Advertisement

Dentin Matrix Protein 4, a Novel Secretory Calcium-binding Protein That Modulates Odontoblast Differentiation*

Open AccessPublished:March 16, 2007DOI:https://doi.org/10.1074/jbc.M701547200
      Formation of calcified tissues is a well regulated process. In dentin, the odontoblasts synthesize several biomolecules that function as nucleators or inhibitors of mineralization. To identify genes that are odontoblast-specific, a subtractive hybridization technique was employed that resulted in the identification of a previously undescribed novel gene synthesized by the odontoblasts. Based on the nomenclature in our laboratory, this gene has been named dentin matrix protein 4 (DMP4). The protein encoded by mouse DMP4 cDNA contained 579 amino acids, including a 26-amino acid signal peptide. Analysis of the protein sequence demonstrated the presence of a Greek key calcium-binding domain and one conserved domain of unknown function in all the species examined thus far. Calcium binding property was confirmed by 45Ca binding assays and the corresponding change in conformation by far-ultraviolet circular dichroism. Northern analysis demonstrated high expression levels of a single 3-kb mRNA transcript in tooth, whereas low expression levels were detected in other tissues. In situ hybridization analysis showed high expression levels of DMP4 in odontoblasts and low levels in osteoblasts and ameloblasts during tooth development. Gain and loss of function experiments demonstrated that DMP4 had the potential to differentiate mesenchymal precursor cells into functional odontoblast-like cells.
      Mammalian tooth development requires a series of reciprocal epithelial-mesenchymal interactions during the odontogenic developmental program (
      • Thesleff I.
      • Keranen S.
      • Jernvall J.
      ). These interactions are necessary for the polarization and terminal differentiation of odontoblasts. Two important components of the dentin matrix are type I collagen, which acts as an inert scaffold, and the noncollagenous proteins, which function as regulators of mineralization (
      • Weiner S.
      • Wagner H.D.
      ). During dentin formation, the odontoblast cells synthesize a number of noncollagenous proteins. Most of these proteins are proteolytically cleaved or post-translationally modified before being secreted to the extracellular matrix. It is generally believed that the noncollagenous proteins are involved in the organization of the matrix and the regulation of the mineralization process (
      • He G.
      • Dahl T.
      • Veis A.
      • George A.
      ,
      • He G.
      • George A.
      ).
      Several noncollagenous proteins from the dentin matrix have been identified in our laboratory, and based on their tissue localization they were named “dentin matrix proteins” (DMPs).
      The abbreviations used are: DMP, dentin matrix protein; DSPP, dentin sialophosphoprotein; DGI-II, dentinogenesis imperfecta type II; PBS, phosphate-buffered saline; RT, reverse transcription; siRNA, small interfering RNA; BSA, bovine serum albumin; shRNA, small hairpin RNA.
      2The abbreviations used are: DMP, dentin matrix protein; DSPP, dentin sialophosphoprotein; DGI-II, dentinogenesis imperfecta type II; PBS, phosphate-buffered saline; RT, reverse transcription; siRNA, small interfering RNA; BSA, bovine serum albumin; shRNA, small hairpin RNA.
      The complete sequences of dentin matrix proteins 1, 2, and 3 were deduced from cDNAs derived from a rat odontoblast library. DMP1 is an acidic protein present in the mineralized matrix of bone and dentin (
      • George A.
      • Sabsay B.
      • Simonian P.A.
      • Veis A.
      ). DMP2 is highly acidic and is mostly present in the dentin matrix and contains predominantly aspartic acid and phosphoserine amino acids (
      • George A.
      • Bannon L.
      • Sabsay B.
      • Dillon J.W.
      • Malone J.
      • Veis A.
      ). DMP3, also known as dentin sialophosphoprotein (DSPP) (mouse), is a compound protein of dentin sialoprotein and a phosphophoryn like domain at the C-terminal end (
      • George A.
      • Srinivasan R.
      • Thotakura R.S.
      • Liu K.
      • Veis A.
      ,
      • MacDougall M.
      • Simmons D.
      • Luan X.
      • Nydegger J.
      • Feng J.
      • Gu T.T.
      ). However, the processing of this protein in vivo is still unclear.
      Dentinogenesis imperfecta type II (DGI-II) is an autosomal dominant disorder of the tooth that primarily affects dentin biomineralization. Opalescent dentin with obliterated pulp chambers are the characteristic features in DGI-II (
      • Witkop Jr., C.J.
      • MacLean C.J.
      • Schmidt P.J.
      • Henry J.L.
      ). Recently, DSPP has been associated with the pathogenesis of DGI-II with and without progressive hearing loss in four independent Chinese families (
      • Zhang X.
      • Zhao J.
      • Li C.
      • Gao S.
      • Qiu C.
      • Liu P.
      • Wu G.
      • Qiang B.
      • Lo W.H.
      • Shen Y.
      ,
      • Xiao S.
      • Yu C.
      • Chou X.
      • Yuan W.
      • Wang Y.
      • Bu L.
      • Fu G.
      • Qian M.
      • Yang J.
      • Shi Y.
      • Hu L.
      • Han B.
      • Wang Z.
      • Huang W.
      • Liu J.
      • Chen Z.
      • Zhao G.
      • Kong X.
      ). These mutations have not been found in Caucasian families of European decent diagnosed with DGI-II (
      • MacDougall M.
      • Thiemann F.
      • Ta H.
      • Hsu P.
      • Chen L.S.
      • Snead M.L.
      ). These data suggest the presence of other genes synthesized by the odontoblasts and expressed during dentinogenesis, residing within the gene cluster.
      In our efforts to better describe the molecular mechanisms by which odontoblasts control dentin homeostasis, a subtractive cDNA library was constructed to identify genes that were differentially expressed in rat immortalized odontoblast cells (T4-4) versus osteoblastic cells (MC3T3-E1) (
      • Hao J.
      • He G.
      • Narayanan K.
      • Zou B.
      • Lin L.
      • Muni T.
      • Ramachandran A.
      • George A.
