Eur. J. Biochem. 207,649 - 659 (1 992)

0 FEBS 1992

Characterization of the promoter region of the porcine upn (osteopontin, secreted phosphoprotein 1) gene Identification of positive and negative regulatory elements and a ‘silent’ second promoter Qi ZHANG, Jeffrey L. WRANA and Jaro SODEK MRC Group in Periodontal Physiology and the Department of Biochemistry, University of Toronto, Canada (Received February 24/May 4, 1992) - EJB 920244

Osteopontin (secreted phosphoprotein-I, Opn) is a phosphorylated glycoprotein expressed by transformed cells, macrophages, activated T-lymphocytes, specialized epithelial cells and bone cells that is characteristically enriched in milk and in the mineralized matrix of bone. The synthesis of Opn by bone cells is regulated by glucocorticoids and growth factors, which promote bone formation, and by the osteotropic hormone calcitriol(1,25-dihydroxycholecalciferol)and retinoic acid, which mediate bone resorption, indicating a bifunctional role for this protein in bone remodelling. To study the transcriptional regulation of the opn gene, two genomic clones (10 and 15 kb) encoding the opn gene were isolated from a porcine liver genomic library cloned into /z phage. From the 15-kb clone a 4-kb EcoRI fragment containing the first two exons and 2.6 kb of the 5’ flanking region of the opn gene was sequenced, and the transcriptional start site determined by primer extension analysis and S1 nuclease mapping. To identify the opn promoter, chimeric chloramphenicol acetyltransferase constructs were prepared using fragments from the first intron and the 5’ flanking region of the opn gene. Transient transfection of porcine bone cells with these constructs showed strong promoter activity located within 74 bp upstream from the transcription initiation site. Within this region a TATA sequence, TTTAAA, was identified at positions - 26 to - 31. However, the highest transcription rate was observed in a construct extending 180 bp upstream that included a CCGCCC Spl binding sequence (- 63 to - 68), and an API site ( - 74 to - 80). Further upstream in the 5‘ flanking region and within the first intron of the opn, a number of consensus sequences could be identified. Chimeric constructs containing a GGGTCAtatGGTTCA direct repeat consensus sequence for a vitamin D3 response element located at nucleotides -2245 to -2259 responded to the addition of 0.1 pM calcitriol by a 2.5-fold stimulation of transcription, although a > 2-fold increase was also observed in shorter constructs - 180 to -905 lacking such a consensus sequence. Promoter activity was also exhibited by a region containing a TTTAAA sequence in the first intron that corresponded to the putative promoter site reported for mouse opn in macrophages (Miyazaki, Y., Setoguchi, M., Yoshida, S., Higuchi, Y., Akizuki, S. & Yamamoto, S. (1990) J . Biol. Chem. 265, 14432-14438). However, primer extension and hybridization analysis of both porcine and murine monocyte/macrophage and bone mRNA failed to reveal an Opn mRNA transcribed from the alternative promoter, indicating that the same promoter regulates transcription of the opn gene in monocytes and macrophages as well as in bone.

Osteopontin (secreted phosphoprotein-I, Opn) is a prominent glycoprotein of mineralized bone [I - 51 that binds to hydroxyapatite [l] and has cell attachment properties [2, 61. The protein was first isolated from rat bone and characterized Correspondence to Qi Zhang, MRC Group in Periodontal Physiology, Medical Science Building, University of Toronto, Toronto, Ontario, M5S 1A8, Canada Abbreviations. Opn, osteopontin; opn, gene for osteopontin; sparc, gene for SPARC, secreted protein rich in cysteine (osteonectin); CAT, bacterial chloramphenicol acetyltransferase; calcitriol, 1,25dihydroxycholecalciferol (sometimes known as 1,25-dihydroxyvitamin D3); ROS 1712.8, rat osteosarcoma cell line 1712.8; SV40, simian virus 40. Note. The complete nucleotide sequence for the 4-kb EcoRI fragment of opn gene which contains the promoter, 2.6-kb 5’ flanking sequence, exons 1 and 2 and intron 1 has been submitted to Genbank (access no. M84121).

as a 44-kDa phosphoprotein in which 12 serines and a threonine are phosphorylated [3]. Opn species migrating on SDS/ PAGE at about 67 kDa have also been isolated from human [4], porcine [5]and bovine [7] bone. Comparison of the primary structures of mammalian Opns determined from cDNA nucleotide sequencing [2, 5, 8, 91 has revealed that a repeat of aspartic acid residues [2], thought to mediate the binding to hydroxyapatite, a (G)RGD(S) cell-attachment site, and a number of sites of serine phosphorylation, are all conserved. Although the precise function of Opn in bone is not known, it may mediate the attachment of bone cells [2], including osteoclasts [lo], to the mineralized matrix. In addition, the rapid association of the protein with hydroxyapatite during bone formation [II, 121 together with the specific fragmentation of the protein [5]indicate that Opn may also be involved in regulating the growth and maturation of the hydroxyapatite

