Two-Component Protein Hydrogels Assembled Using an Engineered

Jul 3, 2013 - *Mailing address: 3122 TAMU, Artie McFerrin Department of Chemical Engineering, Texas A&M University, College Station, TX 77843. Phone: ...
0 downloads 11 Views 4MB Size
Download PDF
Article pubs.acs.org/Biomac

Two-Component Protein Hydrogels Assembled Using an Engineered Disulfide-Forming Protein−Ligand Pair Dongli Guan,†,⊥ Miguel Ramirez,†,⊥ Lin Shao,§ Daniel Jacobsen,† Ivan Barrera,† Jodie Lutkenhaus,† and Zhilei Chen*,†,‡ †

Artie McFerrin Department of Chemical Engineering, Texas A&M University, College Station, Texas 77843, United States Department of Microbial and Molecular Pathogenesis, Texas A&M Health Science Center, College Station, Texas 77843, United States § Department of Chemical and Environmental Engineering, Yale University, New Haven, Connecticut 06520, United States ‡

S Supporting Information *

ABSTRACT: We present the development of a two-component self-assembling protein hydrogel. The building blocks of the hydrogel are two liquid-phase protein block copolymers each containing (1) a subunit of the trimeric protein CutA as a crosslinker and (2) one member of a PDZ-domain-containing protein−ligand pair whose interaction was reinforced by an engineered disulfide linkage. Mixing of the two building blocks reconstitutes a self-assembling polypeptide unit, triggering hydrogel formation. This hydrogel exhibits extremely high solution stability at neutral and acidic pHs and in a wide range of temperatures (4−50 °C). Incorporation of a “docking station peptide” binding motif into a hydrogel building block enables functionalization of the hydrogel with target proteins tagged with a “docking protein”. We demonstrated the application of an enzyme-functionalized hydrogel in a direct electron transfer enzymatic biocathode. These disulfide-reinforced protein hydrogels provide a potential new material for diverse applications including industrial biocatalysis, biosynthesis, biofuels, tissue engineering, and controlled drug delivery.

■

from natural sources is nontrivial,4 limiting their broad applicability. In contrast, recombinant proteins are amenable to tuning of physical properties and can be purified on a large scale, making them a more desirable alternative to natural proteins as building blocks for protein-based hydrogels.5−7 Diverse functional motifs can be easily incorporated into recombinant hydrogel-building-block proteins via genetic manipulation resulting in materials with specific and desired biological, chemical and mechanical properties. Recombinant protein hydrogels have been developed for applications in tissue engineering, biosynthesis, and biofuel applications.7−11 In this study, we report the synthesis of a novel selfassembling protein hydrogel based on the bioaffinity of a pair of engineered proteins. Gelation is triggered by the mixing of two liquid-phase protein building blocks, each containing (1) a subunit of a trimeric protein (CutA) and (2) a PDZ domaincontaining protein (Tip1) or its peptide ligand (Tip1lig). CutA serves as a cross-linking agent. A polypeptide containing a cross-linker on only one terminus is incapable of self-assembly into a network. On the other hand, a polypeptide containing

INTRODUCTION Hydrogels are three-dimensional, hydrophilic polymer networks that hold large amounts of water. On the basis of the nature of the cross-linkers, hydrogels can be categorized as either chemical or physical hydrogels.1 Chemically cross-linked hydrogels are usually very stable in solution, typically undergo a large volume change during solution-to-gel transition, and contain covalently joined cross-linkers formed through chemical reactions such as disulfide formation, (photo)polymerization or the reaction between thiols and acrylates or sulfonates.2 Physical hydrogels contain noncovalently joined cross-linkers and often self-assemble in response to external stimuli such as changes in pH and temperature. These noncovalently joined cross-linkers render the physical hydrogel susceptible to shear-thinning under mechanical stress and selfhealing upon cessation of the stress, making these hydrogels suitable for use as injectable materials. Additionally, gelation of self-assembled hydrogels does not rely on organic solvents or extraneous cross-linking reagents, making this material more favorable for biomedical applications including controlled drug delivery and tissue engineering.3,4 Many natural proteins can self-assemble into physical hydrogels, including elastin, collagen, and gelatin. However, natural proteins often have a short shelf life and purification © 2013 American Chemical Society

