Recombinant Spider Silk Genetically Functionalized with Affinity

Mar 30, 2014 - Still, the bioactive domains are concluded to be folded and accessible, ... Citation data is made available by participants in Crossref...
0 downloads 0 Views 459KB Size
Article pubs.acs.org/Biomac

Recombinant Spider Silk Genetically Functionalized with Affinity Domains Ronnie Jansson,† Naresh Thatikonda,† Diana Lindberg,†,⊥ Anna Rising,†,‡ Jan Johansson,†,‡,§ Per-Åke Nygren,∥ and My Hedhammar*,† †

Department of Anatomy, Physiology and Biochemistry, Swedish University of Agricultural Sciences, Biomedical Center, SE-751 23 Uppsala, Sweden ‡ Department of Neurobiology, Care Sciences and Society (NVS), Karolinska Institutet, Novum, fifth floor, SE-141 86 Stockholm, Sweden § Institute of Mathematics and Natural Sciences, Tallinn University, Narva mnt 25, 101 20 Tallinn, Estonia ∥ Division of Protein Technology, School of Biotechnology, KTH Royal Institute of Technology, SE-106 91 Stockholm, Sweden S Supporting Information *

ABSTRACT: Functionalization of biocompatible materials for presentation of active protein domains is an area of growing interest. Herein, we describe a strategy for functionalization of recombinant spider silk via gene fusion to affinity domains of broad biotechnological use. Four affinity domains of different origin and structure; the IgG-binding domains Z and C2, the albumin-binding domain ABD, and the biotin-binding domain M4, were all successfully produced as soluble silk fusion proteins under nondenaturing purification conditions. Silk films and fibers produced from the fusion proteins were demonstrated to be chemically and thermally stable. Still, the bioactive domains are concluded to be folded and accessible, since their respective targets could be selectively captured from complex samples, including rabbit serum and human plasma. Interestingly, materials produced from mixtures of two different silk fusion proteins displayed combined binding properties, suggesting that tailor-made materials with desired stoichiometry and surface distributions of several binding domains can be produced. Further, use of the IgG binding ability as a general mean for presentation of desired biomolecules could be demonstrated for a human vascular endothelial growth factor (hVEGF) model system, via a first capture of anti-VEGF IgG to silk containing the Z-domain, followed by incubation with hVEGF. Taken together, this study demonstrates the potential of recombinant silk, genetically functionalized with affinity domains, for construction of biomaterials capable of presentation of almost any desired biomolecule.



chemical cross-linking.9 Advances in transgenic silkworm technology have demonstrated that additional protein moieties, e.g., enhanced green fluorescent protein (eGFP)10 and basic fibroblast growth factor (bFGF),11 can be produced as fibroin fusion proteins in the silk glands. However, the isolation and purification of silk fibroins generally require multiple harsh steps (including degumming in NaOH and solubilization in, e.g., concentrated lithium bromide at elevated temperatures or organic solvent such as hexafluoroisopropanol) and these treatments irreversibly destroy the biological activity of protein domains that require a native fold. Moreover, although beneficial due to its high availability, silk harvested from silkworm cocoons is associated with inherent drawbacks of all natural materials such as batch-to-batch variability.

INTRODUCTION Silk has been identified as a biomaterial with potential for many applications due to its exceptional mechanical properties in combination with biocompatibility.1−3 Functionalization of silk fibers via chemical or genetic incorporation of peptides or proteins with binding or enzymatic functionalities has been recognized as a promising route forward to generate novel materials for both biotechnological and medical (e.g., regenerative medicine) applications. Silk fibers isolated from cocoons made by the silkworm have been shown useful for several applications, especially after tailoring with desired properties. The silkworm fibroin contain several reactive amino acids (e.g., serine, tyrosine, glutamic acid) that can be modified with known chemistries4 in order to covalently attach, e.g., cell binding peptides (RGD5,6), growth factors (BMP27) or affinity domains (NeutrAvidin8), albeit in a more or less random fashion. Controlled conformational transition during drying onto fibroin films was recently shown to be more efficient for loading of functional antibodies than conventional © 2014 American Chemical Society

Received: January 23, 2014 Revised: March 18, 2014 Published: March 30, 2014 1696

dx.doi.org/10.1021/bm500114e | Biomacromolecules 2014, 15, 1696−1706

Biomacromolecules

Article

the ability of the silk part to form a solid silk-like material, and the ability of the affinity domain to selectively bind its target molecule?

Recent progress in recombinant production of spider silk proteins,12,13 in combination with utilization of the concept of fusion proteins, opens up another possibility to combine the structural properties of spider silk with bioactivities, a combination not found together in nature. Short cell binding peptides14,15 have been shown to improve cell compatibility of recombinant silk. Moreover, longer protein sequences that induce selective mineralization (hydroxyapatite,16 silica,17−19 calcium phosphates20) have been successfully produced as silk fusion proteins, and shown upon the utility of silk for controlled synthesis of inorganic composite materials. Within these systems, up to ∼300 residues have been added to the silk protein and still silk-like materials have been successfully produced, although post-treatment of methanol has been necessary in order to render water-insoluble materials. Recombinant silk in fusion with antimicrobial peptides (AMP) have also been successfully produced in bacterial systems.21 Solutions of these AMP silk-fusions show antibacterial activity, although the effect is diminished by aggregation at higher concentrations. Native spider silk is formed via a pH switch-triggered selfassembly22,23 of specialized soluble proteins, spidroins, containing a long repetitive sequence rich in alanine and glycine, flanked by conserved terminal domains.24 A partial spidroin containing the C-terminal domain and a piece of the repetitive part (4RepCT), has previously been identified and shown to be sufficient for self-assembly into silk-like fibers after recombinant expression in a heterologous host.25,26 Means for how to produce and treat 4RepCT proteins so that they maintain their ability to form silk-like material in various formats such as fibers, films and foam, have been described.2,13,27−29 Cell binding peptides have been genetically incorporated to such recombinant silk proteins to result in functionalized spider silk fibers of potential use as cell growth supports in regenerative medicine applications.30 Peptides attached to the N-terminus of the silk protein do not interfere with the capability to self-assemble in near-physiological buffers, and get well-exposed in the formed silk materials. However, the foreign sequences so far incorporated into these small and wellexpressed partial spidroins, and similar systems, typically correspond to polypeptides without defined three-dimensional structure. A demonstration of functional incorporation of larger proteins with a folding-dependent activity into chimeric 4RepCT spidroins, retaining an ability to form fibers, would pave the way for additional types of functionalized silk-like materials. Thus, we have herein investigated the possibility to use recombinant DNA technology to develop functionalized silk materials that present whole functionally folded protein domains. Four different affinity domains of broad biotechnological relevance were incorporated at the genetic level to the partial spider silk protein, 4RepCT. To show upon the generality of the method, we chose protein domains of different origin and with different fold and binding partners: (1) domain Z, a 58 residue three-helix bundle domain31 derived from staphylococcal protein A (SPA), (2) domain C2, a 56 residue mixed β-sheet/α helix domain32 from streptococcal protein G (SPG) that both bind immunoglobulin G (IgG), (3) ABD, a 46-residue three-helix bundle albumin binding domain33 from SPG, and (4) domain M4, a 159-residue all β-sheet barrel monomer derived from streptavidin,34 that binds biotin. The four different silk fusion proteins were produced and investigated with the aim to answer the following question: Will the silk fusion proteins attain the properties of both parts:



