Sp1 Family of Transcription Factors Regulates the Human α2 (XI) Collagen Gene (COL11A2) in Saos-2 Osteoblastic Cells

Authors


  • The authors state that they have no conflicts of interest.

Abstract

Genes encoding type XI collagen, normally associated with chondrogenesis, are also expressed by osteoblasts. By studying Saos-2 cells, we showed that the transcription factors, Sp1, Sp3, and Sp7 (Osterix), regulate COL11A2 expression through its proximal promoter. The findings indicate both ubiquitous and osteoblast-specific mechanisms of collagen gene regulation.

Introduction: Type XI collagen is essential for skeletal morphogenesis. Collagen XI gene regulation has been studied in chondrocytes but not in osteoblasts.

Materials and Methods: We cultured Saos-2 cells, a human osteosarcoma-derived line of osteoblasts, and analyzed them for α2(XI) protein and COL11A2 regulatory mechanisms.

Results and Conclusions: Although types I and V were the dominant collagens deposited by Saos-2 cells, they expressed COL11A2 mRNA, and α2(XI) chains were present in the extracellular matrix. The COL11A2 promoter region (from −149 to −40) containing three Sp1 binding sites was required for promoter activity in transient transfection assays. All three Sp1 sites were critical for binding by nuclear proteins in electrophoretic mobility shift assays. Further analysis using consensus oligonucleotides and specific antibodies as well as chromatin immunoprecipitation assay implicated Sp1 and Sp3 in binding to this promoter region. Overexpressing Sp1 or Sp3 significantly increased COL11A2 promoter activity and endogenous COL11A2 gene expression, an effect that was suppressed by the Sp1-binding inhibitor mithramycin A. Further experiments showed that Sp1, Sp3, CREB-binding protein (CBP), p300, and histone deacetylase (HDAC) were physically associated and HDAC inhibitors (trichostatin A or NaB) upregulated COL11A2 promoter activity and endogenous gene expression. Another Sp1 family member, Sp7 (Osterix), was expressed in Saos-2 cells, but not in chondrocytes, and was shown by chromatin immunoprecipitation to occupy the COL11A2 promoter. Overexpressing Sp7 increased COL11A2 promoter activity and endogenous gene expression, an effect also blocked by mithramycin A. Using siRNA to knockdown Sp1, Sp3, or Sp7, it was shown that depression of any of them decreased COL11A2 promoter activity and endogenous gene expression. Finally, primary cultures of osteoblasts expressed COL11A2 and Sp7, upregulated COL11A2 promoter activity and endogenous gene expression when Sp1, Sp3, or Sp7 were overexpressed, and downregulated them when Sp1, Sp3, or Sp7 were selectively depressed. The results establish that Sp1 proteins regulate COL11A2 transcription by binding to its proximal promoter and directly interacting with CBP, p300, and HDAC.

INTRODUCTION

Members of the collagen family of extracellular proteins can function not only as structural components but also regulators of development and morphogenesis.(1,2) They include the fibril-forming collagens, types I, II, III, V, and XI, which can be divided into major (I–III) and minor (V and IX) subgroups, on the basis of their abundance in tissues(3) and gene characteristics.(4) Type XI collagen is a heterotrimer of three genetically distinct α chains, α1 (XI), α2 (XI), and α3 (XI), the latter being a post-translational variant from the α1 type II collagen gene.(5–7) Although quantitatively minor, type XI collagen is crucial for collagen fibril formation and skeletal morphogenesis.(8)

Type XI collagen is mostly found in cartilages.(9) We reported that expression of the mouse α2 (XI) gene (Col11a2) is regulated by cis-elements in the promoter region.(10) We also found that the human gene (COL11A2) can be regulated by the Ewing's sarcoma gene (EWS)/ETS-related gene (ERG) fusion protein (produced by Ewing's sarcoma), through a tandem repeat of tcc trinucleotides (TRT) in the proximal promoter.(11) In addition to prominent expression in cartilages, it is known that the α2 (XI) collagen gene is expressed by osteoblasts, for example in calvarium and periosteum in mice,(12) osteosarcomas in humans,(13) fracture callus in rats,(14) and splice variants have been linked genetically to the pathological ossification of human posterior longitudinal ligaments (OPLL) that can occur particularly in Japan.(15)

It has been reported that several Sp1 family proteins, including Sp1, Sp3, and Sp7/Osterix, are expressed in bone.(16–18) Sp1 and Sp3 show ubiquitous expression patterns and regulate a wide range of cellular functions including cell growth, differentiation, apoptosis, and oncogenesis.(16,18,19) Each factor can act as either an activator or repressor, or both, depending on the promoters and co-regulators with which it interacts. However, the significance of Sp1 and Sp3 in bone physiology and pathology is not clear. In contrast, Sp7/Osterix is expressed in all developing bone and is essential for osteoblast differentiation and bone formation.(17,20) It binds to G/C-rich DNA sequences with characteristics in common with other Sp1 family proteins (Sp1 binding sites). However, the amino acid sequence of Sp7 other than its DNA-binding domain, which includes a proline- and serine-rich transcription activation domain, shares no significant similarity with any other Sp1 family member. The bone-specific expression pattern implies that the unique N-terminal structure may, at least in part, be responsible for the distinctive profile of matrix proteins expressed by osteoblasts in vivo.

To understand how transcription of COL11A2 is controlled in osteoblasts, we analyzed Saos-2 cells, a human osteosarcoma-derived osteoblastic cell line, for COL11A2 expression. The results showed that, unlike chondrocytes and Ewing's sarcoma cells, Saos-2 cells expressed COL11A2 mRNA under the control of three functional Sp1 binding sites in the proximal promoter.

MATERIALS AND METHODS

Cell culture and RT-PCR

Saos-2 cells (American Type Culture Collection) and primary cultures of osteoblasts (normal human osteoblasts [NHOst's]); Cambrex, East Rutherford, NJ, USA) were cultured in MEM (Eagle) with 10% FBS, under 5% CO2 at 37°C. RT-PCR of the alternatively spliced N-terminal portion of COL11A2 mRNA was performed as described,(13) and the correct products were confirmed by sequencing. A 415-bp fragment of the cDNA encoding the C terminus of COL11A2 was amplified using the forward primer (exon 63–64) 5′-agagcttcccgatggagagta-3′ and the reverse primer (exon 66) 5′-gtctgagaaggaggcatccag-3′.(11) In addition, a 1296-bp fragment of SP7/Osterix was amplified using the forward primer 5′-atggcgtcctccctgcttgaggag-3′ (exon 1) and the reverse primer 5′-tcagatctccagcaagttgctctg-3′ (exon 2).

