Self-Assembled pH-Responsive Hydrogels Composed of the

May 23, 2008 - Department of Materials Science and Engineering, Nagoya Institute of Technology, Gokiso-cho, Showa-ku, Nagoya 466-8555, Japan; College ...
3 downloads 10 Views 7MB Size
Biomacromolecules 2008, 9, 1511–1518

1511

Self-Assembled pH-Responsive Hydrogels Composed of the RATEA16 Peptide Ying Zhao,†,‡ Hidenori Yokoi,§ Masayoshi Tanaka,† Takatoshi Kinoshita,*,† and Tianwei Tan*,‡ Department of Materials Science and Engineering, Nagoya Institute of Technology, Gokiso-cho, Showa-ku, Nagoya 466-8555, Japan; College of Life Science and Technology, Beijing University of Chemical Technology, Beijing 100029, China; and Menicon Co., Ltd. 5-1-10, Takamorodai, Kasugai, Aichi 487-0032, Japan Received October 16, 2007; Revised Manuscript Received March 12, 2008

Peptide RATEA16 spontaneously self-assembled into higher-order nanofiber hydrogels with extremely high water content (>99.5% (wt/vol)) under physiological condition. The hydrogels could undergo pH-reversible transitions from viscous solution to elastic hydrogel and to precipitate. The supramolecular self-assembly and the three phase transitions are driven by hydrophobic interactions, intermolecular hydrogen bonds, and a combination of attractive or repulsive electrostatic interactions. These hydrogels are rich in β-sheet nanofibers, as demonstrated by CD and FTIR data. Rheological measurements reveal that the viscoelasticity of the material can be tuned by environmental pH and peptide concentration. The storage modulus of the hydrogels increases with increasing peptide concentration, and the self-assembled hydrogels are able to recover from mechanical breakdowns. AFM images show that the elasticity is attributed to a network nanostructure consisting of fibrous self-assemblies. The hydrogels are promising for a variety of possible biomedical applications, including drug delivery.

Introduction Hydrogels are a class of viscoelastic materials that are promising for a variety of sophisticated applications, including drug release and tissue engineering.1–3 Their porous microstructures and three-dimensional networks provide good permeability and mechanical support, respectively. Hydrogels can be obtained by the covalent crosslinking of high molecular weight synthetic polymers and small molecules.4–7 However, the crosslinking approach usually requires additional chemical synthesis that can result in the loss of biocompatibility and biodegradability of the hydrogel.8 An alternative approach is to take advantage of the spontaneous self-assembly of small molecules into three-dimensional network structures. Importantly, such networks can result from noncovalent interactions, do not involve any chemical crosslinking, and can be achieved under physiological condition.9 Proteins, polypeptides, and more recently, oligopeptides2,4,9–18 have been shown to form functional hydrogels by self-assembly. Oligopeptides are particularly attractive compared to proteins or polypeptides considering that they can be straightforwardly prepared via solid-phase synthesis with precise control over sequence, molecular dimension, and derivatization.4 Recently, noncovalent hydrogels based on self-assembly have been investigated more and more for the production of responsive, reversible, and injectable delivery systems.9,19–31 In these systems, aggregation of the small molecules and a gelsol transition behavior could be controlled through delicately balancing some noncovalent interactions, such as hydrophobic interactions and hydrogen bonds. Likewise, we are interested * To whom correspondence should be addressed. Tel. and Fax: 81-52735-5267. (T.K.); 86-10-64416691 (T.T.). E-mail: kinoshita.takatoshi@ nitech.ac.jp. (T.K.); [email protected] (T.T.). † Nagoya Institute of Technology. ‡ Beijing University of Chemical Technology. § Menicon Co., Ltd. 5-1-10.

