Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Dual Functional Roles of Dentin Matrix Protein 1

Dual Functional Roles of Dentin Matrix Protein 1 THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 278, No. 19, Issue of May 9, pp. 17500 –17508, 2003 © 2003 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. IMPLICATIONS IN BIOMINERALIZATION AND GENE TRANSCRIPTION BY ACTIVATION OF INTRACELLULAR Ca STORE* Received for publication, December 12, 2002, and in revised form, February 19, 2003 Published, JBC Papers in Press, March 3, 2003, DOI 10.1074/jbc.M212700200 Karthikeyan Narayanan‡, Amsaveni Ramachandran‡, Jianjun Hao‡, Gen He‡, Kyle Won Park§, Michael Cho§, and Anne George‡ From the ‡Department of Oral Biology, University of Illinois, Chicago, Illinois 60612 and the §Department of Bioengineering, University of Illinois, Chicago, Illinois 60607 mechanisms that control differentiation of osteoblast pheno- Dentin matrix protein 1 (DMP1) is a bone- and teeth- specific protein initially identified from mineralized type during proliferation, maturation, and mineralization is dentin. Here we report that DMP1 is primarily localized necessary for understanding various skeletal disorders. in the nuclear compartment of undifferentiated osteo- MC3T3-E1 cells are a well-established preosteoblast cell line blasts. In the nucleus, DMP1 acts as a transcriptional derived from mouse calvaria and maintain much of the tightly component for activation of osteoblast-specific genes linked controls between proliferation and differentiation. like osteocalcin. During the early phase of osteoblast These cells, when treated with -glycerophosphate and ascor- maturation, Ca surges into the nucleus from the cyto- bic acid, differentiate into mature osteoblast phenotype and plasm, triggering the phosphorylation of DMP1 by a nu- produce a calcifiable matrix that recapitulates in vivo condi- clear isoform of casein kinase II. This phosphorylated tions. Mineralized nodule formation takes place at least 18 –21 DMP1 is then exported out into the extracellular matrix, days after induction of mineralization. During the early stage where it regulates nucleation of hydroxyapatite. Thus, (3–5 days) of induction the preosteoblastic cells undergo prolif- DMP1 is a unique molecule that initiates osteoblast dif- eration, and at later stage (8 –12 days) the cells differentiate to ferentiation by transcription in the nucleus and orches- mature osteoblast capable of synthesis and assembly of miner- trates mineralized matrix formation extracellularly, at alized matrix with increased alkaline phosphatase activity and later stages of osteoblast maturation. The data pre- production of type I collagen. Dentin matrix protein 1 (DMP1) sented here represent a paradigm shift in the under- is a non-collagenous extracellular matrix protein identified standing of DMP1 function. This information is crucial from mineralized matrix of dentin and bone. DMP1 is highly in understanding normal bone formation, remodeling, anionic and rich in aspartic acid, glutamic acid, and serine fracture healing, and skeletal tissue repair. residues. 52% of these serines can be potentially phosphorylated by casein kinase II. Based on its high negative charge, it has been postulated to play an important role in mineralized tissue forma- Mesenchymal stem cells have the potential to differentiate tion, more specifically, by initiation of nucleation and modulation into several cell types that give rise to bone, cartilage, fat, and of mineral phase morphology (6 –9). Recent experiments demon- muscles. Proliferation and differentiation of mesenchymal cells strated that overexpression of DMP1 in embryonic mesenchymal to osteoblastic lineage is regulated by an intrinsic genetically cells resulted in characteristic morphological changes accompa- defined program, which is well-controlled by various transcrip- nied by transcriptional up-regulation of OCN and AP. Blocking tion factors, cytokines, morphogens, and secreted growth fac- tors. There are two known transcription factors, namely Cbfa1 the translation of DMP1 by antisense expression inhibited the expression of OCN and AP genes. Furthermore, stable cell lines and osterix, that regulate osteoblast differentiation and skele- tal formation during embryonic development (1). Cbfa1-defi- overexpressing antisense DMP1 failed to initiate mineralized nodule formation in cell culture systems (10). These experiments cient mice have an osteopenic skeleton (2) and are known to regulate the expression of bone sialoprotein, osteopontin, den- laid the foundation for speculation of a dual functional role for DMP1 during osteoblast differentiation. In this report we dem- tin matrix protein 1, osteocalcin, and collagen type I (3). Re- cently osterix has been shown to act downstream of Cbfa1 and onstrate that DMP1 resides in the nucleus, cytoplasm, and ex- tracellular matrix of osteoblasts depending on their differentia- functions to regulate the differentiation of preosteoblasts into mature osteoblasts (4). Differentiated osteoblasts synthesize a tion state and exhibits pleiotropic effects. Combined with experimental evidence, we suggest a bifunctional role for DMP1 number of calcium-binding proteins like bone sialoprotein, os- teopontin, and osteocalcin and secrete a complex extracellular during osteoblast differentiation and maturation. matrix that has the capacity to nucleate hydroxyapatite crystal formation when adequate amounts of calcium and phosphate The abbreviations used are: DMP1, dentin matrix protein 1; OCN, are supplied (reviewed in Ref. 5). Understanding the regulatory osteocalcin; AP, alkaline phosphatase; FITC, fluorescein isothiocya- nate; PBS, phosphate-buffered saline; BSA, bovine serum albumin; * This work was supported by National Institutes of Health Grants NLS, nuclear localization signal; NES, nuclear export signal; GFP, DE-11657 (to A. G.), DE-13836 (to A. G.), and GM60741 (to M. C.). The green fluorescent protein; ORF, open reading frame; GST, glutathione costs of publication of this article were defrayed in part by the payment S-transferase; CKI/II, casein kinases I and II; DRB, 5,6-dichloro-1--D- of page charges. This article must therefore be hereby marked “adver- ribofuranosylbenzimidazole; HBSS, Hanks’ balanced salt solution; tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate CCD, charge-coupled device; TRITC, tetramethylrhodamine isothiocya- this fact. nate; ECM, extracellular matrix; MEF2, myocyte enhancer factor 2; To whom correspondence should be addressed: Dept. of Oral Biol- HDAC, histone deacetylase; IP , inositol 1,4,5-trisphosphate; ogy, University of Illinois, 801 S. Paulina St., Chicago, IL 60612. Tel.: BAPTA-AM, 1,2-bis(o-aminophenoxy)ethane-N,N,NN-tetraacetic 312-413-0738; Fax: 312-996-6044; E-mail: [email protected]. acid. 17500 This paper is available on line at http://www.jbc.org This is an Open Access article under the CC BY license. Dual Functional Roles of DMP1 17501 MATERIALS AND METHODS Cell Culture and Transfections—The mouse pre-osteoblastic cells, MC3T3-E1, were cultured with Dulbecco’s modified Eagle’s medium supplemented with 10% heat-inactivated fetal bovine serum (Celgro). Transient transfections with reporter plasmids were performed with Superfect (Qiagen) as per the manufacturer’s protocol. Reporter transfections were carried out in triplicates and repeated at least three times to obtain a mean value. All the transfections contained an internal control vector pRLSV40, which contains a Renilla luciferase gene driven by SV40 promoter. Promoter activity at the control point was taken as 100% activity. FITC Labeling of DMP1—100 g of recombinant DMP1 was adjusted to pH 9.5 with phosphate buffer (400 l). 100 l of FITC solution (0.1 FIG.1. a, immunocytochemical localization of DMP1. MC3T3-E1 cells mg/ml) was added and incubated at 4 °C for 2 h. The reaction mixture grown on coverslips were fixed with paraformaldehyde and incubated was passed through a G-25 column to remove excessive FITC. Further- with monoclonal anti-tubulin antibody (green) and affinity-purified an- more, the labeled protein was dialyzed against PBS buffer at 4 °C for ti-DMP1 antibody (red) for 4 h followed by incubation with appropriate 18 h. Bovine serum albumin was labeled in a similar manner and was secondary antibodies. After incubation the cells were washed and used as a control. For the uptake studies labeled proteins were added mounted on a glass slide. DNA was positively stained with Hoechst dye exogenously to the cells at a concentration of 30 g/ml. (blue). Cells were observed under a laser confocal microscope. Bar  10 Antibody Purification, Immunostaining, and Immunoprecipitation— m. In a: upper left, tubulin antibody; upper right, DMP1 antibody; Polyclonal DMP1 antibody was affinity-purified as described earlier lower left, nuclear staining by Hoechst dye; and lower right, composite (11) using rDMP1 coupled to CNBr-activated Sepharose column. For of first three images. b, uptake of DMP1 by osteoblasts. Recombinant immunostaining, cells grown on coverslips were fixed with paraformal- DMP1 was labeled with FITC as described under the “Materials and Methods.” Free uncoupled FITC was removed by dialysis. MC3T3-E1 dehyde. Fixed cells were incubated with DMP1 antibody in the presence cells were grown on coverslips; labeled DMP1 was added exogenously to of 5% BSA for 4 h. Upon washing with PBS containing 1% Triton X-100, the cells at a concentration of 30 g/ml. For the control experiments, the cells were incubated with appropriate secondary antibody (fluores- BSA was labeled in a similar manner and added exogenously at the cent labeled) for 2 h. The coverslips were then mounted and observed same concentration (data not shown). The cells were fixed at different under laser confocal microscope (Zeiss, LSM 510). Monoclonal tubulin time intervals and observed under laser confocal microscope. Panel A, antibody was purchased from Sigma; nucleus staining dye, propidium migration of labeled DMP1 at different time points and its accumula- iodide, and Hoechst dye were purchased from Molecular Probes. Immu- tion in the nucleus; panel B, immunostaining with monoclonal anti- noprecipitation was carried out as described earlier (12). tubulin antibody; and panel C, the composite of A and B. Note the Site-directed Mutagenesis—Site-directed mutagenesis was carried accumulation of the protein around the nuclear envelope at 15 min after out with the following primers to mutate the respective potential NLS the addition of DMP1, and migration of the labeled protein into the sites. The lowercase letters represent the modified bases. NLS1: 724 bp, nucleus was observed within 30 min. Bar  20 m. 5-TCAAGCaGGAcATCCTTCAGAAGGTCCgGGGTCTCT-3, 760 bp; NLS2: 1294 bp, 5-TCTCAGGACAGTAGCgGATCCAcAGAAGAGAGC- then imaged for DMP1 localization. 3, 1327 bp; NLS3: 1384 bp, 5-GCTGACAATgGGAcACTAATAGcT- Calcium Imaging—Cells were seeded onto glass coverslips at least GATGCT-3, 1414 bp. An in vitro site-directed mutagenesis system 24 h prior to use. Attached cells were washed four times with Hanks’ (GeneEditor, Promega Inc.) was used to achieve mutations. Mutated balanced salt solution (HBSS) without calcium chloride. The cells were sites were verified by sequencing. then loaded with Fluo3 (Molecular Probes, Eugene, OR, 10 M final Plasmid Constructs—Double-stranded oligonucleotides synthesized concentration) for 45 min in the dark. They were then washed four for the NLS1 (724 –760 bp), NLS2 (1294 –1327 bp), NLS3 (1384 –1414 times with HBSS, mounted onto a slide with elevated edges allowing to bp), and NES (13– 48 bp) were ligated to the carboxyl-terminal end of be flushed by the buffer (HBSS containing 10 mM -glycerophosphate the GFP protein at SmaI site with ORF to GFP. pEGFP (Clontech, Palo and 100 g/ml ascorbic acid) every 3 min. Images were taken using a Alto, CA) was used in this study. An osteocalcin promoter driving the cooled CCD camera (CoolSnapFX, Roper Scientific) mounted on a Nikon luciferase gene was a gift from Dr. Gerard Karsenty at Baylor College microscope. Metamorph software (Universal Imaging, PA) was used to of Medicine, Houston, TX. DMP1 sense and antisense plasmids were obtain and analyze data. The intensity of Fluo3 and TRITC-DMP1 were constructed as described earlier (10). calculated with appropriate background subtraction. For monitoring GST Pull-down Assays—GST-importin constructs were obtained as a both the DMP1 and Ca , the cells were fed with TRITC-DMP1 for 3 h kind gift from Dr. Stephen Adams, Northwestern University, Chicago. before loading with Fluo3. Inhibition of phosphorylation was carried out Recombinant GST-importin  bound to glutathione-Sepharose beads by 75 M DRB, whereas calcium was inhibited by 30 M BAPTA. was used in GST pull-down assays. 