Received: June 4, 2013 Revised: July 1, 2013 Published: July 3, 2013 2909

dx.doi.org/10.1021/bm400814u | Biomacromolecules 2013, 14, 2909−2916

Biomacromolecules

Article

Luria−Bertani (LB) agar plate containing either 50 μg/mL kanamycin (constructs 1, 3, 5−7) or 100 μg/mL ampicillin (constructs 2, 4, 8). The next day, all the colonies (50−100) from a plate were pooled, resuspended in 5 mL of Luria Broth (LB), and transferred to 1 L of LB containing the specific antibiotic. The cells were incubated at 37 °C with shaking at 250 rpm to an optical absorbance (OD600) of 0.6−0.9. For constructs 2 and 4, protein expression was induced by the addition of isopropyl β-D-1-thiogalactopyranoside (IPTG, 1 mM) followed by incubation at 37 °C for 4 h. For all the other protein expressions, the cell cultures were cooled to 18 °C prior to the addition of 1 mM IPTG and followed by overnight incubation (∼16 h) at 18 °C. After expression, cells were harvested by centrifugation at 8000g and 4 °C for 15 min and the cell pellets were stored at −80 °C until use. For protein purification, cell pellets were resuspended in buffer A (500 mM NaCl, 10 mM Tris-HCl, pH 8.0, 1 g of wet pellet in 10 mL) and disrupted by sonication (QSonica Misonix 200, Amp 40, with 1 s pulse and 5 s pause for 2 min total pulse). For SH3-Slac (construct 5), 0.25 mM CuCl2 was added during protein expression and 1 mM CuCl2 was added to the cell lysate immediately after sonication. Whole cell lysate was centrifuged at 16 000g for 20 min at 4 °C. Target protein in the soluble lysate was purified by nickel affinity chromatography using a 5-mL Ni Sepharose High Performance HisTrap column (GE Healthcare Life Sciences, Piscataway, NJ). After equilibration with buffer A and protein loading, the column was thoroughly washed with buffer A containing 22.5 mM imidazole and the target protein was eluted in buffer A containing 150 mM imidazole. For the purification of CutA-Tip1/dsTip1 (constructs 1 and 3), 2.5 mM EDTA was immediately added to the eluted protein to minimize heavy metal-induced protein precipitation.12 To remove imidazole from the protein samples, purified proteins were bufferexchanged into buffer A via 30-kDa ultrafiltration spin columns (Amicon Ultra, Millipore, Billerica, MA), as appropriate, concentrated to ∼100 mg/mL using the same column, and stored at −80 °C until use. The concentrations of purified proteins were determined by measuring the absorbance at 280 nm using a NanoDrop 1000 (Thermo Fisher Scientific). Proteins were stored at −80 °C until use. For sodium dodecyl sulfate(SDS)-polyacrylamide gel electrophoresis (PAGE) analysis, appropriately diluted protein samples were mixed with 2× SDS sample buffer (0.5 M Tris-HCl, pH 6.8, 20% glycerol, 10% w/v SDS, 0.1% w/v bromophenol blue) that included βmercaptoethanol (2% final) in the case of reducing conditions, and the samples were incubated at 95 °C for 5 min prior to loading. The gels were stained with Coomassie brilliant blue R250 for visualization. Hydrogel Synthesis. The purified hydrogel building blocks CutAdsTip1 (construct 3) and CutA-dsTip1lig (construct 4) were manually mixed with a pipet tip at a 1:1 molar ratio at the specified concentration in buffer B (500 mM NaCl, 10 mM Tris-HCl, pH 8.0, 0.1% w/v NaN3), and the mixture was briefly centrifuged to remove trapped air bubbles prior to incubation at room temperature for 14−20 h. Hydrogels form during incubation. Characterization of Interaction between dsTip1 and dsTip1lig. dsTip1 (construct 7) and GFP-dsTip1lig (construct 8) were mixed at a 1:1 molar ratio (0.2 mM each) in buffer A supplemented with 2 mM dithiothreitol (DTT) in a 1.5 mL centrifuge tube. The tube was incubated at 22 °C in a humidified chamber to facilitate DTT evaporation. Samples were taken at different times, mixed with 2× SDS loading buffer in the presence (reducing) or absence (nonreducing) of β-mercaptoethanol, and incubated at 95 °C for 5 min. Samples were directly loaded onto a 12% SDS-PAGE gel (nonreducing) or incubated for an additional 3 min at 95 °C in the presence of 2% β-mercaptoethanol (reducing). The gels were stained with Coomassie brilliant blue R250 for visualization. Hydrogel Solution-Stability, Pore Size, And Rheological Characterization. These experiments were performed essentially as described previously11 with the following modifications: (1) buffer B was used for hydrogel synthesis, hydrogel solution-stability, and molecular diffusivity studies; (2) the amount of protein present in the suspension buffer used for the solution-stability study was determined by absorbance at 280 nm using a NanoDrop 1000 (Thermo Fisher Scientific).

cross-linkers on both termini is able to self-assemble into a network. Mixing of the two polypeptide building blocks, each containing a CutA cross-linker on one terminus and either Tip1 or Tip1lig on the other terminus, reconstitutes a polypeptide unit containing CutA cross-linker on both termini via the interaction between Tip1 and Tip1lig, triggering a networking cascade and hydrogel formation. This hydrogel is stable at both acidic and neutral pHs (6−8) and in a wide range of temperatures (4−50 °C), exhibits high elasticity, quickly recovers elasticity after shear-induced thinning, and is compatible with tissue culture growth medium. The high solution-stability of the hydrogel is attributed to (1) the introduction of a disulfide bond between Tip1 and Tip1lig which reinforces the linkage between cross-linkers on the selfassembling polypeptide unit and (2) the use of the extremely stable CutA trimer as the cross-linking agent. Incorporation of a docking station peptide (DSP) enables functionalization of the Tip1-mediated hydrogel with docking protein (DP)-tagged target proteins. We demonstrate the application of a Tip1mediated hydrogel functionalized with a small laccase from Streptomyces coelicolor (Slac) in a direct electron transfer enzymatic biocathode, using carbon nanotubes for electron conduction.

■

EXPERIMENTAL SECTION

Chemicals and Bacterial Strains. Ultrapure O 2 and N 2 (>99.99%) were purchased from Acetylene Oxygen Co. (Harlingen, TX). Carboxylated multiwalled carbon nanotubes (cMWNT, outer diameter 10−20 nm, length 10−30 μm, purity >95%) were purchased from CheapTubes.com (Brattleboro, VT). All other chemicals were purchased from Sigma-Aldrich (St. Louis, MO), VWR International (Radnor, PA), or Thermo Fisher Scientific (Waltham, MA) unless otherwise stated. Escherichia coli strains DH5α (Invitrogen, Carlsbad, CA) and BL21(DE3) (New England Biolabs, Ipswich, MA) were used for recombinant DNA cloning and recombinant protein expression, respectively. Plasmid Construction. A schematic depiction of the amino acid sequences of all protein constructs and their numberings is shown in Table 1. Details on the construction of each plasmid can be found in the Supporting Information. Note that, in this paper, the engineered disulfide-forming protein and its ligand, Tip1T58C and Tip1ligD778C, are referred to as dsTip1 and dsTip1lig, respectively. Protein Expression and Purification. Escherichia coli BL21(DE3) was transformed with expression plasmid and plated on a

Table 1. Protein Constructs Used in This Study construct

a

short name

1 2

CutA-Tip16 CutA-Tip1lig

3

CutA-dsTip1

4

CutA-dsTip1lig

5 6 7 8

SH3-Slac SH3-GFP dsTip1 GFP-dsTip1lig

protein sequencea M-(H)7-M-CutA-Tip1 MG(S)2-(H)6-(S)2GLVPRGSHCutA-EAYRDPMG-[(AG)3PEG]10ARMPYVGS-SH3lig-[(G)4S]2-ASTip1lig CutA-EAC-[(G)4S]2-AS-Tip1T58CLE-(H)6 MG(S)2-(H)6-(S)2GLVPRGSHCutA-EAYRDPMG-[(AG)3PEG]10ARMPYVGS-SH3lig-[(G)4S]2-ASTip1ligD778C SH3-KL-[(G)4S]2-AS-Slac-LE-(H)6 SH3-KL-[(G)4S]2-AS-GFP-LE-(H)6 Tip1T58C-LE-(H)6 MG(S)2-(H)6-(S)2GLVPRGSHGFP-[(G)4S]2-AS-Tip1ligD778C

