ARTICLE pubs.acs.org/Langmuir
Cooperative Calcium Phosphate Nucleation within Collagen Fibrils Diana N. Zeiger,† William C. Miles,† Naomi Eidelman,‡,§ and Sheng Lin-Gibson*,† †
Polymers Division and ‡Paffenbarger Research Center, American Dental Association Foundation, National Institute of Standards and Technology, Gaithersburg, Maryland 20899-8543, United States ABSTRACT: Although “chaperone molecules” rich in negatively charged residues (i.e., glutamic and aspartic acid) are known to play important roles in the biomineralization process, the precise mechanism by which type I collagen acquires intrafibrillar mineral via these chaperone molecules remains unknown. This study demonstrates a mechanism of cooperative nucleation in which three key components (collagen, chaperone molecules, and Ca2þ and PO43) interact simultaneously. The mineralization of collagen under conditions in which collagen was exposed to pAsp, Ca2þ, and PO43 simultaneously or pretreated with the chaperone molecule (in this case, poly(aspartic acid)) before any exposure to the mineralizing solution was compared to deduce the mineralization mechanism. Depending on the exact conditions, intrafibrillar mineral formation could be reduced or even eliminated through pretreatment with the chaperone molecule. Through the use of a fluorescently tagged polymer, it was determined that the adsorption of the chaperone molecule to the collagen surface retarded further adsorption of subsequent molecules, explaining the reduced mineralization rate in pretreated samples. This finding is significant because it indicates that chaperone molecules must interact simultaneously with the ions in solution and collagen for biomimetic mineralization to occur and that the rate of mineralization is highly dependent upon the interaction of collagen with its environment.
’ INTRODUCTION Mineralized tissues such as bones and teeth have hierarchical structures, and when these tissues are damaged, they may require grafts to aid healing. Although effective, grafts have significant drawbacks such as availability, pain at the harvest site, and infectious disease transfer. Thus, synthetic grafts are desirable for the treatment of damaged or diseased mineralized tissues. However, because of the complex structures of these tissues, the formation of synthetic grafts is a problem with significant materials science challenges.1 Current approaches have sought to mimic biological processes to form these structures. However, despite decades of research, the precise mechanism by which bones and teeth mineralize has yet to be elucidated. A complete understanding of this mechanism has profound implications for treatments for diseases of mineralized tissue, including osteoporosis, caries, and osteogenesis imperfecta. The key components of bone and dentin are similar and consist of type I collagen and hydroxyapatite2 assembled such that crystals of apatite are positioned within the collagen fibrils (intrafibrillar mineral) with their c axis aligned along the fibrils.35 The mineral is aligned in such a fashion that the degree of mineralization is not the only factor responsible for the strength and toughness of mineralized tissues. Indeed, intrafibrillar mineralization has been found to be more important for the mechanical properties of bone6 and dentin than overall mineral content.7 Chaperone molecules (generally proteins bearing regions rich in acidic moieties) appear to be necessary to the biomineralization process.8,9 These proteins belong to a group known as the SIBLING (small integrin-binding, N-linked glycoprotein) family.10 Carboxylate-rich regions on such SIBLING proteins as dentin and bone sialoprotein have been shown to facilitate the r 2011 American Chemical Society
mineralization of collagen11 and remineralize dentin.12 Acidic macromolecules such as poly(acrylic acid) (pAA) have been shown to induce the formation of intrafibrillar calcium carbonate,13 and poly(aspartic acid) (pAsp) has been shown to stimulate the formation of intrafibrillar calcium carbonate14 and calcium phosphate.15,16 Several mechanisms or partial mechanisms have been proposed within the last several years.11,14,17,18 Recent work by Sommerdijk et al. utilized cryo-TEM to provide significant insight into the mechanism of biomineralization.19,20 These studies have shown the importance of prenucleation clusters in the biomineralization process.20 Additionally, results from cryo-TEM studies convincingly demonstrated the specific region on the collagen fibril at which the infiltration of pAspmediated prenucleation clusters occurs.19 The infiltration process by stabilized ion clusters is proposed to be charge-driven. This research was designed to isolate the key components of the mineralization process and to understand how the interaction between collagen and the chaperone molecule affects mineralization. This was achieved by varying the order of addition of the mineralization components (i.e., collagen, Ca/P, pAsp). We utilized SEM and FTIR reflectance microspectroscopy (FTIRRM) to determine the type, quantity, distribution, and phase of the minerals formed on and in collagen. We also used a model heterobifunctional compound to probe the interaction between COOH and collagen. We show that through pretreatment of the collagen fibers with pAsp, intrafibrillar mineralization can be slowed or even eliminated. This indicates that the interaction of Received: February 11, 2011 Revised: May 31, 2011 Published: June 09, 2011 8263
dx.doi.org/10.1021/la201361e | Langmuir 2011, 27, 8263–8268
Langmuir
ARTICLE
Table 1. Conditions Examined for Determining the Mineralization Mechanism
collagen with the chaperone molecule and the interaction of the chaperone molecule with the solution ions are both critical to achieving intrafibrillar mineralization.
