Graphene oxide/mussel foot protein composites for high-strength and ultra-tough thin films

Graphene oxide (GO)-based composite materials have become widely popular in many applications due to the attractive properties of GO, such as high strength and high electrical conductivity at the nanoscale. Most current GO composites use organic polymer as the matrix material and thus, their synthesis suffers from the use of organic solvents or surfactants, which raise environmental and energy-consumption concerns. Inspired by mussel foot proteins (Mfp) secreted by the saltwater mussel, Mytilus galloprovincialis and by recent advances in microbial protein production, we developed an aqueous-based green synthesis strategy for preparing GO/Mfp film composites. These GO/Mfp films display high tensile strength (134–158 MPa), stretchability (~ 26% elongation), and high toughness (20–24 MJ/m3), beyond the capabilities of many existing GO composites. Renewable production of Mfp proteins and the facile fabrication process described provides a new avenue for composite material synthesis, while the unique combination of mechanical properties of GO/Mfp films will be attractive for a range of applications.


Scientific Reports
| (2020) 10:19082 | https://doi.org/10.1038/s41598-020-76004-6 www.nature.com/scientificreports/ however, when interacting with GO nanosheets, proteins tend to change conformation to redistribute amino acid residues, adopting a new set of entropically favored interactions, such as hydrogen bonding and electrostatic interactions. In some cases, such proteins undergo denaturation and aggregation in the presence of GO, which would lead to undesirable mechanical properties 26 . One unique class of proteins that has not been fully explored with regard to composite synthesis is the intrinsically disordered family of mussel foot proteins (Mfp). Naturally secreted by the marine mussel, Mfp utilize a wide range of molecular interactions to bond to hydrophilic surfaces such as rocks, metals, and glass, as well as hydrophobic surfaces, such as plastics. These strong interactions with surfaces are achieved largely in part due to the side chain of the non-canonical amino acid, 3,4-dihydroxyphenylalanine (DOPA). More interestingly, the tight interaction between Mfp and various surfaces take place underwater. If this aqueous-based molecular bonding can be used to prepare composite materials, it will provide a low-energy, environmentally-friendly process for composite fabrication, which would otherwise involve high temperature processes or organic solvents compatible with organic polymers.
Herein, we describe a facile approach for fabricating GO composite films using Mfp through an aqueousbased processing route. The Mfp were recombinantly synthesized by genetically engineered E. coli with a subsequent post-translational modification step for the generation of DOPA residues 27 . We further developed an aqueous procedure to fabricate free-standing GO/Mfp films. These composite films consisted of up to 20 w/w% of Mfp and displayed high strength and toughness, comparable to or even stronger than previously reported GO composites.

Results
Synthesis of graphene oxide/mussel foot protein composite films. We hypothesize that the unique chemistry of Mfp allows the flexible protein chains to form extensive interactions with GO nanosheets through hydrophobic interactions, π-π stacking (via Mfp aromatic side chains), cation-π interactions, and hydrogen bonding via DOPA-alcohol, DOPA-carboxylate, and bi-DOPA pairs (Fig. 1a,b) [28][29][30][31] . To promote a robust network of interactions between GO and Mfp, we chose Mfp5 from Mytilus galloprovincialis due to its high DOPA www.nature.com/scientificreports/ content and overall positive charge, which can reduce Mfp-Mfp interactions in low ionic strength solvents while promoting electrostatic interactions with the negatively-charged GO nanosheets 32 . Mfp can accomplish the formation of an interaction network in water-rich solvents, which eliminates the need to use harsh organic solvents.
Although some Mfp3 peptides have been shown to form coacervates in aqueous solutions 33 , Mfp5 precipitates at neutral pH and basic conditions, probably due to its higher molecular weight 27 . However, at acidic pH levels, our Mfp5 remains soluble. We thus used pH 4.5 acetate buffer to prepare the Mfp5/GO mixtures. The acidic condition also helps prevent the catechol groups from oxidizing. Additionally, we hypothesized that a longer Mfp chain length will participate in extensive molecular interactions with GO, therefore strengthening the composite's molecular network and resulting in better film mechanical properties 27,34,35 . Thus, Mfp5 (3) , a synthetic protein containing three consecutive repeats of Mfp5 was also used in this study 27,36 . Both Mfp5 and Mfp5 (3) were microbially synthesized by genetically engineered E. coli, purified, and enzymatically modified to convert tyrosine residues to DOPA using our previously developed method 27 . To fabricate the GO/Mfp composites, different ratios of GO and Mfp were mixed to identify the condition, in which both components stayed soluble in aqueous solution. We found that at a GO-to-Mfp ratio of 5:1 and a pH of 4.5, the solution was homogenous after sonication. The solution was then vacuum filtered and dried to obtain free-standing thin films (Fig. 1c). To evaluate whether longer periods of stacking and compacting could lead to mechanically stronger films, different filtration times were used to compare composite films. All of our composite films were black in color and had metallic lusters (Fig. 1d), similar to GO/polymer composites examined in previous studies 37,38 .