100 g of the extracted protein was BAPTA or DRB were added to cells for 30 min prior to imaging by added to the column and washed with 0.1 M NaCl in PBS buffer. GST confocal microscopy. The Fluo3 intensity was used to monitor the mod- beads were boiled in SDS-sample buffer, and the bound proteins were ulation in calcium level. detected by Western blotting using DMP1 antibody. RESULTS In Vitro Phosphorylation of DMP1—Recombinant DMP1 was in vitro phosphorylated by casein kinases I and II mixture as described by the DMP1 Is Localized in the Nucleus of Undifferentiated Osteo- manufacturer (Upstate Biotechnology Inc., Lake Placid, NY). Phospho- blasts—Based on immunohistochemical data from mature bone rylation was confirmed by monoclonal phosphoserine antibody. tissue we had expected to localize the protein in the mineral- Casein Kinase II Assay—An CKII assay was carried out using the ized matrix (7). Contrary to expectations, immunostaining of casein kinase II assay kit (Upstate Biotechnology). Briefly, 20 ng of recombinant DMP1 was incubated with the nuclear extracts of MC3T3-E1 cells with an affinity-purified monospecific DMP1 MC3T3-E1 cells in the presence and absence of CKII-specific peptide antibody indicated DMP1 to be predominantly nuclear-local- (RRRDDDSDDD) or CKII-specific inhibitor (5,6-dichloro-1--D-ribo- ized. However, we also observed a small amount of fluorescence furanosylbenzimidazole, DRB). The reaction was allowed for 30 min at outside the nucleus (Fig. 1a). In a second set of experiments, 30 °C. The proteins were resolved on a 10% SDS-PAGE and dried for FITC-labeled recombinant DMP1, when added to the culture autoradiography. The intensity of phosphorylation was measured using medium of MC3T3-E1 cells, migrated into the nucleus in a Kodak Digital Science software. In other cases, after phosphorylation the proteins were precipitated with 5% trichloroacetic acid followed by time-dependent manner. Labeled DMP1 was found evenly dis- filter binding assay using PE81 filters. tributed in the cytoplasm within 10 min. After 15 min, DMP1 Induction of Mineralization—Mineralization was induced as de- was found to be concentrated around the outer nuclear mem- scribed earlier (10). The microenvironment for maturation of osteo- brane, and optimum localization in the nucleus occurred within blasts and mineral nodule formation was created by treating the cells 30 min as demonstrated by confocal microscopy (Fig. 1B). (80 –90% confluent) with medium supplemented with 5% fetal bovine Therefore, in proliferating preosteoblasts, DMP1 was predom- serum, 10 mM -glycerophosphate, and 100 g/ml ascorbic acid. To inantly nuclear. However, the mechanism by which DMP1 is mimic in vivo biomineralization microenvironment, the cells were treated with cyclopiazonic acid, at a concentration of 1 M for 2 h and taken up by osteoblasts is currently unknown. 17502 Dual Functional Roles of DMP1 DMP1 in the nucleus of preosteoblasts may initiate osteoblastic differentiation. DMP1 Contains a Functional NLS Sequence at the Carboxyl End—The regulated transport of proteins across the nuclear envelope has been recognized as a critical step in vast number of cellular processes (13, 14). For large proteins such as DMP1 (66 kDa), import into the nucleus would likely require the presence of nuclear localization signals (NLS) and its associ- ated transport machinery. Inspection of the primary sequence of DMP1 led to the identification of three potential NLS se- quences: NLS1 residues 242–252 (amino acid sequence SSRKS- FRRSRVS), NLS2 residues 432- 442 (amino acid sequence SQDSSRSKEES), and NLS3 residues 472– 481 (amino acid sequence ADNRKLIVDA). Mutations were introduced into the NLS1, 2, and 3 sequences as described under “Materials and Methods” to identify the functional NLS domain. Mutations in NLS1 and NLS2 did not hinder the transport of DMP1 into the nucleus (Fig. 3). However, mutations in NLS3 resulted in in- tense cytoplasmic accumulation of the labeled protein (Fig. 3). Results from this mutation study clearly indicate that the NLS3 domain is functional and is required for nuclear import of DMP1. To further characterize the transport pathway, we in- vestigated the interaction of DMP1 with -importin by GST-- importin pull-down assay. Binding assays clearly demon- strated that mutations at the NLS3 (N3) site affected the interaction of DMP1 with -importin. However, mutations on FIG.2. A, Northern blot showing the expression of OCN and AP NLS2 (N2) and NLS1 (N1) did not have any effect on importin (alkaline phosphatase) in the MC3T3-E1 cells overexpressing sense and binding (see Fig. 8B, panel IV). Thus, these results indicate the antisense DMP1. Glyceraldehyde-3-phosphate dehydrogenase was used specific interaction of NLS3 domain with -importin leading to as a housekeeping gene. B, effect of DMP1 on the OCN promoter activity: OCN promoter driving the luciferase gene was co-transfected the import of DMP1 into the nucleus. with an increasing concentration of DMP1 (pcDNA3.1-DMP1). Cells DMP1 Is Localized in the Extracellular Matrix during Bi- were lysed and assayed for luciferase. All transfections were carried out omineralization—Immunohistochemical studies have demon- in triplicates and contained an internal control of Renilla luciferase strated the presence of DMP1 in the mineralized extracellular driven by an SV40 promoter. Luciferase activity was normalized with the Renilla luciferase activity. OCN promoter activity in the control matrix of bone and dentin. This suggests that DMP1 must MC3T3 cell was taken as 100%. C, OCN promoter activity during migrate from the nucleus rather than be maintained in a nu- osteoblast differentiation. MC3T3 cells were induced to undergo differ- clear pool. To examine the export mechanism, MC3T3-E1 cells entiation and transfected with OCN promoter driving the luciferase were treated for 2 days (early maturation stage) with ascorbic gene. Transfections were carried out on different days of differentiation as described in B. pcDNA3.1-DMP1 (15 g) plasmid was co-transfected acid and -glycerophosphate, an organic phosphate, to stimu- for the DMP1 samples. Stable cell line overexpressing antisense late differentiation, because Dulbecco’s modified Eagle’s me- DMP1 (10) was also induced to undergo differentiation (AS-DMP1). dium does not have sufficient phosphate ion product to support OCN promoter activity was analyzed as described above. normal mineral formation. Immunostaining of these mineral- izing cultures with a DMP1 antibody demonstrated a striking Novel Function of DMP1 as a Transcriptional Regulator in relocation of DMP1: instead of being in the nucleus, DMP1 was the Nucleus—To address the role of DMP1 in the nucleus, we now located in the cytoplasm and the plasma membrane (Fig. have previously shown that overexpression of DMP1 in 4a). These data corroborated well with published reports re- MC3T3-E1 cells and C3H10T1/2 (embryonic mesenchymal) garding the localization of DMP1 extracellularly, in the bone cells resulted in characteristic morphological changes accom- matrix. Also, in vivo DMP1 can be localized in the nucleus of panied by transcriptional up-regulation of osteocalcin and al- preosteoblasts and in the mineralized matrix of a mature os- kaline phosphatase (10). However, the antisense-mediated re- teoblast (data not shown). pression of DMP1 protein led to an inhibition of osteocalcin and Release of Intracellular Calcium Is Required for Export of alkaline phosphatase expression level (Fig. 2A). Based on this DMP1 to the ECM—One of the main events during osteoblast experimental evidence we investigated if DMP1 could be oper- differentiation and maturation is the release of calcium from ating directly as a transcriptional regulator for matrix genes intracellular stores. We hypothesized that Ca might serve involved in mineralization. To this end, we specifically inves- as a signal for DMP1 export. Therefore, we investigated if the tigated the effect of DMP1 on OCN (osteocalcin) promoter ac- export of DMP1 from the nucleus during mineralized matrix tivity. Results in Fig. 2B demonstrate that there is no signifi- formation might be in response to a stimulus from the cal- cant increase in OCN promoter activity with increasing cium microenvironment. To directly analyze this question, we concentration of DMP1 plasmid. This result is not surprising treated MC3T3-E1 cells with cyclopiazonic acid, a stimulant because it is well established that OCN expression increases for the release of intracellular Ca stores without altering severalfold when preosteoblasts undergo differentiation. Over- the IP levels (15). Confocal imaging in the presence of cyclo- expression of DMP1 during osteoblast differentiation did not piazonic acid (1 M, “”) showed that the release of calcium have any significant effect on OCN promoter activity (Fig. 2C), however, MC3T3 cells overexpressing antisense DMP1 failed to from the endoplasmic reticulum, in fact, triggered the export of DMP1 from the nucleus to the extracellular matrix (Fig. show an increase in OCN promoter activity during differenti- ation (Fig. 2C). These data clearly demonstrates that DMP1 in 4b). This result is in good agreement with the in vivo local- ization of DMP1 in the extracellular matrix of bone and conjunction with other osteoblast-specific transcription factors regulate the expression of osteocalcin gene. We speculate that dentin. Dual Functional Roles of DMP1 17503 FIG.3. Identification of putative NLS in DMP1. Analysis of the primary sequence of DMP1 led to the identifica- tion of three potential NLS sequences (NLS1, NLS2, and NLS3). To investigate the function of these domains, mutations were made on these sequences as de- scribed under “Materials and Methods.” Mutated proteins were expressed recom- binantly and labeled with FITC as de- scribed earlier. The labeled proteins (pan- el A) were added at a concentration of 30 g/ml to the cells grown on coverslips. The cells were fixed with paraformalde- hyde and counterstained with tubulin an- tibody (panel B). The nuclear compart- ment was stained with Hoechst dye (panel C). Panel D is the composite of A–C. NLS1 and NLS2 bar  10 m and NLS3 bar  5 m. Calcium Dynamics and DMP1 Export during Biomineraliza- tion—To investigate the functional role of Ca in the export process of DMP1, MC3T3-E1 cells in mineralizing cultures were loaded first with TRITC-labeled DMP1 followed by the fluorescent calcium-sensitive dye Fluo3. Exposing the cells to a simulated mineralization medium containing -glycerophos- phate (the concentration of inorganic phosphate used promoted biomineralization and not ectopic mineral deposition) initiated osteoblast differentiation. Strikingly, this process evoked a bi- phasic Ca response. Live cell microscopic analysis revealed an initial rapid release of calcium from intracellular stores followed by a massive influx of this pool of Ca into the nucleus. However, after 3 h, the elevated nuclear calcium levels declined and returned to the basal level. Interestingly, this influx of calcium into the nucleus triggered the export of DMP1 from the nucleus to the extracellular matrix (Fig. 5). This translocation of Ca into the nucleus, under mineralization conditions, is probably specific for cells involved in synthesizing a mineralized matrix, because control fibroblastic NIH3T3 cells failed to respond in a similar manner (Fig. 6). During bone formation, extracellular phosphate levels are raised signifi- cantly and induce changes in gene expression (16). We specu- FIG.4. a, localization of DMP1 during mineralization. Cells were induced to undergo mineralization for 2 days with -glycerophos- late that this elevated extracellular phosphate levels triggers phate and ascorbic acid, and the localization of DMP1 was demon- the release of calcium from intracellular stores of osteoblasts strated by the DMP1 antibody (green), nuclear staining by Hoechst resulting in the export of DMP1 from the nucleus of differen- dye. Please note that