Swelling and Mechanical Behaviors of Chemically Cross-Linked

Biomacromolecules 2003, 4, 572-580

572

Swelling and Mechanical Behaviors of Chemically Cross-Linked Hydrogels of Elastin-like Polypeptides Kimberly Trabbic-Carlson, Lori A. Setton, and Ashutosh Chilkoti* Department of Biomedical Engineering, Box 90281, Duke University, Durham, North Carolina 27708-0281 Received September 9, 2002; Revised Manuscript Received December 16, 2002

Genetically engineered elastin-like polypeptides consisting of Val-Pro-Gly-X-Gly repeats, where X was chosen to be Lys every 7 or 17 pentapeptides (otherwise X was Val), were synthesized and expressed in E. coli, purified, and chemically cross-linked using tris-succinimidyl aminotriacetate to produce hydrogels. Swelling experiments indicate hydrogel mass decreases by 80-90% gradually over an approximate 50 °C temperature range. Gels ranged in stiffness from 0.24 to 3.7 kPa at 7 °C and from 1.6 to 15 kPa at 37 °C depending on protein concentration, lysine content, and molecular weight. Changes in gel stiffness and loss angle with cross-linking formulation suggest a low-temperature gel structure that is nearly completely elastic, where force is transmitted almost exclusively through fully extended polypeptide chains and chemical crosslinks, and a high-temperature gel structure, where ELP chains are contracted and force is transmitted through chemical cross-links as well as frictional contact between polypeptide chains. Introduction Environmentally sensitive polymer hydrogels are a promising class of materials for microactuation, sensors, and for the control of cell-material interactions. The mechanics of contractile gels have been well studied for a variety of polymeric gels capable of undergoing volumetric changes through chemical reactions (e.g., polyelectrolytes with ionizable, redox, or photoactive groups), ion exchange phenomena, phase changes, and “order-disorder” transitions induced by pH, ionic strength, or thermal shifts.1,2 This class of polymer hydrogels generates force by physicochemically induced swelling changes, which modify gel-solvent interactions. The force generation capabilities of these hydrogels have the potential to address the need for microactuator materials in emerging areas of technology, such as robotics and microelectromechanical systems (MEMS).3,4 Bioinspired, muscle-like actuators, which mimic the structure-function of native muscle, are of great interest because of the attractive power/weight ratios, linear displacement, and mechanical flexibility of native muscle.5,6 Ideally, an optimal musclelike actuator will exhibit large linear displacements, tunable compliance (or stiffness), high power-to-weight ratios and make use of convenient and environmentally safe energy to initiate actuation.1 Because force generation is governed by fluid movement through the system,5,7,8 control of gel architecture is fundamental to the power generation properties of the gel. Yet, the majority of polymer gels, to date, have been fabricated from synthetic polymers with polydisperse molecular weight distributions, using nonspecific crosslinking techniques, which preclude an ability to determine precise relationships between polymer structure and physical function. * To whom correspondence may be addressed: Department of Biomedical Engineering, Box 90281, Duke University, Durham, NC 27708; phone, (919) 660-5373; fax, (919) 660-5362; e-mail, [email protected].

Genetically encoded biopolymers provide for exquisite control over molecular details such as polymer chain length (molecular weight) and amino acid sequence (chemical specificity for cross-linking sites) and, thus, are outstanding for the detailed investigation of mechanisms governing polymer gel structure formation and associated impact on material performance. Elastin-like polypeptides (ELPs) are a class of polypeptides, inspired by the amino acid sequence of natural elastin, which are composed of oligomeric repeats of the pentapeptide sequence Val-Pro-Gly-X-Gly (VPGXG) (X is any amino acid except Pro). These repetitive biopolymers exhibit conformational and aggregation sensitivity to environmental conditions such as temperature, ionic strenth, pH, and redox state. Below a critical temperature (Tt), monomeric ELPs are soluble in aqueous solution. When the temperature exceeds this critical limit, ELPs undergo a sharp (2-3 °C range) phase transition leading to contraction, desolvation, and aggregation of the polypeptide. The critical temperature at which this transition occurs is altered by amino acid sequence, molecular weight, ELP concentration, ionic strength, pH, and redox state, and the transition is completely reversible upon cooling.9 It has been shown that ELPs can be cross-linked to form hydrogel networks and that these gel networks maintain contractile sensitivity to environmental changes such as temperature. Urry and co-workers have formed ELP hydrogels through nonspecific radical cross-linking of ELPs in the coacervate state using γ-irradiation and have demonstrated preservation of the inverse temperature phase transition.10,11 In preliminary studies from our own laboratory, hydrogels formed in this manner undergo isotropic dimensional changes resulting in a more than 60% reduction in volume as water is expelled through the inverse phase transition, thus demonstrating the potential of ELP cross-linked hydrogels to serve as microactuation materials.12 However, the non-

10.1021/bm025671z CCC: $25.00 © 2003 American Chemical Society Published on Web 03/15/2003