’ MATERIALS AND METHODS Preparation of Collagen Gels and Thin Films. For crosslinked collagen gels, the self-assembly of collagen fibrils was initiated by adding 1 part of 10X phosphate-buffered saline (PBS, Gibco Dulbecco’s Phosphate Buffered Saline, Invitrogen, Carlsbad, CA) to 8 parts of PureCol collagen solution (Advanced BioMatrix, San Diego, CA) and raising the pH to 7.4 to 7.6. The mixture was allowed to gel in an incubator at 37 °C for 4 h. The gelled collagen was concentrated by centrifugation, spinning the bulk material twice for 20 min at 3220g, and holding the temperature at 10 °C. Excess liquid was removed, and the final volume of collagen gel was ∼67% of the starting gel. The concentrated gel was then transferred to a single-well tissue culture plate with a layer thickness of ∼3 mm, cross-linked with 0.2% glutaraldehyde (grade 1 aqueous 70%, Sigma, St. Louis, MO), rinsed thoroughly with deionized water, and cut into 1 cm 1 cm squares. The gels were stored overnight in individual wells of tissue-culture plates in PBS at 4 °C and washed three times with deionized water immediately prior to immersion in mineralizing solution. Collagen thin films were prepared using an established protocol.21 Briefly, a neutralized PureCol solution was diluted to a final concentration of 300 μg/mL with 1X PBS. The solution was pipetted into individual wells of a six-well tissue culture plate and allowed to gel overnight in an incubator. Bulk gel was removed by suction, and the remaining collagen film was washed with PBS and deionized water. The collagen films were stored in PBS at 4 °C. Mineralization of Collagen Gels. HEPES-buffered saline (20 mM HEPES, g99.5%, Sigma, St. Louis, MO; 150 mmol/L NaCl, 99.7%, Fisher, Pittsburgh, PA; HBS) containing 5 mmol/L CaCl2 (approximately 97%, anhydrous, Sigma, St. Louis, MO) and 2.5 mmol/L K2HPO4 (99.9%, Fisher, Pittsburgh, PA) (HBS-m) was used, with various additives, to mineralize the collagen gels. Gels were incubated at 37 °C for 4 h and 3 days (solution changed daily) in HBS-m alone or in HBS-m containing 75 μg/mL poly(aspartic acid) (MP Biomedicals, LLC, Solon, OH; pAsp). Pretreatment of gels was accomplished by incubating them overnight at 37 °C in HBS containing 75 μg/mL pAsp and then washing three times with deionized water prior to immersion in mineralizing solution. Scanning Electron Microscopy (SEM) and Image Analysis. After thorough air drying, samples were sputter coated with an ultrathin layer of gold (Desk II, Denton Vacuum) and visualized on a Hitachi S4700 FE-SEM. The accelerating voltage was 5 kV (collagen control) or 15 kV (mineralized gels). The fibril diameter was evaluated using the Measure and Label plugin for ImageJ. For each sample type, at least 15 different fibrils on each of at least 3 images (magnification = 10 000) were measured. The measurements were pooled by sample type and analyzed by one-way analysis of variance (ANOVA) with a 95% confidence interval.
