Endoplasmic reticulum chaperone protein GRP-78 mediates endocytosis of dentin matrix protein 1

Abstract

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 K(D)=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.

FIGURE 1.

Endocytosis 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.

FIGURE 2.

Identification 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 M_r = 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).

FIGURE 3.

Three-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.

FIGURE 4.

Identification 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.

FIGURE 5.

DMP1 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.

FIGURE 6.

DMP1 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.

FIGURE 7.

Endocytic 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.

FIGURE 8.

Golgi 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.

FIGURE 9.

Immunohistochemistry 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.

FIGURE 10.

A hypothetical model depicting the receptor-mediated endocytosis and transport of DMP1.