Article pubs.acs.org/Langmuir
Construction and Characterization of Protein-Encapsulated Electrospun Fibermats Prepared from a Silica/Poly(γ-glutamate) Hybrid Shuhei Koeda, Kentaro Ichiki, Norihiko Iwanaga, Koji Mizuno, Masahide Shibata, Akiko Obata,* Toshihiro Kasuga, and Toshihisa Mizuno* Graduate School of Engineering, Nagoya Institute of Technology, Gokiso-cho, Showa-ku, Nagoya, Aichi 466-8555, Japan S Supporting Information *
ABSTRACT: Protein-encapsulated fibermats are an attractive platform for proteinbased bioactive materials. However, the choice of methods is still limited and not applicable to a wide range of proteins. In this study, we studied new polymeric materials for constructing protein-encapsulated fibermats, in which protein molecules are encapsulated within the nanofibers of fibermats without causing deleterious changes to protein structure or function. We constructed a protein-encapsulated fibermat using the poly(γ-glutamate) (PGA)/(3-glycidyloxypropyl)-trimethoxysilane (GPTMS) hybrid as a precursor for electrospinning. Because the PGA/GPTMS hybrid is water-soluble, protein molecules can be added to the precursor in an aqueous solution, significantly enhancing protein stability. Polycondensation during electrospinning (in-flight polycondensation) makes the obtained fibermats water-insoluble, which stabilizes the fibermat structure such that it is resistant to degradation in aqueous buffer. The molecular structure of the PGA/ GPTMS hybrid gives rise to unique molecular permeability, which alters the selectivity and specificity of biochemical reactions involving the encapsulated enzymes; lower molecular-weight (MW) substrates can permeate the nanofibers, promoting enzyme activity, but higher MW substrates such as inhibitor peptides cannot permeate the nanofibers, suppressing enzyme activity. We present an effective method of encapsulating bioactive molecules while maintaining their structure and function, increasing the versatility of electrospun fibermats for constructing various bioactive materials.
■
INTRODUCTION A fibermat is a felt-like film in which homogeneous nanofibers generated from organic and/or inorganic polymeric materials are tightly integrated.1 Using the electrospinning method for precursor polymer solutions, the processes of (I) nanofiber formation, (II) solvent evaporation, and (III) integration of nanofibers on a ground collector are sequentially performed to yield electrospun fibermats consisting of micro- or nanofibers with a diameter of 50 nm to 5 μm. Recently, to design bioactive fibermats that could function in cell-incubation,2 drug delivery,3 enzyme-immobilization,4 and cell-immobilization,5−9 protein conjugation has been extensively studied. Structurally, fibermats have a large surface area-to-volume ratio and high solvent accessibility, even inside the film; surface adsorption or covalent attachment of protein molecules onto micro- or nanofibers in fibermats has been thoroughly examined. The encapsulation of protein molecules, especially enzymes, into fibermats has attracted considerable attention because the micro- or nanofibers in fibermats might uniquely influence the encapsulated proteins’ function.4 To fabricate protein-encapsulated fibermats via electrospinning, several critical points must be addressed. Because protein molecules are in the precursor polymer solution before electrospinning, it is important that the polymer does not induce protein denaturation such that protein function is retained; further, as most proteins are only soluble in water © 2015 American Chemical Society
except for membrane proteins, and choosing water as a solvent is indispensable to maintain protein structure and function. Thus, the polymer must also be water-soluble. The fibermat should become water-insoluble after electrospinning such that it is resistant to degradation when subsequently immersed in an aqueous buffer. After immersion in a buffer, the enzymes can function as they would under near-physiological conditions. Post-cross-linking of protein-encapsulated fibermats generated from water-soluble polymers using bifunctional cross-linkers such as glutaraldehyde (GA) has been used to prepare enzymeencapsulated fibermats. Ren et al. reported the encapsulation of glucose oxidase (GO) in the microfibers of a poly vinyl alcohol (PVA)-fibermat.10 After treatment of the GO-encapsulated PVA-fibermat with GA, the GO-encapsulated PVA-fibermat became water-insoluble, and the leaching of GO into an immersing buffer was effectively suppressed. Similarly, Tang et al. reported the encapsulation of α-galactosidase in a PVA/GA fibermat and examined its enzymatic activity.11 However, immersion of the fibermat in an organic solvent such as acetone was necessary to perform hemiacetal cross-linking reactions, which could induce significant denaturation of the encapsulated proteins and limit the types of proteins that could Received: August 3, 2015 Revised: December 17, 2015 Published: December 17, 2015 221