Fourier-Transform Infrared Microspectroscopy in Reflectance Mode (FTIR-RM). Air-dried mineralized collagen gels were pressed between aluminum plates to flatten them, producing nearly
identical surface topography among samples. Samples were mounted flat on glass or Al slides. The FTIR-RM characterization was performed on the surface of the specimens using a Nicolet Continuμm FT-IR microscope (Thermo Scientific) operated in reflectance mode and interfaced to a Nicolet 6700 FT-IR spectrometer. The microscope is equipped with a color video camera, two liquid-nitrogen-cooled mercury cadmium telluride detectors (MCT-A, 11 700 to 650 cm1 and MCT-B, 11 700 to 400 cm1), and a computer-controlled motorized stage programmable in the x and y directions. The FTIR-RM maps were obtained over the entire surface of the specimens in a grid pattern22 with a spatial resolution of 90 μm 90 μm to 150 μm 150 μm in the 650 to 4000 cm1 region with 8 cm1 spectral resolution and 64 scans per spectrum. Spectra were also obtained from distinct mineralized spots in the specimens with the MCTB detector (150 μm 150 μm spot size, spectral resolution of 8 cm-1, 5000 to 20 000 scans per spectrum) in order to detect clearly the small ν4PO4 bands at ∼560 and 605 cm1. The reflectance spectra were proportioned against a gold-coated glass slide background and transformed to absorbance spectra using the Kramers Kronig algorithm.23 The FTIR-RM maps were processed as mineral and collagen maps (area under the ν3 PO4 bands in the 980 to 1130 cm1 range and under the amide I band in the 1600 to 1700 cm1 range), along with their ratios (mineral to collagen), and were displayed as color contour maps along with the visual maps that were obtained with the same FTIR microscope. Spectra that were obtained from distinct mineralized spots with a greater number of scans with the MCTB detector were compared to bone and apatite spectra that were obtained from flat, solid specimens with the same microscope and mode for identifying and characterizing the mineral phase that formed in the collagen gels. Collagen-Binding Assay. Collagen thin films were incubated in a 100 μg/mL solution of heterobifunctional poly(ethylene glycol) (PEG), terminated at one end with a fluorescein isothiocyanate (FITC) and at the other with a carboxylate group (FITC-PEG-COOH; NanoCS, Inc., New York, NY, USA), in PBS for 24 h. The thin films were exposed to the polymer alone for 24 h or were pretreated with 75 μg/mL pAsp in PBS and then exposed to the polymer for up to 72 h. Samples were rinsed three times with PBS and imaged using a Zeiss LSM 510 confocal laser scanning microscope with an Achroplan IR 40X/0.80 W waterimmersion objective. The samples remained fully hydrated throughout all experiments and visualization. For selected experiments, poly(aspartic acid) was modified through the general esterification reaction to convert the carboxylic acid functionalities to methyl ester functionalities. A 150-fold excess of methanol (99.9%, Aldrich) and concentrated HCl (37%, Aldrich) were used to drive the equilibrium reaction. The formation of the methyl ester was monitored via 1H NMR to determine the degree of conversion.
’ RESULTS AND DISCUSSION The order of addition of each of the mineralization components was varied to gain insight into the intrafibrillar mineralization mechanism. Table 1 illustrates the conditions tested for this study: (A) collagen control, (B) CaCl2 and K2HPO4 only, (C) CaCl2, K2HPO4, and pAsp introduced into collagen simultaneously, (D) CaCl2 and K2HPO4 with pAsp adsorbed on collagen, 8264
dx.doi.org/10.1021/la201361e |Langmuir 2011, 27, 8263–8268
Langmuir
ARTICLE
Figure 1. SEM images of collagen gels that were mineralized for three days in HEPES buffer under the conditions listed in Table 1. Scale bar = 1 μm.
Figure 2. Fibril diameters for conditions A, C, and E were measured using ImageJ software. Values shown in the upper right-hand corner of each box represent the diameter mean plus or minus one standard deviation, which is taken as an estimate of the standard uncertainty.