EXPERIMENTAL SECTION

Gene Constructions. PCR primers were designed to generate DNA fragments encoding domain Z, C2, ABD, and M4, respectively, from cloning vectors containing these sequences. Into the forward primers, an NdeI restriction endonuclease recognition site was introduced, whereas an EcoRI restriction site was introduced into the reverse primers. Moreover, a recognition site for protease 3C (amino acid sequence: LEALFQGP), was also incorporated into the reverse primers for Z, C2 and ABD constructs. Generated PCR products and the target vector, a modified pT7 vector containing sequences for a His6-tag and recombinant spider silk, 4RepCT,25 were treated with NdeI and EcoRI endonucleases (Fermentas), and subsequently joined with T4 DNA ligase (Fermentas). Protein Expression and Purification. The four different functionalized spider silk DNA constructs (Z-4RepCT, C2-4RepCT, ABD-4RepCT, M4-4RepCT) were used to transform Escherichia coli BL21(DE3) cells (Merck Biosciences). The cells were grown in Luria−Bertani medium supplemented with 70 μg/mL kanamycin to an OD600 value of 1−1.5, followed by induction of protein expression with 300 μM IPTG (isopropyl β-D-1-thiogalactopyranoside). Cells were harvested by centrifugation and the resulting cell pellet was dissolved in 20 mM Tris-HCl (pH 8.0), supplemented with lysozyme (Sigma) and DNaseI (Sigma) for cell lysis. Cell lysate was recovered after 30 min of centrifugation and loaded onto Chelating Sepharose Fast Flow-Zn2+ matrix (GE Healthcare). Elution of bound proteins from the matrix was performed with 300 mM imidazole in 20 mM Tris-HCl (pH 8.0), after which the proteins were dialyzed against 20 mM Tris-HCl (pH 8.0). Dialyzed proteins were concentrated to a final protein concentration of 1−3 mg/mL. As control, 4RepCT protein (without any additional protein domain) or RGE-4RepCT (with the tripeptide RGE) was used. 4RepCT and RGE-4RepCT were produced and purified as previously described.25 Film and Fiber Formation. Films of Z-4RepCT, C2-4RepCT, ABD-4RepCT, M4−4RepCT, RGE-4RepCT, and 4RepCT were cast in 96-well plates (Tissue Culture Plate, Suspension Cells, 83.1835.500, Sarstedt). Each film was made from 15 μL of 1 mg/mL soluble protein. The films were then allowed to solidify overnight at 30 °C and 25% relative humidity. Films used for the selective binding and stability analyses were cast in 24-well plates (Tissue Culture Plate, 83.1836, Sarstedt) from 100 μL soluble protein (1 mg/mL), and allowed to solidify overnight at ambient temperature. Macroscopic fibers were made from all four silk fusion proteins, and for 4RepCT, by allowing soluble proteins (1 mg/mL) self-assemble by gentle tilting overnight at room temperature, as previously described.26 Light Microscopy. Light microscopy (0.8× magnification) of films was performed using a stereo light microscope (Nikon) with a portable USB camera. For fibers, light microscopy (2× magnification) was carried out using an inverted Nikon Eclipse Ti light microscope. Measurement of Z-4RepCT Film Thickness. A film was cast from 15 μL of soluble Z-4RepCT protein (3 mg/mL) onto a glass slide, and allowed to dry at 30 °C and 25% relative humidity. The thickness of the film was measured using confocal Raman microscopy, by analyzing the Raman profiles at five different positions on the film. N-Terminal Sequencing of Z-4RepCT Fiber. Amino acid sequence analysis of a Z-4RepCT fiber was performed by cycles of sequencer-assisted Edman degradation (Protein Analysis Center, Karolinska Institutet, Sweden). Fourier Transform Infrared Spectroscopy. Films of Z-4RepCT, C2-4RepCT, ABD-4RepCT, M4-4RepCT, and 4RepCT were cast and dried as previously stated. Fibers of all four spider silk fusion proteins, and 4RepCT, were allowed to air-dry in room temperature. Attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR) was recorded in absorbance mode for both films and fibers using a platinum ATR unit from Bruker (Tensor 37). The region investigated was 4000 cm−1 to 850 cm−1, with increments of 2 cm−1. For plotting 1697

dx.doi.org/10.1021/bm500114e | Biomacromolecules 2014, 15, 1696−1706

Biomacromolecules

Article

of graphs, the background absorption (defined at 1720 cm−1 for films and at 1732 cm−1 for fibers) was first subtracted. For calculation of the β/α absorption ratio for each IR spectrum, the absorption value at the β peak was divided by the corresponding absorption value at the α peak. Selective Binding to Functionalized Silk. Films and fibers of Z4RepCT were immersed in 500 μL of five times diluted rabbit serum (National Veterinary Institute, Uppsala, Sweden) or in 500 μL of 50 μg/mL purified rabbit IgG (Vector Laboratories, Inc., Burlingame, CA) for 1 h at room temperature. C2-4RepCT films and fibers were immersed in 500 μL of three times diluted rabbit serum (Normal rabbit serum, Invitrogen), also for 1 h at room temperature. Films and fibers of ABD-4RepCT were instead immersed in 500 μL of five times diluted human blood plasma (Akademiska Hospital, Uppsala, Sweden) for 1 h at room temperature. All films and fibers were washed three times in 600 μL phosphate-buffered saline (PBS), pH 7.4. Bound IgG (for Z-4RepCT and C2−4RepCT) or albumin (for ABD-4RepCT) was then released by 20 min of incubation in 500 μL low-pH buffer (0.5 M acetic acid, 1 M urea, 100 mM NaCl, pH 2.7) at room temperature. Released fractions were analyzed by nonreducing SDSPAGE. Films of M4-4RepCT were washed three times in 20 mM Tris-HCl (pH 8.0) before and after incubation with 50 μL of biotinylated DNA (700 ng/μL) for 15 min. Biotinylated DNA was prepared by performing a PCR reaction using a sense primer with 5′-biotin and EcoRI restriction sequence, and a regular antisense primer. After wash, the biotinylated PCR product bound to M4-4RepCT films was released by incubation with EcoRI for 1 h at 37 °C and analyzed by Agarose Gel Electrophoresis (AGE). Distribution of Bound Target Molecules to Functionalized Silk Films. Films of Z-4RepCT were washed in PBS, followed by blocking with 1% bovine serum albumin (BSA) in PBS for 1 h at room temperature with gentle shaking. For binding of fluorophore-labeled IgG, each film was immersed in 100 pmol of rabbit antimouse IgGAlexa Fluor 488 (Invitrogen) in PBS for 1 h at +4 °C with gentle shaking. All films were washed two times with 0.05% (v/v) Tween 20 in PBS, followed by a final wash with only PBS. Control films of RGE4RepCT were treated in the same way. Fluorescence microscopy of Z4RepCT and RGE-4RepCT films was carried out using an inverted Nikon Eclipse Ti fluorescence microscope using excitation at 455−490 nm and emission at 500−540 nm. C2-4RepCT and ABD-4RepCT films were washed with 0.5% Tween 20 in PBS. Each film was then incubated with 100 μL of biotinylated human blood plasma (diluted five times) containing 0.5% Tween 20 at room temperature for 20 min (C2-4RepCT) or 2 min (ABD-4RepCT). After three times of washing in PBS/0.5% Tween 20, films were incubated with 100 μL of 1 μM streptavidin-Alexa Fluor 488 (Invitrogen) containing 0.5% Tween 20 at room temperature for 2 min. Finally, films were washed three times with PBS/0.5% Tween 20. Control films of 4RepCT were treated in the same way. All films were then visualized by fluorescence microscopy as indicated above. Films of M4-4RepCT were washed once with 0.5% Tween 20 in PBS, followed by immersion in 100 μL of 3 μg/mL Atto 565-biotin (Fluka) containing 0.5% Tween 20 for 30 min at room temperature. Films were then washed three times in PBS/0.5% Tween 20. Control films of 4RepCT were treated in the same way. Fluorescence microscopy of films was carried out using an inverted Nikon Eclipse Ti fluorescence microscope (excitation: 540−580, emission: 605−650 nm). Effect of Film Concentration on Z-4RepCT Binding. Z4RepCT films and RGE-4RepCT control films containing different amounts of protein (506, 400, 300, 200, 126, 84, 56, 28, 14, and 7 × 10−12 mol, respectively) were cast as previously stated. The procedure for binding of rabbit antimouse IgG-Alexa Fluor 488 (Invitrogen) to Z-4RepCT and to control films was same as stated in the previous section. The level of fluorescence for each film was detected using a plate reader (Infinite M200, Tecan) and visualized by fluorescence microscopy using an inverted Nikon Eclipse Ti fluorescence microscope (excitation: 455−490 nm, emission: 500−540 nm).

