2694
Biomacromolecules 2009, 10, 2694–2698
Self-Assembly of Multidomain Peptides: Sequence Variation Allows Control over Cross-Linking and Viscoelasticity Lorenzo Aulisa, He Dong, and Jeffrey D. Hartgerink* Departments of Chemistry and Bioengineering, Rice University, 6100 Main Street, Houston, Texas 77005 Received June 8, 2009; Revised Manuscript Received July 23, 2009
An important goal in supramolecular chemistry is to achieve controlled self-assembly of molecules into welldefined nanostructures and the subsequent control over macroscopic properties resulting from the formation of a nanostructured material. Particularly important to our lab is control over viscoelasticity and bioactivity. Recently we described a multidomain peptide motif that can self-assemble into nanofibers 2 × 6 × 120 nm. In this work we describe how sequence variations in this general motif can be used to create nanofibrous gels that have storage moduli, which range over 2 orders of magnitude and can undergo shear thinning and shear recovery while at the modest concentration of 1% by weight. Gel formation is controlled by addition of oppositely charged multivalent ions such as phosphate and magnesium and can be carried out at physiological pH. We also demonstrate how maximum strength can be obtained via covalent capture of the nanofibers through disulfide bond formation. Together these hydrogel properties are ideally suited as injectable materials for drug and cell delivery.
Introduction Hydrogels are materials extensively used in tissue engineering and drug delivery.1,2 They can be prepared from a variety of materials including both natural and synthetic high molecular weight polymers.3-7 In contrast, hydrogels may also be formed from relatively small molecules that undergo self-assembly to form a physically cross-linked network of self-assembled nanofibers.8 This second approach has advantages because of the small size of the building blocks and reversibility or responsiveness of the gel. In particular, the use of small peptides is favorable due to their automated synthesis, broad range of chemical functionality derived from natural or unnatural amino acids, and biocompatibility. The ability to control precisely the functionalization of the monomer building block is critical for control over all other levels of assembly. Chemical functionality and patterning coded in a peptide’s amino acid sequence lead to control over conformation. Chemistry and conformation together control self-assembly and nanostructure. All of these together determine a material’s bulk mechanical properties. Therefore, careful design of a peptide sequence should, through this bottom-up approach, allow one to control the macroscopic properties of a material. We have designed short peptides that self-assemble into β-sheet nanofibers, which at low concentrations produce hydrogels with potential applications in areas such as drug and cell delivery and cell scaffolds for tissue regeneration.9-11 β-Sheet fibril-forming peptides have been utilized by a number of other researchers. For example, Lynn has focused on a 26amino-acid peptide derived from Alzheimer’s amyloid protein to form fibrils and nanotubes.12 Starting from designed sequences, Aggeli,13 Boden,14 Collier,15 Messersmith,16 Stupp,17,18 Zhang,19 and our group20-22 have prepared a variety of different fibril morphologies ranging from helical ribbons to twisted and cylindrical fibrils. Of particular note, Pochan and Schneider have utilized β-hairpins to create a series of peptides able to form fibrils, stable in different ranges of pH and temperature.23 The approach presented here uses multidomain peptides (MDPs) that * To whom correspondence should be addressed. E-mail:
[email protected].
