Kinetics of Aggregation and Crystallization of Polyaspartic Acid

Apr 22, 2015 - Meng Li , Lijun Wang , Wenjun Zhang , Christine V. Putnis , and Andrew Putnis. Crystal Growth & Design 2016 16 (8), 4509-4518. Abstract...
0 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/Biomac

Kinetics of Aggregation and Crystallization of Polyaspartic Acid Stabilized Calcium Phosphate Particles at High Concentrations Daniel V. Krogstad,† Dongbo Wang, and Sheng Lin-Gibson* Biosystems and Biomaterials Division, National Institute of Standards and Technology, Gaithersburg, Maryland, United States S Supporting Information *

ABSTRACT: Bone is an important material to study due to its exceptional mechanical properties and relevance with respect to hard tissue regeneration and repair. A significant effort has been directed toward understanding the bone formation process and the production of synthetic bone mimicking materials. Here, the formation and structural evolution of calcium phosphate (CaP) was investigated in the presence of relatively high concentrations of calcium, phosphate, and polyaspartic acid (pAsp) using dynamic light scattering (DLS) and cryo-transmission electron microscopy (cryoTEM). The incipient CaP aggregates were comprised of spherical nanoparticles (diameter ≈ 3−4 nm); they became preferentially aligned over time and eventually transformed into nanorods. The nanorods remained stable in suspension with no signs of further aggregation for at least four months. Detailed cryo-TEM suggested that the CaP nanorods formed through an oriented attachment mechanism. These results show that the reaction concentration greatly influences the mechanism and final properties of CaP. Mechanistic insights gained from this study will facilitate better design and fabrication of bioinspired materials.

■

INTRODUCTION The exceptional engineering properties found in mineralized tissues, such as bone and teeth, have been largely attributed to the hierarchical organization between calcium phosphate (CaP) minerals and collagen. The intricate assembly process is directed by noncollagenous proteins (NCPs), a broad class of mineral associated proteins that are key to many important biological functions in vivo.1−3 One common characteristic among NCPs is a high degree of negative charge density expressed as acidic carboxylated domains and phosphorylated domains.1,2 Experimentally, the roles of NCPs are often screened to better understand the efficiency of a particular (bio)molecule with respect to its ability to promote in vitro formation of intrafibrillar mineralized collagen, using bone as a benchmark.4−9 Some of these studies have focused on natural mineral associated proteins, such as dentin matrix protein 1 (DMP1),4 fetuin,5,6 osteopontin,7,9 bone sialoprotein,8,9 and dentin sialophosphoprotein.9 Others have examined synthetic NCP analogs, including poly(acrylic acid) (pAA),10 polyaspartic acid (pAsp),1,10−13 and polyglutamic acid (pGlu).10 Of the synthetic analogs, pAsp has been shown to be one of the most broadly effective.1,10,11,13 In vitro studies have also suggested that the interaction between collagen and the NCP analog is important for intrafibrillar bone-like mineralization, something that cannot be achieved by collagen and CaP alone.11,13 In addition, NCP analogs typically have an inhibitory effect on CaP mineralization in a concentration dependent manner.11,14 To better understand the interaction between the NCP analogs and the mineral phase, studies have been designed to monitor pAsp mediated CaP mineral formation in the absence © XXXX American Chemical Society

