Self-Assembly of Fibronectin Mimetic Peptide-Amphiphile

(2) α5β1 integrin is a common integrin that binds to the ECM protein, fibronectin. .... high performance liquid chromatography (HPLC) on an Agilent ...
1 downloads 0 Views 5MB Size
Download PDF
pubs.acs.org/Langmuir © 2009 American Chemical Society

Self-Assembly of Fibronectin Mimetic Peptide-Amphiphile Nanofibers Emilie L. Rexeisen,† Wei Fan,† Todd O. Pangburn,† Rajiv R. Taribagil,† Frank S. Bates,† Timothy P. Lodge,†,‡ Michael Tsapatsis,† and Efrosini Kokkoli*,† †

Department of Chemical Engineering and Materials Science and ‡Department of Chemistry, University of Minnesota, Minneapolis, Minnesota 55455 Received July 14, 2009. Revised Manuscript Received October 8, 2009

Single-tailed peptide-amphiphiles have been shown to form nanofibers in solution and gel after screening of their electrostatic charges, and those containing cell-binding motifs are promising as tissue engineering scaffolds. A fibronectin-mimetic peptide sequence was developed, containing both the primary binding domain RGD and the synergy binding domain PHSRN, which has shown superior cell adhesion properties over simple RGD sequences and fibronectin in 2D culture. In order to test this sequence in a 3D environment in the future, we have designed a C16 singletailed peptide-amphiphile, PR_g (with a peptide headgroup of GGGSSPHSRN(SG)5RGDSP), that forms nanofibers and a gel in solution without any screening of its positive charge. In this study, we characterized the self-assembly properties of the PR_g peptide-amphiphile via critical micelle concentration (CMC) measurements, circular dichroism (CD) spectroscopy, cryo-transmission electron microscopy (cryo-TEM), small angle neutron scattering (SANS), and rheology measurements. The CMC of the PR_g amphiphile was determined to be 38 μM. CD measurements showed that even though the peptide formed an unordered secondary structure, the peptide-amphiphile’s spectrum after aging resembled more the spectrum of an Rþβ protein. Cryo-TEM images of a 100 μM peptide-amphiphile solution showed individual nanofibers with a diameter of approximately 10 nm and lengths on the order of several micrometers. Images taken at higher concentrations (1 mM) show a high degree of bundling among the nanofibers, and at even higher concentrations (3 and 4 mM) SANS measurements also indicated that the peptide-amphiphile formed rod-shaped structures in solution. The peptide-amphiphile gel was monitored by parallel-plate rheometry, and the elastic modulus (G0 ) was greater than the viscous modulus (G00 ), which indicates that PR_g forms a gel. The shear modulus for a 2 day old gel was measured to be approximately 500 Pa, which is within the modulus range for living tissue; thus, the PR_g gel shows potential as a possible scaffold for tissue engineering.

*Corresponding author. Telephone: 612-626-1185. Fax 612-626-7246. E-mail: [email protected].

angiogenesis5 and is also active in diseases such as adenovirus infection,6 Alzheimer’s disease,7 and breast,8 prostate,9 and colorectal cancers.10 One option to mediate cell adhesion to a tissue engineering scaffold is to include fibronectin in the design. Native proteins are easily recognized by cells’ receptors and are fully biocompatible, but their large size makes it difficult to control the orientation and presentation of the proper binding domain. Proteins are also susceptible to denaturation by heat or chemical treatment, which renders them ineffective.1 Researchers frequently choose to design a mimetic peptide that contains the sequence of the desired binding domain from the protein. Peptides are much less susceptible to denaturation and are easier to control in terms of orientation and presentation of the binding domain to the cells. The most common biomimetic peptide sequence incorporated into tissue engineering scaffolds is RGD, which is an amino acid sequence found in cell binding domains in ECM proteins such as fibronectin, vitronectin, and type I collagen.11 In effort to increase RGD binding specificity, researchers have tried constraining the peptide sequence in a cyclic conformation and flanking with various amino acids designed to increase selectivity for an individual cell receptor.11

(1) Alberts, B.; Johnson, A.; Lewis, J.; Raff, M.; Roberts, K.; Walter, P. Molecular Biology of the Cell, 4th ed.; Garland Science: New York, 2002. (2) Cukierman, E.; Pankov, R.; Stevens, D. R.; Yamada, K. M. Science 2001, 294(5547), 1708–1712. (3) Pierschabacher, M. D.; Ruoslahti, E. Nature 1984, 309, 30–33. (4) Livant, D. L.; Brabec, R. K.; Kurachi, K.; Allen, D. L.; Wu, Y.; Haaseth, R.; Andrews, P.; Ethier, S. P.; Markwart, S. J. Clin. Invest. 2000, 105, 1537–1545. (5) Kim, S.; Bell, K.; Mousa, S. A.; Varner, J. A. Am. J. Pathol. 2000, 156(4), 1345–1362. (6) Davison, E.; Diaz, R. M.; Hart, I. R.; Santis, G.; Marshall, J. F. J. Virol. 1997, 71, 6204–6207.