Quantification of Z-4RepCT Binding. The level of fluorescence measured for the binding of IgG-fluorophore to Z-4RepCT films, described in the previous section, was used to quantify the binding of IgG-Alexa Fluor 488 to 15 μL of 0.85 mg/mL Z-4RepCT films (corresponding to 400 × 10−12 moles of Z-4RepCT). Dilution series of soluble rabbit antimouse IgG-Alexa Fluor 488 (Invitrogen), in which the level of fluorescence was detected using a plate reader (Infinite M200, Tecan), was used to establish a linear regression fit, describing the relationship between the level of fluorescence and the number of soluble IgG-Alexa Fluor 488 molecules present. From this relationship, the total number of bound IgG-Alexa Fluor 488 molecules to the Z-4RepCT films (15 μL, 0.85 mg/mL) and the number of bound IgG-Alexa Fluor 488 molecules per film surface area could be estimated. In addition, for quantification of IgG binding by SDS-PAGE, Z4RepCT films from 0.91 mg/mL soluble protein (corresponding to 430 × 10−12 moles of Z-4RepCT) were cast. Films were washed once with PBS, and then immersed in 100 μL of five times diluted rabbit serum (Normal rabbit serum, Invitrogen) for 1 h at room temperature. Next, films were washed for three times in PBS, and bound IgG released by 20 min of incubation in 50 μL low-pH buffer (0.5 M acetic acid, 1 M urea, 100 mM NaCl, pH 2.7) at room temperature. Released fractions, together with purified rabbit IgG (IgG from rabbit serum, purified immunoglobulin, Sigma) of known amount, were analyzed by nonreducing SDS-PAGE. The amount of released IgG from each Z4RepCT film was estimated from the SDS-PAGE gel using Image Lab software (Bio-Rad Laboratories) with an included standard of pure rabbit IgG of known amounts. Stability of Z-4RepCT. Z-4RepCT films and fibers, previously used once to bind IgG from rabbit serum and pure IgG, were immediately after the first usage and regeneration immersed in 20 mM Tris-HCl (pH 8.0) and stored at +4 °C for 70 days. After storage, the films were again subjected to rabbit serum or pure rabbit IgG, according to the protocol previously stated. Released fractions were analyzed by nonreducing SDS-PAGE. 4RepCT was used as control and treated in the same way. The films and fibers, subjected to rabbit serum two times, were stored for another 44 days at +4 °C, immersed in 20 mM Tris-HCl. Next, these films and fibers were immersed in 8 M urea for 20 min at room temperature followed by two times of washing in 20 mM TrisHCl, and subsequent incubation in 20 mM Tris-HCl at +4 °C for another six months. After storage they were again subjected to rabbit serum according to the protocol previously stated. Released fractions were analyzed by nonreducing SDS-PAGE. 4RepCT was used as the control and treated in the same way. The same Z-4RepCT films and fibers were, after yet another 22 days of storage in 20 mM Tris-HCl at +4 °C, immersed in 0, 0.5, or 1 M NaOH for 20 min at room temperature. After a single wash in PBS, films and fibers were again subjected to rabbit serum according to the protocol previously stated. Released fractions were analyzed by nonreducing SDS-PAGE. 4RepCT was used as control and treated in the same way. Two Z-4RepCT fibers were transferred to tubes containing 20 mM Tris-HCl (pH 8.0), after which one of the fibers was sterilized by autoclaving for 20 min at 121 °C. Both fibers were then incubated in 500 μL of five times diluted human blood plasma for 1 h at room temperature. After three times of washing in 600 μL PBS, bound IgG was eluted in 500 μL of low-pH buffer (pH 2.7) at room temperature. Released fractions were analyzed by nonreducing SDS-PAGE. 4RepCT was used as control and treated in the same way. Combining Two Functionalizations. Soluble Z-4RepCT and soluble ABD-4RepCT silk fusion proteins were mixed so that the resulting solution contained 0.58 mg/mL Z-4RepCT and 0.52 mg/mL ABD-4RepCT. Mixed Z-4RepCT/ABD-4RepCT films were then cast as previously stated. Mixed Z-4RepCT/ABD-4RepCT fibers were also prepared by allowing a mixture of 0.28 mg/mL of Z-4RepCT and 0.25 mg/mL of ABD-4RepCT to self-assemble. Mixed films and mixed fibers were washed once in PBS, followed by immersion in five times diluted human blood plasma for 1 h at room temperature. After washing three times in PBS, bound IgG, and 1698

dx.doi.org/10.1021/bm500114e | Biomacromolecules 2014, 15, 1696−1706

Biomacromolecules

Article

Figure 1. Production of soluble functionalized silk-fusion proteins and characterization of the silk-like materials formed thereof. (a) Schematic representations of four different functionalized silk fusion proteins, Z-4RepCT (32 kDa), C2-4RepCT (31 kDa), ABD-4RepCT (30 kDa), and M44RepCT (41 kDa). Secondary structure elements of the functionalization domains (Protein Data Bank ID: 2spz, 1fcc, 1gjt, and 1swe, respectively) are color-coded; α-helix in light blue, β-sheet in purple. In addition, all four silk fusion proteins contain an N-terminal His6-tag (not shown). The recombinant partial spider silk protein 4RepCT (23 kDa), to which each functionalization domain has been attached, is schematically presented below. (b) SDS-PAGE of eluted fractions from immobilized metal ion affinity chromatography (IMAC) of four soluble functionalized silk fusion proteins, and 4RepCT. Target proteins are indicated by black squares. (c) Light micrographs showing the appearances of functionalized silk fusion protein films and fibers, compared to those of 4RepCT. Scale bars indicate 1 mm. (d) Infrared spectroscopy absorption spectra from films and fibers of four functionalized silk fusion proteins and 4RepCT. Absorption peaks for α-helix and β-sheet in films are indicated by lines. at room temperature for 10 min. Finally, films were washed three times in PBS/0.5% Tween 20. Included in the experiment was also a set of Z-4RepCT films not subjected to the step of antihuman VEGF IgG binding, they were instead subjected to PBS/0.5% Tween 20 in this particular step. Control films of 4RepCT were treated in the same way. All films were visualized by fluorescence microscopy using an inverted Nikon Eclipse Ti fluorescence microscope using excitation at 455−490 nm and emission at 500−540 nm. M4-4RepCT films were washed once with 0.5% Tween 20 in PBS. Absorbance at 354 nm of films in 100 μL of PBS/0.5% Tween 20 was measured using a plate reader (Infinite M200, Tecan), prior to 10 min incubation of films in 100 μL of 10 μM recombinant human EGF (epidermal growth factor), prelabeled with chromophoric biotin. Films

albumin were released by incubation in low-pH buffer for 20 min at room temperature. Released fractions were analyzed by nonreducing SDS-PAGE. Binding of Growth Factors to Functionalized Silk. After an initial wash of Z-4RepCT films with 0.5% Tween 20 in PBS, films were incubated with 100 μL of 400 nM rabbit antihuman VEGF165 (vascular endothelial growth factor) IgG (PeproTech) in 0.5% Tween 20 for 10 min at room temperature. After three times of washing with PBS/ Tween 20, films were immersed in 100 μL of 150 nM biotinylated recombinant human VEGF165 in 0.5% Tween 20 for 10 min at room temperature, followed by an additional three-times wash with PBS/ Tween 20. Next, films were incubated with 100 μL of 1 μM streptavidin-Alexa Fluor 488 (Invitrogen) containing 0.5% Tween 20 1699

dx.doi.org/10.1021/bm500114e | Biomacromolecules 2014, 15, 1696−1706

Biomacromolecules



were washed three times with PBS/0.5% Tween 20, followed by detection of bound biotinylated EGF by measurement of the absorbance at 354 nm. Control films of 4RepCT were treated in the same way. For graph representation, EGF binding is presented as absorbance after EGF binding minus absorbance before EGF binding. Binding of an Enzyme to Functionalized Silk. M4-4RepCT and Z-4RepCT films were washed three times with 20 mM Tris-HCl (pH 8.0) followed by incubation with 80 μL of 0.1 g/L biotinylated xylanase for 15 min. Films were then washed three times with 20 mM Tris-HCl (pH 8.0) before incubation with 100 μL 1% w/v wheat arabinoxylan (Megazyme), pH 6.5 for 20 min at 40 °C. After incubation, each reaction was stopped and color developed, according to the endo-1,4-β-D-xylanase reducing-sugar assay, as described by Megazyme. Finally, 200 μL of each reaction solution was transferred to a 96-well plate and the absorbance at 520 nm measured using a plate reader (Infinite M200, Tecan). Protein Biotinylation. Biotinylation of human blood plasma was performed by mixing 720 μL of human blood plasma (Akademiska Hospital, Uppsala, Sweden), two times diluted in PBS, with 200 μL of 16.4 g/L biotin reagent (EZ-Link Sulfo-NHS-LC-Biotin; Pierce Biotechnology, Rockford, IL), followed by 1 h of incubation at room temperature. Nonreacted biotin reagent was blocked by addition of 40 mM Tris-HCl (pH 8.0)/1% (v/v) Tween 20 to a final volume of 2 mL, and subsequent incubation for 15 min at room temperature. The generated solution of biotinylated human blood plasma thus contained 0.5% Tween 20 and the final dilution of the plasma was five times. To 10 μg of recombinant human VEGF165 (PeproTech) in PBS, 3.2 μg of Sulfo-NHS-LC-Biotin reagent was added, followed by 1 h of incubation at room temperature. Blocking of nonreacted biotin reagent was achieved by the addition of 20 mM Tris-HCl (pH 8.0) containing 0.5% Tween 20 and subsequent incubation 15 min at room temperature. The final concentration of biotinylated VEGF in the solution was 150 nM. To label EGF with chromophoric biotin, 30 μg of recombinant human EGF (Life Technologies) in PBS was mixed with 80 μg of biotin reagent (SureLINK Chromophoric Biotin; KPL, Inc., Gaithersburg, MD) dissolved in dimethylformamide. Biotinylation was allowed to proceed for 2 h at room temperature, after which nonreacted biotin reagent was removed by passing the solution once through a PD SpinTrap G-25 column (GE Healthcare). The final concentration of labeled EGF was 10 μM. To biotinylate 100 μg of recombinantly produced xylanase in PBS, 260 μg of biotin reagent (EZ-Link Sulfo-NHS-LC-Biotin; Pierce Biotechnology, Rockford, IL) was added and incubated for 1 h at room temperature. Residual, nonreacted biotin reagent was blocked by addition of 20 mM Tris-HCl (pH 8.0) followed by 15 min incubation at room temperature. The final concentration of biotinylated xylanase was 5 μM. Release of Captured Molecules by Proteolytic Cleavage. Z4RepCT and ABD-4RepCT films were washed once in PBS and then incubated with 100 μL of 0.44 μM purified rabbit IgG (Sigma) for 30 min and 100 μL of 150 μM human serum albumin (CSL Behring GmbH, Germany) for 15 min, respectively. Films were then washed three times with PBS. Z-4RepCT films were immersed in 50 μL of 10 μM protease 3C (in-house produced), supplemented with 1.3 μM dithiothreitol (DTT) (Sigma), for 22 h at +4°. The corresponding ABD-4RepCT films were instead immersed in 45 μL of 12 μM protease 3C, supplemented with 1.4 μM DTT, for 45 h at +4°. For both Z-4RepCT and ABD-4RepCT, protease 3C cleavage supernatants from 5 to 9 films were pooled, concentrated using Vivaspin sample concentrators (3 kDa molecular weight cutoff) (GE Healthcare) and analyzed on reducing SDS-PAGE. In addition, another set of films for both Z-4RepCT and ABD-4RepCT, without the initial binding of IgG or albumin, was incubated in 50 μL of 10 μM protease 3C, supplemented with 1.3 μM DTT, for 22 h at +4°. Films of Z4RepCT without bound IgG were also used for cleavage with lower DTT concentration (0.3 μM) and shortened cleavage time (2 h), respectively.

Article

RESULTS

Spidroin Affinity Domain Gene Fusion Constructs Can Be Expressed and Purified as Soluble Proteins. A partial spidroin, 4RepCT, was genetically functionalized with four different protein domains. Gene fragments encoding either the IgG-binding domain Z (7 kDa) or C2 (6 kDa), the albuminbinding domain ABD (5 kDa) or the biotin-binding domain M4 (17 kDa) were cloned into an expression vector in-frame with 4RepCT. The four recombinant spider silk fusion constructs (Z-4RepCT, 32 kDa; C2-4RepCT, 31 kDa; ABD4RepCT, 30 kDa; M4-4RepCT, 41 kDa; Figure 1a) were expressed in E. coli and subsequently purified by immobilized metal ion affinity chromatography (IMAC) from whole cell lysate utilizing an incorporated His6-tag. For all constructs, soluble fusion protein was obtained with reasonable purity as judged by SDS-PAGE (Figure 1b). All four silk fusion proteins could be successfully recovered from the soluble fraction. The Silk Fusion Proteins Maintain Their Ability to Form Fiber and Film. After purification, all four silk fusion proteins could self-assemble into films and macroscopic fibers (Figure 1c), demonstrating maintained ability of the 4RepCT part to self-assemble into silk-like materials when expressed as a fusion protein, even though fused to either of the Z, C2, ABD, or M4 domain. Self-assembled films and fibers of the four silk fusion proteins Z-4RepCT, C2-4RepCT, ABD-4RepCT, and M4-4RepCT were visualized by light microscopy, confirming that their macroscopic appearances resembled that of the partial spider silk protein 4RepCT (Figure 1c). To confirm that the N-terminally attached domain was present in the formed fiber, a Z-4RepCT fiber was subjected to N-terminal sequencing, revealing the expected first 20 Nterminal amino acids. IR Spectroscopy Indicates Maintained Secondary Structure of Added Domains. In order to investigate differences in secondary structure of functionalized and nonfunctionalized recombinant spider silk, films and fibers of Z-4RepCT, C2-4RepCT, ABD-4RepCT, M4-4RepCT, and 4RepCT were analyzed by ATR-FTIR, by which it is possible to distinguish α-helical (band position: 1648−1657 cm−1) from β-sheet structure (band position: 1623−1641 cm−1).35 For all silk films, the peak for α-helical and β-sheet signals were present at 1653−1657 cm−1 and around 1630 cm−1, respectively, with the most pronounced α peak seen for Z-4RepCT, followed by ABD-4RepCT, C2-4RepCT, M4-4RepCT, and 4RepCT (Figure 1d, Table S1). For fibers, α-helical and β-sheet signals were present at 1653−1655 cm−1 and 1624−1626 cm−1, respectively. Fibers of Z-4RepCT and ABD-4RepCT showed the most pronounced α-helical peak, while C2-4RepCT, M44RepCT and 4RepCT looked almost the same, with less pronounced peak in the α-helical region. The results reflect the secondary-structure folds of the different domains used for functionalization of 4RepCT, as Z and ABD are both composed of a bundle of three α-helices, while C2 instead is composed of a four-stranded β-sheet crossed diagonally by a single α-helix, and M4 folds as eight β strands. ATR-FTIR spectroscopic analyses thus suggest that the secondary structure contents of the domain Z, C2, ABD and M4 is at least partially maintained after film and probably also fiber formation. Still, a clear β sheet signal, typical of 4RepCT silk, is present for all four silk fusion proteins when in film and fiber form. 1700

