Silk–Silk Interactions between Silkworm Fibroin ... - ACS Publications

Aug 28, 2017 - *E-mail: [email protected] (B.B.M.)., *E-mail: [email protected] ... The coating process is based on a silk assembly to achieve stabl...
0 downloads 0 Views 9MB Size
Research Article www.acsami.org

Silk−Silk Interactions between Silkworm Fibroin and Recombinant Spider Silk Fusion Proteins Enable the Construction of Bioactive Materials Linnea Nilebac̈ k,† Dimple Chouhan,‡ Ronnie Jansson,† Mona Widhe,† Biman B. Mandal,*,‡ and My Hedhammar*,† †

AlbaNova University Center, School of Biotechnology, KTH Royal Institute of Technology, 106 91 Stockholm, Sweden Biomaterial and Tissue Engineering Laboratory, Department of Biosciences and Bioengineering, Indian Institute of Technology Guwahati, Guwahati 781039, Assam, India



S Supporting Information *

ABSTRACT: Natural silk is easily accessible from silkworms and can be processed into different formats suitable as biomaterials and cell culture matrixes. Recombinant DNA technology enables chemical-free functionalization of partial silk proteins through fusion with peptide motifs and protein domains, but this constitutes a less cost-effective production process. Herein, we show that natural silk fibroin (SF) can be used as a bulk material that can be top-coated with a thin layer of the recombinant spider silk protein 4RepCT in fusion with various bioactive motifs and domains. The coating process is based on a silk assembly to achieve stable interactions between the silk types under mild buffer conditions. The assembly process was studied in real time by quartz crystal microbalance with dissipation. Coatings, electrospun mats, and microporous scaffolds were constructed from Antheraea assama and Bombyx mori SFs. The morphology of the fibroin materials before and after coating with recombinant silk proteins was analyzed by scanning electron microscopy and atomic force microscopy. SF materials coated with various bioactive 4RepCT fusion proteins resulted in directed antibody capture, enzymatic activity, and improved cell attachment and spreading, respectively, compared to pristine SF materials. The herein-described procedure allows a fast and easy route for the construction of bioactive materials. KEYWORDS: biomaterial, silk fibroin, recombinant spider silk, functionalization, self-assembly



INTRODUCTION

are relatively low, especially compared to the extraction of silk from silkworm cocoons because silkworm larvae can be sustained at high density.9,14 Silk from the domesticated mulberry silkworm Bombyx mori, belonging to the family Bombycidae, is widely used for textile production and has also been extensively studied for other purposes, especially as a biomaterial, as a result of its degradability.5,14 To a low cost, silk fibers can be harvested from B. mori cocoons or silk proteins (fibroins) can be extracted by dissolving cocoons in denaturing solvents and thereafter be reprocessed into new formats such as films, sponges, hydrogels, and microparticles.2 Nonmulberry

Silk is a protein-based natural material with unique mechanical properties. It can be processed into many different formats, offering a versatility that makes it interesting for various medical applications, including cell culture, tissue engineering, and drug delivery.1−6 Several strategies have been used to tune the silk properties and to add specific functions to silk, which further expands the application areas.7−9 The current investigation aims at constructing cost-effective silk materials possessing bioactive properties by combining different silk types. The bioactivity can be obtained by producing recombinant silk proteins fused with different bioactive motifs and domains such as cell-binding motifs10,11 and antibody fragments.12,13 Bioactive recombinant silk shows great potential, but the production is less cost-effective than that of natural silk. Yields © 2017 American Chemical Society

Received: July 24, 2017 Accepted: August 28, 2017 Published: August 28, 2017 31634

DOI: 10.1021/acsami.7b10874 ACS Appl. Mater. Interfaces 2017, 9, 31634−31644

Research Article

ACS Applied Materials & Interfaces

coating of 4RepCT fusion proteins, leading to bioactive materials. In a previous study, Morgan et al. combined BmSF with a recombinant spider silk fused to RGD by mixing the silk solutions prior to spin coating. The influence of the mixing ratios on the coating properties was analyzed, but no obvious enhancement in cell binding was confirmed.28 Attempts have also been made to extract recombinant spider silk from cocoons from transgenic silkworms,29,30 but because rather harsh conditions are required to degum the threads and reformat them into alternative scaffolds, incorporated bioactive domains might lose their function during the processing. In this work, we coated SF materials in different formats with the recombinant partial spider silk protein 4RepCT fused to three different bioactive moieties; the fibronectin (FN)-derived cell adhesion motif,11 the Protein A-derived Z domain with high affinity for the Fc region of antibodies,31 and the enzyme xylanase, which degrades the polysaccharide xylan.32 We present studies of the interactions between fibroin and recombinant spider silk and show that they result in bioactive materials.