mineralization stimulates nuclear export of tiating osteoblasts. DMP1, and it is now localized in the extracellular matrix. Bar  20 m. b, release of intracellular calcium stimulates nuclear export of Export of DMP1 from the Nucleus to the Cytoplasm Is Im- DMP1. MC3T3-E1 cells were treated with cyclopiazonic acid ()(1 paired by BAPTA-AM, a Calcium Chelator—An important M) for 2 h, fixed, and immunostained with DMP1 antibody. Cells question is whether this Ca influx into the nucleus plays a without cyclopiazonic acid served as the control (). Panel A shows specific role in the export of DMP1 from the nucleus. Mineral- localization of DMP1, panel B shows nuclear staining with propidium iodide, and panel C is the composite of A and B. Note the export of ization stimulus was found to evoke a steady increase in the DMP1 from the nucleus with the release of intracellular calcium. nuclear calcium levels for over 90 min, after which it subse- Bar  10 m. c, localization of phosphorylated DMP1 in MC3T3-E1 quently declined to the initial level. This influx and efflux of cells. Recombinant DMP1 was phosphorylated in vitro by casein calcium from the nucleus coincided with the export of DMP1 kinase mixture (I and II) and used in comparison with the non- from the nucleus. However, addition of BAPTA-AM, a well- phosphorylated DMP1 for uptake studies. Phosphorylated and un- phosphorylated DMP1 were FITC-labeled and added to the cells and known chelator for calcium, led to the accumulation of the fixed after 16 h. Panel A shows the localization of FITC-labeled DMP1 TRITC-DMP1 in the nucleus. Furthermore, the influx of cal- (phosphorylated (P) and unphosphorylated (NP)), panel B shows cium observed under normal mineralizing conditions did not nuclear staining by propidium iodide, and panel C is the composite of take place in the presence of BAPTA-AM (Fig. 7). These results A and B. Bar  20 m 17504 Dual Functional Roles of DMP1 FIG.7. Role of calcium on the export of DMP1. Cells were seeded and processed as described in Fig. 5, except that the cells were treated with BAPTA-AM (30 M) for 30 min before taking the images. Cells were monitored at regular intervals of 10 min. Imaging was done using a CCD camera mounted on a Nikon microscope. Metamorph software was used to obtain and analyze the data. The intensity of Fluo3 and FIG.5. Calcium dynamics during mineralization. Cells were TRITC-DMP1 were calculated with appropriate background subtrac- seeded onto coverslips at least 18 h before use. TRITC-labeled DMP1 tion. Panel A represents the presence of TRITC-DMP1 in the cells. was fed to the cells for 3 h. Cells were washed with HBSS (calcium-free) Panel B represents the calcium indicator Fluo3. Panel C represents the and loaded with Fluo3 for 45 min. Cells were mounted onto a slide with composite of A and B. Bar  5 m. Three different time frames were elevated edges giving access for the buffer to flow through. Cells were shown (0, 60, and 150 min, respectively). Please note the inability of replenished with buffer (HBSS containing -glycerophosphate and calcium to be transported into the nucleus and the retention of DMP1 ascorbic acid) every 3 min. Cells were monitored at regular intervals of within the nucleus. 10 min. Imaging was done using a CCD camera mounted on a Nikon microscope. Metamorph software was used to obtain and analyze the data. The intensity of Fluo3 and TRITC-DMP1 were calculated with appropriate background subtraction. A, presence of TRITC-DMP1 in coordination with calcium release, the sequence of DMP1 was the cells; B, the calcium indicator Fluo3; C, the composite of A and B. examined for a nuclear export signal (NES). Sequence analysis Bar  5 m. Three different time frames are shown (0, 60, and 150 min, indicated the presence of a classic leucine-rich hydrophobic respectively). Please note the movement of calcium (green) and the export of DMP1 (red) from the nucleus. export signal present at the amino-terminal end (5–16 amino acids) of the polypeptide and is conserved in all species identi- fied thus far. The functionality of both the NES and NLS domains in DMP1 was further confirmed by ligating the NES domain (LLTFLWGLSCAL) and the NLS3 sequence, individu- ally to the carboxyl terminus of the GFP construct with an ORF. Transient transfections and confocal images demon- strated that NLS3-GFP hybrid protein accumulated in the nuclear compartment. On the other hand, the NES-GFP hybrid was found to accumulate at the cellular boundary and in the ECM (Fig. 8A). These results confirmed that both NLS3 and FIG.6. Mineralization condition effect on NIH3T3 cells. NES peptide sequences are functional and are necessary for the NIH3T3 cells were seeded onto the coverslips at least 18 h before use. rapid import and export of DMP1. Cells were washed with HBSS (calcium-free) and loaded with Fluo3 for 45 min. Cells were mounted onto a slide with elevated edges giving Phosphorylation of DMP1 Is Necessary for Nucleocytoplasmic access for the buffer to flow through. Cells were replenished with buffer Transport—DMP1 is a phosphoprotein with a high negative (HBSS containing -glycerophosphate and ascorbic acid) every 3 min. charge. Phosphate groups confer a very high capacity to DMP1 Cells were monitored at regular intervals of 10 min. Imaging was done for binding calcium ions, which is important for its potential using a CCD camera mounted on a Nikon microscope. Metamorph function in mineralization. If fully phosphorylated, DMP1 software was used to obtain and analyze the data. Bar  30 m. Three different time frames were shown (A, B, and C correspond to 0, 90, and would bear a net charge of 175 per molecule of 473 residues. 120 min, respectively). To examine whether DMP1 is differentially phosphorylated in vivo, we immunoprecipitated DMP1 from the nucleus, cytosol, suggest that, in differentiating osteoblasts, DMP1 export into and extracellular matrix of MC3T3-E1 cells and cross-blotted it the extracellular matrix is directly or indirectly related to the with an anti-phosphoserine antibody. It was observed, that local release of Ca . DMP1 from the cytosol and extracellular matrix were phospho- Transport of DMP1 Is Regulated by Functional NES and rylated, however, the nuclear DMP1 was unphosphorylated NLS—To investigate the signal sequence responsible for con- (Fig. 8B, panel II). To investigate the role of phosphorylation in trolling the export of DMP1 into the extracellular matrix in the nucleocytoplasmic transport of DMP1, we investigated the Dual Functional Roles of DMP1 17505 FIG.8. A, characterization of the NLS and NES sequences in DMP1. The oligonucleotides corresponding to the putative NLS and NES domains were cloned into the 5-end of the GFP plasmid with an ORF. MC3T3-E1 cells were transiently transfected with the GFP hybrids, and localization was monitored using confocal microscopy. Note the localization of NLS-pEGFP hybrid in the nucleus and the localization of the NES-pEGFP hybrid at the periphery of the cellular membrane and in the extracellular matrix. Bars  20 (NLS-pEGFP), 10 (pEGFP and NES-pEGFP), and 3 m (NES-pEGFP, enlarged view). In B: panel I, Western blot analysis of DMP1 expressed in different cellular compartments of MC3T3-E1 cells probed with DMP1 antibody; panel II, DMP1 proteins from different compartments were immunoprecipitated using DMP1 antibody and probed for phosphorylation with a monoclonal phosphoserine antibody (note the absence of phosphorylated DMP1 in the nucleus); panel III, protein extracts from different cellular compartments of MC3T3-E1 cells were loaded onto GST- importin column to identify its interaction with DMP1. The binding of DMP1 was revealed by cross-blotting the proteins that bound to the beads using DMP1 antibody. For panels I, II, and III: T, C, N, and E represent total, cytosol, nuclear, and extracellular proteins. Panel IV, recombinant proteins were loaded onto a GST--importin column, washed, and eluted with SDS-PAGE sample buffer. Eluted proteins were analyzed for the presence of DMP1 using DMP1 antibody. For panel IV: C control, N1  NLS1-mutated DMP1, N2  NLS2-mutated DMP1, and N3  NLS3-mutated DMP1. binding of DMP1 to importin, a soluble transport factor. GST (1–5mM) to the reaction mixture combined with the nuclear pull-down assay performed using the GST-importin complex extracts increased the phosphorylating activity of CKII. Fur- showed that the DMP1 from the nuclear compartment was able thermore, this increase can be suppressed by the addition of to bind to importin, while no detectable binding was observed BAPTA-AM (30 M) (Fig. 9A). Moreover, CKII activity during for DMP1 isolated from the cytosol and extracellular compart- the mineralization process was shown to increase at least 2- to ments (Fig. 8B, panel III). Results from these two studies, 3-fold within the 24 h after mineralization induction (Fig. 9B). namely phosphorylation and importin binding assay, indicate Together, the specificity of these reactions clearly demon- that, upon phosphorylation, DMP1 might undergo a conforma- strates the presence of casein kinase II-like enzyme in the tional change enabling it to expose the NES domain leading to nuclear extracts of MC3T3-E1 cells, and their phosphorylating its export into the ECM. In a different approach, recombinant activity on DMP1 can be augmented by the addition of calcium. DMP1 was in vitro phosphorylated by CKII enzyme and added Next we addressed the question of whether blocking DMP1 exogenously to the cells. Confocal microscopy demonstrated phosphorylation can inhibit its export. For this the cells were that there was no uptake of phosphorylated DMP1 by treated with DRB and then monitored for calcium and DMP1 MC3T3-E1 cells. On the contrary, unphosphorylated recombi- dynamics using Fluo3, a Ca indicator. Confocal results nant DMP1 localized in the nucleus within 15 min (Fig. 4c). clearly indicate the movement of calcium into the nucleus, at Thus phosphorylation of DMP1 is necessary for nucleocytoplas- the same time DMP1 was retained in the nuclear compartment mic transport. (Fig. 10). Interestingly, these results demonstrate a direct cor- DMP1 Is Phosphorylated by CKII in the Nucleus Prior to Its relation between inhibition of DMP1 phosphorylation and its Export into ECM—DMP1 has several consensus sites for phos- retention within the nuclear compartment. phorylation by CKII. To identify the casein kinase responsible We further explored the role of Ca on the mechanism of for in vivo phosphorylation of DMP1, proteins were extracted DMP1 phosphorylation. DMP1 has a high affinity for binding from the nucleus and cytosol and analyzed for their phospho- 32 to calcium ions (17). Addition of calcium chloride to nuclear rylating activity in the presence of [- P]ATP as the phospho- extracts increased the phosphorylating activity of CKII in a ryl donor. Initial results demonstrate that the kinase in the dose-dependent manner. This increase could be suppressed by nuclear extract had a greater phosphorylating potential when the addition of BAPTA-AM (Fig. 9A). Reviewing the results compared with the components from the cytosol fraction (data obtained from calcium dynamics in the presence of BAPTA, not shown). The specificity of this phosphorylating activity was DRB and phosphorylation studies clearly indicate the necessity confirmed by DRB (5,6-dichloro-1--ribofuranosylbenzimida- for the presence of Ca and phosphorylation by CKII to zole, 75 M), a well-characterized specific inhibitor for CKII and achieve the export of DMP1 from the nucleus to ECM by competition with excess CKII-specific peptide. Interestingly, addition of exogenous calcium in the form of calcium chloride during mineralization. 