650 crystals. Such a role in biological mineralization is also indicated by the stimulation of Opn synthesis by the osteotropic hormone calcitriol (1,25-dihydroxycholecalciferol)and retinoic acid [13, 14, 271. Although Opn is abundant in bone and is typically expressed by osteoblastic cells [15, 161, it is also synthesized by odontoblasts in dentin, chondrocytes in mineralizing cartilage, and by kidney and epithelial cells involved in Ca2+ metabolism 117, IS]. Moreover, the transformation-associated protein known as pp69, which is regulated by tumor promoters and is reported to be expressed by disseminating carcinomas [19], is a product of the same gene 1201 which is located on chromosome 5 in mouse [21]. The presence of Opn in milk [22] and in the implantation tissues of the uterus [18] indicate further that this protein has a multiplicity of functions that may involve the ability of Opn to mediate cell attachment and bind Ca2+ions. These functions could, in part, be modulated by variations in the post-translational modification of the protein, including sulfation [23j and glycosylation [24] as well as phosphorylation [24- 261. Recent studies have shown that at least two forms of the protein that can be separated on SDSjPAGE are synthesized by rat bone [25, 261 and kidney cells [24]. The faster migrating form produced by rat bone cells corresponds to the 44-kDa protein extracted from rat bone and is more highly phosphorylated 125, 261. A third, non-phosphorylated, form of Opn is also synthesized by kidney cells [24]. The control of the opn gene transcription appears to be influenced by growth and differentiation factors, as well as by osteotropic hormones and tumor promoters, since the expression of Opn in bone [27] and other tissues appears to be developmentally regulated [18]. Recently, the exonjintron organization of the murine opn gene was reported and a putative promoter region identified [28]. In this study six exons were identified based on a cDNA sequence obtained from RNA prepared from activated T-lymphocytes [29]. However, an additional exon at the 5‘ end of the mouse gene was demonstrated in a more recent study in which the gene structure was based on a cDNA isolated from epidermal cells [30]. Since the 5’ untranslated region of rat [2] and porcine [31] bone cDNAs are homologous to the additional exon and promoter activity that has been demonstrated in the 5’ flanking region of this exon [30], the apparent discrepancies in the exon organization of the opn gene could reflect the utilization of alternative promoters in different tissues. To investigate the transcriptional regulation of the opn gene, we have isolated a 4-kb gene fragment containing the promoter and its 3’ and 5’ flanking regions from an EcoRI restriction digest of a Z 5-kb porcine genomic opn clone. We show that this promoter directs transcription in monocytes/ macrophages as well as bone cells and that it is regulated by caicitriol.

MATERIALS AND METHODS Tissue distribution of Opn mRNA by Northern analysis The mRNA was purified from a variety of tissues freshly dissected from fetal pigs, using the procedure as described previously [5]. Poly(A)-rich RNA was then purified by oligo(dT)-cellulose chromatography [32] and the purity and quantity determined spectroscopically. Aliquots containing 5 pg mRNA were separated on 1.2% formaldehyde/agarose gels and transferred to a BioTrans nylon membrane under vacuum (Vacuumblot, Pharmacia, Uppsala, Sweden). The

RNA was probed with porcine Opn cDNA [31] using previously described conditions [5]. Hybridization was detected by exposure of the blots to XAR-5 film (Kodak, Rochester, NY) using two intensifying screens at -70°C for 3-24 h. Total RNA from various cells were also isolated 1261and 20 pg of each RNA was subjected to Northern analysis probed with porcine or rat Opn cDNA as described above. Northern hybridization was also used to determine whether an alternative transcript is generated through the utilization of a second promoter corresponding to the putative promoter reported by Miyazaki et al. [28]. Samples of total RNA (20 pg) isolated from porcine calvariae, from porcine calvarial cells grown in culture and from monocytic cells isolated from fresh porcine blood and stimulated by incubation with lipopolysaccharide (1 pg/ml) were separated as described above and transferred to BioTrans nylon membrane by capillary action. The RNA was hybridized with a 500 bp AvrII-XbaI fragment immediately upstream of exon 2, corresponding to the untranslated region of exon 1 reported by Miyazaki [28]. The same blots were also probed with a full length porcine Opn cDNA [31] generated by random priming. Isolation of the porcine opn gene A porcine liver genomic library constructed in EMBL3 was purchased from Stratagene (La Jolla, CA) and screened with a full-length Opn cDNA [31] labeled with [32P]dCTP (Amersham, IL) using the random priming method [32]. From a screen of approximately 5.0 x lo5 plaques, two positive plaques were detected and, after secondary and tertiary screening, each phage clone was amplified by a large-scale preparation [32]. DNA purified from the two phages was digested by various restriction enzymes (Pharmacia) and mapped by Southern hybridization analysis using the fulllength cDNA as probe. One of the phage clones was found to contain two EcoRI fragments of 3 and 3.5 kb which did not hybridize to the Opn cDNA probe and a 4.0-kb EcoRI fragment which hybridized strongly. Each EcoRI fragment was then subcloned into plasmid pT7T3 (Pharmacia). Preliminary sequencing of the 4-kb subclone revealed that it contained the 5‘ region of the opn gene. The complete sequence of the 4-kb fragment was then obtained employing the dideoxynucleotide chain termination procedure [34] with Sequenase (United States Biochemical, OH) and using a series of nested deletions generated by timed exonuclease 111 (Pharmacia) digestions [351. Primer extension analysis A synthetic oligonucleotide of 20 nucleotides with the sequence S’-GGAGCAATGCTGATGCGGCT-3’corresponding to the antisense sequence of porcine bone Opn cDNA [31] from positions 31 to + 50 was used for primer extension analysis of the transcriptional start site. Poly(A)-rich RNA and total RNA were isolated from fetal porcine calvarial bone as previously described [36]. To determine the presence of an alternative transcriptional start site, another 20-residue synthetic oligonucleotide, 5’-GAGGCACAGTTGATGTCTTG-3’ corresponding to the antisense sequence of the untranslated region immediately upstream from the coding region of exon 1, positions + 166 to + 147 as reported for the mouse gene by Miyazaki 1281, and a 20-residue synthetic oligonucleotide, S’-CTTGGCTGGTTTCCTCCGAG-3’ corresponding to the antisense sequence of exon Z immediately upstream from intron 1, positions + 79 to + 60 as reported