(7) Matter, M. L.; Zhang, Z.; Nordstedt, C.; Ruoslahti, E. J. Cell Biol. 1998, 141, 1019–1030. (8) Pec, M. K.; Artwohl, M.; Fernandez, J. J.; Souto, M. L.; de la Rosa, D. A.; Giraldez, T.; Valenzuela-Fernandez, A.; Diaz-Gonzalez, F. Exp. Cell Res. 2007, 313(6), 1121–1134. (9) van Golen, K. L.; Bao, L. W.; Brewer, G. J.; Pienta, K. L.; Kamradt, J. M.; Livant, D. L.; Merajver, S. D. Neoplasia 2002, 4, 373–379. (10) Varner, J. A.; Cheresh, D. A. Curr. Opin. Cell Biol. 1996, 8, 724–730. (11) Tirrell, M.; Kokkoli, E.; Biesalski, M. Surf. Sci. 2002, 500(1-3), 61–83.

Introduction Tissue engineering is a promising field of research combining living cells in engineered tissue for the replacement of diseased tissue. In order to facilitate the growth of cells, the engineered scaffold needs to provide structure to the tissue, support for the cells, and a mechanism to allow cell attachment. In vivo, cells are surrounded by the extracellular matrix (ECM), which is a network of proteins that provide structure and support to the tissue. Cells bind to the ECM through membrane receptors called integrins.1 In 3D culture, the adhesion domains appear to be a colocalization of focal adhesions and fibrillar adhesions. These cell-matrix adhesions have been termed 3D adhesions, and they require a 3D environment, R5β1 integrin, fibronectin, and pliability for formation.2 R5β1 integrin is a common integrin that binds to the ECM protein, fibronectin.1 Specifically, it binds to the arginine-glycine-aspartic acid (RGD) primary binding domain and proline-histidine-serine-arginine-asparagine (PHSRN) synergy sequence found in the 9-10 repeats of type III fibronectin.3 The R5β1 integrin is necessary for wound healing4 and

Langmuir 2010, 26(3), 1953–1959

Published on Web 10/30/2009

DOI: 10.1021/la902571q

1953

Article

A fibronectin mimetic peptide, PR_b (KSSPHSRN(SG)5RGDSP), was synthesized that incorporates both the primary RGD binding domain and the synergy binding site PHSRN. The PR_b peptide was designed to take into account the distance and the overall hydrophobicity/hydrophilicity of the amino acids between the two binding domains in fibronectin.12 X-ray crystallography has shown that the two binding domains are 30-40 A˚ apart.13 Fully extended peptides measure 3.7 A˚/amino acid.14,15 PR_b contains 10 amino acids linking the two domains and thus has a maximum theoretical linker length of 37 A˚. If the native fibronectin protein is unwound and the amino acids located between the two binding domains are counted, the ratio of hydrophobic to hydrophilic amino acids is nearly 1:1,16 so PR_b’s linker was designed with an alternating serine/glycine motif. The PR_b peptide was coupled to a double hydrocarbon tail, forming a peptide-amphiphile. The PR_b peptide-amphiphile was deposited onto hydrophobized mica surfaces utilizing the LangmuirBlodgett technique, and in both atomic force microscopy (AFM) adhesion experiments and 2-D cell adhesion experiments PR_b functionalized surfaces have been shown to outperform fibronectin surfaces and other fibronectin-mimetic peptide surfaces in terms of cell adhesion and spreading, cytoskeletal organization, and ECM fibronectin production.12,17 Blocking experiments were performed and showed that the sequence is specific to R5β1 integrin.12 Several other fibronectin mimetic peptide sequences similar to PR_b were designed (PR_a, PR_c, PR_d, PR_e, and PR_f) that varied the hydrophilicity and length of the linker between the two binding domains and the length of the spacer at the N-terminus of the peptides, and results showed that the PR_b peptide-amphiphile functionalized surfaces outperformed all the other peptide designs in terms of AFM adhesion forces and 2-D cell adhesion.17 Peptide-amphiphiles can self-assemble into a number of different structures including micelles, vesicles, monolayers, bilayers, nanofibers, and ribbons.18,19 Peptide-amphiphiles with a single tail of 10-22 carbons tend to form nanofibers in solution and show promise as a useful structure for tissue engineering, as they have been shown to form self-supporting gels.20,21 Nanofibers are also called cylindrical micelles and have their hydrocarbon tail situated in the core of the cylinder, and the peptide headgroup forms the outer layer of the cylinder. Self-assembly of the singletailed peptide-amphiphiles is usually triggered by a change in pH,22 by addition of multivalent ions,23 or by mixing oppositely charged peptide-amphiphiles.24 Each of these methods screens the electrostatic repulsion between the peptide-amphiphile molecules and allows for self-assembly. Peptide-amphiphile nanofibers have

Rexeisen et al.