DOI: 10.1021/acs.langmuir.5b02862 Langmuir 2016, 32, 221−229
Article
Langmuir be encapsulated. Patel et al. constructed a fibermat entrapping horseradish peroxidase (HRP) within silica nanofibers (100− 200 nm), which were prepared from an aqueous solution of hydrolyzed orthosilicates including HRP, polyethylene glycol (PEG), and D-glucose. After nanofiber construction by electrospinning, the porous structures within the silica nanofibers remained, facilitating the approach of the substrate to the entrapped HRP, which resulted in moderate enzymatic activity.12 The entrapment of protein molecules in nanofibers causes molecular crowding because of the high concentration of biomolecules within the narrow nanofibers, which could contribute to an increase in resistance to thermal denaturation. Similarly, the enzyme tyrosinase has been encapsulated in a silica nanofiber fibermat.13 In these cases, though, the use of hydrolyzed orthosilicate limited the choice of proteins that were resistant to denaturation when exposed to the harsh conditions present in the construction of protein-encapsulated fibermats. To encapsulate enzyme molecules in the hollow nanostructure in nanofibers, the “emulsion electrospinning method” using the amphiphilic block copolymers has also been examined by Dai et al.,14 Li et al.,15 and Pinto et al.16 Despite advances in techniques to encapsulate protein molecules, the choice of methods is still limited, and previously described methods are not broadly amenable to a wide range of proteins, warranting the development of other polymeric materials to expand the applications of protein-encapsulated fibermats. We have recently reported the preparation and characterization of electrospun fibermats generated from a silica/poly(γ-glutamate) (PGA) hybrid.17 The precursor silica/ PGA hybrid solution was prepared using PGA and 3(glycidyloxypropyl) trimethoxysilane (GPTMS). Specifically, by tethering the epoxy groups in GPTMS to carboxylate groups in PGA, trimethoxysilyl (TMS)-modified PGA was produced. Following polycondensation between silicates, we obtained three-dimensional PGA network structures, aided by silicate junctions, in a viscous aqueous solution. The electrospinning method was applied to this precursor solution to fabricate fibermats containing homogeneous nanofibers, which become water-insoluble enough to analyze protein functions upon immersing in a buffer solution. Fibermats prepared from the silica/PGA hybrids are expected to meet the requirements for constructing proteinencapsulated fibermats. For example, the precursor solution is water-soluble; various biopolymers such as peptides, proteins, and DNA can be included without losing original secondary and tertiary structures. Due to the three-dimensional network structure of the silica/PGA hybrid, the nanofibers could maintain MW-dependence in molecular permeability. Indeed, because enzymes typically have high molecular weights (MW > 10 kDa) they are likely to dissociate slowly from the nanofibers and remain encapsulated, whereas the low MW substrates should quickly penetrate the nanofibers such that enzymatic reactions can occur. Here, we describe the construction of an enzyme-encapsulated fibermat prepared from a silica/PGA hybrid and the function of encapsulated enzymes.
■