dx.doi.org/10.1021/bm500114e | Biomacromolecules 2014, 15, 1696−1706

Biomacromolecules

Article

Figure 2. Selective binding of target molecules to functionalized silk materials. (a) Schematic drawing of the principle for analysis of binding of intended protein targets to films and fibers of functionalized silk. Functionalization domains are either the IgG-binding domains Z or C2, or the albumin-binding domain ABD. After binding of target molecules from a complex sample, the silks are washed before target release. (b) Nonreduced SDS-PAGE of released fractions. Rabbit serum (lane 1) was used as complex sample to bind IgG to Z-4RepCT films and fibers (lane 2 and 3, respectively) and to C2-4RepCT films and fibers (lane 4 and 5, respectively). For analysis of albumin binding to ABD-4RepCT, human blood plasma (lane 8) was used as albumin source, and released fractions from film and fiber are shown in lane 6 and 7, respectively. Bands of IgG and albumin are indicated with arrows. (c) Schematics showing the principle of selective binding of biotinylated DNA to M4-4RepCT silk functionalized with the biotin-binding domain M4. A biotinylated DNA primer was used together with DNA template in a PCR reaction, whereupon the PCR reaction mixture was applied to M4-4RepCT silk films. After binding of the biotinylated PCR product, the silk was washed prior to enzymatic release. (d) Agarose gel of the PCR reaction mixture (left lane) and enzymatically released biotinylated PCR product after binding to M4-4RepCT silk (right lane). Bands of DNA template and biotinylated PCR product are indicated by arrows.

affinity for rabbit albumin36 SDS-PAGE of released fractions from both film and fiber shows albumin bands (∼67 kDa) (Figure 2b), implying functionally folded and exposed ABD domains in ABD-4RepCT films and fibers. To examine the selective binding of biotinylated DNA to M4-4RepCT films, a reaction mixture containing a biotinylated PCR product was used (Figure 2c). After wash and treatment with restriction enzyme, analysis of release fractions from M44RepCT films on an agarose gel showed appearance of a band corresponding to the expected size of the PCR product (∼789 bp) (Figure 2d), highlighting the functionality of the M44RepCT films. These results do not only demonstrate that films and fibers of the silk fusion proteins Z-4RepCT, C2-4RepCT, ABD4RepCT, and M4-4RepCT are able to bind their intended targets, it also shows that the binding is selective. Generally, we observed a higher amount of bound target to the fibers, which could reflect the increased surface area, or that a larger amount of recombinant protein was used in the analysis when fibers where used, due to difficulties in estimating the exact protein content in fiber. However, some unspecific binding of albumin could be seen in released fractions from C2-4RepCT in fiber form. Probably this is due to that in this setup the fibers were less efficiently washed (just moved to new buffer), which could result in unspecific capture of the most abundant proteins. Binding is Distributed All over the Surface of Functionalized Silk Films. The distribution of IgG binding

The Functionalized Silks Selectively Bind the Intended Target Molecules. To explore the functionality and selectivity of binding of domain Z, C2, and ABD in the functionalized silk, films and fibers thereof were subjected to a complex protein solution containing the molecule to which the respective domain has affinity (Figure 2a). The films and fibers were then thoroughly washed before release and analysis of the fractions on SDS-PAGE. Domain Z and C2 have both affinity toward IgG, while domain ABD binds albumin. The interaction between Z/C2 and IgG and that between ABD and albumin can all be broken at low pH (pH ∼ 3). For analysis of IgG binding to Z-4RepCT and C2-4RepCT, rabbit serum was used as complex solution containing IgG. Released fractions run on SDS-PAGE showed IgG (∼150 kDa) bands from Z-4RepCT and C2−4RepCT films and fibers (Figure 2b). This implies that both domain Z and domain C2 are exposed and have maintained ability to bind IgG, even after incorporation into recombinant spider silk via gene fusion. It should be noted that although both domain Z and C2 bind IgG, they have different folds: Z is composed of three α-helices and C2 of a four-stranded β-sheet crossed diagonally by a single αhelix. Despite the differences in secondary structure, they are both able to adopt a functional fold and subsequently bind IgG from serum in both film and fiber format. To analyze binding of albumin to ABD-4RepCT, human blood plasma was used as complex solution containing albumin, since ABD has a high affinity for human albumin but low 1701

dx.doi.org/10.1021/bm500114e | Biomacromolecules 2014, 15, 1696−1706

Biomacromolecules

Article

molecules in films up to approximately 200 pmol. Thereafter the increase in IgG binding starts to level out, possibly due to steric hindrance. Control films of 4RepCT showed low and constant fluorescence signals throughout the entire interval investigated. Also, visualization of the intensity of the different film sets by fluorescence microscopy showed good correlation with the fluorescence measurements (Figure 3b, lower fluorographs). Quantification of IgG Binding to Z-4RepCT Films. Two different approaches were utilized to quantify the binding of IgG to Z-4RepCT films. By comparing the intensities of fluorophore-labeled IgG bound to films with a standard of soluble IgG, a binding capability of 18 ± 0.4 μg IgG/cm2 film was calculated. By instead comparing eluted IgG fractions run on SDS-PAGE with a standard of known IgG amounts, a binding capability of 59 ± 5 μg IgG/cm2 film was estimated. The molar amount of bound IgG per film thus corresponds to 5−15% of the total molar amount of Z-4RepCT molecules within each film. According to Raman microscopy, films of 3 mg/mL protein concentration are approximately 1.5 μm thick. Possibly, the IgG molecules are only able to bind to Z molecules on the surface of the films. The Affinity Domains Are Stable in the Formed Silk. Stability of Z-4RepCT films and fibers was determined by investigation of the functionality after reusage, long-term storage, treatment with urea and sodium hydroxide (NaOH), and autoclaving (only fibers). Both films and fibers, previously used for capture of IgG and thereafter stored for 70 days in buffer, showed the same ability of IgG binding as nonused films and fibers, as revealed by SDSPAGE (data not shown). Harsh treatments with, for example, 8 M urea and/or 0.5−1 M NaOH, are considered necessary for cleaning-in-place protocols of matrices used for chromatography. Therefore, the effect of urea and NaOH treatment of Z4RepCT films and fibers was investigated. Both films and fibers showed maintained ability to bind IgG after treatment with 8 M urea as well as 1 M NaOH, although with slightly reduced capacity (data not shown). Moreover, a Z-4RepCT fiber subjected to sterilization by autoclave treatment did bind IgG from human blood plasma to the same extent as a corresponding untreated fiber (data not shown). Silk with Dual Functions Can Be Obtained by Mixing of Silk Fusion Proteins. To investigate the possibility to obtain silk materials with dual functionality, mixed Z-4RepCT/ ABD-4RepCT film and fibers were prepared by mixing soluble Z-4RepCT with soluble ABD-4RepCT. Light microscopy of such film and fiber revealed macroscopic appearances that resembled that of Z-4RepCT and ABD-4RepCT film and fiber, respectively (Figure 4a). As intended, released fractions from mixed Z-4RepCT/ABD-4RepCT film and fiber, after subjection to a complex sample (plasma), showed simultaneous binding of both IgG and albumin (Figure 4b). From the mixed fiber an additional vague band below 140 kDa could also be seen, although in minority compared to the intended IgG and albumin. Functionalized Silk Can Be Used for Presentation of Growth Factors. By using films or fibers made from the Z4RepCT silk fusion protein it is possible to bind, and thereby decorate the silk material with IgG molecules. As IgG molecules directed toward an almost unlimited repertoire of targets are commercially available, the type of IgG used for decoration of Z-4RepCT materials can be chosen depending on target of interest. Herein, films of Z-4RepCT were used for decoration

to Z-4RepCT and C2-4RepCT films, and also the distribution of albumin and biotin binding to ABD-4RepCT and M44RepCT films, respectively, was investigated using directly or indirectly fluorophore-labeled target molecules, followed by fluorescence microscopy. For all four variants of functionalized silk films a fluorescence signal was seen, as compared to the corresponding control films of 4RepCT protein (Figure 3a).