Chemically Cross-Linked ELP Hydrogels

specific nature of radical cross-linking with γ-radiation limits our ability to understand how molecular details such as crosslink density affect the mechanical and microactuation properties of ELP hydrogels. Conticello and co-workers showed that hydrogels could be fabricated from ELPs with engineered chemoselectivity for cross-link formation at precisely spaced periodicities along the polypeptide backbone.13,14 The gene sequence for a single ELP was designed with lysine residues spaced every six pentapeptides, and the protein was microbially expressed. Purified ELP was chemically cross-linked using commercially available homobifunctional amine-reactive crosslinkers to produce ELP hydrogels that, similar to the behavior of γ-radiated gels, also exhibited reversible temperaturedependent expansion and contraction. Additionally, they demonstrated that cross-linking in organic solvent, where ELP molecules exhibit no inverse phase transition, leads to the formation of hydrogels that are more uniform in structure. Welsh and Tirrell also showed that the stiffness of ELP gels could be controlled through chemoselective cross-linking of dried thin films of ELP.15 In this study, cross-link density was controlled by varying the chain length of ELP molecules with chemically reactive end groups. These studies provided a good first step toward the development of engineered biomimetic microactuators but did not explore the effects of gel formulation on the stimuli responsive mechanical properties in any detail. In this study, we systematically investigate the effects of temperature, ELP molecular weight, concentration, and reactive group spacing on the swelling and mechanical behavior of chemically cross-linked ELP hydrogels, including measures of swelling change, dynamic shear modulus (G*), and loss angle (δ). By independently changing each formulation variable, we hope to gain a fundamental knowledge of how engineered molecular details and gel formulation affect the mechanical and swelling properties of chemically crosslinked, thermally responsive ELP hydrogels. Advances in our knowledge of the relationships between gel formulation and physical properties will open a new avenue for the design and manipulation of bioinspired materials for microactuation, sensing, and cell-material interactions. Experimental Section Monomer Gene Synthesis and Gene Oligomerization. We synthesized two gene monomers encoding for ELP polymers with different lysine contents. Monomer genes encoding for VPGKG(VPGVG)n, where n is 6 (ELP[KV6]) or 16 (ELP[KV16]) were assembled in pUC19 using standard molecular biology techniques16 from chemically synthesized oligonucleotides. The monomer genes were then oligomerized in pUC19 by a protocol called recursive directional ligation.17 In the nomenclature used in this paper, the chemical composition (lysine content) of each polymer is defined by the stoichiometry of the guest residue, X, in the ELP pentapeptide sequence, VPGXG, and the length of the ELP polymer is denoted by the number of pentapeptides in polymer. For example, ELP[KV6-224] denotes an ELP having a total length of 224 pentapeptides with a 1:6 ratio

Biomacromolecules, Vol. 4, No. 3, 2003 573

of lysine to valine residues in the X position. Both ELP[KV6] and ELP[KV16] monomer genes were oligomerized to produce genes encoding ELP polymers ranging from 3.1 to 93.4 kDa and from 7.7 to 84.8 kDa, respectively. Genes encoding each polymer were excised from their respective cloning vectors and separately ligated into a pET-25b expression vector (Novagen, Madison, WI), which had been previously modified to introduce a unique SfiI site for ELP insertion and a C-terminal Trp for spectrophotometric detection of the expressed protein. The resulting modified pET vectors encoding for lysine-based ELP polymers were individually transformed into the Escherichia coli strain, BLR(DE3) (Novagen, Madison, WI) for expression. ELP Expression and Purification. ELP[KV6] and ELP[KV16] gene products encoding for polymers of varying molecular weight were expressed in 1 L cultures of CircleGrow media (Qbiogene, Carlsbad, CA), which had been supplemented with 100 µg/mL ampicillin. These 1 L cultures were inoculated with cells from 10 mL of a starter culture (250 mL flask containing 50 mL of medium supplemented with 100 µg/mL ampicillin) that was inoculated from frozen (-80 °C) DMSO stocks and grown overnight. The 1 L cultures were grown without induction for 24 h at 37 °C with shaking at approximately 200 rpm. This procedure, utilizing gentle shaking without the use of IPTG to induce expression of the target protein, results in the efficient expression of ELPs18 with typical yields of 200-400 mg/L of culture media for these lysine-based ELPs, depending on molecular weight. Each ELP was purified from soluble E. coli lysate using inverse transition cycling, a protein purification technique that we have used and discussed previously.19 In a typical purification, the temperature and/or ionic strength of the soluble lysate was increased to cause aggregation of the ELP in the cell lysate, and the aggregated ELP was separated from soluble E. coli proteins by centrifugation. The pellet containing the ELP coacervate was resuspended in cold PBS and centrifuged to remove insoluble contaminants. This technique of thermal cycling and centrifugation was repeated (usually three times) until the ELP was determined to be approximately 95% pure of E. coli contamination by visualization of copper stained SDS-PAGE gels. Characterization of Expressed ELP Proteins. The concentration of ELPs were measured using a Shimadzu Scientific Instruments UC-1601 spectrophotometer and the molar extinction coefficient at 280 nm (5690 M-1 cm-1). Gross estimation of purity and molecular weight were characterized using SDS-PAGE (BioRad, Hercules, CA) with copper staining to visualize ELP proteins. Matrixassisted laser desorption/ionization mass spectrometry (MALDI-MS) was used for precise molecular weight determination of each expressed ELP using a PE Biosystems Voyager-DE instrument equipped with a nitrogen laser (337 nm). The MALDI-MS samples were prepared in an aqueous 50% acetonitrile solution containing 0.1% trifluoroacetic acid, using a sinapinic acid matrix. The solution turbidity of the lysine-based ELPs was measured as a function of temperature to determine the effects of amino acid sequence, molecular weight, and ionic

