A Density Functional Theory Study of the Interaction of Collagen

Aug 23, 2010 - Ongoing and future work in our group will take into account the effect of water on the adsorption of collagen at hydroxyapatite surface...
11 downloads 7 Views 5MB Size
Download PDF
pubs.acs.org/Langmuir © 2010 American Chemical Society

A Density Functional Theory Study of the Interaction of Collagen Peptides with Hydroxyapatite Surfaces Neyvis Almora-Barrios† and Nora H. de Leeuw*,†,‡ †

Department of Chemistry, University College London, 20 Gordon Street, London WC1H 0AJ, United Kingdom, and ‡Institute of Orthopaedics and Musculoskeletal Science, University College London, Brockley Hill, Stanmore HA7 4LP, Middlesex, United Kingdom Received March 23, 2010. Revised Manuscript Received July 13, 2010 Density functional theory calculations were applied to investigate the binding of four peptide strands, which are important in the collagen protein, to the bone and tooth mineral hydroxyapatite: amphiphilic PRO-HYP-GLY and HYP-PRO-GLY, and hydrophobic PRO-LYS-GLY and PRO-HYL-GLY. The particular peptide sequences are chosen for their different functional groups, containing (i) hydrophobic; (ii) uncharged polar; and (iii) charged polar side groups, thus allowing direct comparison of the general effect of these carboxylic acid and amine functional groups, as well as hydroxylation and charge, on their interactions with two major hydroxyapatite surfaces, (0001) and (0110). The calculated results are consistent with experiments, confirming that the terminal carboxyl groups and amine groups mainly contribute to the adsorption of the peptides to the hydroxyapatite surfaces and primarily to the (0110) surface rather than the dominant (0001) plane. Of the side groups in the tripeptide motifs representing the collagen protein, the -OH and positively charged -NH3þ groups in particular bind strongly to the surfaces, and their presence should therefore promote hydroxyapatite growth.

1. Introduction Carbonated hydroxyapatite (HA) is the major mineral phase in natural bone and teeth1-3 and is therefore an attractive material for use in tissue replacement applications.1-4 However, its successful application as a biomaterial requires a detailed understanding of the interaction with complex biological macromolecules.5-7 Bone is a protein-mineral composite, where the hydroxyapatite grows as nanosized mineral platelets at nucleation sites on a protein template, which is predominantly type I collagen.2,3 The HA platelets, with average lengths and widths of 50 nm  25 nm,8 are very thin indeed, typically 2-4 nm,3 and these crystallites are arranged in an ordered fashion within and around collagen fibrils, where the apatite mineral is aligned with its c-axis, the (0001) direction, along the fibril.2,8 Recent studies have demonstrated that the growth process of HA is controlled by the interaction between the collagen matrix and HA crystal,3,5-7,9 which determines the eventual morphology of the HA platelets in the bone. Thermodynamically, the (0001) surface is the most stable HA surface,10,11 but the (0110) plane is dominant in the morphology of the biological material,12 due to the growthdirecting effect of the collagen matrix. High-resolution electron microscopy studies of HA crystals in tooth enamel, for example, *[email protected]. (1) Narasaraju, T. S. B.; Phebe, D. E. J. Mater. Sci. 1996, 31, 1–21. (2) Weiner, S.; Wagner, H. D. Annu. Rev. Mater. Sci. 1998, 28, 271–298. (3) Fratzl, P.; Gupta, H. S.; Paschalis, E. P.; Roschger, P. J. Mater. Chem. 2004, 14, 2115–2123. (4) Ratner, B. D. Polymers for biomedical applications: Improvement of the interface compatibility. Biomed. Appl./Polym. Blends, 1999 1999, 149, 1-57. (5) Elbert, D. L.; Hubbell, J. A. Annu. Rev. Mater. Sci. 1996, 26, 365–394. (6) Klee, D.; Hocker, H. Adv. Polym. Sci 1999, 149, 1–57. (7) Griffith, L. G. Acta Mater. 2000, 48, 263–277. (8) Weiner, S.; Arad, T.; Traub, W. Febs Lett. 1991, 285, 49–54. (9) Prockop, D. J.; Kivirikko, K. I. Annu. Rev. Biochem. 1995, 64, 403–434. (10) Deer, W. A.; Howie, R. A.; Zussman, J. An introduction to the rock-forming minerals, 2nd ed.; Longman Scientific & Technical; Pearson: Harlow, 1992; p xvi. (11) Mkhonto, D.; de Leeuw, N. H. J. Mater. Chem. 2002, 12, 2633–2642. (12) Rohanizadeh, R.; Trecant-Viana, M.; Daculsi, G. Calcif. Tiss. Int. 1999, 64, 430–436.