and (E) CaCl2, K2HPO4, and pAsp introduced simultaneously with pAsp adsorbed on collagen. Scanning electron microscopy was used to visualize the mineralization under the different conditions described above. Water was completely (or nearly completely) removed during the SEM experiments; studies have strongly suggested that water within the collagen fibrils is gradually replaced by mineral as the collagen is mineralized.24 We therefore compared the difference
in fiber diameter to assess the degree of mineralization. SEM images (Figure 1) show clear differences in morphology depending on the mineralization condition. Collagen control (A) showed the expected fibrillar structure. Clusters of needle-shaped crystalline mineral formed directly on the collagen gels exposed to CaCl2 and K2HPO4 only (B), with collagen fibrils visible under the CaP; the fiber diameter is nearly identical to that of the collagen control. No crystalline deposits are visible outside the collagen fibrils when pAsp is present in the mineralizing solution (conditions C and E). For condition C, the fibrils are substantially greater in diameter and more defined than those of the collagen control samples, indicating the formation of intrafibrillar mineralization. These results are consistent with previously demonstrated intrafibrillar mineralization validated by TEM under similar mineralization conditions.15,16 Condition E shows a thickening of some of the fibers, indicative of mineralization; however, smaller fibers were also present, which is characteristic of unmineralized collagen. Collagen that had pAsp adsorbed on the surface but did not have pAsp in the mineralizing solution (condition D) showed globular mineral deposits on the surface of the sample. The pronounced difference in morphology produced under condition D indicates that pAsp had either been adsorbed onto (via chemical interaction) or trapped within (via physical interaction) the collagen during the pretreatment. Had the pAsp not been adsorbed/trapped, the mineralized sample would have assumed the morphology of condition B. Furthermore, these results support the notion that pAsp plays a critical role in mediating the intrafibrillar mineralization in solution. Indeed, the observation that the mineral uniformly coats the collagen surface indicates that mineral nucleation does not appear to differentiate between collagen and noncollagen surfaces. Collectively, these results suggest that collagen, CaCl2, K2HPO4, and pAsp must interact in solution and work cooperatively for intrafibrillar mineralization to occur. Collagen fibril diameters were determined on the basis of SEM images to assess the extent of mineralization (Figure 2). Collagen that had pAsp adsorbed onto the surface (condition D) could not be analyzed because individual fibrils were not visible to allow fibril size analysis. The control collagen (condition A) had an average fibril diameter of 85 nm. Collagen exposed to all three components (CaCl2, K2HPO4, and pAsp simultaneously (condition C)) had an average fibril diameter of 128 nm at 4 h and 181 nm at 72 h (all significantly different at p e 0.05). For collagen with surface-adsorbed pAsp and exposed to all three 8265
dx.doi.org/10.1021/la201361e |Langmuir 2011, 27, 8263–8268
Langmuir
Figure 3. (A) FTIR-RM spectra of collagen mineralized under the previously discussed conditions. (B) FTIR-RM maps are also provided for these conditions: maps indicate the ratio of calcium phosphate to collagen (scale at right) and are adjusted to the minimum significant value obtained for the pretreated samples incubated in mineralizing solution with pAsp.
components simultaneously for 72 h (condition E), there was an increase in fibril diameter compared to that of the collagen control (156 nm), which is evidence of intrafibrillar mineralization. However, the diameter distribution was substantially broader than condition C and skewed toward the lower values, indicative of uneven mineralization (i.e., only a fraction of the fibrils mineralized and therefore increased in diameter). Lessmineralized collagen fibrils that resemble the collagen control were also present in significant quantities. These results suggest that the mineralization kinetics have been slowed through the surface adsorption of pAsp but not completely eliminated as long as there is free pAsp in solution with the other mineralization components. FTIR-RM was used to determine the chemical composition, relative quantity, and mineral phase information for the top several micrometers of the sample (Figure 3A). FTIR-RM spectra were collected over the entire sample to generate maps containing information regarding the relative chemical content (Figure 3B). Specifically, the relative ratio of phosphate to collagen was computed to determine the extent of surface mineralization and
ARTICLE