Protein analysis

Saos-2 cells were cultured as monolayers in McCoy's medium containing 10% FBS and 50 μg/ml ascorbate. Medium was changed three times a week, and the cultures were maintained for a month. Spent culture media were collected and frozen at −20°C. The cell layer was extracted in 1 M NaCl, 0.05 M Tris-HCl, pH 7.5, containing proteinase inhibitors (0.01 M EDTA, 0.001 M phenylmethyl sulfonyl fluoride [PMSF]) for 24 h at 4°C.(21) Collagen precipitated at 4.5 M NaCl was digested with 100 μg/ml pepsin in 0.5 M acetic acid(22) and further fractionated by precipitation at 0.8 M NaCl and 2.0 M NaCl in 0.5 M acetic acid. Collagen chains were resolved by SDS-PAGE and stained with Coomassie Brilliant Blue.(23,24) Collagen in the medium was precipitated at 30% ethanol saturation for 1 h at 4°C.(25) Conditioned media from a human chondrosarcoma cell line, CH1,(26) and human fetal chondrocytes were used as references.

For Western blot detection of the α2(XI) chain, an equivalent gel was transblotted to polyvinilidene difluoride (PVDF) membrane, blocked with 5% milk powder, and developed using chemiluminescence (Amersham Pharmacia Biotech, Piscataway, NJ, USA).(23) The primary antiserum was raised against the bacterially expressed thrombospondin-N domain (TSPN) of human α2(XI).

Plasmid constructs

Various lengths of the DNA sequences for mouse Col11a2 promoter regions(10) were cloned into MluI/XhoI sites of pGL3 basic luciferase reporter vector (Promega, Madison, WI, USA).(11) To characterize the human COL11A2 promoter sequence between −149 and +27 bp (H149 construct), a deletion construct was created as H40 (–40 to +27 bp). In addition, mutations were introduced into the H149 human COL11A2 promoter construct as follows. mSp1–1 contains a 5-base mutation from gggtgg to TTTtTT located between −108 and −103 bp of the H149 construct; mSp1–2 contains an 8-base mutation from gggcgggcgg to TTTcTTTcTT located between −82 and −73 bp of the H149 construct; mSp1–3 contains a 5-base mutation from gggcgg to TTTcTT located between −50 and −45 bp of the H149 construct; mSp1–1/2 contains two mutations described in the mSp1–1 and the mSp1–2; mSp1–1/3 contains two mutations described in the mSp1–1 and the mSp1–3; mSp1–2/3 contains two mutations described in the mSp1–2 and the mSp1–3; mSp1–1/2/3 contains three mutations described in the mSp1–1, the mSp1–2, and the mSp1–3.

Transfection and luciferase assay

For promoter analysis, two duplicate wells of 70% confluent Saos-2 cells or primary cultures of osteoblasts in 6-well plates were transfected with 2 μg of pGL3 reporter plasmid and 0.2 μg of pRL-SV40 Renilla luciferase control using 4 μl of FuGENE 6 transfection reagent (Roche Molecular Biochemicals, Indianapolis, IN, USA). Forty-eight hours after transfection, cells were assayed for luciferase activity using Dual Luciferase Reporter Assay System (Promega) on a luminometer (Lumat LB9507; Berthold, Tokyo, Japan). Transfection in duplicate was repeated at least three times, and the luciferase activity was normalized to internal control. The results are shown as average ± SE.

In a subset of experiments, Saos-2 cells or primary cultures of osteoblasts were transfected with 0.25 μg of pGL3 reporter plasmid, 1.75 μg of expression plasmid for Sp1 (pN3 Sp1), Sp3 (pN3 Sp3), or Sp7 (pFLAGCMV2 Sp7), or control vector (pN3 empty or pFLAGCMV2 empty). Transfection in duplicate was repeated at least three times, and the luciferase activity was normalized to the amount of protein in the cell layer. The results are shown as average ± SE.

pN3 Sp1, pN3 Sp3, and pN3 empty were generous gifts from Guntram Suske (Institut fuer Molekularbiologie und Tumorforschung, Philipps-Universität, Marburg, Germany).(27) pFLAGCMV2 Sp7, containing a full-length cDNA for the mouse Sp7/Osterix and N-terminal FLAG epitope tag, was from Toshihisa Komori (Division of Oral Cytology and Cell Biology, Department of Developmental and Reconstructive Medicine, Nagasaki University Graduate School of Biomedical Sciences, Nagasaki, Japan). Correct overexpression of Sp1, Sp3, and Sp7 in Saos-2 cells was confirmed by Western blotting using PEP2 anti-Sp1 antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA), D-20 anti-Sp3 antibody (Santa Cruz Biotechnology), and anti-Sp7/Osterix antibody (raised against the BSA-conjugated synthetic polypeptides containing a C-terminal 12 amino acid sequence of Sp7/Osterix [GGSPEQSNLLEI]; Toshihisa Komori and Takashi Fujita, Nagasaki, Japan), respectively.

For inhibition of histone deacetylase, transfected cells were treated with 100 ng/ml of trichostatin A (TSA) or 5 mM sodium butyrate (NaB; Sigma, St Louis, MO, USA) 24 h before luciferase assay and compared with transfected cells without TSA or NaB treatment. Transfection in duplicate was repeated at least three times, and the luciferase activity was normalized to the amount of protein in the cell layer. The results are shown as average ± SE.

In some experiments, Saos-2 cells were also treated with 100 nM of a GC-box binding inhibitor mithramycin A at the time of transfection or addition of histone deacetylase (HDAC) inhibitor (TSA or NaB).

EMSA

Saos-2 nuclear extract was prepared as described previously.(28) Double-stranded oligonucleotides containing wildtype (WT) human COL11A2 promoter sequence between −117 and −92 bp (O1WT), between −94 and −63 bp (O2WT), and between −59 and −36 bp (O3WT) were used in the EMSA. Otherwise identical probes mutating gggtgg to TTTtTT (O1Mut), gggcgggcg to TTTcTTTcT (O2Mut), and gggcgg to TTTcTT (O3Mut) were also used in the assay. [32P]labeled DNA probe was incubated with Saos-2 nuclear extract (1 μg of protein) for 20 minutes at room temperature in the presence of poly(dI-dC) (50 μg/ml). The DNA-protein complexes were analyzed on a 6% nondenaturing polyacrylamide gel. For competition assay, unlabeled WT, mutant (Mut), Sp1 consensus, or Ap1 consensus oligonucleotide was added to the reaction 10 minutes before the addition of the [32P]labeled WT. Supershift experiments were done by the addition of PEP2 rabbit polyclonal anti-Sp1 antibody (Santa Cruz Biotechnology) or D-20 rabbit polyclonal anti-Sp3 antibody (Santa Cruz Biotechnology) to the reaction mixture 1 h after the formation of the DNA–protein complexes. The reaction was continued for 1 h at 4°C. Equal amount of a normal rabbit IgG (Santa Cruz Biotechnology) was added to a control sample to exclude nonspecific interactions.