Langmuir 2010, 26(18), 14535–14542

have shown that matrix proteins only affect the growth of the (0110) surface and not the other HA surfaces present.13 The collagen fibrils are built up of collagen molecules, which each consist of three polypeptide chains of amino acid residues, which are coiled around each other in a triple helix, a structure which places the functional groups of the amino acids on the outside of the collagen molecule. Approximately one-third of amino acid residues are glycine (GLY), with proline (PRO) or hydroxyproline (HYP) making up another quarter or so.9 Lysine (LYS) and hydroxylysine (HYL) residues in different peptide chains form bifunctional interchain cross-links through a relatively complex condensation reaction, which continues as the bone matures, thereby rendering the collagen of older mammals more difficult to extract from tissue.4 A number of experimental and theoretical studies of the interaction of amino acids and proteins with hydroxyapatite surfaces have demonstrated that amino acids are effective regulators in processes involving both nucleation and crystal growth of apatites,14-16 where carboxyl groups were found to form the primary association between molecules and HA surfaces.17-22 However, although terminal carboxyl groups may be important in the binding of amino acids or short peptides to HA surfaces, the 300-nm-long collagen molecules contain few carboxyl groups, and we therefore need to consider the interaction of relevant collagen side groups with the HA surfaces. (13) Bres, E. F.; Hutchison, J. L. J. Biomed. Mater. Res. 2002, 63, 433–440. (14) Pan, H. H.; Tao, J. H.; Xu, X. R.; Tang, R. K. Langmuir 2007, 23, 8972– 8981. (15) Landis, W. J.; Silver, F. H. Cell Tiss. Org. 2009, 189, 20–24. (16) Shaw, W. J.; Ferris, K. J. Phys. Chem. B 2008, 112, 16975–16981. (17) Almora-Barrios, N.; Austen, K. F.; de Leeuw, N. H. Langmuir 2009, 25, 5018–5025. (18) Ikawa, N.; Kimura, T.; Oumi, Y.; Sano, T. J. Mater. Chem. 2009, 19, 4906– 4913. (19) Koutsopoulos, S.; Dalas, E. Langmuir 2000, 16, 6739–6744. (20) Koutsopoulos, S.; Dalas, E. J. Cryst. Growth 2000, 216, 443–449. (21) Koutsopoulos, S.; Dalas, E. Langmuir 2001, 17, 1074–1079. (22) Rimola, A.; Corno, M.; Zicovich-Wilson, C.; Ugliengo, P. J. Am. Chem. Soc. 2008, 130, 16181–16183.

Published on Web 08/23/2010

DOI: 10.1021/la101151e

14535

Article

The role of collagen in the HA mineral has received extensive attention by a number of research groups. Several experimental studies have analyzed tissue structural features at the nanoscale and their relation to bone tissue failure.23,24 Computer modeling, however, can aid experiment by investigating at the atomic level highly complex structures, properties, and processes, which are still difficult or impossible to access with experimental techniques. Interactions between HA surfaces and peptide molecules have been investigated primarily with large-scale classical Molecular Dynamics simulations, e.g., refs 14,25-31, including simulations of collagen to HA surfaces.32-38 However, owing to the unprecedented growth in computing power over the last decades, electronic structure calculations are increasingly used to investigate surface adsorption processes, although thus far, investigations of the interaction of adsorbates with hydroxyapatite have been limited to water and small organic molecules.17,22,39-42 In this paper, we describe our density functional theory (DFT) study of the interaction of four peptide molecules with the (0001) and (0110) surfaces of hydroxyapatite. Each peptide is built up of three amino acid residues found in the collagen molecule, i.e., glycine (GLY), proline (PRO), hydroxyproline (HYP), lysine (LYS), and hydroxylysine (HYL), whose different side groups bestow different functionalities on the peptides. Glycine, proline, and hydroxyproline are major constituents of collagen, and by alternating the order of the residues in the peptide, we are able to isolate the effects of the individual functional groups and thus obtain a quantitative comparison of the strengths of binding of proline with hydroxyproline and lysine with hydroxylysine, without the spurious interactions of terminal -NH2 or -COOH groups with the mineral surface, which were shown in our earlier study on the adsorption of individual amino acids to dominate the surface-adsorbate interactions.17 Our comprehensive study thus provides a direct, quantitative investigation of the effects of three types of peptide functional group, namely, hydrophobic, uncharged polar, and charged polar side groups, on the interaction of the residues with the HA surfaces, thus allowing generalization of the results beyond the behavior of individual adsorbate molecules. (23) Fantner, G. E.; Hassenkam, T.; Kindt, J. H.; Weaver, J. C.; Birkedal, H.; Pechenik, L.; Cutroni, J. A.; Cidade, G. A. G.; Stucky, G. D.; Morse, D. E.; Hansma, P. K. Nat. Mater. 2005, 4, 612–616. (24) Thurner, P. J.; Erickson, B.; Jungmann, R.; Schriock, Z.; Weaver, J. C.; Fantner, G. E.; Schitter, G.; Morse, D. E.; Hansma, P. K. Eng. Fract. Mech. 2007, 74, 1928–1941. (25) Huq, N. L.; Cross, K. J.; Reynolds, E. C. J. Mol. Model. 2000, 6, 35–47. (26) Chen, X.; Wang, Q.; Shen, J. W.; Pan, H. H.; Wu, T. J. Phys. Chem. C 2007, 111, 1284–1290. (27) Zhou, H.; Wu, T.; Dong, X.; Wang, Q.; Shen, J. W. Biochem. Biophys. Res. Commun. 2007, 361, 91–96. (28) Dong, X. L.; Wang, Q.; Wu, T.; Pan, H. H. Biophys. J. 2007, 93, 750–759. (29) Shen, J. W.; Wu, T.; Wang, Q.; Pan, H. H. Biomaterials 2008, 29, 513–532. (30) Zhang, H. P.; Lu, X.; Leng, Y.; Fang, L. M.; Qu, S. X.; Feng, B.; Weng, J.; Wang, J. X. Acta Biomater. 2009, 5, 1169–1181. (31) Almora-Barrios, N.; de Leeuw, N. H. CrystEngComm 12, 960-967. (32) Makrodimitris, K.; Masica, D. L.; Kim, E. T.; Gray, J. J. J. Am. Chem. Soc. 2007, 129, 13713–13722. (33) Skepo, M. J. Chem. Phys. 2008, 129, 185101. (34) Bhowmik, R.; Katti, K. S.; Katti, D. R. J. Mater. Sci. 2007, 42, 8795–8803. (35) Bhowmik, R.; Katti, K. S.; Katti, D. R. J. Eng. Mech-Asce 2009, 135, 413– 421. (36) Dubey, D. K.; Tomar, V. Acta Biomater. 2009, 5, 2704–2716. (37) Dubey, D. K.; Tomar, V. Mater. Sci. Eng., C: Mater. Bio. Appl. 2009, 29, 2133–2140. (38) Dubey, D. K.; Tomar, V. J. Phys.: Condens. Matter 2009, 21, 205103. (39) Astala, R.; Stott, M. J. Phys. Rev. B 2008, 78, 075427. (40) Corno, M.; Busco, C.; Bolis, V.; Tosoni, S.; Ugliengo, P. Langmuir 2009, 25, 2188–2198. (41) Rimola, A.; Corno, M.; Zicovich-Wilson, C. M.; Ugliengo, P. Phys. Chem. Chem. Phys. 2009, 11, 9005–9007. (42) Wierzbicki, A.; Cheung, H. S. THEOCHEM 2000, 529, 73–82.