of collagen. While a number of these studies showed the formation of a stable apatite phase, the rate at which the reaction reached completion and the final CaP structure appear to vary greatly depending on the experimental conditions.5,14,15 More specifically, these studies show that polyanions, such as pAsp, can drastically alter the pathway by which stable crystals form. The changes in pathways are characterized by differences in the stability, aggregation kinetics, and structure of the precursor phases, which then affect the structure and morphology of the final crystalline product. Some studies have shown that the formation of stabilized aggregates (diameter = 30−70 nm) consisted of 1 nm amorphous CaP particles for at least 6 h using a low pAsp concentration (10 μg/ mL). These particles eventually crystallized into apatite rods.5 We recently showed that CaP formation in the presence of a range of pAsp concentrations occurred through a directed process in which there was an initial particle growth regime followed by an aggregation regime.14 Small changes in the pAsp concentration (from ≈35 to ≈95 μg/mL) did not change the particle induction time, but significantly affected the kinetics of precursor particle aggregation.14 In contrast, an experiment that uses a double-jet mixing setup showed a cascade of organic/ inorganic structures, including polymer stabilized globule (r ≈ 100 nm), long needle-like crystals, and spherical clusters of aggregated crystals (diameter ≈ 500 nm).15 Together, these results show pAsp mediated CaP mineralization depends highly on the reaction conditions, that is, the polymer and ion Received: January 27, 2015 Revised: March 26, 2015

A

DOI: 10.1021/bm501725t Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules

Figure 1. (a) DLS showed the changes in the intensity averaged hydrodynamic radius of the particles over 24 h. (b−f) Cryo-TEM correlated these changes to a conversion of globular aggregates of spherical nanoparticles to apatite nanorods: (b) 0.25 h, spheres; (c) 6 h, spheres and rods; (d) 8.75 h, spheres and rods; (e) 11.5 h, rods; (f) 24 h, rods. The scale bar is equal to 50 nm for all images. pAsp/Ca solution or the pAsp/P solution, respectively. These solutions were equilibrated at 37 °C for at least 1 h before filtered using a 0.2 μm Millex-LG PTFE filter. We defined the start of the mineralization reaction (t = 0 s) as the time of mixing of pAsp/Ca and pAsp/P solutions (500 μL each). The reaction was mixed with a vortex for ≈10 s. Using this protocol, the combined mixture consisted of 9.375 mg/mL pAsp, 25 mM CaCl2·2H2O, and 12.5 mM KH2PO4. This general procedure was used for all of the concentrations studied; only the amounts of the initial stock solutions were changed to adjust the final concentrations. For all samples, the Ca/P ratio was 2:1. Dynamic Light Scattering (DLS). For DLS experiments (Malvern Zetasizer Nano-ZS), samples were transferred to a small volume polystyrene cuvette, and measurements were taken at 5 min intervals starting 5 min after mixing. The samples were kept at 37 °C throughout the experiment. Cryogenic Transmission Electron Microscopy (Cryo-TEM). Samples for the cryo-TEM experiments were prepared using a controlled environment vitrification system16 at 100% humidity and room temperature. Aliquots were taken from the DLS experiment at select time points and a 3.5 μL droplet was pipetted onto a plasma cleaned lacey Formvar coated copper TEM grid (200 mesh) from Ted Pella. The sample was hand-blotted three times using Whatman 40 filter paper (Sigma-Aldrich) before the grid was plunged into liquid ethane which was cooled by liquid nitrogen to vitrify the sample. The samples were then stored in liquid nitrogen until imaging. The vitrified samples were transferred into a Gatan 914 cryo transfer holder and were imaged using low dose imaging conditions at 300 keV with a FEI Titan 80−300 TEM. The images were processed and the particles were measured using ImageJ. For each time point, more than 20 images were recorded. Particle measurements were performed on 2−6 images per time point, and at least 300 measurements were made for each particle dimension. All particle measurements were made on images taken at a magnification of 34000× and a defocus of approximately 2.5 μm. While this large defocus value may lead to particle measurements that are slightly larger than the true particle sizes, it is necessary to achieve good contrast. By

concentrations, the polymer structure and molecular mass, as well as the process used. While most studies have used ion concentrations similar to those found in the body (1−5 mM calcium) to better mimic the biological process, more concentrated solutions may be desirable for large-scale production of biomimetic materials. This study investigates the kinetics and particle structure of pAsp stabilized CaP formation at an expanded pAsp and salt concentration (9.375 mg/mL pAsp, 25 mM calcium, 12.5 mM phosphate) by dynamic light scattering (DLS), cryo-transmission electron microscopy (cryo-TEM), and high resolution TEM.