+

651 for the mouse gene by Craig et al. [30], were utilized for primer extension using total RNA isolated from mouse monocytic and macrophage cell lines 5774, RAW267, and IC21. The extension was performed as described in detail by Sambrook et al. [32]. In brief, the primer labeled at the 5’ end using [y32P]ATP(ICN, CA) and polynucleotide kinase (Pharmacia) was annealed to 10 pg poly(A)-rich RNA or 50 pg total RNA at 45°C for 8 h in reaction buffer (40 mM Pipes, 1.O mM EDTA, 0.4 M NaCl, 80% formamide), and extended at 37°C for 2 h in reverse transcriptase buffer (50 mM Tris, pH 7.6, containing 60mM KCl, 1 0 m M MgC12, 1.0mM dithiothreitol, 1.O unit/pl placental RNase inhibitor, 50 pg/ml actinomycin D) and 50 units asian myeloblastosis virus (AMV) reverse transcriptase with 1.O mM dNTPs (Pharmacia). To provide a marker for determining the exact length of the extended primers, the DNA sequencing reaction was carried out. The extended primer products and the marker sequencing reaction were analyzed on a denaturing 6% polyacrylamide/urea sequencing gel followed by an overnight exposure to XAR-5 X-ray film (Eastman-Kodak, Rochester, NY).

S1 nuclease mapping A SphI-SmaI fragment of the opn gene was first dephosphorylated at the 5‘ end by calf intestinal alkaline phosphatase (Pharmacia) and then radiolabeled with [y-”P]ATP by bacterial phage T4 polynucleotide kinase (Pharmacia).The 32P-labeledSphI -SrnaI fragment was hybridized with 50 pg total RNA from porcine calvariae in 40 mM Pipes, 1.0 mM EDTA, 0.4 M NaCl, and 80% formamide. The hybridization mix was first heated to 85°C for 10 min and subsequently incubated for 16 h at 50°C. Following hybridization, the mixture was diluted with 300 p1 nuclease-S1 mapping buffer (0.28 M NaCl, 0.05 M sodium acetate, p H 4.5) containing 4.5 mM ZnSO,, 20 pg/ml single-stranded carrier DNA, 500 units nuclease S1 (Pharmacia), and incubated at 3 7 T for 2 h. The S1 digestion was terminated by chilling the reaction to 0°C and adding 80 p1 nuclease-S1 stop mixture (4 M ammonium acetate, 50 mM EDTA, pH 8.0, 50 pg/ml tRNA). The nuclease-S1 resistant DNARNA hybrid was extracted with phenol/chloroform, precipitated by ethanol, and analyzed on a 6% polyacrylamide/urea sequencing gel. Bacterial chloramphenicol acetyltransferase vector constructions

A vector (pCATB) containing the gene for bacterial chloramphenicol acetyltransferase (CAT) was purchased from Promega (Madison, WI). The promoterless/enhancerless basic vector contained a polylinker with a variety of restriction sites at the 5’ end of the coding region of the CAT gene. To construct popnCAT+51/-905, a 1596-bp SphI( -905)-XbaI (+691) fragment of the porcine opn gene was first ligated into the SphI-XbaI site of the pCATB vector. This construct was then digested with restriction enzymes SmaI ( + 51) and XbaI (+691/on pCATB vector) and separated from the SmaI XbaI fragment on 1O/O low-melting-point agarose gel (Bethesda Research Laboratories, MD). The linearized plasmid vector containing the SphI-SmaI fragment of opn (-905 to +51) was circularized by ligation, using Klenow DNA polymerase and T4 DNA ligase, in the low-melting-point gel. To generate popnCAT - 180/- 905, the construct containing the SphI XbaI fragment (see above) was digested with restriction enzymes PstI (- 180) and XbaI (+691/on pCATB vector) and