the distribution of mineralization over the entire sample. FTIRRM spectra of hydroxyapatite (HA) and bone are included for comparison. HA and bone show the expected ν4PO4 peaks at 560 and 600 cm1 and the ν3PO4 main peak at 1030 cm1; the bone spectrum shows an additional set of peaks at 1540 and 1660 cm1 corresponding to the amide I and II regions of the collagen, respectively. A representative FTIR-RM spectrum collected at a single location for the collagen control (condition A) shows pronounced peaks in the expected amide I and II regions. As shown by the FTIR-RM maps of the mineralized samples, the mineral-to-collagen ratio is quite heterogeneous. For comparison purposes, an FTIR-RM spectrum for a location under condition B with extensive mineralization (which appears red on the maps) is illustrated in Figure 3A. This spot showed strong phosphate peaks relative to the amide bands, indicating a high concentration of apatite on the surface relative to collagen. The phosphate peaks were in the same general region and had the same general shape as the bone and apatite standards, consistent with the hydroxyapatite phase. (The PO4 peak positions of all mineralized collagen samples were slightly shifted from those found for synthetic apatite and bone. These peak shifts may be a result of the dispersion correction of the reflectance spectra for rougher samples.) Furthermore, FTIR-RM maps for samples prepared under condition B contained areas rich in either collagen or mineral, consistent with random mineral precipitation rather than the orderly mineralization of collagen fibrils. A representative spectrum of condition C (selected from a green area of the map) shows the phosphate peaks in the same region as those of bone as well as pronounced peaks in the amide I and II ranges. The spectra were relatively consistent over the entire sample surface, as indicated by the FTIR-RM maps, supporting the notion that the mineral is colocalized with the collagen. This colocalization phenomenon was not observed for condition D (pretreated collagen). The spectra for this sample are similar to those measured for condition B, suggesting that mineralization is essentially random. The FTIR map showed highly heterogeneous phosphate-to-collagen ratios, similar to those under condition B, although the exact values of the phosphate-to-collagen ratios are quite low, indicating that surface-adsorbed pAsp has hindered specific mineralization at the collagen fibrils. This also suggests that the globular mineral structures seen by SEM are likely present only on the surface. Spectra of condition E are similar to those for condition C and show pronounced amide I and II peaks as well as phosphate peaks, suggesting colocalization. However, the phosphate peaks for this condition show significant peak broadening, indicative of the formation of lessmature hydroxyapatite. This, in conjunction with the lower overall phosphate-to-amide ratio as well as the observed smaller average fibril size and wider fibril distribution for condition E, further indicates that the surface adsorption of pAsp on collagen slows the mineralization kinetics. In the spectra for conditions D and E, a small peak is present at about 1750 cm1 in spectrum D and a larger one is present at the same position in spectrum E, suggestive of the presence of adsorbed pAsp. Our use of the MCT-B detector for performing FTIR-RM was particularly important because its range (11 700 to 400 cm1) is beyond that of the traditional MCT-A detector. We were therefore able to observe the “fingerprints” of hydroxyapatite (the ν4PO4 peaks at ∼604 and ∼563 cm1). A comparison of conditions B and E raises the question as to why the adsorption of pAsp onto the collagen surface might retard the mineralization kinetics. We hypothesized that pAsp 8266
dx.doi.org/10.1021/la201361e |Langmuir 2011, 27, 8263–8268
Langmuir
Figure 4. Representative confocal laser scanning microscopy images of (A) collagen fibers collected under reflectance, (B) collagen thin films incubated with FITC-PEG-COOH for 24 h, (C) pAsp-pretreated collagen thin films incubated with FITC-PEG-COOH for 24 h, and (D) pAsp-pretreated collagen thin films incubated with FITC-PEGCOOH for 72 h. Scale bar = 10 μm.
binds to collagen through a specific interaction of COOH on pAsp with collagen, which then prevents or drastically reduces the adsorption of pAsp in solution (that has been interacting with calcium and/or phosphate ions) to the collagen surface. To test this, a fluorescently tagged PEG terminated with a carboxylic acid on one end and a fluorescein on the other end (FITC-PEGCOOH) was used to label thin films of fibrillar collagen. We used thin films rather than bulk gels to be able to view individual fibrils; the method we chose for the preparation of the films specifies that the thickness of the resulting films is consistent with that of a collagen monolayer.21 Figure 4B shows fluorescence following the collagen fibril (a reflectance image of a fibril shown in Figure 4A), indicating specific binding between FITC-PEG-COOH and the collagen fibrils. To verify that COOH is in fact responsible for the binding, COOH was esterified to form a methyl ester (COOCH3), and the binding of this molecule was determined. In the case of the esterified polymer, no fluorescence was observed even though the fluorescence of the polymer was not affected by the esterification reaction (image not shown), confirming that COOH is the active binding site with collagen. The same fluorescent PEG molecule was used to assess the binding strength and desorption kinetics of COOH/collagen binding. Figure 4C,D shows the binding of FITC-PEG-COOH to fibrils with pAsp adsorbed onto the surface after 24 and 72 h, respectively. As described earlier, collagen/FITC-PEG-COOH (without pAsp adsorbed to its surface) shows substantial fluorescence. For collagen pretreated with pAsp, essentially no fluorescence was detected after 24 h of incubation with FITCPEG-COOH, indicating that the FITC-PEG-COOH polymer is initially unable to bind to the pretreated collagen surface; however, after 72 h of incubation, the collagen shows some slight fluorescence. These results suggest that that the binding between COOH and collagen is relatively strong and that the desorption process, although possible, occurs very slowly.
