Calcium ion-induced formation of β-sheet/-turn structure leading to alteration of osteogenic activity of bone morphogenetic protein-2

Preserving bioactivity of bone morphogenetic protein 2 (BMP-2) still remains a challenge in protein-based therapy. It is not known how Ca2+ released from extracellular matrix or existing in physiological environment influences bioactivity in situ till now. Here, effects of extracellular Ca2+ on conformation and osteogenic bioactivity of recombinant human BMP-2 (rhBMP-2) were investigated systematically. In vitro results indicated that Ca2+ could bind rhBMP-2 rapidly and had no obvious effect on cell behaviors. Low concentration of Ca2+ (0.18 mM) enhanced rhBMP-2-induced osteogenic differentiation, while high Ca2+ concentration (>1.80 mM) exerted negative effect. In vivo ectopic bone formation exhibited similar trend. Further studies by circular dichroism spectroscopy, fluorescence spectroscopy, together with cell culture experiments revealed at low concentration, weak interaction of Ca2+ and rhBMP-2 slightly increased β-sheet/-turn content and facilitated recognition of BMP-2 and BMPRIA. But, high Ca2+ concentration (>1.8 mM) induced formation of Ca-rhBMP-2 complex and markedly increased content of β-sheet/-turn, which led to inhibition binding of rhBMP-2 and BMPRIA and thus suppression of downstream Smad1/5/8, ERK1/2 and p38 mitogen-associated protein kinase signaling pathways. Our work suggests osteogenic bioactivity of BMP-2 can be adjusted via extracellular Ca2+, which should provide guide and assist for development of BMP-2-based materials for bone regeneration.


Results
In this study, to exclude the influence of the cations (1.8 mM of Ca 2+ and 0.8 mM of Mg 2+ ) in the culture medium, a specific DMEM without Ca 2+ and Mg 2+ was applied during the interaction of rhBMP-2 and cells. Considering the fact that BMP-2-induced osteogenic differentiation (including alkaline phosphatase activity (ALP), etc.) is often quantitatively evaluated after 72 h culture, a two-stage cell culture process, first with rhBMP-2-containing specific DMEM followed by the normal DMEM without rhBMP-2, was applied to circumvent any adverse effects of Ca 2+ -/Mg 2+ -free medium on cell viability and morphology over such extended period. In order to determine the specific time exposure to BMP-2-containing specific DMEM, we first studied the change of the ALP with the elongation of cultivation time with BMP-2-containing specific DMEM (the total culture time was fixed at 72 h). The results indicated that rhBMP-2, in the absence of cations, induced osteogenic differentiation in a time-dependent manner in the initial 12 h. An obvious increase in ALP activity was observed after 2 h interaction with rhBMP-2 and the highest ALP activity detected after a 6 h exposure ( Supplementary Fig. S1a). No additional increase was observed with longer culture times. This increase in ALP activity was not due to the cell proliferation, since cell viability and morphology were unaffected in the two-stage culture process ( Supplementary Fig.  S1b,c). Therefore, in the following section, C2C12 cells were exposed in rhBMP-2-containing specific DMEM medium in the first 6 h and thereafter cultured in regular DMEM for another 66 h or 162 h for evaluation of the Ca 2+ effect on the osteogesis of C2C12.
In the preliminary experiment, we found that with the above two-stage cell culture process, the OD value of ALP was within 0.5-1.0 when the rhBMP-2 level was in the range of 1.0-3.0 μ g/mL, thereafter, the rhBMP-2 level was fixed at 2.0 μ g/mL. Considering that the concentration of Ca 2+ in normal culture medium is 1.80 mM, the extracellular calcium concentration was fixed at 0.18 mM, 1.80 mM, 5.40 mM and 18.0 mM in this study. Cells were seeded in normal medium as negative control, and cultured in rhBMP-2-containing specific medium as positive control. In subsequent experiments, 0.18 mM-and 18.0 mM-treated rhBMP-2 were named [Ca 2+ ] L /rhBMP-2 and [Ca 2+ ] H /rhBMP-2, respectively and were used to explore the interaction of Ca 2+ and rhBMP-2 in greater detail.
Interaction of rhBMP-2 and Ca 2+ . Prior to investigate the effect of Ca 2+ on the rhBMP-2-induced osteoactivity, the interaction between rhBMP-2 and Ca 2+ was studied. By SPR analysis, it can be observed that over the immobilized rhBMP-2 molecules, Ca 2+ with concentration ranging from 1.5 × 10 −3 to 18.0 mM exhibited a rapid binding and attained an equilibrium state within 30 s (Supplementary Fig.  S2a). Equilibrium binding data from the highest to lowest Ca 2+ concentrations yielded average responses of 398.6 to 1.7 relative units that were clearly discernible above the background noise. An equilibrium dissociation constant of 0.023 M was obtained with only one binding site identified.
Next, we investigated the change of the free Ca 2+ concentration of the [Ca 2+ ] L /rhBMP-2 and [Ca 2+ ] H / rhBMP-2 in the media without calcium by inductively-coupled plasma atomic emission spectrometry (ICP-AES). The concentrations of free Ca 2+ reduced with the prolongation of incubation time Scientific RepoRts | 5:12694 | DOi: 10.1038/srep12694 ( Supplementary Fig. S2b). Specifically, the concentration of calcium in [Ca 2+ ] H /rhBMP-2 was rapidly decreased from 18.0 mM to 2.3 mM.
To further explore the effect of calcium on the C2C12 cells, we studied the change of the free Ca 2+ concentration inside/outside the C2C12 cells after exposed to the [Ca 2+ ] L /rhBMP-2 and [Ca 2+ ] H /rhBMP-2. Obtained from analysis of the extracellular calcium concentration, the results further confirmed that limited concentration of free Ca 2+ existed in the media, even for the [Ca 2+ ] H /rhBMP-2 (As shown in Supplementary Fig. S2c). Moreover, the concentrations of intracellular Ca 2+ were maintained regardless of the concentrations of extracellular Ca 2+ (Supplementary Fig. S2d). The results demonstrated that calcium could bind with rhBMP-2 and did not lead to obvious alteration of the concentration of free calcium ion inside/outside C2C12 cell.
Calcium effect on cell morphology. First, we created an in vitro environment with various extracellular Ca 2+ concentrations in specific culture media to evaluate the response of C2C12 cells to Ca 2+ alone. Concentrations in the range from 0.18 to 18.0 mM of extracellular Ca 2+ had no obvious effect on cell viability, morphology or ALP activity ( Supplementary Fig. S3).
Changes in cell morphology are associated with differentiation [18][19][20] . Then, we investigated whether different Ca 2+ would affect cell shape in the presence of rhBMP-2. As shown in Supplementary Fig. S4a, cells cultured in the presence of 0.18 mM Ca 2+ had a well spread morphology, suggested high cellular activity 18 . At higher Ca 2+ concentrations, cells increasingly became elongated and spindle-shaped. After 6 h of culture, cells exposed to 0.18 mM Ca 2+ were larger area and longer perimeter than those in other groups. These results indicated that changes in cell morphology were induced by exposure to rhBMP-2 combined with low Ca 2+ concentration.
Osteogenic differentiation as a function of Ca 2+ concentration. The effect of Ca 2+ ions on the rhBMP-2-induced ALP activity was confirmed by an ALP colorimetric assay ( Fig. 1a-b) and soluble substrate p-nitrophenylphosphate (PNPP-Na) (Fig. 1c). Figure 1a,b indicated that cells exposed to 0.18 mM Ca 2+ were induced to differentiate into ALP-positive cells; differentiation decreased relative to control at Ca 2+ concentrations > 5.40 mM (Fig. 1a,b). As shown in Fig. 1c, compared with the cells in the DMEM (control), the ALP activity of cells was significantly enhanced in the rhBMP-2-added media. And a Ca 2+ concentration-dependent effect on rhBMP-2-induced ALP activity was observed (Fig. 1c). Specifically, cells treated with 0.18 mM Ca 2+ had higher ALP activity as compared to cells exposed to other Ca 2+ concentrations and rhBMP-2 only (0 mM, positive control). At early time points (day 1 and 3), ALP activity of cells cultured in 1.80 and 5.40 mM Ca 2+ was comparable to that without Ca 2+ , while ALP value was higher on day 7 for cells cultured in 1.80 mM Ca 2+ . In contrast, ALP activity was diminished at all time points for cells exposed to 18.0 mM Ca 2+ . Moreover, when cells were incubated in osteogenic media, no significant enhancement was observed with various concentration of calcium compared with control, suggesting that the calcium ions, both in non-osteogenic and osteogenic media, all had little influence on the ALP activity of C2C12 cells.
Expression of osteogenic gene as a function of Ca 2+ . Osteogenic induction potential was investigated by assessing the expression of the osteogenic genes ALP, collagen I (Col I), osteocalcin (OCN), and runt-related transcription factor 2 (Runx2) on day 3 and 7 by real-time quantitative reverse transcription-polymerase chain reaction (RT-qPCR). ALP and Col I are early and osteocalcin is a late marker for osteoblastic differentiation, while Runx2 is a bone-specific transcription factor that regulates the commitment of pluripotent mesenchymal cells to the osteoblast lineage 21,22 .
Compared with control, a statistically significant difference was observed for the osteo-related gene expression in rhBMP-2-containing groups. Comparison of different concentrations of calcium in rhBMP-2-added groups, ALP mRNA expression was markedly higher in cells cultured in 0.18 mM Ca 2+ but lower for other concentrations relative to free rhBMP-2 on both days ( Fig. 1d-g). Similarly, Col I expression was higher at 0.18 mM but lower at 18.0 mM Ca 2+ on day 7 as compared to positive control (free rhBMP-2), OCN and Runx2 transcript levels were the highest in cells exposed to 0.18 mM Ca 2+ as compared to the other conditions. These results indicated that the expression of osteogenic genes is stimulated at low Ca 2+ (0.18 mM) and suppressed at high Ca 2+ (> 1.80 mM) concentrations.
To determine whether BMP-2 and Ca 2+ act synergistically to induce osteogenesis, the expression levels of osteogenic genes were compared in cells cultured in rhBMP-2 and Ca 2+ alone or in combination. Compared to free rhBMP-2, 0.18 and 18.0 mM Ca 2+ alone decreased the expression of ALP, Col I, OCN, and Runx2 after 3 and 7 days ( Supplementary Fig. S5). However, ALP, OCN (day 3 and 7), Col I, and Runx2 (day 7) were up-regulated in the presence of rhBMP-2 and 0.18 mM Ca 2+ but were down-regulated (except for OCN) upon exposure to rhBMP-2 and high (18.0 mM) Ca 2+ relative to the free rhBMP-2 group on day 3 and 7. These results suggested that a low concentration of Ca 2+ potentiated the osteogenic induction ability of BMP-2. Compared with rhBMP-2-added groups, calcium alone has little effect on the osteogenic gene expression of C2C12 cells.
RhBMP-2 localization and receptor binding. The first step in Ca 2+ -mediated osteogenic differentiation induced by BMP-2 is the binding of BMP-2 to its receptors 23 . In the present study, the rhBMP-2 localization on the cell membranes was evaluated by immunocytochemistry (Fig. 2a) all treatment conditions, cells were positive for rhBMP-2 immunoreactivity after 6 h; however, the signal intensity was increased in the presence of Ca 2+ in a concentration-dependent manner, with the strongest signal observed for 0.18 mM Ca 2+ , suggesting that Ca 2+ potentiated the localization of rhBMP-2 to cell membranes. Higher Ca 2+ concentrations had a negative effect on the binding efficiency of rhBMP-2 to cell membranes.
The binding capacity of rhBMP-2 to its receptors was further investigated by quartz crystal microbalance (QCM) technique and typical QCM shifts in frequency ∆f of the adsorption of rhBMP-2, [Ca 2+ ] L / rhBMP-2 and [Ca 2+ ] H /rhBMP-2 and corresponding BMPRIA (type I) were shown in Fig. 2b-d. The baseline at the beginning of the adsorption process corresponded to the baseline signal from pure PBS prior to protein injection. After the initial injection of rhBMP-2 with elevated calcium concentration, an immediate decrease in frequency was observed and ∆f was comparable with the value around 33.51 to 34.58 Hz. However, after a subsequent BMPRIA injection, The changes of frequency which are proportional to the mass changes were 14.44 Hz, 18.92 Hz and 9.12 Hz in rhBMP-2, [Ca 2+ ] L /rhBMP-2 and [Ca 2+ ] H /rhBMP-2, respectively. The results indicated that the binding efficiency of rhBMP-2 to BMPRIA was influenced by the presence of calcium ions. However, the binding availability of other BMPRs (BMPR-IB and BMPR-II) varied slightly for the [Ca 2+ ] L /rhBMP-2 and [Ca 2+ ] H /rhBMP-2 (Supplementary Table S1). (c) RhBMP-2 induced ALP activity of C2C12 cells in the presence of Ca 2+ . Cells were cultured with 2.0 μ g/mL rhBMP-2 and graded Ca 2+ concentrations in specific DMEM for 6 h, and then exposed to normal DMEM without rhBMP-2 for another 18 h, 66 h and 162 h, respectively. ALP was measured using soluble substrate p-nitrophenylphosphate. The values represent the mean ± standard deviation (n = 5, ## p < 0.01, compared with control (cells cultivated in the DMEM), *p < 0.05, **p < 0.01, compared with free rhBMP-2). (d-g) Effects of Ca 2+ ions on ALP (d), Col I(e), OCN (f) and Runx2 (g) mRNA expression induced by rhBMP-2. Cells were cultured in rhBMP-2 (2.0 μ g/mL) media with various calcium in specific DMEM for 6 h, and then exposed to normal DMEM without rhBMP-2 for another 66 h and 162 h, respectively (n = 3, # p < 0.05, compared with control (cells incubated in DMEM), *p < 0.05, compared with free rhBMP-2).
Scientific RepoRts | 5:12694 | DOi: 10.1038/srep12694 Signaling pathway analysis. Smad 1, 5 and 8 (Smad 1/5/8) are effectors of the canonical BMP-2 signaling pathway 24 . To determine whether Ca 2+ induces the phosphorylation of these factors via BMP-2, cells treated with Ca 2+ and rhBMP-2 were analyzed by western blotting. Smad 1/5/8 phosphorylation was increased at the low concentration of Ca 2+ (0.18 mM) relative to the positive control (free rhBMP-2). Conversely, phospho-Smad levels were decreased when the concentration of Ca 2+ was higher than 5.40 mM (Fig. 3a), although the total Smad 5 protein level was unchanged (Fig. 3b). Similarly, an increase in Ca 2+ concentration from 0.18 to 18.0 mM led to a reduction in ERK1/2 and p38 mitogen-associated protein kinase (MAPK) phosphorylation, while the total ERK1/2 and p38 protein level remained the same ( Fig. 3c-f). Therefore, these results indicated the simultaneous involvement of Smad and MAPK (ERK1/2 and p38) signal pathway in the regulation of rhBMP-2-induced osteogenic differentiation of C2C12 cells with various concentrations of Ca 2+ ions. To clarify the relationship between the Smad and ERK pathway and Ca 2+ influx via Ca 2+ channels, cells were pretreated with nifedipine, an inhibitor of L-type Ca 2+ channels and then stimulated with rhBMP-2/Ca 2+ to evaluate the levels of Smad/ERK phosphorylation ( Supplementary Fig. S6). The results showed that rhBMP-2-induced Smad and ERK phosphorylation was not obviously regulated in the presence of nifedipine, indicating that the Smad/ ERK signaling triggered by rhBMP-2 in the presence of calcium would operate independently of Ca 2+ channels.
The electron densities of O and N atoms in the Ca-rhBMP-2 complex were decreased since they were electron pair donors. However, the electron density of Ca 2+ , electron pair acceptor, in the Ca-rhBMP-2 complex was increased relative to that of pure Ca 2+ , indicating the formation of a coordinate covalent Ca-O bond in Ca-rhBMP-2 via an interaction between Ca 2+ and rhBMP-2.
The contributions of particular bonds of rhBMP-2 with O were characterized by quantitative evaluation of O1s. The high-resolution spectra of O1s in rhBMP-2 and [Ca 2+ ] L /rhBMP-2 were resolved into three components with peaks at 533, 532, and 531 eV, corresponding to O=C, O-C, and O-H bonds, respectively. However, in the case of [Ca 2+ ] H /rhBMP-2, two peaks appeared, which could arise from the O of carboxylate groups interacting with Ca 2+ (Fig. 4e-g).
The Ca 2p3/2 core level spectrum for [Ca 2+ ] H /rhBMP-2 was composed of two peaks: one at 347.8 eV representing CaCl 2 25 and one at 347.3 eV representing Ca-O bonds 27 from the interaction between calcium ions and carboxylate group of rhBMP-2 (Fig. 4d). These results confirmed that the Ca-rhBMP-2 complex was formed by the interaction between Ca 2+ and the O of the carboxylate group in rhBMP-2.
Ca 2+ -dependent ectopic bone formation induced by rhBMP-2. The Ca 2+ -dependent induction of osteogenesis by rhBMP-2 was further confirmed by the formation of ectopic bone in the thigh muscle pouches of mice. In the study, we added rhBMP-2 and various CaCl 2 on FDA-approved gelatin sponge. The rhBMP-2 dosage was based on the reports of our group and the ratio of rhBMP-2 and Ca 2+ was kept the same as that in vitro cell culture. Prior to in vivo experiment, we investigated the release of Ca 2+ and the effect Ca 2+ on the rhBMP-2-induced osteoactivity from/on the matrix (Supplementary Fig. S7). Release kinetic of calcium from the matrix indicated that the calcium was rapidly released from sponges in the initial 8 h (Supplementary Fig. S7a). But, after addition of rhBMP-2, the release of calcium was greatly inhibited. The result of ALP showed a similar trend with that obtained in 2D culture, confirming again that Ca regulated the rhBMP-2-induced osteogenic differentiation (Supplementary Fig. S7b).
