| Sign In to gain access to subscriptions and/or personal tools. |
Amelogenin-guided Crystal Growth on Fluoroapatite Glass-ceramics
1 Department of Preventive and Restorative Dental Sciences, University of California, 707 Parnassus Avenue, D-2260, San Francisco, CA 94143-0758, USA; and Correspondence: * corresponding author, shabeli{at}itsa.ucsf.edu
The formation of aligned fibrous apatite crystals in enamel is predominantly attributed to the involvement of amelogenin proteins. We developed a model to study interactions of matrix proteins with apatite mineral in vitro and tested the hypothesis that amelogenin solubility affects the ability to induce protein-guided mineralization. Crystal growth experiments were performed on fluoroapatite (FAP) glass-ceramics in mineralizing solutions containing recombinant full-length amelogenin (rH174) at different concentrations. Using atomic force microscopy, we observed that mineral precipitated randomly on the substrate, but also formed thin layers (height, 10 nm) on FAP within 24 hrs. This growth pattern was unaffected when 0.4 mg/mL of rH174 was added. In contrast, crystals grew on FAP at a rate up to 20 times higher, at 1.6 mg/mL protein. Furthermore, nanospheres and mineral bound specifically to FAP and aligned in strings approximately parallel to the c-axis of FAP, leading us to the conclusion that amelogenin proteins indeed control direction and rate of growth of apatite in enamel.
Key Words: enamel • amelogenin • biomimetics • apatite • atomic force microscopy
Enamel develops through complex interactions among organic and inorganic components in an aqueous fluid and gradually transforms from a proteinaceous substance into a hard and durable tissue of 95% mineral (Robinson et al., 1998; Smith, 1998; Fincham et al., 1999). Ameloblasts secrete structural proteins and proteases, of which at least 90% are amelogenins, including the full-length protein of 175 amino acids (Fukae et al., 1980; Fincham et al., 1999). The nascent molecule is bipolar, with a hydrophilic carboxyl terminus, but is hydrophobic over most of its length (Snead et al., 1985; Salih et al., 1998). It is thus different from the highly acidic and phosphorylated matrix macromolecules that control biomineralization in bone and dentin (George et al., 1993; MacDougall et al., 1997; Saito et al., 2000). Amelogenin assembles extracellularly into nanospheres of approximately 20 nm and forms a supramolecular structural framework (Fincham et al., 1999). In vivo and in vitro investigations have revealed some putative functions of amelogenins (Robinson et al., 1998; Moradian-Oldak, 2001). Amelogenins inhibit apatite nucleation (Doi et al., 1984; Aoba et al., 1987; Hunter et al., 1999), but appear to direct crystal growth almost exclusively in the c-axis direction, controlling crystal orientation and texture formation. In vitro studies on seeded apatite crystals with the use of extracted or recombinant amelogenin resulted in protein adsorption and altered crystal morphology (Doi et al., 1984; Moradian-Oldak et al., 1998). Wen et al. (1999a, 2000) introduced mineralization studies on bioactive glasses using amelogenin. The presence of up to 50 µg/mL recombinant murine amelogenin during crystallization resulted in the formation of bundles of elongated apatite crystals. Hunter et al.(1999), however, using up to 30 µg/mL recombinant amelogenin, did not observe significant effects in the de novo formation of hydroxyapatite. The ionic environment in which enamel matures appears to be controlled by ameloblasts regulating predominantly the calcium concentration to unusually low levels (Aoba and Moreno, 1987). It is therefore speculated that physical-chemical conditions, e.g., saturation, may affect the ability of mineralizing enamel matrix proteins to interact with inorganic phases. Amelogenin proteins are secreted in high quantities and most likely precipitate in the enamel matrix, since the limit of solubility is only 0.7 mg/mL at physiological pH (Tan et al., 1998). Amelogenin self-assembly into nanospheres and the ability to guide octacalcium phosphate crystal growth are dose-dependent (Wen et al., 2001). This study tested whether amelogenin protein interactions with the forming apatite mineral were affected by protein concentrations above the solubility limit and determined if protein-controlled mineralization on apatite templates is dependent on the crystallographic orientation of the template.
Substrate Preparation A novel glass-ceramic containing rod-like FAP crystals (0.7–1 µm wide, 5–10 µm long) served as substrates for the study of amelogenin-mineral interactions (Moisescu et al., 1999a; Hoche et al., 2001). This glass-ceramic is unique due to the uni-axial alignment of FAP crystals obtained by a high-temperature extrusion process (Moisescu et al., 1999b). We prepared the substrates by sectioning and polishing (0.1 µm, diamond paste) the extruded glass-ceramic rods. Sections perpendicular or parallel to the extrusion axis exposed predominantly (001)- or (hk0)-planes of FAP, respectively. A (001)- crystal plane of FAP (dark) appeared 4–8 nm below the glass level (bright) (Figs. 1a  , 1b). These substrates are flat enough to reveal assembled proteins by AFM, similar to glass or mica substrates (Wallwork et al., 2001; Wen et al., 2001) and allow for in situ studies of biomimetic mineralization, since apatite is the mineral phase in most vertebrate calcified tissues. Furthermore, the height difference between glass matrix and the FAP crystals serves as a reference for the precise measurement of crystal growth.