17506 Dual Functional Roles of DMP1 FIG. 10. Role of CKII on DMP1 export. Cells were seeded onto the coverslips at least 18 h before use. TRITC-labeled DMP1 was fed to the cells for 3 h. Cells were washed with HBSS and loaded with Fluo3 for 45 min. Cells were loaded with DRB (75 M) for 30 min before taking images. Cells were replenished with buffer every 3 min. Imaging was done using a CCD camera mounted on a Nikon microscope. Metamorph software was used to obtain and analyze data. The intensities of Fluo3 and TRITC-DMP1 were calculated with appropriate background sub- traction. Panel A represents the presence of TRITC-DMP1 in the cells. Panel B represents the calcium indicator Fluo3. Panel C represents the composite of A and B. Bar  20 m. Three different time frames were FIG.9. A, phosphorylation of DMP1 by the nuclear extract and the shown (0, 60, and 150 min, respectively). role of calcium during this process. Nuclear extracts from MC3T3-E1 were incubated with recombinant DMP1 to monitor the phosphoryla- The export of DMP1 from the nucleus during maturation of tion of DMP1. Lane C represents the control; lane DRB represents osteoblasts was found to be in response to a stimulus from phosphorylation of DMP1 in the presence of CKII-specific inhibitor, DRB (75 M). Phosphorylation was also investigated in the presence of calcium ions. Calcium ions have been reported as ubiquitous increasing concentrations of calcium. 1, 2, and 5 mM CaCl solutions second messengers. The modulated release of Ca from inter- were used in this experiment. Phosphorylation was also studied in the nal stores such as the endoplasmic reticulum can be trans- presence of BAPTA. Experiments were conducted in the presence of 5 duced into an intracellular response. This signaling mecha- mM CaCl along with a 30 M concentration of BAPTA. The proteins were resolved on a 10% SDS-PAGE and dried for autoradiography. The nism might convey information across individual cells and intensity of phosphorylation was measured using Kodak Digital Science between connected cells by targeting a variety of calcium-sen- software and presented as histograms. B, CKII activity during the sitive elements that are pathway-specific. In this study Ca induced mineralization. Mineralization was induced as described ear- was found to serve as a signal for DMP1 export. It has been lier, and nuclear extracts were made at different time intervals during reported that, under mineralization conditions, cells release the mineralization process. CKII activity was measured by the incor- poration of labeled phosphate from [- P]ATP. The reaction was calcium from their intracellular stores (18). For the first time, stopped by the addition of 5% trichloroacetic acid, and a filter binding we were able to show that in mature osteoblasts there is a assay was carried out using PE81. The x axis represents the time calcium influx into the nucleus, induced by the mineralization in hours. stimulus. This influx of Ca into the nucleus is probably specific for cells involved in mineralized tissue formation, be- DISCUSSION cause the same effect was not observed in fibroblastic NIH3T3 For the first time, the results presented in this study shed cells. Two well-defined pathways have been reported for the new light on the mechanism by which nuclear localization of entry of calcium into the nucleus. Malviya and co-workers (19) DMP1 might play an important role in the regulation of specific reported the existence of IP receptors on the inner nuclear genes that control osteoblast differentiation. Experimental ev- membrane, and ryanodine receptors in the isolated liver nuclei idence presented here and elsewhere indicates a dual biological (20) have also been found to be involved in calcium transport. function for DMP1 both as a transcriptional signal during early More recently, Zadi and co-workers (21) clearly demonstrated differentiation of osteoblasts and as an initiator of mineraliza- the presence of IP and ryanodine receptors on the nuclear tion during the terminal differentiation of osteoblasts (17). membrane of osteoblast cell line (MC3T3-E1), which were re- Different spatial and temporal profiles of DMP1 lead to differ- sponsible for the transport of calcium into the nucleus. We ent pleiotropic effects. The nuclear import of DMP1 is depend- speculate that the accumulated cytoplasmic Ca gets trans- ent on the NLS3 domain present at its carboxyl end. Positive ported to the nucleus either through the IP /ryanodine receptor interactions of GST--importin with the NLS3 domain of on the nuclear membrane or by an unknown carrier protein. DMP1 further indicate the regulated, unidirectional import of However, to date there are no reports indicating the role of DMP1 into the nucleus. This phenomenon exhibited by osteo- nuclear Ca dynamics with the induction of biomineralization. blasts is cell type-specific, because fibroblasts and epithelial Casein kinase II is a messenger-independent Ser/Thr kinase cells failed to respond in a similar manner (data not shown). found predominantly in the nucleus of most cells (22). We However, the mechanism of DMP1 uptake in cells remains speculate that phosphorylation of DMP1 by casein kinase II, unclear. present in the nuclear extracts of osteoblasts, might induce a Dual Functional Roles of DMP1 17507 FIG. 11. A hypothetical model. The export of DMP1 from the nucleus to the extracellular matrix in a differentiating osteoblast is illustrated. In polarized osteoblasts, the nucleocytoplasmic transport of DMP1 is mediated by calcium ions. The phosphates present at the time of calcified tissue formation can trigger the intracellular calcium stores to release calcium. The released calcium gets transported to the nucleus either by an unknown carrier protein or through the IP /ryanodine receptors on the nuclear membrane. In the nucleus, calcium binds to DMP1, which undergoes structural modification (I). Casein kinase II then phosphorylates DMP1, leading to a conformational change that exposes the NES sequence (II). This triggers the export of the DMP1Ca complex to the extracellular matrix where the phosphorylated, highly anionic DMP1 initiates the nucleation of hydroxyapatite formation. conformational change in the protein and prevent its re-entry tissue formation. In polarized osteoblasts, the nucleocytoplas- back into the nucleus from the cytoplasm. Confocal microscopy mic transport of DMP1 is mediated by calcium ions. Addition- analysis demonstrated that phosphorylated DMP1 failed to ally, the phosphates present at the time of calcified tissue migrate into the nucleus, whereas unphosphorylated rDMP1 formation can trigger the release calcium from intracellular became localized in the nucleus within 15 min. Thus DMP1 stores. The released calcium gets transported to the nucleus must be phosphorylated to be exported out into the ECM. This either by an unknown carrier protein or through the IP /ryan- result corroborates well with a functional role for phosphoryl- odine receptors, present on the nuclear membrane of osteo- ated DMP1 in the ECM. Based on charge density, phosphoryl- blasts. In the nucleus, calcium binds to DMP1 and undergoes ated DMP1 is highly acidic and has the capacity to bind Ca structural modifications and phosphorylation is facilitated by ions. Extracellularly, DMP1 has been hypothesized to play a activating casein kinase II. The overall conformational change regulatory role in the nucleation of hydroxyapatite within the leads to the exposure of the NES sequence and the export of the collagenous matrix of bone and dentin (6). Thus, both Ca 2 DMP1Ca complex to the extracellular matrix where phos- signaling and phosphorylation of DMP1 by CKII are essential phorylated DMP1 with high anionic characteristics initiates regulatory components required for the export of DMP1 the nucleation of hydroxyapatite and orchestrates calcified tis- during biomineralization. sue formation. Movement of transcription factors between the The novel finding that DMP1 is localized in the nucleus and nucleus and the cytoplasm has been shown to control various acts as a transcriptional regulator for OCN expression is ap- cellular activities (23). Such tight regulation has important pealing, because it is consistent with the observation that OCN functional consequences for cell metabolism and differentia- might be one of the downstream genes known, to be regulated tion. Recently, the mechanism involved in the repression of by DMP1. Based on our overexpression studies, we also spec- myogenesis differentiation in the presence of histone deacety- ulate that DMP1 may directly activate transcriptional path- lase has been reported. Briefly, the differentiation of myoblast ways leading to expression of alkaline phosphatase in osteo- depends on the availability of free myocyte enhancer factor 2 blasts (10). Thus our findings suggest a bifunctional role for (MEF2). Histone deacetylase (HDAC) represses myogenesis by DMP1 during biomineralization and support a model in which forming a complex with MEF2. Calcium/calmodulin-dependent Ca is actively supplied to the mineralized matrix. The model protein kinase induces the myoblast differentiation by phos- (Fig. 11) predicts that DMP1, synthesized by preosteoblasts, is phorylating HDAC, thereby disrupting the complex with transported into the nucleus by binding to soluble transport MEF2, and the phosphorylated HDAC gets exported to the factors such as -importin. In the nucleus DMP1 is responsible for the transcription of matrix genes involved in mineralized cytoplasm (25). Phosphorylation has also been reported to play 17508 Dual Functional Roles of DMP1 11. Srinivasan, R., Chen, B., Gorski J. P., and George, A. (1999) Connect. Tissue a major role in the transport of factors into the nucleus and in Res. 40, 251–258 export to the cytoplasm (24 –28). 12. Harlow, E., and Lane, D. (1998) Antibodies: A Laboratory Manual, pp. 421– 470, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY Acknowledgments—We thank Dr. Trisha Gura and Dr. Primal de 13. Jans, D. A., and Hubner, S. (1996) Physiol. Rev. 76, 651– 685 Lanerolle for critical editing and helpful comments on the manuscript. 14. Misteli, T. (2001) Science 291, 843– 847 15. Demaurex, N., Lew, D. P., and Krause, K. H. (1992) J. Biol. Chem. 267, We acknowledge Dr. Mei Lin at the confocal microscope center for her 2318 –2324 assistance. 16. Beck, G. R., Jr., Zerler, B., and Moran, E. (2000) Proc. Natl. Acad. Sci. U. S. A. REFERENCES 97, 8352– 8357 17. He, G., Dahl, T., Veis, A., and George, A. (2003) Connect. Tissue Res. 44, Suppl. 1. Ducy, P., Zhang, R., Geoffroy, V., Ridall, A. L., and Karsenty, G. (1997) Cell 89, 1, 240 –245 747–754 18. Lundgren, T., Nilsson, M., Ritchie, H. H., and Linde, A. (2001) Calcif. Tissue 2. Otto, F., Thornell, A. P., Crompton, T., Denzel, A., Gilmour, K. C., Rosewell, Int. 68, 192–201 I. R., Stamp, G. W., Beddington, R. S., Mundlos, S., Olsen, B. R., Selby, 19. Humbert, J. P., Matter, N., Artault, J. C., Koppler, P., and Malviya, A. N. P. B., and Owen, M. J. (1997) Cell 89, 765–771 (1996) J. Biol. Chem. 271, 478 – 485 3. Ducy, P., Starbuck, M., Priemel, M., Shen, J., Pinero, G., Geoffroy, V., Amling, 20. Gerasimenko, O. V., Gerasimenko, J. V., Tepikin, A. V., and Petersen, O. H. M., and Karsenty, G. (1999) Genes Dev. 13, 1025–1036 (1995) Cell 80, 439 – 444 4. Nakashima, K., Zhou, X., Kunkel, G., Zhang, Z., Deng, J. M., Behringer, R. R., 21. Adebanjo, O. A., Biswas, G., Moonga, B. S., Anandatheerthavarada, H. K., and de Crombrugghe, B. (2002) Cell 108, 17–29 Sun, L., Bevis, P. J., Sodam, B. R., Lai, F. A., Avadhani, N. G., and Zaidi, M. 5. Karsenty, G. (2001) Endocrinology 142, 2731–2733 (2000) Am. J. Physiol. 278, F784 –F791 6. George, A., Sabsay, B., Simonian, P. A. L., and Veis A. (1993) J. Biol. Chem. 22. Faust, M., and Montenarh, M. (2000) Cell Tissue Res. 301, 329 –340 268, 12624 –12630 23. Yoneda, Y. (2000) Genes Cells 5, 777–787 7. George, A., Gui, J., Jenkins, N. A., Gilbert, D. J., Copeland, N. G., and Veis, A. 24. McKinsey, T. A., Chun-Li, Z., Lu, J., and Olson, E. N. (2000) Nature 408, (1994) J. Histochem. Cytochem. 42, 1527–1531 106 –111 8. D’Souza, R. N., Cavender, A., Sunavala, S. G., Alvarez, J., Ohshima, T., 25. Boyd, S. D., Tsai, K. Y., and Jacks, T. (2000) Nat. Cell Biol. 2, 563–568 Kulkarni, A. B., and MacDougall, M. (1997) J. Bone Miner. Res. 12, 2040 –2049 26. Geyer, R. K., Yu, Z. K., and Maki, C. G. (2000) Nat. Cell Biol. 2, 569 –573 27. Kao, H. Y., Verdel, A., Tsai, C. C., Simon, C., Juguilon, H., and Khochbin, S. 9. MacDougall, M., Gu, T. T., Luan, X., Simmons, D., and Chen, J. (1998) J. Bone Miner. Res. 13, 422– 431 (2001) J. Biol. Chem. 276, 47496 – 47507 10. Narayanan, K., Srinivas, R., Ramachandran, A., Hao, J., Quinn, B., and 28. Timmerman, L. A., Clipstone, N. A., Ho, S. N., Northrop, J. P., and Crabtree, George, A. (2001) Proc. Natl. Acad. Sci. U. S. A. 98, 4516 – 4521 G. R. (1996) Nature 383, 837– 840 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Biological Chemistry Unpaywall

Dual Functional Roles of Dentin Matrix Protein 1

Download PDF
9 pages

Loading next page...