The X-ray images revealed bone formation in the muscle in mice of Groups A, B and C, visible as an opaque area around the site of implantation. The digital images confirmed the formation of bone in Groups A to C at 2 weeks and in Groups A to D at 4 weeks. The implantation of a 10 μ g rhBMP-2/gelatin sponge with or without 0.1 mg CaCl 2 stimulated new bone formation. There was no bone formation in Group E animals (Fig. 5a,b). Based on the harvested implants (Fig. 5d,e), bone wet weight in the Group    B (rhBMP-2/0.1 mg CaCl 2 implanted) was higher than Group A (rhBMP-2 only) at 2 weeks. At 4 weeks, bone wet weights were significantly reduced in Group E, compared with Group A (p < 0.05). Moreover, Micro-CT images allowed the bone tissues to be distinguished from other tissues in the ectopic bone (Fig. 5c,f). At 4 weeks, the Group B showed a higher content of bone tissues than other groups. However, reduced bone volume/total volume was observed in Group E. In the free rhBMP-2 group, ectopic bone pellets were filled with fibrous (F, purple area) and some bone tissue (TB, blue area) at 2 weeks, and contained bone marrow (yellow and green circle) and some bone tissue at 4 weeks. With the addition of 0.1 mg CaCl 2 , not only bone marrow but large amounts of trabecular bone and lipocytes (also known as medulla ossium flava, yellow circle, white cells) were observed in the pellets (Fig. 5g). At 4 weeks, the pellets mainly contained lamellar bone within the rich medulla ossium rubra (green circle, red cells). With the addition of 1.0 or 3.0 mg CaCl 2 , fibrous and bone tissue were all observed at each time point. However, there was no ectopic bone formation in mice with the addition of 10.0 mg CaCl 2 . These results indicated that adding low concentrations of Ca 2+ promoted bone formation, while high concentrations had adverse effects.

Discussion
Efficient utilization of BMP-2 and avoidance of supraphysiological concentrations remain challenges in the area of regenerative medicine in part because of changes in secondary/tertiary structure and consequent denaturation of BMP-2 during the application in vitro and in vivo. The effects of the delivery systems, and the environmental factors on the bioactivity of rhBMP-2 have been described previously 5,[8][9][10]28 . Here, we have shown that calcium ion, a ubiquitous cation in bone repair biomaterials and in the physiological environment, could induce formation of β -sheet/-turn structure and thus modulate the recognition of BMP receptors and osteogenic activity of rhBMP-2.
Data obtained here indicated that Ca 2+ ion could bind rhBMP-2 rapidly. Even at 18.0 mM, the concentration of rhBMP-2 decreased to 2.3 mM in 30 s and had no obvious effect on the Ca 2+ concentration inside and outside of cells. Inconsistent with the previous report 29 , Ca 2+ ions alone even at higher concentrations had little effect on cell viability, cell morphology and osteogenic differentiation both in normal culture media and standard osteogenic medium (Supplementary Fig. S3). These results excluded the possibility that Ca 2+ ions itself directly mediated osteodifferentiation of C2C12. Together with the fact that BMP-2 can differentiate C2C12 into osteoblast cells, it can be confirmed that the effect of Ca 2+ on the osteodifferentiation of C2C12 is via mediating the rhBMP-2-induced osteogenesis. The ALP activity and expressions of the osteogenic genes ALP, Col I, OCN, and Runx2, along with the formation of ectopic bone in vivo all indicated that Ca 2+ exerted a typical hormetic effect on rhBMP-2-induced osteogenic differentiation: an enhancement at low Ca 2+ concentrations (0.18 mM) and an inhibition at high concentrations (especially > 5.40 mM).