■

MATERIALS AND METHODS

Materials. Poly-L-aspartic acid sodium salt (pAsp, MW = 14000 g/ mol) was purchased from Alamanda Polymers. Calcium chloride dihydrate (CaCl2·2H2O), potassium phosphate monobasic (KH2PO4), sodium chloride (NaCl), HEPES, sodium hydroxide (NaOH), and hydrogen chloride (HCl) were purchased from Sigma-Aldrich. All materials were used as received. Certain commercial equipment, instruments, or materials are identified in this paper in order to specify the experimental procedure adequately. Such identification is not intended to imply recommendation or endorsement by the National Institute of Standards and Technology, nor is it intended to imply that the materials or equipment identified are necessarily the best available for the purpose. Sample Preparation. All solutions were prepared with ultrapure water with a resistivity of 18.2 MΩ (Milli-Q). Stock solutions of pAsp (37.5 mg/mL), calcium (1 M CaCl2·2H2O), and phosphate (500 mM KH2PO4) were each prepared in HEPES-buffered saline (HBS, 100 mM HEPES, 150 mM NaCl) and the pH was adjusted to 7.4 with NaOH or HCl. Samples for the mineralization studies were prepared by first mixing the pAsp stock solution (150 μL), 420 μL of HBS, and either 30 μL of the calcium or phosphate stock solution to form the B

DOI: 10.1021/bm501725t Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules using the same defocus for every image, the procedure is selfconsistent and the particle sizes can be compared with one another. For spherical particle measurements, the circle tool in ImageJ was used to outline the circumference of the particle to measure the particle diameter. For the rod length and width, the line tool was used to measure the particle dimensions. The line measurements for the rod widths were taken perpendicular to the rod direction. High Resolution Transmission Electron Microscopy (HRTEM). For the HRTEM experiments, a 3.5 μL droplet of the reaction solution was placed on a plasma cleaned TEM grid (ultrathin carbon film with holey carbon support from Ted Pella) for 5 min. Filter paper was touched to the edge of the grid to wick away the excess solution. The grid was then dipped in ultrapure water 3 times in order to wash away the excess salt. Blotting paper was once again used to wick away the excess solution and then the grid was placed under vacuum overnight. The sample was imaged using low dose imaging conditions at 300 keV with a FEI Titan 80−300 TEM. The images were processed using ImageJ.

quantity of nanorods. The ratio of rods to globular aggregates increased with time until approximately the 12 h time point (Figure 1e) when no appreciable amount of globular aggregates were observed. For this reaction composition, cryo-TEM measurements on samples aged for 18 weeks showed little or no change in the particle structure as compared to the 12 h image, indicating that the rods were quite stable in solution (Figure S3). The diameter of the spherical CaP nanoparticles and the dimensions (width and length) of the nanorods were quantified to better assess the transition process (Figure 2). Interestingly,