the subsequent procedures were the same as those used in the construction of popnCAT+ 51/- 905. To generate popnCAT 51/ 180, which corresponded to a PstI - SmaI fragment of the opn gene (-180 to +51), popnCAT+51/ - 905 was digested with SphI (- 905/on pCATB vector) and PstI ( - 180) and then recircularized as described above. To construct popnCAT+1191/-2615, an AvvII (+1191) -Hind111 (on pT7T3 vector) fragment was released from the 4-kb opn gene subcloned in pT7T3 vector and ligated into the Hind111 - XbaI site in pCATB vector. This construct retained the pT7T3 polylinker region at the 5’ end of the opn gene. A similar method was used to construct popnCAT + 691/ 261 5. Prior to digestion of popnCAT + 691/ - 261 5 with XbaI (+691/on pCATB vector) and SmaI (+51) to generate popnCAT 511- 261 5, the pT7T3 polylinker region was digested from the popnCAT+691/-2615. To construct popnCAT 51/ - 74, an Ex0111 time-deleted 4-kb opn gene containing -74 to +1328 was digested with HindIII (on pT7T3 vector) and XbaI (+691), and the HindIII - XbaI fragment was ligated into the HindIII-XbaI site in the pCATB vector. The pCATB vector containing the HindIII - XbaI fragment was then digested with XbaI (+691/on pCATB vector) and SmaI (+51). The construct popnCAT + 149/ + 1191 was generated by using a similar method. All constructs were confirmed by restriction mapping and sequencing. The constructs were transfected into competent NM 522 cells (Pharmacia), amplified by large-scale preparation and purified by passage through a Sepharose ‘Fast Q’ resin (Pharmacia) operated on an FPLC (Pharmacia) system. Supercoiled and relaxed DNAs, which did not bind to the resin, were collected in the void volume. The promoterless/ enhancerless basic vector (pCATB) and the SV40 (simian virus 40) promoter/enhancer-containing vector (pCATC) were used as negative and positive controls respectively for the transfection assays.

+

-

-

+ +

Cell culture and transfections

Porcine bone cells were obtained from fetal calvariae utilizing sequential digestions with collagenase. The latereluting osteoblastic cell populations were used in these experiments. The porcine bone cells and ROS 17/2.8 cells (rat osteosarcoma cell line 17/23) were cultured in a-minimal essential medium, supplemented with 5% fetal bovine serum and antibiotics. The day before transfection, exponentially growing cells were seeded at a density of 3.5 x 10’ (ROS 17/ 2.8) and 4 x lo5 (porcine calvarial cells) cells/60-mm dish. Each dish was transfected with 10 pg of one of the popnCAT constructs described above using the calcium phosphate coprecipitation method [32]. This was followed 24 h later by a 3-min shock with 15% glycerol, and a further 48-h incubation in culture medium containing 5% fetal bovine serum and antibiotics prior to the preparation of cell extracts for enzyme assays or a 12-h incubation in the above culture medium followed by a 48-h incubation in the same medium with 0.1 pM calcitriol. The cell culturing with calcitriol stimulation was also performed at 12,24 and 72 h. Control for transfection efficiency was achieved by co-transfecting cells with 3 kg plasmid vector pCHl10 (Pharmacia) containing the pgalactosidase gene under the control of the SV40 early promoter and enhancer. Mouse macrophage and monocyte cell lines 5774, RAW264 and IC21 (ATCC) were cultured in Dulbecco’s modified Eagle’s medium containing 5% fetal bovine serum and antibiotics (J774 and RAW264) and RPMI 1640 medium supplemented with 5% fetal bovine serum and

652 antibiotics (IC21). A porcine monocyte preparation isolated from fresh porcine blood by isopycnic centrifugation [33] was cultured in RPMI 1640 medium supplemented with 5% fetal bovine serum and antibiotics.

A

18s

Assays for CAT and P-galactosidase activities Assays for the activities of both CAT and P-galactosidase were carried out essentially as described by Sambrook [32]. To assay for CAT activity, 50 pl cell extract (total volume 150 pl) was heated for 10 min at 62OC to inactivate endogenous deacetylases, then mixed with 60 p1 0.25 M Tris/HCl, pH 7.8, 10 p1 acetyl coenzyme A (5 mg/ml, Promega) and 5 pl [ ''C]chloramphenicol (0.025 mCi/ml, Amersham). The cell extract was then incubated at 37°C for 2-8 h. For direct quantitation of the acetylated chloramphenicol, the reaction was terminated by addition of 300 p1 xylene. After a brief centrifugation, the entire upper xylene phase containing the acetylated chloramphenicol was collected and back extracted twice with 100 pl 0.25 M Tris/HCl, pH 8.0. The radioactivity in a 2 0 0 4 aliquot of the extracted xylene phase was then determined by liquid scintillation counting. To analyze the reaction products by thin-layer chromatography, the reaction was terminated by the addition of 500 p1 ethyl acetate. The ethyl acetate organic phase was collected, vacuum-dried and redissolved in 20 p1 ethyl acetate from which a volume normalized by B-galactosidase activity was analyzed by thin-layer chromatography on silica plates (Baker, NJ). To assay for P-galactosidase activity, 3 pl of a magnesium solution (0.1 M MgCl,, 4.5 M 2-mercaptoethanol), 66 p1 onitrophenyl ,8-d-galactopyranoside (Sigma, MO) in 0.1 M sodium phosphate, pH 7.5, SO p1 cell extract, and 181 p1 0.1 M sodium phosphate were mixed and incubated for 8 h at 37°C. The reaction was stopped by adding 500 pl 1 M Na,CO, and quantitated by measuring the absorbance at 420 nm. The /Igalactosidase activity determined per dish was used to normalize the total CAT activity obtained from the dish to correct for transfection efficiency. The relative CAT activity was thereby expressed as CAT activitylb-galatosidase activity.

