblast Specificity of the Mouse Bsp Promoter13916

trix. Such a complex, in conjunction with other factors such asCbfa1, might function as a central determinant of osteoblastdifferentiation. Our data, combined with the recent descriptionof the Dlx5 knockout mice mentioned above, strongly support arole for Dlx5 as one such determinant. However, in consideringthis scenario, it is important to recognize that the results wepresent here, as well as the aforementioned osteocalcin pro-moter studies, merely demonstrate that Dlx5, or a relatedhomeoprotein, may regulate osteoblast-specific transcription invivo; they do not positively identify Dlx5, or any other knownfactor, as such a regulator. Furthermore, the apparently im-precise DNA binding specificities of homeoproteins in vitroprohibits their identification merely by examination of the se-quences to which they bind in individual promoters. Our ongo-ing studies are aimed at identifying the factor that binds thenewly characterized osteoblast-specific element in the Bsp pro-moter and at elucidating its role in the expression of the osteo-blast phenotype.

REFERENCES

1. Fujisawa, R., Nodasaka, Y., and Kuboki, Y. (1995) Calcif. Tissue Int. 56,140–144

2. Goldberg, H. A., Warner, K. J., Stillman, M. J., and Hunter, G. K. (1996)Connect Tissue Res. 35, 385–392

3. Chen, J., Shapiro, H. S., and Sodek, J. (1992) J. Bone Miner. Res. 7, 987–9974. Ducy, P., and Karsenty, G. (1995) Mol. Cell. Biol. 15, 1858–18695. Komori, T., Yagi, H., Nomura, S., Yamaguchi, A., Sasaki, K., Deguchi, K.,

Shimizu, Y., Bronson, R. T., Gao, Y. H., Inada, M., Sato, M., Okamoto, R.,Kitamura, Y., Yoshiki, S., and Kishimoto, T. (1997) Cell 89, 755–764

6. Benson, M. D., Aubin, J. E., Xiao, G., Thomas, P. E., and Franceschi, R. T.(1999) J. Bone Miner. Res. 14, 396–405

7. Ducy, P., Desbois, C., Boyce, B., Pinero, G., Story, B., Dunstan, C., Smith, E.,Bonadio, J., Goldstein, S., Gundberg, C., Bradley, A., and Karsenty, G.(1996) Nature 382, 448–452

8. Defranco, D. J., Lian, J. B., and Glowacki, J. (1992) Endocrinology 131,114–121

9. Kim, R. H., Li, J. J., Ogata, Y., Yamauchi, M., Freedman, L. P., and Sodek, J.(1996) Biochem. J. 318, 219–226

10. Ogata, Y., Niisato, N., Furuyama, S., Cheifetz, S., Kim, R. H., Sugiya, H., andSodek, J. (1997) J. Cell. Biochem. 65, 501–512

11. Ogata, Y., Yamauchi, M., Kim, R. H., Li, J. J., Freedman, L. P., and Sodek, J.(1995) Eur J Biochem 230, 183–192

12. Yang, R., and Gerstenfeld, L. C. (1996) J. Biol. Chem. 271, 29839–2984613. Kerr, J. M., Hiscock, D. R., Grzesik, W., Robey, P. G., and Young, M. F. (1997)

Calcif. Tissue Int. 60, 276–28214. Sudo, H., Kodama, H.-A., Amagai, Y., Yamamoto, S., and Kasai, S. (1983)

J. Cell Biol. 96, 191–198

15. Xiao, G., Cui, Y., Ducy, P., Karsenty, G., and Franceschi, R. T. (1997) Mol.Endocrinol. 11, 1103–1113

16. Midura, R. J., McQuillan, D. J., Benham, K. J., Fisher, L. W., and Hascall,V. C. (1990) J. Biol. Chem. 265, 5285–5291

17. Yaffe, D., and Saxel, O. (1977) Nature 270, 725–72718. Majeska, R. J., Rodan, S. B., and Rodan, G. A. (1980) Endocrinology 107,