 
/lp/unpaywall/dual-functional-roles-of-dentin-matrix-protein-1-0KbVb9013s

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Unpaywall
ISSN
0021-9258
DOI
10.1074/jbc.m212700200
Publisher site
See Article on Publisher Site

Abstract

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 278, No. 19, Issue of May 9, pp. 17500 –17508, 2003 © 2003 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. IMPLICATIONS IN BIOMINERALIZATION AND GENE TRANSCRIPTION BY ACTIVATION OF INTRACELLULAR Ca STORE* Received for publication, December 12, 2002, and in revised form, February 19, 2003 Published, JBC Papers in Press, March 3, 2003, DOI 10.1074/jbc.M212700200 Karthikeyan Narayanan‡, Amsaveni Ramachandran‡, Jianjun Hao‡, Gen He‡, Kyle Won Park§, Michael Cho§, and Anne George‡ From the ‡Department of Oral Biology, University of Illinois, Chicago, Illinois 60612 and the §Department of Bioengineering, University of Illinois, Chicago, Illinois 60607 mechanisms that control differentiation of osteoblast pheno- Dentin matrix protein 1 (DMP1) is a bone- and teeth- specific protein initially identified from mineralized type during proliferation, maturation, and mineralization is dentin. Here we report that DMP1 is primarily localized necessary for understanding various skeletal disorders. in the nuclear compartment of undifferentiated osteo- MC3T3-E1 cells are a well-established preosteoblast cell line blasts. In the nucleus, DMP1 acts as a transcriptional derived from mouse calvaria and maintain much of the tightly component for activation of osteoblast-specific genes linked controls between proliferation and differentiation. like osteocalcin. During the early phase of osteoblast These cells, when treated with -glycerophosphate and ascor- maturation, Ca surges into the nucleus from the cyto- bic acid, differentiate into mature osteoblast phenotype and plasm, triggering the phosphorylation of DMP1 by a nu- produce a calcifiable matrix that recapitulates in vivo condi- clear isoform of casein kinase II. This phosphorylated tions. Mineralized nodule formation takes place at least 18 –21 DMP1 is then exported out into the extracellular matrix, days after induction of mineralization. During the early stage where it regulates nucleation of hydroxyapatite. Thus, (3–5 days) of induction the preosteoblastic cells undergo prolif- DMP1 is a unique molecule that initiates osteoblast dif- eration, and at later stage (8 –12 days) the cells differentiate to ferentiation by transcription in the nucleus and orches- mature osteoblast capable of synthesis and assembly of miner- trates mineralized matrix formation extracellularly, at alized matrix with increased alkaline phosphatase activity and later stages of osteoblast maturation. The data pre- production of type I collagen. Dentin matrix protein 1 (DMP1) sented here represent a paradigm shift in the under- is a non-collagenous extracellular matrix protein identified standing of DMP1 function. This information is crucial from mineralized matrix of dentin and bone. DMP1 is highly in understanding normal bone formation, remodeling, anionic and rich in aspartic acid, glutamic acid, and serine fracture healing, and skeletal tissue repair. residues. 52% of these serines can be potentially phosphorylated by casein kinase II. Based on its high negative charge, it has been postulated to play an important role in mineralized tissue forma- Mesenchymal stem cells have the potential to differentiate tion, more specifically, by initiation of nucleation and modulation into several cell types that give rise to bone, cartilage, fat, and of mineral phase morphology (6 –9). Recent experiments demon- muscles. Proliferation and differentiation of mesenchymal cells strated that overexpression of DMP1 in embryonic mesenchymal to osteoblastic lineage is regulated by an intrinsic genetically cells resulted in characteristic morphological changes accompa- defined program, which is well-controlled by various transcrip- nied by transcriptional up-regulation of OCN and AP. Blocking tion factors, cytokines, morphogens, and secreted growth fac- tors. There are two known transcription factors, namely Cbfa1 the translation of DMP1 by antisense expression inhibited the expression of OCN and AP genes. Furthermore, stable cell lines and osterix, that regulate osteoblast differentiation and skele- tal formation during embryonic development (1). Cbfa1-defi- overexpressing antisense DMP1 failed to initiate mineralized nodule formation in cell culture systems (10). These experiments cient mice have an osteopenic skeleton (2) and are known to regulate the expression of bone sialoprotein, osteopontin, den- laid the foundation for speculation of a dual functional role for DMP1 during osteoblast differentiation. In this report we dem- tin matrix protein 1, osteocalcin, and collagen type I (3). Re- cently osterix has been shown to act downstream of Cbfa1 and onstrate that DMP1 resides in the nucleus, cytoplasm, and ex- tracellular matrix of osteoblasts depending on their differentia- functions to regulate the differentiation of preosteoblasts into mature osteoblasts (4). Differentiated osteoblasts synthesize a tion state and exhibits pleiotropic effects. Combined with experimental evidence, we suggest a bifunctional role for DMP1 number of calcium-binding proteins like bone sialoprotein, os- teopontin, and osteocalcin and secrete a complex extracellular during osteoblast differentiation and maturation. matrix that has the capacity to nucleate hydroxyapatite crystal formation when adequate amounts of calcium and phosphate The abbreviations used are: DMP1, dentin matrix protein 1; OCN, are supplied (reviewed in Ref. 5). Understanding the regulatory osteocalcin; AP, alkaline phosphatase; FITC, fluorescein isothiocya- nate; PBS, phosphate-buffered saline; BSA, bovine serum albumin; * This work was supported by National Institutes of Health Grants NLS, nuclear localization signal; NES, nuclear export signal; GFP, DE-11657 (to A. G.), DE-13836 (to A. G.), and GM60741 (to M. C.). The green fluorescent protein; ORF, open reading frame; GST, glutathione costs of publication of this article were defrayed in part by the payment S-transferase; CKI/II, casein kinases I and II; DRB, 5,6-dichloro-1--D- of page charges. This article must therefore be hereby marked “adver- ribofuranosylbenzimidazole; HBSS, Hanks’ balanced salt solution; tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate CCD, charge-coupled device; TRITC, tetramethylrhodamine isothiocya- this fact. nate; ECM, extracellular matrix; MEF2, myocyte enhancer factor 2; To whom correspondence should be addressed: Dept. of Oral Biol- HDAC, histone deacetylase; IP , inositol 1,4,5-trisphosphate; ogy, University of Illinois, 801 S. Paulina St., Chicago, IL 60612. Tel.: BAPTA-AM, 1,2-bis(o-aminophenoxy)ethane-N,N,NN-tetraacetic 312-413-0738; Fax: 312-996-6044; E-mail: [email protected]. acid. 17500 This paper is available on line at http://www.jbc.org This is an Open Access article under the CC BY license. Dual Functional Roles of DMP1 17501 MATERIALS AND METHODS Cell Culture and Transfections—The mouse pre-osteoblastic cells, MC3T3-E1, were cultured with Dulbecco’s modified Eagle’s medium supplemented with 10% heat-inactivated fetal bovine serum (Celgro). Transient transfections with reporter plasmids were performed with Superfect (Qiagen) as per the manufacturer’s protocol. Reporter transfections were carried out in triplicates and repeated at least three times to obtain a mean value. All the transfections contained an internal control vector pRLSV40, which contains a Renilla luciferase gene driven by SV40 promoter. Promoter activity at the control point was taken as 100% activity. FITC Labeling of DMP1—100 g of recombinant DMP1 was adjusted to pH 9.5 with phosphate buffer (400 l). 100 l of FITC solution (0.1 FIG.1. a, immunocytochemical localization of DMP1. MC3T3-E1 cells mg/ml) was added and incubated at 4 °C for 2 h. The reaction mixture grown on coverslips were fixed with paraformaldehyde and incubated was passed through a G-25 column to remove excessive FITC. Further- with monoclonal anti-tubulin antibody (green) and affinity-purified an- more, the labeled protein was dialyzed against PBS buffer at 4 °C for ti-DMP1 antibody (red) for 4 h followed by incubation with appropriate 18 h. Bovine serum albumin was labeled in a similar manner and was secondary antibodies. After incubation the cells were washed and used as a control. For the uptake studies labeled proteins were added mounted on a glass slide. DNA was positively stained with Hoechst dye exogenously to the cells at a concentration of 30 g/ml. (blue). Cells were observed under a laser confocal microscope. Bar  10 Antibody Purification, Immunostaining, and Immunoprecipitation— m. In a: upper left, tubulin antibody; upper right, DMP1 antibody; Polyclonal DMP1 antibody was affinity-purified as described earlier lower left, nuclear staining by Hoechst dye; and lower right, composite (11) using rDMP1 coupled to CNBr-activated Sepharose column. For of first three images. b, uptake of DMP1 by osteoblasts. Recombinant immunostaining, cells grown on coverslips were fixed with paraformal- DMP1 was labeled with FITC as described under the “Materials and Methods.” Free uncoupled FITC was removed by dialysis. MC3T3-E1 dehyde. Fixed cells were incubated with DMP1 antibody in the presence cells were grown on coverslips; labeled DMP1 was added exogenously to of 5% BSA for 4 h. Upon washing with PBS containing 1% Triton X-100, the cells at a concentration of 30 g/ml. For the control experiments, the cells were incubated with appropriate secondary antibody (fluores- BSA was labeled in a similar manner and added exogenously at the cent labeled) for 2 h. The coverslips were then mounted and observed same concentration (data not shown). The cells were fixed at different under laser confocal microscope (Zeiss, LSM 510). Monoclonal tubulin time intervals and observed under laser confocal microscope. Panel A, antibody was purchased from Sigma; nucleus staining dye, propidium migration of labeled DMP1 at different time points and its accumula- iodide, and Hoechst dye were purchased from Molecular Probes. Immu- tion in the nucleus; panel B, immunostaining with monoclonal anti- noprecipitation was carried out as described earlier (12). tubulin antibody; and panel C, the composite of A and B. Note the Site-directed Mutagenesis—Site-directed mutagenesis was carried accumulation of the protein around the nuclear envelope at 15 min after out with the following primers to mutate the respective potential NLS the addition of DMP1, and migration of the labeled protein into the sites. The lowercase letters represent the modified bases. NLS1: 724 bp, nucleus was observed within 30 min. Bar  20 m. 5-TCAAGCaGGAcATCCTTCAGAAGGTCCgGGGTCTCT-3, 760 bp; NLS2: 1294 bp, 5-TCTCAGGACAGTAGCgGATCCAcAGAAGAGAGC- then imaged for DMP1 localization. 3, 1327 bp; NLS3: 1384 bp, 5-GCTGACAATgGGAcACTAATAGcT- Calcium Imaging—Cells were seeded onto glass coverslips at least GATGCT-3, 1414 bp. An in vitro site-directed mutagenesis system 24 h prior to use. Attached cells were washed four times with Hanks’ (GeneEditor, Promega Inc.) was used to achieve mutations. Mutated balanced salt solution (HBSS) without calcium chloride. The cells were sites were verified by sequencing. then loaded with Fluo3 (Molecular Probes, Eugene, OR, 10 M final Plasmid Constructs—Double-stranded oligonucleotides synthesized concentration) for 45 min in the dark. They were then washed four for the NLS1 (724 –760 bp), NLS2 (1294 –1327 bp), NLS3 (1384 –1414 times with HBSS, mounted onto a slide with elevated edges allowing to bp), and NES (13– 48 bp) were ligated to the carboxyl-terminal end of be flushed by the buffer (HBSS containing 10 mM -glycerophosphate the GFP protein at SmaI site with ORF to GFP. pEGFP (Clontech, Palo and 100 g/ml ascorbic acid) every 3 min. Images were taken using a Alto, CA) was used in this study. An osteocalcin promoter driving the cooled CCD camera (CoolSnapFX, Roper Scientific) mounted on a Nikon luciferase gene was a gift from Dr. Gerard Karsenty at Baylor College microscope. Metamorph software (Universal Imaging, PA) was used to of Medicine, Houston, TX. DMP1 sense and antisense plasmids were obtain and analyze data. The intensity of Fluo3 and TRITC-DMP1 were constructed as described earlier (10). calculated with appropriate background subtraction. For monitoring GST Pull-down Assays—GST-importin constructs were obtained as a both the DMP1 and Ca , the cells were fed with TRITC-DMP1 for 3 h kind gift from Dr. Stephen Adams, Northwestern University, Chicago. before loading with Fluo3. Inhibition of phosphorylation was carried out Recombinant GST-importin  bound to glutathione-Sepharose beads by 75 M DRB, whereas calcium was inhibited by 30 M BAPTA. was used in GST pull-down assays. 