RESULTS Poly(A)-rich RNA from various tissues was analyzed by Northern hybridization for the presence of Opn mRNA using a radiolabeled full-length porcine Opn cDNA probe [31]. Strong hybridization was observed in calvarial bone, with a significantly weaker signal in skin, kidney and placenta. N o hybridization was evident in RNA samples from pancreas lung, small intestine, liver, muscle and brain (Fig. 1A). These results were consistent with the distribution of Opn mRNA observed previously in rat tissues [18, 271. The expression of Opn was also analyzed in bone cells and various monocytes and macrophages maintained in culture (Fig. 1B). Strong hybridization was obtained for Opn mRNA in the mouse monocyte/macrophage lines 5774 and RAW 264 with slightly lower hybridization in the mouse macrophage cell line IC21. Similar levels of hybridization were evident in porcine calvarial bone cells and rat osteosarcoma cells (ROS 17/2.8), and a relatively lower level of hybridization in porcine blood monocytes activated with lipopolysaccharide. However, Opn mRNA could not be detected in unstimulated blood monocytes (results not shown). From a screen of 5 x lo5 plaques prepared from a porcine liver genomic library in 1 phage, two clones encoding about

+

OPN +

B

* OPN * 18s

Fig. 1. Tissue distribution and cellular expression of Opn mRNA. Total RNA extracted from various porcine tissues (A), from porcine and murine bone cells and monocytes/macrophages (B) was probed with Opn [32P]cDNAas described in Materials and Methods. The position o f the 18s ribosomal RNA (1 8 S) in the RNA preparation visualized by ethidium bromide staining is indicated, as is the position of the porcine Opn mRNA (OPN) that specifically hybridized with the probe. Comparable quantities of RNA are present in each lane as indicated by the intensity of the ribosomal RNAs stained with ethidium bromide (not shown). Exon 1

E

-2000

-1500

-1WO

Exon 2

-500

Fig. 2. Structure of the cloned EcoRl opn gene fragment. A schematic representation of the 4.0-kb EcoRI fragment containing the 5' region o f opn gene. The clone contains two exons (exon 1 and 2) separated by an intron of N 1.1 kb, and 2.6-kb 5' flanking sequence of the opn gene. Exon 1 contains the 5' untranslated region of Opn mRNA, whereas exon 2 comprises the signal peptide and the N-terminal amino acids of the Opn protein. The translation start codon, ATG, and major restriction sites are indicated.

15 kb of opn genomic DNA were identified using the fulllength Opn cDNA radiolabeled probe [31]. Each clone was isolated by secondary and tertiary screening. From the initial sequencing of subclones obtained from one of the original 15kb clones, a 4-kb subclone, which hybridized to the Opn cDNA in a Southern blot analysis, was recognized to encode for the first 18 amino acids of pre-Opn together with a portion of the 5' untranslated region of Opn mRNA. Complete sequencing of this fragment revealed that it contained two exons (exons 1 and 2) separated by an intron of about 1.I kb, and 2.6 kb of 5' flanking region of the opn gene (Fig. 2). Exon 1 contained 84 bp of a 5' untranslated region of Opn mRNA, whereas exon 2 contained 13 bp of untranslated sequence followed by 54 bp that encoded a signal peptide of 16 amino acids and the first two amino acids of the secreted Opn protein

PI.

Both primer extension analysis and S1 nuclease mapping of the Opn mRNA were performed to determine precisely the transcriptional start site for the opn gene (Fig. 3). Primer extension was performed using a synthetic 20-residue

653 A

1

G A T C

B

2

TATAbax +I SmaI

+3

+z

-1 -?

-3-

-

O p n 5’ flanking region 1

Protected fragments

probe

-905

+5 1

+2 +1 -2

-

I

J

S1 Mapping

- 1 --i)

3

4

D

Primerextended products

Primer extension

Primer

Fig. 3. Determination of the transcription initiation site of the opn gene. Both primer extension and S1 nuclease mapping were performed to determine the transcription start sites. (A) Results of primer extension (lanc 1 ) and S1 mapping (lane 2) are compared with a marker sequence (see Materials and Methods) as indicated. The major transcription start site determined by primer extension (indicated as 1) is consistent with the cDNA sequence and is highlighted by a bold arrow. (B) Detailed schematic representation of the results of the primer extension and S1 protection experiments. Primer extensions were carried out using both total RNA and mRNA preparations from porcine calvarial bone as template. S1 protection was performed using a Sp/d - Srnd fragment of the opn gene as probe.

+

oligonucleotide that was annealed to both poly(A)-rich RNA and total RNA isolated from fetal porcine calvariae. One major and three minor transcription start sites were detected (Fig. 3). The length of the extended nucleotides was determined by comparing the sizes of the extended primer and marker nucleotide sequences (see Materials and Methods). Since the extended nucleotide sequence of the major transcript was consistent with the sequence of Opn cDNA [31], the 5’ end of the major transcript was considered to be the start of exon 1 and is indicated as + 1 in Fig. 4. The minor transcripts were within three nucleotides at positions +2, -2 and -3 (Fig. 3A). The transcriptional start site was also analyzed by S1 nuclease protection mapping (Fig. 3). In addition to the sites obtained by primer extension, several minor fragments were observed at +3, + 4 and + 5 (Fig. 3A). Further, the + 2 start site generated by S1 nuclease mapping was more prominent than the + I site identified by primer extension. The differences evident in the S1 nuclease analysis may be the result of incomplete protection of nucleotides at the extreme 3’ end of the protected fragment as has been reported pre-