As a member of the TGF-β superfamily, BMP-2 binds to BMPR-IA/IB with high affinity, and subsequently recruits the lower affinity BMPR-II to form a signaling complex, which in turn initiates a downstream signal transduction cascade (Smad and MAPK signaling pathways) [30][31][32][33][34][35] . Therefore, the role of Ca 2+ in the regulation of BMP-2-induced osteogenic differentiation requires an understanding of BMP-2 binding to its receptors and expression levels of associated signaling pathways at various Ca 2+ concentrations. The results obtained here indicated that not only the binding efficiency of rhBMP-2 to cell membranes, but also the binding capacity to BMPR-IA/IB with high affinity were increased by a low concentration of Ca 2+ , but were inhibited at a high concentration. High Ca 2+ also decreased the levels of phosphorylated Smad 1/5/8, ERK1/2 and p38. Ca 2+ , a ligand for several G-protein coupled receptors (GPCRs), can enter the cell via ion channels such as voltage-gated Ca 2+ channels (VGCCs), acid sensing ion channels (ASIC), etc, and thus results in phosphorylation of extracellular signal pathway 29 . L-type VGCCs (L-VGCCs) could often be activated by CaP crystals or the dissolved Ca 2+ and thus led to an increase in ALP expression. Here, Nifedipine, a L-VGCC blocker 36 , was found almost no effect on the osteogenesis of C2C12, confirming that Ca 2+ exerted effect via binding to rhBMP-2 and rhBMP-2-mediated signaling pathway, but not a negative feedback loop through the calcium receptors ( Supplementary Fig. S6). In light of these findings, it can therefore be concluded that Smad and MAPK (ERK1/2 and p38) signaling pathways play critical role in Ca 2+ -mediated alteration of osteogenic differentiation of rhBMP-2. Morphological changes may be an additional factor that influences osteogenic differentiation ( Supplementary Fig. S4) 29 .
Some cations, such as La 3+ , Ca 2+ , and Mg 2+ often mediate activity of proteins via alteration of the secondary structure 37,38 . As anticipated, the CD spectra indicated that Ca 2+ ions had little effect on α -helix, but induced the formation of β -turn/-sheet, especially at high concentrations. Similar phenomena have previously been observed by Nagano, A, where the β -sheet of silk protein was promoted by the binding of Ca 2+ or Zn 2+ with proteins 39,40 .The steady-state fluorescence spectroscopic analysis showed a clear blue shift and enhanced fluorescence intensity at high Ca 2+ concentrations, indicating an improvement of the hydrophobic environment for tryptophan residues, which is related to protein aggregation and conformational change. These results confirm that high concentration Ca 2+ leads to increased order within aggregation of BMP-2 molecules.
It is reported that Ca 2+ interacts with proteins through negatively charged carboxylate ions of Glu and Asp [41][42][43][44][45] . And 3-4 acidic residues in osteopontin are often needed to chelate divalent cations. It is possible that water molecules, carbonyl groups of the protein backbone, or other amino acid side chains form Scientific RepoRts | 5:12694 | DOi: 10.1038/srep12694 coordinating bonds 45 . As a homodimer linked with inter-chain disulfide bridges, the images of the charge distribution in the surface and band structure of the BMP-2 molecule, which are calculated using the Electrostatic Potential module in the software of Pymol, are shown in Fig. 6a,b. BMP-2 molecule contains 12 Asp, 10 Glu, and 2 terminal carboxyl residues. That is to say, there are 24 carboxylate groups in the BMP-2 might register with Ca 2+ and in turn mediate the conformation of BMP-2. The structural analyses α -helices were indicated as spiral, β -strands as arrow. The subunits were color-coded blue and green, respectively. The amino acid residues of negative charge represented for magenta color. Calcium ions were color-coded pink. (c) Effect of Ca 2+ on the rhBMP-2-induced osteogenic differentiation. Ca 2+ ion at low concentration facilitated for the binding capacity of BMP-2 with its receptors on cell membrane and thus enhanced Smad1/5/8, and MAPK signaling transduction, which further stimulated expression of osteogenic marker genes and the ALP activity. Additionally, increasing Ca 2+ attenuated the binding capacity to BMPR-IA, down-regulated the signaling transductions of Smad1/5/8, MAP Kinases and consequently reduced the rhBMP-2-induced gene expression and ALP activity.
provide evidence for the coordination of O in the carboxylate group of the rhBMP-2 molecule with Ca 2+ , resulting in the formation of a Ca-rhBMP-2 in the presence of high concentration of Ca 2+ . Cation/ protein complexes have been proven to facilitate the formation of inter-or intramolecular hydrogen bonds within proteins and thereby cause an amino acid chain to fold into β -turn/-sheet 40,46,47 . With these reports and the data obtained here, it can be inferred that formation of a Ca-rhBMP-2 complex may be responsible for the increase in the ordered structure of rhBMP-2 at high Ca 2+ concentrations. However, given our findings, the site specific binding information between calcium and BMP-2 is unclear.