■

RESULTS AND DISCUSSION Dynamic light scattering (DLS) was used to estimate the calcium phosphate (CaP) particle sizes formed over a range of pAsp concentrations (1.875−9.375 mg/mL) and salt concentrations (10−25 mM CaCl2, 5−12.5 mM KH2PO4). Specifically, the distributions of the intensity averaged hydrodynamic radius (z-average Rh) were measured at the 1 h time point (Figure S1) to assess the overall ability for pAsp to form stabilized CaP particles. The relationship between the pAsp concentration and ion concentration strongly influenced the stability of the mixtures, that is, at 1 h reaction time, CaP aggregate size decreased when the polymer concentration is increased or the salt concentration is decreased. In addition, reactions that contained lower pAsp to ion ratios resulted in the formation of visible precipitates after several hours, while those containing higher pAsp to ion ratios remained clear over several months. Apatite precipitates rapidly in the absence of pAsp; control experiments showed the formation of large aggregates consisting of apatite platelets within 15 min (Figure S2). These results are consistent with previous findings that pAsp and similar NCP analogs have a concentration dependent inhibitory effect on CaP apatite mineralization.11,14 One specific composition (9.375 mg/mL pAsp, 25 mM CaCl2 and 12.5 mM KH2PO4) was selected for detailed investigation of the aggregation mechanism. This sample formed CaP aggregates (radius ≈ 13 nm after 1 h) and became a stable particle suspension at approximately 12 h, making it particularly suitable for monitoring the structural evolution over time. For this sample, DLS was used to determine the reaction kinetics by measuring the hydrodynamic radius (Rh) every 5 min over time. At selected time points, aliquots were collected from a parallel reaction solution for investigation with cryo-TEM (Figure 1). Interestingly, the average Rh reached ≈13 nm before the first measurement could be collected at 5 min (Figure 1a). Rh gradually increased to ≈15 nm over the next 6 h, at which point a transition period, with a more rapid increase in Rh, ensued. The Rh increased significantly over this transition period until it finally stabilized at ≈12 h when the Rh reached ≈27 nm. Cryo-TEM images collected at various time points, corresponding to different stages of the DLS profile, showed the detailed structural evolution of CaP mineralization (Figure 1b−f). The initial structures consisted of globular aggregates comprised of CaP nanoparticles (diameter ≈ 3.5 nm; Figure 1b). At the 6 h time point, cryo-TEM images (Figure 1c) showed mostly CaP globular aggregates but also a small

Figure 2. Measurements of the (a) sphere diameter, (b) rod width, and the (c) rod length of the particles from the cryo-TEM images show that the particles did not change much with time. (a, b) The error bars represent the standard deviation of the measurements and the dashed lines represent the average radius of all of the measured particles.

the size distribution of the diameter of the spherical CaP nanoparticles (Figure 2a) and the width of the nanorods (Figure 2b) were relatively narrow and did not change significantly over time. The average width of the nanorods (2.8 ± 0.5 nm) was slightly smaller than the diameter of the spherical CaP particles (3.7 ± 0.6 nm). The distributions for the rod length were broader and took on a log-normal distribution, but also remained relatively constant with time (Figure 2c); that is, there was no evidence for elongation of the rods. To better understand the structural evolution process, cryoTEM images collected at the intermediate time points (6 and 8.75 h) were examined more closely. Images at these time points revealed many examples in which the spherical particles within globular aggregates became preferentially aligned to form elongated structures, strongly suggesting the presence of an intermediate structure for the conversion of the globular aggregates to rods (Figure 3). These observations, along with the fact that the rod dimensions (average size and size distribution) remained nearly constant with respect to the reaction time, indicated that the transition to the crystalline rods did not proceed through a classical dissolution/ reprecipitation mechanism, that is, the ions from spherical nanoparticles did not undergo dissolution followed by C