Chromatin immunoprecipitation assay

To cross-link DNA and protein, 1 × 108 of Saos-2 cells were treated with 1% formaldehyde for overnight at 4°C. Chromatin solution was prepared as described previously.(29) For immunoprecipitation, 3 μg of antibody against acetylated histone H3 (anti-Ac-H3; Upstate Biotechnology, Lake Placid, NY, USA), antibody against acetylated histone H4 (anti-Ac-H4; Upstate Biotechnology), antibody against Sp1 (PEP2), antibody against Sp3 (D-20), antibody against Sp7/Osterix (from Toshihisa Komori and Takashi Fujita), or a normal rabbit IgG (Santa Cruz Biotechnology) was incubated with chromatin solutions overnight at 4°C on a rotating wheel. The immunocomplexes were collected with salmon sperm DNA/Protein A-Sepharose beads and washed sequentially with 1 ml of each of the following buffers containing protease inhibitor mixture (Sigma): once with low salt immune complex wash buffer (20 mM Tris, pH 8.1, 2 mM EDTA, 0.1% SDS, 1% Triton X-100, 150 mM NaCl), once with high salt immune complex wash buffer (20 mM Tris, pH 8.1, 2 mM EDTA, 0.1% SDS, 1% Triton X-100, 500 mM NaCl), once with LiCl immune complex wash buffer (10 mM Tris, pH 8.1, 1 mM EDTA, 1% NP40, 0.25 M LiCl, 1% sodium deoxycholate), and twice with PBS. Formaldehyde cross-linking was reversed by overnight incubation at 65°C in 0.2 M NaCl plus 200 μg/ml of proteinase K (Sigma). The mixture was treated with phenol/chloroform, and the DNA was precipitated with ethanol and resuspended in 20 μl of H2O. One microliter of the DNA template was used for PCR amplification of COL11A2 promoter regions. Primers for COL11A2 were 5′-caggagagagcgagcgatag-3′ (–167 to −147 bp from the transcriptional start site) and 5′-cccggcccggcccccgcctccagccgcccgcccacagcca-3′ (–89 to −50 bp from the transcriptional start site). The resulting 118-bp products for COL11A2 were separated by agarose gel electrophoresis.

Immunoprecipitation and Western blotting

Saos-2 cells in a 10-cm dish were lysed with 1.2 ml of buffer A (10 mM Tris, pH 7.4, 100 mM NaCl, 2.5 mM MgCl2, 0.5% Triton X-100, 10 mM DTT) supplemented with protease inhibitor cocktail and phosphatase inhibitor cocktail (Sigma). The PEP2 rabbit polyclonal anti-Sp1 antibody, the D-20 rabbit polyclonal anti-Sp3 antibody, the C-19 goat polyclonal anti-HDAC1 antibody (Santa Cruz Biotechnology), the A22 rabbit polyclonal anti-cAMP-responsive element-binding protein (CREB)-binding protein (CBP) antibody (Santa Cruz Biotechnology), the N-15 rabbit polyclonal anti-p300 antibody (Santa Cruz Biotechnology), or a normal rabbit IgG (Santa Cruz Biotechnology) was incubated with 40 μl of Protein A/G PLUS-Agarose (Santa Cruz Biotechnology) and 0.2 ml of buffer A for 50 minutes at 4°C. The complex was incubated with 0.2 ml of fresh cell lysate for 60 minutes at 4°C on a rotating wheel. After three washes with RIPA buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS) at 4°C, 50 μl of Laemmli sample buffer (Bio-Rad, Hercules, CA, USA) with 0.1 M dithiothreitol (DTT) was added to the agarose beads. The protein samples were denatured for 5 minutes at 95°C, separated by 8% SDS-PAGE, transferred onto a PVDF membrane, and subjected to Western blotting with the PEP2 rabbit polyclonal anti-Sp1 antibody or the D-20 rabbit polyclonal anti-Sp3 antibody. Protein bands were visualized using the ECL Western Blotting Analysis System (Amersham Pharmacia Biotech).

Real-time PCR

In a subset of experiments, after transfection of expression constructs, treatment with TSA, NaB, and/or mithramycin A, or small interfering RNA (siRNA) treatment, total RNA was isolated using Isogen (Nippon Gene, Tokyo, Japan) and treated with DNase I (TaKaRa, Shiga, Japan) according to the manufacturer's protocol. cDNA was made from 1.5 μg of total RNA using SuperScript II reverse transcriptase (Invitrogen) with random hexamers as a primer. For quantitative analysis of the expression levels of COL11A2 mRNAs, Sp1 mRNA, Sp3 mRNA, or Sp7/Osterix mRNA, real-time PCR was performed using TaqMan gene expression assays (Applied Biosystems) and a thermal cycler (ABI 7500 Real Time PCR System). Experiment in duplicate was repeated at least three times, and levels of β2-microglobulin transcripts were used to normalize COL11A2 expression levels. Primer sequences for β2-microglobulin were 5′-GCCGTGTGAACCATGTGACTT-3′ and 5′-CAAACCTCCATGATGCTGCTT-3′. The results are shown as average ± SE.

siRNA

siRNA cocktails targeting human Sp1, human Sp3, or human Sp7 were purchased from B-Bridge International (Sunnyvale, CA, USA). The sequences for these siRNA duplexes were as follows: gcaaaaagaaggagagcaaTT, cggaugagcuacagaggcaTT, and ccagaagaguggugagauaTT for Sp1; gauagauaguacagguauaTT, ggacagguaacuuggcaaaTT, and acauggaacucuagggauaTT for Sp3; ccuacuccauggugggauaTT, agggagugguggagccaaaTT, and cggaugagcuggagcgucaTT for Sp7. An siRNA cocktail for negative control was also purchased from B-Bridge International. Saos-2 cells or primary cultures of osteoblasts were transfected using Lipofectamine 2000 (Invitrogen) to achieve a final concentration of siRNA of 50 nM. Total RNA was isolated using Isogen (Nippon Gene) and treated with DNase I (TaKaRa) according to the manufacture's protocol. cDNA was made from 200 ng of total RNA using Quantitect Reverse Transcription (QIAGEN). The specificity of Sp1, Sp3, or Sp7 knockdown was confirmed by PCR and real-time PCR. For promoter analysis, transfection including luciferase reporter construct was allowed to proceed for 48 h before analysis. Experiment in duplicate was repeated at least three times, and the luciferase activity was normalized to the amount of protein in the cell layer. The results are shown as average ± SE. For quantitative analysis of the expression levels of COL11A2 mRNA, real-time PCR was performed. Experiment in duplicate was repeated at least three times, and levels of β2-microglobulin transcripts were used to normalize COL11A2 expression levels. The results are shown as average ± SE.