1494–150319. Green, H., and Meuth, M. (1974) Cell 3, 127–13320. Strickland, S., and Mahdavi, V. (1978) Cell 15, 393–40321. Gluzman, Y. (1981) Cell 23, 175–18222. Franceschi, R. T., Iyer, B. S., Parikh, U. K., and Pineron, G. J. (1992) J. Bone

Miner. Res. 7, S21823. Newberry, E. P., Latifi, T., and Towler, D. A. (1998) Biochemistry 37,

16360–1636824. Chomczynski, P., and Sacchi, N. (1987) Analytical Biochemistry 162, 156–15925. Thomas, P. S. (1980) Proc. Natl. Acad. Sci. U. S. A. 77, 5201–520526. Young, M. F., Ibaraki, K., Kerr, J. M., Lyu, M. S., and Kozak, C. A. (1994)

Mamm. Genome 5, 108–11127. Franceschi, R. T., Iyer, B. S., and Cui, Y. (1994) J. Bone Miner. Res. 9, 843–85428. Dignam, J. D., Martin, P. L., Shastry, S., and Roeder, R. G. (1983) Methods

Enzymol. 101, 582–59829. Maxam, A. M., and Gilbert, W. (1977) Proc. Natl. Acad. Sci. U. S. A. 74,

560–56430. Chen, J., Thomas, H. F., Jin, H., Jiang, H., and Sodek, J. (1996) J. Bone Miner.

Res. 11, 654–66431. Franceschi, R. T., and Iyer, B. S. (1992) J. Bone Miner. Res. 7, 235–24632. Schinke, T., and Karsenty, G. (1999) J. Biol. Chem. 274, 30182–3018933. Ohkuma, Y., Horikoshi, M., Roeder, R. G., and Desplan, C. (1990) Proc. Natl.

Acad. Sci. U. S. A. 87, 2289–229334. Catron, K. M., Iler, N., and Abate, C. (1993) Mol. Cell. Biol. 13, 2354–236535. Ryoo, H. M., Hoffmann, H. M., Beumer, T., Frenkel, B., Towler, D. A., Stein,

G. S., Stein, J. L., van Wijnen, A. J., and Lian, J. B. (1997) Mol. Endocrinol.11, 1681–1694

36. Acampora, D., Merlo, G. R., Paleari, L., Zerega, B., Postiglione, M. P., Mantero,S., Bober, E., Barbieri, O., Simeone, A., and Levi, G. (1999) Development126, 3795–3809

37. Franceschi, R. T. (1999) Crit. Rev. Oral Biol. Med. 10, 40–5738. Satokata, I., and Maas, R. (1994) Nat. Genet. 6, 348–35639. Liu, Y. H., Kundu, R., Wu, L., Luo, W., Ignelzi, M. A., Jr., Snead, M. L., and

Maxson, R. E., Jr. (1995) Proc. Natl. Acad. Sci. U. S. A. 92, 6137–614140. Qiu, M., Bulfone, A., Martinez, S., Meneses, J. J., Shimamura, K., Pedersen,

R. A., and Rubenstein, J. L. (1995) Genes Dev. 9, 2523–253841. Qiu, M., Bulfone, A., Ghattas, I., Meneses, J. J., Christensen, L., Sharpe, P. T.,

Presley, R., Pedersen, R. A., and Rubenstein, J. L. (1997) Dev. Biol. 185,165–184

42. Thomas, B. L., Tucker, A. S., Qui, M., Ferguson, C. A., Hardcastle, Z.,Rubenstein, J. L., and Sharpe, P. T. (1997) Development 124, 4811–4818

43. Towler, D. A., Rutledge, S. J., and Rodan, G. A. (1994) Mol. Endocrinol 8,1484–1493

44. Bedalov, A., Salvatori, R., Dodig, M., Kronenberg, M. S., Kapural, B.,Bogdanovic, Z., Kream, B. E., Woody, C. O., Clark, S. H., Mack, K., Rowe,D. W., and Lichtler, A. C. (1995) J. Bone Miner. Res. 10, 1443–1451

45. Rossert, J. A., Chen, S. S., Eberspaecher, H., Smith, C. N., and de Crombrugghe,B. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 1027–1031

Osteoblast Specificity of the Mouse Bsp Promoter 13917