Calcium Phosphate and Protein Solutions Thus far, in vitro studies on protein-guided mineralization have been performed at calcium and phosphate concentrations significantly elevated from those in in vivo conditions. We wanted to replicate biological conditions and therefore used human recombinant full-length amelogenin in an ionic environment similar to the fluid of developing enamel (Aoba and Moreno, 1987). The recombinant full-length human amelogenin, rH174, was synthesized by Escherichia coli bacteria, as described previously (Li et al., 2003). The protein was dissolved in 0.1% trifluoroacetic acid (TFA) at a concentration of 2.2 mg/mL. We prepared protein solutions by adding 50 mM Tris total and adjusting the pH to 7.4 using HCl, then diluting to concentrations of 1.6 and 0.4 mg/mL. Mineralizing solutions were prepared from Ca(NO3)2, KH2PO4, and KCl and dissolved in double-de-ionized water. The concentrations of Ca2+ and PO43- were 0.5 and 2.5 mM, respectively, chosen according to Aoba and Moreno’s (1987) finding of total ion concentration in the enamel fluid of secretory-stage porcine enamel. The buffered calcium and phosphate solutions were prepared and stored separately at eight-fold concentrations in 600 mM Tris/HCl at pH 7.4 (± 0.1) at 37°C. KCl was added at 1200 mM (eight-fold) to the Ca2+ solution only. The degree of saturation with respect to hydroxyapatite (pK = 58.6) was 7.55 (Larsen, 2001).
Crystal Growth Experiments
AFM Imaging and Raman Spectroscopy We used micro-Raman spectroscopy with a 20-mW He-Ne laser at a wavelength of 632.8 nm (HR 800, Jobin Ivon, Horiba Group, Tokyo, Japan) through an optical lens at 50X magnification to obtain spectra in the range from 600 to 1800 cm–1.
The topography of the glass-ceramic substrate is shown in Fig. 1a  . The polishing procedure resulted in highly flat substrates with FAP crystals at a level of about 5 nm below the surrounding glass matrix (Fig. 1b). Substrates immersed in calcium- and phosphate-free solutions containing 0.4 mg/mL rH174 revealed an almost complete monolayer of nanospheres adhering non-specifically to both FAP crystals and glass matrix at 1 hr (Fig. 1c). The number of nanospheres increased over time. High-resolution images of the spheres at 24 hrs of immersion (Fig. 1d) exhibited diameters between 20 and 30 nm, characteristic of amelogenin (Fincham et al., 1999). When a protein concentration of 1.6 mg/mL was used, protein precipitated within 5 to 30 min of pH adjustment, resulting in an opaque gel, presumably similar to that observed from enamel matrix extracts (Wen et al., 1999b). Imaging of the FAP substrates showed larger spherical aggregates of about 150–250 nm, randomly distributed across the surface (Fig. 1e). High-resolution imaging did not show formation and adherence of characteristic 20-nm spheres at this protein concentration (Fig. 1f).
When the FAP substrates were immersed in CaP-sol without proteins, mineral precipitated sparsely and randomly as 15- to 20-nm spheres on the surface. Furthermore, a thin layer of mineral formed on the (001)-planes of FAP crystals, which continuously increased in height until they overgrew the surrounding glass matrix by about 3 to 5 nm at 24 hrs (Fig. 2a
When rh174 at 0.4 mg/mL was added to CaP-sol, formation of nanospheres and random precipitation of mineral covering the entire surface were observed. In agreement with the protein-free solution, FAP crystals grew above the glass-layer by a few nanometers when 0.4 mg/mL rH174 was added (Fig. 2b ). In contrast, the growth pattern changed when substrates were immersed in 1.6 mg/mL rH174 mixed with CaP-sol. Protein gelation was observed within 5 to 30 min. AFM imaging of the substrate revealed a significant increase in height of FAP crystals. (001)- planes of FAP crystals grew to an average height of 191 ± 73 nm above the glass matrix (Fig. 2c). The hexagonal shape of the crystal was still recognizable; however, the crystal edges were more irregular and diffuse, indicating the incorporation of an organic phase. Amelogenin nanospheres were not observed on these substrates. It appears that the reaction of the solution with the substrate was restricted to the FAP crystals. AFM revealed that the layers formed on FAP consisted of elongated structures of about 50 nm width that did not appear to have a strong preferred orientation or texture (Fig. 2d). This growth pattern occurred on all FAP crystals over the entire surface, with minor variations in FAP-layer height (Fig. 2e).
