Advertisement

Endoplasmic Reticulum Chaperone Protein GRP-78 Mediates Endocytosis of Dentin Matrix Protein 1*

Open AccessPublished:August 28, 2008DOI:https://doi.org/10.1074/jbc.M800786200
      Dentin matrix protein 1 (DMP1), a phosphorylated protein present in the mineral phase of both vertebrates and invertebrates, is a key regulatory protein during biogenic formation of mineral deposits. Previously we showed that DMP1 is localized in the nuclear compartment of preosteoblasts and preodontoblasts. In the nucleus DMP1 might play an important role in the regulation of genes that control osteoblast or odontoblast differentiation. Here, we show that cellular uptake of DMP1 occurs through endocytosis. Interestingly, this process is initiated by DMP1 binding to the glucose-regulated protein-78 (GRP-78) localized on the plasma membrane of preodontoblast cells. Binding of DMP1 to GRP-78 receptor was determined to be specific and saturable with a binding dissociation constant KD = 85 nm. We further depict a road map for the endocytosed DMP1 and demonstrate that the internalization is mediated primarily by caveolae and that the vesicles containing DMP1 are routed to the nucleus along microtubules. Immunohistochemical analysis and binding studies performed with biotin-labeled DMP1 confirm spatial co-localization of DMP1 and GRP-78 in the preodontoblasts of a developing mouse molar. Co-localization of DMP1 with GRP-78 was also observed in T4-4 preodontoblast cells, dental pulp stem cells, and primary preodontoblasts. By small interfering RNA techniques, we demonstrate that the receptor for DMP1 is GRP-78. Therefore, binding of DMP1 with GRP-78 receptor might be an important mechanism by which DMP1 is internalized and transported to the nucleus during bone and tooth development.
      Acidic noncollagenous proteins play a pivotal role during biomineral formation. Dentin matrix protein 1 (DMP1)
      The abbreviations used are: DMP1, dentin matrix protein 1; DPSC, dental pulp stem cells; GST, glutathione S-transferase; TRITC, tetramethylrhodamine isothiocyanate; R-WGA, rhodamine-conjugated wheat germ agglutinin; NLS, nuclear localization signal; DAPI, 4,6-diamidino-2-phenylindole; BSA, bovine serum albumin; DMEM, Dulbecco's modified Eagle's medium; PBS, phosphate-buffered saline; HRP, horseradish peroxidase; shRNA, short hairpin RNA; MS/MS, tandem mass spectroscopy; FITC, fluorescein isothiocyanate.
      2The abbreviations used are: DMP1, dentin matrix protein 1; DPSC, dental pulp stem cells; GST, glutathione S-transferase; TRITC, tetramethylrhodamine isothiocyanate; R-WGA, rhodamine-conjugated wheat germ agglutinin; NLS, nuclear localization signal; DAPI, 4,6-diamidino-2-phenylindole; BSA, bovine serum albumin; DMEM, Dulbecco's modified Eagle's medium; PBS, phosphate-buffered saline; HRP, horseradish peroxidase; shRNA, short hairpin RNA; MS/MS, tandem mass spectroscopy; FITC, fluorescein isothiocyanate.
      was the first of the acidic proteins cloned from the dentin matrix. Although initially isolated from the dentin matrix and thought to be unique to dentin and named accordingly, DMP1 has now been found to be present in all mineralized tissues of the vertebrate system (
      • Narayanan K.
      • Srinivas R.
      • Ramachandran A.
      • Hao J.
      • Quinn B.
      • George A.
      ,
      • Narayanan K.
      • Gajjeraman S.
      • Ramachandran A.
      • Hao J.
      • George A.
      ).
      We previously showed that DMP1 may act as a transcriptional regulator as DMP1 is localized in the nucleus of preosteoblasts (
      • Narayanan K.
      • Ramachandran A.
      • Hao J.
      • He G.
      • Park K.W.
      • Cho M.
      • George A.
      ). In the nucleus DMP1 played a regulatory role in the regulation of specific genes that control osteoblast and odontoblast differentiation (
      • Narayanan K.
      • Ramachandran A.
      • Hao J.
      • He G.
      • Park K.W.
      • Cho M.
      • George A.
      ,
      • Narayanan K.
      • Gajjeraman S.
      • Ramachandran A.
      • Hao J.
      • George A.
      ). Specifically, we showed that DMP1 functions as a transcriptional regulator of the odontoblast specific gene dentin sialophosphoprotein (
      • Narayanan K.
      • Gajjeraman S.
      • Ramachandran A.
      • Hao J.
      • George A.
      ). The export of DMP1 from the nucleus during maturation of osteoblasts was found to be in response to a stimulus from calcium ions. In the extracellular matrix DMP1 can nucleate the formation and growth of hydroxyapatite (
      • He G.
      • Dahl T.
      • Veis A.
      • George A.
      ,
      • He G.
      • George A.
      ). Initial observations from the DMP1-deficient mouse showed no apparent skeletal or tooth phenotype during early development, suggesting that DMP1 function may be redundant. However, postnatal DMP1-null mice developed tooth formation defects, characterized by a partial failure of maturation from predentin to dentin (
      • Ye L.
      • MacDougall M.
      • Zhang S.
      • Xie Y.
      • Zhang J.
      • Li Z.
      • Lu Y.
      • Mishina Y.
      • Feng J.Q.
      ). Collectively, these data indicate that DMP1 may play a pivotal role in regulating mineralized matrix formation. In the present study we demonstrate the mechanism by which DMP1 is endocytosed and translocated to the nucleus of preodontoblasts and dental pulp stem cells (DPSC).
      Endocytosis is a complex and diverse process and is usually classified into two distinct pathways, namely clathrin-dependent and clathrin-independent pathways. The clathrin-mediated pathway targets its cargo to early endosomes and is an important pathway for down-regulation of receptors (
      • Katzmann D.J.
      • Babst M.
      • Emr S.D.
      ). The clathrin-independent endocytosis is further divided into dynamin-dependent and dynamin-independent pathways. However, the more characterized of the clathrin-independent pathways is the lipid-raft and the caveolar pathway. Lipid rafts and caveolae are involved in the internalization and trafficking of cholesterol, numerous receptors and signaling molecules and also toxins and viruses (
      • Anderson R.G.
      • Kamen B.A.
      • Rothberg K.G.
      • Lacey S.W.
      ,
      • Pelkmans L.
      • Kartenbeck J.
      • Helenius A.
      ,
      • Nichols B.J.
      ). Caveolae are 50–80-nm plasma membrane invaginations that contain caveolin-1 as their major constituent. Therefore, caveolin-1 serves as a marker for caveolae-mediated endocytosis.
      In this study we show that internalization of DMP1 is a caveolae-associated, receptor-mediated event. Three major steps highlighted in this study are as follows; GRP-78 functions as a cell surface receptor in DMP1 endocytosis, caveolae mediate the intracellular transport of DMP1, and DMP1-containing vesicles are transported intracellularly to the Golgi and the nucleus via the microtubules.

      MATERIALS AND METHODS

      Cell Culture—Three different cell types were used in this study, namely, the rat T4-4 preodontoblast cell line (
      • Hao K.
      • Narayanan A.
      • Ramachandran G.
      • He A.
      • Almushayt Evans C.
      • George A.
      ), the human DPSC (
      • Gronthos S.
      • Mankani M.
      • Brahim J.
      • Robey P.G.
      • Shi S.
      ,
      • Batouli S.
      • Miura M.
      • Brahim J.
      • Tsutsui T.W.
      • Fisher L.W.
      • Gronthos S.
      • Robey P.G.
      • Shi S.
      ), and the primary preodontoblasts obtained from mouse day 3 post-natal pups. The T4-4 cell line has been previously characterized by our laboratory and exhibits odontogenic potential both in vitro and in vivo. The DPSC are a gift from Dr. Songtao Shi (University of Southern California) and were isolated from the perivascular niche of human dental pulp (
      • Gronthos S.
      • Mankani M.
      • Brahim J.
      • Robey P.G.
      • Shi S.