100 g of the extracted protein was BAPTA or DRB were added to cells for 30 min prior to imaging by added to the column and washed with 0.1 M NaCl in PBS buffer. GST confocal microscopy. The Fluo3 intensity was used to monitor the mod- beads were boiled in SDS-sample buffer, and the bound proteins were ulation in calcium level. detected by Western blotting using DMP1 antibody. RESULTS In Vitro Phosphorylation of DMP1—Recombinant DMP1 was in vitro phosphorylated by casein kinases I and II mixture as described by the DMP1 Is Localized in the Nucleus of Undifferentiated Osteo- manufacturer (Upstate Biotechnology Inc., Lake Placid, NY). Phospho- blasts—Based on immunohistochemical data from mature bone rylation was confirmed by monoclonal phosphoserine antibody. tissue we had expected to localize the protein in the mineral- Casein Kinase II Assay—An CKII assay was carried out using the ized matrix (7). Contrary to expectations, immunostaining of casein kinase II assay kit (Upstate Biotechnology). Briefly, 20 ng of recombinant DMP1 was incubated with the nuclear extracts of MC3T3-E1 cells with an affinity-purified monospecific DMP1 MC3T3-E1 cells in the presence and absence of CKII-specific peptide antibody indicated DMP1 to be predominantly nuclear-local- (RRRDDDSDDD) or CKII-specific inhibitor (5,6-dichloro-1--D-ribo- ized. However, we also observed a small amount of fluorescence furanosylbenzimidazole, DRB). The reaction was allowed for 30 min at outside the nucleus (Fig. 1a). In a second set of experiments, 30 °C. The proteins were resolved on a 10% SDS-PAGE and dried for FITC-labeled recombinant DMP1, when added to the culture autoradiography. The intensity of phosphorylation was measured using medium of MC3T3-E1 cells, migrated into the nucleus in a Kodak Digital Science software. In other cases, after phosphorylation the proteins were precipitated with 5% trichloroacetic acid followed by time-dependent manner. Labeled DMP1 was found evenly dis- filter binding assay using PE81 filters. tributed in the cytoplasm within 10 min. After 15 min, DMP1 Induction of Mineralization—Mineralization was induced as de- was found to be concentrated around the outer nuclear mem- scribed earlier (10). The microenvironment for maturation of osteo- brane, and optimum localization in the nucleus occurred within blasts and mineral nodule formation was created by treating the cells 30 min as demonstrated by confocal microscopy (Fig. 1B). (80 –90% confluent) with medium supplemented with 5% fetal bovine Therefore, in proliferating preosteoblasts, DMP1 was predom- serum, 10 mM -glycerophosphate, and 100 g/ml ascorbic acid. To inantly nuclear. However, the mechanism by which DMP1 is mimic in vivo biomineralization microenvironment, the cells were treated with cyclopiazonic acid, at a concentration of 1 M for 2 h and taken up by osteoblasts is currently unknown. 17502 Dual Functional Roles of DMP1 DMP1 in the nucleus of preosteoblasts may initiate osteoblastic differentiation. DMP1 Contains a Functional NLS Sequence at the Carboxyl End—The regulated transport of proteins across the nuclear envelope has been recognized as a critical step in vast number of cellular processes (13, 14). For large proteins such as DMP1 (66 kDa), import into the nucleus would likely require the presence of nuclear localization signals (NLS) and its associ- ated transport machinery. Inspection of the primary sequence of DMP1 led to the identification of three potential NLS se- quences: NLS1 residues 242–252 (amino acid sequence SSRKS- FRRSRVS), NLS2 residues 432- 442 (amino acid sequence SQDSSRSKEES), and NLS3 residues 472– 481 (amino acid sequence ADNRKLIVDA). Mutations were introduced into the NLS1, 2, and 3 sequences as described under “Materials and Methods” to identify the functional NLS domain. Mutations in NLS1 and NLS2 did not hinder the transport of DMP1 into the nucleus (Fig. 3). However, mutations in NLS3 resulted in in- tense cytoplasmic accumulation of the labeled protein (Fig. 3). Results from this mutation study clearly indicate that the NLS3 domain is functional and is required for nuclear import of DMP1. To further characterize the transport pathway, we in- vestigated the interaction of DMP1 with -importin by GST-- importin pull-down assay. Binding assays clearly demon- strated that mutations at the NLS3 (N3) site affected the interaction of DMP1 with -importin. However, mutations on FIG.2. A, Northern blot showing the expression of OCN and AP NLS2 (N2) and NLS1 (N1) did not have any effect on importin (alkaline phosphatase) in the MC3T3-E1 cells overexpressing sense and binding (see Fig. 8B, panel IV). Thus, these results indicate the antisense DMP1. Glyceraldehyde-3-phosphate dehydrogenase was used specific interaction of NLS3 domain with -importin leading to as a housekeeping gene. B, effect of DMP1 on the OCN promoter activity: OCN promoter driving the luciferase gene was co-transfected the import of DMP1 into the nucleus. with an increasing concentration of DMP1 (pcDNA3.1-DMP1). Cells DMP1 Is Localized in the Extracellular Matrix during Bi- were lysed and assayed for luciferase. All transfections were carried out omineralization—Immunohistochemical studies have demon- in triplicates and contained an internal control of Renilla luciferase strated the presence of DMP1 in the mineralized extracellular driven by an SV40 promoter. Luciferase activity was normalized with the Renilla luciferase activity. OCN promoter activity in the control matrix of bone and dentin. This suggests that DMP1 must MC3T3 cell was taken as 100%. C, OCN promoter activity during migrate from the nucleus rather than be maintained in a nu- osteoblast differentiation. MC3T3 cells were induced to undergo differ- clear pool. To examine the export mechanism, MC3T3-E1 cells entiation and transfected with OCN promoter driving the luciferase were treated for 2 days (early maturation stage) with ascorbic gene. Transfections were carried out on different days of differentiation as described in B. pcDNA3.1-DMP1 (15 g) plasmid was co-transfected acid and -glycerophosphate, an organic phosphate, to stimu- for the DMP1 samples. Stable cell line overexpressing antisense late differentiation, because Dulbecco’s modified Eagle’s me- DMP1 (10) was also induced to undergo differentiation (AS-DMP1). dium does not have sufficient phosphate ion product to support OCN promoter activity was analyzed as described above. normal mineral formation. Immunostaining of these mineral- izing cultures with a DMP1 antibody demonstrated a striking Novel Function of DMP1 as a Transcriptional Regulator in relocation of DMP1: instead of being in the nucleus, DMP1 was the Nucleus—To address the role of DMP1 in the nucleus, we now located in the cytoplasm and the plasma membrane (Fig. have previously shown that overexpression of DMP1 in 4a). These data corroborated well with published reports re- MC3T3-E1 cells and C3H10T1/2 (embryonic mesenchymal) garding the localization of DMP1 extracellularly, in the bone cells resulted in characteristic morphological changes accom- matrix. Also, in vivo DMP1 can be localized in the nucleus of panied by transcriptional up-regulation of osteocalcin and al- preosteoblasts and in the mineralized matrix of a mature os- kaline phosphatase (10). However, the antisense-mediated re- teoblast (data not shown). pression of DMP1 protein led to an inhibition of osteocalcin and Release of Intracellular Calcium Is Required for Export of alkaline phosphatase expression level (Fig. 2A). Based on this DMP1 to the ECM—One of the main events during osteoblast experimental evidence we investigated if DMP1 could be oper- differentiation and maturation is the release of calcium from ating directly as a transcriptional regulator for matrix genes intracellular stores. We hypothesized that Ca might serve involved in mineralization. To this end, we specifically inves- as a signal for DMP1 export. Therefore, we investigated if the tigated the effect of DMP1 on OCN (osteocalcin) promoter ac- export of DMP1 from the nucleus during mineralized matrix tivity. Results in Fig. 2B demonstrate that there is no signifi- formation might be in response to a stimulus from the cal- cant increase in OCN promoter activity with increasing cium microenvironment. To directly analyze this question, we concentration of DMP1 plasmid. This result is not surprising treated MC3T3-E1 cells with cyclopiazonic acid, a stimulant because it is well established that OCN expression increases for the release of intracellular Ca stores without altering severalfold when preosteoblasts undergo differentiation. Over- the IP levels (15). Confocal imaging in the presence of cyclo- expression of DMP1 during osteoblast differentiation did not piazonic acid (1 M, “”) showed that the release of calcium have any significant effect on OCN promoter activity (Fig. 2C), however, MC3T3 cells overexpressing antisense DMP1 failed to from the endoplasmic reticulum, in fact, triggered the export of DMP1 from the nucleus to the extracellular matrix (Fig. show an increase in OCN promoter activity during differenti- ation (Fig. 2C). These data clearly demonstrates that DMP1 in 4b). This result is in good agreement with the in vivo local- ization of DMP1 in the extracellular matrix of bone and conjunction with other osteoblast-specific transcription factors regulate the expression of osteocalcin gene. We speculate that dentin. Dual Functional Roles of DMP1 17503 FIG.3. Identification of putative NLS in DMP1. Analysis of the primary sequence of DMP1 led to the identifica- tion of three potential NLS sequences (NLS1, NLS2, and NLS3). To investigate the function of these domains, mutations were made on these sequences as de- scribed under “Materials and Methods.” Mutated proteins were expressed recom- binantly and labeled with FITC as de- scribed earlier. The labeled proteins (pan- el A) were added at a concentration of 30 g/ml to the cells grown on coverslips. The cells were fixed with paraformalde- hyde and counterstained with tubulin an- tibody (panel B). The nuclear compart- ment was stained with Hoechst dye (panel C). Panel D is the composite of A–C. NLS1 and NLS2 bar  10 m and NLS3 bar  5 m. Calcium Dynamics and DMP1 Export during Biomineraliza- tion—To investigate the functional role of Ca in the export process of DMP1, MC3T3-E1 cells in mineralizing cultures were loaded first with TRITC-labeled DMP1 followed by the fluorescent calcium-sensitive dye Fluo3. Exposing the cells to a simulated mineralization medium containing -glycerophos- phate (the concentration of inorganic phosphate used promoted biomineralization and not ectopic mineral deposition) initiated osteoblast differentiation. Strikingly, this process evoked a bi- phasic Ca response. Live cell microscopic analysis revealed an initial rapid release of calcium from intracellular stores followed by a massive influx of this pool of Ca into the nucleus. However, after 3 h, the elevated nuclear calcium levels declined and returned to the basal level. Interestingly, this influx of calcium into the nucleus triggered the export of DMP1 from the nucleus to the extracellular matrix (Fig. 5). This translocation of Ca into the nucleus, under mineralization conditions, is probably specific for cells involved in synthesizing a mineralized matrix, because control fibroblastic NIH3T3 cells failed to respond in a similar manner (Fig. 6). During bone formation, extracellular phosphate levels are raised signifi- cantly and induce changes in gene expression (16). We specu- FIG.4. a, localization of DMP1 during mineralization. Cells were induced to undergo mineralization for 2 days with -glycerophos- late that this elevated extracellular phosphate levels triggers phate and ascorbic acid, and the localization of DMP1 was demon- the release of calcium from intracellular stores of osteoblasts strated by the DMP1 antibody (green), nuclear staining by Hoechst resulting in the