574

Biomacromolecules, Vol. 4, No. 3, 2003

strength on the inverse transition temperature. ELP solutions (25 µM ELP in PBS and in PBS + 1 M NaCl) were heated at a rate of 1 °C/min, and the optical density at 350 nm (OD350) was measured by a Cary 300 UV-visible spectrophotometer equipped with a multicell thermoelectric temperature controller (Varian Instruments, Walnut Creek, CA). The inverse transition temperature was defined as the temperature at which the OD350 reached 5% of its maximum. Gel Formation. Chemical Cross-Linking. A low (2224 kDa), medium (43-47 kDa), and high (85-93 kDa) molecular weight variant of both the high lysine content polymer, ELP[KV6], and the lower lysine content polymer, ELP[KV16], were cross-linked to form hydrogels. Because the inverse transition temperature is highly sensitive to the hydrophobicity of the ELP,20 it was confirmed in preliminary experiments (results not shown) that chemical cross-linking in an aqueous medium results in the formation of inhomogeneous gels due to the local occurrence of the ELP inverse phase transition and phase separation of the cross-linked ELP from the aqueous solvent as charged lysine residues are consumed in the cross-linking reaction. To eliminate this complication, ELPs were transferred to an organic solvent where they exhibit no conformational sensitivity to temperature or charge state. Preliminary experiments also indicated that tris-succinimidyl aminotriacetate (TSAT) (Pierce Endogen) formed gels at lower protein concentrations than a chemically similar bifunctional cross-linking agent. This is not unexpected because a gel network can be formed if only two of the three available reactive groups are involved in intermolecular cross-links. The third reactive group can further facilitate the formation of a more densely cross-linked network. Thus, TSAT was chosen as the cross-linker of choice for gelation experiments. Purified, aqueous ELP solutions were dialyzed against distilled water at 4 °C for 48 h to remove excess salt, lyophilized, and resuspended in a solution of 15% dimethylformamide (DMF) and 85% anhydrous dimethyl sulfoxide (DMSO) to make stock solutions with ELP concentrations ranging from 280 to 420 mg/mL, which were stored frozen at -20 °C until crosslinking. ELPs were chemically cross-linked using TSAT, which was dissolved in the same organic solvent (15% DMF in DMSO) to a concentration of 50 mg/mL cross-linker. Aliquots of TSAT solution were stored at -80 °C until crosslinking. Prior to cross-linking, ELP stock solutions were diluted with additional solvent (15% DMF in DMSO) such that the addition of the TSAT solution at a 1:1 ratio between ELP amine and succinimidyl ester groups resulted in final ELP concentrations of 50, 75, 100, 125, 150, and 175 mg/ mL. The solvent, 15% DMF in DMSO, was chosen to balance the need for high ELP solubility (DMSO) with a reasonably low solution freezing point (DMF lowers the freezing point of DMSO) such that the reactants could be chilled on ice prior to cross-linking to slow the chemical reaction and allow better mixing. Cylindrical ELP gels (∼8 mm in diameter and 2.5 mm in thickness) were created by rapid vortex mixing of chilled ELP and TSAT solutions (125 µL reaction volume) in a 2 mL polypropylene microcentrifuge tube where the conical bottom had been filled with

Trabbic-Carlson et al.