viously in mapping the transcriptional start sites of the thrombospondin gene [37]. The DNA sequence of the 4-kb subclone including two exons (exon 1 and 2), an z 1.1-kb intron and 2.6 kb of the 5’ flanking region was determined by sequencing both complementary DNA strands and is shown in Fig. 4. Consensus sequences were searched using DNA Inspector IIe operated on a Macintosh computer. The promoter region contained a TATA-like consensus sequence TTTAAA [37] at positions - 26 to - 31, a CCGCCC (- 63 to - 68) Spl binding site [38, 391, and a TGACACA AP1 (-74 to -80) site [40]. No CCAAT box was present in the immediate promoter region, although a reversed TAACC sequence was identified at positions -138 to -142. Further upstream a GGAAGTG E4TF1[41] binding site (-117 to -123) and a stretch (- 146 to - 173) of the repeating (14x) dinucleotide TG were observed. Other consensus sequences revealed include two putative OTFljOCTl [42] octamer protein binding sequences (-207 to -215, and +486 to +493), several CCAAT boxes [39], and a second Spl site that overlaps a CGCCCCCCGC sequence (-284 to -296) which contains an extra C compared to the recently identified early growth response (Egr-I) element [43], two additional AP1 sites at -2070 and +505, two type I1 collagen gene ‘silencer’ sequences at -682 and - 1715, a CAMP [44] and two glucocorticoid response elements [45] at -332, -658 and +203, respectively, and a potential bcd binding site [46] at + 988. In addition, a calcitriol response element consensus sequence was identified at nucleotides 2245 to 2259. Analysis of the exon/intron boundaries revealed that the 5’ and 3’ ends of the introns contained, respectively, the strictly conserved G T and AG dinucleotides [47]. A more extensive analysis of the extended sequences at the splice boundaries further demonstrated a close similarity to the consensus sequence. The donor splice site of intron 1 AG/GTAAGC matches the consensus nucleotides AG/GT(A/G)AGT at all positions except the last nucleotide, whereas the donor splice site of intron 2 (CA/GTGAGT) is not as well matched. However, the CA dinucleotide preceding the splice point is present in a number of mammalian splice sites such as ovalbumin intron 4 [48]. To investigate whether the 5’ flanking region of the opn gene possesses promoter activity, eight different test sequences from the region were placed immediately upstream of the reporter gene for bacterial chloramphenicol acetyltransferase (CAT) constructed in a promoterless/enhancerless SV40 basic vector (see Materials and Methods). Transient transfection of porcine calvarial cells (Fig. 5 ) with a construct extending upstream from the SmaI site (+51) in exon I , popnCAT + 51/ - 74 which includes the TATA-like sequence, revealed good promoter activity. The strongest promoter activity, however, was obtained with popnCAT+ 51/- 180, which includes an API site and an Spl site. In contrast, no CAT activity was obtained with popnCAT - 180/ - 905, which extends upstream from the fragment in popnCAT 511 - 180 and does not include a TATA-like sequence (Fig. 4). Indeed the popnCAT + 51/ -905 construct, which includes a fragment that spans the fragments in popnCAT+51/- 180 and popnCAT - 180/ -905, had lower promoter activity than popnCAT 511- 180, indicating that there are negative regulators in the - 180 to -905 region. Notably, a ‘silencer’ sequence [49] is evident in position - 682 to - 688 (Fig. 4) and a second ‘silencer’ sequence (-1715 to -1721) is located further upstream in the popnCAT + 51/ -261 5 construct which displays a further reduction in basal promoter activity. -

-

+

+

654 -2615 -2515

-2415

-23 15 -2215 -2115 -20 15

-1915 -1815 -1715 -1615 -1515 -1415 -1315 -1215 -1115 -1015 -915 -8 15

-715 -615 -515 -415 -315 -215 -115 -15 +86

*286

TCAAATATAAAACTTAAARATATGTCCACGGAGTCCTCARAAGAATTATAACTACTTCTTGCCATCAG~TAAAATTCTATGCCTATGTATT~.~

*3S6 A T G T A G C T A C G A T G G T A T C T G C A T C A T T T T A A A T G G A C A G G C T G G T G A T A

I ,486 ~

~

T

T

T

~

T

G

*586 T . ~ C C C A C A T G A C A T T A T G A G A C T T A C R R A T A G G C T T T T A G A A C A C C T T A C T T ~ T T A + 6 S G CTAACTCTAGAGACAGCTTCATTTCACTTAAGTAGCACCTTTT~TAATTT~GCTGAAARATCGCCCTTG~TGCATGCTGGAAAATGGAGACAGCA Xbbn I SPh 1 *786 A G T T T C T T T C T C T C T T T C C T T T T A T T T T C C C T C T T T C T C T T T G T A T T T T T C G T C T T T G ~ T G T G T T C C C C T T C T A G C T T A T T A T T T T ~ T T ~ 0 *886 TTACTGTTGATCTGTTTTTAGGTTTAGACGGCTGGAGATATCGGGTAGTGATGGCATATCTCTGAAACTCTACA~~~GGGACT~TAAGACTTG