been used as templates to orient hydroxyapatite mineralization,22 as magnetic resonance imaging (MRI) contrast agents,25 and as scaffolds for cell encapsulation23 and wound repair.26 Previous research in the literature has probed the structure of single-tailed peptide-amphiphiles and has shown that the outer surface of the nanofiber is highly hydrated, whereas significantly less water is found in the hydrophobic core.27 Polarization modulation infrared reflection-absorption spectroscopy has indicated that peptide-amphiphile nanofibers exhibit extensive β-sheet-like hydrogen bonding between neighboring peptide-amphiphiles along the length of the nanofiber.28 This hydrogen bonding is especially important between the four amino acids closest to the hydrocarbon tail and has been shown to be vital for gelation.29 The goal of this study was to design a fibronectin-mimetic peptide-amphiphile which would form nanofibers in solution and gel, allowing for future cell encapsulation experiments. PR_b (KSSPHSRN(SG)5RGDSP) has shown excellent cell adhesion properties, but the PR_b peptide sequence is not optimal for nanofiber formation because it contains both lysine (K) and proline (P) residues at its N-terminus (lysine and proline are not desirable for the design of peptide-amphiphiles that form nanofibers with β-sheet characteristics, as they tend to break the β-sheet formation in proteins30). Thus, a new peptide was designed and called PR_g (GGGSSPHSRN(SG)5RGDSP); in this design, the lysine was removed from the spacer and just three glycine residues were added at its N-terminus to push the proline to the sixth residue without further increasing the peptide length. Glycine was chosen, as it is a small amino acid and will not prevent β-sheet formation. We hypothesized that a peptideamphiphile comprising the PR_g peptide coupled to a single hydrocarbon tail would form nanofibers in solution and selfsupporting gels, in the absence of any external stimuli such as pH changes or addition of ions, which is different from what has been shown before for other peptide-amphiphiles that self-assemble into nanofibers that gel. Thus, the focus of this work was to synthesize monoalkyl PR_g peptide-amphiphiles and characterize their self-assembly. We determined the critical micelle concentration (CMC) and zeta potential of the PR_g peptide-amphiphile and investigated the secondary structure of the PR_g peptide and peptide-amphiphile via CD measurements. Cryo-transmission electron microscopy (cryo-TEM) and small angle neutron scattering (SANS) were employed to study the self-assembled structure that the amphiphiles form, and rheology was used to study the mechanical properties and gelation of the PR_g peptide-amphiphile.

Materials and Methods (12) Mardilovich, A.; Craig, J. A.; McCammon, M. Q.; Garg, A.; Kokkoli, E. Langmuir 2006, 22, 3259–3264. (13) Leahy, D. J.; Hendrickson, W. A.; Aukhil, I.; Erickson, H. P. Science 1992, 258(5084), 987–991. (14) Idiris, A.; Alam, M. T.; Ikai, A. Protein Eng. 2000, 13(11), 763–770. (15) Kokkoli, E.; Ochsenhirt, S. E.; Tirrell, M. Langmuir 2004, 20(6), 2397–2404. (16) Mardilovich, A.; Kokkoli, E. Biomacromolecules 2004, 5, 950–957. (17) Craig, J. A.; Rexeisen, E. L.; Mardilovich, A.; Shroff, K.; Kokkoli, E. Langmuir 2008, 24(18), 10282–10292. (18) Israelachvili, J. N. Intermolecular and surface forces; Academic Press: San Diego, 1992. (19) Kokkoli, E.; Mardilovich, A.; Wedekind, A.; Rexeisen, E. L.; Garg, A.; Craig, J. A. Soft Matter 2006, 2(12), 1015–1024. (20) Jun, H. W.; Paramonov, S. E.; Hartgerink, J. D. Soft Matter 2006, 2(3), 177–181. (21) Hartgerink, J. D.; Beniash, E.; Stupp, S. I. Proc. Natl. Acad. Sci. U.S.A. 2002, 99(8), 5133–5138. (22) Hartgerink, J. D.; Beniash, E.; Stupp, S. I. Science 2001, 294(5547), 1684– 1688. (23) Beniash, E.; Hartgerink, J. D.; Storrie, H.; Stendahl, J. C.; Stupp, S. I. Acta Biomater. 2005, 1(4), 387–397. (24) Behanna, H. A.; Donners, J. J. J. M.; Gordon, A. C.; Stupp, S. I. J. Am. Chem. Soc. 2005, 127(4), 1193–1200.

1954 DOI: 10.1021/la902571q

All synthesis chemicals were obtained from Sigma-Aldrich with the exception of O-(benzotriazole-1-yl)-N,N,N0 ,N0 -tetramethyluronium hexofluorophosphate (HBTU), which was obtained from EMD Biosciences, and diisopropylethylamine, which was from Fisher Scientific. Water was obtained from a Milli-Q water system and was purified to a resistivity of 18.2 MΩ 3 cm. Peptide-Amphiphile Synthesis and Purification. The protected peptide (GGGSSPHSRN(SG)5RGDSP) was synthesized by the Oligonucleotide and Peptide Synthesis Facility at the (25) Bull, S. R.; Guler, M. O.; Bras, R. E.; Meade, T. J.; Stupp, S. I. Nano Lett. 2005, 5(1), 1–4. (26) Tysseling-Mattiace, V. M.; Sahni, V.; Niece, K. L.; Birch, D.; Czeisler, C.; Fehlings, M. G.; Stupp, S. I.; Kessler, J. A. J. Neurosci. 2008, 28(14), 3814–3823. (27) Tovar, J. D.; Claussen, R. C.; Stupp, S. I. J. Am. Chem. Soc. 2005, 127(20), 7337–7345. (28) Jiang, H. Z.; Guler, M. O.; Stupp, S. I. Soft Matter 2007, 3(4), 454–462. (29) Paramonov, S. E.; Jun, H. W.; Hartgerink, J. D. J. Am. Chem. Soc. 2006, 128(22), 7291–7298. (30) Levitt, M. Biochemistry 1978, 17(20), 4277–4284.