14536 DOI: 10.1021/la101151e

Almora-Barrios and de Leeuw

2. Theoretical Methods All calculations were carried out using the SIESTA code,43 which employs a numerical basis set to solve the Kohn-Sham equations of the density functional theory (DFT) in a periodic system, using standard norm-conserving scalar-relativistc pseudopotentials generated in the Troullier-Martins manner.44 The basis sets used for all atoms were of the DZP type (double ζ with polarization), apart from the oxygen and hydrogen in the hydroxy group, where the basis sets were obtained from the optimization of water at 0.2 GPa.45,46 For the oxygen, carbon, nitrogen, and hydrogen in the peptides, the basis sets were optimized with respect to the interaction of guanine-cytosine at 0.2 GPa.47 The unit-cell was optimized using a cutoff energy of 250 Ry, sampling was taken at the Γ point, and a force tolerance of 0.01 eV/A˚ was used. The suitabilities of cutoff energy, k points, and the force tolerance were evaluated by monitoring the convergence of the total energy with respect to the various parameters and validated against the structural properties of hydroxyapatite. The crystal structure of hydroxyapatite (HA) has the P63/m space group with the 4e Wyckoff positions occupied by two hydroxy oxygen atoms, each with 1/2 occupancy.48 In order to translate this structure into a model with full occupancies for the hydroxy groups, as is required to carry out the calculations, we have assigned alternate 0 and 1 occupancies to these sites along the hydroxy channels in the c direction, thus changing the space group of the hydroxyapatite unit cell from P63/m to P63. Since there is only one hydroxy channel per unit cell, all the hydroxy groups in the periodic structure are oriented in the same direction, creating a net electric polarization per unit cell. In the real structure, the disorder in the relative orientation of the parallel OH channels means that electric polarization is not present in the material. Therefore, in order to create a more realistic structure, we have doubled the cell in the b direction, and assigned opposite orientations to the two OH channels in this supercell. This antiparallel orientation of the hydroxy groups is calculated to be 19 kJ mol-1 more stable than the parallel configuration, in agreement with previous theoretical results.49,50 The configuration with antiparallel hydroxy groups columns, shown in Figure 1, coincides with the arrangement of the hydroxy groups in pure, synthetic HA, which crystallizes in an ordered monoclinic space group with a double unit cell compared to the hexagonal structure.48 However, as natural HA has the hexagonal space group, we have used this structure for our calculations, thus allowing direct comparison of the simulations with experimentally determined surfaces. The geometry-optimized bulk unit cell of HA is used as the basis to construct all slabs for the surface calculations. On the basis of the results of our previous work, each slab was 10 A˚ thick, which was shown to be sufficient to obtain convergence of the surface properties.17 Parallel to the surface, the supercell consists of a 2  2 array of surface unit cells, and the slabs, containing 176 atoms, are separated from their images in the neighboring cells by a vacuum region of approximately 50 A˚. In this study, we have concentrated on two of the most significant surfaces of the apatite (43) Ordejon, P.; Artacho, E.; Soler, J. M. Phys. Rev. B 1996, 53, 10441–10444. (44) Troullier, N.; Martins, J. L. Phys. Rev. B 1991, 43, 8861–8869. (45) Anglada, E.; Soler, J. M.; Junquera, J.; Artacho, E. Phys. Rev. B 2002, 66, 205101. (46) Junquera, J.; Paz, O.; Sanchez-Portal, D.; Artacho, E. Phys. Rev. B 2001, 64, 235111. (47) Fernandez-Serra, M. V.; Junquera, J.; Jelsch, C.; Lecomte, C.; Artacho, E. Solid State Commun. 2000, 116, 395–400. (48) Stork, L.; Muller, P.; Dronskowski, R.; Ortlepp, J. R. Z. Kristallogr. 2005, 220, 201–205. (49) de Leeuw, N. H. Chem. Commun. 2001, 1646–1647. (50) de Leeuw, N. H. Phys. Chem. Chem. Phys. 2002, 4, 3865–3871.

Langmuir 2010, 26(18), 14535–14542

Almora-Barrios and de Leeuw

Article

Figure 2. Scheme of amino acids: (a) Proline (hydrophobic side group), (b) hydroxyproline (uncharged polar side group), (c) glycine (no side group), (d) lysine (charged polar side group), and (e) hydroxylysine (charged polar side group). Figure 1. Plane and side views of the hydroxy apatite structure, showing the OH- groups in hexagonal channels surrounded by triangles of Ca ions, (Ca = green, O = red, P = purple, H = white).

bone mineral, namely, the (0001) and (0110) surfaces. These two surfaces are important in the morphology of the apatite mineral platelets, and there is experimental evidence that these faces act as the binding sites for many ionic species, small molecules, and polymer-modified cell surfaces.51 The basic building block of the (0001) surface consists of three phosphate groups and three calcium ions, which is translated along the a and b axes to form the infinite surface. Previous energy minimization calculations of the different terminations of a number of HA surfaces52 have shown that this (0001) termination is the most stable, hence its use in this work. It is not possible to construct an HA slab of the (0110) surface with the same surface termination on both sides, even though the total dipole moment across the slab perpendicular to the surface is zero, which is a requisite for a stable surface. As a result, the bottom plane of the slab contains two calcium ions only, whereas the top plane has an extra two hydroxy groups. The (Ca-Ca) termination is more stable, but the (OH-OH-Ca-Ca) has a greater variation of surface species, which could form different interactions with the adsorbates,52 and both planes will occur during the HA growth process. The surface energy of this slab will be an average of the two (0110) terminations, but the adsorption energy will be specific to the particular plane used as substrate. The peptide chains were constructed from a combination of the amino acid residues proline (PRO), hydroxyproline (HYP), glycine (GLY), lysine (LYS), and hydroxylysine (HYL) (shown in Figure 2). We have designed a number of different tripeptides containing hydrophobic, uncharged polar side groups and charged polar side groups, namely, HYP-PRO-GLY, PRO-HYP-GLY, PRO-LYS-GLY, and PRO-HYL-GLY. In our investigation of (51) Kirkham, J.; Brookes, S. J.; Shore, R. C.; Wood, S. R.; Smith, D. A.; Zhang, J.; Chen, H. F.; Robinson, C. Curr. Opin. Colloid Interface Sci. 2002, 7, 124–132. (52) Filgueiras, M. R. T.; Mkhonto, D.; de Leeuw, N. H. J. Cryst. Growth 2006, 294, 60–68.