DOI: 10.1021/bm501725t Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules

pattern consistent with the apatite structure with the long-axis of the nanorods parallel to the ⟨110⟩ direction. Additionally, the solution pH dropped from ≈7.4 to ≈7.0 immediately after mixing and then remained relatively stable over the course of 2 days (Table S1). Previous studies have shown hydroxyapatite to be the most thermodynamically stable phase (or the least soluble phase) over this concentration regime and pH range.17 We note that traditional understanding of CaP mineralization suggests that CaP first forms an amorphous structure before crystallizing into apatite, but no other phases are expected for this process. For the present study, aggregates from earlier time points could not be measured using HRTEM because previous studies have shown substantial changes in both the structure and phase of amorphous particles upon drying,18 rendering results from such experiments inconclusive. Regardless of the structure of the initial particles, there must be a significant driving force for the particles to align into the less energetically favorable elongated aggregates, especially considering the polymer rearrangements that would be necessary for such assembly. Previous work involving amelogenin (enamel protein) directed CaP mineralization showed assembly of nanorods from primary particles via specific protein−protein interactions,19 such mechanisms are not possible for pAsp. One likely mechanism that could explain this structural transformation is oriented attachment. Oriented attachment, the process in which small crystallites combine to form larger particles with a uniform crystallographic order, was first described in the late 1990s in titanium dioxide.20,21 This mechanism has since been attributed to dozens of material systems including other oxides, sulfides, selenides, and metal oxyhydroxides,22 as well as calcium carbonate.23 Oriented attachment has also been suggested to occur in CaP;24,25 however, it was suggested that complicated synthesis methods were needed in order to produce CaP particles through this mechanism. The primary driving force for oriented attachment is thought to be Coulombic interactions, although van der Waals interactions may play a role at long distances in salt solutions.26 These driving forces cause the particles to assemble and align in the crystallographically preferred orientation. Once the particles align, they coalesce, resulting in an overall reduction in the energy of all of the atoms in the particles, not just the atoms at the surface.26 Particle fusion in oriented attachment is thought to be a rapid process where slight misalignments between constituent particles could result in twins or stacking faults.22,26 As such, one way to test this mechanism is to locate periodic defects in the assembled particles via high resolution TEM (HRTEM).22 A representative image of a nanorod clearly shows two key hallmarks associated with oriented attachment (Figure 5). First, the image showed a distinct bend along the long axis of the nanorod, which is not commonly associated with crystalline materials produced through classical crystal growth mechanisms.20 Second, FFT of the HRTEM images showed misaligned crystal structures at different locations along the same nanorod. These characteristics, caused by the misalignment of individual crystalline domains, strongly suggests that the rods were formed from the coalescence of aligned particles, where low angle misorientations between constituent particle domains are common.22 Additionally, the fact that the rod length and length distribution remained constant with time (Figure 2) and that the rod lengths dropped off significantly at an aspect ratio of ≈16 are consistent with simulations of nanorod formation through oriented attachment.27

Figure 3. Cryo-TEM images show the alignment of the spherical particles before rod formation. The images are 20 × 20 nm: (a, b) 6 h, (c, d) 8.5 h.

precipitation into nanorods. It would be expected of classical dissolution/reprecipitation that nanorods grow with time. In this system, the particle sizes did not change, but the relative proportion of nanorods increased with time, consistent with a mechanism in which nanorods are formed via the rearrangement of existing spherical particles into elongated aggregates followed by particle fusion. Attempts to use electron diffraction under cryo-TEM conditions to determine the phase of the particles failed due to the large amount of amorphous scattering from vitreous water that dominated the diffraction pattern. Therefore, dried samples of the stable nanorods were collected at the 24 h time point and imaged using HRTEM (Figure 4). Fast Fourier transform (FFT) of the particles showed a clear diffraction

Figure 4. HRTEM image showing a CaP rod that was dried on a TEM grid 24 h after mixing. An FFT of the area within the white box shows that the particles are apatite. This particle is aligned in the ⟨110⟩ direction. D

DOI: 10.1021/bm501725t Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules

apatite nanorods form. As time passes, the crystal structure of the CaP nanoparticle may become sufficiently ordered for the Coulombic interactions to drive the particles to align through the oriented attachment mechanism, overcoming the stability provided by the polymers. The limited number of aggregates in this configuration observed at any given time indicated that this structure is energetically unfavorable and the nanoparticles will likely coalesce into rods very quickly. Once a critical point is reached (≈6 h for the current system), the number of aggregates that are transformed significantly increases until the sample contains only rods (≈12 h for the current system). These rods remain unchanged for months after they are formed, although we suspect that the stability of the final rods will also depend on the polymer and salt concentration (Figure S3).