Langmuir 2010, 26(3), 1953–1959

Rexeisen et al. University of Minnesota using standard fluorenylmethoxycarbonyl (Fmoc) solid phase peptide synthesis on a PAL resin. This resin chemistry leaves an amide group on the C-terminus upon cleavage. The Fmoc protecting group was removed in a 20% solution of piperidine in dimethylformamide (DMF). A hexadecanoic acid (C16) tail was coupled to the N-terminus of the peptide amide with 3 mol equiv of the fatty acid, 3 mol equiv of HBTU, and 4.5 mol equiv of diisopropylethylamine in DMF for 4.5 h. The Kaiser test showed complete coupling of tails to the peptide. The resin beads were washed with DMF, dichloromethane, and methanol and then dried under vacuum for 1 h. The peptideamphiphiles were cleaved from the resin, and the amino acid side groups were deprotected in a solution of 90% trifluoroacetic acid (TFA), 5% thioanisole, 3% 1,2 ethanedithiol, and 2% anisole for 2.5 h. The solution was rotoevaporated until the product started to precipitate. Ten times excess cold isopropyl ether was added, and the resulting slurry was filtered. The peptide-amphiphile was redissolved in Milli-Q water, lyophilized, and purified using reversed phase high performance liquid chromatography (HPLC) on an Agilent 1100 Series system using a Waters XTerra Prep MS C18 column (5 μm beads, 125 A˚ pore size, 10  150 mm). The elution gradient used 0.1% TFA in Milli-Q water and 0.1% TFA in acetonitrile. The purified product was analyzed with matrixassisted laser desorption/ionization time-of-flight (MALDITOF) on a Bruker Reflex III system to ensure the product had the expected molecular mass. Zeta Potential. The zeta potential of a 1 mM PR_g peptideamphiphile solution was measured on a ZetaPALS zeta potential analyzer (Brookhaven Instruments Corporation) using a clear polystyrene cuvette with 10 mm path length. The measurement was taken 20 min after the sample was mixed and again after 48 h. Data for each run were collected until the relative residual was less than 0.01. The mobility of the nanofibers was measured and was converted to zeta potential by the software using the Smoluchowski equation. Critical Micelle Concentration. The CMC was determined using the lipophilic dye, Nile red.31 A solution of 2.5 mg/mL Nile red in methanol was diluted to 200 μM Nile red in 50 mM TrisHCl buffer. A total of 12.5 μL of final dye solution was added to 12.5 μL peptide-amphiphile dissolved in Milli-Q water, and the resulting solution was pipetted into a black-walled 96-well plate to make peptide-amphiphile concentrations varying from 0 to 500 μM. The well plate was wrapped in aluminum foil and left to equilibrate for 18 h. The fluorescence was measured using a SpectraMAX GeminiXS plate reader, Molecular Devices, with excitation 550 nm and emission 635 nm. Circular Dichroism Spectroscopy. Measurements were taken on a JASCO J815 CD spectrapolarimeter at room temperature. Samples were either prepared in Milli-Q water at 100 μM or diluted to 100 μM from a 1 mM or 10 mM stock solution. The spectra were obtained from 300 to 180 nm at a bandwidth of 10 nm and were the average of five individual scans. A background spectra of Milli-Q water was subtracted from each spectrum. Small Angle Neutron Scattering. SANS measurements were performed at NIST on a 30 m instrument (NG-7). Samples of 1 wt % (4.12 mM) PR_g and 0.75 wt % (3.09 mM) PR_g were prepared with D2O the night before the experiment. Solutions were pipetted into banjo cells with a path length of ∼1 mm after the overnight equilibration. Sample to detector distances were 1, 3, and 13 m, giving a q-range of 0.003 > q (1/A˚) > 0.331. Neutrons with a wavelength of λ = 7 A˚ were used with a distribution of Δλ/λ = 0.11. The raw data were reduced to correct for D2O background, empty cell, and background scattering, sample transmission, and detector response. Cryo-Transmission Electron Microscopy. All samples were prepared in a controlled environment vitrification system. Sample solution (10 μL) was pipetted onto a copper TEM grid (31) Lau, C.; Bitton, R.; Bianco-Peled, H.; Schultz, D. G.; Cookson, D. J.; Grosser, S. T.; Schneider, J. W. J. Phys. Chem. B 2006, 110(18), 9027–9033.