Langmuir 2010, 26(18), 14535–14542

the sorption of peptides at the HA surfaces, we have adsorbed one molecule per surface simulation cell at two initial surface sites to identify the energetically most favorable location. In addition, peptides were adsorbed at the surfaces in their optimized geometries in a number of different starting configurations, before the complete surface/adsorbate systems were geometry-optimized again and the adsorption energies calculated. The adsorption energies due to the interaction of the peptides with the HA surfaces were calculated according to eq 1 Eads ¼ Esystem - ðEsurf þ Epep Þ

ð1Þ

where Esystem is the energy of the surface with adsorbed peptide, Esurf is the energy of the simulation cell containing the surface only, and Epep is the self-energy of the peptide molecule, calculated using the same conditions. A negative adsorption energy thus indicates that adsorption of the peptide at the surfaces is thermodynamically favorable.

3. Result and Discussion 3.1. Hydroxyapatite (0001) and (0110) Surface Models. As described, the (0001) and (0110) surfaces (Figure 3) were derived from the HA bulk structure. In previous work, the stability of the surfaces as a function of slab thickness has been checked carefully.17,53 Upon geometry optimization of the (0001) surface, we observe that the phosphate groups display a little rotation of the oxygen toward the calcium ions, although the P-O bond length remains 1.60 A˚. The Ca 3 3 3 O bond length to the oxygen of the phosphate groups is 2.43 A˚, but the surface calcium ions have also relaxed further into the surface, decreasing the distances to oxygens of the phosphate groups by 6%. The column of hydroxy groups has become distorted with the topmost group relaxing out of the surface, the OH 3 3 3 OH hydrogen-bond distances shorten to an average value of 2.09 A˚, and hydrogen bonds between OH 3 3 3 OPO3 are formed at about 2.44 A˚. In the (0110) surfaces (Figure 3 b), the columns of hydroxy groups lie in the plane of the surface, rather than perpendicular to (53) Corno, M.; Orlando, R.; Civalleri, B.; Ugliengo, P. Eur. J. Mineral. 2007, 19, 757–767.

DOI: 10.1021/la101151e

14537

Article

Almora-Barrios and de Leeuw

Figure 3. Relaxed surfaces of hydroxyapatite: (a) PO4-CaPO4-PO4-Ca-Ca termination of the (0001) surface, and (b) CaCa in the bottom and (OH-OH-Ca-Ca) in the top of the (0110) surface (Ca = green, O = red, P = purple, H = white).

the surface as in the (0001) surface. The hydroxy groups remain more or less parallel to the surface plane, although half of the groups rotate to interact more closely with neighboring OH groups with average OH-OH distances of 2.31 A˚, whereas the resulting OH-OPO3 distances are about 2.13 A˚. The surface calcium ions interact with the oxygen atoms of the hydroxy groups, at a Ca-OH bond length of 2.2 A˚. There is no significant movement of the subsurface calcium ions and phosphate groups. In this surface, the disordering of hydroxy groups is more significant than in the (0001) surface, probably due to their initial orientation parallel to the surface plane, although the surface area of the (0001) is greater than that of the (0110) surface and will therefore allow more surface relaxation, if required. The calculated surface energies (γ) of the (0001) and (0110) surfaces are 1.01 and 1.32 J m2-, respectively, defined as γ ¼ ðEsurf - Ebulk Þ=A

ð2Þ

where Esurf is the energy of the simulation cell containing the surface, Ebulkis the energy of an equivalent number of bulk stoichiometric units, and A is the surface area. The (0001) surface thus is more stable than the average of the two (0110) terminations, in agreement with ab initio and classical calculations.39,40,52,54 The less stable surface is likely to be more reactive toward to the adsorption of peptides and other adsorbates.17,40,52,54 3.2. Peptide Models. We first modeled the peptide molecules as zwitterions, which is their preferred structure in aqueous environment, and in contrast to isolated amino acids,17 it is also the most stable configuration in vacuo. The structures of the peptides are largely due to the formation of intramolecular hydrogen bonds, which are shown in Figure 4. The geometries of PROHYP-GLY and HYP-PRO-GLY are similar, although there is a small but signicant difference between their conformational energies (7.8 kJ mol-1). In general, the oxygen atoms in the (54) de Leeuw, N. H.; Rabone, J. A. L. CrystEngComm 2007, 9, 1178–1186.