are designed to have forces favoring the assembly (hydrophobic packing and hydrogen bonding), balanced by forces working against assembly (electrostatic repulsion). With proper balancing of these forces it is possible to control the assembly of the peptides into nanofibers (Figure 1).9,10 Peptide Design. MDPs are organized with an ABA structure in which the central B domain contains alternating hydrophilic and hydrophobic amino acids. When the peptide is an extended β-sheet conformation, this creates a facial amphiphile as shown in Figure 1.9,24,25 The hydrophobic amino acids are the primary driving force for self-assembly in an aqueous solution, as packing with another MDP to form a “hydrophobic sandwich” hiding the hydrophobic residues from water is entropically favorable. The peripheral A block is composed of charged amino acids that, due to electrostatic repulsion at neutral pH, work against the central block’s propensity to associate. Previously we demonstrated that peptide 1, under low salt conditions, makes an ideal balance between these forces and forms nanofibers approximately 120 nm long, 6 nm wide, and 2 nm high.9 In this study, we examine the use of the negatively charged glutamic acid (E) in the peripheral A domain (peptides 2, 4, and 5) and the use of serine (S) as the hydrophilic amino acid (peptides 3-5) in the central B domain. Our aim was to find peptides that could assemble under physiological conditions into nanofibers and subsequently form a viscoelastic hydrogel. It was believed that the use of a negatively charged amino acid in domain A would result in nanofibers that could be triggered to undergo a second step of self-assembly into a cross-linked hydrogel in the presence of multivalent cations such as Mg2+,26 commonly present in cell culture media. The selection of serine as an alternative neutral, yet hydrophilic, amino acid was expected to change the quality of hydrogen bonding between nanofibers, which should play a large role in determining the mechanical properties of a gel. Finally, peptide 5 was designed based on peptide 4 in which three of the serine residues were replaced with cysteine. Because cysteine can form disulfide cross-links under mild oxidative conditions, we expected this covalent linkage to result in stronger, more stable nanofibrous peptide hydrogels. Together these peptides illustrate three
10.1021/bm900634x CCC: $40.75 2009 American Chemical Society Published on Web 08/25/2009
Self-Assembly of Multidomain Peptides
Biomacromolecules, Vol. 10, No. 9, 2009
2695
Figure 1. Dimerization driven by hydrophobic packing results in the parallel arrangement of two peptides forming a “hydrophobic sandwich”. The backbone amides are thus optimally oriented for fiber formation in which the hydrogen bonding axis is parallel to the growing fiber.
different methods of cross-linking assembled nanofibers to form a hydrogel with a broad range of mechanical properties. Peptides 1 and 3 can be cross-linked with negatively charged ions such as PO4-3, peptides 2 and 4 can be cross-linked with positively charged ions such as Mg2+, and peptide 5 can be cross-linked under mild oxidative conditions.
Experimental Section Peptide Synthesis and Characterization. The peptides were prepared on an Advanced ChemTech 396 multipeptide automated synthesizer. The scale of the synthesis was 0.15 mmol. A Rink Amide MBHA low loading resin was used to generate a terminal amide. Fmoc (fluorenylmethoxycarbonyl) protected amino acids were purchased from NovaBiochem. Serine and glutamic acid side chains were t-butyl protected, lysine side chain was Boc (t-butoxycarbonyl) protected, and cysteine was trityl protected. All the amino acids and other reagents were dissolved in a mixture of 50% DMF (dimethylformamide) and 50% DMSO (dimethylsulfoxide). Amino acid coupling cycles were 60 min in length, with the following proportions of reagents: 4 equiv of amino acid, 4 equiv of HATU (O-(7-azabenzotriazole-1-yl)-N,N,N,N′tetramethyluronium hexafluoroposphate), and 6 equiv of DiEA (diisopropylethylamine). Fmoc protecting group was removed with two 7 min treatments of 20% (by volume) piperidine and 2% (by volume) DBU (1,8-diazabicyclo[5.4.0]-undec-7-ene) in DMF and DMSO. After the peptide sequence was completed, the N-terminus was acetylated by acetic anhydride and DiEA in DMF for 2 h. A ninhydrin test was used to monitor the acetylation step.27 After the reaction was finished, the resin was washed several times with DCM (dichloromethane) before running the cleavage step. Cleavage of the peptides were accomplished by shaking the resin with 20 mL of TFA (trifluoroacetic acid)/ triisopropylsilane/anisole/ethane dithiol /H20 (18:0.5:0.5:0.5:0.5 by volume) for 3 h at room temperature. The solution was collected by filtration followed by rinsing the