Langmuir 2010, 26(3), 1953–1959

Article

Figure 1. PR_g peptide-amphiphile gel. A 10 mM solution of PR_g was prepared with Milli-Q water and allowed to sit overnight at room temperature. Figure 1A shows the peptide-amphiphile at the bottom of the screw-cap vial. Figure 1B shows the same vial tipped upside down to illustrate that the gel is self-supporting. coated with a polymer film (Ted Pella, Inc.). The sample was blotted to create a thin film and plunged into liquid ethane to vitrify the water. Images were collected using a JEOL JEM-1210 transmission electron microscope operating at 120 kV. The temperature in the transmission electron microscope was kept at -179 °C. A Gatan 724 multiscan digital camera was used to capture the images. Rheology. Rheological properties were measured using an AR-G2 apparatus from TA Instruments. Time sweep experiments were performed using a 40 mm parallel plate. PR_g amphiphile solutions (12 mM) in Milli-Q water were prepared and immediately put onto a Peltier plate at a temperature of 25 °C. The top parallel plate was lowered to make contact with the sample, and the gap was decreased until the sample completely filled the gap. A time sweep was performed at 0.5% strain and a frequency of 1 rad/s. A frequency sweep was performed on a 10 mM PR_g gel, which had been allowed to sit at room temperature in a sealed chamber before measurement. The sealed chamber was constructed by adhering a silicone O-ring to a glass coverslip using silicone rubber and covering to the top of the chamber with a glass slide lightly coated with vacuum grease. The glass coverslip was adhered to a Peltier plate with vacuum grease, and the Peltier plate kept the sample at 25 °C for the rheology measurement. A 13.07 mm parallel plate was oscillated at 1% strain, and the frequency was varied from 0.01 to 100 rad/s.

Results and Discussion The PR_g peptide was prepared using standard Fmoc-peptide synthesis chemistry, and a hexadecanoic acid (C16) tail was coupled to the N-terminus. When the peptide-amphiphile was dissolved in pure Milli-Q water at a concentration of 10 mM, it readily formed a self-supporting gel (Figure 1) without the addition of any charge-screening ions or oppositely charged peptide-amphiphiles as needed in previous studies.22-24 The zeta potential of a 1 mM solution of the PR_g peptide-amphiphile in Milli-Q water was 38.8 ( 1.8 and 42.9 ( 1.8 mV for the same sample measured at 20 min and 48 h, respectively, after dissolution. This indicates that the PR_g peptide-amphiphile is positively charged in Milli-Q water and that its zeta potential is not changing significantly with time. These numbers are consistent with values reported in the literature for similar molecules.32 The critical micelle concentration is the concentration at which micelles start to form spontaneously in solution. Micellization (32) Niece, K. L.; Czeisler, C.; Sahni, V.; Tysseling-Mattiace, V.; Pashuck, E. T.; Kessler, J. A.; Stupp, S. I. Biomaterials 2008, 29(34), 4501–4509.

DOI: 10.1021/la902571q

1955

Article

Figure 2. Critical micelle concentration: fluorescence intensity of the Nile red dye solubilized in PR_g micelles versus the PR_g amphiphile concentration. Samples were left to equilibrate for 18 h before measurements were taken. A linear regression fit (solid line) to the data gives a CMC of 38 μM. Background fluorescence was subtracted, and it is denoted by the dotted line. The inset graph is a zoomed in area focusing on the critical micelle concentration. The results are representative of n = 2 (two independent experiments performed on different days), but data are shown from a single experiment.

induces a change in many solution properties including surface tension, osmotic pressure, solubilization, and turbidity.33 A plot of the effect of concentration on any of these physical properties shows a marked change in slope at the CMC, and in this study we employed the dye solubilization method31 to determine the CMC. Nile red is a lipophilic dye that is quenched in water. When micelles form, the dye is solubilized in the hydrophobic interior of the micelle, and a fluorescent signal is detected. PR_g amphiphile solutions were prepared in an aqueous solution containing Nile red with the peptide-amphiphile concentration varying from 0 to 500 μM. Figure 2 shows the fluorescence intensity as a function of PR_g amphiphile concentration. Above the CMC, the fluorescence intensity increases linearly with concentration. A linear regression was fit to the data, and the CMC was determined to be the point at which the line crossed the intensity of the background fluorescence. The background fluorescence was calculated as the average intensity of the lowest prepared concentrations below the CMC. The critical micelle concentration of PR_g was 38 μM, and subsequent experiments were conducted at a minimum concentration of 100 μM to ensure micellization. The secondary structure of PR_g was examined with CD spectroscopy. Proteins with prominent R-helix domains exhibit CD spectra with a negative minimum at 222 nm, a weaker negative minimum at 208 nm, a positive maximum between 190 and 195 nm, and a crossover from positive to negative below 172 nm.34,35 Proteins that are primarily β-sheets display a negative minimum between 210 and 220 nm, a positive signal at ∼200 nm, a crossover from positive to negative at 185 nm, and a negative minimum around 170-180 nm.34,36 Many proteins contain both motifs and are called R-β proteins. Two specific types of R-β proteins are Rþβ proteins and R/β proteins. Rþβ proteins contain separate regions of R-helix and β-sheet, and an example of this type is lysozyme.37 The R/β proteins, such as leucine-rich-repeat proteins, contain alternating regions of R-helix and β-sheet.38 (33) Evans, D. F.; Wennerstr€om, H. The Colloidal Domain; VCH Publishers: New York, 1994. (34) Manavalan, P.; Johnson, W. C., Jr. Nature 1983, 305, 831–832. (35) Su, J. Y.; Hodges, R. S.; Kay, C. M. Biochemistry 1994, 33, 15501–15510. (36) Manning, M. C.; Woody, R. W. Biopolymers 1987, 26(10), 1731–1752. (37) Brahms, S.; Brahms, J. J. Mol. Biol. 1980, 138(2), 149–178. (38) Bella, J.; Hindle, K. L.; McEwan, P. A.; Lovell, S. C. Cell. Mol. Life Sci. 2008, 65(15), 2307–2333.