+9s6

TPITGTAATCC~CTCTCTCTTGCCTAACAGTAAGAGATGG~TAGAGGTGCCCTARC~TATTAACTC~GGATCAT~TTAAARAG~T

* l o 8 6 TTTTCTCTAAGTAGTAGAGAGTATTTCTATAGG~TATATATATATATTTTTCGTGATTATTTTGTARTGTGGTGGCTTG~GATGTCATTGT *1186 T T T A A C C T A G G A G A A G A T C A A A T A T T T C T T A C A A A G T A T T Avr I 1 M R I A V I A F C L W G F A *I 286 CTCTGCCCTTCCAGTGAGTACAGCTGAATCTTAAACAGAATTC S

A

L

P

Fig. 4. Nucleotide sequence of the 4.0-kb opn gene fragment. The sequences of exons 1 and 2 are presented in bold italics whereas the sequence o f intron 1 and the 5' flanking region are in plain type. The amino acid sequence is printed under the nucleotide sequence using the singlelettcr code. Major restriction sites and the dinucleotide G T repeat arc underlined. The transcription start site is designated as + 1 . Consensus elements arc boxed and the symbols above the boxes represent the following: (0)TATA box; ( 0 )CAAT box; ( x ) GC box; (n)Egr-I site; ( m ) APlsite; ( 0 )OCTI site; (+) glucocorticoid response; ( + ) C A M P response; (*) bcd site; ( 0 )E4TF1 site; ( # ) type I1 collagen silencer sequences; ( A )calcitriol response element.

The additional presence of the 5' half of intron I in pop/7CAT+ 691/ -2615 did not affect promoter activity significantly but when all of intron 1 was included ( p o p C A T + 1191/ - 261 5 ) promoter activity reached maximal levcls. Notably, in a construct comprising intron I alone,

popnCAT+149/+1191, significant promoter activity approaching that of popnCATf 51/-74 was also evident. This activity appeared to be generated from a region immediately 5' to exon 2 that is equivalent to the putative promoter of mouse macrophages [28]. This region also contained a

~

655 A

-2

Smal

B +1191 to-2615 6 9 1 to -2615 +51 to -2615 +51 to -905 +51 to-180 +51 to -14

-180 to -905 +1191 t0+149 0

CAT Constructs

10000

30000

2oooo

Relative CAT Activity

C

3-ACM

1-ACM

EM

Fig.5. Analysis of promoter activity. Various sized gene fragments generated from the 4-kb opn gene subclone (A) were ligated to the promoterless CAT vector (pCATB) as described in Materials and Methods, producing heterologous gene constructs (B, left) for assaying promoter activity. The black box represents the CAT gene (B, left). Each of the eight constructs shown was used in transient transfection assays of porcine calvarial bone cells. (B, right) promoter activities obtained from liquid scintillation counting. (C) A representative autoradiogram showing the thin-layer chromatographic separation of the chloramphenicol substrate from the reaction products in the analyses of the promoter activity in the various constructs. Triplicate analyses are shown of samples normalized with j-galactosidase activity. The radiolabeled 3-acetyl chloramphenicol (3-ACM), 1-acetyl chloramphenicol (1 -ACM) and chloramphenicol (CM) are indicated.

TTTAAA sequence at position + 961 to + 966 (Fig. 4). Since the promoter activity in popnCAT + 1191/ - 261 5 was greater than the combined activities of popnCAT + 51/ 261 5 and popnCAT + 149/ + 1191 it would appear that there are strong positive enhancers in the 3' half of intron 1. To determine whether or not the region immediately 5' to exon 2 might act as an alternative promoter in murine monocytic cells, a 20-residue oligonucleotide (primer 1) complementary to nucleotides 79 to + 60 in exon 1 of Craig et al. [30] and a 20-residue synthetic oligonucleotide (primer 2) coding the antisense sequence of the untranslated region sequence immediately upstream from the coding region ( 166 to +147) of exon 1 of Miyazaki et al. [28], were used for primer extension analysis of RNAs obtained from mouse monocytic cell lines RAW 264 and 5774 as shown in Fig 6. Whereas primer 1 initiated chain elongation to nucleotides + 2 and + 3 of Craig et al. [30] with the RNAs from both cell lines as well as mouse macrophage line IC21 (result not shown), no chain elongation was observed with primer 2 in either cell line indicating the absence of a transcript derived from the second putative promoter site in murine monocytes/ macrophages (Fig. 6A). To confirm these observations in porcine monocytic cells and bone cells, total RNA from lipopolysaccharide-stimulated porcine blood monocytes, as ~

+

+

well as RNA from fetal porcine calvariae and calvarial bone cells, were probed with a complementary 500-bp fragment extending immediately upstream from exon 2 and corresponding to the untranslated region of exon 1 of mouse macrophages [28]. Whereas Opn mRNA could be readily detected in all three samples with a full-length cDNA probe, the 500-bp fragment only hybridized to a band identified as Opn heterogeneous nuclear RNA (Fig. 6B), indicating that the second promoter site is not normally utilized in porcine monocytic cells and bone cells. To determine whether the chimeric constructs of the opn promoter would respond to calcitriol, ROS 17/23 cells were transfected with various constructs and stimuhted for 48 h with 0.1 pM calcitriol. Studies were carried out in ROS 17/23 cells since these cells respond consistently to calcitriol [13, 141 whereas the porcine calvarial cells showed a variable response to the hormone. Basal promoter activity in the ROS 17/23 cells was similar to the results obtained from the porcine calvarial cells except that transcription was greater in popnCAT + 511- 905 indicating that the silencer sequence may not be operative in the rat cells. There was no significant change in transcription rates with calcitriol stimulation in the smallest construct (popnCAT + 51/ - 74) but the transcription rate was increased I .4-fold with popnCAT 51/ - 180 (Fig. 7).