      ). This paper describes the identification and characterization of a novel gene obtained using this technique. In keeping with the systematic nomenclature in our laboratory, we have assigned the name DMP4 to the protein and the corresponding gene as DMP4. The results from this study suggest that DMP4 might function during odontoblast differentiation and dentin matrix formation.

      MATERIALS AND METHODS

      Odontoblast and Osteoblast Cell Culture—The immortalized rat odontoblast, T4-4 (
      • Hao J.
      • Narayanan K.
      • Ramachandran A.
      • He G.
      • Almushayt A.
      • Evans C.
      • George A.
      ), and mouse preosteoblastic cells, MC3T3-E1, were cultured separately with Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum (Celgro). These cells were grown to subconfluence, and poly(A)+ RNA was extracted using a FastTrack Kit (Invitrogen) according to the manufacturer's protocol.
      Suppressive Subtractive Hybridization and Cloning of Fulllength Mouse DMP4—Suppressive subtractive hybridization screen was performed following the manufacturer's instructions (Clontech). Briefly mRNA from T4-4 cells was used as a “tester” and mRNA from MC3T3-E1 cells was used as a “driver.” From this screen, we identified a partial cDNA sequence (450 bp) that corresponded to a partial coding region of mouse DMP4 (
      • Hao J.
      • He G.
      • Narayanan K.
      • Zou B.
      • Lin L.
      • Muni T.
      • Ramachandran A.
      • George A.
      ). The upstream sequence was obtained from mouse tooth germ mRNA by performing a 5′-rapid amplification of cDNA ends technique (Clontech) utilizing a series of gene-specific oligonucleotide primers. The complete mouse nucleotide sequence was deposited in the GenBank™ (accession number AY778962).
      RNA Isolation and Northern Blot Analysis—For conventional Northern blot analysis, poly(A) RNA (2 μg) from T4-4 and MC3T3-E1 cells or total RNA (10 μg) from adult mouse tissues were isolated and fractionated by electrophoresis through 0.8% agarose gels containing formaldehyde and blotted onto Hybond nylon membrane (Amersham Biosciences). Mouse full-length DMP4 cDNA was labeled with [α-32P]dCTP using a random labeling system (Ambion). Hybridization was performed in ExpressHyb hybridization buffer (Clontech) at 68 °C overnight. The membranes were washed twice with 2× SSC, 0.1% SDS at 42 °C for 20 min; washed twice in 0.1× SSC, 0.1% SDS at 56 °C for 30 min; and exposed to x-ray film at –80 °C overnight. The density of each band was quantified by a Kodak digital imaging system.
      Expression of DMP4 during Mineralized Nodule Formation—The mineralization microenvironment was created by treating T4-4 cells at passage 35 (80–90% confluent) with 100 μg/ml ascorbic acid and 10 mm β-glycerophosphate along with 10 nm dexamethasone for 45 days. mRNA was extracted from the T4-4 cells at 0, 15, 30, and 45 days and subjected to Northern blot analysis to check the expression of DMP4 and other dentin matrix proteins during mineralized nodule formation.
      In Situ Hybridization—The 3-kb full-length DMP4 cDNA was linearized for in vitro transcription using appropriate enzymes. The antisense and sense cRNA were prepared by in vitro transcription using T3 or T7 polymerase (Promega) in the presence of 35S-labeled dUTP (ICN). The radiolabeled probes were hydrolyzed to a final length of ∼300 bp. The experimental protocol for hybridization was conducted as described before (
      • Hao J.
      • Zou B.
      • Narayanan K.
      • George A.
      ). Briefly, tissues from 3-day-old postnatal mice were fixed overnight with 4% paraformaldehyde in phosphate-buffered saline (PBS), pH 7.4, at 4 °C, washed overnight with PBS at 4 °C, dehydrated, and embedded in paraffin. Sections of 5 μm were cut, mounted on charged glass slides, dewaxed in xylene, and rehydrated. After washing in PBS these sections were digested with 200 milliunits/ml proteinase K, postfixed, and acetylated with 0.25% acetic anhydride. The sections were hybridized overnight at 55 °C with radiation-labeled riboprobes. After hybridization, the sections were washed with 50% formamide, 2× SSC (0.3 m NaCl containing 0.03 sodium citrate) for 30 min at 55 °C, digested with RNase A, washed once with 2× SSC, and washed twice with 0.2× SSC for 20 min at 55 °C. The signal was visualized with NBT2 silver emulsion (Kodak). The developed slides were counterstained with hematoxylin and eosin, dehydrated, and mounted. The corresponding sense probes served as negative controls.
      Establishment of C3H10T1/2, HAT-7, and MC3T3-E1 Stable Cell Lines Transfected with DMP4—The entire coding sequence of DMP4 cDNA was amplified by PCR using Pfu DNA polymerase (Stratagene) and inserted into a mammalian expression vector p3XFLAG-CMV-14 (pCMV) (Sigma) containing cytomegalovirus promoter with a 3XFLAG epitope tag at the C terminus (pCMV-DMP4). The subcloned cDNA fragment was confirmed by sequencing. Transfections were performed in 6-well culture plates containing either C3H10T1/2 (embryonic mesenchymal mouse cell-line), HAT-7(undifferentiated mouse ameloblast cell-line) (
      • Harada H.
      • Ichimori Y.
      • Yokohama-Tamaki T.
      • Ohshima H.
      • Kawano S.
      • Katsube K.
      • Wakisaka S.
      ,
      • Xu L.
      • Harada H.
      • Yokohama-Tamaki T.
      • Matsumoto S.
      • Tanaka J.
      • Taniguchi A.
      ), or MC3T3-E1 cells (mouse preosteoblast cell line). pCMV-DMP4 or the mock vector (pCMV) was transfected into cells by lipofection using SuperFect Transfection Reagent (Qiagen) according to the manufacturer's instructions. After 48 h, the cells were incubated in a maintenance medium containing 700 μg/ml G418 (Invitrogen). The drug-resistant cells were isolated after 2 weeks. To confirm the overexpression of DMP4 in the transfected cells, we performed Northern blot, Western blot, and immunohistochemistry.