MW (kDa) 24.8 26.1

25.9 26.1

45.9 35.5 12.6 28.8

Hydrogel functional motifs are in bold. 2910

dx.doi.org/10.1021/bm400814u | Biomacromolecules 2013, 14, 2909−2916

Biomacromolecules

Article

Rheological characterization was carried out on a Paar-Physica MCR-300 (Anton Paar, Ashland, VA) parallel plate rheometer with a 25 mm plate fixture (PP25). Purified CutA-dsTip1 and CutA-dsTip1 were manually mixed on top of the rheometer plate by pipetting up and down. The final mixture contained 1.9 mM of each protein in 200 μL of buffer B. The measuring gap was set at 0.2 mm. The rheometer plate fixture was positioned in a humid chamber, and mineral oil was added to the outer edge of the fixture to minimize evaporation. The sample on the plate fixture was incubated at room temperature overnight using settings of 1 rad/s and 1% strain. The next day, strain sweeps and frequency sweeps were performed at 10 rad/s frequency and 10% strain amplitude, respectively. The large-amplitude oscillatory shear (LAOS) experiment was performed at 10 rad/s frequency. Compatibility of Hydrogel with Tissue Culture Growth Medium. A dsTip1 hydrogel (40 μL) containing 1.9 mM of CutAdsTip1 (construct 3) and CutA-dsTip1lig (construct 4) was immersed in 400 μL of tissue culture growth medium Dulbecco’s modified eagle medium (DMEM) supplemented with 0.1% NaN3 for 24 h at 22 °C. The integrity of the hydrogel was monitored by visual inspection. Protein Immobilization in Hydrogel. The enzymatic assay of purified SH3-Slac was carried out in buffer C (50 mM NaPOi, pH 6.0, 0.1% NaN3) containing 1 mM 2,6-dimethoxyphenol (DMP) as the substrate.13 Product formation was monitored by measuring the absorbance at 477 nm every minute for the initial 40 min in a SpectraMax 340PC384 plate reader (Molecular Devices, Sunnyvale, CA). For protein immobilization, SH3-Slac (construct 5) or SH3-GFP (construct 6) was mixed with the hydrogel building block CutAdsTip1lig (construct 4, 1.3 mM in buffer B) at the desired molar ratio. CutA-dsTip1lig contains the DSP SH3lig (Table 1). Purified CutAdsTip1 (construct 3, 1.3 mM in buffer B) was then added to the mixture to induce hydrogel formation. The mixture was manually mixed using a pipet tip and incubated at room temperature for 14−20 h for hydrogel formation. To study the leaching kinetics of the immobilized enzymes, this hydrogel (40 μL) was immersed in 400 μL of buffer B. At the specified times, the buffer was replaced with fresh buffer B. The amount of SH3-Slac or SH3-GFP present in the buffer was quantified by measuring the Slac enzymatic activity or the GFP fluorescence intensity with excitation/emission wavelengths of 485/ 538 nm in a SpectraMax 340PC384 plate reader. Protein quantities were normalized to the activity of the same proteins incubated at the same conditions as the hydrogels. Electrochemical Measurements. CutA-dsTip1, CutA-dsTip1lig, and SH3-Slac were each mixed with carboxylated multiwall carbon nanotubes (cMWNT) to achieve a final cMWNT concentration of 140 mg/mL. Enzyme loading and hydrogel synthesis were carried out as described above on the surface of a pyrolytic graphite working electrode (5 mm, Pine Research Instrumentation, Durham, NC). The hydrogel-loaded electrode was covered with a close fit lid and incubated at room temperature. The next day, the hydrogel was fastened onto the electrode with a cellulose membrane (12−14 kDa, Spectrum Laboratories, Rancho Dominguez, CA) secured by a rubber band. All electrochemical experiments were carried out at room temperature in a conventional three-electrode cell containing 50 mL of buffer C or D (50 mM NaPOi, pH 7.0, 0.1% NaN3). Measurements were performed using a WaveNow USB potentiostat (Pine Research Instrumentation). A pyrolytic graphite electrode and a platinum wire (CH Instruments, Austin, TX) were used as the working and counter electrodes, respectively. Potentials were measured relative to an Ag/ AgCl (saturated KCl) reference electrode (Pine Research Instrumentation). The chronoamperometric response was recorded at 0.2 V versus the standard hydrogel electrode (SHE) with O2 or N2 sparging through the solution.

block copolymers CutA-Tip1 and CutA-Tip1lig (Figure 1, Table 1). CutA, a small trimeric protein (12 kDa) from Pyrococcus

Figure 1. Schematic of hydrogel formation mediated by the interaction between two protein building blocks, each containing the trimeric cross-linker CutA and one member of the protein−ligand interaction pair Tip1/Tip1lig.

horikoshii,12 was chosen as the cross-linking protein due to its extremely high stability. CutA exhibits a denaturation temperature of nearly 150 °C and is able to retain a stable trimeric structure on SDS-PAGE even after exposure to 0.1% SDS and boiling for more than 1 h.14,15 The tax-interacting protein-1 (Tip1) and its peptide ligand (Tip1lig, QLAWFDTDL)16 were used to reconstitute a self-assembling polypeptide unit by connecting CutA cross-linkers located on the two different building blocks. A nonstructural hydrophilic linker, S fragment ([(AlaGly)3ProGluGly]10), was inserted between CutA and Tiplig to facilitate water retention17,18 (Table 1). To enable target protein immobilization, we inserted the small peptide SH3lig (PPPALPPKRRR) between the S fragment and Tip1lig as the DSP. SH3lig exhibits high affinity toward the Src homology 3 domain (SH3) protein19 and thus mediates the immobilization of SH3-tagged target proteins. The proteins CutA-Tip1 and CutA-Tip1lig were purified from E. coli lysate. Mixing of these proteins at concentrations >2 mM leads to the formation of a jellylike material that weakly adheres to the containment vessel after inversion.20 However, this material is completely solubilized within 1 h when suspended in buffer B (data not shown). Previously we successfully prepared a stable protein hydrogel using CutA as the cross-linker,11 and in this stable hydrogel different cross-linkers were covalently linked. Thus, the low solution-stability of the CutA-Tip1/CutATip1lig material was attributed to the weak affinity between the wild-type Tip1 and Tip1lig (Kd ∼ 0.19 μM),16 making the material prone to intermolecular domain swapping.21 Rational Design of a Disulfide Bond between Tip1 and Tip1lig. Disulfide bonds can conditionally form between cysteine residues on proteins under nonreducing conditions, providing a convenient way to seal a linkage between a cargo protein and carrier protein. Previously, Miyagi et al. reversibly conjugated a δPKC inhibitor peptide to a cell penetration peptide TAT via a disulfide linkage to enable the delivery of δPKC into mice.22 Upon delivery to the reducing environment of the cytosol, the disulfide linkage breaks, releasing free δPKC. To increase the solution-stability of the self-assembled CutATip1/CutA-Tip1lig material, we sought to convert the non-