Figure 3. Target binding is distributed all over the surface of functionalized silk films. (a) Fluorescence micrographs showing functionalized silk fusion films with their respective bound target, visualized by directly or indirectly fluorophore-labeled targets; fluorophore-labeled IgG (Z), IgG-biotin from biotinylated blood plasma (C2), albumin-biotin from biotinylated blood plasma (ABD) and fluorophore-labeled biotin (M4), respectively (upper panels). For biotinylated targets, fluorophore-labeled streptavidin was used for visualization. As controls, 4RepCT films without functionalizations were used (lower panels). Scale bar indicates 1 mm. (b) Effect of varying number of Z-4RepCT molecules on the binding capacity is shown in the graph (blue squares). Films of 4RepCT were used as control (black squares). Fluorescence micrographs of Z-4RepCT films of 80, 200, and 500 pmol are shown below. (n = 2, in duplicates).

These results, again, demonstrates that the functionalization domains in silk fusion films are exposed and able to bind its respective target. Moreover, it shows that binding of the target molecule is distributed all over the surface area of the functionalized silk films. The IgG Binding Is Increased with the Number of Molecules in the Film. In order to test the effect of different amounts of Z-4RepCT molecules in films on the binding of IgG, ten different sets of Z-4RepCT films were prepared, ranging from 7 to 500 pmol of Z-4RepCT molecules per film. The fluorescence signals from binding of fluorophore-labeled IgG to these films indicate an increased IgG binding with increased number of Z-4RepCT molecules in the films (Figure 3b, upper graph). The increase in IgG binding showed an almost linear relationship with the number of Z-4RepCT 1702

dx.doi.org/10.1021/bm500114e | Biomacromolecules 2014, 15, 1696−1706

Biomacromolecules

Article

4RepCT films, nor when excluding the anti-VEGF IgG (Figure 5b). In general, such a two-step approach could be utilized for decoration of Z-4RepCT silk materials with IgG directed toward any suitable target. Another approach for presentation of proteins onto silk fusion materials is to utilize the biotin binding ability of M44RepCT. Herein, films of M4-4RepCT were used for binding of human EGF, prelabeled with chromophoric biotin (Figure 5c). Detection of chromophoric biotin-EGF revealed stronger signals from M4-4RepCT films than from control 4RepCT films (Figure 5d), which implies that the intended binding to M4-4RepCT was successful. Functionalized Silk Can Be Used for Presentation of Active Enzyme. In order to verify that a biotinylated protein can be presented on M4−4RepCT films in a way that preserves the inherent activity of the presented molecule, the enzyme xylanase37 was chosen. Recombinantly produced xylanase was biotinylated and incubated onto M4-4RepCT films as well as on control films (Z-4RepCT). The thereafter measured ability to convert a substrate to colored product (Figure 5e) confirmed that the activity of xylanase was maintained while presented on the M4-4RepCT film (Figure 5f). M4-4RepCT silk fusion materials thus likely allow for additional functionalization of the material with a wide repertoire of biomolecules, since biotinylated proteins, peptides, and nucleotides are easily obtained (and commercially available). Functionalized Silk Allows for Release of Captured Molecules by Proteolytic Cleavage. Two functionalized silk fusion proteins, Z-4RepCT and ABD-4RepCT, have been constructed to contain a recognition site for proteolytic cleavage. The site is situated in-between the silk part

Figure 4. Dual-functionalized silk materials can simultaneously bind two targets. (a) Light micrographs of films and macroscopic fibers prepared by mixing soluble Z-4RepCT with soluble ABD-4RepCT protein. Scale bars indicate 1 mm. (b) The ability of mixed Z4RepCT/ABD-4RepCT film and fiber to simultaneously bind IgG (via the Z domain) and albumin (via the ABD domain) was explored according to the same scheme as described in Figure 2a. Human blood plasma served as both IgG and albumin source. Released fractions were analyzed by SDS-PAGE from mixed film and mixed fiber. Bands of IgG and albumin are indicated with arrows.

with an IgG directed to the human vascular endothelial growth factor (hVEGF) (Figure 5a). Following binding of anti-VEGF IgG and biotinylated hVEGF, fluorophore-labeled streptavidin was used to visualize binding of VEGF to IgG-decorated Z4RepCT films, whereas no binding was seen for control

Figure 5. The use of functionalized silk films for presentation of growth factors and enzyme. (a) Schematics of a two-step approach for presentation of the growth factor VEGF onto the surface of functionalized silk films. In the first step, silk film of Z-4RepCT is decorated with IgG directed toward VEGF, and in the second step biotinylated VEGF is captured onto the film surface by the anti-VEGF IgG. Visualization of bound VEGF is achieved by fluorophore-labeled streptavidin. (b) Fluorescence micrographs of Z-4RepCT and control 4RepCT films with bound anti-VEGF IgG, VEGFbiotin and fluorophore-streptavidin (upper micrographs). An additional set of Z-4RepCT and 4RepCT films were also used, with the first step of anti-VEGF IgG binding omitted (lower micrographs). (c) Schematics of an approach for presentation of biotinylated EGF onto M4-4RepCT. Bound EGF is detected by measuring absorbance of chromophoric biotin (d) Graph showing signal from chromophoric biotin-EGF to M4-4RepCT and 4RepCT control films, respectively (n = 2, in duplicates). (e) Principle for presentation of a biotinylated enzyme on M4-4RepCT silk, where the activity of bound xylanase is determined by measuring its ability to convert substrate to a colored product. (f) Graph showing absorbance of product obtained by biotinylated xylanase bound to M4-4RepCT and Z-4RepCT control films, respectively (n = 1, in triplicates). 1703

dx.doi.org/10.1021/bm500114e | Biomacromolecules 2014, 15, 1696−1706

Biomacromolecules

Article

proteins are assembled into solid films. Furthermore, the fact that also prebound IgG and albumin appeared in the cleavage supernatant demonstrates that proteolytic cleavage can be utilized for release of the bound targets.

(4RepCT) and the respective functionalization domain (i.e., Z or ABD), and is composed of an amino acid sequence recognized by protease 3C. We have herein investigated whether it is possible to release the functionalization domain (Z or ABD) from films of Z-4RepCT and ABD-4RepCT, respectively, by proteolytic cleavage (Figure 6a).