      Generation of DMP4 Knockdown Stable Cell Lines—Three target sequences of the mouse DMP4 gene were selected and designed using DHARMACON software (siRNA 793–811, siRNA 1057–1075, and siRNA 1836–1854). The oligonucleotides were synthesized by IDT DNA (Coralvill, IA), annealed, and inserted into the pSIREN (Clontech) vector to generate three shRNAs. All three plasmids were tested for down-regulation of DMP4 expression by transient transfection with RT-PCR analysis (supplemental Fig. S4). One mDMP4-siRNA (siRNA 1836–1854) was selected for all experiements based on knockdown of DMP4 expression level. The sequence of the small hairpin mDMP4 (DMP4-shRNA) was as follows: sense, 5′-gatccACCGAATCCTGGACATCATTTCAAGAGAATGATGTCCAGGATTCGGTTTTTTTACGCGTg-3′; antisense, 5′-aattcACGCGTAAAAAAACCGAATCCTGGACATCATTCTCTTGAAATGATGTCCAGGATTCGGTg-3′. The sequence was designed to target mDMP4 at ACCGAATCCTGGACATCAT (position in gene sequence: nucleotides 1836–1854). Vector with scrambled shRNA was also constructed to serve as a control (Control-shRNA). The scrambled shRNA has been tested by BD Clontech and does not match with any known gene sequences. All three cell lines overexpressing DMP4, namely C3H10T1/2, HAT-7, and MC3T3-E1, were transfected with pSIREN-DMP4-shRNA and pSIREN-control-shRNA, respectively, by lipofection as mentioned above. The transfected cells were selected with culture medium containing 100 μg/ml hygromycin B (Cellgro) for 2 weeks. Western blot and RT-PCR were used to analyze gene silencing efficacy (see Figs. 2 and 7a), and the functional effect of siRNA was evaluated by mineralized nodule formation assay (Fig. 7b).
      Figure thumbnail gr2
      FIGURE 2DMP4 is a secretory protein. a, the cell lysates (cell) and culture medium (medium) from C3H10T1/2, HAT-7 and MC3T3-E1 cell lines transfected with a plasmid carrying FLAG-tagged mouse DMP4 protein or pSIREN-DMP4-shRNA were immunoblotteded and subjected to Western blot analysis, using anti-FLAG antibody. A distinct band around 66 kDa was detected by anti-FLAG antibody in cell lysates (Cell) and culture medium (Medium). Lanes 1, DMP4 overexpression cell; lanes 2, mock cell; lanes 3, DMP4 shRNA cell. b, a densitometric representation of the Western blot data is shown.
      Figure thumbnail gr7
      FIGURE 7a, expression of odontoblastic and osteoblastic biochemical markers in DMP4 overexpressing cell lines. Stably transfected cell lines were analyzed by RT-PCR as described under “Materials and Methods.” Lanes 1, stable C3H10T1/2, HAT-7, and MC3T3-E1 cell lines overexpressing DMP4; lanes 2, mock C3H10T1/2, HAT-7, and MC3T3-E1; lanes 3, DMP4 overexpressed cells stably transfected with the DMP4 shRNA. PCR products were resolved on a 2% agarose gel, and odontoblast and osteoblast markers were identified. A densitometric representation of the RT-PCR data is shown in the panel ab. b, DMP4 overexpression and mineralization in different cell lines with von Kossa staining. The mineralized nodule formed after 45 days was photographed with a Nikon camera.
      RT-PCR and Expression of Matrix Genes—DNase I (RNase-free, RQ1; Promega)-treated RNA was used for all RT-PCRs. 3 μg of total RNA from stably transfected and mock cells was reverse transcribed for 90 min at 42 °C with Superscript II (Invitrogen). PCR Supermix (Invitrogen) was used in all of the PCRs. Primers for the matrix genes used in the PCR were designed from available sequences at the National Center for Biotechnology Information gene data bank (supplemental Fig. S1). The PCR products were verified either by sequencing or by restriction mapping.
      Western Blot Analysis—Stably transfected cells with DMP4 were cultured in 6-well plates (Nunc) in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. When cultures were 80% confluent, the serum-containing medium was replaced with serum-free medium (Invitrogen) followed by further culturing at 37 °C in the presence of 5% CO2 for 3 days. Total protein was extracted from the cells and medium as described by Harlow and Lane (
      • Harlow E.
      • Lane D.
      Antibodies: A Laboratory Manual.
      ), and the concentration was estimated by Bradford micro assay (Bio-Rad). 20 μg of the total protein was fractionated on a 10% SDS-PAGE gel. Transfer of proteins to nitrocellulose was performed according to Towbin et al. (
      • Towbin H.
      • Staehelin T.
      • Gordon J.
      ). The blots were incubated with polyclonal anti-FLAG antibody (Sigma) and then with alkaline phosphatase-conjugated anti-rabbit secondary antibody (Sigma). The blots were washed with distilled water and then developed in 20 mm Tris-Cl, pH 7.5, 200 mm NaCl using alkaline phosphatase-conjugated substrate kit (Bio-Rad).
      Immunohistochemical Analysis—For immunostaining, 50% confluent stably transfected cells from the second passage were grown on coverslips and fixed with paraformaldehyde. Fixed cells were incubated with anti-FLAG polyclonal antibody in the presence of 5% BSA for 4 h. Upon washing with PBS containing 1% Triton X-100, the cells were incubated with appropriate secondary antibody (fluorescent labeled) for 2 h. The coverslips were then mounted and observed under laser confocal microscope (Zeiss; LSM 510). Oregon, propidium iodide, and Hoechst dye were purchased from Molecular Probes Inc. (Eugene).
      von Kossa Staining for Mineralized Nodules—The C3H10T1/2, HAT-7, and MC3T3-E1 cells or stable transfected cells that were allowed to undergo differentiation under mineralization conditions (45 days in culture) were fixed with 10% formalin in neutral buffer (Sigma) for 15 min. The slides were washed with distilled water and then treated with 1% AgNO3 for 1 h, washed again with distilled water, and treated with 2.5% sodium thiosulfate for 5 min. The specimens were then examined under a light microscope. Postnatal day 3, rat head was used as positive control.