      ). Both the cell types were grown in DMEM/F-12 (Cellgro) containing 10% fetal bovine serum (Cellgro) unless otherwise specified with 1% antibiotic-antimycotics. The primary preodontoblasts were obtained from the developing molar tooth germ and were grown in DMEM/F-12 medium with 20% fetal bovine serum.
      Expression and Purification of DMP1 and the N (N-DMP1)- and C (C-DMP1)-terminal Fragments of DMP1—The recombinant DMP1 protein without any post-translational modifications was expressed and purified as published earlier (
      • Srinivasan R.
      • Chen B.
      • Gorski J.P.
      • George A.
      ). The N (amino acids 1–334)- and C (amino acids 335–489)-terminal fragments of DMP1 were amplified from the full-length DMP1 cDNA, cloned into pGEX-4T-3 (Invitrogen), and expressed as glutathione S-transferase (GST) fusion proteins in BL21-DE3 cells. The fusion proteins were then purified by binding to a glutathione-Sepharose column, cleaved using thrombin to remove the fusion tag, and eluted through a benzamidine column to remove thrombin and obtain the respective purified fragments. The recombinant full-length form of DMP1, the N-terminal fragment (N-DMP1), and the C-terminal fragment (C-DMP1) were fluorescently labeled with fluorescein isothiocyanate (FITC) according to published protocols (
      • Narayanan K.
      • Gajjeraman S.
      • Ramachandran A.
      • Hao J.
      • George A.
      ).
      Binding of DMP1 to T4-4 Cells—Recombinant DMP1 was biotinylated using the Sulfo-NHS biotinylation kit (Pierce) as per the manufacturer's instructions. T4-4 cells were plated onto 96-well flat-bottomed tissue culture plates at a density of 20,000 cells/well and allowed to attach for about 16–20 h. These experiments were performed at 4 °C, which did not permit endocytosis. Before the start of the binding assay, the plate was cooled to 4 °C for 30 min, and then the cells were washed 3× with ice-cold phosphate-buffered saline (PBS) (Invitrogen) and once with ice cold growth medium (DMEM/F-12) without serum. Cells were then incubated for 1 h at 4 °C in serum-free DMEM/F12 containing increasing concentrations of biotinylated DMP1 (0, 0.2, 0.4, 0.5, 0.8, 1.0, 1.5, 2, 3, 4, 5 μg/100 μl). After binding, the cells were gently rinsed 3× with ice-cold PBS and incubated with streptavidin-conjugated with horseradish peroxidase (HRP) (0.1 μg/ml) for 1 h at 4 °C. After rinsing the cells 3× in ice-cold PBS, 100 μl of an HRP substrate (1-step Turbo TMB-ELISA from Pierce) was added to the wells and incubated for 20 min. The reaction was then stopped by the addition of 1 m sulfuric acid, and the plate was read at 450 nm in a Bio-Tek microtiter plate reader. The absorbance reading after subtraction of background (biotinylated DMP1 in the absence of cells) was then used to determine the amount of DMP1 bound to the cell surface. A standard curve was constructed from known concentrations of DMP1 bound to the cell surface. Results were analyzed using the Sigma Plot software to derive binding curves.
      Endocytosis of DMP1—Endocytosis of DMP1 was analyzed by two approaches. In the first method T4-4 cells were plated at a density of 20,000 cells/well in 96-well flat-bottomed tissue culture plates. Cells were incubated with biotinylated DMP1 at varying concentrations as mentioned previously in serum-free DMEM/F-12 and incubated at 37 °C for a period of 30 min. They were then washed extensively in PBS and fixed with 4% paraformaldehyde, permeabilized with PBS containing 0.5% Triton X-100, and incubated with streptavidin-HRP conjugate (0.1 μg/ml) for 30 min at room temperature. The cells were then washed extensively with PBS and then incubated with HRP substrate. Absorbance was measured at 450 nm using a BioTek microtiter plate reader. The concentration of DMP1 endocytosed at saturation was estimated from the standard curve.
      In the second method T4-4 cells, DPSC, or primary odontoblasts were grown on coverslips placed in 6-well tissue culture dishes in DMEM/F-12 containing 10% fetal bovine serum for at least 24 h. Before the start of the experiment the cells were washed with PBS and then incubated with 10 μg/ml FITC-labeled DMP1 at 37 °C for 30 min. Protocol for FITC labeling was as published earlier (
      • Narayanan K.
      • Gajjeraman S.
      • Ramachandran A.
      • Hao J.
      • George A.
      ). For the time course experiments, T4-4 cells were incubated with FITC-DMP1 for 60 min at 4 °C before incubation for the indicated periods of time at 37 °C. The cells were washed extensively, fixed, permeabilized, and mounted on glass slides using mounting media with or without DAPI (Vector Laboratories). Fluorescence in the cells was observed by confocal laser scanning microscopy. Fluorescence quantification was performed for experiments involving 30- and 60-min incubations with T4-4 cells to obtain nuclear and cytoplasmic distribution of the endocytosed FITC-DMP1. Briefly, the mean fluorescence from an outlined area (nuclear or cytoplasmic) of the confocal images was multiplied with the area under consideration. This yielded the total fluorescence from the two regions. These data were used to estimate the percentage of nuclear and cytoplasmic pools of DMP1. Uptake experiments were also performed with FITC-N-DMP1 and FITC-C-DMP1 at the 45-min time point. FITC-bovine serum albumin (BSA) was used as a negative control.
      Endocytosis experiments were also performed in the presence of various endocytic inhibitors. T4-4 cells were pretreated with the specified concentrations of the inhibitors for 60 min at 37 °C before treatment with FITC-DMP1. The different inhibitors used were hyperosmotic sucrose (0.45 m) (
      • Inal J.
      • Miot S.
      • Schifferli J.A.
      ), Ikg (2 μm) (
      • Luo T.
      • Fredericksen B.L.
      • Hasumi K.
      • Endo A.
      • Garcia J.V.
      ), methyl-β-cyclodextrin (1%) (
      • Schousboe I.
      • Thomsen P.
      • van Deurs B.
      ), and colchicine (20 μm) (
      • Baron D.A.
      • Burch R.M.
      • Halushka P.V.
      • Spicer S.S.
      ) according to published concentrations. Endocytosis was also performed after pretreatments either with anti-rabbit GRP-78 antibody (1:100), 2 mm RGD (Arg-Gly-Asp) peptide, or CD44 antibody (1/100).
      Immunostaining and Confocal Microscopy—For immunostaining, T4-4 cells were treated with various reagents as mentioned above, fixed, permeabilized, blocked with 5% BSA in PBS for 1 h at room temperature, and incubated with the indicated concentration of the primary antibody overnight at 4 °C. After washing extensively with PBS, the cells were incubated for 1 h at room temperature with the appropriate secondary antibody. All secondary antibody incubations were performed in the dark to minimize loss of fluorescent signal due to photo bleaching. The cells were then washed extensively in PBS and mounted on glass slides. The antibodies used were anti-α-tubulin (1:5000) (Sigma), anti-caveolin1 (1:75) (Santa Cruz), anti-clathrin (1:50) (Santa Cruz), and anti-GRP-78 serum (1:100). Golgi staining was performed after permeabilization by incubating the cells with R-WGA (rhodamine-conjugated wheat germ agglutinin, 10 μg/ml) (Vector Laboratories) in PBS for 60 min at room temperature in the dark. Actin was stained after permeabilization by incubating the cells with phalloidin-TRITC (Molecular Probes) for 15 min at room temperature.
      For confocal microscopy the images were acquired using a Zeiss LSM 510 or 510 META confocal microscope equipped with a 63× water immersion objective. 100× oil immersion objective was only used for FITC-DMP1 and GRP-78 z-stack images. All the confocal images are representatives of experiments conducted in triplicates. The imaging parameters were maintained constant whenever intensity of fluorescence was to be compared. Three-dimensional reconstruction of the z-stack images was performed using the Zeiss LSM imaging software.