The data obtained and discussed above led us to a proposed schematic illustration for the effect of Ca 2+ on BMP-2-induced osteogenic differentiation and was shown in Fig. 6c. At low Ca 2+ concentrations, a weak interaction between Ca 2+ and rhBMP-2 leads to a slight increase in β -sheet/-turn content and a relaxation of the rhBMP-2 molecule, which increases the exposure of active sites in the protein 26 and thereby enhances its bioactivity. While an increase in Ca 2+ concentration leads to the formation of a Ca-BMP-2 complex followed by a marked increase in the β -sheet/-turn content of the BMP-2 molecule. As a result, the binding of BMP-2 to its receptor and the activation Smad1/5/8 and MAPK signaling pathways, such as ERK1/2, p38, are inhibited, which attenuates the capacity for rhBMP-2 to induce osteogenic differentiation.
Several factors including pH, delivery molecules, and release mode among others have been shown to exert effect on the BMP-2-induced osteogenesis bioactivity 28,48 . The present study for the first time demonstrated that the extracellular Ca 2+ level can also modulate the conformation of BMP-2 and thus the capacity for inducing osteogenic differentiation. This concept, therefore, has significant implications in understanding the interactions of cation ion and BMP-2 and assisting in the design and fabrication of BMP-2-based scaffold for bone regeneration.

Methods
Alkaline phosphatase (ALP) activity assay. A two-stage cell culture process was used to detect ALP activity. C2C12 was first cultured in rhBMP-2-containing specific DMEM for 6 h and followed by the normal DMEM without rhBMP-2 for another 18 h, 66 h and 162 h. After that, cells were lysed in 50 μ L of 1% Nonidet P-40 (NP-40) solution and mixed with 100 μ L of 1 mg/mL p-nitrophenylphosphate (PNPP-Na) solution. After 15 min incubation at 37 o C, the reaction was ended by adding 50 μ L of 0.1 M NaOH. The optical density (OD) was measured at the wavelength of 405 nm using an enzyme-linked immunoadsorbent assay plate reader (SPECTRAmax 384, Molecular Devices, USA). The total protein content was quantified by the bicinchoninic acid protein assay kit (Beyotime, Shanghai, China). The ALP activity was expressed as OD values divided by the reaction time and total protein quantity (n = 5).
ALP stain assay was also employed to determine the ALP activity histochemically. After fixation of C2C12 cells in 2.5% glutaraldehyde for 10 min on ice, cells were incubated in alkaline phosphatase stain solution (Beyotime, Shanghai, China) for 30 min at room temperature. Stained cells were visualized with inverted light microscope (TE2000U, Nikon Corp., Japan) and quantified using NIH Image J software (NIH, Bethesda, MD, USA).
Gene expression analysis. Total RNA was extracted using Trizol Reagent (Takara, Tokyo, Japan) and concentration was determined by measuring optical absorbance at 260 nm. First strand complementary DNA (cDNA) was synthesized from 2 μ g RNA using PrimeScript RT reagent Kit (Takara, Tokyo, Japan). ALP, Collagen type I (Col I), osteocalcin (OCN) and runt-related transcription factor 2 (Runx2) were evaluated, while GAPDH gene was used as the house keeping gene. All experiments were performed in triplicate. The forward and reverse primer sequences were listed in Supplementary Table S2.
Binding capacity to cell membranes of rhBMP-2. The binding efficiency of rhBMP-2 to cell membranes was assessed using the immunofluorescence assay. C2C12 cells were seeded as mentioned before. Briefly, the cells were fixed with 2.5% glutaraldehyde for 15 min at 4 o C and nonspecific binding sites were blocked with 3% FBS/PBS solution for 1 h at 37 o C. Afterwards, cells were incubated with anti-BMP-2 antibody (R&D systems Inc., Minneapolis, USA) and a dilution of FITC-conjugated goat-anti-mouse IgG (Sigma, St. Louis, USA) to stain BMP-2. Cell nucleus were stained with DAPI solution for 15 min at RT. RhBMP-2 and cell nucleus were measured using confocal microscope at the wavelength of 488 nm and 405 nm, respectively. Quartz Crystal Microbalance (QCM) measurements. The QCM (E1, Q-Sense AB) measurements were performed at 25.0 ± 0.05 o C by real time in situ monitoring of Δ f at 25 MHz in the QCM cell chamber. After PBS filling by an injection of 1.0 mL, the Δ f curves were completely stabilized for 10-20 min. The adsorption was measured after the exchange with 2.0 mL of 50 μ g/mL rhBMP-2 or mixed solution of rhBMP-2/Ca 2+ dispersed in PBS to stabilize the Δ f curves for 80-100 min. The adsorption of the rhBMP-2 receptors after the exchange with 10 μ g/mL BMPRIA, BMPRIB was monitored for 100 min.