      Expression and Purification of Recombinant DMP4—Mouse DMP4 protein spanning the entire coding region was expressed in bacteria using pGEX-4T3 (Amersham Biosciences) (
      • Srinivasan R.
      • Chen B.
      • Gorski J.P.
      • George A.
      ). Recombinant Escherichia coli BL-21 cells containing mouse DMP4 gene (pGEX4T3/mDMP4) were grown in 50 ml of LB media with 50 μg/ml ampicillin overnight. This seed culture was inoculated in 1 liter of LB (without ampicillin) and incubated on a shaker for 2 h. The cells were induced with 1 mm isopropyl β-d-thiogalactopyranoside (Fisher) and grown further for 4 h. The cells were then centrifuged and lysed by sonication, and the expressed protein was purified using glutathione S-transferase fusion protein purification method as per the manufacturer's protocol. The protein was either eluted from glutathione-Sepharose beads using 10 mm reduced glutathione or after digestion (on column) with 1 unit of thrombin/100 μg of protein (Amersham Biosciences) for 20 min at 37 °C. In the case of thrombin digestion, the protein was eluted in 1 mm phenylmethylsulfonyl fluoride (Fisher) in PBS. The eluted protein was subjected to dialysis for 30 h at 4 °C with three successive changes against 0.5× PBS.
      45Ca Binding Assay—The 45Ca binding assay was based on the procedure described previously (
      • He G.
      • Dahl T.
      • Veis A.
      • George A.
      ). Specifically, 1 μg of recombinant DMP4 protein was dotted onto polyvinylidene difluoride membrane (Bio-Rad) and washed with a solution containing 60 mm KCl and 10 mm imidazole-HCl, pH 7.4, four times at 15 min each. Afterward, the membrane was incubated in the same buffer containing 1 mCi/liter of 45CaCl2 for 15 min, then rinsed with 50% ethanol for 10 min, and air-dried. Autoradiographs of the membrane were obtained by exposure of the dried membrane to Kodak XAR-5 x-ray film overnight. Recombinant DMP1, calmodulin, and BSA were used, in equivalent amounts, as controls.
      Far-UV Circular Dichroism—The spectra were obtained using 5 mm recombinant DMP4 in 0.1 mm HEPES, pH 7.4, or the same buffer containing 5 mm CaCl2 in a 1-mm-pathlength cell at 20 °C. The data were collected at 0.5-nm intervals at 2 nm/min. Five such runs were averaged and smoothened using the accompanying software program (
      • He G.
      • Dahl T.
      • Veis A.
      • George A.
      ).

      RESULTS

      Cloning and Characterization of the Primary Structure of DMP4—We had previously reported the construction of an odontoblast-enriched cDNA library using the suppressive subtractive hybridization technique (
      • Hao J.
      • He G.
      • Narayanan K.
      • Zou B.
      • Lin L.
      • Muni T.
      • Ramachandran A.
      • George A.
      ). One such gene present in this enriched pool is a previously undescribed 450-bp fragment named DMP4. Primers were made, and a full-length fragment was obtained from the 5′-rapid amplification of cDNA ends product of mouse tooth germ mRNA. This 3018-bp fragment contained an open reading frame of 1740 bp flanked by 536 and 742 bp of untranslated sequences at the 5′ and 3′ ends, respectively (Fig. 1). These sequences were also confirmed in a mouse IMAGE clone 5101405 obtained from ATCC that contained the full-length DMP4. Sequence similarity was verified using the GenBank™ data base. Delineation of the corresponding genomic locus was enabled by BLAST searches of the mouse genome sequence (www.ensembl.org). The gene is located on chromosome 5G2 (syntenic to human chromosome region 7p22). Protein expression and Western blot analysis identified a single 66-kDa band, and this corresponded to the full-length open reading frame of mouse DMP4 (Fig. 2).
      Figure thumbnail gr1
      FIGURE 1Characterization of mouse DMP4. Nucleotide sequence of mouse DMP4 (3018 bp) was aligned with the predicted amino acid sequence. The underlined sequence from 1–26 represents the signal peptide. The Greek key domain is illustrated in bold and italic type. The asterisk denotes the termination codon in the putative amino acid sequence of DMP4. The bold letters indicate the polyadenylation signal in DMP4.
      Bioinformatic analysis of the primary sequence using CBS Signal P V2.0 World Wide Web Prediction Server at the Center for Biological Sequence (www.cbs.dtu.dk) indicated the presence of a signal peptide cleavage site between residues 26 and 27 (Fig. 1). The calculated pI is 7.10, hence different from the other DMPs. Of the 579 amino acids, only 111 residues are comprised of aspartic acid, glutamic acid, and serines. DMP4 also contains one RGD site located at position 530 in humans (accession number AAQ09019) and at position 150 in rat (accession number XP_221975). Three potential sites for N-linked glycosylation, four tyrosine sulfation sites, and 11 cysteine residues were also identified (Table 1). The other characteristic protein domains that might be of functional importance include one Greek key domain (
      • Rajini B.
      • Shridas P.
      • Sundari C.S.
      • Muralidhar D.
      • Chandani S.
      • Thomas F.
      • Sharma Y.
      ) and one conserved domain in all species examined thus far (
      • Nalbant D.
      • Youn H.
      • Nalbant S.I.
      • Sharma S.
      • Cobos E.
      • Beale E.G.
      • Du Y.
      • Williams S.C.
      ). This conserved region is designated as the conserved C-terminal domain, and it is listed in the C-terminal domain at NCBI as DUF1193. So far, no function has been attributed to this domain.