      Isolation of DMP1 Plasma Membrane Binding Partner—Recombinant GST-DMP1 was bound to GST-Sepharose beads and used as a bait column. Membrane protein lysate from T4-4 cells were prepared using the MEM-PER membrane protein extraction kit (Pierce) as per the manufacturer's protocol. The bait column was then incubated with the membrane proteins isolated from T4-4 cells at 4 °C overnight. The column was then washed extensively with PBS containing 150 mm NaCl and eluted using PBS containing 1 m NaCl. The eluted proteins were extensively dialyzed against 1 mm Tris-HCl, pH 7.4, lyophilized, and dissolved in 100 μl of PBS. 40 μl of the protein sample was subjected to SDS-PAGE analysis. Additionally, a T-75 flask of T4-4 cells was treated with recombinant GST-DMP1 (10 μg/ml) for 1 h at 4 °C to ensure binding but not endocytosis. The cells were washed with ice-cold PBS three times after which the membrane proteins were isolated as described previously. The isolated membrane proteins were incubated with glutathione-Sepharose beads overnight at 4 °C. The beads were then washed extensively in PBS containing 150 mm NaCl. The bound proteins were eluted with glutathione, extensively dialyzed in 1 mm Tris-HCl, pH 7.4, lyophilized, and redissolved in 50 μl of PBS. The entire sample was then subjected to SDS-PAGE and transferred to nitrocellulose. Finally, immunoblotting was performed with anti-GRP-78.
      Protein Sequencing Using Mass Spectrometry—SDS-PAGE analysis identified 4 protein bands at ∼72, 37, 33, and 30 kDa. Each of these bands were excised from the gel and subjected to in-gel proteolytic digestion with trypsin. The recovered peptides were separated and analyzed using an LTQ mass spectrometer to obtain MS/MS spectra at the Proteomics facility at University of Illinois at Chicago. The sequences of the peptides were inferred by matching the MS/MS spectra to protein sequence data base using Sequest and Mascot data base search engines.
      GRP-78 shRNA Knockdown—Two target sequences of GRP-78 gene were used. One sequence that was effective in silencing GRP-78 was used in this experiment. This sequence targeted rat GRP-78 at the 3′-untranslated region. T4-4 cells were plated onto 6-well tissue culture plates at 80% confluence and allowed to attach for 24 h. They were then transiently transfected with the GRP-78 shRNA plasmid pLKO.1 (Sigma) containing the sequence CCGGCTCGAATGTAATTGGAATCTTCTCGAGAAGATTCCAATTACATTCGAGTTTTT using Superfect transfection reagent (Qiagen) as per the manufacturer's protocol. 24 h post-transfection, the cells were incubated with 10 μg/ml FITC-DMP1 for 10 min. The cells were fixed, permeabilized, and immunostained with GRP-78 antibody as described previously. The cells were imaged using a Zeiss Axio Observer-D1 fluorescence microscope equipped with the appropriate filter sets to view green and red fluorescence and Axiovision imaging software.
      GRP-78 and DMP1 Immunohistochemistry—5-Day post-natal mice heads were sectioned along the midline and embedded in paraffin. Sections were cut at a thickness of 5 μm, with each slide containing 2 sections. The sections were deparaffinized with xylene and hydrated through graded ethanol. They were then incubated in 3% H2O2 for 30 min to quench endogenous peroxidase activity and blocked with serum in PBS (Vecta stain ABC peroxidase kit) for 1 h at room temperature. Sections were then incubated overnight at 4 °C with either PBS containing 30 μg/ml of biotinylated DMP1, FITC-DMP1, anti-DMP1 antibody (1/200), or anti-GRP-78 antibody (1/100) in PBS containing 3% BSA. Slides were then washed 4× in PBS (15 min each wash). Sections that were incubated with biotinylated DMP1 were incubated with peroxidase-conjugated streptavidin (Vectorstain ABC peroxidase kit) and developed with the DAB kit (Vector Laboratories). Sections incubated with FITC-DMP1 were mounted with fluorescence mounting solution (Vector Laboratories) and imaged using a Zeiss Axio Observer-D1 fluorescence microscope equipped with the appropriate filter sets to view green fluorescence and Axiovision imaging software. The sections that were incubated with the primary antibodies were incubated with their biotin-conjugated secondary antibodies, washed 4× in PBS, incubated with peroxidase-conjugated streptavidin (Vecta stain ABC peroxidase kit), and developed with the DAB kit (Vector Laboratories).
      Preparation of Detergent-resistant Membrane Fractions—T4-4 cells were incubated briefly with recombinant DMP1. The cells were then lysed, and the detergent-resistant membrane fractions were prepared as described by Macdonald et al. (
      • Macdonald J.L.
      • Pike L.J.
      ). 50 μl of each fraction was subjected to SDS-PAGE, and immunoblotting was performed with anti-caveolin-1 and anti-DMP1 antibodies.
      Immunoblot Analysis—Proteins were separated by SDS-PAGE, transferred to a nitrocellulose membrane, and blocked with PBS containing 3% BSA. The membrane was then incubated overnight with rabbit anti-GRP-78 antibody (1:1000), anti-caveolin-1 antibody (1/750), or anti-DMP1 antibody (1/500) at 4 °C. The membrane was then washed 3× for 15 min with PBS-Tween at room temperature and then incubated with anti-rabbit HRP-conjugated secondary antibody for GRP-78 blots and with anti-rabbit alkaline phosphatase-conjugated secondary antibody for DMP1 and caveolin-1 blots. The membrane was then washed 3× with PBS-Tween and then developed using a chemiluminescent substrate (Pierce) for HRP or with an Bio-Rad alkaline phosphatase developer kit.

      RESULTS

      Binding of Biotinylated DMP1 to T4-4 Cells—In this experiment we show that there is specific binding of DMP1 to the cell surface. We performed binding experiments using biotinylated DMP1 on T4-4 cells. These experiments were conducted at 4 °C, which did not permit internalization. Fig. 1a shows the inverse Langmuir isotherm plot (rectangular hyperbola) fit to the data from the binding experiment, and Fig. 1b shows the corresponding Scatchard plot. These results demonstrate that the binding was saturable, and the dissociation constant was estimated to be 85 nm, indicating a tight binding of DMP1 to the cell surface with 9 × 106 molecules of biotinylated DMP1 bound per cell (estimated using Avogadro constant).
      Figure thumbnail gr1
      FIGURE 1Endocytosis of DMP1. a, binding curve depicting specific binding of DMP1 to the surface of T4-4 cells. The x axis represents the total ligand concentration, and the y axis represents the amount of DMP1 bound. The data points were fit to an inverse Langmuir isotherm, and the dissociation constant was estimated to be 85 nm. b, Scatchard plot for the data shown in a. c, endocytosis of biotinylated DMP1 after incubation with the same at 37 °C for 30 min. Note the saturation at 0.217 μm DMP1. The error bars indicate the mean ± S.E. and (n ≥ 6). d, negative control showing the absence of endocytosis of FITC-labeled BSA. Scale bar, 50 μm. e, f, and g, confocal images showing endocytosed FITC-DMP1 in T4-4 preodontoblasts. e, scale bar, 10 μm; endocytosis in DPSC. f, scale bar, 20 μm; endocytosis in primary odontoblasts. g, scale bar, 20 μm; after 30 min of incubation at 37 °C. Note the intracellular vesicles showing the presence of vesicular transport. h, time course confocal images of FITC-DMP1 after incubation at 4 °C for an hour and internalization observed at 37 °C for 2, 5, 10, 15, and 30 min. Scale bar, 5 μm.
      Endocytosis of DMP1—To demonstrate the cellular uptake of DMP1, biotinylated DMP1 was added exogenously in increasing concentrations to T4-4 preodontoblast cells and incubated for 30 min at 37 °C. Quantification of biotin uptake indicated that DMP1 was internalized in a concentration-dependent manner; Fig. 1c. We estimated that at saturation 16.1 nm was endocytosed by 20,000 T4-4 cells. To further validate the internalization process, T4-4 cells, DPSC, and primary odontoblasts were treated for 30 min with FITC-DMP1. Analysis by confocal microscopy showed that DMP1 was endocytosed and transported to the perinuclear region (Fig. 1, e, f, and g). However, FITC labeled BSA, which served as the control was not internalized (Fig. 1d).