■

RESULTS AND DISCUSSION Self-Assembling Protein Hydrogel Design. The initial building blocks (before the incorporation of disulfide-forming functionality) of our self-assembling hydrogel are two protein 2911

dx.doi.org/10.1021/bm400814u | Biomacromolecules 2013, 14, 2909−2916

Biomacromolecules

Article

pretation of SDS-PAGE results (see Figure 3A vs Figure S1 in the Supporting Information). We anticipated that, as DTT evaporates over time, intermolecular disulfide bonds between dsTip1 and dsTip1lig should form gradually. Samples taken at different times were boiled in sample buffer either containing or lacking the reducing agent β-mercaptoethanol and analyzed via SDS-PAGE (Figure 3). A band corresponding to the combined molecular weight of dsTip1 and GFP-dsTip1lig (41.4 kDa) appeared only on the gel loaded with samples prepared under nonreducing conditions. The emergence of this band coincided with the disappearance of the monomers GFP-dsTip1lig and dsTip1 on the same gel, suggesting that this band corresponds to the disulfide-linked complex of GFP-dsTip1lig and dsTip1. Additional bands of higher molecular weight also appeared on the nonreducing gel, likely indicating homodimers and/or multimers mediated by intercysteine disulfide linkage. No new bands or changes in band intensity were observed in the reducing gel. These results confirm that the engineered dsTip1 covalently attaches to its ligand dsTip1lig via a disulfide bond. Hydrogel Stability Characterization. The wild-type Tip1 and Tip1lig in the hydrogel building block were replaced with dsTip1 and dsTip1lig to form CutA-dsTip1 and CutA-dsTip1lig (constructs 3 and 4, Table 1). Both CutA-dsTip1 and CutAdsTip1lig can be conveniently purified by one-step immobilized ion metal affinity chromatography (IMAC). Most of the purified CutA-dsTip1 and CutA-dsTip1lig exist as higher molecular weight multimers (Figure 4A), with only a small fraction existing as monomer. This result is consistent with the observation that CutA retains trimeric structure during SDSPAGE analysis even after boiling.12 Unlike CutA-dsTip1, which shows a single trimer band in the SDS-PAGE analysis, multiple high molecular weight bands coexist on the gel for CutAdsTip1lig. The appearance of these multiple bands could be due to the presence of the nonstructured S-fragment, which can cause associated proteins to migrate variably on SDS-PAGE gels (data not shown). For hydrogel formation, purified CutA-dsTip1 and CutAdsTip1lig were manually mixed at a 1:1 molar ratio (1.3 mM each) using a pipet tip in buffer B (500 mM NaCl, 10 mM TrisHCl, pH 8.0, 0.1% w/v NaN3) either lacking or containing the reducing agent DTT. The vials were subsequently sealed. A highly viscous material immediately formed under both conditions. Both materials appeared to only weakly adhere to

covalent Tip1−Tip1lig interaction to a covalent disulfide bond through the introduction of a pair of cysteine substitutions, one in the Tip1 binding pocket and one in Tip1lig. The crystal structure of the Tip1 protein in complex with the Tip1lig, (PDB 3IDW)16 reveals that the Tip1lig interacts with Tip1 through a combination of hydrophobic interactions and hydrogen bonds. Particularly, the oxygen (OD2) of D778 is within hydrogen bonding distance of the hydrogen (HG1) of T58 (Figure 2A).

Figure 2. Crystal structure of Tip1-Tip1lig (PDB code, 3IDW). (A) Representation of wild-type Tip1−Tip1lig interaction. Tip1 is shown in purple surface mode. Residues in Tip1 that are within 5 Å of Tip1lig (green) and are shown in pink color. Tiplig is show in green licorice mode. Residues T58 and D778 are shown in element colors. (B) Representation of dsTip1 and dsTip1lig interaction containing the substitutions T58C and D778C.

We hypothesized that substituting both these residues with cysteines will facilitate disulfide bond formation (Figure 2B) between the protein−peptide ligand pair. Tip1 (T58C) and Tip1lig (D778C) are referred as dsTip1 and dsTip1lig, respectively, in this paper. To confirm the formation of disulfide bonds in vitro, we fused dsTip1lig to the C-terminus of GFP to form GFP-dsTip1lig (construct 8, Table 1). dsTip1 was expressed by itself with a 6× histidine tag at the C-terminus (construct 7). dsTip1 and GFPdsTip1lig were mixed in the presence of dithiothreitol (DTT, 2 mM) in an open container. The incorporation of reducing agent in the beginning is to minimize disulfide-mediated homodimeric/multimeric species, thus facilitating the inter-