1956 DOI: 10.1021/la902571q

Rexeisen et al.

Figure 3. CD spectroscopy of the PR_g peptide and PR_g peptide-amphiphile (PA). The peptide sample was diluted to 100 μM from a 1 mM stock solution and shows an unordered structure. A 1 mM stock solution of the PA was prepared and was allowed to age for 0, 6, and 48 h after which a PA sample was diluted to 100 μM. CD spectra were taken immediately after dilution. After 48 h, the PA shows a spectrum that resembles more the spectrum of an Rþβ protein. The results are representative of n = 3 (three independent experiments performed on different days), but data are shown from a single experiment.

Both types of R-β protein display a similar CD spectra to the all R proteins in that they show negative minima at approximately 222 and 208 nm and a positive signal between 190 and 195 nm, but they show a crossover from positive to negative above 172 nm. For Rþβ proteins, such as lysozyme, the minimum around 208 nm dominates the spectrum and the minimum at 222 nm is usually a very shallow minimum.34 Figure 3 shows the CD spectra of the PR_g peptide and PR_g amphiphile. Spectra were collected down to 185 nm below which the water starts to absorb strongly, and the data are no longer accurate. All measurements were performed at 100 μM. A stock solution of 1 mM PR_g peptide (no hydrocarbon tail) was prepared and then diluted to 100 μM immediately prior to measurement. The PR_g peptide shows a minimum at 196 nm (Figure 3, red line). The peptide’s CD spectrum was not dependent on the age of the stock solution, and it showed the same spectra after aging the stock solution for 20 days (data not shown). The peptide shows no particular secondary structure and is classified as an unordered sequence. The peptide-amphiphile, however, shows a more interesting structure, as the amphiphile’s CD spectrum depended upon the age of the stock solution. The CD spectrum of a 100 μM sample prepared from a 1 mM stock solution immediately before the CD measurement (Figure 3, green line) is similar to the peptide’s spectrum. However, when the 1 mM stock solution was prepared and allowed to age for 6 h (Figure 3, blue line) before diluting to 100 μM for measurement, the spectrum shows a slight shift (the signal at 219 nm is much stronger, and the minimum has shifted right to 198 nm), and upon further aging of the 1 mM solution for 48 h (Figure 3, black line) or 20 days (data not shown) prior to dilution to 100 μM the spectrum shows a minimum at 203 nm and a less prominent minimum at 220 nm, thus resembling more the spectrum of an Rþβ protein than the spectrum of an all-β protein. To test further whether the CD spectrum was affected by the concentration of the stock solution, a 10 mM PR_g amphiphile stock solution was prepared and diluted to 100 μM at several time points for CD measurement (Figure 4). The solution exhibits the unordered secondary structure upon aging of the stock solution for 6 min (Figure 4, red line) and 1 h before dilution (Figure 4, green line), whereas after 2 h (Figure 4, blue line) the spectrum is similar to the spectrum observed in Figure 3 (black line) for the Langmuir 2010, 26(3), 1953–1959

Rexeisen et al.

Article

Figure 4. CD spectroscopy of a 100 μM PR_g amphiphile solution diluted from a 10 mM stock solution. Stock solution was diluted at 0.1, 1, 2, and 5 h after dissolution. At 0.1 h, the CD spectrum shows an unordered secondary structure, and by 2 h the spectrum resembles that of an Rþβ protein. At 5 h, the intensity of the signal has increased. The results are representative of n = 3 (three independent experiments performed on different days), but data are shown from a single experiment.

1 mM stock solution aged for 2 days. Upon further aging of the 10 mM solution for 5 h (Figure 4, black line) the CD signal intensity strengthened. The results from Figures 3 and 4 show that the CD spectra of the PR_g peptide-amphiphile match the spectra of Rþβ proteins, indicating that the PR_g peptide-amphiphile assembled structures contain separate regions of R-helix and β-sheet. This differs from the CD spectra of other nanofiber-forming peptide-amphiphiles. Previous research shows that nanofiber-forming peptideamphiphiles exhibit predominantly β-sheet secondary structure after the addition of charge-screening ions or mixing with oppositely charged peptide-amphiphiles.24,32 One study reported peptide-amphiphiles containing some R-helical structures at pH 10 in addition to β-sheet structures.39 Prior to addition of charge screening molecules, the peptide-amphiphiles typically show an unordered secondary structure similar to the one seen in this study for the PR_g peptide.24,32 In our case, however, the secondary structure forms over time without adding any charge-screening agents. Next, cryo-TEM was performed to visualize the self-assembled structures of the PR_g peptide-amphiphiles. We examined PR_g amphiphile solutions at 100 μM and 1 mM (Figure 5). The 100 μM peptide-amphiphile solution was prepared by diluting a 2 day old 1 mM stock solution. Figure 5A shows single nanofibers that all appear to have a diameter of approximately 10 nm and lengths in the micrometer range. The theoretical length of the peptideamphiphile molecule depends upon the structure. In a fully extended conformation, the unordered peptide headgroup measures 3.7 A˚/amino acid.15 The fully extended hydrocarbon tail is assumed to have C-C bonds 1.54 A˚ long and a bond angle of 109°. Therefore, a peptide-amphiphile with a 25 amino acid peptide headgroup and a C16 tail in a fully extended conformation would measure ∼11.25 nm long. If the peptide headgroup was in an all-β-sheet conformation, the length is 3.5 A˚/amino acid.1 Assuming the peptide headgroup adopts an R-helical formation, the length would only be 1.5 A˚/amino acid.1 This leads to a length of ∼5.75 nm for the entire amphiphile. Thus, the theoretical length of the peptide-amphiphile is estimated to be between 5.75 and 11.25 nm. The 10 nm nanofiber diameter measured by cryoTEM falls within the theoretical expectations and suggests that (39) Guler, M. O.; Claussen, R. C.; Stupp, S. I. J. Mater. Chem. 2005, 15(42), 4507–4512.