■

CONCLUSIONS We characterized the structural evolution of CaP nanoparticle formation in the presence of polyaspartic acid at significantly higher concentrations than previously investigated. The samples initially formed aggregates of CaP nanoparticles before they transformed into apatite nanorods. Cryo-TEM provided clear evidence for the alignment of spherical particles before rod formation. Depending on the polymer and salt concentration, CaP nanorods can remain stable in suspension for at least several months. HRTEM of nanorods showed distinct features consistent with an oriented attachment mechanism. Collectively, we provided experimental evidence for oriented attachment mechanism in the CaP particle growth under high ionic and polymer concentrations.

Figure 5. HRTEM image shows a rod that is bent indicating that the particle was formed from the aggregation of smaller particles. FFT at the ends of the particles show that the crystal structure was not aligned along the length of the particle. The white lines in the FFT are to guide the eye to see that the angles are quite different.

The mechanism we propose for CaP particle growth at high polymer and salt concentration (Figure 6) is quite different from previous low concentration experiments.5,11,14,15 We suggest that the phosphate ions and carboxyl groups on the pAsp compete for calcium ions immediately after mixing. Calcium ions can bind to both the polymer and the phosphate, which results in pAsp stabilized CaP particles. The polymers can potentially bind to multiple CaP particles causing the particles to aggregate as shown by the cryo-TEM images (Figure 1b−f). Additionally, the reaction occurs in an excess of calcium relative to phosphate. Excess divalent calcium ions may bind to the polymers and decrease the electrostatic repulsion of the polymers to physically cross-link the polymers. The CaP aggregates form very quickly, well within the first 5 min after mixing and before the first DLS measurement. The aggregates appeared to be relatively stable for a few hours (in a concentration-dependent manner) before undergoing structural changes. During this time, the individual particle size does not change significantly, as was indicated by the particle analysis. We propose that aggregates of CaP nanoparticles are in a kinetically trapped state before the thermodynamically stable

■

ASSOCIATED CONTENT

* Supporting Information S

The DLS results for all nine of the initial reactions (S1), an SEM and TEM image of the no pAsp control sample (S2), the pH of the reactions (Table S1), and the 18 week cryo-TEM image (S3). The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/ bm501725t.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel.: (301) 975-6765.

Figure 6. Proposed mechanism for the structural evolution of the CaP particles showed that the initial competition between the phosphate ions and the carboxyl groups of the pAsp for the calcium ions led to the formation of spherical nanoparticles coated in pAsp. The particles formed globular aggregates due to the ability of the polymers to bind to multiple particles and because excess divalent calcium ions served as bridges between the polymers. The spheres in the aggregates aligned in a row before coalescing into pAsp coated apatite rods through oriented attachment. E

DOI: 10.1021/bm501725t Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules Present Address

(25) Tao, J. H.; Pan, H. H.; Zeng, Y. W.; Xu, X. R.; Tang, R. K. J. Phys. Chem. B 2007, 111, 13410−13418. (26) Zhang, H. Z.; Banfield, J. F. CrystEngComm 2014, 16, 1568− 1578. (27) He, W. D. CrystEngComm 2014, 16, 1439−1442.

†

Illinois Applied Research Institute, University of Illinois at Urbana−Champaign (D.V.K.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work is supported by an Interagency Agreement between National Institute of Dental and Craniofacial Research (NIDCR) and NIST (Y1-DE-7005-01). D.V.K. acknowledges the NIST-National Research Council (NRC) Research Associate Program for postdoctoral support. Research performed in part at the NIST Center for Nanoscale Science and Technology. We thank Steven Hudson, John Bonevich, and Alline Meyer for their assistance with cryo-TEM and HRTEM. We thank Vivek Prabhu for assistance with DLS. Official contribution of NIST; not subject to copyrights in U.S.A.