      Fig. 1h shows the time course of FITC-DMP1 uptake in T4-4 cells. The intracellular transport of DMP1 was seen in vesicles and localized around the perinuclear region by 15 min and localized in the nucleus by 30 min.
      Additionally, endocytosis was also performed after pretreatment of both DPSC and T4-4 cells with 2 mm RGD polypeptide to block the integrins and also with anti-CD44 (1/100) antibody (Santa Cruz) for 1 h at 37 °Cas αvβ3 integrin, αvβ5 integrin, and CD44 have been previously suggested to be putative DMP1 cell surface binding partners. The endocytosis of FITC-DMP1 was unaffected by these treatments (see supplemental Fig. 1).
      Identification of GRP-78 as a Cell Surface Receptor—The results mentioned above indicate that DMP1 might be endocytosed through a receptor-mediated mechanism. To identify a receptor, the total membrane protein lysate from T4-4 cells was incubated with a GST-DMP1 bait column. Eluates from the column were separated on an SDS-PAGE gel. Fig. 2a shows four bands at ∼75, 40, 33, and 30 kDa. Each of these bands was subjected to trypsin digestion and MS/MS analysis. The 33- and 30-kDa bands were identified as tropomyosin fragments. The 40-kDa band was identified as actin, and the 75-kDa band was identified as GRP-78. Among these proteins, GRP-78 seemed a likely candidate for being a possible DMP1 binding partner. Although GRP-78 is one of the endoplasmic reticulum chaperone proteins, it is also present at the cell surface of many cell types (
      • Delpino A.
      • Castelli M.
      ,
      • Altmeyer A.
      • Maki R.G.
      • Feldweg A.M.
      • Heike M.
      • Protopopov V.P.
      • Masur S.K.
      • Srivastava P.K.
      ,
      • Weist D.I.
      • Bhandoola A.
      • Punt J.
      • Kreibich G.
      • McKean D.
      • Singer A.
      ,
      • Xiao G.
      • Chung T.-F.
      • Pyun H.Y.
      • Fine R.E.
      • Johnson R.J.
      ,
      • McMichael B.K.
      • Kotadiya P.S.
      • Holliday T.
      • Shannon L.
      • Lee B.S.
      ). Immunoblot analysis performed with anti-GRP-78 antibody on the eluate from GST-DMP1 bait column confirmed the membrane protein to be GRP-78 (Fig. 2c). Tropomyosin and actin that were also identified have been reported to be co-localized on the plasma membrane of osteoclasts (240). To further confirm that DMP1 binds specifically to GRP-78, the membrane fraction of T4-4 cells incubated with GST-DMP1 at 4 °C and subsequently purified on a GST column was subjected to immunoblotting with GRP-78 antibody. Results confirmed that DMP1 bound specifically to GRP-78 localized on the plasma membrane of T4-4 cells (Fig. 2d). Together, these observations indicate that GRP-78 is the cell surface DMP1-binding protein.
      Figure thumbnail gr2
      FIGURE 2Identification of a receptor mediating DMP1 endocytosis. a, eluate from T4-4 membrane protein lysate was passed through a glutathione column bound with GST-DMP1 fusion protein. Four polypeptides of Mr = 75, 50, 37, and 25 were obtained and sequenced. b, peptides identified with greater than 95% accuracy by Mascot (italics), Sequest (normal), or both (bold) search engines fed with the raw MS/MS data from the mass spectrometer present in the boxed protein band from a subjected to in-gel trypsin digestion identifying the DMP1-binding protein to be GRP-78. c, immunoblot of the eluate from the GST-DMP1 bait column with anti GRP-78 antibody, thus confirming the DMP1-binding protein to be GRP-78. d, immunoblot of membrane proteins isolated from T4-4 cells incubated with GST-DMP1 passed through a GST column with GRP-78 antibody to confirm GRP-78 as a DMP1-binding protein. e, confocal images of T4-4 cells treated with anti GRP-78 antibody showing cytoplasmic as well as plasma membrane localization of GRP-78 (white arrows point to membranous staining). Scale bar, 10 μm. f, g, and h, confocal images colocalizing FITC-DMP1 with GRP-78, respectively, in T4-4 (f), primary odontoblasts (g), and DPSC (h), respectively. White arrows indicate sites of DMP1 and GRP-78 colocalization. i, j, and k, confocal images representing endocytosis of FITC-DMP1 in the absence of anti-GRP78 antibody (i) and pretreated with anti GRP-78 antibody (j) and control IgG serum (k). Scale bar, 20 μm. l, endocytosis of FITC-DMP1 in T4-4 cells in the presence of GRP-78 small interfering RNA transfection. Arrowheads point to untransfected cells where the intensity of endocytosed FITC-DMP1 is high, and arrows point to transfected cells expressing lower amounts of GRP-78 leading to lower levels of DMP1 endocytosis. m, quantification of fluorescence data for T4-4 cells transfected with GRP-78 small interfering RNA and subsequently subjected to FITC-DMP1 incubation for 10 min at 37 °C. Error bars indicate mean ± S.E. (n ≥ = 40).
      Localization of GRP-78 on the plasma membrane of T4-4 cells was examined by indirect immunofluorescence using anti-GRP-78 antibody. Fig. 2e demonstrates that indeed a subpopulation of GRP-78 was localized on the plasma membrane of T4-4 cells. The co-localization of GRP-78 with the endocytosed FITC-DMP1 was also examined in T4-4 cells, primary odontoblasts, and DPSC by confocal microscopy. Results in Fig. 2, f–h, show the co-localization of GRP-78 with FITC-DMP1 after 10 min of exogenous addition of each, respectively. Also extensive co-localization of DMP1 and GRP-78 was observed near the plasma membrane, indicating probable interaction between DMP1 and GRP-78 on the plasma membrane.
      To confirm that GRP-78 is a bona fide receptor responsible for mediating endocytosis of DMP1, cells were treated with anti-GRP-78 antibody and assessed for DMP1 internalization. Results in Fig. 2, i and j, show that DMP1 internalization was abrogated with GRP-78 antibody confirming GRP-78-mediated endocytosis of DMP1. The image in Fig. 2i was acquired with maximum laser intensity. This intensity was maintained constant when imaging Fig. 2j. Negative control with IgG alone did not inhibit FITC-DMP1 endocytosis (Fig. 2k). Additionally, T4-4 cells were transiently transfected with GRP-78 shRNA. 24 h post-transfection, the cells were incubated for 10 min with FITC-DMP1 and then immunostained with GRP-78 antibody. Fig. 2l shows a representative image from this experiment. The arrows show a few cells that have been transfected with the shRNA plasmid, and the arrowheads show untransfected cells. The transfected cells show significantly weaker staining for GRP-78 when compared with the neighboring untransfected cells. Correspondingly, the amount of internalized FITC-DMP1 was less in the GRP-78 knockdown cells. Fig. 2m is a quantification of the fluorescence data that shows the effect of GRP-78 knockdown and the subsequent intensity of internalized FITC-DMP1. Results show that an ∼60% loss of GRP-78 expression translates into a 4 (∼75%)-fold decrease in DMP1 internalization. Overall, these results confirm that endocytosis of DMP1 is mediated by GRP-78.
      Furthermore, confocal analysis was performed to confirm the interaction between the endocytosed DMP1 and GRP-78 in three dimensions. Fig. 3a is a confocal image similar to that in Fig. 2f. The image depicts DMP1 co-localizing with GRP-78 both on the cell membrane as well as intracellularly in T4-4 cells. The white boxed area in Fig. 3a was subjected to z-stack confocal imaging (3 μm with an interval of 0.3 μm between each slice) and reconstituted. Fig. 3b is a snapshot of the reconstituted image, showing the close proximity between DMP1 and GRP-78. Fig. 3, c, d, and e, show three-dimensional snapshots of intracellular endocytic vesicles containing FITC-DMP1, GRP78, and both, respectively. This projection depicts the close association of DMP1 with GRP-78. Fig. 3f is a series of z-stack images of a T4-4 cell showing co-localization of FITC-DMP1 and GRP-78. Note the extensive co-localization of the two proteins. Fig. 3, g and h, are orthogonal projections of the images in Fig. 3f. The boxed regions in the figures show membrane co-localization of FITC-DMP1 and GRP-78 in the x-z and y-z planes, respectively.