paraffin wax to create a flat reaction surface. Reaction tubes were briefly centrifuged in a swinging bucket microcentrifuge to remove air bubbles and were allowed to gel undisturbed for at least 24 h at room temperature. Gels were removed from the microcentrifuge tubes by adding 1 mL of a concentrated NaCl solution (PBS + 4 M NaCl) to each gel. This caused the gels to contract as the organic solvent was exchanged for the salt solution, and the contracted gels could be easily extracted from reaction vessels and be placed in individual wells containing 15 mL of PBS. The gels were washed with gentle mixing at 4 °C for 48 h in two exchanges of PBS with 0.02% sodium azide (a 240-fold volume excess of the cross-linking volume). Gels were stored in PBS with 0.2% sodium azide at 4 °C until testing. Cross-linked hydrogels of ELP[KV6-112] and ELP[KV16-102] were fabricated in triplicate for physical characterization. Characterization of Gel Swelling. The effect of temperature on the swelling behavior of ELP hydrogels was determined from the mass of the ELP hydrogels as a function of temperature. Gels were incubated in PBS at successively increasing temperatures between 4 and 50 °C. Gels were incubated for at least 2 h at each temperature allowing ELP contraction to reach steady state before weighing. The swelling ratio was calculated from the normalization of gel mass at each temperature to the estimated mass of hydrated gel at the casting volume of 125 µL. Since at least 80% of the volume of the cast gel is occupied by solvent, we assume that in PBS all ELP gels have a density of approximately 1 g/mL at the casting volume. Thus, the swelling ratio was calculated as the ratio of gel mass relative to 0.125 g. Swelling ratios greater than 1 indicate that gels have imbibed water and have a volume greater than their casting volume, and swelling ratios less than 1 indicate that gels have expelled water and have contracted to a volumes smaller than the casting volume. Material Property Characterization. Mechanical tests were performed on a strain-controlled rheometer (ARES, Rheometric Scientific, Piscataway, NJ) equipped with torque (0.2-200 g cm) and normal force (0-2000 g) transducers in a parallel plate configuration. Sample temperature was maintained by a circulating fluid bath of PBS at one of two test temperatures, 7 or 37 °C. Samples were first incubated at the appropriate test temperature for at least 2 h prior to testing. The thickness of the ELP test sample was measured as the distance between the two parallel plates with the equilibrated sample under a 10 gf tare load, and thickness varied from 1.0 to 3.1 mm for all samples. The diameter of parallel plates used for each test, which varied from 6 to 12 mm, was chosen to ensure that the ELP test sample fully covered the plates. In protocol development, a dynamic strain sweep (sinusoidal strain of 0.0001-0.10 maximum strain amplitude, ω ) 10 rad/s) was performed to determine the range of strain amplitudes for which ELP gels exhibit linear stress-strain behaviors. Results confirmed that linear behaviors were observed up to 0.08 strain amplitude, so that all subsequent tests were performed for a maximum strain amplitude of 0.05. To measure the dynamic shear modulus (a measure of dynamic matrix stiffness) and loss angle (a measure of energy



dissipation) in the ELP gel constructs, a series of dynamic frequency sweep tests were performed. A sinusoidal shear strain profile with an amplitude of 0.05 strain was applied to the test samples at angular frequencies between 0.1 and 100 rad/s. The ELP hydrogel constructs exhibited insignificant frequency-sensitive behaviors; thus, the magnitude of the dynamic shear modulus (|G*|) and loss angle (δ) are reported from data at a single frequency, ω ) 10 rad/s, to compare physical properties among gels of varying polypeptide concentration, molecular weight, and amino acid composition. Results and Discussion Characterization of Un-Cross-Linked ELP Polymers. The ELPs were expressed in and purified from E. coli such that only a single monomeric band was visible on copperstained SDS-PAGE gels. This was generally accomplished after three rounds of inverse transition cycling, and we estimate that the ELP proteins are at least 95% pure of other E. coli protein contaminants. Mass spectrometry verified that the expressed ELP molecular weight correlated well with the anticipated molecular weights from gene design. The molecular weights of all constructs expressed as determined by MALDI-MS were within 0.1% of the calculated values, and no additional peaks were observed, further confirming the purity of the ELPs. It was anticipated that lysine-based ELPs would exhibit inverse transition behavior typical of other ELP proteins, and their sensitivity to temperature was explored through turbidity measurements. Figure 1A shows, as an example, the turbidity profiles of the medium molecular weight variant of both ELP[KV6] and ELP[KV16]. Not surprisingly, lysine-based ELPs exhibit a solution inverse phase transition characterized by a sharp increase in solution turbidity at a critical temperature, which is completely reversible upon cooling, although heating and cooling turbidity profiles indicate that aggregation and solubilization follow different pathways. Figure 1B summarizes the transition temperatures of uncross-linked (25 µM) solutions of ELP[KV6] and ELP[KV16], showing the effects of lysine content, ionic strength, and molecular weight on the inverse transition temperature. The inverse transition temperature of both lysine-based ELPs decreases with molecular weight. This is consistent with the behavior of other ELP sequences, which have been systematically oligomerized by recursive directional ligation,17 and qualitatively parallels the observations of Urry et al. 21 Polypeptides of ELP[KV6], which have over 2-fold more charged amine groups, exhibit inverse transition temperatures approximately 10 °C higher than ELP[KV16] polymers of similar molecular weight. The higher charge density of ELP[KV6] also results in a greater sensitivity of the inverse transition temperature to ionic strength. Increasing the NaCl concentration by 1 M results in 15-20 °C decrease in the inverse transition temperature of ELP[KV16] polymers and a 20-30 °C decrease for ELP[KV6] polymers. Chemical Cross-Linking/Hydrogel Formation. Crosslinking of lysine-based ELPs did not result in the formation of hydrogels amenable to swelling and mechanical property

Figure 1. (A) Heating (solid) and cooling (dotted) turbidity profiles of ELP[KV6-112] and ELP[KV16-102] polymers showing reversibility of the inverse phase transition. (B) The effects of polypeptide length, primary sequence, and ionic strength on the inverse transition temperature of 25 µM solutions of ELP[KV6] and ELP[KV16] polymers in PBS (solid markers) and PBS + 1 M NaCl (open markers). Table 1. ELP Formulations at Which TSAT Cross-Linked Hydrogels Were Obtained low Mw

medium Mw

high Mw

KV6-56 KV16-51 KV6-112 KV16-102 KV6-224 KV16-204

Mw (kDa) no. of amines 50 mg/mL ELP 75 mg/mL ELP 100 mg/mL ELP 125 mg/mL ELP 150 mg/mL ELP 175 mg/mL ELP

23.9 10

gel gel gel

21.7 5

47.1 18

42.7 8

gel

gel gel gel gel gel

gel gel gel gel gel

93.4 34 gel gel gel gel gel gel

84.8 14

gel gel gel

characterizations (Table 1) under all cross-linking formulations. None of the constructs formed hydrogels at concentrations less than 50 mg/mL. SDS-PAGE analysis of crosslinking reactions that did not result in gel formation revealed the formation of oligomers, indicating that there are insufficient intermolecular contacts to promote the formation of a solidlike hydrogel network. ELP solutions with concentrations greater than 175 mg/mL were too viscous to allow adequate mixing during the cross-linking reaction, resulting in the formation of inhomogeneous gels. At the lower concentrations, only higher molecular weight and higher lysine content ELPs formed hydrogels. Gels were formed at concentrations as low as 50 mg/mL from ELP[KV6-224], the highest molecular weight polymer with the highest lysine content. Gels were not formed at this concentration for either the medium or low molecular weight variant of ELP[KV6], suggesting that the increase in ELP molecular weight promotes the formation of intermolecular contacts and cross-links, leading to the formation of a gel

576



Figure 2. Swelling behavior of hydrogels constructed from 150 mg/ mL solutions (A) ELP[KV6] and (B) ELP[KV16] polymers of varying molecular weight. Mass of ELP gels gradually decreases with temperature as a result of ELP contraction and exclusion of water. Gels of lower lysine content, ELP[KV16], and lower molecular weight undergo greater degrees of both swelling and contraction.

Figure 3. Swelling measurements of (A) ELP[KV6-112] and (B) ELP[KV16-102] as a function of increasing temperature and ELP concentration. Mass of ELP gels gradually decreases with temperature as a result of ELP contraction and exclusion of water. Gels of lower lysine content, ELP[KV16-102], and lower ELP concentration undergo greater degrees of both swelling and contraction.

network. Similarly, increasing the number of lysine residues increases the total number of cross-links and, in the process, the total number of intermolecular cross-links formed at each protein concentration. ELP[KV16-51] only formed gels at the highest ELP concentration investigated, 175 mg/mL, while its higher lysine content counterpart, ELP[KV6-56], formed gels at concentrations as low as 125 mg/mL. It is unclear why ELP[KV16-102] forms gels at the concentrations as low as 75 mg/mL. According to trends at lower and higher molecular weights, we would not expect ELP[KV16-102] to gel at concentrations less than 125-150 mg/mL. Hydrogel Swelling. The inverse phase transition of ELP hydrogels was studied by the degree of gel swelling and contraction in PBS solution as a function of temperature (Figures 2 and 3). Swelling ratio is a useful measure of ELP gel dimensional changes because gels both swell and contract to varying degrees depending on gel formulation. Ratios greater than 1 indicate that gels that have swollen and have masses greater than 0.125 g, the hydrated mass of the casting volume, and ratios less than 1 indicate that gels have contracted and have masses less than 0.125 g. As has been observed for other cross-linked ELP gels,10-14 these chemically cross-linked ELP gels contract in response to increasing temperature. At cold temperatures, ELP gels swell in aqueous solvent, and the ELP accounts for approximately 2-8% of the gel mass. As they are heated, they contract and shed water resulting in a 3- to 35-fold increase in protein concentration depending on the gel formulation. At 37 °C, the ELP

accounts for 25-65% of the gel mass depending on ELP concentration, molecular weight, and lysine content. The most striking feature of the swelling profile is the gradual nature of ELP hydrogel contraction and water loss with increasing temperature. Unlike turbidity measurements, which suggest that the ELP inverse transition is a discrete event that takes place over a very narrow temperature range, swelling measurements suggest that ELP molecules undergo molecular rearrangement over a much broader temperature range. This gradual ELP contraction is not unique to these ELP hydrogels and has been observed in ELP gels crosslinked by γ-radiation as well,12,22,23 yet the reasoning behind the difference in thermal response has not been fully explained. It is possible that the stochastic nature of the crosslink formation between ELP chains broadens the transition by introducing a range of effective ELP segment lengths, each with its own discrete transition temperature. A similar phenomenon has been documented for mixtures of ELPs, which exhibit separate and distinct aggregation events corresponding to the transition temperatures of the individual solution components.24 In this case, the broad gel transition observed could reflect the superposition of these independent ELP segment transitions. An alternative interpretation is that turbidity and swelling measurements probe two different aspects of the ELP inverse transition. Molecular dynamics simulations of single ELP molecules in aqueous solution suggest that ELP contraction is a gradual process with temperature.25 While it is possible that no critical phase behavior was observed in these single