in employing oligopeptides for the preparation of responsive hydrogels. It is well-known that one of the properties of peptides is their ability to become electrostatically charged moieties in response to changes in pH values, which is important for peptide hydrogels to form intelligent and pH-responsive materials.4,12,32–35 Peptides that self-assembled into hydrogels exhibited the common feature that they are composed of both hydrophilic and hydrophobic units. For instance, some protein-based peptides presenting coiled-coil motif have been used to prepare stimuli-responsive hydrogels.36,37 Peptides with alternating hydrophilic and hydrophobic residues that are amphipathic with respect to a β-sheet or β-strand structure have also been used as building blocks of hydrogels. For example, the peptide MAX19,12 could form smart pH-responsive hydrogels under basic condition, through pH-dependent intramolecular folding of a β-hairpin structure.12 This pH-responsive behavior was attributed to the protonation/deprotonation of the charged lysine, resulting in switching between the hydrogel and the solution phase. Similarly, we previously reported the efficient selfassembly of an oligopeptide, RADA16 ([CH3CONH]RADARADARADARADA-[CONH2]), into a hydrogel.38 This peptide formed nanofibers that spontaneously self-assembled in aqueous solution into higher-order nanofiber hydrogels with extremely high water content. The RADA16 hydrogel is applied as a scaffold for 3D cell culture and is commercially available as PuraMatrix (3DM Inc.). However, to prepare a RADA16 hydrogel at physiological condition, a complex procedure is required to gradually change the pH of the RADA16 solution from acidic to neutral by a solvent substitution.39 Furthermore, once the hydrogel is stirred under neutral condition, electrostatic interactions between protonated Arg+ and deprotonated Gluoperate strongly to give precipitate. From this perspective, we examined here a new selfassembling peptide, RATEA16 ([CH3CONH]-RATARAEARATARAEA-[CONH2]), which forms a stable transparent

10.1021/bm701143g CCC: $40.75  2008 American Chemical Society Published on Web 05/23/2008

1512

Biomacromolecules, Vol. 9, No. 6, 2008

hydrogel without complex procedure at neutral pH. RATEA16 is composed of four cationic and two anionic residues, thus, it would form a different self-organized structure at neutral pH. Especially, the polar but uncharged threonine was introduced, resulting in hydrogels that formed under physiological condition. Furthermore, it is expected that the peptide RATEA16, which contains two kinds of charged amino acids (glutamic acid and arginine), will exhibit more complicated protonation/deprotonation processes compared to the MAX1 peptide. This might be advantageous for applications such as drug delivery. In fact, we show that hydrogels of the peptide RATEA16 can undergo pH-reversible transitions from viscous solution to elastic hydrogel and to precipitate. Of course, hydrophobic interaction was the main driving force for self-assembly of the peptide RATEA16. Electrostatic interaction between the oppositely charged side chains of arginine and glutamic acid would be decisive for the self-aggregation, although it is not a necessity for some other short and intermediate-sized peptides.40 Therefore, another focus of this work is to demonstrate how noncovalent interactions lead to the spontaneous self-assembly of peptide nanofibers and to pH-reversible transitions.