Statistical analysis

An ANOVA followed by Dunnett's posthoc test was used to determine statistical significance. A value of p < 0.05 was considered significant.

RESULTS

Saos-2 cells express COL11A2 mRNA and α2(XI) protein chains

Previous RT-PCR analysis showed that the expression and splicing patterns of COL11A2 are linked to a chondrocytic phenotype in osteochondrogenic tumors.(13) The human osteosarcoma-derived osteoblastic cell line Saos-2 also expressed COL11A2 transcripts, but the splicing pattern of COL11A2 in Saos-2 cells differed from that in chondrocytes (Fig. 1A, lanes 1 and 2). Several prominent COL11A2 transcripts retaining combinations of the alternative exons 6–8 were evident in Saos-2 cells but not in chondrocytes.

Figure Figure 1.

Expression of COL11A2 transcripts and α2(XI) chains by Saos-2 osteoblastic cells. (A) RT-PCR analysis of COL11A2 transcripts from Saos-2 cells (lane 1), human chondrocytes (lane 2), and HeLa cells (lane 3). Splice variants containing different combinations of exons 6–8 covering the N-terminal COL11A2-coding region are identified. Exon 0 lacks all three exons. (B) SDS-PAGE analysis of collagens in the matrix of the Saos-2 cell layer. Types I and V molecules are the main collagens deposited in the cell layer. Lane 1, 1 M NaCl extract of the Saos-2 cell layer; lane 2, type I collagen in the 0.8 M NaCl precipitate; lane 3, type V collagen in the 2.0 M NaCl precipitate; lane 4, standard type II and type XI collagen from cartilage. The gel was stained with Coomassie Brilliant Blue. All lanes were electrophoresed under reducing conditions (+DTT). (C) Detection of pNα2(XI) chains in the cell layer extract by Western blot analysis. Lane 1, conditioned medium from a human chondrosarcoma cell line, CH1; lane 2, Saos-2 conditioned medium; lane 3, 1 M NaCl extract of Saos-2 cell layer; lane 4, Saos-2 cell lysate; lane 5, conditioned medium from human fetal chondrocytes. All lanes were electrophoresed under reducing conditions. The migration positions of α1(II) and its dimer β1(II) are shown.

Types I and V collagens were the major collagen types expressed by Saos-2 cells (Fig. 1B). A 1 M NaCl extract of the cell layer showed fully processed α1(I) and α2(I) chains of type I collagen and fainter bands in the position of type V collagen chains (lane 1). Digestion with pepsin and salt precipitation separated type I collagen in the 0.8 M precipitate (lane 2) from type V collagen [α1(V) and α2(V) chains] in the 2.0 M precipitate (lane 3). No α2(XI) chains were visible by Coomassie staining. The four chain identities [α1(I), α2(I), α1(V), and α2(V)] were confirmed by mass spectrometry after in-gel digestion with trypsin (data not shown).

By Western blot analysis, type XI N-procollagen α 2 [pNα2(XI)] chains were detected as distinct bands (Fig. 1C, lane 3) that matched the predicted size of the different splice forms detected by RT-PCR (Fig. 1A). The polyclonal anti-serum used for this Western blot is specific to the TSPN domain located at the amino-terminus of the variably spliced exonic sequences in pNα2(XI) variants. Therefore, all splice variants will be detected. No bands were detected by Western analysis of Saos-2 medium (lane 2) or Saos-2 cell lysate (lane 4), indicating the incorporation of expressed α2(XI) into the extracellular matrix.

It has been shown that the mouse Col11a2 promoter sequence, −742 to +380 bp from the transcriptional start site, is sufficient to drive reporter gene expression,(10) and multiple subregions between −742 and −453 bp are known to be involved in cartilage- and neural tissue–specific promoter activities.(30) The mouse Col11a2 gene is located next to the retinoic acid receptor β (Rxrb) gene in a head to tail alignment (Fig. 2A, top line).(10) Human COL11A2 and RXRB are arranged similarly and the intergenic DNA sequences of mice and humans are highly homologous (Fig. 2A, bottom line).(31) The transcriptional start site in mouse corresponds to −327 bp from the human transcriptional start site (Fig. 2A).(10,32)

Figure Figure 2.

Identification of three Sp1 sites in the proximal promoter of COL11A2 gene functional in Saos-2 cells. (A) The head-to-tail arrangement of the mouse Col11a2 and Rxrb genes (top) and the human COL11A2 and RXRB genes (bottom). The transcription start site of mouse Col11a2 corresponds to −327 bp from the human COL11A2 transcription start site. Cartilage-specific (•) and neural tissue-specific (•) cis-elements are shown in the mouse Col11a2 promoter sequence. Horizontal lines mark the mouse Col11a2 and human COL11A2 promoter constructs to scale. (B) Firefly luciferase reporter constructs were transfected into Saos-2 cells, and the promoter activity was normalized to the Renilla luciferase control. Transfection in duplicate was repeated at least three times, and the results are shown as average ± SE. *p < 0.05 compared with empty vector; **p < 0.05 compared with −742, −530, −500, −453, or +133. (C) The human COL11A2 promoter construct H149 is shown with potential Sp1-binding sites underlined. The location of the tandem repeat of tcc trinucleotides (TRT), which was dispensable for the promoter activity in Saos-2 cells, is also designated. Mutations of three Sp1 sites in a various combination are shown. The locations of the nucleotide sequences used for electrophoretic mobility shift assay (EMSA) are shown (top). (D) Firefly luciferase reporter constructs were transfected into Saos-2 cells, and the promoter activity was normalized to the Renilla luciferase control. Transfection in duplicate was repeated at least three times, and the results are shown as average ± SE. *p < 0.05 compared with empty vector; **p < 0.05 compared with H149; ***p < 0.05 compared with mSp1–1, mSp1–2, or mSp1–3.

