|Home | About | Journals | Submit | Contact Us | Français|
Matrix extracellular phosphoglycoprotein (MEPE) is an inhibitor of mineralization in situ and in cell cultures where altered expression is associated with oncogenic osteomalacia and hypophosphatemic rickets. The purpose of this study was to determine whether the intact protein or the peptide(s) originating from this protein was responsible for the inhibition. The ability of the intact protein and the acidic, serine- and aspartate-rich MEPE-associated motif (ASARM) peptide to promote or inhibit de novo hydroxyapatite formation and growth of hydroxyapatite seed crystals, in both phosphorylated and dephosphorylated forms, was assessed at room temperature in a dynamic gel diffusion system at 3.5 and 5 days. The most effective nucleator concentration was also examined when associated with fibrillar type I collagen. The phosphorylated intact protein was an effective promoter of mineralization in the gelatin gel diffusion system, while the ASARM peptide was an effective inhibitor. When dephosphorylated both the intact protein and the ASARM peptide had no effect on mineralization. Associated with collagen fibrils, some of the effect of the intact protein was lost. This study demonstrates the importance of posttranslational modification for the site-specific activity of MEPE and its ASARM peptide.
Matrix extracellular phosphoglycoprotein (MEPE or OF45) was first cloned as a candidate substrate for PHEX (phosphaturic hormone with endopeptidase activity on the X chromosome) . MEPE is a member of the SIBLING (small integrin binding ligand N-glycosylated) family of extracellular matrix proteins and, similar to the other SIBLING proteins, is believed to be multifunctional, playing roles in cell signaling, mineral homeostasis, and mineralization. In situ MEPE exists as smaller peptides, including the C-terminal ASARM (acidic, serine-, and aspartic acid-rich MEPE motif) peptide, which is analogous to the C terminus of the salivary protein statherin . In bone, MEPE is mainly expressed by osteocytes .
Based on studies in mice in which MEPE was ablated  leading to enhancement of bone formation, MEPE is believed to inhibit mineralization. This was confirmed in cell culture experiments [6, 7] as well as in cell-free solution studies . In the cell-free studies, it was the ASARM peptide, a substrate and ligand for PHEX, that provided the inhibition [6, 8, 9, 10]. Similar to the other SIBLING proteins, we hypothesized that posttranslational modification of MEPE would alter its effects on mineralization and that the intact protein would have a distinct effect compared to the peptides. Thus, we tested the intact protein and the ASARM peptide in the dynamic gelatin gel system  to validate this hypothesis. We also investigated whether phosphorylation and dephosphorylation would alter the effects of these proteins and peptides on the mineralization process.
MEPE protein was generated by expression in insect cells using the sf9 system, as previously described . ASARM peptides (phosphorylated and without phosphate) were synthesized and purchased from Neo-MPS (now called PolyPeptide Laboratories, San Diego, CA). Dephosphorylation of the intact protein in Tris buffer was performed with alkaline phosphatase-agarose beads (Sigma, St. Louis, MO) by incubating overnight at room temperature.
Hydroxyapatite formation and growth were monitored in the dynamic collagen gel hydroxyapatite growth system . In this double diffusion system, Ca2+ and HPO42− (Pi) ions circulate at room temperature and diffuse into opposite ends of a 6-cm-long 10% gelatin gel (Bloom 275; Fisher Chemicals, Fair Lawn, NJ). The pH of the 10% gelatin solution, prepared in tris-(hydroxymethyl)amino-methane (Tris) buffer by stirring at 50°C, was adjusted to 7.40 with 0.1 N NaOH. At a site 3.4 cm from the Pi entrance point, the gel included a 100-μL band containing the protein/peptide to be tested. This band consisted of a specified concentration of MEPE or ASARM peptide, in 0.15 M pH Tris. The rest of the 100-μL band was made of 20% gelatin so that the final concentration in the band was 10%. The experimental gels contained a range of 0–50 μg/mL protein or peptide. Once this gel had hardened (~30 min in a cold room), it was covered without any interface with the 10% gelatin. This provided a “sandwich” of blank gel, with the protein in the experimental tubes and without the protein in the control tubes. All experiments were performed at ambient temperatures.