cyanate isomer I, chymotrypsin, trypsin-chymotrypsin inhibitor (MW 8000), glucose oxidase (GO) from Aspergillus niger, and horseradish peroxidase (HRP) were purchased from Sigma-Aldrich (USA). The amino PEGs, mPEG-NH2 (MW 1000 and 2000), were purchased from Nanocs, Inc. (USA). The fluorescein isothiocyanate (FITC)modified PEGs (MW 1000 and 2000) were synthesized using fluorescein isothiocyanate isomer I and mPEG-NH2 (MW 1000 and 2000, respectively).18 Thrombin was purchased from Mochida Pharmaceutical Co. Ltd. (Japan). Green fluorescent protein (GFP) was expressed in an Escherichia coli BL21(DE3) strain using an expression vector harboring the superfolder GFP gene,19 and purified using a Ni-affinity chromatography with His-Bind Resin (Novagen, Merck, USA). Unless otherwise stated, all chemicals and reagents were obtained commercially and used without further purification. Preparation of the Protein-Encapsulated Electrospun Fibermats. Poly(γ-glutamic acid) (500 mg) and Ca(OH)2 (128 mg) were dissolved in 2.5 mL of H2O before GPTMS (188 mg) was added, and the solution was stirred at ambient room temperature for 2.5 h. In doing so, the modification of GPTMS to carboxylate groups in poly(γglutamic acid), and the subsequent hydrolysis and polycondensation of alkoxysilyl moieties, progressed to yield the viscous precursor solution. At this stage, lyophilized protein (GFP, chymotrypsin, or thrombin at 15 wt % in the PGA/GPTMS precursor) dissolved in 0.5 mL of H2O, was added to the solution and stirred for 0.5 h to achieve homogeneous mixing. To prepare the GO/HRP-encapsulated fibermat, lyophilized GO (12.5 mg) and HRP (25 mg) were used. The final precursor solution was electrospun to obtain a fibermat with protein molecules encapsulated inside the nanofibers using a Nanofiber Electrospinning Unit (Kato Tech Co, Japan). A high tension electric field of 20 kV was applied to the needle where the tip of the needle was positioned 150 mm from a grounded rotating drum (rotating speed: 200 mm/min) that was used as the fiber collector. The average nanofiber diameter and standard deviation were measured from the SEM images (n = 30). Preparation of Electrospun Fibermats Encapsulating Fluorescent Molecules. The methods used for preparing fluorescent molecules encapsulated in fibermats were identical to those described for encapsulating proteins in electrospun fibermats described above, except that each fluorescent molecule, rather than protein, was dissolved in 0.5 mL of H2O and added to the precursor solution. Surface Morphologies of the Protein- and FITC-PEGEncapsulated Electrospun Fibermats Using a Scanning Electron Microscope (SEM).17 Fibermats were coated with amorphous osmium using plasma chemical vapor deposition (CVD) with vaporized OsO4, and their morphology was observed using fieldemission scanning electron microscopy (SEM) (JSM-6301F, JEOL, Japan). The mean diameter of the nanofibers in fibermats was evaluated using images from SEM. Attenuated-Total Reflection Fourier-Transformed Infrared (ATR-FTIR) spectra of the Silica/PGA fibermats. ATR- FTIR spectra of the silica/PGA fibermats were acquired using a FT-IR-4000 instrument (JASCO, Japan) at ambient temperature. Fluorescence Spectra of the GFP-Encapsulated Fibermats. First, the fibermats encapsulating GFP were wetted in 20 mM Tris HCl (pH 7.8) and adhered to a coverslip (20 × 20 mm, Eagle XG, Corning Inc., USA). The coverslip was then placed into the sample chamber of a FluoroMax-4 spectrofluorometer (HORIBA Scientific, Japan), and the fluorescence emission of the GFP-encapsulated fibermat was acquired with photo excitation at 460 nm. Confocal Laser Scanning Microscopic Observation of Fluorescence Emission from the GFP-Encapsulated Electrospun Fibermat. A confocal laser scanning microscope (LSM880, Zeiss, Germany) was used to observe the fluorescence emission of the GFP-encapsulated electrospun fibermats. Before observation, a fibermat was attached onto a clean glass coverslip. Samples were imaged using a 63× objective lens. Leakage Behaviors of Encapsulated Fluorescent Molecules When Fibermats Were Immersed in a Buffer Solution. The molecule-encapsulated fibermat (1.5 mg) was cut into defined pieces (2 mm × 2 mm), which were immersed in 3 mL of 20 mM HEPESNaOH buffer (pH 7). To monitor leakage over time, fibermat pieces
EXPERIMENTAL SECTION
Materials. Poly(γ-glutamic acid) (Mn 300 000−500 000), Ca(OH)2, tris(hydroxymethyl)aminomethane (Tris), p-nitrophenyl acetate (p-NAc), 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride (AEBSF-HCl), and 4-aminoantipyrine (4-AAP) were purchased from Wako Pure Chemical Ind. Ltd. (Japan). 3(Glycidyloxypropyl) trimethoxysilane (GPTMS), fluorescein isothio222
DOI: 10.1021/acs.langmuir.5b02862 Langmuir 2016, 32, 221−229
Article
Langmuir Scheme 1. Preparation of the Protein-Encapsulated Fibermat Using Electrospinning