Langmuir 2010, 26(3), 1953–1959

Figure 5. Cryo-TEM images of PR_g nanofibers. (a) A 1 mM PR_g peptide-amphiphile stock solution was aged for 2 days then diluted to 100 μM and vitrified. The image shows several nanofibers in a “W” shape. The diameter of the fibers is approximately 10 nm. (b) A grid was prepared from a 2 day old 1 mM PR_g amphiphile. Fibers appear to be interacting in bundles and show some branching. These interactions at sufficiently high concentrations could give rise to gelation behavior.

the peptide headgroup may be tilted. Another grid was prepared with a 2 day old 1 mM solution (Figure 5B). At this higher concentration, PR_g forms nanofibers bundled together. These bundles branch and entangle with other nearby bundles. The neighboring nanofibers could form noncovalent cross-links, and this could give rise to the gelation of the PR_g amphiphile. The diameter of the bundled nanofibers can be up to 0.2 μm thick and many micrometers long. The images still show some single nanofibers with diameters of ∼10 nm that are not assembled into larger bundles. Cryo-TEM at higher concentrations becomes challenging due to difficulties associated with sample preparation. Therefore, at larger concentrations, we employed SANS, which in turn could not be used at very low concentrations because of insufficient signal-to-noise ratio. In future studies, we will attempt to study the same samples using both cryo-TEM and SANS. Two concentrations were used for the SANS experiments: 1 wt % (4.12 mM) and 0.75 wt % (3.09 mM) of PR_g peptide-amphiphile. When the intensity (I(q)) was normalized with respect to concentration and I(q) was plotted against the scattering vector, q, the two sets of data overlapped (data not shown), indicating that the scattering intensity is proportional to concentration. For clarity, Figure 6 shows the 1 wt % data only. The data show that the Guinier limit is not reached in the q range 0.003 > q (1/A˚) > 0.331. This implies that the radius of gyration for PR_g DOI: 10.1021/la902571q

1957

Article

Rexeisen et al.

Figure 6. SANS data of 1 wt % (4.12 mM) PR_g amphiphile dissolved in 100% D2O, plotted as intensity (I(q)) against scattering vector (q). The intensity first decays as q-1, and in the Porod scattering region the intensity decays as q-3.133. The results are representative of n = 2 (two independent experiments performed on different days), but data are shown from a single experiment.

Figure 7. PR_g gelation of a 12 mM sample in Milli-Q water monitored by oscillating rheometry at a frequency of 1 rad/s. Elastic modulus, G0 (open circles), and viscous modulus, G00 (filled circles), of a PR_g solution measured as a function of time. Gelation point (signified by G0 > G00 ) is observed at 30 min. The inset shows the crossover between G0 and G00 in greater detail. For clarity, only every 300th data point in the larger graph and every 20th data point in the inset were plotted. The results are representative of n = 3 (three independent experiments performed on different days), but data are shown from a single experiment.

is quite large. For cylinders, the radius of gyration, Rg, is related to the length, L, and radius, R, of the cylinder by eq 1. Rg 2 ¼