DISCUSSION Herein, we have investigated a system for recombinant production of functionalized silk, based on fusion of different affinity protein domains to the partial spider silk protein 4RepCT. The conserved C-terminal domain of 4RepCT guides self-assembly into silk-like materials that are water-stable.25 This is performed under near-physiological conditions, and the formed fibers do not require any further treatment, a feature which is ideal for incorporation of whole protein domains that require a natively folded structure for their activity. The 4RepCT domain attains a random/helical fold in solution25 but is transformed into a mainly β-sheet structure during the selfassembly into silk-like materials.30 The present study shows that moderately sized protein domains (6−17 kDa) do not hinder this silk assembly process. The silk-like fibers are formed within the same time frame and acquire the typical β-sheet structure, as when the wild-type 4RepCT is investigated alone. It is not completely known what drives the transformation of 4RepCT from random/helical in solution to β-sheet in solid form. However, the process is efficient in physiological-like buffers when a nondenatured spidroin C-terminal domain is included. We have herein seen that all four investigated protein domains fused to 4RepCT maintain their own fold in the fibers, as shown by their retained biological activities, and are thus not dragged along into the β-sheet conversion during self-assembly. Notably, the domains were fused to the N-terminus of 4RepCT, thus opposite to the C-terminal domain, which has been suggested to be responsible for organization of fiber formation.25 Two of the produced silk-fusion variants contain IgG binding domains, derived from staphylococcal protein A (SPA) or streptococcal protein G (SPG), respectively.38 Traditionally, affinity matrices for IgG purification based on these proteins have been based on polysaccharide-based matrices (e.g., Sepharose) or paramagnetic beads. There is a wide range of coupling chemistries available for covalent immobilization of protein domains to such solid supports. Most of these methods utilize accessible functional groups (e.g., −NH2, −COOH, −SH) naturally occurring or added to the protein via genetic engineering. Obviously, it is desired that the protein domains are immobilized in a way that allows the binding regions to be exposed and free to interact with and bind to the target molecules. The performance of the affinity ligand can be affected by multisite attachment due to steric hindrance and unfavorable orientation of the protein domain. Ideally, the immobilization method attaches the protein via defined functional groups on the opposite side from the binding site. Moreover, the protein domain must not be negatively affected by the buffer conditions used during the immobilization procedure. The strategy described in this study to generate a solid fiber-based “resin”, including genetically incorporated and thus robustly anchored affinity domains, could be an attractive alternative to traditional procedures to immobilize such domains. The standard matrix for affinity purification is in the bead format. With the silk fusion proteins also other formats are possible, such as fiber meshes and cast film, which opens up for alternative options such as filters and membranes. The 4RepCT fiber is in itself very stable, both thermally (up to

Figure 6. Release of bound targets from functionalized silk using proteolytic cleavage. (a) Schematic drawing showing the principle for release of bound molecules from functionalized silk films by proteolytic cleavage. The release approach utilizes an incorporated protease recognition site in-between the silk part (4RepCT) and the respective functionalization domain (Z or ABD). (b) Pictures from reducing SDS-PAGE of protease 3C cleavage of Z-4RepCT (left picture) and ABD-4RepCT (right picture) films. Films of Z-4RepCT were treated with protease 3C, either with prebound IgG or without. Resulting supernatants (i.e., released proteins) from the two different conditions are shown in the right picture. In the middle lane, reduced IgG is shown for comparison. Similarly, ABD-4RepCT films, with and without prebound human albumin, were also treated with protease 3C followed by analysis of cleavage supernatants by SDS-PAGE (left picture). In the middle lane, pure human albumin is shown for comparison.

After incubation of films with a solution containing protease 3C, analysis of this supernatant by SDS-PAGE revealed proteolytically released Z and ABD domains, respectively (Figure 6b). Proteolytic cleavage was also performed on Z4RepCT films with bound IgG and on ABD-4RepCT films with bound albumin. Cleavage supernatants then revealed both IgG and Z for Z-4RepCT films, and albumin and ABD for the corresponding ABD-4RepCT films (Figure 6b). This shows that the recognition site for proteolytic cleavage in films of silk fusion proteins is still accessible for cleavage, although the silk 1704

dx.doi.org/10.1021/bm500114e | Biomacromolecules 2014, 15, 1696−1706

Biomacromolecules

Article

267 °C) and chemically.28 In fact, the Z-4RepCT matrices maintain affinity for IgG also after long-term storage, treatment with urea and NaOH as well as after autoclaving, which opens up the possibility of efficient sterilization of the materials. For some applications, such as use of fibers as framework for promoting growth of cells in three dimensions, it would be desirable to utilize fibers decorated with relatively complex mammalian proteins, such as growth factors, not easily produced as 4RepCT fusion proteins in bacterial expression systems. In those cases, a two-step approach using Z-4RepCT and antigrowth factor antibodies can be used instead. As antibodies directed toward an almost unlimited repertoire of proteins are commercially available, the type of IgG used for decoration of Z-4RepCT materials can in those cases be chosen to suit the target of interest. We have herein shown a proof-ofconcept for such functionalization, with IgG directed to the growth factor VEGF, which is widely used in cell culture and tissue engineering applications due to its promotion of angiogenesis.39 Growth factors are often indispensable for advanced cell culturing although they constitute a major cost, especially if needed to be added in the cell culture media. Matrices functionalized with the Z domain could potentially be used as a general strategy for immobilization of specific growth factors, which then would be presented to the adherent cells and remain attached to the matrix upon each media exchange. Alternatively, biotinylated factors can be immobilized to M44RepCT matrices, with similar benefits. If mild and nondenaturing release of the bound factor, or even attached cells, is wanted, the silk fusion strategy used herein allows for the possibility to proteolytically release the added domain and thereby also its bound target. The fact that the cleavage site, situated in-between the silk and the added domain, is accessible for the protease further points at a well-defined fold of the fused domain also when connected to 4RepCT in silk form. The possibility to generate silk films and fibers with multiple functionalities, by simple mixing of different silk fusion proteins, broadens the areas of application for functionalized silk materials. With this new immobilization strategy, it is not only possible to control the stoichiometry of domain versus scaffold, but it is also possible to define a selected ratio between different added domains. For example, this could be useful for construction of scaffolds for advanced tissue engineering, where a defined mixture of cell types should be recruited by selected cell adhesion motifs and growth factors. Notably, the Z and ABD domains have both been subject of protein engineering via combinatorial protein library methodology to generate novel variants of the domains with binding activity to non-native targets, including cell-surface receptors.40−43 The successful incorporation of the wild type Z and ABD domains described in the present study suggests that also such engineered variants, of the same fold and structure, should be possible to incorporate. This could provide the silk materials with a direct binding ability to different targets, as opposed to the indirect strategy relying on antibody-Z/C2 interactions.

the specific affinity of the respective added domain are retained in matrices of all silk fusion constructs.



ASSOCIATED CONTENT

S Supporting Information *

Table S1 is available free of charge via the Internet at http:// pubs.acs.org



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Present Address ⊥

Nofima, P.O. Box 6122, N-9291 Tromsø, Norway.

Author Contributions

The manuscript was written through contributions of all authors. R.J. and M.H. designed and performed experiments, analyzed data and wrote the manuscript. N.T. performed experiments and analyzed data. D.L. performed experiments using preceding M4-constructs. A.R. designed preceding M4constructs. J.J. designed experiments. P.-Å.N. designed experiments and wrote the manuscript. Notes

The authors declare the following competing financial interest(s): A.R., J.J., and M.H. have shares in Spiber Technologies AB, a company that aims to commercialize recombinant spider silk.



ACKNOWLEDGMENTS The authors would like to thank Spiber Technologies AB for providing RGE-4RepCT and 4RepCT protein. A plasmid containing the gene for xylanase was kindly provided by Christophe M Courtin and the protein expressed and purified by Eva-Lena Andersson. Andreas Barth and Nadja Eremina are acknowledged for help with the FTIR measurements, as well as Erik Hermansson for light micrographs of silk films, and Mats Sandgren for valuable discussions. The Swedish Research Council, Vinnova, and Magnus Bergvall’s foundation supported this work.