ARTICLE
Our results provide strong evidence that the rate of mineral nucleation is controlled by the rate of adsorption of the stabilized prenucleation cluster to the collagen surface whereas the overall mechanism is one of cooperative nucleation for pAsp-mediated intrafibrillar mineralization. The nature of the COOH/collagen interaction is likely to be driven by charge, as proposed by Sommerdijk et al.19 Although this article showed that collagen is indeed an active part of the mineralization process, providing site-specific cues in the form of regions of favorable charge for the entry of stabilized prenucleation clusters, our work gives additional insight into the strength and reversibility of this interaction. Because the rate of adsorption is retarded by the pretreatment of the collagen surface with pAsp, the mineralization kinetics slowed considerably. This explains why mineralization condition E showed smaller fibril diameters and less-mature apatite than did condition C. In other words, the pAsp in solution had to move to the collagen surface to induce mineralization; the relevant spaces were already occupied by pAsp from the pretreatment procedure, thereby markedly reducing the rate at which the solution pAsp could reach the collagen. Our results also indicate that the desorption of pAsp from the collagen surface is a slow process, which explains the difference in morphology and the lack of colocalization of the collagen and mineral for condition D. In this case, it is likely that pAsp took so long to desorb from the surface that the mineral formed randomly (because of precipitation under supersaturated conditions) rather than at the specific sites at which pAsp had adsorbed. For the design of synthetic mineralized tissues, the concentration of the chaperone molecule in relation to the calcium and phosphate ions is critical to controlling the mineralization rate. Additionally, because this is an adsorption/desorption phenomenon, the molecular mass of the chaperone molecule is expected to play a significant role. Specifically, we would expect the use of smaller molecular mass chaperone molecules to translate to more mineralization because of an increased ability to bind to the collagen surface. Indeed, a decrease in mineralization has been observed with increasing pAsp molecular weight in the literature.25 These conclusions have far-reaching implications for developing bone and tooth mimetics.
’ CONCLUSIONS Taken in concert, the results presented in this study provide critical insights into the mechanism of pAsp-mediated intrafibrillar mineralization of collagen with calcium phosphate. The overall mechanism is one of cooperative nucleation in which all three mineralization components (collagen, Ca2þ/PO43, and chaperone molecule) must be combined in solution for intrafibrillar mineralization to occur. This cooperative mechanism involves specific binding between collagen and COOH from the pAsp molecule; this was demonstrated by adsorption studies via a model fluorescently tagged FITC-PEG-COOH polymer. Our results provide additional evidence to support the notion that charge is the driving force for prenucleation cluster/collagen interaction as proposed by Sommerdijk et al.19 Our adsorption studies further demonstrate that collagen surface adsorption is the rate-limiting step; the desorption of pAsp from the collagen surface is relatively slow and can affect the overall mineralization process. These conclusions have far-reaching implications for developing bone and tooth mimetics where the location specificity and kinetics can be tuned on the basis of a known mineralization mechanism. 8267
dx.doi.org/10.1021/la201361e |Langmuir 2011, 27, 8263–8268
Langmuir
’ AUTHOR INFORMATION Corresponding Author
*Phone: (þ1) 301-975-6765. Fax: (þ1) 301-975-4977. E-mail:
[email protected]. Present Addresses §
Current address: Environmental Autoimmunity Group, National Institute of Environmental Health Sciences, National Institutes of Health, Bethesda, Maryland 20892, United States.