      TABLE 1Characterization of the predicted mouse DMP4 amino acid sequence
      Amino acid positionDescription
      1-26Signal peptide
      27Cleavage site
      530 (human), 150 (rat)RGD site, cell attachment sequence
      96, 330, 465N-Glycosylation site
      296, 300, 420, 430Tyrosine sulfation sites
      21, 46, 48, 357, 362, 366, 373, 421, 495, 496, 554Cysteine residue
      To further investigate the significance of DMP4 expression in the dentin matrix, we determined whether this gene was conserved in other species. Human (EAL23705) and rat (XP_221975) full-length cDNAs and a partial cDNA from bovine (XP_601319) have been submitted to GenBank™. Overall amino acid homology between different species is high; for example, human and mouse amino acid sequences have a 69% homology (73% similar) (supplemental Fig. S2), rat and mouse are 90% homologous, and bovine and mouse sequences are 59% homologous (71% similar). In addition, a similar sequence was found in the genomes of chimpanzee, Caenorhabditis elegens, dog, Drosophila, Fugu rubripes, red jungle fowl, and tetraodon.
      DMP4 Is a Secretory Protein—Signal sequences are commonly found on proteins that are directed to the endoplasmic reticulum and either retained there or processed and transported into the Golgi apparatus and secreted from the cell. Many proteins are glycosylated during their transit through the endoplasmic reticulum and Golgi apparatus, and the mouse DMP4 protein contains three potential sites for N-glycosylation (Table 1) and might be glycosylated during the transport process. Because DMP4 does not contain an endoplasmic reticulum retention signal, we predicted that it should be detected in the medium of expressing cells; therefore, a mammalian expression vector was constructed that contained the full-length mouse DMP4 coding sequence fused to a C-terminal 3X FLAG tag. The plasmid was transfected into three cell lines: a mouse multipotential mesenchymal cell line (C3H10T1/2), a dental epithelial cell line (HAT-7), and a preosteoblast cell line (MC3T3-E1). Higher expression levels of DMP4 mRNA were detected in all three transfected cell lines by Northern blot (supplemental Fig. S3). Total protein was isolated from both the cells and the cell medium. Both protein samples were analyzed by immunoblotting using an anti-serum specific for the 3X FLAG epitope. The molecular mass of the processed form of mDMP4 was determined to be 66 kDa (Fig. 2).
      We then demonstrated the subcellular localization of the mDMP4 protein in the three DMP4 overexpressing cell lines and the control mock cells using anti-FLAG antibody. All three transfected cells displayed perinuclear and cytoplasmic staining consistent with endoplasmic reticulum localization (Fig. 3, arrowhead). In contrast, in the control mock cells DMP4 was exclusively localized within the cytoplasm.
      Figure thumbnail gr3
      FIGURE 3Immunostaining of DMP4 in stably transfected cell lines. The expression of DMP4 was analyzed by immunostaining with anti-FLAG polyclonal antibodies using confocal microscopy. A represents staining with anti-FLAG antibody, B represents staining for the nucleus (Hoechst dye), C represents staining of actin filaments (Oregon green) and D is a composite of A–C. Bar, 20 μm.
      DMP4 Expression in Vivo and during Tooth Development—Tissue expression of DMP4 was then determined by Northern blot analysis and in situ hybridization techniques using a mouse model. Northern blot detected DMP4 at very high expression level in the dental pulp-odontoblast complex. This level of expression was in concurrence with the data obtained from T4-4 odontoblast cell line, which was 10 times higher when compared with the expression in both precursor and fully differentiated MC3T3-E1 cells (Fig. 4A). It was also moderately expressed in bone, and low levels were present in kidney, liver, brain, and lung. No expression or very low expression levels were detected in spleen and skeletal muscle (Fig. 4B). The size of the message of mDMP4 in all tissues was confirmed to be about 3.0 kb.
      Figure thumbnail gr4
      FIGURE 4Expression of DMP4 in odontoblasts, osteoblasts, and various mouse tissues. A, expression of DMP4 in odontoblasts and osteoblasts by Northern blots. Poly(A) RNA from T4-4, MC3T3-E1 without mineralization media (MC3T3 W/O M), and MC3T3-E1 with mineralization solution for 30 days (MC3T3 W M) was isolated by FastTrack kit (Invitrogen). 2 μg of mRNA was resolved on a 0.8% agarose gel under denaturing conditions and transferred. The membrane was probed with random [32P]dCTP-labeled DMP4 cDNA, alkaline phosphates (ALP), or glyceraldehyde-3-phosphate dehydrogenase (GAPDH). c, a densitometric representation of the Northern data is shown. B, multiple tissues Northern blot. The panel shows 10 μg of total RNA/lane isolated from spleen, liver, lung, kidney, bone (calvaria), tooth, brain, and skeletal muscle. The blot was hybridized with DMP4 and GAPDH. The molecular sizes of DMP4 (arrowhead) are indicated on the right side. c, a densitometric representation of the Northern data is shown.
      In situ hybridization results performed on sagital sections of postnatal 3-day-old mice with antisense DMP4 probe showed high expression levels in the differentiated odontoblasts (Fig. 5, arrow). Expression was also detected in the ameloblasts (Fig. 5, arrowhead) and osteoblasts in alveolar bone (Fig. 5, asterisk). However, the amount was lower than that in odontoblasts. No expression was detected in the dental pulp. Sense probe was used as a control, and no expression was observed (data not shown).
      Figure thumbnail gr5
      FIGURE 5Expression of DMP4 during tooth development. In situ hybridization was performed with antisense DMP4 probes on sagital sections of postnatal day 3 mouse. DMP4 mRNA was highly expressed in fully differentiated and secretory odontoblasts (a and b, arrows) in the first molar. The expression level is relatively lower in ameloblasts (arrowhead) and osteoblasts of alveolar (asterisk). a′ and b′ were dark field images of a and b. b and b′ are the higher magnification images from the corresponding boxed portion from a and a′.
      Overexpression of DMP4 Accelerates Odontoblast Differentiation Process and Mineralized Nodule Formation—To monitor changes in the DMP4 mRNA expression levels during the mineralization of T4-4 in vitro, the cells were cultured in the presence of differentiation medium containing ascorbic acid, β-glycerophosphate, and dexamethasone for 45 days. Results from the Northern blot analysis demonstrated a steady increase in DMP4 mRNA expression during terminal differentiation of odontoblasts (Fig. 6). This process led to an up-regulation of the differentiation markers for odontoblasts namely DSPP and DMP2. However, fully differentiated MC3T3 cells expressed low amounts of DMP4 (Fig. 4A).