Structural characterization of composite films.
Micro-scale structures of the composite films were first studied using scanning electron microscopy (SEM). The cross-sections of the GO, GO/Mfp5 and GO/ Mfp5 (3) films all show dense packing with minimal voids and cracks, which would otherwise negatively affect mechanical strength via crack propagation (Fig. 2). Both GO/Mfp5 and GO/Mfp5 (3) films have comparable thicknesses of 25.1 ± 2.6 and 24.8 ± 4.9 μm, respectively; however, these films were both slightly thinner than pure GO films (27.8 ± 2.9 μm, P < 0.05). In addition, we found that filtration time did not affect overall film thicknesses (Supplementary Figure S1). X-ray diffraction (XRD) analysis was conducted to examine the extent of nanosheet alignment within the films and interlayer spacing between nanosheets. Previous studies of pure GO films showed characteristic peaks at 2θ values of 10-11 12,39 . Similar peaks at 10.6° and 11.8° were observed in our GO/Mfp5 and GO/Mfp5 (3) films, respectively (Fig. 3), suggesting good alignment of GO nanosheets in the composite films. The average interplanar distances, d, between GO nanosheets were estimated using the measured 2θ values. The d values of GO/Mfp5 and GO/Mfp5 (3) were 8.1 Å and 7.4 Å, respectively, and were slightly lower than the 8-9 Å interplanar distance of pure GO measured here and in previous studies [40][41][42] . Decreased interplanar distance is likely to be caused by the Mfp's natural ability to displace water molecules hydrating intercalated space between GO nanosheets, a layer that has been shown to be as thick as 1.2 nm 32,[43][44][45][46] . This is consistent with the decreasing trend in film thicknesses as observed in SEM and consistent with the decreased interplanar distance as molecular weight of Mfp increases.
Thermogravimetric analysis (TGA) was further conducted to confirm Mfp incorporation in our composite films, as well as to test their thermal stabilities. Our pure GO film exhibited similar thermal degradation kinetics as other studies, for which there is a rapid decrease in mass at around 200 °C due to evaporation of associated water molecules (Supplementary Figure S2) 9,47 . Pure lyophilized proteins, which were used as single-component controls, degraded rapidly at temperatures higher than ~ 150 °C (Supplementary Figure S3). However, the proteins that were incorporated into our GO films do not contribute to a higher rate of degradation in the overall film and remain stable within the range of 200-550 °C, which is also beneficial for many practical applications. Fourier transform infrared (FTIR) spectra of our pure GO and GO/Mfp films closely resemble that of pure graphene oxide (Supplementary Figure S4). Additionally, the FTIR spectra show that there were no peak shifts or significant changes in peak intensities throughout the measured wavenumber range, suggesting the lack of covalent interactions between GO and Mfp as well as a lack of new protein secondary structures induced during composite formation. Rather, the Mfp molecules bind on and between the GO nanosheets through non-covalent interactions. shaped strips of as-synthesized films. We obtained the ultimate tensile strength, toughness, and Young's modulus