silks have recently started to be explored for use as biopolymers, showing characteristics slightly different from those of B. mori silk. Antheraea assama, a nonmulberry silkworm under the Saturniidae family, is mainly found in the Assam region in India. The A. assama silk is sought-after because of its gold-shimmering silk.15,16 In contrast to B. mori silk fibroins (BmSFs), A. assama silk fibroins (AaSFs) contain arginine−glycine−aspartic acid (RGD) motifs,15 which promotes cell adhesion. AaSF-based materials in various formats similar to BmSF-derived materials can be generated from silk proteins extracted directly from the glands of A. assama or from solutions of the naturally spun fibers after dissolution in denaturing solvents. However, isolation of nonmulberry silk fibroin (SF) from spun fibers using conventional denaturing solvents yields a very low amount of protein.17 Therefore, isolation from glands is a more viable option. Previous in vivo studies of such silk matrixes have revealed nonimmunogenicity of the material.18 Vascularization and tissue remodeling were observed under in vivo conditions, making it a suitable substrate for tissue engineering applications.19 A recent study also demonstrated the immunomodulatory properties of threedimensional (3D) silk biomaterials, where the macrophage polarity was controlled in order to tailor the biological response.20 For wound healing applications and skin patches, electrospinning of SF can be employed for the construction of materials with micrometer to millimeter thickness consisting of networks of fibers with diameters in the nanometer range.18 It is hypothesized that a fiber network may be an advantage when used as a biomaterial because it resembles the fibrous structure of a tissue extracellular matrix.21 For 3D cell culturing, microporous structures are considered to be promising as scaffolds. Freeze-drying can be employed to transform SF into suitable scaffold dimensions,19 although these typically lack bioactivity, which would be desired for enhanced interactions with cells and biomolecules. SF materials can be rendered bioactive by physical adsorption or, in the case of electrospinning, by co-spinning with functional biomolecules.6,18 The loaded substances will then diffuse out of the material over time. To achieve a prolonged effect, immobilization of functional motifs onto the material is preferred. This can be obtained by the chemical coupling of active substances onto the amino acid residues exposed in SF.9,22 However, such a process leads to a risk for the denaturation of folded protein domains (e.g., enzymes) or entrapment in a conformation that could impair the activity. The construction of fusion proteins, where the DNA encoding a functional protein domain is fused to a silk protein sequence at the genetic level, allows production under native conditions upon translation into fusion proteins in host cells. The functional domain is then attached at a defined site and is retained in its active form during assembly into bioactive silk materials.23,24 It was recently shown that the recombinant silk protein 4RepCT25,26 assembles into thin coatings on various surfaces under mild buffer conditions and that the peptide motifs fused to them are functionally presented in the obtained coatings.27 We hypothesized that 4RepCT proteins are able to assemble also onto SF materials via silk−silk interactions. In that way, the advantages of both silk types can be combined; the costefficient extraction of silk from silkworms for the construction of large scaffolds in a wide variety of formats and the possibility of adding functional groups to the material via a thin top