L 2 R2 þ 12 8

ð1Þ

Cryo-TEM showed that the length of the nanofibers is significantly greater than the radius, so the Rg will be primarily a function of the nanofiber length. The length of the nanofibers appears to be at least a micrometer long, so using a conservative estimate for Rg of 1 μm or higher, the Guinier limit would be q < 0.0001 A˚-1, which is outside the range of the present instrument. In the q-range from 0.0037 to 0.021 A˚-1, intensity decays with q-1. This indicates that PR_g forms an elongated rodlike structure and supports the hypothesis that PR_g forms cylindrical micelles. In the q-range from 0.043 to 0.183 A˚-1, intensity decays with q-3.133, which deviates from the expected q-4 for particles with sharp interfaces. This could be a result of a diffuse interface between the nanofiber and the solvent, or due to a polydispersity in the nanofiber radius, or is a result from a significant structure factor, or any combination of the above. When PR_g is dissolved in water at higher concentrations (6 mM or higher), it forms a self-supporting gel. To verify that PR_g actually forms a gel, rheology experiments were performed. The rheometer uses a Peltier plate as the lower plate that was kept at 25 °C. First, a 12 mM solution of PR_g was prepared and immediately pipetted onto the Peltier plate. A time sweep was performed using a frequency of 1 rad/s and a strain of 0.5% (Figure 7). At the gelation point the viscous modulus, G00 , is less than the elastic modulus, G0 , and this crossover is the defining characteristic between a gel and a viscous liquid. The gelation time of a 12 mM solution of PR_g amphiphile is approximately 30 min. The moduli continue to increase over the course of the measurement, indicating that the gel was still forming cross-links. After about 2 days, the moduli remain fairly constant. Another important characteristic of gels with a high degree of cross-linking is that G0 and G00 are minimally sensitive to the oscillation frequency. To test the degree of PR_g cross-links in the gel, a 10 mM solution of PR_g amphiphile was allowed to age in a sealed chamber for 48 h, and a frequency sweep was performed using a 13.07 mm parallel plate oscillating at 1% strain from 0.01 to 1958 DOI: 10.1021/la902571q

Figure 8. G0 and G00 of a PR_g gel. A 10 mM PR_g gel was prepared and aged for 48 h (black circles) or for 1 month (red squares) in a sealed chamber before rheometry measurement. Both G0 (open symbols) and G00 (filled symbols) were not significantly affected by changes in oscillation frequency. The results are representative of n = 2 (two independent experiments performed on different days), but data are shown from a single experiment.

100 rad/s (Figure 8). The data show that G0 and G00 are minimally affected by the frequency change between 0.1 and 10 rad/s and the moduli do not cross in the frequency range tested. This indicates that the PR_g nanofibers are forming cross-links between nanofibers and the gel behavior is not merely due to entangled nanofibers. In the case of PR_g, these cross-links are most likely noncovalent bonds, such as hydrogen bonds. To see if the gel would continue to stiffen over time, a 10 mM PR_g amphiphile gel was allowed to age in the sealed chamber for 1 month and then another frequency sweep was taken. The gel showed a slight increase in both G0 and G00 . Over the course of 1 month, some water evaporation is expected, but the use of a sealed chamber prevented most of the evaporation. An important consideration in tissue design is mimicking the mechanical properties of native tissues. Neural tissue has a modulus of approximately E ≈ 0.1 kPa and muscle tissue is E ≈ 10 kPa.40 At a frequency of 1 rad/s, the G0 of PR_g is ∼500 Pa. Thus, the modulus of elasticity (E) is approximately 1.5 kPa, which is within the range modulus of elasticity of living tissues. (40) Engler, A. J.; Griffin, M. A.; Sen, S.; Bonnetnann, C. G.; Sweeney, H. L.; Discher, D. E. J. Cell Biol. 2004, 166(6), 877–887.

Langmuir 2010, 26(3), 1953–1959

Rexeisen et al.

Article

Conclusion A fibronectin mimetic peptide-amphiphile was designed to contain both the primary and synergy binding domains of the fibronectin-R5β1 integrin binding site. The peptide, PR_g, was coupled to a single hydrocarbon tail to make a peptide-amphiphile that would self-assemble into nanofibers in solution. The critical micelle concentration of the PR_g amphiphile was found to be 38 μM, and both the peptide and peptide-amphiphile were analyzed using CD spectroscopy to determine their secondary structure. The peptide showed no ordered structure, whereas the peptide-amphiphile showed an Rþβ secondary structure after aging. Cryo-TEM was performed on PR_g amphiphile samples and showed single nanofibers with diameters of ∼10 nm and lengths well within the micrometer range. At higher concentrations, cryo-TEM suggests that the nanofibers bundle and twist together. SANS experiments were performed at even higher PR_g peptide-amphiphile concentrations, and analysis of the data indicates agreement with the hypothesis that the peptide-amphiphile forms a fundamental rodlike structure. However, the SANS data deviate significantly from the behavior expected for cylinders, indicating contributions due to polydispersity, due to a diffuse interface between the nanofiber and the solvent, or a structure factor. Finally, rheological measurements of PR_g gels indicate that the elastic modulus approaches that of a native

Langmuir 2010, 26(3), 1953–1959

tissue, which is important for future consideration as a tissue scaffold. Acknowledgment. This work was supported in part by the National Institute of Biomedical Imaging and Bioengineering (R03EB006125), the Camille Dreyfus Teacher-Scholar Awards Program, and by the MRSEC Program of the National Science Foundation under Award Number DMR-0819885. The cryoTEM was carried out in the Institute of Technology Characterization Facility, University of Minnesota, which has received capital equipment funding from the NSF through the MRSEC, ERC, and MRI programs. Part of this work was carried out in the Department of Chemistry Molecular Characterization Facilities, University of Minnesota, which have received capital equipment funding from the NSF through the MRSEC. We acknowledge the support of the National Institute of Standards and Technology, U.S. Department of Commerce, in providing the neutron research facilities used in this work. The CD spectroscopy work was carried out in the Biophysical Spectroscopy Facility in the Department of Biochemistry, Molecular Biology, and Biophysics at the University of Minnesota. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Biomedical Imaging and Bioengineering or the National Institutes of Health.

DOI: 10.1021/la902571q

1959