of each film from the measured stress-strain curves (Fig. 4a,e). Compared to pure GO film, GO/Mfp5 composite film displayed a 1.9-and 4.1-fold higher tensile strength and toughness, respectively, and 40% decrease in Young's modulus. Thus, Mfp is participating in the formation of an extensive interaction network, allowing the films to withstand higher stress before fracture and to absorb a higher amount of energy before deformation. We also observed that incorporating the higher molecular weight Mfp5 (3) further enhanced ultimate tensile strength by 2.3-fold and decreased Young's modulus by 14% with respect to the pure GO film control, while maintaining toughness at a similar level to GO/Mfp5 films (Fig. 4b-d). Consistent with our design, the higher tensile strength of the GO/Mfp5 (3) compared to the GO/Mfp5 film indicates the formation of a stronger protein-GO interaction network. Further, the observed higher strain of the GO/Mfp composite compared to that of GO film suggests that under tensile stress, the Mfp protein chains are straightened, sliding along the GO nanosheets, thereby absorbing energy and contributing to a higher film toughness. Additionally, we fabricated GO/Mfp5 films with  www.nature.com/scientificreports/ varied filtration times. We hypothesized that longer filtration times could potentially increase nanosheet alignment and promote tighter packing of the nanosheets. Indeed, films filtered for 4 days exhibited a 1.5-fold higher tensile strength and 1.9-fold higher toughness, than those of films filtered for 3 days. When films were filtered for 5 days, no further enhancement on film strength and toughness was observed, suggesting an optimal filtration time was reached (Fig. 4f-h).
Our approach for synthesizing our GO composites allowed us to create films that were significantly tougher than polymeric materials while comparable in tensile strength to many types of metals, metal alloys, and ceramics, as well as other GO composites synthesized with divalent ions or synthetic polymer crosslinkers (Fig. 5). The toughness of our films is 1-2 orders of magnitude higher than those of similar GO composites. The unique Mfp matrix endowed the GO/Mfp films with the ability to absorb a large amount of energy to deform without fracturing. This high material toughness may open new applications in protection and energy absorption (Fig. 5a). It is also important to note that that the GO/Mfp films also have lower Young's moduli, which make them less stiff and more flexible due to the use of soft protein matrices (Fig. 5b) 10,37,38,[48][49][50] . Such a combination of toughness and flexibility is highly desired in environmental GO applications, for which there are many [51][52][53] . When subject to reducing conditions, our film was moderately conductive, exhibiting conductivities of 0.6-1.5 S/m. If further processed and optimized, our film, given its flexible nature could potentially be used in bioelectronic applications, for example, as a wearable device that can convert and transmit physical resistance into electrical signals 54 . Such property combinations underpin GO/Mfp films as a unique material, demonstrating properties that are not possible through existing GO composite strategies.