■

REFERENCES

(1) 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. (2) Clarke, B. Clin. J. Am. Soc. Nephrol. 2008, 3, S131−S139. (3) Boonrungsiman, S.; Gentleman, E.; Carzaniga, R.; Evans, N. D.; McComb, D. W.; Porter, A. E.; Stevens, M. M. Proc. Natl. Acad. Sci. U.S.A. 2012, 109, 14170−14175. (4) He, G.; George, A. J. Biol. Chem. 2004, 279, 11649−11656. (5) Nudelman, F.; Pieterse, K.; George, A.; Bomans, P. H. H.; Friedrich, H.; Brylka, L. J.; Hilbers, P. A. J.; de With, G.; Sommerdijk, N. A. J. M. Nat. Mater. 2010, 9, 1004−1009. (6) Toroian, D.; Price, P. A. Calcif. Tissue Int. 2008, 82, 116−126. (7) Rodriguez, D. E.; Thula-Mata, T.; Toro, E. J.; Yeh, Y. W.; Holt, C.; Holliday, L. S.; Gower, L. B. Acta Biomater. 2014, 10, 494−507. (8) Baht, G. S.; Hunter, G. K.; Goldberg, H. A. Matrix Biol. 2008, 27, 600−608. (9) Zurick, K. M.; Qin, C. L.; Bernards, M. T. J. Biomed. Mater. Res., Part A 2013, 101, 1571−1581. (10) Thula, T. T.; Svedlund, F.; Rodriguez, D. E.; Podschun, J.; Pendi, L.; Gower, L. B. Polymers 2011, 3, 10−35. (11) Deshpande, A. S.; Beniash, E. Cryst. Growth Des. 2008, 8, 3084− 3090. (12) Jee, S. S.; Thula, T. T.; Gower, L. B. Acta Biomater. 2010, 6, 3676−3686. (13) Zeiger, D. N.; Miles, W. C.; Eidelman, N.; Lin-Gibson, S. Langmuir 2011, 27, 8263−8268. (14) Ye, J.; Wang, D. B.; Zeiger, D. N.; Miles, W. C.; Lin-Gibson, S. Biomacromolecules 2013, 14, 3417−3422. (15) Peytcheva, A.; Colfen, H.; Schnablegger, H.; Antonietti, M. Colloid Polym. Sci. 2002, 280, 218−227. (16) Bellare, J. R.; Davis, H. T.; Scriven, L. E.; Talmon, Y. J. Electron Microsc. Technol. 1988, 10, 87−111. (17) Chow, L. C.; Eanes, E. D. Octacalcium Phosphate; Karger: Basel; New York, 2001; p 167. (18) Habraken, W.; Tao, J. H.; Brylka, L. J.; Friedrich, H.; Bertinetti, L.; Schenk, A. S.; Verch, A.; Dmitrovic, V.; Bomans, P. H. H.; Frederik, P. M.; Laven, J.; van der Schoot, P.; Aichmayer, B.; de With, G.; DeYoreo, J. J.; Sommerdijk, N. Nat. Commun. 2013, 4, 12. (19) Fang, P. A.; Conway, J. F.; Margolis, H. C.; Simmer, J. P.; Beniash, E. Proc. Natl. Acad. Sci. U.S.A. 2011, 108, 14097−14102. (20) Penn, R. L.; Banfield, J. F. Science 1998, 281, 969−971. (21) Penn, R. L.; Banfield, J. F. Am. Mineral. 1998, 83, 1077−1082. (22) Penn, R. L.; Soltis, J. A. CrystEngComm 2014, 16, 1409−1418. (23) Li, D. S.; Nielsen, M. H.; Lee, J. R. I.; Frandsen, C.; Banfield, J. F.; De Yoreo, J. J. Science 2012, 336, 1014−1018. (24) Wang, L. J.; Guan, X. Y.; Yin, H. Y.; Moradian-Oldak, J.; Nancollas, G. H. J. Phys. Chem. C 2008, 112, 5892−5899. F

DOI: 10.1021/bm501725t Biomacromolecules XXXX, XXX, XXX−XXX