export of DMP1 from the nucleus of differen- dye. Please note that mineralization stimulates nuclear export of tiating osteoblasts. DMP1, and it is now localized in the extracellular matrix. Bar  20 m. b, release of intracellular calcium stimulates nuclear export of Export of DMP1 from the Nucleus to the Cytoplasm Is Im- DMP1. MC3T3-E1 cells were treated with cyclopiazonic acid ()(1 paired by BAPTA-AM, a Calcium Chelator—An important M) for 2 h, fixed, and immunostained with DMP1 antibody. Cells question is whether this Ca influx into the nucleus plays a without cyclopiazonic acid served as the control (). Panel A shows specific role in the export of DMP1 from the nucleus. Mineral- localization of DMP1, panel B shows nuclear staining with propidium iodide, and panel C is the composite of A and B. Note the export of ization stimulus was found to evoke a steady increase in the DMP1 from the nucleus with the release of intracellular calcium. nuclear calcium levels for over 90 min, after which it subse- Bar  10 m. c, localization of phosphorylated DMP1 in MC3T3-E1 quently declined to the initial level. This influx and efflux of cells. Recombinant DMP1 was phosphorylated in vitro by casein calcium from the nucleus coincided with the export of DMP1 kinase mixture (I and II) and used in comparison with the non- from the nucleus. However, addition of BAPTA-AM, a well- phosphorylated DMP1 for uptake studies. Phosphorylated and un- phosphorylated DMP1 were FITC-labeled and added to the cells and known chelator for calcium, led to the accumulation of the fixed after 16 h. Panel A shows the localization of FITC-labeled DMP1 TRITC-DMP1 in the nucleus. Furthermore, the influx of cal- (phosphorylated (P) and unphosphorylated (NP)), panel B shows cium observed under normal mineralizing conditions did not nuclear staining by propidium iodide, and panel C is the composite of take place in the presence of BAPTA-AM (Fig. 7). These results A and B. Bar  20 m 17504 Dual Functional Roles of DMP1 FIG.7. Role of calcium on the export of DMP1. Cells were seeded and processed as described in Fig. 5, except that the cells were treated with BAPTA-AM (30 M) for 30 min before taking the images. Cells were monitored at regular intervals of 10 min. Imaging was done using a CCD camera mounted on a Nikon microscope. Metamorph software was used to obtain and analyze the data. The intensity of Fluo3 and FIG.5. Calcium dynamics during mineralization. Cells were TRITC-DMP1 were calculated with appropriate background subtrac- seeded onto coverslips at least 18 h before use. TRITC-labeled DMP1 tion. Panel A represents the presence of TRITC-DMP1 in the cells. was fed to the cells for 3 h. Cells were washed with HBSS (calcium-free) Panel B represents the calcium indicator Fluo3. Panel C represents the and loaded with Fluo3 for 45 min. Cells were mounted onto a slide with composite of A and B. Bar  5 m. Three different time frames were elevated edges giving access for the buffer to flow through. Cells were shown (0, 60, and 150 min, respectively). Please note the inability of replenished with buffer (HBSS containing -glycerophosphate and calcium to be transported into the nucleus and the retention of DMP1 ascorbic acid) every 3 min. Cells were monitored at regular intervals of within the nucleus. 10 min. Imaging was done using a CCD camera mounted on a Nikon microscope. Metamorph software was used to obtain and analyze the data. The intensity of Fluo3 and TRITC-DMP1 were calculated with appropriate background subtraction. A, presence of TRITC-DMP1 in coordination with calcium release, the sequence of DMP1 was the cells; B, the calcium indicator Fluo3; C, the composite of A and B. examined for a nuclear export signal (NES). Sequence analysis Bar  5 m. Three different time frames are shown (0, 60, and 150 min, indicated the presence of a classic leucine-rich hydrophobic respectively). Please note the movement of calcium (green) and the export of DMP1 (red) from the nucleus. export signal present at the amino-terminal end (5–16 amino acids) of the polypeptide and is conserved in all species identi- fied thus far. The functionality of both the NES and NLS domains in DMP1 was further confirmed by ligating the NES domain (LLTFLWGLSCAL) and the NLS3 sequence, individu- ally to the carboxyl terminus of the GFP construct with an ORF. Transient transfections and confocal images demon- strated that NLS3-GFP hybrid protein accumulated in the nuclear compartment. On the other hand, the NES-GFP hybrid was found to accumulate at the cellular boundary and in the ECM (Fig. 8A). These results confirmed that both NLS3 and FIG.6. Mineralization condition effect on NIH3T3 cells. NES peptide sequences are functional and are necessary for the NIH3T3 cells were seeded onto the coverslips at least 18 h before use. rapid import and export of DMP1. Cells were washed with HBSS (calcium-free) and loaded with Fluo3 for 45 min. Cells were mounted onto a slide with elevated edges giving Phosphorylation of DMP1 Is Necessary for Nucleocytoplasmic access for the buffer to flow through. Cells were replenished with buffer Transport—DMP1 is a phosphoprotein with a high negative (HBSS containing -glycerophosphate and ascorbic acid) every 3 min. charge. Phosphate groups confer a very high capacity to DMP1 Cells were monitored at regular intervals of 10 min. Imaging was done for binding calcium ions, which is important for its potential using a CCD camera mounted on a Nikon microscope. Metamorph function in mineralization. If fully phosphorylated, DMP1 software was used to obtain and analyze the data. Bar  30 m. Three different time frames were shown (A, B, and C correspond to 0, 90, and would bear a net charge of 175 per molecule of 473 residues. 120 min, respectively). To examine whether DMP1 is differentially phosphorylated in vivo, we immunoprecipitated DMP1 from the nucleus, cytosol, suggest that, in differentiating osteoblasts, DMP1 export into and extracellular matrix of MC3T3-E1 cells and cross-blotted it the extracellular matrix is directly or indirectly related to the with an anti-phosphoserine antibody. It was observed, that local release of Ca . DMP1 from the cytosol and extracellular matrix were phospho- Transport of DMP1 Is Regulated by Functional NES and rylated, however, the nuclear DMP1 was unphosphorylated NLS—To investigate the signal sequence responsible for con- (Fig. 8B, panel II). To investigate the role of phosphorylation in trolling the export of DMP1 into the extracellular matrix in the nucleocytoplasmic transport of DMP1, we investigated the Dual Functional Roles of DMP1 17505 FIG.8. A, characterization of the NLS and NES sequences in DMP1. The oligonucleotides corresponding to the putative NLS and NES domains were cloned into the 5-end of the GFP plasmid with an ORF. MC3T3-E1 cells were transiently transfected with the GFP hybrids, and localization was monitored using confocal microscopy. Note the localization of NLS-pEGFP hybrid in the nucleus and the localization of the NES-pEGFP hybrid at the periphery of the cellular membrane and in the extracellular matrix. Bars  20 (NLS-pEGFP), 10 (pEGFP and NES-pEGFP), and 3 m (NES-pEGFP, enlarged view). In B: panel I, Western blot analysis of DMP1 expressed in different cellular compartments of MC3T3-E1 cells probed with DMP1 antibody; panel II, DMP1 proteins from different compartments were immunoprecipitated using DMP1 antibody and probed for phosphorylation with a monoclonal phosphoserine antibody (note the absence of phosphorylated DMP1 in the nucleus); panel III, protein extracts from different cellular compartments of MC3T3-E1 cells were loaded onto GST- importin column to identify its interaction with DMP1. The binding of DMP1 was revealed by cross-blotting the proteins that bound to the beads using DMP1 antibody. For panels I, II, and III: T, C, N, and E represent total, cytosol, nuclear, and extracellular proteins. Panel IV, recombinant proteins were loaded onto a GST--importin column, washed, and eluted with SDS-PAGE sample buffer. Eluted proteins were analyzed for the presence of DMP1 using DMP1 antibody. For panel IV: C control, N1  NLS1-mutated DMP1, N2  NLS2-mutated DMP1, and N3  NLS3-mutated DMP1. binding of DMP1 to importin, a soluble transport factor. GST (1–5mM) to the reaction mixture combined with the nuclear pull-down assay performed using the GST-importin complex extracts increased the phosphorylating activity of CKII. Fur- showed that the DMP1 from the nuclear compartment was able thermore, this increase can be suppressed by the addition of to bind to importin, while no detectable binding was observed BAPTA-AM (30 M) (Fig. 9A). Moreover, CKII activity during for DMP1 isolated from the cytosol and extracellular compart- the mineralization process was shown to increase at least 2- to ments (Fig. 8B, panel III). Results from these two studies, 3-fold within the 24 h after mineralization induction (Fig. 9B). namely phosphorylation and importin binding assay, indicate Together, the specificity of these reactions clearly demon- that, upon phosphorylation, DMP1 might undergo a conforma- strates the presence of casein kinase II-like enzyme in the tional change enabling it to expose the NES domain leading to nuclear extracts of MC3T3-E1 cells, and their phosphorylating its export into the ECM. In a different approach, recombinant activity on DMP1 can be augmented by the addition of calcium. DMP1 was in vitro phosphorylated by CKII enzyme and added Next we addressed the question of whether blocking DMP1 exogenously to the cells. Confocal microscopy demonstrated phosphorylation can inhibit its export. For this the cells were that there was no uptake of phosphorylated DMP1 by treated with DRB and then monitored for calcium and DMP1 MC3T3-E1 cells. On the contrary, unphosphorylated recombi- dynamics using Fluo3, a Ca indicator. Confocal results nant DMP1 localized in the nucleus within 15 min (Fig. 4c). clearly indicate the movement of calcium into the nucleus, at Thus phosphorylation of DMP1 is necessary for nucleocytoplas- the same time DMP1 was retained in the nuclear compartment mic transport. (Fig. 10). Interestingly, these results demonstrate a direct cor- DMP1 Is Phosphorylated by CKII in the Nucleus Prior to Its relation between inhibition of DMP1 phosphorylation and its Export into ECM—DMP1 has several consensus sites for phos- retention within the nuclear compartment. phorylation by CKII. To identify the casein kinase responsible We further explored the role of Ca on the mechanism of for in vivo phosphorylation of DMP1, proteins were extracted DMP1 phosphorylation. DMP1 has a high affinity for binding from the nucleus and cytosol and analyzed for their phospho- 32 to calcium ions (17). Addition of calcium chloride to nuclear rylating activity in the presence of [- P]ATP as the phospho- extracts increased the phosphorylating activity of CKII in a ryl donor. Initial results demonstrate that the kinase in the dose-dependent manner. This increase could be suppressed by nuclear extract had a greater phosphorylating potential when the addition of BAPTA-AM (Fig. 9A). Reviewing the results compared with the components from the cytosol fraction (data obtained from calcium dynamics in the presence of BAPTA, not shown). The specificity of this phosphorylating activity was DRB and phosphorylation studies clearly indicate the necessity confirmed by DRB (5,6-dichloro-1--ribofuranosylbenzimida- for the presence of Ca and phosphorylation by CKII to zole, 75 M), a well-characterized specific inhibitor for CKII and achieve the export of DMP1 from the nucleus to ECM by competition with excess CKII-specific peptide. Interestingly, addition of exogenous calcium in the form of calcium chloride during mineralization. 