      Figure thumbnail gr6
      FIGURE 6Expression of DMPs during mineralization. T4-4 cells were induced to undergo mineralization as described earlier. mRNAs were isolated at specified time points during mineralization. Expression of DMP1, DMP2, DSPP, and DMP4 were analyzed by Northern blotting. A densitometric representation of the Northern data is shown in the right panel. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as the standard.
      Gain of function studies were performed to elucidate the function of DMP4 in dentin biomineralization. Stable clones overexpressing DMP4 were used in this study. Gene expression analyses by RT-PCR during the differentiation of overexpressed DMP4 cells with that of control mock cells were first compared. Expression of alkaline phosphatase, which is known to be synthesized during the early phase of odontoblast differentiation, was constitutively seen before and after induction, and the expression was not significantly different between overexpressing cells and mock cells (Fig. 7a). However, expression of DMP1 and DSPP, molecular markers of terminally differentiated odontoblasts, was strongly expressed in overexpressing cells and not in mock cells. We also performed cell nodule formation assays and compared the mineralization rate by von Kossa staining techniques. Results demonstrate that von Kossa staining was significantly stronger in overexpressing cells when compared with mock cells (Fig. 7b). Taken together, these results strongly suggest that DMP4 might have a regulatory role during odontoblast differentiation and in the mineralization process. These effects seemed to corroborate well with elevated DMP1 expression during odontoblast differentiation.
      Silencing DMP4 in Vitro Inhibited Odontoblast Differentiation and Mineralization—We have successfully silenced DMP4 gene expression using synthesized siRNA, which was demonstrated by RT-PCR and Western blot analysis. The results revealed that siRNA 1836–1854 target led to a remarkable suppression of DMP4 expression (lanes 3 in Figs. 2 and 7a), whereas no change was observed with the plasmid containing scrambled siRNA (data not shown).
      To further confirm the effect of pSIREN-DMP4-shRNA on odontoblast differentiation, the mRNA expression of odontoblastic phenotypic markers in stable transfection cell lines were detected by RT-PCR. The expression of DMP1 and DSPP in DMP4 shRNA cell lines (lanes 3 in Figs. 2 and 7a) was significantly down-regulated compared with DMP4 transfected cell lines (lanes 1 in Figs. 2 and 7a). However, there was no significant reduction of ALP activity in DMP4 knockdown cell lines (Fig. 7a). To further confirm the gene silencing efficacy, terminal differentiation of odontoblast was also assessed by von Kossa staining, and the results showed that mineralization was significantly suppressed in the presence of pSIREN-DMP4-shRNA (Fig. 7b). These results obtained from gene silencing experiments corroborated well with the overexpression studies.
      Silencing DMP4 in normal C3H10T1/2, HAT-7, MC3T3-E1, and even T4-4 cell lines caused the cells to undergo senescence or underwent apoptosis (data not shown). We are currently working on the molecular mechanisms regulating these events.
      DMP4 Is a Calcium-binding Protein—Calcium binding is a key property necessary for proteins involved in the mineralization process. Because the putative Greek key domain was the only known calcium-binding domain identified in mouse DMP4, we therefore tested the functionality of this domain. The potential calcium binding ability of the recombinant mDMP4 protein was confirmed by the use of a dot blot technique. The results demonstrate that recombinant DMP4 clearly bound 45Ca (Fig. 8a). Calmodulin and recombinant DMP1 were used as positive controls, and BSA was used as a negative control.
      Figure thumbnail gr8
      FIGURE 8Calcium binding property of DMP4. a, the recombinant mouse DMP4 was spotted onto a polyvinylidene difluoride membrane and detected by 45Ca overlay. Calmodulin and recombinant DMP1 were used as positive controls; BSA was used as a negative control. b, Far-UV circular dichroism spectrum of recombinant full-length DMP4. 5 mm recombinant DMP4 in 0.1 mm HEPES (pH 7.4) before (solid line) and after the addition of 5 mm CaCl2 (dotted line). Strong absorbance at 217 nm indicates the appearance of β-sheet formation (arrow).
      Conformational Change Associated with Calcium Binding—Secondary structural changes are a good indication of protein function. Therefore, Far-UV CD measurements were performed with calcium to DMP4 ratio of 1000:1. The results demonstrate a strong negative absorption at 201 nm, indicating that the protein backbone predominantly adopts a random coil structure that is highly flexible. Upon calcium binding, there was a clear shoulder at 217 nm. This change indicted that the structure possessed a type II β-turn (Fig. 8b, arrow), which might be an important structural feature for the protein to provide the necessary microenvironment for apatite nucleation.

      DISCUSSION

      This paper describes the cloning and characterization of a novel dentin matrix protein that we have named DMP4. Computer algorithms predict a highly conserved domain of unknown function at the C-terminal domain (
      • Nalbant D.
      • Youn H.
      • Nalbant S.I.
      • Sharma S.
      • Cobos E.
      • Beale E.G.
      • Du Y.
      • Williams S.C.
      ) and one Greek key domain (
      • Rajini B.
      • Shridas P.
      • Sundari C.S.
      • Muralidhar D.
      • Chandani S.
      • Thomas F.
      • Sharma Y.
      ). This domain is conserved in the DMP4 molecule from C. elegans to humans. Analysis of the predicted amino acid sequence suggests that DMP4 shares many features with the other dentin matrix proteins. These features include a secretory protein that is highly charged, few potential serine phosphorylation sites, and a putative integrin recognition RGD domain that is present in rats and humans and absent in the mice (
      • Blundell T.
      • Lindley P.
      • Miller L.
      • Moss D.
      • Slingsby C.
      • Tickle I.
      • Turnell B.
      • Wistow G.
      ).
      The calcium-binding domains are unique in each of the four dentin matrix proteins. DMP1 has several acidic domains that contain an abundance of phosphorylation consensus sites for casein kinases I and II. One of the most prominent domains is SSSES. Overall the phosphorylated molecule has a higher charge density in the C-terminal region (
      • George A.
      • Sabsay B.
      • Simonian P.A.
      • Veis A.
      ). The DMP2 C-terminal sequence has two specific subdomains, one with a unique repetitive triplet sequence motif of aspartic acid and serines [DSS]m and the other with a smaller repetitive doublet motif sequence, [S.D.]m. These domains form excellent substrates for phosphorylation by casein kinases I and II. The majority of the serine residues in DMP2 are phosphorylated (
      • George A.
      • Bannon L.
      • Sabsay B.
      • Dillon J.W.
      • Malone J.
      • Veis A.
      ) in vivo. Dentin sialoprotein on the other hand does not exhibit the calcium binding property.
      A characteristic functional domain identified in the DMP4 sequence is the Greek key calcium-binding motif found at positions 457–488, similar to the βγ-crystallin superfamily. γ-Crystallin is a well studied protein, and the Greek key crystallin fold was first described in this protein (
      • Blundell T.
      • Lindley P.
      • Miller L.
      • Moss D.
      • Slingsby C.
      • Tickle I.
      • Turnell B.
      • Wistow G.
      ). It was later found in another lens protein, β-crystallin, and in several other non-lens proteins, which were together classified as the βγ-crystallin superfamily (
      • Wistow G.
      • Turnell B.
      • Summers L.
      • Slingsby C.
      • Moss D.
      • Miller L.
      • Lindley P.
      • Blundell T.
      ,
      • Wistow G.
      • Piatigorsky J.
      ). The three-dimensional structure of this domain has been recently solved (
      • Blundell T.
      • Lindley P.
      • Miller L.
      • Moss D.
      • Slingsby C.
      • Tickle I.
      • Turnell B.
      • Wistow G.
      ). The crystallin fold, also called the βγ motif, is a supersecondary structure formed from the symmetrical association of the two Greek key motifs that are organized as two four-stranded anti-parallel β-sheets (
      • Blundell T.
      • Lindley P.
      • Miller L.
      • Moss D.
      • Slingsby C.
      • Tickle I.
      • Turnell B.
      • Wistow G.
      ,
      • Wistow G.
      • Turnell B.
      • Summers L.
      • Slingsby C.
      • Moss D.
      • Miller L.
      • Lindley P.
      • Blundell T.
      ). The crystallin fold is a protein domain in which aromatic residues Tyr/Phe/Trp at position 1 and Gly at position 8 constitute the conserved sequence (Y/F/W)XXXXXXG, followed by a Ser at positions 28–34 from the first Y/F/W residue, and this sequence is repeated within 40 residues. Gly-8 is irreplaceable and is required for forming a dihedral angle, which is not possible with any other amino acid. These residues are required for the stabilization of the folded hairpin of the βγ motif (
      • Wistow G.
      • Piatigorsky J.
      ). Between Gly-8 and Ser-34 lie two charge clusters of alternate signs (
      • Hemmingsen J.M.
      • Gernert K.M.
      • Richardson J.S.
      • Richardson D.C.
      ). Calcium homeostasis plays an important role in lens transparency, opacification, and cataractogenesis. DMP4 has at least one Greek key domain, and several similar domains repeat in the mouse DMP4 protein sequence. Conformational analysis indicates that the entire molecule adopts a flexible random structure, but in the presence of calcium, the Greek key domain could bind calcium ions resulting in ordered structures. Future studies would be directed to delete or mutate these domains and determine the function of DMP4 in the process of biomineralization.
      Expression analysis of DMP4 in mouse tissues demonstrates a wide expression pattern with high levels in the tooth. In situ hybridization results demonstrate that in a developing tooth, initial expression is seen in the odontoblasts at all stages of development and a restricted expression pattern in the ameloblasts at later stages of development. DMP4 was also present in the alveolar bone, implying that this protein might function in the assembly of mineralized matrices. Other matrix proteins such as DMP1 and DSPP (
      • Qin C.
      • Brunn J.C.
      • Cadena E.
      • Ridall A.
      • Butler W.T.
      ), which were initially thought to be dentin-specific, have also been identified in bone.
      Currently, the biological function of DMP4 is unknown. Similarity searches in the data base have identified another hypothetical protein FAM20 with sequence similarity to DMP4. Employing representational difference analysis, Nalbant et al. (
      • Nalbant D.
      • Youn H.
      • Nalbant S.I.
      • Sharma S.
      • Cobos E.
      • Beale E.G.
      • Du Y.
      • Williams S.C.
      ) identified several genes during experimentally induced myeloid differentiation in a mouse hematopoietic stem cell line. FAM20 family members were identified during this process. They described the identification of a clone derived from an uncharacterized putative secreted protein. This 541-amino acid protein is a member of family of predicted proteins that has been named FAM20 (family with sequence similarity 20) with three members (FAM20A, FAM20B, and FAM20C) in mammals. This family is highly conserved in human, mouse, rat, worm, and zebra fish (
      • Nalbant D.
      • Youn H.
      • Nalbant S.I.
      • Sharma S.
      • Cobos E.
      • Beale E.G.
      • Du Y.
      • Williams S.C.
      ), and however, different family members were identified in certain genomes. The expansion in gene number in different species suggests that the family has evolved as a result of several gene duplication events that have occurred in both vertebrates and invertebrates. Overall amino acid homology between different species is high (supplemental Fig. S2); for example, predicted human FAM20C and mouse DMP4 amino acid sequences are 69% identical (73% similar). Expression analysis revealed that the Fam20A gene was indeed differentially expressed during hematopoietic differentiation and that the other two family members (Fam20B and Fam20C) were also expressed during hematopoiesis. DMP4 on the other hand showed a single 3-kb band in the Northern blot in several of the examined tissues. Thus, isoforms of DMP4 were not detected by traditional Northern blot.