14538 DOI: 10.1021/la101151e

Figure 4. Optimized structures of the peptides in their zwitterionic forms, i.e., with all molecules terminated by þH2N- and -COOgroups (a) [HYP-PRO-GLY], (b) [PRO-HYP-GLY], (c) [PROLYS-GLY]þ, (d) [PRO-HYL-GLY]þ, showing intramolecular hydrogen bonding (C = gray, O = red, N = blue, and H = white).

carbonyl and carboxylate groups form hydrogen bonds with hydrogen atoms of the amine groups. However, in PRO-HYLGLY hydrogen bonds are formed between the amine group and the hydroxy group of the side chain, as well as between the hydroxy group and the carboxylate oxygens, where the hydroxy thus forms a bridge between amino and carboxylate groups. Having optimized the structures of the free peptides, we next placed them above the surfaces, in a range of different starting positions and at a variety of surface sites. Once located near the surface, the complete system was geometry-optimized once again, allowing translation, rotation, and distortion during the relaxation to optimize the interactions of the peptides with the surfaces. The calculations show that the peptides form strong interactions with the hydroxyapatite surfaces, which confirms the affinity of the NH3þ and COO- groups for HA, consistent with the estimates reported for different protein/HA systems.32,55-60 3.3. Adsorption of HYP-PRO-GLY Peptide. The lowestenergy geometry of the HYP-PRO-GLY peptide adsorbed at the dehydrated (0001) surface is shown in Figure 5a,b, where one oxygen atom of the carboxylate group interacts simultaneously with two calcium ions (Ca 3 3 3 O = 2.41, 2.61 A˚), and the second oxygen with a further calcium ion (Ca 3 3 3 O = 2.63 A˚). One carbonyl group interacts with other surface calcium ions (Ca 3 3 3 O = 2.40 A˚), (55) Chen, P. H.; Tseng, Y. H.; Mou, Y.; Tsai, Y. L.; Guo, S. M.; Huang, S. J.; Yu, S. S. F.; Chan, J. C. C. J. Am. Chem. Soc. 2008, 130, 2862–2868. (56) Gibson, J. M.; Popham, J. M.; Raghunathan, V.; Stayton, P. S.; Drobny, G. P. J. Am. Chem. Soc. 2006, 128, 5364–5370. (57) Goobes, R.; Goobes, G.; Shaw, W. J.; Drobny, G. P.; Campbell, C. T.; Stayton, P. S. Biochemistry 2007, 46, 4725–4733. (58) Long, J. R.; Shaw, W. J.; Stayton, P. S.; Drobny, G. P. Biochemistry 2001, 40, 15451–15455. (59) Shaw, W. J.; Long, J. R.; Dindot, J. L.; Campbell, A. A.; Stayton, P. S.; Drobny, G. P. J. Am. Chem. Soc. 2000, 122, 1709–1716. (60) Capriotti, L. A.; Beebe, T. P.; Schneider, J. P. J. Am. Chem. Soc. 2007, 129, 5281–5287.

Langmuir 2010, 26(18), 14535–14542

Almora-Barrios and de Leeuw

Figure 5. Geometry-optimized structure of [HYP-PRO-GLY] adsorbed at the hydroxyapatite (0001) surface: (a) overall and (b) detailed view, showing interaction distances (Ca = green, P = purple, C = gray, O = red, N = blue, and H = white).

Figure 6. Geometry-optimized structure of [HYP-PRO-GLY] adsorbed at the hydroxyapatite (0110) surface: (a) overall and (b) detailed view, showing interaction distances (Ca = green, P = purple, C = gray, O = red, N = blue, and H = white).

while a range of hydrogen bonds between the hydrogen atoms of the amine groups and oxygen atoms of the phosphate groups further enhance the binding between the peptide and the (0001) surface (NH 3 3 3 OPO3 = 1.4-2.7 A˚). Other less favorable adsorption configurations are reported in the Supporting Information. In all cases, the interaction of the adsorbed peptide with the surface is similar, i.e., the -COO- group interacts with surface calcium ions and the >NH2þ group forms hydrogen bonds to oxygen atoms of the surface phosphate groups. Also, the peptide is adsorbed at the hydroxyapatite (0001) surface in its zwitterionic form, similar to the adsorption of individual amino acids (glycine, proline, and hydroxyproline) to the same surface.17 Figure 6a,b shows the peptide molecule adsorbed at the (0110) surface. The oxygen atoms of the carboxylate group are shared among two surface calcium ions (Ca 3 3 3 OCO- = 2.38, 2.88 A˚, Ca 3 3 3 OdC = 2.31, 2.76 A˚), whereas the nitrogen atom interacts with calcium at a distance of 2.70 A˚. In addition, the hydrogen atoms of the amine group form hydrogen bonds with the oxygen atom of the phosphate group (NH 3 3 3 OPO3 = 2.3 A˚). The mechanism of adsorption on this surface differs from the (0001) surface, as now the hydrogen atom of the amine group (>NH2þ) migrates to the oxygen atom of a surface phosphate group, leaving the peptide negatively charged. The second lowest-energy configuration is similar to the lowest-energy configuration, but has fewer interactions between the oxygen atoms of the carboxylic acid group and surface calcium ions. However, there is one case where the peptide is adsorbed as a neutral zwitterion (shown in the SI), which is ∼36 kJ mol-1 higher in energy than the anionic form. In the least favorable configuration, the proton migrates to a surface hydroxy group, similar to what was found for glycine adsorption,17 to form a molecule of water that competes with the anionic form of the peptide for adsorption on the surface of hydroxyapatite, but at an energy penalty of 111 kJ mol-1 compared to the lowest-energy configuration. Langmuir 2010, 26(18), 14535–14542

Article

Figure 7. Geometry-optimized structure of [PRO-HYP-GLY] adsorbed at the hydroxyapatite (0001) surface: (a) overall and (b) detailed view, showing interaction distances (Ca = green, P = purple, C = gray, O = red, N = blue, and H = white).

Since both -COO- and >NH2þ groups are involved in peptide bonds in the chains of the collagen molecule and thus are not free to interact with the surface, we next changed the position of the amino acid residues in the peptide in order to study the effect of side groups on the interaction with the surfaces of hydroxyapatite. 3.4. Adsorption of PRO-HYP-GLY Peptide. The most stable configuration of the PRO-HYP-GLY peptide adsorbed at the (0001) hydroxyapatite surface is shown in Figure 7. Again, the peptide is located on the edge of the step on the surface, with the proline and hydroxyproline residues accommodated within the gap. The oxygen atoms of the carboxylate and carboxyl groups interact with surface calcium ions (Ca 3 3 3 OCO-=2.312.81 A˚, Ca 3 3 3 OdC = 2.55 A˚), with hydrogen bonding taking place between hydrogen atoms of the amine group and phosphate and hydroxy oxygen atoms (NH 3 3 3 OPO3 = 1.41-2.99 A˚). The central hydroxyproline residue appears to be in close contact with the surface, where a number of interactions are formed between the hydroxy group and surface species (OH 3 3 3 OPO3 = 2.8 A˚ and HO 3 3 3 Ca = 2.46 A˚). In the second most favorable configuration, the hydrogen atom of the amine group migrates to the nitrogen atom of a surface phosphate group, leaving the peptide in its anionic form. However, this configuration is 40 kJ mol-1 higher in energy than the adsorption of the peptide as a zwitterion. Both the third and fourth configurations (see the Supporting Information) are similar to the lowest-energy configuration (Figure 7), which forms, however, more interactions between the carboxylic acid group and surface calcium ions and hydrogen bonds. Figure 8 shows the most favorable configuration of the PROHYP-GLY peptide adsorbed at the (0110) surface. The amine proton migrates to the oxygen atom of a surface phosphate group, and one oxygen of the carboxyl group is shared among two calcium ions, whereas the nitrogen atom interacts with calcium at a distance of 2.51 A˚. In addition, the oxygen of one carbonyl group is engaged in an interaction with the calcium ion and the other forms hydrogen bonds with the surface hydroxy group (Ca 3 3 3 OCO-=2.39 A˚, Ca 3 3 3 OdC=2.53 A˚, -CdO 3 3 3 HO- = 2.62 A˚, NH 3 3 3 OPO3 = 2.77 A˚, NH 3 3 3 OH- = 2.85 A˚). In the second and third most stable configurations, the peptide is adsorbed in its zwitterionic form but at an energetic cost of 26 and 89 kJ mol-1, respectively. However, in the least favorable configuration, the proton migrates to a surface hydroxy group, with the formation of a molecule of water that competes with the peptide for adsorption on the surface of hydroxyapatite, leading to an energetically even less favorable adsorption mode. The primary association between the PRO-HYP-GLY peptide and the hydroxyapatite surfaces is again via -COO- and >NH2þ groups, suggesting that amino side chains carrying either DOI: 10.1021/la101151e

14539

Article

Figure 8. Geometry-optimized structure of [PRO-HYP-GLY] adsorbed at the hydroxyapatite (0110) surface: (a) overall and (b) detailed view, showing interaction distances (Ca = green, P = purple, C = gray, O = red, N = blue, and H = white).

Almora-Barrios and de Leeuw

Figure 10. Geometry-optimized structure of [PRO-LYS-GLY]þ adsorbed at the hydroxyapatite (0110) surface: (a) overall and (b) detailed view, showing interaction distances (Ca = green, P = purple, C = gray, O = red, N = blue, and H = white).

Figure 9. Geometry-optimized structure of [PRO-LYS-GLY]þ adsorbed at the hydroxyapatite (0001) surface: (a) overall and (b) detailed view, showing interaction distances (Ca = green, P = purple, C = gray, O = red, N = blue, and H = white).

acidic or basic residues are likely to interact strongly with hydroxyapatite surfaces. We next studied the peptide containing lysine in the PRO-LYS-GLY sequence, which has a positively charged side group, thus affording a direct comparison of the surface sorption affects of hydroxyproline and lysine. 3.5. Adsorption of PRO-LYS-GLY Peptide. The adsorbed PRO-LYS-GLY peptide is in close contact with the (0001) surface (see Figure 9). Again, the most favorable mode of adsorption for the peptide is through a number of interactions between the oxygen atoms of the peptide’s carboxylate and carbonyl groups and surface calcium ions (Ca 3 3 3 OCO- = 2.46-2.65 A˚, and Ca 3 3 3 OdC=2.49, 2.80 A˚). In addition, hydrogen bonding takes place between hydrogen atoms and the oxygen atoms of the surface phosphate group (NH 3 3 3 OPO3=1.52-2.94 A˚), whereas the positive charge associated with the -NH3þ of the side group of the lysine residue has increased the adsorption of the peptide to the surface. As a result, the hydrogen atom of the amine group of the lysine migrates to the oxygen atom of the surface phosphate group, leaving the peptide charge-neutral. The second configuration is 58 kJ mol-1 less favorable compared to the lowest-energy configuration, where the main difference is that now the peptide is adsorbed in its original cationic form. The third configuration also leaves the peptide in its cationic form, but in a different orientation. The fourth configuration exhibits features similar to the lowest-energy geometry, but the peptide has lost the interactions between the proline residue and the surface species (all shown in Figure S5 in the Supporting Information). On the (0110) surface, the most favorable mode of adsorption for the peptide is through its carboxylate and amine groups, shown in Figure 10. Upon adsorption to the surface, the hydrogen atoms of the amine groups of both lysine and proline migrate to the oxygen atoms of the surface phosphate groups, and the peptide is adsorbed at the surface in its anionic form. One carboxylate 14540 DOI: 10.1021/la101151e

oxygen is coordinated to calcium at a distance of 2.38 A˚ and to another calcium at a distance of 2.52 A˚, whereas the other carboxylate oxygen is coordinated to a calcium at a distance of 2.34 A˚. Also, the carbonyl oxygen and amine nitrogen form a number of interactions with the topmost calcium ions (Ca 3 3 3 OdC = 2.45, 2.62 A˚, and Ca 3 3 3 N = 2.69 A˚), as well as a range of hydrogen-bonded interactions through the hydrogen atoms of the amine group (NH 3 3 3 OPO3 = 1.76-2.12 A˚). The second configuration is very similar to the first (but higher in energy by 37 kJ mol-1), with the main difference being the more numerous interactions among carbonyl oxygen, amine nitrogen and calcium ions and hydrogen bonding in the most stable geometry (shown in Figure 10). The third configuration has lost the interaction between the amine group of the proline residue and the surface, as shown in Figure S6 in the Supporting Information. In the last configuration, we see that the proline residue is located away from the surface, so that the peptide interacts only through the carboxylate group of the glycine and amine group of the lysine. However, in all geometries the amine proton of the lysine residue migrates to an oxygen atom of a surface phosphate group. 3.6. Adsorption of PRO-HYL-GLY Peptide. The hydroxylation of the lysine residue adds to the number and strength of the interactions between the peptide and the surface species. These interactions involve the formation of hydrogen bonds between the hydroxy group of the hydroxylysine residue and phosphate oxygen atoms, whereas its hydroxy oxygen can interact with surface calcium ions. Figure 11 shows the lowest-energy configuration on the (0001) surface, where the interaction with the surface occurs mainly via oxygen atoms of carboxylate and carbonyl groups, and hydrogens of the amine groups (Ca 3 3 3 OCO- =2.41-2.61 A˚, Ca 3 3 3 OdC = 2.61 A˚, NH 3 3 3 OPO3 = 1.69, 2.15 A˚). The oxygen atoms of the HYL hydroxy group interact with surface calcium ions (HO 3 3 3 Ca = 2.40 A˚), whereas the hydrogen atom of the hydroxy group forms hydrogen-bonding interactions to a phosphate oxygen atom (-OH 3 3 3 OPO3 = 1.59 A˚). In all configurations, electrostatic attractions are likely to be important between the surface calcium ions and the oxygen atoms of the carboxylate and carbonyl groups, as well as between the oxygen atoms of the surface phosphate groups and the positively charged -NH3þ of the side group of the hydroxylysine residue (see Supporting Information, Figure S7). However, only in the lowest-energy configuration does the amine proton of the hydroxylysine residue migrate to the oxygen atom of a surface phosphate group. The lowest-energy geometry of the PRO-HYL-GLY peptide adsorbed at the (0110) surface is shown in Figure 12. The peptide adsorbs in a bidentate fashion, coordinating via both oxygen atoms of the carboxylate group to the same surface calcium atom, Langmuir 2010, 26(18), 14535–14542

Almora-Barrios and de Leeuw

with Ca 3 3 3 O distances of 2.43 A˚ and 2.53 A˚, whereas the oxygen atoms of its carbonyl and hydroxy groups interact at distances of 2.28 A˚ and 2.60 A˚, respectively. The hydroxy and amine hydrogen atoms are both weakly hydrogen bonded to surface oxygen atoms, with (-OH 3 3 3 OPO3) distances of 1.51 and (NH 3 3 3 OPO3) 2.14-2.57 A˚. As happened with PRO-LYS-GLY, the hydrogen atoms of the amine group of both hydroxylysine and proline migrate to oxygen atoms of surface phosphate groups, leaving the peptide negatively charged. In the second configuration, the peptide is adsorbed as a neutral zwitterion (shown in the Supporting Information), and only a hydrogen atom of the amine in the side group of the hydroxylysine residue migrates, this time first to a carboxylate group, followed by a hop to the oxygen atom of a surface phosphate group. The third configuration is similar to the second configuration, but with fewer interactions to surface species. The last configuration is the only one in this study, where proton transfer between peptide and (0110) surface is not observed. However, the peptide interacts strongly with surface calcium ions and forms hydrogen bonds between the hydrogen atoms in the amine group of the side of the hydroxylysine residue and oxygen atoms of the surface phosphate groups (see the Supporting Information). 3.7. Adsorption Energies. The adsorption energies of the peptides to the hydroxyapatite surfaces, calculated according to eq 1, are presented in Table 1. When we compare the adsorption energies, we see that the energies change with the position of the constituent amino acid residues in the peptide, as well as with the type of side group. This behavior is not surprising as the side groups determine the number and strength of the interactions between the peptides and the surface species. All peptides interact strongly with the two HA surfaces; however, this interaction is stronger on the (0110) surface than the (0001) surface. The (0110)

Figure 11. Geometry-optimized structure of [PRO-HYL-GLY]þ adsorbed at the hydroxyapatite (0001) surface: (a) overall and (b) detailed view, showing interaction distances (Ca = green, P = purple, C = gray, O = red, N = blue, and H = white).

Article

surface is less stable than the (0001) surface, and the adsorption process on the (0110) surface always includes chemical reactions; both factors that lead to the high adsorption energies calculated for this surface. As mentioned, the nature of the side chain of the amino acid residues that form the peptides is one of the major factors determining the adsorption energy. When we compare the adsorption energies for the different peptides, we see that the presence of a basic residue in the chain of the PRO-LYS-GLY and PRO-HYLGLY peptides favors a strong electrostatic interaction between the peptide and hydroxyapatite. As a result, the energies released upon adsorption of PRO-LYS-GLY and PRO-HYL-GLY are larger than for HYP-PRO-GLY and PRO-HYP-GLY, due to the positively charged NH3þ functional group of the lysine residue. The hydroxylation of the lysine residue creates further binding opportunities to calcium and phosphate ions, and sorption of the PRO-HYL-GLY peptide is hence energetically the most favorable. The position of the hydroxyproline residue in the HYP-PROGLY and PRO-HYP-GLY peptides does not affect the adsorption energies significantly, as the geometry of both peptides is very similar and the adsorption energies depend on the number of interactions that can be formed to the surface species. We see in Table S1 of the Supporting Information that in the case of the (0001) surface PRO-HYP-GLY forms a greater number of close interactions between the carboxylate group and surface calcium ions, while HYP-PRO-GLY has more interactions with the (0110) surface.

4. Conclusions In this work, we have presented a detailed simulation study of the adsorption mechanism of four peptides, made up of amino acid residues with different functional groups, HYP-PRO-GLY, PRO-HYP-GLY, PRO-LYS-GLY, and PRO-HYL-GLY, to the hydroxyapatite (0001) and (0110) surfaces. All four peptides adsorb more strongly to the less stable (0110) surface than the (0001) surface, with proton transfer generally occurring from the peptides to the reactive (0110) surface. On the (0001) surface, proton transfer from the peptide only takes place when the amino acid residue has a charged polar side group, i.e., in PRO-LYSGLY and PRO-HYL-GLY, where the proton of the lysine and hydroxylysine amine group migrates to the basic phosphate group. Moreover, on the (0001) surface all peptides are adsorbed as neutral zwitterions or in their cationic form, whereas on the (0110) surface, the peptides are generally adsorbed in their anionic form or as a neutral zwitterion, i.e., the charge distribution in the adsorbates is affected by the mineral surface. Thus, the process of

Figure 12. Geometry-optimized structure of [PRO-HYL-GLY]þ adsorbed at the hydroxyapatite (0110) surface: (a) overall and (b) detailed view, showing interaction distances (Ca = green, P = purple, C = gray, O = red, N = blue, and H = white). Langmuir 2010, 26(18), 14535–14542

DOI: 10.1021/la101151e

14541

Article

Almora-Barrios and de Leeuw

Table 1. Calculated Adsorption Energies of Peptides at Hydroxyapatite Surfaces adsorption energies (kJ mol-1) structure HYP-PRO-GLY

PRO-HYP-GLY

PRO-LYS-GLY

PRO-HYL-GLY

1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4

(0001)

(0110)

-429 -421 -407 -403 -450 -410 -378 -362 -657 -599 -537 -371 -694 -644 -621 -466

-666 -654 -637 -555 -625 -599 -536 -508 -756 -719 -589 -485 -872 -640 -596 -494

peptide adsorption onto HA depends on the different crystallographic faces of HA with which they interact. Similar results were obtained by Katti et al. who studied the mechanical response of a collagen molecule in the proximity of the (0001) and (0110) surfaces of HA by means of classical molecular dynamics simulations.61 All peptides form multiple interactions with the surface species, where the -COO- and >NH2þ groups were found to bind most closely to the hydroxyapatite surfaces, in agreement with experimental studies18-21 and calculations of the adsorption of single amino acids, peptides, and collagen to hydroxyapatite.17,22,36 For example, Dubey and Tomar36-38,62-64 have shown that a high degree of toughness and strain hardening behavior of the collagen-hydroxyapatite composite is due to reconstitution of Coulombic interactions between the NH3þ and COO- groups in collagen molecules and the surface ions of HA (Ca2þ, PO43-, OH-). However, in the extended 300-nm-long collagen protein, these groups are no longer terminal groups that are free to interact with the surfaces, but instead are involved in the peptide bonds between the amino acid residues, whereas only the functional side groups of the amino acid residues are available to interact with the hydroxyapatite surfaces. The positively charged -NH3þ group in the lysine and hydroxylysine side groups of PRO-LYS-GLY and (61) (62) (63) (64)

Katti, D. R.; Pradhan, S. M.; Katti, K. S. J. Biomech. 2010, 43, 1723–30. Dubey, D. K.; Tomar, V. J. Mech. Phys. Solids 2009, 57, 1702–1717. Dubey, D. K.; Tomar, V. Appl. Phys. Lett. 2010, 96, 023703. Dubey, D. K.; Tomar, V. J. Comput. Theor. Nanosci. 2010, 7, 1306–1316.

14542 DOI: 10.1021/la101151e

PRO-HYL-GLY generally interacts strongly with the surfaces, where lysine hydroxylation in PRO-HYL-GLY leads to even stronger interactions owing to the extra hydroxy group. This stabilizing effect of the hydroxy group was also seen in the hydroxyproline residue in PRO-HYP-GLY adsorption, where hydrogen bonds were formed between the uncharged polar side groups (OH), compared to the hydrophobic side group in HYPPRO-GLY, which moves away from the surface. Hence, our calculations suggest that -OH and >NH3þ side groups, in particular, should promote hydroxyapatite growth, both in the biomineralized natural material and when synthesized in the presence of growth modifiers. Ongoing and future work in our group will take into account the effect of water on the adsorption of collagen at hydroxyapatite surfaces, which clearly takes place in an aqueous environment. The presence and competition of surrounding water molecules may well affect the strength of adsorption of peptides to the HA surfaces and hence hydroxyapatite growth. Work in progress therefore includes classical molecular dynamics simulations of the adsorption of peptides at hydroxyapatite surfaces in an aqueous environment. Although these methods cannot model chemical processes such as proton transfer, they can provide insight into hydration effects and the influence of temperature on the strength of adsorption.29 The ab initio results reported in this study will be invaluable for the derivation of reliable forcefields for classical Molecular Dynamics simulations and evaluation of more approximate methods, which are necessarily used to model much larger collagen structures, e.g., refs 31,34,35,65-67. Acknowledgment. NAB is grateful to UCL for a Dorothy Hodgkin Overseas Research Studentship and EPSRC PhDþ award. We acknowledge computational resources on the UK national High Performance Computing facilities HPCx and HECToR, provided through the EPSRC High End Computing Programme and our membership of the UK’s HPC Materials Chemistry Consortium (EPSRC portfolio grant EP/D504872 and EP/F067496). Supporting Information Available: Interatomic HA-peptide distances and geometry-optimized structures of the higher energy configurations of the peptides adsorbed at hydroxyapatite surfaces. This material is available free of charge via the Internet at http://pubs.acs.org. (65) Buehler, M. J. J. Mater. Res. 2006, 21, 1947–1961. (66) Buehler, M. J. J. Mech. Behavior Biomed. Mater. 2008, 1, 59–67. (67) Gautieri, A; Russo, A.; Vesentini, S.; Redaelli, A.; Buehler, M. J. J. Chem. Theory Comput. 2010, 6, 1210–1218.

Langmuir 2010, 26(18), 14535–14542