Figure 3. dsTip1 forms a disulfide bond with dsTip1lig. A mixture containing dsTip1 (12.6 kDa), GFP-dsTip1lig (28.8 kDa) and 2 mM DTT were analyzed at different times after initial mixing using SDS-PAGE conducted under nonreducing (A) and reducing (B) conditions. The DTT in the original mixture evaporates over time, simulating oxidizing conditions. The black triangle indicates the band corresponding to the disulfide-mediated complex of dsTip1:GFP-dsTip1lig (41.4 kDa). A band close to twice the molecular weight of GFP-dsTip1lig is also observed (57.6 kDa). Higher molecular weight bands observed on nonreducing gel were not identified. 2912

dx.doi.org/10.1021/bm400814u | Biomacromolecules 2013, 14, 2909−2916

Biomacromolecules

Article

use of the hyper-thermostable CutA trimer as the cross-linker and the use of a covalent (disulfide) linkage to bridge the two hydrogel building blocks. To our knowledge, this dsTip1 hydrogel is one of the most stable artificial protein hydrogels engineered to date.7,21 The stability of the dsTip1 hydrogel is not affected by acidic pH but is significantly weakened by basic pH (Figure 5B). This phenomenon is likely caused by hydrolysis of disulfide bonds under alkaline conditions,23,24 further underscoring the contribution of disulfide bonds to hydrogel stability. A dsTip1 hydrogel with as low as 1.3 mM of each building block (6.7% w/v total protein) displayed a slightly increased rate of erosion compared to the same hydrogel with 2.2 mM of each building block (11.3% w/v total protein) (Figure 5C). Hydrogels with even lower building block concentrations could potentially be prepared but were not tested in this study. Rheological Characterization of the dsTip1 Hydrogel. The plateau storage modulus (G′∞) for a dsTip1 hydrogel containing 1.9 mM of each building block is 262 ± 54 Pa, 15fold greater than the plateau loss modulus (G″∞) for the same gel (17 ± 1 Pa) (Figure 6A), consistent with gel-like materials.

Figure 4. Characterization of 1.3 mM dsTip1 hydrogel. (A) SDSPAGE analysis (12% acrylamide) of purified CutA-dsTip1 and CutAdsTip1lig. “+” denotes unidentified proteins. (B) Formation of jellylike material in the presence and absence of DTT. The material formed in the absence of DTT appeared to more strongly adhere to the bottom of glass vial after inversion.

the bottom of the glass vial upon inversion, analogous to the observation with the hydrogel composed of wild-type Tip1 and Tip1lig (data not shown). These hydrogels were then incubated at room temperature for 20 h, and the vial test was repeated. At the later time point, the material formed in the absence of DTT appeared to more firmly adhere to the bottom of the vial upon inversion, while the material formed in the presence of DTT still appeared relatively fluid (Figure 4B). This result indicates that the hydrogel formed in the absence of DTT grew stronger over time likely due to the increased formation of disulfide bonds between dsTip1 and dsTip1lig (Figure S2 in the Supporting Information). For convenience, we refer to the material formed in the absence of DTT as dsTip1 hydrogel. Unlike the material containing wild-type Tip1 and Tip1lig, which is completely solubilized within 1 h after immersion in buffer, the dsTip1 hydrogel is highly stable. After being immersed in buffer for 1 month at temperatures up to 50 °C, only 31% of the total protein was lost by erosion (Figure 5A). The high solution-stability of the hydrogel is attributed to the

Figure 6. Rheological characterization of a 1.9 mM dsTip1 hydrogel: (A) angular frequency sweep at 10% strain, (B) strain sweep at 10 rad/ s, and (C) large-amplitude oscillatory shear cycles at 10 rad/s.

The G′∞ value reflects the cross-linker density and strength. Hydrogels with a higher G′∞ value can potentially be obtained by increasing the building block concentration and/or using cross-linker proteins with a higher order of multimerization. To evaluate the ability of dsTip1 hydrogel (1.9 mM) to reform after shear-induced thinning, we conducted a largeamplitude oscillatory shear (LAOS) experiment. Although a permanent loss of storages modulus G′ (∼15%) occurred after the first cycle, no further decrease in G′ was observed in subsequent LAOS cycles (Figure 6B). Hydrogel thinning is attributed to the dissociation of CutA trimer into monomers under shear stress. Upon cessation of stress, disengaged monomers can rapidly reassociate to form new trimeric crosslinkers, restoring the hydrogel to full strength. The ability of the dsTip1 hydrogel to regain mechanical strength points to its suitability for applications that require injection, including controlled drug delivery and tissue engineering.10,25 Characterization of Molecular Diffusion between dsTip1 Hydrogel and Surroundings and Hydrogel

Figure 5. Solution stability of dsTip1 hydrogel: (A) erosion profiles of hydrogels (1.9 mM) at different temperatures in pH 8 buffer, (B) erosion profiles of hydrogels (1.9 mM) at different pHs at 22 °C, and (C) erosion profiles of hydrogels containing different protein concentrations in pH 8 buffer at 22 °C. 2913

dx.doi.org/10.1021/bm400814u | Biomacromolecules 2013, 14, 2909−2916

Biomacromolecules

Article

Compatibility with Tissue Culture Growth Medium. The hydrogel pore size directly affects the rate of mass transfer between the hydrogel interior and the surroundings and is therefore a critical consideration for applications such as tissue engineering and biocatalysis. Fluorescent molecules of 20 kDa (fluorescein isothiocyanate (FITC)-labeled dextran) and 524 Da (pyranine) were incorporated into a 1.9 mM dsTip1 hydrogel (40 μL), and the rate of molecular release from the hydrogel was compared to that from a 1% agarose gel. As shown in Figure 7A, the diffusion profiles of 20 kDa dextran

engineering from compatibility and molecular diffusivity standpoints. dsTip1 Hydrogel as a Scaffold for the Immobilization of Globular Proteins. To determine the suitability of the dsTip1 hydrogel for protein immobilization, we fused the docking protein (DP) SH3 to green fluorescent protein (GFP) and a small laccase from Streptomyces coelicolor (Slac)13 to form SH3-GFP and SH3-Slac, respectively (Table 1). Purified CutAdsTip1lig,which contains a SH3lig motif, was first mixed with SH3-GFP or SH3-Slac at 1:1 or 3:1 molar ratio, and the mixture was combined with CutA-dsTip1 at a 1:1 CutAdsTip1lig /CutA-dsTip1 molar ratio to induce hydrogel formation as described above (Figure 8A). The loading of

Figure 7. dsTip1 hydrogel (1.9 mM) supports the diffusion of small molecules. Diffusion profiles of FITC-dextran (20 kDa) (A) and pyranine (524 Da) (B) from the hydrogel. The hydrogel was suspended in buffer B at 22 °C. 1% agarose gel was used as a control. The inset in part B shows the hydrogel containing pyranine under UV exposure at t = 0, 1, 3, and 5 days. (C) The growth medium DMEM supplemented with 0.1% NaN3 did not disrupt the dsTip1 hydrogel and supported the molecular influx of phenol red at 22 °C.