ELP molecule simulations simply because the inverse phase transition is an intermolecular phenomenon, it is also possible that aggregation is a secondary event, which is the end result of a gradual change in ELP conformation and dehydration with temperature. Since it would be impossible for ELP molecules in a gel to aggregate in a manner similar to their solution behavior, we suggest an alternative hypothesis that the gradual nature of ELP gel contraction is reflective of the thermal behavior of ELP monomer in the absence of aggregation. Figure 2 shows the effect of molecular weight on the temperature-dependent swelling ratio of hydrogels produced from ELP[KV6] and ELP[KV16] polymers. In contrast to the behavior of monomeric ELP solutions, where the inverse transition temperature can vary by more than 30 °C depending on molecular weight, ELP gels undergo the majority of their contraction over a similar temperature range regardless of the molecular weight of the ELPs from which they were constructed. This suggests that the gels are more chemically similar than are the polymers from which they were produced. Regardless of composition and molecular weight of the starting material, cross-linking reduces the difference between ELPs by consuming charged lysine groups and producing polymer networks of essentially infinite molecular weight, where only the total protein content and segment length between cross-links differentiate the gels and their swelling properties. Gels constructed from ELP[KV6] show that the degree of swelling and contraction is inversely correlated with molecular weight (Figure 2A); gels constructed from the lowest molecular weight polymers swell and contract to the greatest degree. Trends in the behavior of ELP[KV16] gels with polymer molecular weight (Figure 2B) are not as clear since gels could only be obtained from the medium and high molecular weight constructs. At 7 °C gels cast from the lower lysine density polymer, ELP[KV16], swell 20 to 50% more than gels of ELP[KV6]. Figure 3 shows the effect of increasing temperature on the swelling ratio of gels constructed from differing concentrations of ELP[KV6-112] and ELP[KV16-102] polymers. At 7 °C, gels with the lowest ELP concentration swell 4565% more than gels with the highest ELP concentration. These highly swollen, low ELP concentration gels also contract to the greatest degree exhibiting the lowest swelling ratio at high temperature. Gels cast from the low lysine content polymer, ELP[KV16-102], swell 20-50% more than gels cast from high lysine content polymer, ELP[KV6-112], and the difference is inversely correlated with the ELP concentration. ELP concentration can slightly alter the temperature range of gel contraction. Gels fabricated from the lower lysine content polymer, ELP[KV16-102], show their maximum contraction by 37 °C, and no appreciable water loss is observed above this temperature for any of the concentrations investigated (Figure 3B). Gels constructed from the higher lysine content polymer, ELP[KV6], exhibit a slightly broader transition and continue to contract and lose water at temperatures above 37 °C especially at higher ELP concentrations (Figure 3A). This behavior may be evidence of a small fraction of residual charged lysine groups which have not


Figure 4. Effect of molecular weight on the (A) complex modulus and (B) loss angle of ELP hydrogels constructed from 125 mg/mL solutions of ELP[KV6] at 7 °C (open circles) and 37 °C (solid circles). Error bars represent the first standard deviation calculated from tests of three gels of ELP[KV6-112].

been consumed in the cross-linking reaction, which is consistent with the findings of McMillan and Conticello, who observed that approximately 15% of the lysine groups were not consumed during a similar cross-linking reaction.14 At high temperatures, the propensity for ELP molecules to gain free energy by contracting and shedding bound water would be balanced by energetically unfavorable repulsion between charged lysine residues as they are brought in closer proximity to one another. If residual charged residues are responsible for broadening the inverse phase transition of ELP gels and producing gels that are more expanded at higher temperatures, the deliberate incorporation of charged residues could be exploited as a means of controlling gel pore size at higher temperatures. Mechanical Properties. The effects of temperature and molecular weight on the magnitude of the dynamic shear modulus and phase angle of ELP hydrogels constructed from 125 mg/mL solutions of ELP[KV6] are shown in Figure 4. Raising the temperature from 7 to 37 °C generally results in a 2- to 3-fold increase in the magnitude of the dynamic shear modulus for all molecular weights, giving evidence of stiffer hydrogels at the higher temperatures (Figure 4A). Phase angles are very low and nearly zero (Figure 4B) at both high and low temperatures, characteristic of a highly elastic, energy-storing hydrogels. Slightly larger phase angles at high temperature are observed for each molecular weight indicating an increase in viscous frictional losses; however, this increase does not represent a significant change in the overall elastic nature of the ELP hydrogel.

578


Figure 5. Effect of ELP concentration on the (A) complex modulus and (B) loss angle of ELP[KV6-112] (circles) and ELP[KV16-102] (squares) at 7 °C (open symbols) and 37 °C (solid symbols). Error bars represent the first standard deviation calculated from tests of three gels for each ELP polymer at each ELP concentration.