Experimental Section Preparation of Peptide Hydrogel. The amphiphilic peptide, RATEA16 [CH3CONH]-RATARAEARATARAEA-[CONH2], was synthesized by a solid-phase method using standard Fmoc strategy on CLEAR-Amid resin (see Supporting Information). Based on the result of MALDI-TOF-MS, the peptide used for the hydrogel formation was crude. Dissolving peptide RATEA16 with a final 0.5% (wt/vol; 5 mg/ mL) of concentration in Milli-Q water and Tris · HCl buffer (pH 7.5) immediately resulted in the formation of hydrogels under neutral pH condition. This transparent hydrogel was considered to be selfsupporting only when it was possible to invert the container without any collapsing.1 AFM Observation. A total of 1 µL of hydrogel was deposited on a freshly cleaved mica surface. After about 15 s, the hydrogel on the mica was washed with 50 µL of Milli-Q water and then the mica was air-dried. To study the influence of the drying process on the network structures, some hydrogels were placed on mica and then vacuumdried. The images were obtained by scanning the mica surface in air by AFM (Nanoscope IIIa, Digital Instruments, Santa Barbara, CA) using silicon cantilever (NCH-10V, Veeco Instruments) operated in tapping mode. CD Measurement. Circular dichroism (CD) spectra were obtained using a JASCO J-820K spectropolarimeter, which was flushed with nitrogen during operation. Wavelength scans were recorded at 0.1 nm intervals from 260 to 190 nm, using a 0.1 cm path length quartz cuvette. The spectra obtained were averaged from 16 consecutive scans and subtracted from the background. Ellipticity measurement was shown to mean residue ellipticity ([θ], in deg · cm2 · dmol-1).41 FTIR Spectroscopy. Fourier transform infrared (FTIR) spectra were collected in absorbance mode on a Perkin-Elmer Spectrum 2000 FTIR spectrometer. Aliquots of hydrogel samples prepared in D2O were placed on CaF2 plates and dried in vacuum. For each sample, 16 scans were collected and averaged to obtain a good signal-to-noise ratio. Rheological Characterization. Viscoelastic property was determined with a rheometer (AR1000, TA Instruments). A 40-mm-diameter stainless-steel parallel plate and a 40-mm-diameter/1-degree cone-plate were used. Rheological studies of samples were conducted by loading 310 µL of the freshly prepared samples on the sample platform of the instrument, followed immediately by a series of rheometrical tests at 25 °C. For the determination of viscoelastic properties and to avoid the influence of mixing and breakdown during loading for hydrogel and precipitate samples, they were immediately subjected to a 3 h time sweep test. After the data reached a well-defined equilibrium, a series

Zhao et al.

Figure 1. Peptide RATEA16. (a) Molecular model of RATEA16. (b) Photograph of RATEA16 hydrogel (0.5% (wt/vol) of peptide concentration, pH 7.4).

of frequency sweep tests and strain amplitude sweep tests were sequentially carried out. Dynamic time sweep experiments were performed with a frequency and strain amplitude of 6 rad/s and 0.1, respectively. The oscillatory frequency sweep tests (frequency ω ) 0.1-100 rad/s) were studied to measure G′ and G′′, the linear viscoelastic storage and loss modulus, respectively. The oscillatory strain amplitude sweep tests (strain amplitude γ0 ) 0.001-10) at a fixed frequency (6 rad/s) were investigated. To characterize the responsiveness of the hydrogel from repeated strain-induced breakdowns, the hydrogel was pretested in time sweep for 3 h (under 0.1 amplitude strain and 6 rad/s frequency) and consequently subjected to a routine of recovery for five cycles. Each cycle consists of a 2 min break period with a continuous 6 amplitude sine-wave strain and a 30 min recovery period with a constant 0.1 strain at 6 rad/s frequency. The magnitude of the strain force (6 amplitude) and the duration of the break and recovery periods (2 and 30 min, respectively) were based on results from the strain and time sweep tests. pH-Response. Tris · HCl buffer with pH 7.5 was used to study pHresponse of peptide RATEA16 hydrogels. The desired pHs (3.5, 7.4, and 12.5) were adjusted by NaOH and HCl solutions. To avoid the influence of ionic strength on self-assembly, sodium chloride was used to adjust ionic strength of peptide solutions to 0.2 M; the pH values were precisely checked by a pH meter (B-212, accuracy ( 0.01, HORIBA Ltd., Kyoto, Japan). Drug Release Study. Vitamin B1 (337 Da), supplied by SigmaAldrich (Missouri) was used as guest molecule. The peptide RATEA16 self-assembled into a hydrogel in the aqueous solution, and the final concentrations of the RATEA16 and vitamin B1 in the mixture were 0.5% (wt/vol) and 0.5 mM, respectively. In a controlled release caused by pH-response, the hydrogel (250 µL) containing vitamin B1 was exposed to HCl vapor for 5 h, with a 15 min interval and then was covered with 2.5 mL of Milli-Q water; finally, the water was collected after 15 min. Acidification was carried out by placing a beaker containing 35% HCl (wt/vol) and a vial with the peptide hydrogel in a desiccator. In a control experiment, the hydrogel (250 µL) was covered with 2.5 mL of Milli-Q water, and the water was collected every 15 min. The concentration of the drug released from the hydrogel was determined using a calibration curve at the wavelength where vitamin B1 showed its maximum absorbance (246 nm), quantitatively determined by UV/visible spectrophotometer (JASCO V-550). The averages from three independent determinations were plotted. The release ratio was calculated by eq 1

release ratio(% ) )

Mt × 100% M∞

(1)

where M∞ and Mt represent the initial amount of drug loaded and the cumulative amount of drug released at time t.42

Results and Discussion Nanofiber Hydrogel Formation. The molecular weight of peptide RATEA16 (Figure 1a) was determined by MALDI-

Hydrogels Composed of the RATEA16 Peptide

Biomacromolecules, Vol. 9, No. 6, 2008

1513

Figure 2. Representative AFM images of RATEA16 nanofibers. (a) 0.3% (wt/vol), pH 7.4; (b) 0.5% (wt/vol), pH 7.4; (c) 1.0% (wt/vol), pH 7.4; (d) by vacuum-drying (0.5% (wt/vol), pH 7.4); (e) a bilayer structure (0.5% (wt/vol), pH 7.4); and (f) AFM cross-section profile of (e).

TOF-MS (see Supporting Information). The experimental value (1713.9) agreed very well with the calculated value (1712.848). The process of hydrogel formation was extremely simple. A transparent elastic hydrogel (Figure 1b) consisting of interconnecting nanofibers (Figure 2) immediately appeared when the pH of the peptide solution was adjusted to neutrality using Tris · HCl buffer (pH 7.5). The gelation behavior was found to depend heavily on pH, with transitions from hydrogel to solution or to precipitate observed under acidic or basic media, respectively. The reversibility of the three phase transitions (solution-hydrogel-precipitate) was studied through changing the pH of media (see Supporting Information), demonstrating that the self-assembly process was reversible. In addition, it is found that a tremendous amount of water can be “immobilized” by the hydrogel structure. Although there was only less than 1% (wt/vol) peptide component in the hydrogel, the three-dimensional spacious structure was kept exceptionally well and was stable over a broad temperature range (see Supporting Information) or in the presence of solute. Moreover, the hydrogel could be fabricated into different geometric morphologies according to the vessel in which it was assembled. In fact, it could be specifically injected through a needle after self-assembly in a syringe, making it become an appealing biomaterial for biomedical

applications, which involve the injection of a biomaterial into injured tissue within a living organism.1 Nanostructure of the Hydrogels. We used atomic force microscopy (AFM) to obtain better views for the nanofibers, which constitute the scaffold structures. The images of the selfassembling hydrogel with various concentrations of peptide were shown in Figure 2a-c. Although the nanostructures in these three images were different from each other, all of them consisted of crosslinked network structures, suggesting that the obtained hydrogel was constructed by interconnecting nanofibers. The crosslinked network structure was found to be denser as the concentration of the sample increases. On the other hand, it should be pointed out that the image of the vacuum-dried sample (Figure 2d) suggested that the complicated network construction of the hydrogel partially collapsed during the drying process. This phenomenon was attributed to the fact that the stability of the fibrous scaffold can be disrupted by the volatilization of water.1 In addition, Figure 2b clearly showed fibrous structures with widths of 6∼8 nm, which was in good agreement with the molecular dimensions of the peptide (Figure 1a). It was interesting to observe that the nanofibers possessed a bilayer structure (Figure 2e and Supporting Information); on the basis of AFM cross-section profile (Figure 2f), we found

1514

Biomacromolecules, Vol. 9, No. 6, 2008

Zhao et al.

Figure 3. CD examination. (a) The secondary structure of peptide RATEA16 under various concentrations and time intervals (environmental pH ∼ 7.4): 100 µM, 5 min ((); 100 µM, 126 min ()); 200 µM, 5 min (2); 200 µM, 126 min (4). (b) Reversibility of the self-assembly process as a function of environmental pH: 7.4 (()f3.5 (9)f12.5 (2)f7.4 ()).

that the difference of these two layers heights is about 1.0∼1.3 nm, similar to a single layer thickness of the RATEA16 molecule. Secondary Structure. It is well-known that the existence of an ordered secondary structure of building block, such as R-helix or β-sheet, could be a dominating factor for gelation. The secondary structure of the peptide that constituted the hydrogel was characterized using circular dichroism (CD) and Fourier transform infrared (FTIR) spectroscopies. Of course, CD spectroscopy gives information about the secondary structure of the proteins. It can be considered that the forming of the β-sheet structure is itself kind of a self-assembly, which provides the peptides with the possibility to further self-assemble into a gel scaffold. Therefore, the CD results reported below will be discussed in terms of intramolecular self-assembly of β-sheet structures and the formation of intermolecular bilayers. In CD spectra, β-sheet structure is characterized by a negative band near 215 nm and a strong positive band between 195 and 200 nm.4 Based on our previous report,38 we believe that the typical β-sheet spectra of peptide RATEA16 should be observed under neutral pH condition, where one face is lined with hydrophobic Ala residues and the other face is an array of hydrophilic Arg, Thr, and Glu residues. However, variations in the peptide sequence can result in significant changes in the properties of obtained self-assembled materials;2 therefore, CD measurement is absolutely indispensable. The peptide concentration used in the CD measurements was low and, although there was not enough peptide to cause gelation, the self-assembly of peptide nanofibers still happened, making CD studies possible. The CD results were shown in Figure 3a and supported our expectation that the typical β-sheet spectra were observed at all concentration systems. Moreover, the variety of amino acid sequences did not alter the secondary structure, compared to peptide RADA16. Time-dependent CD measurements revealed a fast self-assembling process, and the content of the β-sheet remained nearly identical, between 5 and 126 min time intervals. Furthermore, the observed value at θ216 depended on concentration, indicating that more peptide facilitated the β-sheet formation. The reversibility of the obtained supramolecular structure was also investigated by measuring CD as a function of pH (Figure 3b). The β-sheet structure was quite stable at pH 7.4 in the course of pH changes. It should be pointed out that some β-sheets were still present even at pH 3.5 and pH 12.5. However, the strong electrostatic repulsions between Arg+ residues or Glu- residues prevented the nanofibers from

Figure 4. FTIR spectrum of peptide RATEA16 hydrogel with 0.5% (wt/vol) of peptide concentration at environmental pH 7.4.

becoming longer and crosslinking. Consequently, hydrogels did not appear at pH 3.5 and pH 12.5. According to CD measurements, it was apparent that peptide RATEA16 nanofibers were rich in β-sheet structure under neutral condition. To further estimate the presence of β-sheet nanofibers within the hydrogel network, the hydrogel prepared with D2O was examined by FTIR. The FTIR spectrum of the gel showed two obvious peaks at 1621 and 1684 cm-1, which are indicative of intermolecular antiparallel β-sheet structure (Figure 4).12,14,43 Rheological Properties. CD experiments, which characterized the intramolecular self-assembly of β-sheet structure, and the formation of intermolecular bilayers, and rheology measurements, which reflected the self-assembly of peptide nanofibers into a hydrogel scaffold, combine to illuminate how material properties can be attributed to molecular self-assembly behavior.12 The viscoelasticity of the RATEA16 self-assembled hydrogels was characterized by rheological measurement. Frequency sweep measurements were used for studying the effects of the environmental pH and concentration on the hydrogel strength. Figure 5a showed the mechanical properties of the peptide hydrogels that were prepared from various pH conditions, with a concentration of 0.5% (wt/vol). As seen, the storage modulus (G′ elastic response) values over the entire frequency range exceeded those of the loss modulus (G′′ viscous response) by 1 order of magnitude at pH 7.4, indicating the underlying network structure of a typical elastic hydrogel.1,3 The rheological data showed that G′ and G′′ values decreased at pH 12.5, suggesting that the crosslinked structure was weakened by electrostatic repulsion between Glu- residues. The

Hydrogels Composed of the RATEA16 Peptide

Biomacromolecules, Vol. 9, No. 6, 2008

1515

Figure 5. Storage modulus (G′, solid) and loss modulus (G′′, open) as (a) frequency sweeps for 0.5% (wt/vol) samples with various pH: (2) pH 3.5, (() pH 7.4, and (9) pH 12.5; (b) frequency sweeps for pH 7.4 samples with various concentration: (b) 0.2% (wt/vol), (2) 0.3% (wt/vol), (9) 0.5% (wt/vol), and (() 1.0% (wt/vol), compared with the storage moduli of representative self-assembling hydrogels from the literature: β-sheet by (+) Caplan et al.,12,45 (s) fibrillar gel by Aggeli et al.,12,15 (-) coiled-coil proteinaceous gel by Petka et al.,12,46 (×) peptide RADA16 hydrogel38 and (*) wormlike micelle gel by Won et al.;12,47 (c) time sweeps for pH 7.4 hydrogels with various concentration: ( · · · · ) 0.3% (wt/vol), (- - -) 0.5% (wt/vol), and (s) 1.0% (wt/vol); (d) strain sweeps for pH 7.4 hydrogels with various concentration: (2) 0.3% (wt/vol) sample collapsed at 1.6 strain amplitude, (9) 0.5% (wt/vol) sample collapsed at 1.0 strain amplitude, and (() 1.0% (wt/vol) sample collapsed at 0.5 strain amplitude; (e) five cycles of time sweeps for the hydrogel recovered from strain-induced breakdowns (0.5% (wt/vol) of peptide concentration, pH 7.4).

samples were actually not hydrogels at two extreme pHs, 3.5 and 12.5, but rather a solution and a precipitate, respectively (Figure 5a). The apparent rises of G′ and G′′ at high frequencies for the pH 3.5 and 12.5 samples are experimental artifacts of the cone-plate geometry.44 In addition, as expected, decreasing peptide concentration induced a decrease in both G′ and G′′ values (Figure 5b). In the range of 0.3-1.0% (wt/vol), G′ values were also at least 1 order of magnitude higher than G′′ values, and both G′ and G′′ were essentially independent of the frequency in the entire range, revealing no crossover point between G′ and G′′. The phenomena suggested that the preponderant relaxations of the viscoelastic network were at lower frequencies; it meant that the relaxation time, τ, of the network was long, which matched the presence of a crosslinked scaffold.9,14 Viscoelasticity measurements at lower peptide concentrations demonstrated that the critical concentration separating a liquid-like (G′′ > G′) and a gel-like (G′ > G′′) response12 was about 0.3% (wt/vol), and the sample was a very viscous liquid up to this concentration (Figure 5b). Several

examples of hydrogel moduli from the literature are also shown in Figure 5b for comparison. The modulus of peptide RATEA16 hydrogel was moderate, compared with these physically similar systems measured in the concentration range of 0.5-5% (wt/ vol). The time sweep profiles of the hydrogels with different concentrations were presented in Figure 5c, indicating the gelation process was smoother for the lower concentration sample. In all samples, gelation was kinetically characterized by two regions: an elastic burst region followed by a slower elastic growth phase.32 Analogous to results of β-sheet formation observed by CD, rheology studies revealed that the gelation process significantly progressed within the first 5 min, continuing to mature after 2 h, and faster self-assembly kinetics due to higher peptide concentration induced more elastic hydrogels. Therefore, the increase in elasticity resulted from more highly crosslinked networks (more crosslinked points).9

1516

Biomacromolecules, Vol. 9, No. 6, 2008

Figure 6. Self-assembly of peptide RATEA16 at different environmental pH. (red circles) Glu-; (blue circles) Arg+, the size represents the extent of dissociation; the big circles mean protonated Arg at pH 3.5 and 7.4, and the small circles mean almost deprotonated Arg at pH 12.5. In the bottom right scheme, the crossed circles represent positive charge.

Two characteristics of a self-assembled hydrogel are the breakdown of networks under a certain mechanical force and subsequent recovery after the force has ended.12 Strain amplitude sweeps were investigated at a fixed frequency of 6 rad/s (Figure 5d). It was shown that G′ decreased rapidly above a concentration-dependent critical strain amplitude γ0c, indicating a breakdown of any physical crosslinked points by the application of strain.12,44 Moreover, they also demonstrated that the γ0c of the low concentration hydrogel was greater than that of the high concentration one, attributed to a much greater resistance to deformation for the former. The hydrogel with increasing concentration, thus, became more brittle, proved by the collapse of network structure at smaller strains. Significantly, selfassembled hydrogels are able to reform after mechanical destruction. To show this, five cycles of hydrogel recovery were performed immediately after the strain-induced breakdowns (Figure 5e), revealing that 90% of the equilibrium G′ modulus was recovered after 30 min. Therefore, this fibrous network structure, the rapid gelation process, and the recoverable essence would be promising characteristics for in vivo applications such as controlled drug release and soft tissue engineering. pH-Responsive Behavior. The formation of hydrogels based on low molecular weight molecules by self-assembly process is quite different from that of polymeric networks. The selfassembly of peptide RATEA16 was driven by hydrophobic interactions, inter- and intramolecular electrostatic interactions, and intermolecular hydrogen bonds (see Supporting Information). To understand the dynamic process of pH-response, Figure 6 was used to interpret the differences in the kinetics of selfassembly at various pH values. Aqueous dispersions of these amphiphilic bilayer structures provided the possibility of nanoassemblies. However, under the acidic environment, Glu was protonated and usually hydrogel could not be formed because of the strong inter- and intramolecular electrostatic repulsions between Arg+ residues, which prevented the nanofibers from becoming longer (Figure 7a) and crosslinking. On the other hand, when the net charge of peptide molecule neared zero (basic condition), the extent of β-sheet was very weak (Figure 3b), and the facial charges were shielded by the formation of inter- and intramolecular ion pairs. The strong electrostatic attractions made hydrophilic amino acids

Zhao et al.

aggregate together, resulting in the exposure of hydrophobic Ala arrays to water. Therefore, precipitate appeared (Figure 7b), owing to loss of solubility. It was expected that samples should not precipitate but remain in solution at pH 12.5, considering that the pK of Arg is 12.1. However, the electrostatic attractive forces between neighboring Glu- and Arg+ residues in the peptide shifted the effective pK values of the Glu and Arg units to lower and higher values, respectively, than those of the free amino acids.33 Consequently, a more basic condition could result in solution again. Under neutral pH condition, the favorable electrostatic attraction between Glu- and Arg+ residues largely contributed to the stability of the scaffold. Although there are two positive net charges in each peptide molecule, the intermolecular hydrogen bonds and the antiparallel β-sheet gave sufficient association energy and structural support to resist the electrostatic repulsions and to form longer nanofibers. The reduction of the charge repulsions between the fibrous assemblies must facilitate the formation of a scaffold structure when compared with acidic media. The nanofibers were crosslinked to each other by physical junction points, possibly due to simple entanglements of fibers and facial electrostatic interactions. Moreover, two positive net charges per one peptide sequence, given by the protonated Arg, were required for ordered self-assembly without precipitate. They also provided enough osmotic pressure (ionic pressure) between the network and the external water to promote the formation of the hydrogel. It may be reasonable to propose that, in addition to increasing osmotic pressure, the charged nanoassemblies could also provide an interfibrous repulsion to resist precipitation and to yield a swollen hydrogel. In contrast, the decrease of such an electrostatic repulsion and an osmotic pressure caused by a higher pH medium led to aggregation upon the charge neutralization.48 In addition, according to our previous report,38 this work emphasized the two units of net charge per peptide molecule were necessary in the case of our self-attractive peptide. The above explanation was correlated well with the rheological study (Figure 5a). The three phase transitions (solution-hydrogel-precipitate) could be achieved by reversibly changing the pH, indicating that the pHresponse did not result from the chemical decomposition of the component but from a reversible and physicochemical property change in the hydrogel.48 The reformation of the hydrogel was strongly attributed to the reformation of long and crosslinked fibrous structures (see Supporting Information). In other words, peptide RATEA16 was prone to selfassembly, allowing environmentally sensitive hydrogels to be formed. In addition, because the gelation process was a reversible process governed by pH, consequent hydrogel collapse resulted from simple acidification and basification. It is significant to know this pH-responsive behavior for further biomedical application. Controlled Drug Release. The hydrogels stable at physiological pH were subsequently tested for their ability to be applied in the biomedical field as drug carriers. This capability was attributed to the unique structure comprising a dilute twocomponent system in which the minor (solid) and major (fluid) components form a separate, three-dimensional, continuous phase.49 As observed from our AFM images, the fibrous network contains pores in which drug molecules can be entrapped. To explore the potential of this pH-responsive network structure for a diagnostic or therapeutic material, vitamin B1 was used in the process of hydrogel preparation. The environmental pH of the hydrogel was gradually acidified to control the release by pH-response; another control experiment was also investi-

Hydrogels Composed of the RATEA16 Peptide

Biomacromolecules, Vol. 9, No. 6, 2008

1517

Figure 7. AFM images for 0.5% (wt/vol) samples at acidic and basic environments: (a) pH 3.5 and (b) pH 12.5.

the scaffolds were released. Therefore, the controlled release can be performed in this pH-responsive system, and the diffusion kinetics of the drug depends on their molecular structures, such as the available charged groups and hydrogen bonds. Finally, the self-assembled hydrogel could be formed under physiological pH condition, so it also has a promising prospect in other biomedical applications, such as tissue engineering.

Conclusions

Figure 8. Time courses of vitamin B1 release (() caused by pH-response, and ()) caused by control experiment. Mean represents average of three repeats; error bars are (1 standard deviation.

gated without acidification. The time courses of release ratio (Figure 8) were quantitatively determined using UV/visible spectroscopy. It is interesting to find that vitamin B1, whose molecular weight is much lower than that of peptide RATEA16, was released very slowly in the control experiment, probably due to formation of hydrogen bonds between the amine and the hydroxyl groups of vitamin B1 and amino acids of the peptide nanofibers at neutral pH. In addition, vitamin molecules were released faster in the acidification experiment, indicating that the electrostatic interactions between the solutes and the nanofibers in networks had a significant effect on the rate of release. The increasing release rate was attributed to the increase in electrostatic repulsions under acidic condition, rendering the vitamins increasingly free to diffuse through the hydrogel network. Furthermore, the protonated vitamins and the positively charged nanofibers repulsed each other with the decreasing environmental pH, resulting in a faster release. In fact, vitamins were released concurrent with network collapse; the hydrogels underwent a slow gel-to-sol transformation and changed into a very viscous liquid after being exposed to an acidic atmosphere of HCl vapor for a long time, and the final pH of the system was near 3. Hydrogels that undergo pH-dependent sol-to-gel transformation have been most frequently used to develop controlledrelease formulations for oral administration. The pH in the stomach (