Conclusions
In summary, we report a new type of GO composite material using Mfp as a novel matrix. Mfp matrices were microbially synthesized from renewable feedstock, and the composite film was made through an environmentally-friendly, aqueous-based process route. As shown, GO/Mfp composites integrate the unique chemistry of Mfp, as well as the versatility of GO. These GO/Mfp films have low stiffness, high tensile strength, and ultra-high toughness, comparable to or exceeding previously reported GO-based materials 55 . The simple green synthesis process will also open new avenues for composite preparation, and when coupled with unique mechanical properties of the GO/Mfp, material adoption is thus attractive for a variety of applications.

Chemicals and reagents. Unless otherwise noted, all chemicals and reagents were obtained from Sigma
Aldrich (Saint Louis, MO, USA). Plasmid purification and gel extraction kits were obtained from iNtRON Biotechnology (Seoul, South Korea). Restriction enzymes and DNA ligase were purchased from Thermo Fisher Scientific (Austin, TX, USA). Graphene oxide (GO) nanosheets were synthesized from graphite in solution using a modified Hummers' method [56][57][58] . Previous studies using similar modified Hummers' methods resulted in 54-59% oxygenated carbons out of total carbon, according to XPS 12,57 . Hydrogen peroxide was added dropwise to reduce residual permanganate. The graphite oxide was thoroughly washed multiple times with pure DI water to eliminate residual strong organic acids and other Hummers' reagents. The, graphite oxide was dried, resuspended in DI water, and exfoliated into GO nanosheets. Solutions of GO nanosheets were diluted to a concentration of 100 ppm, confirmed by dry weight.

Synthesis and Purification of Mfp.
The two proteins used in this study (Mfp5 and Mfp5 (3) ) were designed and recombinantly expressed using the methods from our previous study (Supplementary Table S1-S4) 27 . Mfp5 was directly expressed in E. coli strain BL21 (DE3). Mfp5 (3) was post-translationally spliced together in vitro from an Mfp5 protein with a C-terminal Cfa N split intein domain and an Mfp5 (2) protein with an N-terminal www.nature.com/scientificreports/ Cfa C split intein domain. Both Mfp5 and Mfp5 (3) proteins were purified using nickel affinity chromatography columns and were reacted with mushroom tyrosinase to convert tyrosine residues to DOPA residues. After purification and conversion of tyrosine residues, proteins were finally dialyzed in 0.5% v/v acetic acid.

Synthesis of GO/Mfp film composites. Approximately 50 mL of GO solution (containing ~ 5 mg GO)
was mixed with 1 mg of either Mfp5 or Mfp5 (3) protein suspended in 0.5% acetic acid (or equal volume of 0.5% acetic acid for a pure GO control film). If necessary, pH was adjusted to 4.5 with additional acetic acid. The mixture is sonicated on ice for a total of ~ 2 h cycling between 6 s on and 4 s off. After sonication, the solution was poured on top of a PES support membrane (Sterlitech, Kent, WA) inside an Advantech glass microanalysis filter holder (Cole-Palmer, Vernon Hills, IL). The solution was passed through the membrane using vacuum filtration. After filtration, the GO film was soft-baked at 37 °C for at least one hour, then peeled off the PES membrane.
Thermogravimetric analysis (TGA). A 100 μL platinum-high temperature pan (TA Instruments, New Castle, DE) was tared and GO film and purified lyophilized protein crosslinker samples were weighed prior to heating using a Q5000 IR thermogravimetric analyzer (TA Instruments). All measurements were conducted in nitrogen (AirGas, Radnor, PA) at a purge flow rate of 25 mL/min over a temperature range of 30-750 °C with a ramp rate of 10 °C/min.

Scanning electron microscopy (SEM).
Films were mounted on a stainless steel sample holder using black carbon tape as an adhesive backing. Samples were coated with 10 nm Au using a Leica EM ACE600 high-vacuum sputter coater (Leica Microsystems, Wetzlar, Germany). The films were imaged with the Nova NanoSEM 230 field emission scanning electron microscope (Field Electron and Ion Company, FEI, Hillsboro, Oregon).
Fourier transform infrared spectroscopy (FTIR). FTIR spectra of the samples were collected using a Thermo Nicolet Nexus 470 (Thermo Scientific, Waltham, MA) following previous methods 59,60 . Specifically, spectra were acquired between wavenumbers of 500 cm −1 and 4000 cm −1 . Peaks were assigned and compared to specific bonds according to previous studies of similar materials 12,24,55,61-63 .
Mechanical testing. Mechanical properties, such as ultimate tensile strength and toughness, were measured using an MTS Criterion Model 41 universal test frame fitted with a 25 N load cell (MTS Systems Corporation, Eden Prairie, MN). Tests were conducted at a crosshead speed of 2.5 mm/min. The maximum force at fracture was divided by the cross-sectional area of the film strip to determine the ultimate tensile strength.