EXPERIMENTAL SECTION

Preparation of SF Solutions from B. mori Silk Cocoons and A. assama Silk Glands. B. mori silk cocoons were used to isolate a BmSF solution following a standard protocol.33 Briefly, cocoons were cut into small pieces for degumming to remove sericin. Pieces of cocoons were boiled with 0.02 M sodium carbonate for 2 × 15 min. The obtained degummed silk fibers were dried, detangled, and dissolved in a 9.3 M lithium bromide (LiBr) solution at 60 °C. The dissolved SF/LiBr solution was incubated at 60 °C for 4 h to become highly viscous and subsequently dialyzed against Milli-Q water for 3 days using a cellulose dialysis membrane (MWCO 12 kDa, SigmaAldrich) to obtain aqueous BmSF. AaSF was isolated from the silk glands of A. assama silkworms as previously described.17,18 Briefly, silkworms in the fifth instar larvae stage were sacrificed. The silk glands were collected, and SF protein was extruded by squeezing the glands using forceps. The protein was gently washed with water three times and then dissolved in a 1% (w/v) aqueous sodium dodecyl sulfate (SDS) solution. The dissolved SF solution was centrifuged at 5000 rpm for 10 min at 4 °C. The supernatant SF solution was further dialyzed against Milli-Q water using a 12 kDa cellulose dialysis membrane for 4 h at 4 °C to obtain an aqueous AaSF solution. The SF concentration was measured using a gravimetric method and further diluted to the desired concentration. Fabrication of SF Mats, Scaffolds, and Coatings. Nanofibrous mats were prepared using electrospinning. Poly(vinyl alcohol) (PVA; Loba Chemie Pvt. Ltd.; 1700−1800 degrees of polymerization and 98−99 mol % hydrolysis) was mixed with SF to obtain the optimum viscosity for electrospinning.18 A highly viscous 13% (w/v) PVA solution was prepared by dissolving PVA granules in lukewarm Milli-Q water. The aqueous PVA solution was blended separately with the two different SF solutions (AaSF and BmSF) in a ratio of 4:1 PVA/SF (w/ w). The electrospinning unit Super ES1 (E-spin nanotech) was used at the following settings: voltage = 25 ± 3 kV, flow rate = 0.800 ± 0.100 mL/h, tip-to-collector distance (d) = 15 cm, and rotating speed of the drum collector = 500 rpm. Two types of nanofibrous mats were prepared: one hybrid mat of PVA + AaSF (PVA/AaSF, also called the AaSF mat) and one hybrid mat from a blend of PVA + BmSF (PVA/ BmSF, also called the BmSF mat). Hybrid electrospun mats were incubated in 70% ethanol for 2 h to render the mats water insoluble by the induction of β-sheet formation and to sterilize the materials. Subsequently, they were washed three to four times with sterile water to remove residual ethanol. Wet mats were treated under a UV lamp for 20−30 min in two cycles to further sterilize both sides of the electrospun mats. The mats were dried overnight at 37 °C and stored at −20 °C until use. 31635