+

656 B

A

*

+ +

m

+,Exonl o*

Primer -1 Extendedproducu-b

Exon 2

n* f Primer2

+3* O* +2

-*

f No extension

1

2

3

4

5

6

Fig. 6. Analysis of alternative promoter activity. The existence of a second promoter site utilized by monocytic cells was investigated using primcr extension analysis of murine mRNA (A) and Northern hybridization of porcine mRNA (B). A schematic representation of the mouse o p i gene 5' portion which contained promotcr (PI), exon 1 according to Craig et al. [30] and the potential promoter (P2), and exon 1 (including Craig's exon 2 and part of intron 1) suggested by Miyazaki et al. [28], is given (A, lower). Primer extension was performed using RNA prepared from the mouse monocytic cell lines (A, upper) 5774 (lanes 1 and 2) and RAW 264 (lanes 3 and 4). For Northern hybridizations (B), total RNA (20 pg) prepared from porcine calvarial bone (lanes 1 and 4), porcine calvarial cells (lanes 2 and S ) , and porcine blood monocytes stimulated with lipopolysaccharidc (lanes 3 and 6) was probed with a full-length porcine cDNA probe (lanes 1-3) and an Avrll-XbaI fragment from the first intron region of the porcine opn gene (lanes 4-6). The latter probe hybridized only to a band (arrow) identified as heterogeneous nuclear RNA in porcine calvarial bone RNA preparation.

25000,

for a calcitriol response element. However, calcitriol also reproducibly stimulated a twofold transcription by the construct popnCAT+51/-905 which has no apparent calcitriol response element (Fig. 7). The nucleotide sequences of porcine and murine opn [30] extending upstream from the splice site of exon 1 to 900 of the 5' flanking region were compared by dot matrix analysis using D N A Inspector IIe run on a Macintosh computer. A small segment of the 5' flanking region from the start of transcription to -220 which encompasses the promoter of porcine opn and the promoter of the mouse gene was found to be highly conserved with approximately XO% identity between the two species (Fig. 8). Of note, the TATA element, TTTAAA, at -26 and the AP-1 site at -74 are conserved between species as is the exon l/intron 1 boundary. However, there are several important differences between the two species. In particular, the inverse CCAAT box present in mouse is absent in porcine opn while the Spl site present in the porcine gene is absent in mouse gene. Further, the T G dinucleotide repeat in porcine opn is not present in murine opn which instead contains a polypyrimidine stretch that could function as an inverse GAGA box. Of interest, in another bone protein gene, mouse spare (secreted protein rich in cysteine), GAGA boxes have been implicated as transcriptional regulatory elements [SO]. Together, these data indicate that while pig and mouse rnay share some regulatory mechanisms that control constitutive, basal opn transcription, there may also be important differences in regulation of Opn expression between these species. We have also identified three short stretches of sequence further upstream that are highly conserved and it is possible that transcription factors mediating tissue-specific expression of Opn rnay interact with these conserved sequences.

-

P

CAT Constructs

Fig. 7. Regulation of promoter activity by calcitriol. Five heterologous constructs that included the promoter were used to determine the rcgion ofcalcitriol rcgulation of transcription in the opn gene. Individual dishes of ROS 17/2.X cells were transfected with one of the constructs and incubatcd in the presence of 0.1 pM calcitriol for 48 h as dcscrtbed i n Materials and Methods. Control cells (not stimulated with calcitriol) were maintained for the same time in culture medium. Lcvel of CAT activity was measured by scintillation counting and normalized by /I-galactosidase activity. Means of the triplicates for individual constructs are presented in bars (white, control; black, calcitriol stimulation) with i SEM. Different scales on the right and left vertical axes have been used to present the results of two separate expcrimcnts. the results of which werc reproducible in three replicates.

However. the greatest stimulation ( z 2 -2.5-fold) was observed with the constructs popnCAT+ 1191/-2615, and popfrCAT+ 511- 261 5, which include the consensus sequence

~

657 -761

/Do"Se

*.

x L . ~ C

-i_se

,,***********.,,

~

-511

-861

. ........ .

. * . A~ . *. , .>. ~ .I,

F

..ACARAACAGAGC..

..***C*"****'**-***., -871

*

r.,

~

*-. T

-535

-546

-741

..TCAGAAACTGCTTGACTT..

porcine

-131

-349

..TTTTTCTTCTAATTCAG.. .....................

-500

-321

-241

..C .,T -223

-305

TGA. 'CAC' * * ' * * ' T C A W L C T ~ T C T ~ R X % T T

l

4" L

T

PnmCWLGWLIIlIC...-.--

I

C

.I