ELP molecular weight alters the material properties of cross-linked hydrogels at both low and high temperature. Gels become stiffer with molecular weight at both high and low temperature, although the protein concentration is held constant. This suggests that increasing ELP molecular weight promotes the formation of a tighter network upon crosslinking, with a greater degree of intermolecular cross-links. Phase angle remains relatively constant with molecular weight at low temperature. Although trends are not as clear at high temperature, the loss angle appears to decrease with increasing molecular weight. The concentration-dependent behaviors of the ELP hydrogels are shown in Figure 5 for gels formed from ELP[KV6-112] and ELP[KV16-102]. At 7 °C, gels of both ELP polymers have very similar mechanical properties and exhibit increases in the magnitude of dynamic shear moduli (Figure 5A), but not loss angle (Figure 5B), with increasing concentration. At 37 °C, dynamic shear modulus increases with ELP concentration, and dynamic shear moduli of all gels are 3- to 4-fold higher than their respective values at 7 °C. However, significant differences were also apparent between the two ELP formulations, with larger values for the loss angle associated with hydrogels formed from the lower lysine-content gels of ELP[KV16-102] (Figure 5B). Since viscous energy losses likely result from friction between polypeptide chains, the elevated loss angles of ELP[KV16-102] gels suggest that the lower cross-link density allows greater molecular rearrangement in response to applied strain. Furthermore, loss angles of both ELP[KV6112] and ELP[KV16-102] gels exhibit moderate sensitivity


to concentration at the higher temperature, suggesting that tighter, more elastic networks are formed in the higher concentration ELP formulations. Structural Model of Temperature-Sensitive ELP Gels. Because there are multiple cross-linking sites periodically spaced along the length of the ELP molecules, cross-linking is a random process likely resulting in a wide range of molecular weights between cross-links and gels probably contain a mixture of intra- and intermolecular cross-links. The swelling and mechanical property characterizations are consistent with the formation of two types of cross-links in the ELP hydrogels: functional cross-links that transmit force through the gel and nonfunctional cross-links that are not directly involved in the transmission of force. Nonfunctional cross-links could serve to incorporate protein in the gel, likely through the formation of loops and non-load-bearing polypeptide segments. A schematic representation of gel structure showing the functional and nonfunctional cross-links and their structural role in response to applied strain and temperature can be seen in Figure 6. At 7 °C, ELP gels are swollen with buffer and the ELP segments are fully extended (Figure 6A). In response to strain, force is transmitted largely through these fully extended polypeptide chains and chemical cross-links (Figure 6B). Very low loss angles suggest that polypeptide chains in the low-temperature gels do not have much contact with one another regardless of gel formulation, and chain reorientation in response to applied strain occurs in the absence of substantial intermolecular frictional interactions. At 37 °C, gels are contracted with resultant increases in ELP concentration due to water loss (Figure 6C). The higher values for loss angle under these conditions suggest a contracted ELP structure where functional cross-links no longer bear the entire load and that some portion of the load is transferred by friction between interacting ELP chains. Figure 6 presents a rather simplistic model of ELP gel structure and in no way addresses the physical distance between functional cross-links and the relative numbers of functional and nonfunctional cross-links. However, studying how gel formulation affects swelling and mechanical properties yields insights into likely trends in these structural features. Intermolecular contact between macromolecules increases with the molecular weight and concentration of polymers in solution.26 Thus, it is reasonable to assume that increasing ELP molecular weight and concentration promotes intermolecular, functional cross-links, resulting in tighter gel networks with shorter segment lengths between cross-links. This is consistent with our observations that swelling and high-temperature loss angles decrease and that stiffness increases with increasing ELP concentration and molecular weight. Conversely, conditions that minimize intermolecular interactions (e.g., low molecular weight and concentration) promote the formation of intramolecular and likely nonfunctional cross-links leading to longer average segment lengths between functional cross-links, increased swelling and high temperature loss angles, and reduced stiffness. Increased lysine content also results in stiffer, more elastic gels, but simply because the total number of cross-links is increased. Higher shear moduli and lower loss angles 37 °C for ELP-



Figure 6. Proposed structural model that schematically represents the effects of temperature and applied strain on ELP hydrogels. Panels A and B represent proposed gel structure at 7 °C in the absence and presence of applied strain, respectively. Panels C and D represent the proposed structure of the gel at 37 °C in the absence and presence of applied strain, respectively. At 37 °C, greater frictional energy losses are observed since ELP chains are contracted and far less water is present to lubricate chain reorientation in response to applied strain.

[KV6] gels are consistent with a shorter segment length between cross-links. Conclusions In this study, we have demonstrated the ability to design ELPs with precisely defined molecular weight and reactive group spacing. We have efficiently expressed and purified these polymers and devised a scheme for chemical crosslinking that allows for the formation of gels amenable to swelling and mechanical property characterization. The results of these studies suggest that structural differences in the cross-linked hydrogel are directly related to parameters of the constituent ELP formulation, such as molecular weight, solution concentration, and lysine content. At low temperatures, the ELP hydrogels are completely elastic in nature, transmitting force through extended ELP segments and functional, intermolecular cross-links formed by the chemical cross-linking reaction. As the temperature is increased, gels contract and shed water and viscous frictional energy losses arise from intermolecular interactions between condensed polypeptide segments. The degree of this viscous energy loss is most significant in gels that were formed at low ELP concentration, from low molecular weight polymers, and with lower lysine content. These conditions promote the formation of intramolecular and likely nonfunctional cross-links, generally leading to weaker gels and longer segments between functional cross-links. The results of this study demonstrate that genetically engineered ELP polymers may be chemically cross-linked through precisely spaced reactive groups to yield thermally

responsive hydrogels with “tunable” physical properties. The three parameters studied heresELP molecular weight, concentration, and lysine contentshad distinct contributions to hydrogel properties that would impact the materials’ performance as a microactuator, sensor, or scaffold. The degree of swelling and contraction as well as materials’ stiffness and energy dissipation properties were highly sensitive to all three gel formulation parameters. Broadening of the inverse transition at high lysine content and ELP concentrations, where charge repulsion between unreacted lysine residues would be most significant, suggests that the deliberate incorporation of charged residues may be used to tune ELP porosity at high temperatures. Stiffness differences among formulations at low temperature were generally amplified at higher temperature, demonstrating a sensitivity that could be exploited to obtain tunable mechanical properties that may be attractive for the design of bioinspired microactuators. In conclusion, these results suggest the potential to synthesize an array of hydrogel materials with well-defined and varied physical properties to address a wide range of design specifications for applications in microactuation and tissue engineering. Acknowledgment. We thank the Office of Naval Research (N00014-00-1-0184) and the National Institutes of Health (AR47442) for supporting this research. References and Notes (1) DeRossi, D.; Suzuki, M.; Osada, Y.; Morasso, P. J. Intell. Mater. Syst. Struct. 1992, 3, 75. (2) Hirasa, O. J. Intell. Mater. Syst. Struct. 1993, 4, 259.

580


(3) Fatikow, S.; Rembold, U. Microsystem Technology and Microbiotics; Springer: Berlin, 1997. (4) Fujimasa, I. Micromachines: A New Era in Mechanical Engineering; Oxford University Press: Oxford, 1996. (5) DeRossi, D.; Chiarelli, P. In Macro-Ion Characterization: From Dilute Solutions to Complex Fluids; ACS Symposium Series 548; American Chemical Society: Washington, DC, 1994; pp 517-530. (6) Hunter, I. W.; S., L. Tech. Dig. IEEE Solid-State Sens. Actuator Workshop 1992, 178-185. (7) Chiarelli, P.; Basser, P. J.; DeRossi, D.; Goldstein, S. Biorheology 1992, 29, 383. (8) Chiarelli, P.; DeRossi, D. Polym. Gels Networks 1996, 4, 499. (9) Urry, D. W. J. Phys. Chem. 1997, 101, 11007-11028. (10) Urry, D. W. J. Protein Chem. 1988, 7, 1-34. (11) Urry, D. W.; Haynes, B.; Zhang, Z.; Harris, R. D.; Prasad, K. U. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 3407-3411. (12) Betre, H.; Meyer, D. E.; Chilkoti, A.; Setton, L. A. Trans. Orth. Res. Soc. 2001, 26, 601. (13) McMillan, R. A.; Caran, K. L.; Apkarian, R. P.; Conticello, V. P. Macromolecules 1999, 32, 9007-9010. (14) McMillan, R. A.; Conticello, V. P. Macromolecules 2000, 33, 48094821. (15) Welsh, E. R.; Tirrell, D. A. Biomacromolecules 2000, 1, 23-30.

Trabbic-Carlson et al. (16) Ausubel, F. M.; Brent, R.; Kingston, R. E.; Moore, D. H.; Seidman, J. G.; Smith, J. A.; Struhl, K. Current Protocols in Molecular Biology; John Wiley & Sons: New York, 1995. (17) Meyer, D. E. C., A. Biomacromolecules 2002, 3, 357-367. (18) Guda, C.; Zhang, X.; McPherson, D. T.; Xu, J.; Cherry, J. H.; Urry, D. W.; Daniell, H. Biotechnol. Lett. 1995, 17, 745-750. (19) Meyer, D. E.; Chilkoti, A. Nat. Biotechnol. 1999, 17, 1112-1115. (20) Urry, D. W.; Luan, C. H.; Parker, T. M.; Gowda, D. C.; Prasad, K. U.; Reid, M. C.; Safavy, A. J. Am. Chem. Soc. 1991, 113, 43464348. (21) Urry, D. W.; Trapane, T. L.; Prasad, K. U. Biopolymers 1985, 24, 2345-2356. (22) Urry, D. W.; Shaw, R. G.; Prasad, K. U. Biochem. Biophys. Res. Commun. 1985, 130, 50-57. (23) Urry, D. W.; Haynes, B.; Harris, R. D. Biochem. Biophys. Res. Commun. 1986, 141, 749-755. (24) Meyer, D. E.; Kong, G. A.; Dewhirst, M. W.; Zalutsky, M. R.; Chilkoti, A. Cancer Res. 2001, 61, 1548-1554. (25) Li, B.; Alonso, O. V.; Daggett, V. J. Mol. Biol. 2001, 305, 581592. (26) Schoff, C. K. In Polymers: Polymer Characterization and Analysis; Kroschwitz, J. I., Ed.; John Wiley & Sons: New York, 1990.

BM025671Z

Swelling and Mechanical Behaviors of Chemically Cross-Linked

Recommend Documents