ABBREVIATIONS ABD, albumin binding domain; VEGF, vascular endothelial growth factor; EGF, epidermal growth factor



REFERENCES

(1) Kluge, J. A.; Rabotyagova, O.; Leisk, G. G.; Kaplan, D. L. Trends Biotechnol. 2008, 26, 244−51. (2) Widhe, M.; Johansson, J.; Hedhammar, M.; Rising, A. Biopolymers 2012, 97, 468−78. (3) Spiess, K.; Lammel, A.; Scheibel, T. Macromol. Biosci. 2010, 10, 998−1007. (4) Murphy, A. R.; Kaplan, D. L. J. Mater. Chem. 2009, 19, 6443− 6450. (5) Chen, J.; Altman, G. H.; Karageorgiou, V.; Horan, R.; Collette, A.; Volloch, V.; Colabro, T.; Kaplan, D. L. J. Biomed Mater. Res. A 2003, 67, 559−70. (6) Sofia, S.; McCarthy, M. B.; Gronowicz, G.; Kaplan, D. L. J. Biomed Mater. Res. 2001, 54, 139−48. (7) Karageorgiou, V.; Meinel, L.; Hofmann, S.; Malhotra, A.; Volloch, V.; Kaplan, D. J. Biomed Mater. Res. A 2004, 71, 528−37. (8) Wang, X.; Kaplan, D. L. Macromol. Biosci. 2011, 11, 100−10. (9) Lu, Q.; Wang, X.; Zhu, H.; Kaplan, D. L. Acta Biomater. 2011, 7, 2782−6.



CONCLUSION Herein we describe a novel strategy for immobilization of functional protein domains. Four recombinant spider silk fusion proteins, Z-4RepCT, C2-4RepCT, ABD-4RepCT, and M44RepCT were successfully expressed, purified and assembled into silk-like materials. The four added affinity domains are of different origin, have different folds, and target different molecules. Still, both the silk-like properties of 4RepCT and 1705

dx.doi.org/10.1021/bm500114e | Biomacromolecules 2014, 15, 1696−1706

Biomacromolecules

Article

(10) Inoue, S.; Kanda, T.; Imamura, M.; Quan, G. X.; Kojima, K.; Tanaka, H.; Tomita, M.; Hino, R.; Yoshizato, K.; Mizuno, S.; Tamura, T. Insect Biochem. Mol. Biol. 2005, 35, 51−9. (11) Hino, R.; Tomita, M.; Yoshizato, K. Biomaterials 2006, 27, 5715−24. (12) Heidebrecht, A.; Scheibel, T. Adv. Appl. Microbiol. 2013, 82, 115−53. (13) Rising, A.; Widhe, M.; Johansson, J.; Hedhammar, M. Cell. Mol. Life Sci. 2011, 68, 169−84. (14) Bini, E.; Foo, C. W.; Huang, J.; Karageorgiou, V.; Kitchel, B.; Kaplan, D. L. Biomacromolecules 2006, 7, 3139−45. (15) Wohlrab, S.; Muller, S.; Schmidt, A.; Neubauer, S.; Kessler, H.; Leal-Egana, A.; Scheibel, T. Biomaterials 2012, 33, 6650−9. (16) Huang, J.; Wong, C.; George, A.; Kaplan, D. L. Biomaterials 2007, 28, 2358−67. (17) Mieszawska, A. J.; Fourligas, N.; Georgakoudi, I.; Ouhib, N. M.; Belton, D. J.; Perry, C. C.; Kaplan, D. L. Biomaterials 2010, 31, 8902− 10. (18) Mieszawska, A. J.; Nadkarni, L. D.; Perry, C. C.; Kaplan, D. L. Chem. Mater. 2010, 22, 5780−5785. (19) Wong Po Foo, C.; Patwardhan, S. V.; Belton, D. J.; Kitchel, B.; Anastasiades, D.; Huang, J.; Naik, R. R.; Perry, C. C.; Kaplan, D. L. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 9428−33. (20) Gomes, S.; Numata, K.; Leonor, I. B.; Mano, J. F.; Reis, R. L.; Kaplan, D. L. Biomacromolecules 2011, 12, 1675−85. (21) Gomes, S. C.; Leonor, I. B.; Mano, J. F.; Reis, R. L.; Kaplan, D. L. Biomaterials 2011, 32, 4255−66. (22) Askarieh, G.; Hedhammar, M.; Nordling, K.; Saenz, A.; Casals, C.; Rising, A.; Johansson, J.; Knight, S. D. Nature 2010, 465, 236−8. (23) Hagn, F.; Eisoldt, L.; Hardy, J. G.; Vendrely, C.; Coles, M.; Scheibel, T.; Kessler, H. Nature 2010, 465, 239−42. (24) Ayoub, N. A.; Garb, J. E.; Tinghitella, R. M.; Collin, M. A.; Hayashi, C. Y. PLoS One 2007, 2, e514. (25) Hedhammar, M.; Rising, A.; Grip, S.; Martinez, A. S.; Nordling, K.; Casals, C.; Stark, M.; Johansson, J. Biochemistry 2008, 47, 3407−17. (26) Stark, M.; Grip, S.; Rising, A.; Hedhammar, M.; Engstrom, W.; Hjalm, G.; Johansson, J. Biomacromolecules 2007, 8, 1695−701. (27) Spider silk proteins and methods for producing spider silk proteins WO 2007/078239. (28) Hedhammar, M.; Bramfeldt, H.; Baris, T.; Widhe, M.; Askarieh, G.; Nordling, K.; Aulock, S.; Johansson, J. Biomacromolecules 2010, 11, 953−9. (29) Widhe, M.; Bysell, H.; Nystedt, S.; Schenning, I.; Malmsten, M.; Johansson, J.; Rising, A.; Hedhammar, M. Biomaterials 2010, 31, 9575−85. (30) Widhe, M.; Johansson, U.; Hillerdahl, C. O.; Hedhammar, M. Biomaterials 2013, 34, 8223−34. (31) Nilsson, B.; Moks, T.; Jansson, B.; Abrahmsen, L.; Elmblad, A.; Holmgren, E.; Henrichson, C.; Jones, T. A.; Uhlen, M. Protein Eng. 1987, 1, 107−13. (32) Sauer-Eriksson, A. E.; Kleywegt, G. J.; Uhlen, M.; Jones, T. A. Structure 1995, 3, 265−78. (33) Kraulis, P. J.; Jonasson, P.; Nygren, P. A.; Uhlen, M.; Jendeberg, L.; Nilsson, B.; Kordel, J. FEBS Lett. 1996, 378, 190−4. (34) Wu, S. C.; Wong, S. L. J. Biol. Chem. 2005, 280, 23225−31. (35) Barth, A. Biochim. Biophys. Acta 2007, 1767, 1073−101. (36) Jonsson, A.; Dogan, J.; Herne, N.; Abrahmsen, L.; Nygren, P. A. Protein Eng., Des. Sel. 2008, 21, 515−27. (37) Pollet, A.; Vandermarliere, E.; Lammertyn, J.; Strelkov, S. V.; Delcour, J. A.; Courtin, C. M. Proteins 2009, 77, 395−403. (38) Grodzki, A. C.; Berenstein, E. Methods Mol. Biol. 2010, 588, 33− 41. (39) Nomi, M.; Miyake, H.; Sugita, Y.; Fujisawa, M.; Soker, S. Curr. Stem Cell Res. Ther. 2006, 1, 333−43. (40) Lofblom, J.; Feldwisch, J.; Tolmachev, V.; Carlsson, J.; Stahl, S.; Frejd, F. Y. FEBS Lett. 2010, 584, 2670−80. (41) Nilvebrant, J.; Astrand, M.; Lofblom, J.; Hober, S. Cell. Mol. Life Sci. 2013, 70, 3973−85.

(42) Nord, K.; Gunneriusson, E.; Ringdahl, J.; Stahl, S.; Uhlen, M.; Nygren, P. A. Nat. Biotechnol. 1997, 15, 772−7. (43) Nygren, P. A. FEBS J. 2008, 275, 2668−76.

1706

dx.doi.org/10.1021/bm500114e | Biomacromolecules 2014, 15, 1696−1706