DOI: 10.1021/acsami.7b10874 ACS Appl. Mater. Interfaces 2017, 9, 31634−31644

Research Article

ACS Applied Materials & Interfaces Microporous scaffolds of BmSF and AaSF were fabricated using a freeze-drying process. Briefly, 300 μL of 2.5% (w/v) BmSF and AaSF solutions, respectively, were transferred to a 24-well plate and frozen at −20 °C overnight followed by lyophilizing for 24 h. The freeze-dried scaffolds were treated with 100% ethanol for 1 h, followed by a 70% ethanol treatment overnight to induce β sheets. The scaffolds were washed three to five times with Milli-Q water for the complete removal of ethanol. Water-soaked scaffolds were autoclaved for sterilization prior to use and stored in 70% ethanol at 4 °C. An aqueous SF solution was freeze-dried using the drying manifold Alpha 1-4 LD plus (Martin Christ Gefriertrocknungsanlagen) for storage until preparation of the coatings. Lyophilized SF was dissolved into 1 mg/mL solutions in Milli-Q water for 1 h and centrifuged for 7 min at 10000 rpm to remove nondissolved particles. The supernatants were diluted in 20 mM tris(hydroxymethyl)aminomethane (Tris), pH 8.0, to obtain 0.1 mg/mL solutions. The solutions were used for the preparation of coatings according to the procedures described in each analysis section. Recombinant Spider Silk Proteins. Recombinant spider silk proteins were provided from Spiber Technologies AB and stored frozen in 20 mM Tris buffer, pH 8.0. 4RepCT (23 kDa) was used either as is or as fusion proteins with bioactive motifs and domains; FN−4RepCT (4RepCT fused with a cell-binding peptide motif from fibronectin,11 24 kDa), Z−4RepCT (4RepCT fused with the antibody binding Z domain,23 33 kDa), and Xyl−4RepCT (4RepCT fused with the xylan degrading enzyme xylanase,24 43 kDa). Proteins were thawed, centrifuged briefly to remove aggregates, and diluted in 20 mM Tris, pH 8.0, to obtain 0.1 mg/mL solutions. Quartz Crystal Microbalance with Dissipation (QCM-D). Using AT-cut quartz crystals with metal-film electrodes as sensors and monitoring their oscillating frequencies in response to an applied pulsating voltage, the QCM-D technique allows the observation of small changes at the sensor surfaces, such as the adsorption of biomolecules. Adsorption of mass onto the sensor leads to a decrease in its oscillation frequencies and an increase in the dissipation of oscillations. The viscoelasticity of the adsorbed layer also influences the dissipation.34 Gold-coated QCM-D sensors (Q-Sense QSX 301, Biolin Scientific) were cleaned extensively and rendered hydrophobic by incubation in 2 mM 1-undecanethiol (Sigma-Aldrich) in 99.5% ethanol (Solveco), as previously described.27 Excess alkylthiols were removed by sonication in pure 99.5% ethanol five times, for 5 min each. Sensors were stored in 99.5% ethanol until use. QCM-D was performed using an E4 instrument (Q-sense AB) at 20.0 °C. Buffer and proteins were flowed over the sensors at 20 μL/min. A baseline was obtained with 20 mM Tris. 0.1 mg/mL protein solutions were freshly prepared before each measurement and kept on ice until the modules were entered. First, AaSF or BmSF solutions were flowed over the sensors. Following a buffer flow to regain the baseline, 4RepCT protein variants were flowed through the system. Finally, the baseline was regained with Tris buffer. Except for AaSF followed by 4RepCT, for which one measurement was done, all measurements were performed in duplicate. If not otherwise stated, overtone 3 was chosen for the data representation. Atomic Force Microscopy (AFM). AaSF, BmSF, and 4RepCT coatings were prepared on siliconized hydrophobic coverslips (Paul Marienfeld) by incubation in 0.1 mg/mL protein solutions at room temperature for 60 min, followed by three times wash with 20 mM Tris. Samples with top coatings were prepared by a total adsorption time of 60 min (i.e., 30 min in AaSF or BmSF followed by 30 min in 4RepCT) with a 20 mM Tris wash after incubation in each protein solution. AFM imaging was done in a Bruker Dimension FastScan instrument, using ScanAsyst Fluid+ tips and PeakForce Tapping. All images were obtained in a droplet of 20 mM Tris. Photographs. SF coatings were prepared as described in the AFM section, and mats and scaffolds were prepared as described in the fabrication section. Top coatings were made by incubation in 0.1 mg/ mL 4RepCT at room temperature for 60 min. The coatings were washed at least two times in Tris and dried before being photographed. The macroscopic appearances of AaSF and BmSF