      Figure thumbnail gr3
      FIGURE 3Three-dimensional confocal analysis of the interaction between DMP1 and GRP-78. a, confocal micrograph showing co-localization of FITC-DMP1 and GRP-78. b, a three-dimensional reconstruction of z-stack images of the region enclosed by the white box in a. Interaction between GRP-78 and FITC-DMP1 can be clearly seen from this figure. c, d, and e are three-dimensional reconstructions of z-stack images of endocytic vesicles showing FITC-DMP1 (c), GRP78 (d), and co-localization of both (e). f, z-stack images of a cell representing co-localization of FITC-DMP1 and GRP-78 in all three dimensions. g and h represent orthogonal views of the images shown in f. The boxed regions show co-localization near the plasma membrane in x-z and y-z planes, respectively.
      Because DMP1 is a multidomain protein containing hydroxyapatite nucleation sites, collagen, and integrin binding domains (
      • Narayanan K.
      • Ramachandran A.
      • Hao J.
      • He G.
      • Park K.W.
      • Cho M.
      • George A.
      ,
      • George A.
      • Sabsay B.
      • Simonian P.A.L.
      • Veis A.
      ,
      • George A.
      • Gui J.
      • Jenkins N.A.
      • Gilbert D.J.
      • Copeland N.G.
      • Veis A.
      ,
      • MacDougall M.
      • Gu T.T.
      • Luan X.
      • Simmons D.
      • Chen J.
      ), we next determined if any of the previously identified functional domains contributed to its ability to endocytose. To verify this we created two fragments of DMP1. The C-terminal fragment contained the nuclear localization signal (NLS), RGD, and the calcium binding domains, and the N terminus contained the rest of the protein (Fig. 4a). We hypothesized that if the C-terminal polypeptide was endocytosed, it might result from one of the functional domains, probably the RGD peptide sequence. Therefore, FITC-tagged N- and C-DMP1 were generated (Fig. 4b), and they were tested for their ability to be internalized. Results demonstrate that the N-terminal polypeptide was internalized and localized in the cytoplasm. However, no nuclear localization of FITC-N-DMP1 fragment was observed (Fig. 4c). Results in Fig. 4d show no internalization with FITC-C-DMP1. Similar cellular localization of FITC-N-DMP1 and GRP-78 (Fig. 4d) further confirmed the presence of GRP-78 binding domain within the N-terminal fragment of DMP1. Overall, these data indicate the specific interaction of the N-terminal DMP1 domain with GRP-78 leading to DMP1 endocytosis, whereas the C-terminal domain of DMP1 was not involved in the endocytic process.
      Figure thumbnail gr4
      FIGURE 4Identification of GRP-78 binding domain in DMP1. a, recombinant N- and C-terminal constructs of DMP1 were synthesized as stated under “Materials and Methods.” RGD, NLS, and calcium binding domains have been shown to be present in the C-terminal domain of the construct. b, SDS-PAGE of FITC-labeled N and C DMP1 fragments. M denotes molecular mass markers. c and d are confocal images showing the endocytosis of FITC-N-DMP1 and FITC-C-DMP1, respectively, after 45 min of incubation at 37 °C. Scale bar, 5 and 20 μm, respectively. Note the absence of endocytosis with FITC-C-DMP1. e, confocal image co-localizing FITC-N-DMP1 with GRP78 respectively. White arrows represent regions of colocalization of FITC-N-DMP1 and GRP-78. Note FITC-N-DMP1 colocalize with GRP-78 extensively near the plasma membrane.
      DMP1 Endocytosis Is Cholesterol-sensitive—To further elucidate the characteristics of the endocytic pathway, the effects of inhibitors for clathrin and caveolae-mediated endocytosis were examined. Use of hyperosmotic sucrose (0.45 m sucrose) and 2 μm ikarugamycin, common inhibitors for clathrin-mediated endocytosis, did not have any effect (Fig. 5, b and c). However, pretreatment of the cells with of 1% methyl β-cyclodextrin (MBCD), generally used to disrupt the caveolar integrity by depleting the cholesterol across the plasma membrane, significantly inhibited FITC-DMP1 uptake (Fig. 5d) when compared with its absence in Fig. 5a. Fig. 5e is a quantification of the fluorescence data. Together these results indicate that FITC-DMP1 uptake by T4-4 cells could possibly be mediated via the caveolae. Immunostaining with anti-clathrin antibody after treatment with FITC-DMP1 did not show any co-localization with DMP1 and further confirmed the absence of endocytosis via clathrin-coated pits (Fig. 5f).
      Figure thumbnail gr5
      FIGURE 5DMP1 endocytosis is cholesterol-sensitive. Confocal images of FITC-DMP1 endocytosis in the presence of inhibitors. Shown is the control without any inhibitors (a, scale bar, 10 μm), 0.45 m hyperosmotic sucrose (b, scale bar, 5 μm), 5 μm ikarugamycin (c, scale bar, 5 μm), 1% methyl-beta cyclo dextrin (MBCD) (d, scale bar, 10 μm). Note the intense decrease in fluorescence only when treated with MBCD indicating sensitivity of endocytosis to cholesterol depletion. e, quantification of fluorescence data. Data represent the mean ± S.E. of at least 20 cells counted in duplicate experiments. f, immunostaining of FITC-DMP1 treated cells with clathrin antibody showing no co-localization between the endocytosed DMP1 and clathrin. Arrows point to clathrin staining on the plasma membrane.
      Role of Caveolae in the Internalization of DMP1—To confirm the involvement of caveolae in the endocytosis of DMP1, uptake experiments were performed with FITC-DMP1, and the cells were immunostained with anti-caveolin-1 antibody (Fig. 6, b–d). Analysis by confocal microscopy demonstrated that internalized DMP1 vesicles exhibited co-localization with caveolin-1. Fig. 6a is a confocal image of control T4-4 cells stained with anti-caveolin-1 antibody showing that caveolin-1 was primarily localized on the plasma membrane of T4-4 cells (arrow); however, when cells were treated with FITC-DMP1, caveolin-1 was seen in the cytosol, co-localizing with DMP1 in T4-4 (Fig. 6, b–d) and DPSC (Fig. 6, e and f). Fig. 6, c and d, are enlarged images of the boxed regions of Fig. 6b, respectively. The white arrows in Fig. 6d indicate membrane co-localization of FITC-DMP1 and caveolin-1. As further confirmation of the involvement of caveolae, immunoblot analysis was performed on the proteins isolated from the detergent-resistant membrane fractions of T4-4 cells treated with recombinant DMP1. Fig. 6g shows the result of this experiment. As can be seen from the figure, both caveolin-1 and DMP1 were present in the same detergent-resistant membrane fractions. Together, these observations suggest that endocytosis of DMP1 could be primarily a caveolae-mediated process.
      Figure thumbnail gr6
      FIGURE 6DMP1 is endocytosed via the caveolae. Confocal images showing co-localization of FITC-DMP1 with caveolin-1. a, control showing T44 cells stained with anti-caveolin-1 antibody. The arrow indicates membranous caveolin-1 staining. b, c, and d represent co-localization of FITC-DMP1 with caveolin-1 in T4-4 cells. c and d are enlarged images of the boxed regions in b. Arrows point to co-localization of DMP1 and caveolin-1 on the plasma membrane. Also note the extensive co-localization of intracellular vesicles. e and f show FITC-DMP1 and caveolin-1 co-localization in DPSC. White arrows point to co-localizing intracellular vesicles. g, immunoblot analysis of detergent-resistant membrane fractions of T4-4 cells incubated with DMP1.
      Actin and Tubulin Facilitate Internalization and Transport of DMP1—Next we examined the mode of transport of DMP1 containing endocytic vesicles. Endocytic vesicles are normally transported either through microfilaments or via the microtubules. Microfilaments are used for short range transport, and the microtubules are used for long range transport (
      • Soldati T.
      • Schliwa M.
      ). Results in Fig. 7, a and c, show that the vesicles containing FITC-DMP1 move along the microtubules in the cytoplasm but require actin on the plasma membrane to be internalized. Fig. 7, d and e, are enlarged images of the boxed areas in Fig. 7, c and a, respectively. These images show co-localization of FITC-DMP1 with actin near the plasma membrane and the vesicles moving along the microtubules intracellularly. To confirm microtubule-dependent transport process, we performed the endocytosis assay after disruption of the microtubules with 20 μm colchicine. Results in Fig. 7b demonstrate the disruption in the trafficking process resulting in the accumulation of a considerable amount of FITC-DMP1 along the inner boundary of the plasma membrane. Another interesting feature that was observed repeatedly and illustrated in Fig. 7, a and e, is the formation of “tubulin taps,” which indicate the realignment of the microtubules to the sites of the endocytic vesicles to facilitate transportation of the cargo. These results indicate that microfilaments and microtubules facilitate the endocytosis and transport of DMP1.
      Figure thumbnail gr7
      FIGURE 7Endocytic vesicles are transported via the microtubules. a, confocal image showing endocytosed FITC-DMP1 vesicles on the microtubules in the absence of colchicine. Note the tubulin taps showing realigned microtubules near the plasma membrane to facilitate intracellular transport. b, the same experiment as a after treatment with of 20 μm colchicine, disrupting the cytoskeleton at 37 °C for 60 min. Scale bar, 10 μm. Note the membrane-bound intracellular FITC-DMP1 accumulated on the cytosolic side of the plasma membrane due to lack of transport and the possible route taken by the intracellular FITC-DMP1 (arrowhead). c, confocal image showing the endocytosed FITC-DMP1 on actin filaments near the plasma membrane. Note the absence of intracellular colocalization. d, enlarged image of the boxed area in c. e, enlarged image of the boxed area in a.
      Presence of Endocytosed DMP1 in the Golgi—The Golgi apparatus has been reported to be one of the primary destinations of vesicles endocytosed via the caveolar endocytic pathway (
      • Huang B.
      • Maciejewska I.
      • Sun Y.
      • Peng T.
      • Qin D.
      • Lu Y.
      • Bonewald L.
      • Butler W.T.
      • Feng J.
      • Qin C.
      ). Having observed that DMP1 accumulates in the perinuclear region and the nucleus, we proceeded to investigate if DMP1 was indeed localized in the Golgi apparatus during the endocytic process. When T4-4 cells were incubated with FITC-DMP1 for 45 min and labeled with the Golgi marker, R-WGA, it was observed that FITC-DMP1 is delivered to the Golgi (Fig. 8a). R-WGA has been reported to incorporate effectively in the Golgi apparatus and Golgi-associated vesicles. The plot of the intensities of the different channels across the line drawn through a cell (Fig. 8b) reiterates this point. These results indicate that endocytosed DMP1 was targeted to the Golgi apparatus.
      Figure thumbnail gr8
      FIGURE 8Golgi and nuclear localization of the endocytosed DMP1. a, confocal image of FITC-DMP1 localizing in the Golgi. b, intensity profile across the line drawn from top right to bottom left through the cell for FITC-DMP1, R-WGA, localizing Golgi, and nucleus. Scale bar, 10 μm. Note the match in intensity profiles of FITC-DMP1 and Golgi staining in the perinuclear regions (arrow) and the presence of DMP1 in the nucleus marked by increased FITC-DMP1 signal inside the region of increased nuclear intensity (black arrow). c, confocal image showing nuclear localization of FITC-DMP1 in T4-4 odontoblast cells. Scale bar 10 μm. The insets are confocal images of the regions marked by the boxes below them. Note that the cell to the right is at a different focal plane (diffuse nuclear staining) and shows nuclear localization of FITC-DMP1 when the nucleus is brought in to the confocal plane. d, nuclear localization of DMP1 in DPSC. Scale bar, 20 μm. The arrowheads in c and d indicate FITC-DMP1, nucleus, and the merge of both from left to right. Note the presence of FITC-DMP1 in the nucleus. e, z-stack sections of a cell showing endocytosed FITC-DMP1 entering the nucleus. f, orthogonal view of the z-stack showing clearly the entry of FITC-DMP1 into the nucleus. Scale bar, 5 μm for e and f. g, a three-dimensional reconstruction of the z-stack. The black arrow points to the nuclear penetration of FITC-DMP1. Also note the intense accumulation of FITC-DMP1 in the z direction along the nuclear membrane.
      Endocytosed DMP1 Is Localized in the Nucleus—We had earlier demonstrated the presence of a functional NLS at the C-terminal end of DMP1 (amino acid residues 462–471 of rat DMP1). Based on this, we proceeded to investigate if the internalized DMP1 was translocated to the nucleus. When T4-4 cells and DPSC were incubated with FITC-DMP1 for 30 min at 37 °C, nuclear localization was seen in both T4-4 cells and DPSC (Fig. 8, c and d). Localization within the nucleus was a prominent feature in dividing cells (supplemental Fig. 2). To further validate the nuclear localization process, confocal image analysis was performed. A z section of a cell with the endocytosed FITC-DMP1 shows an early phase of nuclear entry with FITC-DMP1 observed bound to the nuclear membrane. Fig. 8e depicts a panel of the z axis slices at 1-μm intervals (order left to right, top to bottom). Fig. 8f is an orthogonal representation of the z-stack showing the x-y, y-z, and x-z axes. The intersection of the lines in the x-y plot marks the spot of nuclear entry. It can be seen clearly from the x-z plot the presence of FITC-DMP1 to the left of the vertical line, indicating positive nuclear entry. Fig. 8g is a three-dimensional reconstruction of the z-stack images. This image shows DMP1 binding to the nuclear membrane and its subsequent entry into the nucleus (black arrow). Additionally, nuclear localization is also seen in many of the other results when the incubation was long enough (Figs. 1, f and h, and 8, a and b, and supplemental Figs. 1 and 2). Finally, quantification of fluorescence in T4-4 cells after 30 and 60 min of incubation with FITC-DMP1 at 37 °C indicated that 17.81% of the endocytosed FITC-DMP1 was nuclear-localized at 30 min and 28.32% at 60 min (Table 1).
      TABLE 1
      TimeIntracellular location
      CytoplasmNucleus
      min
      30 (n = 31)82.19 (S.E. = 1.04)17.81 (S.E. = 1.04)
      60 (n = 10)71.68 (S.E. = 3.1)28.32 (S.E. = 3.1)
      Tissue Localization of GRP-78 and DMP1—To shed light on the localization of GRP-78 at the tooth tissue level, immunohistochemical analysis of developing mouse molars was performed with anti-GRP-78 (Fig 9, a and b) and anti-DMP1 (Fig. 9, c and d) antibodies. These images show similar spatial localization pattern in the polarized odontoblasts and surprisingly in the ameloblasts as well. Far Western immunohistochemical analysis was also performed by incubating mouse head sections with biotinylated DMP1 (Fig. 9, e and f) or FITC-DMP1 (Fig. 9g). Results demonstrate that the DMP1 binding pattern observed was similar to the GRP-78 immunostaining specifically localizing in the polarizing odontoblasts and the ameloblasts. Negative controls with both secondary antibody and streptavidin peroxidase for the immunohistochemical analysis (Fig. 9, i and j) and with streptavidin peroxidase alone for the experiment with biotinylated DMP1 (Fig. 9, k and l) did not show any reaction with the substrate. Fig. 9h shows background fluorescence from sections incubated with unlabeled DMP1 and serves as a control for Fig. 9g. Overall, this is the first report demonstrating the presence and localization of GRP-78 in the developing tooth.