’ ACKNOWLEDGMENT This work is supported by NIDCR/NIST Interagency Agreement Y1-DE-7005-01. This is a publication of the National Institute of Standards and Technology (NIST), an agency of the U.S. Government, and by statute is not subject to copyright in the United States. D.N.Z. and W.C.M. acknowledge the support of the National Research Council Research Associateship Program. The FTIR microscopy system that was used in this work was purchased with NIH Shared Instrumentation Grant 1S10RR22650-01A2 awarded to N.E. We thank Dr. John Elliott for discussing collagen chemistry and preparation. Certain equipment, instruments, or materials are identified in this article in order to specify the experimental details adequately. Such identification does not imply a recommendation by the National Institute of Standards and Technology nor does it imply that the materials are necessarily the best available for the purpose
ARTICLE
(19) Nudelman, F.; Pieterse, K.; George, A.; Bomans, P. H.; Friedrich, H.; Brylka, L. J.; Hilbers, P. A.; de With, G.; Sommerdijk, N. A. Nat. Mater. 2010, 9, 1004–1009. (20) Dey, A.; Bomans, P. H.; Muller, F. A.; Will, J.; Frederik, P. M.; de With, G.; Sommerdijk, N. A. Nat. Mater. 2010, 9, 1010–1014. (21) Elliott, J. T.; Tona, A.; Woodward, J. T.; Jones, P. L.; Plant, A. L. Langmuir 2003, 19, 1506–1514. (22) Eidelman, N.; Raghavan, D.; Forster, A. M.; Amis, E. J.; Karim, A. Macromol. Rapid Commun. 2004, 25, 259–263. (23) Chalmers, J. M.; Everall, N. J.; Ellison, S. Micron 1996, 27, 315–328. (24) Liu, Y.; Kim, Y. K.; Dai, L.; Li, N.; Khan, S. O.; Pashley, D. H.; Tay, F. R. Biomaterials 2011, 32, 1291–1300. (25) Jee, S. S.; Thula, T. T.; Gower, L. B. Acta Biomater. 2010, 6, 3676–3686.
’ REFERENCES (1) Weiner, S.; Wagner, H. D. Annu. Rev. Mater. Sci. 1998, 28, 271–298. (2) Glimcher, M. J. Rev. Mineral. Geochem. 2006, 64, 223–282. (3) Landis, W. J.; Hodgens, K. J.; Arena, J.; Song, M. J.; McEwen, B. F. Microsc. Res. Tech. 1996, 33, 192–202. (4) Landis, W. J.; Song, M. J.; Leith, A.; Mcewen, L.; Mcewen, B. F. J. Struct. Biol. 1993, 110, 39–54. (5) Fratzl, P.; Gupta, H. S.; Paschalis, E. P.; Roschger, P. J. Mater. Chem. 2004, 14, 2115–2123. (6) Cointry, G. R.; Capozza, R. F.; Chiappe, M. A.; Feldman, S.; Meta, M. D.; Daniele, S. M.; Fracalossi, N. M.; Reina, P.; Ferretti, J. L. J. Bone Miner. Metab. 2005, 23 (Suppl), 30–35. (7) Kinney, J. H.; Habelitz, S.; Marshall, S. J.; Marshall, G. W. J. Dent. Res. 2003, 82, 957–961. (8) Linde, A. Anat. Rec. 1989, 224, 154–166. (9) Butler, W. T.; Ritchie, H. Int. J. Dev. Biol. 1995, 39, 169–179. (10) Fisher, L. W.; Torchia, D. A.; Fohr, B.; Young, M. F.; Fedarko, N. S. Biochem. Biophys. Res. Commun. 2001, 280, 460–465. (11) He, G.; Gajjeraman, S.; Schultz, D.; Cookson, D.; Qin, C.; Butler, W. T.; Hao, J.; George, A. Biochemistry 2005, 44, 16140–16148. (12) Saito, T.; Ito, S.; Abiko, Y.; Matsuda, K.; Crenshaw, M. A. Dentin/Pulp Complex 2002, 159–161. (13) Olszta, M. J.; Douglas, E. P.; Gower, L. B. Calcif. Tissue Int. 2003, 72, 583–591. (14) Olszta, M. J.; Odom, D. J.; Douglas, E. P.; Gower, L. B. Connect. Tissue Res. 2003, 44 (Suppl 1), 326–334. (15) Deshpande, A. S.; Beniash, E. Cryst. Growth Des. 2008, 8, 3084–3090. (16) Olszta, M. J.; Cheng, X. G.; Jee, S. S.; Kumar, R.; Kim, Y. Y.; Kaufman, M. J.; Douglas, E. P.; Gower, L. B. Mater. Sci. Eng., R 2007, 58, 77–116. (17) Glimcher, M. J.; Hodge, A. J.; Schmitt, F. O. Proc. Natl. Acad. Sci. U.S.A. 1957, 43, 860–867. (18) Price, P. A.; Toroian, D.; Lim, J. E. J. Biol. Chem. 2009, 284, 17092–17101. 8268
dx.doi.org/10.1021/la201361e |Langmuir 2011, 27, 8263–8268