To identify the promoter sequence responsible for COL11A2 expression in Saos-2 cells, a range of pGL3 luciferase reporter plasmids containing various lengths of mouse Col11a2 promoter sequence were generated (Fig. 2A). The mouse Col11a2 promoter sequence from −742 to +380 bp was activated when transfected into Saos-2 (Figs. 2A and 2B, −742). Interestingly, in Saos-2 cells, the mouse promoter constructs −530 and −500 containing either the cartilage- or the neural tissue–specific cis-elements (Fig. 2A, top line) showed similar activity to the −453 construct containing only the constitutive promoter sequence between −453 and +380 bp (Figs. 2A and 2B).(30) Further deletion analysis identified a necessary minimum sequence between +133 and +380 bp of the mouse promoter, which corresponds to −178 to +37 bp of the human COL11A2 promoter (Figs. 2A and 2B, compare construct +133 and construct +264). The evolutionally conserved human COL11A2 promoter sequence between −149 and +27 bp was therefore cloned into the luciferase reporter vector and was equivalent in activity to the intact mouse Col11a2 promoter (–742 to +380 bp) when transfected into Saos-2 cells (Fig. 2A, construct H149).

Human COL11A2 promoter contains three functional Sp1 binding sites

Inspection of the DNA sequence between −149 and +27 bp of the human COL11A2 promoter revealed several candidate Sp1 binding sites (Fig. 2C, construct H149). To test them for functional activity, a series of mutation and deletion constructs were generated for transfection (Fig. 2C). Involvement of the three Sp1 sites was shown by mutations from gggtgg to TTTtTT located between −108 and 103 bp (Fig. 2C, Sp1–1), from gggcgggcgg to TTTcTTTcTT located between −82 and −73 bp (Fig. 2C, Sp1–2), and from gggcgg to TTTcTT located between −50 and −45 bp (Fig. 2C, Sp1–3). Mutating just one of the three Sp1 sites resulted in a 50% decrease (Figs. 3A and 3B, constructs mSp1–1, mSp1–2, and mSp1–3), mutating two of the three Sp1 sites resulted in a 70% decrease (Figs. 2C and 2D, constructs mSp1–1/2, mSp1–1/3, and mSp1–2/3), and mutations at all sp1 sites led to a 85% decrease (Figs. 2C and 2D, construct mSp1–1/2/3). Deleting the DNA sequence between −149 and −41 bp nearly abolished the promoter activity (Figs. 2C and 2D, construct H40). Mutation or deletion of a tandem repeat of the tcc trinucleotides (Fig. 2C, TRT), tccctcc located between −120 and −114 bp, which was involved in the COL11A2 regulation by EWS/ERG fusion protein,(11) did not affect promoter activity in Saos-2 cells (data not shown). Together, these results indicate that in Saos-2 cells, three Sp1 sites coordinately regulate COL11A2 promoter activity.

Figure Figure 3.

Interaction of COL11A2 promoter sequences with Saos-2 nuclear proteins. Nuclear extract from Saos-2 cells was subjected to an electrophoretic mobility shift assay using a radiolabeled oligo probe (O1WT in A, O2WT in B, O3WT in C; for locations, see Fig. 2C, top). The DNA-protein complexes (lane 2) were eliminated by the addition of an excess amount of Sp1 sequence (lanes 3 and 5) but not by mutant Sp1 sequence (Mut) or unrelated sequence (Ap1; lanes 4 and 6). Addition of an anti-Sp1 antibody (lane 7) or an anti-Sp3 antibody (lane 8) but not a normal IgG resulted in specific elimination of the DNA–protein complexes and supershifts indicated by asterisks. (D) Chromatin immunoprecipitation (ChIP) assay. ChIP assay of a 118-bp COL11A2 genomic fragment (–167 to −50 bp to the transcriptional start site) was performed with Saos-2 cells. Antibodies used in the assay are indicated. Input DNA represents a portion of the sonicated chromatin before immunoprecipitation, used as the positive control.

To detect potential interactions between Saos-2 nuclear proteins and the Sp1 sites within the human COL11A2 promoter, an EMSA was performed using oligonucleotides matching −117 and −92 bp (O1WT), −94 and −63 bp (O2WT), and −59 and −36 bp (O3WT) from the COL11A2 transcriptional start site (Fig. 2C). Saos-2 nuclear extract and the Sp1 sequences formed DNA–protein complexes indicated by arrows (Figs. 3A–3C, compare lanes 1 and 2). The addition of wildtype Sp1 competitor or Sp1 consensus oligonucleotide prevented formation of the complexes (Figs. 3A–3C, lanes 3 and 5), but an excess of the mutated Sp1 DNA or Ap1 consensus oligonucleotide had little effect on the complex formation (Figs. 3A–3C, lanes 4 and 6). Furthermore, addition of anti-Sp1 antibody or anti-Sp3 antibody resulted in specific elimination of the complex formation and supershifts, which were not observed by addition of normal IgG (Figs. 3A–3C, lanes 7–9). Thus, Sp1 and Sp3 in Saos-2 nuclear extract appear to interact specifically with the Sp1 sequences within the human COL11A2 promoter.

We next performed a chromatin immunoprecipitation assay to determine whether Sp1 and/or Sp3 selectively bind to the COL11A2 promoter in Saos-2 cells. Acetylated histone molecules are known to be enriched in areas near actively transcribed genes such as GAPDH.(29) In Saos-2 cells, the COL11A2 gene is active, and its promoter sequence becomes cross-linked to acetylated histone H3 and acetylated histone H4 after treatment with formaldehyde (Fig. 3D, lanes 2 and 3). The COL11A2 promoter sequence was enriched in immunocomplexes using antibody against Sp1 and Sp3, but not in those using normal IgG (Fig. 3D, lanes 4–6). These results suggest that Sp1 and Sp3 directly interact with COL11A2 promoter in vivo.

Sp1 and Sp3 activate the COL11A2 promoter and induce endogenous COL11A2 expression in Saos-2 cells

To investigate whether Sp1 and Sp3 can transactivate the COL11A2 promoter, Sp1 and Sp3 expression constructs were co-transfected with the COL11A2 promoter reporter construct into Saos-2 cells. Expression of Sp1 or Sp3 potently transactivated human COL11A2 promoter construct H149 (Fig. 4A). To investigate whether the transactivation was dependent on the ability of Sp1 or Sp3 to bind DNA, mithramycin A, which interferes with the binding of the Sp family of transcription factors to GC-rich promoter regions,(33) was used in the promoter assay. Mithramycin A dramatically decreased promoter activity in Saos-2 cells, indicating that action of Sp1 or Sp3 depended on DNA binding. Furthermore, we found that Sp1 or Sp3 significantly increased the endogenous mRNA levels of COL11A2 by 43.4% and 42.9%, respectively, in Saos-2 cells (p < 0.05; Fig. 4B). These increases were cancelled by the addition of mithramycin A (p < 0.05). These results indicate that Sp1 and Sp3 transcription factors upregulate the human COL11A2 gene through Sp1 binding sites in the proximal promoter region.

Figure Figure 4.

Transactivation of the COL11A2 gene by overexpression of Sp1 or Sp3, association of Sp1 and Sp3 with HDAC, CBP, and p300, and transactivation of the COL11A2 gene by HDAC inhibitors in Saos-2 cells. (A) Effect of Sp1 or Sp3 expression construct and/or mithramycin A on the activity of COL11A2 promoter construct H149. Transfection in duplicate was repeated at least three times, and the luciferase activity was normalized to the amount of protein in the cell layer. Results are shown as average ± SE. (B) Effect of Sp1 or Sp3 expression construct and/or mithramycin A on the endogenous mRNA levels of COL11A2. Experiment in duplicate was repeated at least three times, and levels of β2-microglobulin transcript were used to normalize COL11A2 expression levels. Results are shown as average ± SE. (C) Lysates from Saos-2 cells transfected with Sp1 or Sp3 expression construct or with control empty plasmid were electrophoresed and blotted with an anti-Sp1 antibody (left) or an anti-Sp3 antibody (right). *p < 0.05 compared with empty vector; **p < 0.05 compared with the same expression construct without mithramycin A. (D and E) Saos-2 cell lysate (lane 1) was immunoprecipitated (IP) with an anti-Sp1 antibody (lane 2), an anti-Sp3 antibody (lane 3), an anti-HDAC1 antibody (lane 4), an anti- CBP antibody (lane 5), an anti- p300 antibody (lane 6), or a normal rabbit IgG (lane 7). Lysate and IP samples were electrophoresed and blotted with (D) an anti-Sp1 antibody or (E) an anti-Sp3 antibody. (F) Effect of TSA (100 ng/ml) or NaB (5mM) and/or mithramycin A (100 nM) on the activity of COL11A2 promoter construct H149. Transfection in duplicate was repeated at least three times, and the luciferase activity was normalized to the amount of protein in the cell layer. Results are shown as average ± SE. (G) Effect of TSA (100 ng/ml) or NaB (5mM) and/or mithramycin A (100 nM) on the endogenous mRNA levels of COL11A2. Experiment in duplicate was repeated at least three times, and levels of β2-microglobulin transcript were used to normalize COL11A2 expression levels. Results are shown as average ± SE. *p < 0.05 compared with control; **p < 0.05 compared with the same HDAC inhibitor without mithramycin A.

Sp1 and Sp3 are physically associated with HDAC, CBP, and p300 in Saos-2 cells

The Sp1 family of transcription factors is known to function as activators or repressors depending on the coregulators with which they interact.(19) To study this, Saos-2 cell lysates were immunoprecipitated with antibodies to Sp1, Sp3, HDAC1, CBP, or p300. Endogenous Sp1 was co-precipitated with each of the four antibodies, whereas a control rabbit IgG did not immunoprecipitate Sp1 (Fig. 4D, lanes 3–7). Similarly, endogenous Sp3 was also co-precipitated by the Sp1, HDAC1, CBP, or p300 antibodies, but not by a control rabbit IgG (Fig. 4E, lanes 2 and 4–7). These results show that Sp1 and Sp3 bind to HDAC, CBP, and p300 in Saos-2 cells.

HDAC inhibition activates the COL11A2 promoter and induces endogenous COL11A2 expression in Saos-2 cells

The findings that Sp1 and Sp3 upregulated human COL11A2 gene transcription through Sp1 binding sites in the proximal promoter region and interact with CBP and p300, which are prototypical transcriptional co-activators, are consistent with previous reports.(19) We next analyzed the role of HDAC, a prototypical transcriptional co-repressor, on COL11A2 gene regulation as a component of the Sp1/Sp3 protein complex. To study whether inhibition of HDAC transactivates the COL11A2 promoter, Saos-2 cells were transfected with the COL11A2 promoter reporter construct and incubated with two inhibitors of HDAC activity, TSA and NaB. Both potently and independently resulted in transactivation of the human COL11A2 promoter construct H149 (Fig. 4F). In contrast, mithramycin A, which interferes with DNA binding of Sp1 family of transcription factors, dramatically decreased promoter activity in Saos-2 cells, suggesting that COL11A2 activation by HDAC inhibitors requires Sp1/Sp3 binding to DNA in the promoter construct. We also found that TSA and NaB significantly increased the endogenous mRNA levels of COL11A2 by 66.9% and 83.0%, respectively, in Saos-2 cells (p < 0.05; Fig. 4G). These increases were cancelled by the addition of mithramycin A (p < 0.05). These results indicate that HDAC inhibition by TSA or NaB upregulates the human COL11A2 gene through Sp1 binding sites in the proximal promoter region. Together, COL11A2 gene transcription seems to be regulated in a balance between HDAC and CBP/p300.

Saos-2 cells express Sp7/Osterix, which activates the COL11A2 promoter and induces endogenous COL11A2 expression in Saos-2 cells

Sp7/Osterix is a recently identified zinc finger–containing transcription factor, expressed in the osteoblasts of all endochondral and membranous bones. In Osterix-null mutant mice, neither endochondral nor intramembranous bone is formed, and osteoblast differentiation is arrested.(17) Because Sp7/Osterix expression is essential for osteoblast differentiation, and the protein has a zinc finger DNA-binding domain conserved in the Sp1 protein family, we hypothesized that Sp7 is involved in regulating the COL11A2 gene in Saos-2, human osteosarcoma-derived osteoblastic cells. RT-PCR analysis showed that Saos-2 cells expressed SP7/Osterix mRNA, whereas chondrocytes did not (Fig. 5A, lanes 1 and 2). Chromatin immunoprecipitation assay revealed that the COL11A2 promoter sequence was cross-linked to Sp7 after treatment with formaldehyde (Fig. 5C, lane 4), suggesting that Sp7 directly interacts with COL11A2 promoter in vivo. To investigate whether Sp7 can transactivate the COL11A2 promoter, an Sp7 expression construct was co-transfected with the COL11A2 promoter reporter construct into Saos-2 cells. Subsequent expression of Sp7 potently transactivated the human COL11A2 promoter construct H149 (Fig. 5D). Mithramycin A dramatically decreased promoter activity in Saos-2 cells, indicating that action of Sp7 was dependent on the ability of Sp7 to bind DNA. We also found that Sp7 significantly increased the endogenous mRNA levels of COL11A2 by 24.7% in Saos-2 cells (p < 0.05; Fig. 5E). This increase was cancelled by the addition of mithramycin A (p < 0.05). These results indicate that Sp7 upregulates the human COL11A2 gene transcription through Sp1 binding sites in the proximal promoter region.

Figure Figure 5.

Transactivation of the COL11A2 gene by overexpression of Sp7 in Saos-2 cells. (A) RT-PCR analysis of Sp7/Osterix transcripts from Saos-2 cells (lane 1), human chondrocytes (lane 2), and HeLa cells (lane 3). (B) Lysates from Saos-2 cells transfected with Sp7 expression construct or with control empty plasmid were electrophoresed and blotted with an anti-Sp7/Osterix antibody (from Toshihisa Komori and Takashi Fujita, Nagasaki, Japan). (C) ChIP assay. ChIP assay of a 118-bp COL11A2 genomic fragment (–167 to −50 bp to the transcriptional start site) was performed with Saos-2 cells. Antibodies used in the assay are indicated. Input DNA represents a portion of the sonicated chromatin before immunoprecipitation, used as the positive control. (D) Effect of Sp7 expression construct and/or mithramycin A on the activity of COL11A2 promoter construct H149. Transfection in duplicate was repeated at least three times, and the luciferase activity was normalized to the amount of protein in the cell layer. Results are shown as average ± SE. (E) Effect of Sp7 expression construct and/or mithramycin A on the endogenous mRNA levels of COL11A2. Experiment in duplicate was repeated at least three times, and levels of β2-microglobulin transcript were used to normalize COL11A2 expression levels. Results are shown as average ± SE. *p < 0.05 compared with empty vector; **p < 0.05 compared with the same expression construct without mithramycin A.

Selective inhibition of Sp1, Sp3, or Sp7/Osterix by siRNA decreases the COL11A2 promoter activity and endogenous COL11A2 expression in Saos-2 cells

To examine the relative contribution of the three Sp1 family members to COL11A2 transcription in Saos-2 osteoblast-like cells, we selectively inhibited each one in turn using siRNA. Selective inhibition of Sp1, Sp3, or Sp7 decreased its mRNA level by 72.7%, 79.1%, or 84%, respectively (Fig. 6A). siRNA treatment of Sp1, Sp3, or Sp7 significantly decreased the activity of human COL11A2 promoter construct H149 (p < 0.05; Fig. 6B) and the endogenous mRNA levels of COL11A2 (p < 0.05; Fig. 6C). These results imply that COL11A2 gene transcription in Saos-2 cells is under the control of Sp1, Sp3, and Sp7.

Figure Figure 6.

Selective inhibition of Sp1, Sp3 or Sp7/Osterix by small interfering RNA (siRNA) in Saos-2 cells. (A) Relative mRNA levels of Sp1, Sp3, or Sp7 after siRNA treatment (left). Experiment in duplicate was repeated at least three times, and the results are shown as average ± SE. Selective inhibition of Sp1, Sp3, or Sp7 was confirmed by RT-PCR analysis (right) and real-time PCR analysis (data not shown). (B) Effect of selective inhibition of Sp1, Sp3, or Sp7 on the activity of COL11A2 promoter construct H149. Transfection in duplicate was repeated at least three times, and the luciferase activity was normalized to the amount of protein in the cell layer. Results are shown as average ± SE. (C) Effect of selective inhibition of Sp1, Sp3, or Sp7 on the endogenous mRNA levels of COL11A2. Experiment in duplicate was repeated at least three times, and levels of β2-microglobulin transcript were used to normalize COL11A2 expression levels. Results are shown as average ± SE. *p < 0.05 compared with control.

Primary cultures of osteoblasts express both COL11A2 and Sp7/Osterix, and Sp1, Sp3, and Sp7/Osterix upregulate COL11A2 promoter activity and endogenous gene expression in primary cultures of osteoblasts

We analyzed primary culture of osteoblasts to ascertain that our data are not a peculiarity of the Saos-2 transformed cell line. RT-PCR analysis showed that primary cultures of osteoblasts expressed both COL11A2 and Sp7/Osterix mRNA (Fig. 7A). Transient transfection assays using luciferase reporter constructs identified a minimum sequence between +133 and +380 bp of the mouse Col11a2 promoter for the activity in primary cultures of osteoblasts (data not shown). The evolutionally conserved human COL11A2 promoter sequence between −149 and +27 bp (construct H149) showed equivalent activity to the intact mouse Col11a2 promoter (construct −742) when transfected into primary cultures of osteoblasts, and mutating the three Sp1 sites of construct H149 in a various combination significantly decreased the activity (data not shown). To study whether Sp1, Sp3, or Sp7 can transactivate the COL11A2 promoter, expression construct for each Sp1 family protein was co-transfected with the COL11A2 promoter reporter construct into primary cultures of osteoblasts. Expression of Sp1, Sp3, or Sp7 transactivated the human COL11A2 promoter construct H149 (p < 0.05; Fig. 7B). These effects were cancelled by the addition of mithramycin A, suggesting that transactivation by Sp1, Sp3, or Sp7 was dependent on the functional Sp1 binding sites. We also found that Sp1, Sp3, or Sp7 potently increased the endogenous mRNA levels of COL11A2 (Fig. 7C). Mithramycin A significantly decreased the endogenous mRNA levels of COL11A2 (p < 0.05). Furthermore, siRNA treatment of Sp1, Sp3, or Sp7 significantly decreased the activity of human COL11A2 promoter construct H149 (p < 0.05; Fig. 7D) and the endogenous mRNA levels of COL11A2 (p < 0.05; Fig. 7E). These results indicate that Sp1, Sp3, and Sp7 upregulate the human COL11A2 transcription through Sp1 binding sites in the primary cultures of osteoblasts.

Figure Figure 7.

Transactivation of the COL11A2 gene by overexpression of Sp1, Sp3, or Sp7/Osterix in primary cultures of osteoblasts. (A) RT-PCR analysis of COL11A2 (lanes 1 and 2) and Sp7/Osterix (lane 3) from primary cultures of osteoblasts. Splice variants containing different combinations of exons 6–8 covering the N-terminal COL11A2-coding region are identified. Exon 0 lacks all three exons. (B) Effect of Sp1, Sp3, or Sp7 expression construct and/or mithramycin A on the activity of COL11A2 promoter construct H149. Transfection in duplicate was repeated at least three times, and the luciferase activity was normalized to the amount of protein in the cell layer. Results are shown as average ± SE. (C) Effect of Sp1, Sp3, or Sp7 expression construct and/or mithramycin A on the endogenous mRNA levels of COL11A2. Experiment in duplicate was repeated at least three times, and levels of β2-microglobulin transcript were used to normalize COL11A2 expression levels. Results are shown as average ± SE. *p < 0.05 compared with empty vector; **p < 0.05 compared with the same expression construct without mithramycin A. (D) Effect of selective inhibition of Sp1, Sp3, or Sp7 on the activity of COL11A2 promoter construct H149. Transfection in duplicate was repeated at least three times, and the luciferase activity was normalized to the amount of protein in the cell layer. Results are shown as average ± SE. (E) Effect of selective inhibition of Sp1, Sp3, or Sp7 on the endogenous mRNA levels COL11A2. Experiment in duplicate was repeated at least three times, and levels of β2-microglobulin transcript were used to normalize COL11A2 expression levels. Results are shown as average ± SE. *p < 0.05 compared with control.

DISCUSSION

To understand the transcriptional regulation of the human α2 (XI) collagen gene (COL11A2) in Saos-2 osteosarcoma-derived osteoblastic cells, we characterized the function of the −149 to +27-bp segment of the promoter by transfection assay using a luciferase reporter gene and EMSA. The findings showed that three functional Sp1 binding sites, rather than the tandem repeat of the tcc trinucleotides (TRT), tccctcc located between −120 and −114 bp found to be involved in COL11A2 regulation by EWS/ERG,(11) were critical for both transcriptional activity and DNA-nuclear protein binding in Saos-2 cells (Figs. 2 and 3). We showed that Sp1 and Sp3 directly bound to the Sp1 sites in the proximal promoter region of the COL11A2 gene both in vitro and in vivo (Fig. 3). Functional analysis by luciferase constructs and real-time PCR showed that overexpression of Sp1 or Sp3 in Saos-2 cells significantly increased COL11A2 promoter activity and endogenous COL11A2 mRNA levels through functional Sp1 binding sites (Fig. 4). We also showed that Sp1 and Sp3 form a multiprotein complex with CBP/p300 and HDAC in Saos-2 cells, and HDAC inhibition significantly increased COL11A2 promoter activity and endogenous COL11A2 mRNA levels through functional Sp1 binding sites (Fig. 4). Next, we confirmed that Saos-2 cells expressed Sp7/Osterix but chondrocytes prepared from adult human rib cartilage did not. Despite reports of Sp7/Osterix mRNA expression in chondrocytes,(34) our results suggest that not all chondrocytes express Sp7/Osterix. Chromatin immunoprecipitation revealed that Sp7/Osterix occupied COL11A2 promoter in vivo. Functional analysis by luciferase constructs and real-time PCR showed that overexpressing Sp7 in Saos-2 greatly increased COL11A2 promoter activity and endogenous COL11A2 mRNA levels through functional Sp1 binding sites (Fig. 5). Selective inhibition of Sp1, Sp3, or Sp7/Osterix by siRNA decreased COL11A2 promoter activity and endogenous gene expression (Fig. 6). The findings were extended to primary cultures of osteoblasts, which also expressed both COL11A2 and Sp7/Osterix and in which overexpression of Sp1, Sp3, or Sp7/Osterix upregulated COL11A2 promoter activity and endogenous gene expression through Sp1 binding sites. We also confirmed that selective inhibition of Sp1, Sp3 or Sp7 by siRNA decreased COL11A2 promoter activity and endogenous gene expression in these cells (Fig. 7). Working together, both the ubiquitous Sp1 family proteins, Sp1 and Sp3, and the tissue-specific Sp7/Osterix are involved in the transcriptional regulation of the human COL11A2 gene in Saos-2 osteoblast-like cells as well as primary cultures of osteoblasts.

Transcription of collagen genes is controlled by a complex interaction of transcription factors. Previous studies have identified the ubiquitous factors, Sp1 and Sp3, and CCAAT binding factor (CBF)/nuclear factor for Y box (NF-Y) for activity of the α1 (I) and α2 (I) promoters.(35–40) The structures and functions of the two promoters are likely to be similar, because the products function in the same type I collagen molecule [α1(I)2α2(I)]. Type I collagen forms the matrix of bone and type II collagen that of hyaline cartilage, so they are clearly differentially expressed but their basic mechanism of gene regulation seems to be similar, through Sp1 and Sp3.(41–43)

Previous studies have shown that the proximal promoter regions of COL11A1 and COL11A2 resemble those of housekeeping genes, with several GC boxes (candidate Sp1 binding sites) and no TATA box.(31,32,44,45) Although the most abundant molecular form of type XI collagen is the heterotrimer [α1(XI)α2(XI)α3(XI)], the promoters of α1(XI) and α2(XI) differ in structure and properties.(30,45–47) For example, the CCAAT box characteristic of genes for CBF/NF-Y proteins was found to be critical for the activation of the COL11A1 promoter,(48) but is absent from COL11A2. Understanding how the three type XI collagen chains are coordinately or differentially expressed will require further study. The protein products of the collagen V/XI gene family form a spectrum of heterotrimers, not simply type V [α1(V)2α2(V)] and type XI [α1(XI)α2(XI)α3(X)] molecules.(49,50) Tissue-specific differences in function of the different chain combinations seem likely.

In conclusion, these findings show that COL11A2 expression in Saos-2 cells is regulated by the Sp1 family of proteins through ubiquitous and osteoblast-specific interactions. Although the biological meaning of COL11A2 expression in osteoblasts is not clear, there is evidence for a potential role in development,(12) regeneration,(14) oncogenesis,(13) and hyperossification diseases.(15)

Acknowledgements

The authors thank Dr Guntram Suske for providing expression vectors. This work was supported, in part, by Japan Society for the Promotion of Science Grant 16591496, Japan Orthopaedics and Traumatology Foundation Grant 0142, the Nakatomi Foundation, and NIH Grants AR36794 and HD 22657.

Ancillary