coatings, mats, and scaffolds top-coated with 4RepCT were documented by a Canon EOS 5D Mark II camera. Scanning Electron Microscopy (SEM). The morphology of silk coatings, mats, and scaffolds was analyzed using field-emission SEM (FESEM; Sigma 300, Zeiss). SF coatings and 4RepCT top coatings were prepared as described in the previous section. For FESEM imaging, dried samples were placed on a stub covered with carbon tape, coated with gold by 130 s of sputtering, and characterized at an accelerating voltage of 2−4 kV. ImageJ software (Wayne Rasband, National Institute of Health) was used to determine the diameter of nanofibers in mats by measuring 50 randomly selected nanofibers from the magnified images. The pore size of the scaffold was also investigated using ImageJ software. Detection of Hexahistidine in Combined Silk Coatings. SF coatings were prepared from 0.1 mg/mL AaSF or BmSF by incubation for 30 min at room temperature. Top coatings were made by the incubation of SF coatings in Xyl−4RepCT for another 30 min. The coatings were washed at least two times with 20 mM Tris after each incubation. The coatings were immersed in 2 μg/mL of a fluorophorelabeled antihistidine antibody (Penta-His Mouse IgG-Alexa Fluor 488, Product: 35310, Qiagen) in phosphate-buffered saline (PBS) with 0.05% Tween 20 for 60 min at room temperature. After three washes with PBS/0.05% Tween 20, fluorescence microscopy was performed using an inverted Nikon Eclipse Ti fluorescence microscope. As negative controls, coatings of BmSF with 4RepCT and AaSF with 4RepCT were treated in the same way. Coating SF Materials with Recombinant Spider Silk for Cell Culture Studies. SF coatings were prepared in hydrophobic 96-well plates (TC Plate 96 Well, Suspension, F, Sarstedt) by incubation of 0.1 mg/mL AaSF or BmSF solution in wells for 60 min at 37 °C. The coatings were washed two times with 20 mM Tris. SF coatings and mats (prewetted in sterile water) were top-coated by incubation in 0.1 mg/mL FN−4RepCT for 60 min at 37 °C. Excess silk solution was removed, and coatings and mats were washed two times with 20 mM Tris. Cell Culture. Human dermal fibroblasts of neonatal origin (HDF; ECAAC, Salisbury, U.K.) and the human keratinocyte cell line HaCaT (Cell Line Service, Germany) were cultured in a Dulbecco’s modified Eagle’s medium (DMEM) nutrient mixture F12 ham (Sigma) supplemented with 5% fetal bovine serum (Sigma), penicillin (100 U/mL), and streptomycin (100 μg/mL). HDF was used at passage 9. The medium was changed every second to third day. Cellular Stainings. Cells were harvested and seeded onto either coatings in 96-well plates (TC Plate 96 Well, Suspension, F, Sarstedt) or electrospun mats in 24-well plates at a seeding density of 5000/cm2. After 48 h (coatings) or 72 h (mats) of culture, cells were washed in PBS, fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X 100, and stained with Phalloidin-Alexa Flour 488 or 594 diluted 1:500 for F-actin and Dapi 1:1000 for nuclei. Samples were viewed under a Nikon Eclipse Ti inverted fluorescence microscope, and micrographs were taken using NIS elements BR (Nikon). Binding of IgG to Z−Silk-Coated Materials. SF coatings were prepared as described in the photograph section, and coatings, mats, and scaffolds of AaSF and BmSF were top-coated with Z−4RepCT for 60 min at room temperature followed by washing with 20 mM Tris at least two times. All materials were immersed in 0.05 mg/mL IgG/ fluorophore [Rabbit Anti-Mouse IgG (H + L)/Alexa Fluor 488, Product A-11059, Sigma-Aldrich] in PBS for 30 min. After washing with PBS/0.05% Tween 20 three times, PBS was added to all materials, which were then analyzed by fluorescence microscopy using an inverted Nikon Eclipse Ti fluorescence microscope. As control materials, BmSF and AaSF coated with 4RepCT were treated in the same way. Enzymatic Activity Assay. Coatings, mats (prewetted in 20 mM Tris), and scaffolds of AaSF and BmSF top-coated with Xyl−4RepCT were prepared in the same way as those for the hexahistidine detection. The enzymatic activity of xylanase in the silk materials was determined using a colorimetric sugar-reducing assay (Megazyme, Bray, Ireland), based on a method described elsewhere.35,36 In brief, enzyme-functionalized silk materials were immersed in a 1% (w/v) 31636

DOI: 10.1021/acsami.7b10874 ACS Appl. Mater. Interfaces 2017, 9, 31634−31644

Research Article

ACS Applied Materials & Interfaces

Figure 1. Silk−silk coatings formed by physical adsorption. QCM-D studies of the formation of silk coatings from protein solutions of (a) AaSF and (b) BmSF flowed over oscillating sensors. From line i, SF solutions are flowed over the surfaces, followed by a buffer rinse from line ii. From line iii, 4RepCT solutions are flowed over the surfaces, followed by another buffer rinse from line iv. Overtones 3−13 are shown for representative measurements. Note the different scales in parts a and b.