17506 Dual Functional Roles of DMP1 FIG. 10. Role of CKII on DMP1 export. Cells were seeded onto the coverslips at least 18 h before use. TRITC-labeled DMP1 was fed to the cells for 3 h. Cells were washed with HBSS and loaded with Fluo3 for 45 min. Cells were loaded with DRB (75 M) for 30 min before taking images. Cells were replenished with buffer every 3 min. Imaging was done using a CCD camera mounted on a Nikon microscope. Metamorph software was used to obtain and analyze data. The intensities of Fluo3 and TRITC-DMP1 were calculated with appropriate background sub- traction. Panel A represents the presence of TRITC-DMP1 in the cells. Panel B represents the calcium indicator Fluo3. Panel C represents the composite of A and B. Bar  20 m. Three different time frames were FIG.9. A, phosphorylation of DMP1 by the nuclear extract and the shown (0, 60, and 150 min, respectively). role of calcium during this process. Nuclear extracts from MC3T3-E1 were incubated with recombinant DMP1 to monitor the phosphoryla- The export of DMP1 from the nucleus during maturation of tion of DMP1. Lane C represents the control; lane DRB represents osteoblasts was found to be in response to a stimulus from phosphorylation of DMP1 in the presence of CKII-specific inhibitor, DRB (75 M). Phosphorylation was also investigated in the presence of calcium ions. Calcium ions have been reported as ubiquitous increasing concentrations of calcium. 1, 2, and 5 mM CaCl solutions second messengers. The modulated release of Ca from inter- were used in this experiment. Phosphorylation was also studied in the nal stores such as the endoplasmic reticulum can be trans- presence of BAPTA. Experiments were conducted in the presence of 5 duced into an intracellular response. This signaling mecha- mM CaCl along with a 30 M concentration of BAPTA. The proteins were resolved on a 10% SDS-PAGE and dried for autoradiography. The nism might convey information across individual cells and intensity of phosphorylation was measured using Kodak Digital Science between connected cells by targeting a variety of calcium-sen- software and presented as histograms. B, CKII activity during the sitive elements that are pathway-specific. In this study Ca induced mineralization. Mineralization was induced as described ear- was found to serve as a signal for DMP1 export. It has been lier, and nuclear extracts were made at different time intervals during reported that, under mineralization conditions, cells release the mineralization process. CKII activity was measured by the incor- poration of labeled phosphate from [- P]ATP. The reaction was calcium from their intracellular stores (18). For the first time, stopped by the addition of 5% trichloroacetic acid, and a filter binding we were able to show that in mature osteoblasts there is a assay was carried out using PE81. The x axis represents the time calcium influx into the nucleus, induced by the mineralization in hours. stimulus. This influx of Ca into the nucleus is probably specific for cells involved in mineralized tissue formation, be- DISCUSSION cause the same effect was not observed in fibroblastic NIH3T3 For the first time, the results presented in this study shed cells. Two well-defined pathways have been reported for the new light on the mechanism by which nuclear localization of entry of calcium into the nucleus. Malviya and co-workers (19) DMP1 might play an important role in the regulation of specific reported the existence of IP receptors on the inner nuclear genes that control osteoblast differentiation. Experimental ev- membrane, and ryanodine receptors in the isolated liver nuclei idence presented here and elsewhere indicates a dual biological (20) have also been found to be involved in calcium transport. function for DMP1 both as a transcriptional signal during early More recently, Zadi and co-workers (21) clearly demonstrated differentiation of osteoblasts and as an initiator of mineraliza- the presence of IP and ryanodine receptors on the nuclear tion during the terminal differentiation of osteoblasts (17). membrane of osteoblast cell line (MC3T3-E1), which were re- Different spatial and temporal profiles of DMP1 lead to differ- sponsible for the transport of calcium into the nucleus. We ent pleiotropic effects. The nuclear import of DMP1 is depend- speculate that the accumulated cytoplasmic Ca gets trans- ent on the NLS3 domain present at its carboxyl end. Positive ported to the nucleus either through the IP /ryanodine receptor interactions of GST--importin with the NLS3 domain of on the nuclear membrane or by an unknown carrier protein. DMP1 further indicate the regulated, unidirectional import of However, to date there are no reports indicating the role of DMP1 into the nucleus. This phenomenon exhibited by osteo- nuclear Ca dynamics with the induction of biomineralization. blasts is cell type-specific, because fibroblasts and epithelial Casein kinase II is a messenger-independent Ser/Thr kinase cells failed to respond in a similar manner (data not shown). found predominantly in the nucleus of most cells (22). We However, the mechanism of DMP1 uptake in cells remains speculate that phosphorylation of DMP1 by casein kinase II, unclear. present in the nuclear extracts of osteoblasts, might induce a Dual Functional Roles of DMP1 17507 FIG. 11. A hypothetical model. The export of DMP1 from the nucleus to the extracellular matrix in a differentiating osteoblast is illustrated. In polarized osteoblasts, the nucleocytoplasmic transport of DMP1 is mediated by calcium ions. The phosphates present at the time of calcified tissue formation can trigger the intracellular calcium stores to release calcium. The released calcium gets transported to the nucleus either by an unknown carrier protein or through the IP /ryanodine receptors on the nuclear membrane. In the nucleus, calcium binds to DMP1, which undergoes structural modification (I). Casein kinase II then phosphorylates DMP1, leading to a conformational change that exposes the NES sequence (II). This triggers the export of the DMP1Ca complex to the extracellular matrix where the phosphorylated, highly anionic DMP1 initiates the nucleation of hydroxyapatite formation. conformational change in the protein and prevent its re-entry tissue formation. In polarized osteoblasts, the nucleocytoplas- back into the nucleus from the cytoplasm. Confocal microscopy mic transport of DMP1 is mediated by calcium ions. Addition- analysis demonstrated that phosphorylated DMP1 failed to ally, the phosphates present at the time of calcified tissue migrate into the nucleus, whereas unphosphorylated rDMP1 formation can trigger the release calcium from intracellular became localized in the nucleus within 15 min. Thus DMP1 stores. The released calcium gets transported to the nucleus must be phosphorylated to be exported out into the ECM. This either by an unknown carrier protein or through the IP /ryan- result corroborates well with a functional role for phosphoryl- odine receptors, present on the nuclear membrane of osteo- ated DMP1 in the ECM. Based on charge density, phosphoryl- blasts. In the nucleus, calcium binds to DMP1 and undergoes ated DMP1 is highly acidic and has the capacity to bind Ca structural modifications and phosphorylation is facilitated by ions. Extracellularly, DMP1 has been hypothesized to play a activating casein kinase II. The overall conformational change regulatory role in the nucleation of hydroxyapatite within the leads to the exposure of the NES sequence and the export of the collagenous matrix of bone and dentin (6). Thus, both Ca 2 DMP1Ca complex to the extracellular matrix where phos- signaling and phosphorylation of DMP1 by CKII are essential phorylated DMP1 with high anionic characteristics initiates regulatory components required for the export of DMP1 the nucleation of hydroxyapatite and orchestrates calcified tis- during biomineralization. sue formation. Movement of transcription factors between the The novel finding that DMP1 is localized in the nucleus and nucleus and the cytoplasm has been shown to control various acts as a transcriptional regulator for OCN expression is ap- cellular activities (23). Such tight regulation has important pealing, because it is consistent with the observation that OCN functional consequences for cell metabolism and differentia- might be one of the downstream genes known, to be regulated tion. Recently, the mechanism involved in the repression of by DMP1. Based on our overexpression studies, we also spec- myogenesis differentiation in the presence of histone deacety- ulate that DMP1 may directly activate transcriptional path- lase has been reported. Briefly, the differentiation of myoblast ways leading to expression of alkaline phosphatase in osteo- depends on the availability of free myocyte enhancer factor 2 blasts (10). Thus our findings suggest a bifunctional role for (MEF2). Histone deacetylase (HDAC) represses myogenesis by DMP1 during biomineralization and support a model in which forming a complex with MEF2. Calcium/calmodulin-dependent Ca is actively supplied to the mineralized matrix. The model protein kinase induces the myoblast differentiation by phos- (Fig. 11) predicts that DMP1, synthesized by preosteoblasts, is phorylating HDAC, thereby disrupting the complex with transported into the nucleus by binding to soluble transport MEF2, and the phosphorylated HDAC gets exported to the factors such as -importin. In the nucleus DMP1 is responsible for the transcription of matrix genes involved in mineralized cytoplasm (25). Phosphorylation has also been reported to play 17508 Dual Functional Roles of DMP1 11. Srinivasan, R., Chen, B., Gorski J. P., and George, A. (1999) Connect. Tissue a major role in the transport of factors into the nucleus and in Res. 40, 251–258 export to the cytoplasm (24 –28). 12. Harlow, E., and Lane, D. (1998) Antibodies: A Laboratory Manual, pp. 421– 470, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY Acknowledgments—We thank Dr. Trisha Gura and Dr. Primal de 13. Jans, D. A., and Hubner, S. (1996) Physiol. Rev. 76, 651– 685 Lanerolle for critical editing and helpful comments on the manuscript. 14. Misteli, T. (2001) Science 291, 843– 847 15. Demaurex, N., Lew, D. P., and Krause, K. H. (1992) J. Biol. Chem. 267, We acknowledge Dr. Mei Lin at the confocal microscope center for her 2318 –2324 assistance. 16. Beck, G. R., Jr., Zerler, B., and Moran, E. (2000) Proc. Natl. Acad. Sci. U. S. A. REFERENCES 97, 8352– 8357 17. He, G., Dahl, T., Veis, A., and George, A. (2003) Connect. Tissue Res. 44, Suppl. 1. Ducy, P., Zhang, R., Geoffroy, V., Ridall, A. L., and Karsenty, G. (1997) Cell 89, 1, 240 –245 747–754 18. Lundgren, T., Nilsson, M., Ritchie, H. H., and Linde, A. (2001) Calcif. Tissue 2. Otto, F., Thornell, A. P., Crompton, T., Denzel, A., Gilmour, K. C., Rosewell, Int. 68, 192–201 I. R., Stamp, G. W., Beddington, R. S., Mundlos, S., Olsen, B. R., Selby, 19. Humbert, J. P., Matter, N., Artault, J. C., Koppler, P., and Malviya, A. N. P. B., and Owen, M. J. (1997) Cell 89, 765–771 (1996) J. Biol. Chem. 271, 478 – 485 3. Ducy, P., Starbuck, M., Priemel, M., Shen, J., Pinero, G., Geoffroy, V., Amling, 20. Gerasimenko, O. V., Gerasimenko, J. V., Tepikin, A. V., and Petersen, O. H. M., and Karsenty, G. (1999) Genes Dev. 13, 1025–1036 (1995) Cell 80, 439 – 444 4. Nakashima, K., Zhou, X., Kunkel, G., Zhang, Z., Deng, J. M., Behringer, R. R., 21. Adebanjo, O. A., Biswas, G., Moonga, B. S., Anandatheerthavarada, H. K., and de Crombrugghe, B. (2002) Cell 108, 17–29 Sun, L., Bevis, P. J., Sodam, B. R., Lai, F. A., Avadhani, N. G., and Zaidi, M. 5. Karsenty, G. (2001) Endocrinology 142, 2731–2733 (2000) Am. J. Physiol. 278, F784 –F791 6. George, A., Sabsay, B., Simonian, P. A. L., and Veis A. (1993) J. Biol. Chem. 22. Faust, M., and Montenarh, M. (2000) Cell Tissue Res. 301, 329 –340 268, 12624 –12630 23. Yoneda, Y. (2000) Genes Cells 5, 777–787 7. George, A., Gui, J., Jenkins, N. A., Gilbert, D. J., Copeland, N. G., and Veis, A. 24. McKinsey, T. A., Chun-Li, Z., Lu, J., and Olson, E. N. (2000) Nature 408, (1994) J. Histochem. Cytochem. 42, 1527–1531 106 –111 8. D’Souza, R. N., Cavender, A., Sunavala, S. G., Alvarez, J., Ohshima, T., 25. Boyd, S. D., Tsai, K. Y., and Jacks, T. (2000) Nat. Cell Biol. 2, 563–568 Kulkarni, A. B., and MacDougall, M. (1997) J. Bone Miner. Res. 12, 2040 –2049 26. Geyer, R. K., Yu, Z. K., and Maki, C. G. (2000) Nat. Cell Biol. 2, 569 –573 27. Kao, H. Y., Verdel, A., Tsai, C. C., Simon, C., Juguilon, H., and Khochbin, S. 9. MacDougall, M., Gu, T. T., Luan, X., Simmons, D., and Chen, J. (1998) J. Bone Miner. Res. 13, 422– 431 (2001) J. Biol. Chem. 276, 47496 – 47507 10. Narayanan, K., Srinivas, R., Ramachandran, A., Hao, J., Quinn, B., and 28. Timmerman, L. A., Clipstone, N. A., Ho, S. N., Northrop, J. P., and Crabtree, George, A. (2001) Proc. Natl. Acad. Sci. U. S. A. 98, 4516 – 4521 G. R. (1996) Nature 383, 837– 840

Journal

Journal of Biological Chemistry – Unpaywall

Published: May 1, 2003

There are no references for this article.