Figure 8. dsTip1 hydrogel as a scaffold for the immobilization of globular proteins. (A) Schematic depicting the immobilization of globular proteins onto dsTip1 hydrogel. SH3lig-containing CutAdsTip1lig is first mixed with SH3-tagged target protein. CutA-dsTip1 building block is then added to the mixture triggering the formation of a hydrogel with immobilized target protein. A 1.3 mM hydrogel containing SH3-Slac is shown at the bottom right. The blue color of the hydrogel indicates an active enzyme bound to Cu2+. (B) Erosion profile of total proteins and (C) leaching profile of SH3-GFP and SH3-Slac from a 1.3 mM hydrogel. The erosion and leaching studies were carried out in buffer B at 22 °C. CTL: CutA-dsTip1lig.

from a dsTip1 hydrogel and a 1% agarose gel are comparable with over 80% of the entrapped dextran diffusing out of the hydrogel in 3 days. On the other hand, pyranine diffuses more rapidly from a 1% agarose gel than from a dsTip1 hydrogel. Over 90% of the pyranine diffused from the hydrogel after 12 h, while only 56% diffused from the dsTip1 hydrogel in the same period (Figure 7B). The reason for the difference in diffusivity for large and small molecules in the dsTip1 hydrogel versus 1% agarose is at present unclear, but this phenomenon may derive from the dynamic nature of the protein hydrogel caused by the presence of the unstructured S fragment.21,26 We next determined the compatibility of tissue culture growth medium with the dsTip1 hydrogel. A 40 μL dsTip1 hydrogel (1.9 mM) was immersed in Dulbecco’s Modified Eagle Medium (DMEM) at room temperature. As shown in Figure 7C, the growth medium does not physically disrupt the dsTip1 hydrogel, as determined by visual inspection. Furthermore, it is evident from the reddening of the ∼5 mm thick dsTip1 hydrogel over time that the small molecule phenol red (354 Da) readily diffuses into the hydrogel. Given that most tissue engineering applications employ thinner slices of hydrogel for tissue implants,27 this result supports the application of the dsTip1 hydrogel as a scaffold for tissue

target protein onto liquid-phase building blocks prior to hydrogel formation facilitates a relatively even distribution of the target enzyme throughout the hydrogel. Incorporation of SH3-GFP does not affect the hydrogel stability (compare the dotted line with diamonds in Figure 8B). Incorporation of SH3-Slac, on the other hand, slightly improves the hydrogel stability in a dose-dependent manner with 26% and 19% total protein loss from hydrogels harboring 3:1 and 1:1 CutA-dsTip1lig/SH3-Slac molar ratios, respectively (compare dotted line with triangles in Figure 8B). Unlike GFP, a monomer, Slac is a trimeric protein.28 Thus, Slac proteins have the potential to function as secondary cross-linkers in the hydrogel, increasing hydrogel stability. The percentages of SH3-GFP and SH3-Slac leached from the hydrogel for 1:1 CutA-dsTip1lig/SH3-protein molar ratios are 32% and 9%, respectively, and are both ∼15% for 3:1 CutAdsTip1lig:SH3-protein hydrogels (Figure 8C). Interestingly, the percentage of SH3-Slac leached from the hydrogel is less than the percentage of total protein eroded from the hydrogel over 2914

dx.doi.org/10.1021/bm400814u | Biomacromolecules 2013, 14, 2909−2916

Biomacromolecules

Article

a week with the electrode immersed in buffer. The OCP values only varied slightly between 0.47 and 0.54 V versus SHE over a week (Figure 9B). This measured OCP corresponds to the redox potential of the T1 copper site of Slac (∼0.5 V).13,33 The catalytic current density was calculated by subtracting the background current (N2 saturated) from the corresponding plateau current (O2 saturated) (Figure 9A). Interestingly, we consistently observed a slight increase in catalytic current in the initial 2−3 days (from −39.5 to −51.8 μA/cm2), followed by a decrease in the current density to a stable value of −36.5 μA/ cm2 (Figure 9B). This initial increase in catalytic current is likely a consequence of hydrogel swelling following immersion in buffer, which causes an expansion in the hydrogel pore size and facilitates the influx of the substrate O2. The subsequent decrease in current density may derive from loss of the immobilized Slac enzyme due to leaching. These results support the use of the Slac-functionalized dsTip1 hydrogel for stable electrical wiring of immobilized enzymes. The catalytic current generated in our system is comparable to other nonrotating direct electron transfer (DET) enzymatic biocathodes in oxygen-saturated solutions.36,37 Modifications that may further increase the current density include the use of modified MWNT with higher affinity toward the enzyme active site to properly orient the enzyme for more efficient electron transfer,38,39 enhancing the influx of the substrate O2 (e.g., by expanding the hydrogel pore size through incorporation of elongated rigid structural proteins into hydrogel building blocks or increasing the hydrogel surface area by micromolding) and the use of more active laccases.34

the same period of time (compare parts B and C of Figure 8), suggesting that trimeric Slac may be preferentially retained in the hydrogel matrix by virtue of its cross-linking functionality, relative to more loosely incorporated hydrogel elements (e.g., homo/heteromultimers of hydrogel building blocks). Although SDS-PAGE analysis of proteins present in the hydrogel suspension buffer was attempted to determine the identity of the eroded proteins in Figure 8B, the band pattern was too cluttered to make a clear interpretation of band identities (data not shown). It is conceivable that multiple different DSP/DP pairs can be incorporated into the hydrogel in a predefined sequence, enabling the ordered coimmobilization of multiple enzymes to achieve substrate channeling between different enzymes.29−31 Application of a Slac-Functionalized dsTip1 Hydrogel As a Biocathode. Laccase catalyzes the electroreduction of O2 to water at near-neutral pH and ambient temperature and thus has been widely employed as a cathode enzyme in biofuel cells.32−34 In this study, we determined the ability of Slacfunctionalized dsTip1 hydrogel to support current generation. To render the hydrogel conductive, cMWNT were incorporated into the hydrogel. We chose cMWNT because these can be easily suspended in aqueous solution due to the presence of their hydrophilic carboxyl groups.35 Incorporation of cMWNT did not compromise the integrity of the dsTip1 hydrogel or increase the leaching rate of the immobilized Slac (Figure S3 in the Supporting Information). To measure the current generated by Slac, a 1:1 SH3-Slac/dsTip1lig molar ratio hydrogel (1.0 mM) containing ∼140 mg/mL cMWNT was loaded onto a pyrolytic graphite electrode and the chronoamperometric response was monitored at 25 °C. At pH 6.0, a cathodic current maximized at −65.5 μA/cm2 was generated in the presence of O2 sparging (Figure 9A). A much smaller