      Figure thumbnail gr9
      FIGURE 9Immunohistochemistry and Far Western immunohistochemistry. a and b, GRP-78 immunostained sections of 5-day mouse molar. Arrows point to intense staining in the polarizing preodontoblasts and the Tomes process of polarized ameloblasts. c and d, DMP1 immunostained sections of 5-day mouse molar showing a similar pattern of staining to GRP78. Arrows point to similarly stained regions in GRP-78 sections. e and f, 5-day mouse molar of section incubated with biotinylated DMP1 and developed with streptavidin peroxidase. Arrows point to areas stained similarly to GRP-78 immunostained sections (polarized odontoblasts and ameloblast tomes regions). g, 5-day mouse molar of section incubated with FITC-DMP1 showing similar staining to that of biotin-labeled DMP1-treated sections. h, background fluorescence from sections incubated with unlabeled DMP1. i and j, negative controls with both secondary antibody and streptavidin peroxidase for the immunohistochemical analysis. k and l, negative controls with streptavidin peroxidase alone for the experiment with biotinylated DMP1; AM, ameloblasts; OD, odontoblasts.

      DISCUSSION

      DMP1 has been identified in the extracellular matrix of bone and dentin, suggesting a direct role in biomineralization. In a recent study four forms of DMP1 have been identified in the ECM of bone and dentin; namely, the N-terminal fragment, the C-terminal fragment, N-terminal fragment containing a proteoglycan (DMP1-PG), and full-length DMP1 (
      • Huang B.
      • Maciejewska I.
      • Sun Y.
      • Peng T.
      • Qin D.
      • Lu Y.
      • Bonewald L.
      • Butler W.T.
      • Feng J.
      • Qin C.
      ). We have demonstrated earlier that DMP1 contains a NLS at the C-terminal end. This functional NLS domain was responsible for the nuclear import of DMP1 in preodontoblasts and preosteoblasts. In the nucleus DMP1 can regulate the transcription of osteoblast- and odontoblast-specific genes (
      • Narayanan K.
      • Ramachandran A.
      • Hao J.
      • He G.
      • Park K.W.
      • Cho M.
      • George A.
      ,
      • Narayanan K.
      • Gajjeraman S.
      • Ramachandran A.
      • Hao J.
      • George A.
      ). More recently DMP1 has been implicated in the regulation of phosphate homeostasis through fibroblast growth factor 23 (
      • Qin C.
      • D'Souza R.
      • Feng J.Q.
      ). During mineralization, DMP1 is exported to the ECM, where it functions to nucleate hydroxyapatite (
      • He G.
      • Dahl T.
      • Veis A.
      • George A.
      ,
      • He G.
      • George A.
      ). This bimodal pattern of localization provided direct evidence that DMP1 is involved in complex cellular functions during biomineralization.
      In the present study we demonstrate that internalization of full-length DMP1 by preodontoblasts and dental pulp stem cells occurs through endocytosis. This endocytic process is mediated by specific binding of DMP1 with the GRP-78 receptor present on the cell surface of preodontoblasts. The spatial localization patterns for DMP1 and GRP-78 in the developing tooth might implicate their possible interactions. This is further supported by the binding of biotinylated DMP1 to the polarizing odontoblasts in 5-day mouse molars in a pattern resembling GRP-78 immunostaining. Interestingly the N-terminal DMP1 construct used in this study contained the GRP-78 interacting domain to facilitate endocytosis. The N-terminal polypeptide generated in this study does not correspond to the 37-kDa N-terminal DMP1 isolated from bone by Qin et al. (
      • Qin C.
      • D'Souza R.
      • Feng J.Q.
      ,
      • Qin C.
      • Brunn J.C.
      • Cook R.G.
      • Orkiszewski R.S.
      • Malone J.P.
      • Veis A.
      • Butler W.T.
      ,
      • Steiglitz B.M.
      • Ayala M.
      • Narayanan K.
      • George A.
      • Greenspan D.S.
      ). Strikingly, GRP-78 down-regulation by shRNA affected the endocytosis of DMP1 in T4-4 cells, thus confirming the role of GRP-78 in mediating DMP1 endocytosis. Thus, a novel function for GRP-78 as an endocytic receptor for DMP1 in cells responsible for mineralized matrix formation is demonstrated.
      Identification of GRP-78 as a cell surface receptor for DMP1 is particularly interesting as several investigators have reported that GRP-78 induction is a protective response to a variety of stress conditions (
      • Delpino A.
      • Castelli M.
      ,
      • Xiao G.
      • Chung T.-F.
      • Pyun H.Y.
      • Fine R.E.
      • Johnson R.J.
      ). GRP-78, also known as Hspa5, is a member of the heat shock protein 70 (HSP 70) family. Its primary function is in the endoplasmic reticulum to function as a molecular chaperone. However, it has been shown to be translocated to the cell membrane in cell types that are subjected to high levels of stress namely cancer cells (
      • Wu J.
      • Kaufman R.J.
      ). GRP-78 has also been reported to be localized to the detergent-resistant microdomains when it is translocated to the plasma membrane (
      • Broquet A.H.
      • Thomas G.
      • Masliah J.
      • Trugnan G.
      • Bachelet M.
      ). More importantly, GRP-78 functions as a co-receptor for internalization of Coxsackievirus A9 via the lipid rafts (
      • Triantafilou K.
      • Fradelizi D.
      • Wilson K.
      • Triantafilou M.
      ,
      • Triantafilou K.
      • Triantafilou M.
      ), as a receptor for activated α2-macroglobulin (
      • Misra U.K.
      • Deedwania R.
      • Pizzo S.V.
      ), and as a receptor on liver cells for endocytosis of dengue virus serotype 2 (
      • Jindadamrongwech S.
      • Thepparit C.
      • Smith D.R.
      ).
      Previous studies have identified αvβ3 integrin, αvβ5 integrin, and CD44 to be putative DMP1 binding partners (
      • Jain A.
      • Karadag A.
      • Fohr B.
      • Fisher L.W.
      • Fedarko N.S.
      ,
      • Karadag A.
      • Fedarko N.S.
      • Fisher L.W.
      ). However, neither of these studies shows physical binding between DMP1 and αvβ3, αvβ5 integrins or CD44. Instead, their results suggested that cells used DMP1 as a bridge to link MMP-9 and factor H to cell surface integrin and CD44 receptors. Results from the present study exclude a role for integrins or CD44 in regulating DMP1 endocytosis, suggesting that DMP1 does not physically interact with these cell surface receptors. The in vitro binding assay in this study further confirmed the presence of a single cell surface binding site. Finally, results from the membrane protein pulldown experiments did not identify integrins or CD44 in the protein mixture. These results suggest that extracellular DMP1 might interact only with GRP-78 localized on the plasma membrane resulting in endocytosis.
      Once bound to the cell surface, DMP1 is internalized by the caveolae-mediated endocytic machinery. Several studies have shown that the caveolae is a hub for many receptors, and endocytosis via the caveolae is essential for many significant signaling pathways (
      • Roy S.
      • Luetterforst R.
      • Harding A.
      • Apolloni A.
      • Etheridge M.
      • Stang E.
      • Rolls B.
      • Hancock J.F.
      • Parton R.G.
      ,
      • Smart E.J.
      • Graf G.A.
      • McNiven M.A.
      • Sessa W.C.
      • Engelman J.A.
      • Scherer P.E.
      • Okamoto T.
      • Lisanti M.P.
      ,
      • Prevostel C.
      • Alice V.
      • Joubert D.
      • Parker P.J.
      ). Most cell surface caveolae show limited motility and dynamics. However, when triggered by an appropriate ligand, they can be internalized (
      • Tagawa A.
      • Mezzacasa A.
      • Hayer A.
      • Longatti A.
      • Pelkmans L.
      • Helenius A.
      ,
      • Sharma D.K.
      • Brown J.C.
      • Choudhury A.
      • Peterson T.E.
      • Holicky E.
      • Marks D.L.
      • Simari R.
      • Parton R.G.
      • Pagano R.E.
      ). In this study endocytosis of DMP1 activated the caveolin-1 mobility at the cell surface. It is possible that caveolae-mediated endocytosis of DMP1 might initiate signaling cascades necessary for downstream events leading to odontoblast differentiation.
      The transport of endocytosed DMP1 was determined to be a well regulated event. Results from confocal laser scanning microscopy show the involvement of microfilaments and microtubules as tracks for the transport of endocytic vesicles containing DMP1 and targeted toward the Golgi apparatus. Fig. 10 summarizes the endocytic pathway taken by DMP1.