For the de novo formation and growth studies, Ca and Pi accumulation in the precipitant band containing Tris buffer was compared to experimental gels with ASARM or MEPE at 3.5 and 5.0 days. Identical “single-diffusion” gels, into which only Ca or Pi diffused, served as additional controls to correct for ion accretion due only to diffusion or to bonding of the ion to the protein or peptide. Control tubes and experimental tubes were run parallel to each other. Calcium and phosphate solutions were circulated at a continuous rate under nitrogen pressure, in separate loops at opposite ends of the gel. The calcium and phosphate solutions were maintained at 2 L, providing an “infinite reservoir” relative to the 3 mL contents of the 18 gels used for each experiment. The system was not thermostated; however, the ambient temperature around the gels was monitored, and experiments were not done if the temperature was outside the range of 20–25°C. Physiological temperatures could not be used since the gelatin began to melt at 35°C. After 3.5 days, a thin opaque band (precipitant band) formed perpendicular to the direction of the flow and either the precipitant band was cut out for mineral analysis (X-ray diffraction) or the entire gel was cut into slices (0.30 mL vol) and the Ca  and Pi  contents of each slice were determined following hydrolysis in 1 N HCl. Values in single-diffusion tubes were subtracted from those from double-diffusion tubes, as described previously .
For measurement of seeded growth, the protein or peptide was incubated overnight in 100 μL Tris with 0.5 mg hydroxyapatite, and the entire 100 μl was transferred to the position of the precipitant band. Controls containing 0.5 mg hydroxyapatite without protein or peptide were used. Analyses were performed, as above, at 5.0 days.
For studies of the interaction with fibrillar collagen, the most effective MEPE concentration (25 μg/mL) was mixed with fibrillar type I collagen (BD Biosciences, San Jose, CA), which was slowly neutralized with 0.1 N NaOH. MEPE was not added to the solution until the first signs of fibril formation were visible. After addition of MEPE, the pH was readjusted to 7.4. The control for this experiment contained only neutralized fibrillar collagen without MEPE. MEPE and the collagen fibrils or the collagen fibrils alone were incubated overnight at 37°C prior to suspension in the 100-μl band of the gelatin gel. The fibrillar collagen + MEPE was compared to fibrillar collagen alone after 5.0 days, as described above.
To account for any time or temperature fluctuations during the course of analyses, all data were normalized to the controls run at the same time and expressed as experimental/control. Mean values for each set of concentrations were calculated for each experimental/control data set, and these were compared to 1.0 using the Dunnett test (GraphPad Instat2, La Jolla, CA). Data were excluded if the Ca/Pi ratio in the experimental or control tubes differed from 1.7 by >15%. In that case, additional experiments were added so that for some concentrations n = 6 and for others n = 3. Differences with P < 0.05 were considered significant. Where only single concentrations were tested, a one-sided t-test was used, assuming the samples represented a gaussian distribution of variables.
The ASARM peptide inhibited both de novo hydroxyapatite formation and apatite-seeded growth in a dose-dependent fashion (Fig. 1a); this inhibition was lost when the peptide was dephosphorylated (Fig. 1b). The mineral formed in all cases was shown by X-ray diffraction to be a poorly crystalline hydroxyapatite (data not shown), with Ca/Pi concentrations in the experimental and control precipitant bands being 1.63 ± 0.20, and 1.56 ± 0.16, respectively.
In contrast, at low concentrations, the intact MEPE promoted hydroxyapatite formation and growth (Fig. 2a), although at higher concentrations it behaved like the ASARM peptide and was an inhibitor. The Ca/Pi concentrations in the experimental and control precipitant bands were 1.78 ± 0.27 and 1.64 ± 0.08, respectively. Similar to the ASARM peptides, once dephosphorylated there was no significant effect of the intact protein on mineralization (Fig. 2b).
When 25 μg/mL MEPE was combined with fibrillar collagen, to mimic what might occur in bone or dentin, much of its potential to promote hydroxyapatite formation was lost; however, the increased yield of mineral was still statistically significant. There was a difference between the yield with MEPE alone and MEPE + fibrillar collagen for Ca uptake, but phosphate uptake showed more scatter and did not differ significantly (Fig. 3).