      In summary, DMP4 is a novel macromolecule found abundantly in the dentin matrix. High expression levels in the odontoblasts strongly suggest that DMP4 represents a novel soluble regulator of dentin mineralization. Gain of function results obtained from this study showed a significant up-regulation of dentin matrix genes responsible for dentin mineralization. Similarly, silencing DMP4 showed down-regulation in the mineralization cascade. However, the precise molecular function of DMP4 in dentinogenesis in vivo remains unclear. In future, the generation of knock-out mice in which DMP4 expression is completely attenuated would further characterize the molecular function of DMP4. Determining the function and activity of DMP4 during dentin mineralization would provide new avenues for dentin regeneration and therapeutic approaches.

      Acknowledgments

      We thank Verna J. Brown for histological staining, Sherry Lin for DNA sequencing, and Dr. A. Moneim Zaki for assistance with dark field microscopy.

      Supplementary Material

      References

        • Thesleff I.
        • Keranen S.
        • Jernvall J.
        Adv. Dent Res. 2001; 15: 14-18
        • Weiner S.
        • Wagner H.D.
        Annu. Rev. Mater. Sci. 1998; 28: 271-298
        • He G.
        • Dahl T.
        • Veis A.
        • George A.
        Nat. Mater. 2003; 2: 552-558
        • He G.
        • George A.
        J. Biol. Chem. 2004; 279: 11649-11656
        • George A.
        • Sabsay B.
        • Simonian P.A.
        • Veis A.
        J. Biol. Chem. 1993; 268: 12624-12630
        • George A.
        • Bannon L.
        • Sabsay B.
        • Dillon J.W.
        • Malone J.
        • Veis A.
        J. Biol. Chem. 1996; 271: 32869-32873
        • George A.
        • Srinivasan R.
        • Thotakura R.S.
        • Liu K.
        • Veis A.
        Connect. Tissue Res. 1999; 40: 49-57
        • MacDougall M.
        • Simmons D.
        • Luan X.
        • Nydegger J.
        • Feng J.
        • Gu T.T.
        J. Biol. Chem. 1997; 72: 835-842
        • Witkop Jr., C.J.
        • MacLean C.J.
        • Schmidt P.J.
        • Henry J.L.
        Ala. J. Med. Sci. 1966; 3: 382-403
        • Zhang X.
        • Zhao J.
        • Li C.
        • Gao S.
        • Qiu C.
        • Liu P.
        • Wu G.
        • Qiang B.
        • Lo W.H.
        • Shen Y.
        Nat. Genet. 2001; 27: 151-152
        • Xiao S.
        • Yu C.
        • Chou X.
        • Yuan W.
        • Wang Y.
        • Bu L.
        • Fu G.
        • Qian M.
        • Yang J.
        • Shi Y.
        • Hu L.
        • Han B.
        • Wang Z.
        • Huang W.
        • Liu J.
        • Chen Z.
        • Zhao G.
        • Kong X.
        Nat. Genet. 2001; 27: 201-204
        • MacDougall M.
        • Thiemann F.
        • Ta H.
        • Hsu P.
        • Chen L.S.
        • Snead M.L.
        Connect. Tissue Res. 1995; 22: 71-77
        • Hao J.
        • He G.
        • Narayanan K.
        • Zou B.
        • Lin L.
        • Muni T.
        • Ramachandran A.
        • George A.
        Bone. 2005; 37: 578-588
        • Hao J.
        • Narayanan K.
        • Ramachandran A.
        • He G.
        • Almushayt A.
        • Evans C.
        • George A.
        J. Biol. Chem. 2002; 277: 19976-19981
        • Hao J.
        • Zou B.
        • Narayanan K.
        • George A.
        Bone. 2004; 34: 921-932
        • Harada H.
        • Ichimori Y.
        • Yokohama-Tamaki T.
        • Ohshima H.
        • Kawano S.
        • Katsube K.
        • Wakisaka S.
        Biochem. Biophys. Res. Commun. 2006; 340: 611-616
        • Xu L.
        • Harada H.
        • Yokohama-Tamaki T.
        • Matsumoto S.
        • Tanaka J.
        • Taniguchi A.
        J. Biol. Chem. 2006; 281: 2257-2262
        • Harlow E.
        • Lane D.
        Antibodies: A Laboratory Manual.
        Cold Spring Harbor Laboratory, Cold Spring Harbor, NY1998: 421-470
        • Towbin H.
        • Staehelin T.
        • Gordon J.
        Proc. Natl. Acad. Sci. U. S. A. 1979; 76: 4350-4354
        • Srinivasan R.
        • Chen B.
        • Gorski J.P.
        • George A.
        Connect. Tissue Res. 1999; 40: 251-258
        • Rajini B.
        • Shridas P.
        • Sundari C.S.
        • Muralidhar D.
        • Chandani S.
        • Thomas F.
        • Sharma Y.
        J. Biol. Chem. 2001; 276: 38464-38471
        • Nalbant D.
        • Youn H.
        • Nalbant S.I.
        • Sharma S.
        • Cobos E.
        • Beale E.G.
        • Du Y.
        • Williams S.C.
        BMC Genomics. 2005; 6: 11
        • Blundell T.
        • Lindley P.
        • Miller L.
        • Moss D.
        • Slingsby C.
        • Tickle I.
        • Turnell B.
        • Wistow G.
        Nature. 1981; 289: 771-777
        • Wistow G.
        • Turnell B.
        • Summers L.
        • Slingsby C.
        • Moss D.
        • Miller L.
        • Lindley P.
        • Blundell T.
        J. Mol. Biol. 1983; 170: 175-202
        • Wistow G.
        • Piatigorsky J.
        Annu. Rev. Biochem. 1988; 57: 479-504
        • Hemmingsen J.M.
        • Gernert K.M.
        • Richardson J.S.
        • Richardson D.C.
        Protein Sci. 1994; 3: 1927-1937
        • Qin C.
        • Brunn J.C.
        • Cadena E.
        • Ridall A.
        • Butler W.T.
        Connect. Tissue Res. 2003; 44: 179-183