Figure 2. Nanotopography of silk coatings. AFM images of (a) AaSF coating, (b) BmSF coating, (c) AaSF with a 4RepCT top coating, and (d) BmSF with a 4RepCT top coating. Height curves from sections along the marked arrows are shown beneath each image. Images are 1 μm wide. substrate solution (Wheat Arabinoxylan, Product P-WAXYM, Megazyme) at pH 6.3. A 20 mM Tris buffer was used as the negative control. The enzymatic reaction was allowed to proceed for 60 min at room temperature, after which the reaction was terminated by the addition of a copper reagent. The enzymatic product was visualized in a color-development step by the addition of sodium arsenate, followed

by detection of the product by measuring the absorbance at 520 nm using a CLARIOstar microplate reader (BMG LABTECH). Image Analysis of Fluorophore-Labeled Samples. The fluorescence intensity of Xyl−4RepCT top-coated SF coatings after staining with an antihistidine antibody was measured using the Intensity profile tool in NIS elements BR software. The intensity 31637

DOI: 10.1021/acsami.7b10874 ACS Appl. Mater. Interfaces 2017, 9, 31634−31644

Research Article

ACS Applied Materials & Interfaces

Figure 3. Various formats of AaSF coated with 4RepCT. Photographs of AaSF as (a) coating (7 mm wide) on a circular coverslip, (b) electrospun mat (12 mm wide), and (c) microporous scaffold (6 mm wide) with 4RepCT top coating. SEM images of AaSF coating, electrospun mat, and microporous scaffold alone (d−f) and with 4RepCT top coating (g−i). Scale bars are 2 μm for coatings and mats and 100 μm for microporous scaffolds.

1) show that a fibroin layer is formed for both AaSF and BmSF and that 4RepCT proteins readily adsorb onto the fibroin coatings. In the subsequent buffer flow, the frequency and dissipation signals are constant, which means that no proteins are detached, indicating stable silk−silk interactions. We thus show that fibroin easily can be coated with recombinant spider silk in mild buffer (20 mM Tris, pH 8.0) by physical adsorption. AaSF assembles faster than BmSF (Figure 1), forming thicker and more viscous coatings, as concluded from a larger change in the frequency and dissipation signals for AaSF. While BmSF is extracted from B. mori cocoons by solubilization under denaturing conditions, the AaSF protein dope is taken directly from A. assama glands in its native form. To elucidate whether the denaturing conditions used for the extraction of BmSF are the reason for the slower assembly at surfaces, a sample of AaSF was analyzed with QCM-D after being prepared with the same denaturing protocol as BmSF, using 9.3 M LiBr with subsequent dialysis against water. The same fast assembly occurs for AaSF treated according to the BmSF extraction protocol (Figure S1) as that after direct extraction from glands (Figure 1a), indicating that the intrinsic differences between these types of fibroins affect their surface assembly properties rather than their preparation procedures. Nanotopography Differs for Coatings Made from the Three Types of Silk Proteins. Coatings made by the physical adsorption of AaSF, BmSF, and 4RepCT, as well as AaSF and BmSF with 4RepCT top coatings, were imaged by AFM to reveal the nanotopography of the silk types. All three silk types form nanofibrillar structures, but their characteristics differ (Figure 2, 1 μm images; Figure S2, 1 and 10 μm images). While 4RepCT form distinct nanofibrils (Figure S2), BmSF forms much shorter structures in a single layer (Figure 2b). The AaSF

profile across the image was measured at 10 different positions, and a pixel mean intensity was calculated for each image. For visual representation, the brightness and contrast of the micrographs were enhanced by 100 and 80, respectively, using Adobe Photoshop CC 2015.0.1. However, this was not applied when the intensity data for quantification and subsequent statistical analyses were extracted. The fluorescence intensities of bound IgG/fluorophore to Z−silkcoated SF materials were extracted from fluorescence images as mean gray values (MGVs) using ImageJ software. After removal of the bright signal artifacts using the Remove outliers tool (radius, 10 pixels; threshold, 30), the mean intensity MGV values were calculated by ImageJ using four different areas per material. Similar to the images of the samples stained with antihistidine antibody, the brightness and contrast of IgG-stained micrographs for visual representation were enhanced by 100 and 80, respectively, for coatings and mats using Adobe Photoshop CC 2015.0.1. For scaffolds, the contrast was enhanced by 80 and the brightness was left unmodified. The modifications were not applied when the intensity data for quantification and subsequent statistical analyses were extracted. The cell coverage area (F-actin) on the coatings was measured using NIS elements BR software. For each condition, 4−7 different areas from duplicate samples were analyzed. Statistics. The Student’s t test (or Mann−Whitney U test) was used for statistical analysis of the fluorescence intensity, spectrophotometric absorbance from enzymatic activity, and cell coverage area. P values of