As a control, substrates were immersed for 24 hrs in a BSA-solution mixed with CaP-sol. In comparison with the protein-free CaP-sol, the growth of layers on FAP appeared to be reduced to less than a 3-nm increase in height when BSA was added (Fig. 2f
Glass-ceramic substrates that exposed predominantly (hk0)- surfaces (Fig. 3a
Micro-Raman spectra of a series of FAP substrates were compared (Fig. 4 ). The sharp peak at 963 cm–1 is characteristic of the PO43- groups in apatites and is present before (curve a) and after the mineralization treatments (curves b and c) (Penel et al., 1998). No alterations of the spectra were observed when substrates were exposed to CaP-sol only, or to CaP-sol with the addition of 0.4 mg/mL rH174 (curve b). However, a series of new peaks appeared when the high concentration of rH174 was used during mineralization. The spectrum exhibited bands at 708 and 1384 cm–1, indicating that CO32- ions were incorporated into the apatite lattice. Furthermore, bands around 1450 and 1670 cm–1 were observed and attributed to bending modes of C-H and N-H groups of the amelogenin protein, respectively (Zheng et al., 1987; Penel et al., 1998).
It is well-established that the full-length amelogenin protein assembles into nanospheres, with a variety of diameters, depending on the pH, as has been shown by AFM and dynamic light-scattering (Moradian-Oldak et al., 2000; Wen et al., 2001). Our observation of the formation and adhesion of 20- to 30-nm nanospheres confirms these findings for a concentration of 0.4 mg/mL recombinant amelogenin at pH 7.4. Nanospheres adhered non-specifically to the glass matrix as well as to the FAP crystals, covering the substrate as a monolayer within 1 hr (Fig. 1b ), and gradually increased in density over time (Fig. 1c) (Moradian-Oldak, 2003). The full-length amelogenin protein has a relatively low solubility limit at 0.7 mg/mL (Tan et al., 1998). This concentration is presumably exceeded in vivo, and amelogenin proteins therefore precipitate in the developing enamel as a gel (Wen et al., 1999b). We do not know the effect of amelogenin precipitation on protein folding, assembly, and function. This study shows, however, that the interaction of amelogenin with the forming mineral can be drastically influenced by protein concentration and precipitation. Once precipitation of 1.6 mg/mL rH174 was induced by a pH change, characteristic amelogenin nanospheres were no longer observed on the substrates. Instead, larger spherical aggregates with a diameter of about 150 nm adhered to the substrate randomly at a low density. It is unclear if these larger spheres comprise the main portion of the amelogenin precipitate.
When the substrates were immersed in calcium phosphate solutions, we observed that, despite the relatively low degree of saturation, mineral formation was induced. Both heterogeneous nucleation and precipitation of nanoparticles, as well as growth of the FAP crystals to levels above the glass level, were observed. We assume that the latter was a result of homogeneous nucleation of apatite on FAP. In contrast, homogeneous nucleation was not clearly identified on (hk0)- planes of FAP (Fig. 3b While the addition of 0.4 mg/mL rH174 to the mineralizing solution had no significant effect on mineral precipitation and crystal growth, growth patterns were strongly altered when 1.6 mg/mL rH174 were supplemented. Layers grown on (001)- planes of FAP were, on average, about 20 times higher than those at low rH174 concentrations, indicating that the precipitated form of rH174 interacted differently with the forming mineral and accelerated crystal growth. The mechanisms for increased crystallization rates are unknown, but this study showed that the affinity of rH174 to apatite increased strongly when the protein precipitated in a supersaturated calcium phosphate solution. Under these conditions, reactions of the mineralizing solution became specific to the FAP crystals of the substrate. Only minor precipitation was observed on the surrounding glass. The origin of fibrous enamel crystals has been attributed to the ability of amelogenin proteins to interact with specific crystal planes. Kirkham et al.(2000) showed that certain domains exist on extracted enamel crystals, which may relate to specific binding sites for amelogenin proteins, as reported by Wallwork et al.(2001). A current model (Robinson et al., 1998; Fincham et al., 1999) suggests that the bound protein blocks growth perpendicularly to the c-axis, preventing widening and early crystal fusion. Given the limitations of in vitro experiments, this study provided evidence for this mechanism, since nanoparticles of about 40 to 60 nm aligned along the c-axis of FAP, indicating specific binding sites for assembled amelogenin proteins on apatite. Thus, the assembled proteins formed short strings that may protect the crystal from lateral growth and could indeed provide the guidance required to produce a fibrous and aligned apatite crystal as observed in vivo, which is further enhanced by the amelogenin-induced accelerated growth of apatite parallel to the c-axis. However, this mechanism appears to be effective at high protein concentrations only. Future studies will require the addition of other matrix proteins, as well as proteolytic degradation of amelogenin by proteinases, to establish a more complete model of enamel formation on the molecular level and, eventually, to create artificial enamel in vitro by a biomimetic approach.
The authors thank Drs. C. Rüssel and C. Moisescu, Otto-Schott-Institute, Friedrich-Schiller-University, Jena (Germany), for providing the glass-ceramic substrates, and Dr. J.D.B. Featherstone, University of California, San Francisco, Department of Preventive and Restorative Dental Sciences, for advice on mineralization experiments. This work was supported by NIH/NIDCR P01-DE09859, R21-DE015416, R01-DE13029, and UCSF-DDCF 12. Received for publication February 5, 2003. Revision received May 27, 2004. Accepted for publication June 24, 2004.
Journal of Dental Research, Vol. 83, No. 9,
698-702 (2004)
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION
