NANO LETTERS
Metal Recognition of Septapeptides via Polypod Molecular Architecture
2005 Vol. 5, No. 3 415-419
Ersin Emre Oren,† Candan Tamerler,§ and Mehmet Sarikaya*,†,‡ Materials Science & Engineering and Chemical Engineering, UniVersity of Washington, Seattle, Washington 98195, and Molecular Biology and Genetics, Istanbul Technical UniVersity, Maslak, Istanbul, Turkey Received September 24, 2004; Revised Manuscript Received November 16, 2004
ABSTRACT The understanding of the nature of recognition of inorganic materials by proteins is one of the core elements of and has profound implications in biological materials science and engineering. Using combinatorial display methods, a considerable number of short polypeptides have recently been selected with affinity to engineering materials. During these selections, more than several polypeptides are identified with binding specificity to a chosen inorganic material. Understanding the nature of surface recognition of materials by polypeptides is essential for rational design and biomimetic engineering of these inorganic-binding polypeptides for use as linkers, catalyzers, and growth modifiers in nanotechnology and nanobiotechnology. Although there may not be direct homology among the amino acids constituting the polypeptides, their function may come from conserved molecular architecture. Here we study crystallographic surface recognition of platinum metal-binding septapeptides by conformational analysis. We find that the septapeptides conform into certain molecular architectures containing multiple protrusions (polypods) that spatially match with the crystallographic metal surfaces. While the physical recognition may originate from how well the molecular polypods spatially match a given crystallographic surface, the degree of binding may be due to the reactive groups that form the polypods, e.g., charged or polar groups (e.g., hydroxyl and amine). These results are highly consistent with the experimental binding characteristics of the Pt binders with various degrees of affinities.
Structural control of inorganics at the molecular scale is a key to the production of materials with new and improved properties used in a wide range of systems from biomaterials to nanotechnological systems.1-4 Biological hard tissues are striking examples that may serve as conceptual models for future biomimetically engineered materials.5 Hard tissues are biocomposites, i.e., hybrid materials containing biomacromolecules (such as proteins) and bioinorganics (e.g., calcite, magnetite, and silica) resulting in highly functional properties (magnetic, mechanical, and photonic). They have intricate nano- and microarchitectures controlled at the molecular level by macromolecules, often proteins via molecular recognition with high affinity and specificity.5-7 Intermolecular interactions between proteins and inorganics are mediated via multiple weak interactions (e.g., electrostatic, polar, hydrogen bonds), which collectively may approach the strength of covalent bonds. It would be desirable to engineer polypeptides with sequences that recognize inorganics with high specificity and to use them as nucleators, growth modifiers, and catalyzers in producing controlled material structures.8,9 Recently combinatorial biology methods, such as phage and cell-surface display, have been adapted to select inorganic binding polypeptides.9-13 Although the nature of * Corresponding author. E-mail:
[email protected]. † Materials Science & Engineering, University of Washington. ‡ Chemical Engineering, University of Washington. § Istanbul Technical University. 10.1021/nl048425x CCC: $30.25 Published on Web 02/16/2005
© 2005 American Chemical Society
binding of a polypeptide to a given inorganic has not been well understood yet, many short polypeptide sequences specific to metals (Au, Ag, Pt), oxides (silica, calcite, magnetite, zinc oxide, cuprous oxide), and semiconductors (cadmium sulfide, zinc sulfide) have been identified and used in the proof-of-principle synthesis, morphogenesis, and assembly of inorganics, building blocks for future engineering materials.14 Many inorganic-binding sequences are often selected with varying affinities to the same material (e.g., recognized by immunofluorescence microscopy) but without a high degree of specific intermolecular homology.9-13 The nature of molecular recognition of a given polypeptide, which leads to binding with a certain degree of affinity and specificity, has to be known so that it can be of utility in rational applications as a molecular building block.8,9,14 This knowledge is also essential for post-selection engineering of the inorganic-binding polypeptide for further tailoring its structure and, therefore, its function. Experimental as well as modeling studies toward this understanding have so far been lacking.15,16 Here we present a preliminary computational approach that would shed more light into the understanding of molecular recognition of inorganic-binding polypeptides. Specifically, we used phage display selected septapeptides with different levels of affinity (as determined by immunofluorescence microscopy) to metallic platinum.17 The presence of interface-reactive side chains, their special arrangement, and backbone stereochemistry represent plau-
sible determinants of peptide-interface recognition and adsorption orientation on crystalline surfaces.18 The conformational analyses of the selected sequences were carried out in vacuum for simplicity. In these analyses, HyperChem 7.5 molecular modeling system19 and the CHARMM 22 force field parameters were used.20 Five different polypeptides were chosen as representatives of these three affinity groups to understand polypeptide-substrate binding mechanism: strong binders, SD152 (PTSTGQA) and SD60 (QSVTSTK); moderate binders, SD128 (LGPSGPK); and weak binders, SD1 (APPLGQA) and SD6 (LNDGHNY).17 In these calculations the following assumptions were taken into account: there is no strong chemical bonding between the substrate and the polypeptide and for simplicity the substrate Pt atoms were taken as neutral in terms of electronic charges. The missing force field parameters between the protein and substrate atoms21 were calculated using Lorentz-Berthelot + rmin and i,j ) combination rules;22 namely, σi,j ) rmin i j xij, where is an energy scale (the minimum of the potential), σ is a length scale (the interatomic distance at which the potential is zero), and rmin is the van der Waals radii for a given atom i and j. In our modeling studies, we modified the selected Ptbinder sequences by the addition of cysteine residues at both N- and C-termini and manually constrained with a disulfide bridge to mimic experimentally used constrained peptide libraries.17 The energy of the resulting constrained peptide was then minimized in vacuum using the conformational analysis tool of the HyperChem 7.5 molecular modeling system.19 The “conformational search” module is a program for finding low energy conformations of molecular systems in HyperChem by varying the chosen dihedral angles. The method involves random variation of dihedral angles to generate new initial structures to perform energy minimization. During the energy minimization, low-energy unique conformations are stored while high-energy or duplicate structures are discarded. By using the conformational search module, 1000 different local minima were found on the potential energy surface and the lowest one was chosen as the global minimum or the lowest energy conformation for each peptide. To explore the binding mechanism between selected polypeptides and the inorganic Pt powder substrate, three different crystallographic surfaces (representing the powders used) were prepared using the fcc lattice parameter of 3.920 Å with periodic boundary conditions in the lateral dimensions. The crystallographic substrate thickness was determined such that the Lennard-Jones 6-12 potential, seen by a single C atom when it approaches the given Pt surface, is converged with the increase in the atomic layers, i.e., about five atomic Pt layers as seen in Figure 1a for the Pt (100) case. The overall periodic potentials of a given Pt surface that mimic the surface atomic periodicities can be seen in Figure 1b. The energy minimized peptide structures were placed on different symmetry centers of the Pt crystal surfaces at 5 Å separation, at which the polypeptide starts to sense the substrate (Figure 1c). At least 72 different initial configurations were prepared by rotating (θ) and tilting (φ) 416
Figure 1. (a) Lennard-Jones 6-12 potential between a single C atom and the substrate as a whole for various substrate thicknesses. (b) Surface potentials of Pt (100), (110), and (111). (c) Preparation of initial configurations of SD152-Pt substrate by rotating (θ) and tilting (φ) the polypeptide on top of substrate.
the molecule while keeping the substrate rigid for each peptide-substrate pair. The overall system of these structures was then energy minimized by allowing the protein to relax only on the solid surface. The binding energies are calculated by using the following formula: ∆Ebinding ) (Esubstrate + Epeptide) - Esubstrate - peptide
(1)
where, E represents the most favorable energy state for all three cases. Nano Lett., Vol. 5, No. 3, 2005
Table 1. Normalized Peptide-Substrate Binding Energies for Different Peptides on Top of Different Crystallographic Pt Surfaces ∆Ebinding
Figure 2. Comparison of (a) SD152 and (b) SD60 polypeptide molecular architecture with (opaque) and without (transparent) the platinum crystallographic surface.
The binding energy is related to the size of the peptide, in particular to the area on the peptide facing, and in close proximity to, the inorganic surface. The overall binding energies, therefore, were normalized with respect to N, the number of atoms in the peptide relatively closer to the surface, i.e., within a thickness of 5 Å, beyond which the effect of the substrate diminishes (Figure 1c): ∆Ebinding )
∆Ebinding N
(2)
The normalized peptide-substrate binding energies are listed in Table 1. These normalized binding energies have a good agreement with the relative binding strengths experimentally determined qualitatively by fluorescence microscopy. It should be noted that these energies are calculated with assumed force field parameters (described above), without the effects of electrostatic interaction, which may be negligible in the present case, between the peptide and the substrate, and the effect of water or solvation. Therefore, the values in the table are only relative energies compared to each other, and should be used as such. Nano Lett., Vol. 5, No. 3, 2005
strong
moderate
weak
Pt crystallographic surface
SD60
SD152
SD128
SD1
SD6
(100) (110) (111)
0.607 0.584 0.625
0.551 0.557 0.473
0.520 0.513 0.512
0.466 0.461 0.484
0.444 0.425 0.455
These normalized binding energies show that SD60 binds more strongly to all three crystallographic surfaces than does any other peptide. Although SD60 binds to the (111) surface better than the other surfaces, SD152 seems to bind both (100) and (110) better than the (111) surface. These differences may originate from how well the polypod architecture of a given polypeptide spatially matches the atomic grooves and rows of an inorganic crystallographic surface. As shown in Figure 2, the relaxation process is a result of reconformation of the polypeptides on the solid surface by the readjustment of its backbone and side chains to conform into the underlying via the fitting of its protrusions along the channel and shrinking in the direction perpendicular to the surface. The most stable protein-substrate binding configuration was obtained by comparing these structures in terms of overall energy of the system. In this comparison, the cysteine binding configurations were excluded since these amino acids do not play any role in the binding because of the constrained nature of the peptide. The displayed peptide is fused to pIII phage coat protein via the cysteine residues. The model structure of SD60 (one of the strong binding polypeptides) is seen on different Pt crystallographic surfaces in Figure 3, displaying top, front, and side views of the lowest energy conformation. A close inspection shows formation of protruding molecular structure through the peptide atoms closest to the underlying Pt substrate. For a comparison, the top views of conformations displaying similar architectures are shown in Table 2 for all binders chosen for this study. In all cases, the septapeptides conform to certain molecular architectures containing polypod structures that spatially match with atomic grooves and pits on the surfaces. As we analyze below, while the molecular recognition may originate from how well the polypod polypeptide architecture spatially matches the chosen crystallographic surfaces, the degree of binding may originate from the types of side chains in the polypods and degree of proximity of the charged or polar groups (e.g., hydroxyl and amine) to the metal surface. The observation of the presence of polypod molecular architecture in all three groups of Pt binders may seem to be surprising at first sight. First of all, this molecular conformation may provide a physical stability to any of the molecules on the inorganic surface (Figure 4). Second, it should consequently be the reactive groups that are in close contact with the inorganic surface that could bring about the main difference in (chemical) binding affinity. It is well417
Table 2. List of Binding Amino Acids, Groups and Atoms that Form the Polypod Molecular Architecturea
Figure 3. Top, front and side views of SD60 on top of (a) (100), (b) (110) and (c) (111) Pt surfaces. The side chains are colored according to CQSVTSTKC (C)yellow, Q)dark blue, S)red, V)gray, T)light blue, K)pink) . The Pt atoms were rendered by overlapping spheres whereas the peptides were rendered van der Waals radii of the atoms of the secondary structure of the protein.
known that any difference in the properties of amino acids, such as size, polarity, or charge, would significantly affect the folded conformation of a protein and its interactions with other molecules. To understand binding mechanisms we combined our conformational analysis with the experimental binding affinity results by assuming that these minimum energy conformations are also valid in real experimental conditions, even though these calculations were performed by considering only van der Waals interactions. A detailed analysis of the contact points of the amino acids among the three groups of polypeptides (Table 2) shows that the polar groups may be playing an important role, i.e., as the binding strength increases, polar amino acid content increases as well. Close inspection of the molecular architectures shows that even strong binders among themselves display significant differences in their conformations on the metal surfaces. Specifically, two oxygen and one nitrogen atoms in the footprint of SD152 on Pt(110) might bring a tighter binding compared to SD60 having one oxygen from the hydroxyl of serine forming a polypod architecture. Among the weak binders, the common feature was that neither oxygen nor nitrogen atoms were observed at contact points, except the case of one nitrogen atom of an amine group of glutamine in SD1 and only on the Pt(110) surface. In SD128, a moderate binder, only one oxygen, from the hydroxyl of serine is observed at contact point, although the peptide forms many contacts with the Pt surface at neutral protrusions. In the amino acid analysis, we have also observed glycine and proline, which are unique in influencing the conformation of the polypeptides. Because glycine does not have a side chain it provides a local flexibility to polypeptides for 418
a Protein atoms represented by color spheres are the legs of the polypods touching the Pt(110) surface.
Figure 4. Pseudo 3D view of polypod molecular architectures of (a) SD152 (CPTSTGQAC) and (b) SD60 (CQSVTSTKC) on Pt(110) surface. (C)yellow, P)gold, Q)dark blue, S)red, V)gray, T)light blue, K)pink, G)green, A)brown).
adapting to many conformations including turns, as it is the case in the constrained library. In contrast, another frequently Nano Lett., Vol. 5, No. 3, 2005
observed amino acid is proline, especially among the weaker binders. It is a rigid amino acid and therefore restricts flexibility to the peptide providing a rigid backbone to the polypod architecture without reactive (polar) contact points. Modeling of proteins on inorganic surfaces has so far been limited mostly to globular proteins on substrates of lower atomic number elements, which allows the problem to be treated with semiempirical methods.23,24 Many of these approaches considered only electrostatic and van der Waals interactions.25 To understand the binding mechanism further in short polypeptides, other interactions between the contact atoms and the metal surface need to be examined.26,27 Future models need to be performed in water and by considering more accurate force field parameters between protein and metal atoms. As we show here, even in vacuum conditions, platinum binders conform into polypod molecular architectures and that are directed by the lattice structure of the crystallographic metal surfaces. Molecular architectures on Pt (110) surface, for example, are good examples to observe the effect of atomic arrangement such that pods are elongated along the channels in the 〈110〉 directions. The reactive groups on the polypod, through which the peptide binds to the inorganic surface, then direct the strength of the binding. To our knowledge, polypod molecular architecture of engineered short polypeptides that conform onto metal surfaces is a novel concept. As shown here, molecular architecture derives rationally designed short polypeptides to bind and align in predictable geometry on crystallographic solid surfaces. These polypeptides could then be used as molecular erectors for functional brace proteins that can be engineered recombinantly and fused genetically, a great utility for protein-based supramolecular structures and microarrays such as in biosensors and proteomics.28,29 Acknowledgment. This research was supported by the U.S. Army Research Office through the Defense University Research Initiative in NanoTechnology (ARO-DURINT: DAAD19-01-1-04999). Supporting Information Available: Top, front and side views of SD152, 60, 128, 1 and 6 on top of (100), (110) and (111) Pt surfaces and detailed list of binding amino acids, groups and atoms that form the polypod molecular architecture on top of (100), (110) and (111) Pt surfaces. This material is available free of charge via the Internet at http:// pubs.acs.org.
Nano Lett., Vol. 5, No. 3, 2005
References (1) Ratner, B.; Schoen, F.; Hoffman, A.; Lemons, J. (Eds.) Biomaterials Science: Introduction to Materials in Medicine; Academic Press: San Diego, 1996. (2) Sakiyama-Elbert, S. E.; Hubbell, J. A. Annu. ReV. Mater. Res. 2001, 31, 183-201. (3) Niemeyer, C. M. Angew. Chem., Int. Ed. 2001, 40, 4128-4158. (4) Mrksich, M. Curr. Opin. Chem. Biol. 2002, 6, 794-797. (5) Sarikaya, M. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 14183-14185. (6) Mann, S. Nature 1988, 332, 119. (7) Weiner S.; Addadi, L. J. Mater. Chem. 1997, 7, 689-702. (8) Seeman, N. C.; Belcher, A. M. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 6452-6455. (9) Sarikaya, M.; Tamerler, C.; Schulten, K.; Jen, A.; Baneyx, F. Nature Mater. 2003, 2, 577. (10) Brown, S. Nature Biotechnol. 1997, 15, 269-272. (11) Whaley, S. R.; English, D. S.; Hu, E. L.; Barbara, P. F.; Belcher, M. A. Nature 2000, 405, 665-668. (12) Naik, R. R.; Brott, L.; Carlson, S. J.; Stone, M. O. J. Nanosci. Nanotechnol. 2002, 2, 1-6. (13) Thai, C. K.; Dai, H.; Sastry M. S. R.; Sarikaya, M.; Schwartz, D. T.; Baneyx, F. Biotechnol. Bioeng. 2004, 87, 129. (14) Sarikaya, M.; Tamerler, C.; Schwartz, D. T.; Baneyx, F. Annu. ReV. Mater. Res. 2004, 34, 377. (15) Zhou, J.; Chen, S.; Jiang, S. Langmuir 2003, 19, 3472-3478. (16) Gray, J. J. Curr. Opin. Struct. Biol. 2004, 14, 110. (17) Dincer, S.; Tamerler, C.; Heidel, D.; Oren, E. E.; Sarikaya, M. 2004, submitted. (18) Michenfelder, M.; Fu, G.; Lawrence, C.; Weaver, J. C.; Wustman, B. A.; Taranto, L.; Evans, J. S.; Morse, D. E. Biopolymers 2003, 70, 522-533. (19) Hyperchem 7.5, Hypercube Inc., Gainesville, FL, 2003. (20) MacKerell, A. D., Jr.; Bashford, D.; Bellott, M.; Dunbrack, R. L., Jr.; Evanseck, J. D.; Field, M. J.; Fischer, S.; Gao, J.; Guo, H.; Ha, S.; Joseph-McCarthy, D.; Kuchnir, L.; Kuczera, K.; Lau, F. T. K.; Mattos, C.; Michnick, S.; Ngo, T.; Nguyen, D. T.; Prodhom, B.; Reiher, W. E., III; Roux, B.; Schlenkrich, M.; Smith, J.C.; Stote, R; Straub, J.; Watanabe, M.; Wio´rkiewicz-Kuczera, J.; Yin, D.; Karplus, M. J. Phys. Chem. B 1998, 102, 3586-3616. (21) Allen, M. P.; Tildesley, D. J. Computer Simulations of Liquids; Clarendon Press: Oxford, U.K., 1987; p 21. (22) Zhu, S.-B.; Philpott M. R. J. Chem. Phys. 1994, 100, 6961-6968. (23) Shen Y.-J.; Tsao, H.-K.; Zhou, J.; Jiang, S. Phys. ReV. E. 2002, 66, 011911. (24) Roth C. M.; Lenhoff, A. M., Surf. Sci. Ser. 1998, 75, 89-118. (25) Castells, V.; Yang, S.; Van Tassel, P. R. Phys. ReV. E. 2002, 65, 031912. (26) Braun, R.; Sarikaya, M.; Schulten, K. S. J. Biomater. Sci. Polym. Ed. 2002, 13, 747-757. (27) Brown, C. L.; Aksay, I. A.; Saville, D. A.; Hechst M. H. J. Am. Chem. Soc. 2002, 124, 6846. (28) Chicurel, M. E.; Dalma-Weiszhausz, D. D. Pharmacogenomics 2002, 3, 589-601. (29) Yarmush, M. L.; Yarayam, A. Annu. ReV. Biomed. Eng. 2002, 4, 349-373.
NL048425X
419