Signaling pathway analysis. Western blot assay was employed to analyze signaling pathway of C2C12 cells triggered by rhBMP-2 with elevated [Ca 2+ ]. The cells were extracted by RIPA lysis buffer. Equal amount of proteins (50 μ g) were then subjected to 8% SDS-polyacrylamide gel electrophoresis (PAGE) and then electro-transferred to a polyvinylidene difluoride membrane (PVDF, Bio-Rad, USA).
Far-ultraviolet circular dichroism spectroscopy. The secondary structures of rhBMP-2 affected by calcium ions were detected by far-UV circular dichroism spectroscopy (Model J-715, Tokyo, Japan). Far-UV CD spectra were acquired at the wavelength of 190-260 nm in quartz cell with 0.1 cm path length and a time constant of 0.25 s at a scan rate of 100 nm/min. Three such runs were averaged, calculated net of buffer, and smoothed using the accompanying software program. CDNN Deconvolution software (version 2.1) was employed for estimation of secondary structure content.
Steady-state fluorescence spectroscopy. Steady-state fluorescence spectroscopy measurement was recorded with a spectrofluorometer (RF-5301, Shimadzu, Japan) at 25 o C in a thermostated cuvette holder. The measurements were performed in a 1 cm × 1 cm quartz cuvette with 3 mL of sample. The intrinsic protein fluorescence was monitored between 300 and 450 nm with an excitation wavelength of 281 nm. The spectra were recorded with 0.1 s integration time per 1 nm increment. The excitation and emission slits were set to 0.5 mm and 0.5 mm separately. X-ray photoelectron spectroscopy (XPS) spectra. XPS spectra were carried out on an ESCALAB 250Xi X-ray photoelectron spectrophotometer (Thermo Fisher, USA) with an Al Kα source. The specimens were prepared as follow: 1 cm × 1 cm PVDF membranes were rinsed with water and immediately immersed in the solution of the rhBMP-2 and/or Ca 2+ . After 24 h incubation at 4 o C, the membranes were dried in the vacuum condition overnight.
Ectopic bone formation. This study was performed in strict accordance with the NIH guidelines for the care and use of laboratory animals (NIH Publication No. 85e23 Rev. 1985) and was approved by the Research Center for Laboratory Animal of Shanghai University of Traditional Chinese Medicine. Ectopic bone formation was evaluated using the mouse hindlimb muscle pocket model. Forty female C57BL/6 mice (eight weeks old, Silaike Inc. Shanghai, China) were randomly allocated into week 2 and 4 time points. At each time point, twenty mice received surgery that the gelatin pellets prepared before were surgically implanted into right leg muscle pouches and randomly assigned to one of the following five groups: Group A, free rhBMP-2/gelatin sponge (n = 4); Group B, 0.1 mg CaCl 2 /rhBMP-2/gelatin sponge (n = 4); Group C, 1.0 mg CaCl 2 /rhBMP-2/gelatin sponge (n = 4); Group D, 3.0 mg CaCl 2 /rhBMP-2/gelatin sponge (n = 4); Group E, 10.0 mg CaCl 2 /rhBMP-2/gelatin sponge (n = 4).
At 2 and 4 weeks after surgery, the mice were scanned using a Kodak multimodel imaging system (Carestream Health Inc., USA) with X-ray to evaluate the ectopic bone formation. Then the mice were sacrificed and the implants were harvested. The weights of all these samples were obtained at wet contains. All procedures were carried out under aseptic conditions. The harvested bone specimens were imaged with μ CT (GE Explore Locus SPmicroCT, USA) to determine the new bone volume. The specimens were used for histological evaluation. Briefly, ectopic bone samples were fixed with 4% neutral buffered formalin for 48 h at 4 o C and then decalcified in 12.5% EDTA. Following dehydration in a graded series of alcohol, the specimens were embedded in paraffin and sectioned at 4-mm thickness. Masson/trichrome staining was performed on serial sections and images were taken using inverted light microscope.

Data analysis.
All numerical data were expressed as the mean ± standard deviation. Statistical analysis was performed with one-way analysis of variance (ANOVA). A value of p < 0.05 was considered as statistical significance.