■

CONCLUSION We demonstrated the synthesis of a disulfide-reinforced protein hydrogel customizable with globular proteins of interest that self-assembles upon mixing of two liquid-phase building blocks. Specifically, mixing of the two components reconstitutes a single self-assembling polypeptide unit via a Tip1 protein− ligand interaction that is reinforced by an engineered disulfide linkage. The use of a universal docking-station-peptide/ docking-protein (DSP/DP) pair enables convenient incorporation of globular proteins in a “plug and play” fashion without changing the molecular architecture of the hydrogel backbone. This hydrogel retains its integrity under a wide variety of conditions and does not require extraneous chemicals or alterations in pH/temperature/ionic strength to induce the gelation process, making this hydrogel especially suitable for tissue engineering as live cells can be conveniently encapsulated without any environmental triggers.40 When functionalized with the enzyme laccase, the disulfide-reinforced hydrogel serves as an effective bioelectrode for current generation. The hydrogel can easily be functionalized with other proteins depending on the application. For example, growth hormones can potentially be incorporated to generate bioactive scaffolds for tissue engineering, and enzymes can be incorporated to create solid-phase biocatalysts for industrial biosynthesis. With the use of multiple orthogonal DSP/DP pairs, it is conceivable that an entire metabolic pathway can be synthesized on the hydrogel for in vitro pathway construction.29,41 Finally, we anticipate that the physical properties of the hydrogel can be further tailored through the employment of linkers with varying structural rigidities and cross-linker proteins with different multimeric states,21 paving the way for the development of even more sophisticated hydrogels.

Figure 9. Characterization of dsTip1 hydrogel biocathode. (A) Chronoamperometric response (0.2 V vs SHE) for biocathodes either lacking SH3-Slac or with SH3-Slac loaded at a 1:1 SH3-Slac/dsTip1lig molar ratio at pH 6.0 and pH 7.0 in the presence of sparging with N2 or O2. (B) Time evolution of open circuit potential (OCP) and catalytic current density at 0.2 V (vs SHE) of the biocathode with SH3-Slac at pH 6.0. The OCP of the biocathode without SH3-Slac is also indicated. Experiments were carried out in either buffer C (50 mM NaPOi, pH 6.0, 0.1% NaN3) or D (50 mM NaPOi, pH 7.0, 0.1% NaN3).

current (maximal −23.7 μA/cm2) was observed at pH 7.0 (Figure 9A), consistent with previously reported data indicating a pH optimum of 6.0 for Slac activity.33 Hydrogel lacking SH3Slac did not generate any current. These results demonstrate that enzyme-functionalized dsTip1 hydrogel has the potential to be used for current generation in biocatalytic biofuel cells. To evaluate the ability of the Slac-functionalized dsTip1 hydrogel to retain biocathodic function over time, the cathodic current and open circuit potential (OCP) were monitored over 2915

dx.doi.org/10.1021/bm400814u | Biomacromolecules 2013, 14, 2909−2916

Biomacromolecules

■

Article

(16) Zhang, J.; Yan, X.; Shi, C.; Yang, X.; Guo, Y.; Tian, C.; Long, J.; Shen, Y. J. Mol. Biol. 2008, 384, 255−263. (17) Fournier, M. J.; Creel, H. S.; Krejchi, M. T.; Mason, T. L.; Tirrell, D. A.; Mcgrath, K. P.; Atkins, E. D. T. J. Bioact. Compat. Polym. 1991, 6, 326−338. (18) Petka, W. A.; Harden, J. L.; McGrath, K. P.; Wirtz, D.; Tirrell, D. A. Science 1998, 281, 389−392. (19) Wu, X.; Knudsen, B.; Feller, S. M.; Zheng, J.; Sali, A.; Cowburn, D.; Hanafusa, H.; Kuriyan, J. Structure 1995, 3, 215−226. (20) Lu, H. D.; Wheeldon, I. R.; Banta, S. Protein Eng., Des. Sel. 2010, 23, 559−566. (21) Shen, W.; Zhang, K.; Kornfield, J. A.; Tirrell, D. A. Nat. Mater. 2006, 5, 153−158. (22) Miyaji, Y.; Walter, S.; Chen, L.; Kurihara, A.; Ishizuka, T.; Saito, M.; Kawai, K.; Okazaki, O. Drug Metab. Dispos. 2011, 39, 1946−1953. (23) Andersson, L. O.; Berg, G. Biochim. Biophys. Acta 1969, 192, 534−536. (24) Du, C.; Martin, P. A.; Nickerson, K. W. Appl. Environ. Microbiol. 1994, 60, 3847−3853. (25) Yu, L.; Ding, J. Chem. Soc. Rev. 2008, 37, 1473−1481. (26) Shen, W. Structure, dynamics, and properties of artificial protein hydrogels assembled through coiled-coil domains. Ph.D. Thesis, California Institute of Technology, Pasadena, CA, 2005. (27) Annabi, N.; Nichol, J. W.; Zhong, X.; Ji, C.; Koshy, S.; Khademhosseini, A.; Dehghani, F. Tissue Eng., Part B: Rev. 2010, 16, 371−383. (28) Skalova, T.; Dohnalek, J.; Ostergaard, L. H.; Ostergaard, P. R.; Kolenko, P.; Duskova, J.; Stepankova, A.; Hasek, J. J. Mol. Biol. 2009, 385, 1165−1178. (29) You, C.; Myung, S.; Zhang, Y. H. Angew. Chem., Int. Ed. 2012, 51, 8787−8790. (30) Dueber, J. E.; Wu, G. C.; Malmirchegini, G. R.; Moon, T. S.; Petzold, C. J.; Ullal, A. V.; Prather, K. L.; Keasling, J. D. Nat. Biotechnol. 2009, 27, 753−759. (31) Conrado, R. J.; Varner, J. D.; DeLisa, M. P. Curr. Opin. Biotechnol. 2008, 19, 492−499. (32) Zebda, A.; Gondran, C.; Le Goff, A.; Holzinger, M.; Cinquin, P.; Cosnier, S. Nat. Commun. 2011, 2, 370. (33) Gallaway, J.; Wheeldon, I.; Rincon, R.; Atanassov, P.; Banta, S.; Barton, S. C. Biosens. Bioelectron. 2008, 23, 1229−1235. (34) Opallo, M. Biocathodes for Dioxygen Reduction in Biofuel Cells. In Fuel Cell Science: Theory, Fundamentals, and Biocatalysis; Wiechowski, A., Nørskov, J. K., Eds.; John Wiley & Sons, Inc.: Hoboken, NJ, 2010; pp 169−214. (35) Yu, C. M.; Yen, M. J.; Chen, L. C. Biosens. Bioelectron. 2010, 25, 2515−2521. (36) Ramasamy, R. P.; Luckarift, H. R.; Ivnitski, D. M.; Atanassov, P. B.; Johnson, G. R. Chem. Commun. (Cambridge, U. K.) 2010, 46, 6045−6047. (37) Rubenwolf, S.; Strohmeier, O.; Kloke, A.; Kerzenmacher, S.; Zengerle, R.; von Stetten, F. Biosens. Bioelectron. 2010, 26, 841−845. (38) Meredith, M. T.; Minson, M.; Hickey, D.; Artyushkova, K.; Glatzhofer, D. T.; Minteer, S. D. ACS Catal. 2011, 1, 1683−1690. (39) Minson, M.; Meredith, M. T.; Shrier, A.; Giroud, F.; Hickey, D.; Glatzhofer, D. T.; Minteer, S. D. J. Electrochem. Soc. 2012, 159, G166− G170. (40) Wong Po Foo, C. T.; Lee, J. S.; Mulyasasmita, W.; Parisi-Amon, A.; Heilshorn, S. C. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 22067− 22072. (41) Zhang, Y. H.; Evans, B. R.; Mielenz, J. R.; Hopkins, R. C.; Adams, M. W. PLoS ONE 2007, 2, e456.