were immersed in a buffer for 0−6 h. Once fibermat pieces were removed from the buffer, the A492 for FITC-PEG1000, FITC-PEG2000, and fluorescein and the A370 for 1-anilino-8-naphthalene sulfonic acid (ANS) were monitored to determine the degree of leakage for each encapsulated molecule. To estimate the total amount of encapsulated molecules in 1.5 mg of fibermat, we performed homogeneous solubilization of the molecule-encapsulated fibermats using 1 mL of NaOH(aq) (1 mol/L). The A492 (FITC-PEG1000, FITC-PEG2000, and fluorescein) and A370 (ANS) determined in aqueous solution and after adjustment to pH 7 were used to calculate the percent leakage (%) of encapsulated molecules. Analyses of Chymotrypsin Activity in Electrospun Fibermats Using p-NAc as a Substrate. Chymotrypsin-encapsulated fibermats with various masses (0.375, 0.75, and 2.25 mg) were cut into pieces of a defined size (2 mm × 2 mm). The amount of chymotrypsin in each fibermat was evaluated using the initial ratio of chymotrypsin to the amount of PGA and GPTMS in the precursor solution (69.8, 140, and 419 μg, respectively). The fibermat pieces were immersed in 3 mL of 50 mM Tris HCl (pH 7.8) and gently stirred for 30 min at 37 °C. To initiate the enzymatic hydrolysis reaction of p-NAc, a solution of 20 μL p-NAc (60 mM) in acetonitrile was added such that the final concentration of p-NAc was 0.4 mM in all cases. Enzymatic reactions were performed for 10 min at 37 °C, and then fibermats were removed from the solution and the A400 of the supernatants was monitored. Using ε400 (18384 L mol−1 cm−1)20 as the molar extinction coefficient of p-nitrophenol in an aqueous solution at pH 7.8, we determined the amount of p-nitrophenol produced for each mass of chymotrypsinencapsulated fibermat. We plotted the amount of p-nitrophenol as a function of chymotrypsin and used linear-fitting analysis to determine enzymatic activity (mmol g−1 min−1). As a control experiment, we determined the enzymatic activity of chymotrypsin that was not encapsulated in a fibermat using similar methods. Analyses of Thrombin Activity in Fibermats Using a Substrate That Is a Fusion Protein Consisting of GFP and Trx Connected by a Linker Peptide Containing the Thrombin Digestion Site. The thrombin-encapsulated fibermats (2.7 mg, with 0.5 mg of thrombin included in the fibermat) were cut into pieces (2 mm × 2 mm) and immersed in 300 μL of thrombin cleavage buffer (20 mM Tris HCl, 150 mM NaCl, 2.5 mM CaCl2, pH 8.4) at 25 °C. The GFP and Trx fusion protein, connected by a linker peptide containing the thrombin cleavage site (LeuValProArgGlySer), was
added to this solution and the enzymatic digestion was performed for 1.5 h at 25 °C. Once the fibermat was removed from solution, the supernatant was treated with dithiothreitol (DTT) and then subjected to SDS-PAGE to analyze the resulting protein fragments. In a control experiment, digestion of the fusion protein by thrombin was performed using 0.5 mg of thrombin in 300 μL of thrombin cleavage buffer at 25 °C for 1.5 h. Chymotrypsin Activity in the Fibermat in the Presence of Trypsin−Chymotrypsin Inhibitor. Chymotrypsin-encapsulated fibermats (0.375, 0.75, and 2.25 mg) containing different amounts of chymotrypsin were cut into pieces (2 mm × 2 mm), and the amount of chymotrypsin was evaluated using the method described above. The fibermat pieces were immersed in 3 mL of 50 mM Tris HCl (pH 7.8) containing a trypsin−chymotrypsin inhibitor (0.42 mg), and the solution was gently stirred for 30 min at 37 °C. The enzymatic hydrolysis reaction of p-NAc was initiated and performed using methods identical to those described above, and the supernatants’ A400 and the amount of p-nitrophenol produced were also determined as described above. As a control experiment, a trypsin−chymotrypsin inhibitor was used to inhibit chymotrypsin activity in solution (not in a fibermat). Glucose Detection in the GO/HRP-Encapsulated Fibermat. The GO/HRP-encapsulated fibermat was cut to a defined area (11 × 11 mm) and then mounted onto a coverslip (50 × 70 mm, Matsunami Glass Ind. Ltd., Japan) using hydrophobic tape to form a fibermat window (6 × 6 mm). First, 35 μL of 4-AAP (6.6 mM) and TOOS (6.6 mM) in 100 mM Tris HCl (pH 7.0) was spotted onto the fibermat window. Subsequently, 35 μL of D-glucose (66.6, 116, 196, 500, and 666 mM) in 100 mM Tris HCl (pH 7.0) was spotted onto the fibermat. After 30 min, the A555 values were recorded using a fiberscope-type UV−vis spectrometer (EPP2000-UVN-SR, StellarNet Inc., U.S.A.) equipped with a halogen light source (SL5-CUV, StellarNet Inc., U.S.A.). For detection of glucose in the rabbit blood sample (Nippon Bio-Test Laboratories Inc., Japan), the blood was diluted 5-fold with 100 mM Tris HCl (pH 7.0), and 35 μL of solution was spotted onto the fibermat window. Absorbance was measured as for pure glucose. 223