resin twice with 20 mL of neat TFA. All washings were combined and rotoevaporated to achieve a thick solution. The peptide was triturated by addition of cold diethyl ether. The precipitate was collected by centrifugation. The crude peptides were purified through a precipitation method in which each peptide was washed several times with cold ether, and after that they were dried. Peptides 1 and 3 were subsequently washed with 0.1 M NaOH solution. The supernatant was discarded and the precipitated material was collected, dried, and redissolved in deionized water, with the addition of 0.1 M HCl to adjust the pH to slightly acidic (around pH 5). The solution was frozen and lyophilized. Peptides 2, 4, and 5 were washed with 0.1 M HCl. The supernatant was discarded and the precipitated material was collected, dried, and redissolved in deionized water using 0.1 M NaOH to adjust the pH to approximately pH 8. The solution was frozen and lyophilized. MALDI-TOF (matrix assisted laser desorption ionization time-of-flight) analysis was used to characterize the mass of the final products. See Supporting Information for spectra. Circular Dichroism (CD). A JASCO J-810 spectropolarimeter was used to perform CD measurements. The peptides were dissolved in a 10 mM Tris buffer, pH 7. The spectra were recorded from 250 to 180 nm with a scan speed of 20 nm/min, the signal was averaged over 10 scans. mDeg of rotation was converted to mean residue ellipticity using the formula θ ) (θ × Mw)/(c × 10 × l × n), where c is the concentration of the peptide expressed in grams/liter, l is the thickness
of the cell used in cm, n is the number of amino acids in the sequence, and Mw is the molecular weight of the peptide. Temperature was controlled using a peltier heater. Samples were prepared at a concentration of 0.25% by weight. Spectra for peptides 4 and 5 were significantly influenced by linear dichroism from fiber formation at this concentration. See Supporting Information for spectra. Fourier Transform Infrared Spectroscopy (FTIR). A Jasco FT/ IR 660 plus spectrometer was used for the measurements. The peptides were all analyzed at 1% w/w concentration. Aliquots of the peptide were deposited on a “Golden Gate” diamond ATR (attenuated total reflectance) and dried under nitrogen to generate a thin layer of peptide. Collected spectra were linear baseline corrected and normalized for ease of comparison. Vitreous Ice Cryo-Transmission Electron Microscopy. A total of 2 µL of peptide was placed on a holey carbon grid (Quantifoil R 1.2/1.3), blotted for 2 s, and then plunged in liquid ethane, cooled by liquid nitrogen, using a controlled environment vitrification system (Vitrobot, FEI). The samples obtained were transferred in a cryoholder (Gatan 626DH) without any further modification and subsequently analyzed on a JEM 2010 microscope, equipped with a CCD camera (Gatan 2 × 2K). During the sample analysis, the temperature was kept at -176 °C. Rheometry. A TA AR-G2 rheometer was used to perform the measurements on all the peptides. Dynamic stress, frequency, and time sweep analysis was performed using an 8 mm diameter parallel plate geometry at 25 °C, with a fixed gap of 350 µm. A dynamic stress sweep test was used to determine the linear viscoelastic region. Gel properties were monitored through frequency sweep measurements with a fixed strain amplitude (1%) and a variable frequency (0.1-10 Hz) to measure storage and loss modulus. Time sweep tests were performed to measure the gel recovery properties of the materials. After an initial step, in which the sample was oscillated with low strain amplitude to reach equilibrium, a strong deformation (100% strain for 60 s) was applied. After 60 s, the deformation was stopped and the recovery analyzed through time sweep measurements performed at constant strain and frequency. To ensure that shear, rather than slipping, was occurring, the phase angle was monitored and found to remain within acceptable bounds (δ e 90°).
Results and Discussion Peptides were synthesized utilizing a modified Fmoc solid phase strategy (Table 1). It was found that our typical synthetic protocol, which uses a coupling cycle with 4 equiv of HBTU (O-benzotriazole-N,N,N′,N′-tetramethyluronium hexafluorophosphate), 4 equiv HOBt (hydroxybenzotriazole), 6 equiv of DiEA reacting with the peptide N-termini for 30 min, and deprotection cycle using 20% piperidine in DMF for 2 × 7 min resulted in poor yield and many deletions. In retrospect, this is not surprising considering the strong propensity of peptides with this design to form β-sheet structure, which is considered to be one of the major culprits for failure in solid phase peptide synthesis. To alleviate these problems, we made the following modifications to standard solid phase protocol: low loading (