      Figure thumbnail gr10
      FIGURE 10A hypothetical model depicting the receptor-mediated endocytosis and transport of DMP1.
      In conclusion, these results strongly suggest the mechanism by which DMP1 is internalized and transported to the nucleus. This is an important step toward understanding the complex signaling mechanisms that occur during differentiation of precursor mesenchymal cells to calcified matrix producing cells. Furthermore, we have made the novel identification of GRP-78 as a cell surface receptor for DMP1.

      Acknowledgments

      We thank the University of Illinois at Chicago Proteomics core facility for carrying out protein identification analysis.

      Supplementary Material

      References

        • Narayanan K.
        • Srinivas R.
        • Ramachandran A.
        • Hao J.
        • Quinn B.
        • George A.
        Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 4516-4521
        • Narayanan K.
        • Ramachandran A.
        • Hao J.
        • He G.
        • Park K.W.
        • Cho M.
        • George A.
        J. Biol. Chem. 2003; 278: 17500-17508
        • Narayanan K.
        • Gajjeraman S.
        • Ramachandran A.
        • Hao J.
        • George A.
        J. Biol. Chem. 2006; 281: 19064-19071
        • 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
        • Ye L.
        • MacDougall M.
        • Zhang S.
        • Xie Y.
        • Zhang J.
        • Li Z.
        • Lu Y.
        • Mishina Y.
        • Feng J.Q.
        J. Biol. Chem. 2004; 279: 19141-19148
        • Katzmann D.J.
        • Babst M.
        • Emr S.D.
        Cell. 2001; 106: 145-155
        • Anderson R.G.
        • Kamen B.A.
        • Rothberg K.G.
        • Lacey S.W.
        Science. 1992; 255: 410-411
        • Pelkmans L.
        • Kartenbeck J.
        • Helenius A.
        Nat. Cell Biol. 2001; 3: 473-483
        • Nichols B.J.
        Nat. Cell Biol. 2002; 4: 374-378
        • Hao K.
        • Narayanan A.
        • Ramachandran G.
        • He A.
        • Almushayt Evans C.
        • George A.
        J. Biol. Chem. 2002; 277: 19976-19981
        • Gronthos S.
        • Mankani M.
        • Brahim J.
        • Robey P.G.
        • Shi S.
        Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 13625-13630
        • Batouli S.
        • Miura M.
        • Brahim J.
        • Tsutsui T.W.
        • Fisher L.W.
        • Gronthos S.
        • Robey P.G.
        • Shi S.
        J. Dent Res. 2003; 82: 976-981
        • Srinivasan R.
        • Chen B.
        • Gorski J.P.
        • George A.
        Connect. Tissue Res. 1999; 40: 251-258
        • Inal J.
        • Miot S.
        • Schifferli J.A.
        Exp. Cell Res. 2005; 310: 54-65
        • Luo T.
        • Fredericksen B.L.
        • Hasumi K.
        • Endo A.
        • Garcia J.V.
        J. Virol. 2001; 75: 2488-2492
        • Schousboe I.
        • Thomsen P.
        • van Deurs B.
        Eur. J. Biochem. 2004; 271: 2998-3005
        • Baron D.A.
        • Burch R.M.
        • Halushka P.V.
        • Spicer S.S.
        Am. J. Physiol. Renal Physiol. 1985; 249: 464-469
        • Macdonald J.L.
        • Pike L.J.
        J. Lipid Res. 2005; 46: 1061-1067
        • Delpino A.
        • Castelli M.
        Biosci. Rep. 2002; 22: 407-420
        • Altmeyer A.
        • Maki R.G.
        • Feldweg A.M.
        • Heike M.
        • Protopopov V.P.
        • Masur S.K.
        • Srivastava P.K.
        Int. J. Cancer. 1996; 69: 340-349
        • Weist D.I.
        • Bhandoola A.
        • Punt J.
        • Kreibich G.
        • McKean D.
        • Singer A.
        Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 1884-1889
        • Xiao G.
        • Chung T.-F.
        • Pyun H.Y.
        • Fine R.E.
        • Johnson R.J.
        Mol. Brain Res. 1999; 72: 121-128
        • McMichael B.K.
        • Kotadiya P.S.
        • Holliday T.
        • Shannon L.
        • Lee B.S.
        Bone (NY). 2006; 39: 694-705
        • George A.
        • Sabsay B.
        • Simonian P.A.L.
        • Veis A.
        J. Biol. Chem. 1993; 268: 12624-12630
        • George A.
        • Gui J.
        • Jenkins N.A.
        • Gilbert D.J.
        • Copeland N.G.
        • Veis A.
        J. Histochem. Cytochem. 1994; 42: 1527-1531
        • MacDougall M.
        • Gu T.T.
        • Luan X.
        • Simmons D.
        • Chen J.
        J. Bone Miner. Res. 1998; 13: 422-431
        • Soldati T.
        • Schliwa M.
        Nat. Rev. Mol. Cell Biol. 2006; 12: 897-908
        • Huang B.
        • Maciejewska I.
        • Sun Y.
        • Peng T.
        • Qin D.
        • Lu Y.
        • Bonewald L.
        • Butler W.T.
        • Feng J.
        • Qin C.
        Calcif Tissue Int. 2008; 82: 401-410
        • Qin C.
        • D'Souza R.
        • Feng J.Q.
        J. Dent. Res. 2007; 12: 1134-1141
        • Tagawa A.
        • Mezzacasa A.
        • Hayer A.
        • Longatti A.
        • Pelkmans L.
        • Helenius A.
        J. Cell Biol. 2005; 170: 769-779
        • Sharma D.K.
        • Brown J.C.
        • Choudhury A.
        • Peterson T.E.
        • Holicky E.
        • Marks D.L.
        • Simari R.
        • Parton R.G.
        • Pagano R.E.
        Mol. Biol. Cell. 2004; 15: 3114-3122
        • Qin C.
        • Brunn J.C.
        • Cook R.G.
        • Orkiszewski R.S.
        • Malone J.P.
        • Veis A.
        • Butler W.T.
        J. Biol. Chem. 2003; 278: 34700-34708
        • Steiglitz B.M.
        • Ayala M.
        • Narayanan K.
        • George A.
        • Greenspan D.S.
        J. Biol. Chem. 2004; 279: 980-986
        • Roy S.
        • Luetterforst R.
        • Harding A.
        • Apolloni A.
        • Etheridge M.
        • Stang E.
        • Rolls B.
        • Hancock J.F.
        • Parton R.G.
        Nat. Cell Biol. 1999; 1: 98-105
        • Wu J.
        • Kaufman R.J.
        Cell Death Differ. 2006; 13: 374-384
        • Broquet A.H.
        • Thomas G.
        • Masliah J.
        • Trugnan G.
        • Bachelet M.
        J. Biol. Chem. 2003; 278: 21601-21606
        • Triantafilou K.
        • Fradelizi D.
        • Wilson K.
        • Triantafilou M.
        J. Virol. 2002; 76: 633-643
        • Triantafilou K.
        • Triantafilou M.
        Virology. 2003; 317: 128-135
        • Misra U.K.
        • Deedwania R.
        • Pizzo S.V.
        J. Biol. Chem. 2005; 280: 26278-26286
        • Jindadamrongwech S.
        • Thepparit C.
        • Smith D.R.
        Arch. Virol. 2004; 149: 915-927
        • Jain A.
        • Karadag A.
        • Fohr B.
        • Fisher L.W.
        • Fedarko N.S.
        J. Biol. Chem. 2002; 277: 13700-13708
        • Karadag A.
        • Fedarko N.S.
        • Fisher L.W.
        Cancer Res. 2005; 65: 11545-11552
        • Smart E.J.
        • Graf G.A.
        • McNiven M.A.
        • Sessa W.C.
        • Engelman J.A.
        • Scherer P.E.
        • Okamoto T.
        • Lisanti M.P.
        Cell Biol. 1999; 19: 7289-7304
        • Prevostel C.
        • Alice V.
        • Joubert D.
        • Parker P.J.
        J. Cell Sci. 2000; 113: 2575-2584