DOI: 10.1021/acs.langmuir.5b02862 Langmuir 2016, 32, 221−229
Article
Langmuir
■
RESULTS AND DISCUSSION Construction and Characterization of the GFPEncapsulated Silica/PGA Fibermat. Before we characterized fibermats in which enzyme molecules were encapsulated in the nanofibers, we first examined the effect of encapsulation on protein tertiary structure by encapsulating GFP in nanofibers. Several studies have attempted GFP encapsulation in polymeric materials. Brantley et al. constructed a GFP-encapsulated poly(methyl methacrylate) (PMMA) film.21 Although the construction conditions were harsh for biomolecules, as radical polymerization of MMA was performed in the presence of GFP and plasticizers, GFP could be encapsulated in PMMA film without protein denaturation. Encapsulation of GFP inside nanofibers of several types of fibermats was also reported by Canbolat et al.5 and Dror et al.22 Using GFP, it is possible to investigate protein denaturation and mechanical stresses to protein structures by analyzing the spectroscopic properties of GFP.21 ATR-FTIR was used to measure the extent of Si−O−Si bond formation. The GFP-encapsulated silica/PGA fibermats were prepared using methods similar to those presented in our previous study.17 In short, an aqueous solution of silica/PGA hybrid and GFP was electrospun at ambient temperature, at which point the fibermat became water-insoluble, but could be characterized upon immersion in an aqueous buffer (Scheme 1). We measured the maximum GFP content that could be encapsulated in the base polymeric material, silica/PGA, and, for GFP content less than or equal to 15 wt %, we succeeded in constructing GFP-encapsulated fibermats with homogeneous nanofibers. As the GFP-encapsulated fibermats could become water-insoluble, they could be characterized either in their dry state or after immersion in a buffer (20 mM phosphate buffer, pH 7). Images of the dry-state GFP-encapsulated fibermat and its nanostructure, acquired with an SEM, are presented in Figure 1a. As a reference, the fibermat without GFP is shown in Figure 1b. The structures of nanofibers were homogeneous and their average diameters were ∼300 nm, which was similar to those in the absence of encapsulated GFP. The fibermat was not fragile, but relatively soft because of the hybrid’s organic component, PGA. We determined whether GFP encapsulation in fibermats altered the extent of Si−O−Si bond formation
using ATR-FTIR spectroscopy. ATR-FTIR spectra of the silica/PGA fibermats with and without 15 wt % GFP are summarized in Figure 1c. The IR band at 1630 cm −1 corresponds to the CO vibration of the amide bond (amide I) in PGA.17,23 The IR bands at 1570 and 1410 cm−1 correspond to the CO vibration in the carboxylate in PGA.17,24The IR bands at 1100 and 1030 cm−1 correspond to Si−O−Si bonds (Si−O stretch) in cyclic and random network configurations, respectively.25 Because GFP and PGA are both a polypeptide and contain amide bonds in the main chain, the IR peaks assigned to amide bonds were similar to those with and without GFP. While, significant differences in the IR peaks for linear Si−O−Si (1030 cm−1) and cyclic Si−O−Si (1030 cm−1) were not observed, meaning that GFP encapsulation of less than 15 wt % did not induce significant effects on crosslinking in the silica/PGA 3D network. To characterize GFP encapsulation, fluorescence emission with UV irradiation (312 nm) or laser excitation (405 nm) and the fluorescence spectrum of the GFP-encapsulated silica/PGA fibermat were measured (Figure 2). Both measurements were obtained after wetting of the fibermats with 20 mM phosphate buffer (pH 7). As shown in Figure 2b,c, fluorescence emission from the silica/PGA fibermat indicated homogeneous emission from the whole nanofibers, which meant the GFP molecules in the nanofibers of the fibermat were homogeneously dispersed. For a detailed analysis, the fluorescence spectrum of GFP in the silica/PGA fibermat was compared to the spectrum of GFP in 20 mM phosphate buffer (pH 7) (Figure 2c). Although a slight red shift (