Our study demonstrated that, similar to the other SIBLING proteins, MEPE’s mineralizing activity is a function of its state of phosphorylation and of the extent of cleavage. This is in agreement with other SIBLING proteins, osteopontin, dentin matrix protein 1, and bone sialoprotein (Table 1), which have different activities depending on their extent of cleavage and extent of phosphorylation, although in some cases the intact protein is the inhibitor and the cleaved protein is the promoter of mineralization.
It is interesting to note that recently MEPE with the ASARM peptide deleted was found to promote bone formation in culture and in mice ; similarly, a midterminal fragment of MEPE has been found to enhance cell binding  and pulp repair . Thus, distinct components of MEPE appear to have different effects on mineralized tissue formation and signaling processes. It is also important to note that in dentin pulp cells and in osteoblasts MEPE expression decreases as the cells differentiate [17, 18].
MEPE’s interaction with collagen, at the concentration tested, slightly diminished its effect as a promoter of mineralization. This is in contrast to another SIBLING nucleator, bone sialoprotein (BSP), which has enhanced activity when bound to collagen . Although this might be due to different methods of binding the protein to the collagen fibrils, it is likely that it reflects the relative importance of the interaction of BSP and MEPE with collagen.
Because MEPE was cloned as the substrate for PHEX , it is useful to compare the findings of the present study and the account of Addison et al.  for the hypophosphatemic phenotype in mice lacking PHEX. Cathepsin B is proposed to be a candidate protease for MEPE cleavage in bone . A targeted “bone-specific” cathepsin B null mouse is required to resolve this hypothesis. However, one can explain the phenotype of the MEPE knockout and MEPE transgenics as well as the hypophosphatemic mouse based on the finding that the intact protein is a nucleator and the cleaved peptide an inhibitor. When MEPE is present, it and the other SIBLING proteins regulate initial mineral deposition in the osteoid and predentin of the mineralizing tissues. Their activities are regulated by phosphorylation, dephosphorylation, and, of course, cleavage. Once cleaved, the ASARM peptide is protected from binding to the mineral via its interaction with MEPE .
In hypophosphatemic rickets, when PHEX is absent, the ASARM peptides bind to the mineral and inhibit further mineralization. In the MEPE knockout the enhanced mineralization most likely results from the absence of the PHEX–ASARM interaction, while in the MEPE transgenic the increased release and accumulation of the ASARM peptide results in decreased mineralization .
While this study supports the hypothesis that posttranslational modifications of the SIBLING proteins determine their tissue-specific effects , it was limited in the following ways. First, we did not have sufficient intact protein to do assays at all concentrations, especially in the presence of fibrillar collagen. Second, we looked at only two time points (3.5 and 5.0 days), preventing discovery of what would have happened at shorter and longer times. Finally, this was a solution-based study, where the concentrations of the other SIBLING proteins did not vary, as they do in the knockout and transgenic. While results are thus easier to interpret, a caveat must be that in the animal the interactions of all these SIBLING proteins are linked in some way.
This work was supported by NIH grants DE04141 (to A. L. B.) and AR51598-01 (to P. S. N. R.). The authors express their thanks to all the students from Columbia University School of Dentistry who worked on this study over the past few years.
The authors have stated that they have no conflicts of interest.
Adele L. Boskey, Musculoskeletal Integrity Program, Hospital for Special Surgery, 535 East 70th Street, New York, NY 10021, USA.
Phyllis Chiang, Musculoskeletal Integrity Program, Hospital for Special Surgery, 535 East 70th Street, New York, NY 10021, USA.
Alexis Fermanis, Musculoskeletal Integrity Program, Hospital for Special Surgery, 535 East 70th Street, New York, NY 10021, USA.
Jared Brown, Musculoskeletal Integrity Program, Hospital for Special Surgery, 535 East 70th Street, New York, NY 10021, USA.
Hayat Taleb, Musculoskeletal Integrity Program, Hospital for Special Surgery, 535 East 70th Street, New York, NY 10021, USA.
Valentin David, The Kidney Institute, Kansas University Medical Center, 3901 Rainbow Boulevard, Kansas City, KS 66160, USA.
Peter S. N. Rowe, The Kidney Institute, Kansas University Medical Center, 3901 Rainbow Boulevard, Kansas City, KS 66160, USA.