ASSOCIATED CONTENT

S Supporting Information *

A description of the process for the plasmid constructs used in this study and the primers used. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Mailing address: 3122 TAMU, Artie McFerrin Department of Chemical Engineering, Texas A&M University, College Station, TX 77843. Phone: 979-862-1610. Fax: 979-845-6446. E-mail: [email protected]. Author Contributions ⊥

Contributed equally. All authors have given approval to the final version of the manuscript.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Funding for this work was provided by the Norman Hackerman Advanced Research Program, National Science Foundation and Air Force Office of Scientific Research. The authors would like to thank Dr. David Tirrell (California Institute of Technology), Dr. Takehisa Matsuda (Kanazawa Institute of Technology, Japan), and Prof. Gerard Canters (Leiden University, Netherlands) for their kind gifts of the plasmid pQE9AC10Atrp,5 pET30-CutA-Tip1,6 and pSLAC1,13 respectively. We also thank Dr. Victor Ugaz (Texas A&M University) for his assistance with rheological characterization and Dr. Karuppiah Chockalingam for critical reading and editing of the manuscript.

■

REFERENCES

(1) Maity, G. C. J. Phys. Sci. 2008, 12, 173−186. (2) Veronese, F. M.; Pasut, G. Drug Discovery Today 2005, 10, 1451− 1458. (3) Nguyen, M. K.; Lee, D. S. Macromol. Biosci. 2010, 10, 563−579. (4) Jonker, A. M.; Lowik, D. W. P. M.; van Hest, J. C. M. Chem. Mater. 2012, 24, 759−773. (5) Shen, W.; Lammertink, R. G. H.; Sakata, J. K.; Kornfield, J. A.; Tirrell, D. A. Macromolecules 2005, 38, 3909−3916. (6) Ito, F.; Usui, K.; Kawahara, D.; Suenaga, A.; Maki, T.; Kidoaki, S.; Suzuki, H.; Taiji, M.; Itoh, M.; Hayashizaki, Y.; Matsuda, T. Biomaterials 2010, 31, 58−66. (7) Wheeldon, I. R.; Gallaway, J. W.; Barton, S. C.; Banta, S. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 15275−15280. (8) Byun, J. Y.; Lee, K. H.; Lee, K. Y.; Kim, M. G.; Kim, D. M. Lab Chip 2013, 13, 886−891. (9) Vazquez-Duhalt, R.; Tinoco, R.; D’Antonio, P.; Topoleski, L. D.; Payne, G. F. Bioconjugate Chem. 2001, 12, 301−306. (10) Olsen, B. D.; Kornfield, J. A.; Tirrell, D. A. Macromolecules 2010, 43, 9094−9099. (11) Ramirez, M.; Guan, D.; Ugaz, V.; Chen, Z. J. Am. Chem. Soc. 2013, 135, 5290−5293. (12) Tanaka, Y.; Tsumoto, K.; Nakanishi, T.; Yasutake, Y.; Sakai, N.; Yao, M.; Tanaka, I.; Kumagai, I. FEBS Lett. 2004, 556, 167−174. (13) Machczynski, M. C.; Vijgenboom, E.; Samyn, B.; Canters, G. W. Protein Sci. 2004, 13, 2388−2397. (14) Tanaka, T.; Sawano, M.; Ogasahara, K.; Sakaguchi, Y.; Bagautdinov, B.; Katoh, E.; Kuroishi, C.; Shinkai, A.; Yokoyama, S.; Yutani, K. FEBS Lett. 2006, 580, 4224−4230. (15) Sawano, M.; Yamamoto, H.; Ogasahara, K.; Kidokoro, S.; Katoh, S.; Ohnuma, T.; Katoh, E.; Yokoyama, S.; Yutani, K. Biochemistry 2008, 47, 721−730. 2916

dx.doi.org/10.1021/bm400814u | Biomacromolecules 2013, 14, 2909−2916