Introduction

The performances of functional biomaterials are defined from nano to micro scale by their chemical composition, structural morphology, penetrability, crystallinity and hierarchical structural design1. Over the years biomaterials have been developed for a variety of applications, such as medical therapy, ecological monitoring, food engineering as well as flexible circuit fabrication.2,3,4,5,6.

Microbes form biofilms by the secretion and assembly of various extracellular matrix components such as polysaccharides, proteins, and DNA5,7. Among the various polysaccharides secreted by bacteria, cellulose produced by bacteria such as Acetobacter, Achromobacter, Pseudomonas, Rhizobium, Salmonella and Azotobacter has garnered much attention8,9. Interests in bacterial cellulose (BC) develops in material science due to its unique properties including high hydrophilicity, high flexibility and optical transparency, good mechanical strength, excellent biocompatibility, and environmental degradability10,11.

Bacterial cellulose comprises of a highly porous nanoscale network of β-(1-4)-glucan chains linked by glycosidic bonds. These chains, secreted through pores on the bacterial cell membranes, fuse together to form microfibrils with 40–60 nm width in the form of ribbon-like bundles. Microfibrils form the floating cellulose matrix produced by the aerobe where there is higher oxygen concentration at the air-water interface12,13,14. Bacterial cellulose is chemically pure and highly crystalline as compared to plant cellulose and is devoid of lignin and hemicellulose, ensuring easy processability9,15. Moreover, bacterial cellulose can be modified in situ during fermentation to produce various forms of structures such as thin fibers, films, or microspheres16,17,18. The presence of a multiple hydroxyl groups on the bacterial cellulose backbone provides chemical handles for its functionalization to produce bacterial cellulose-based hybrid functional materials19,20. These highly packed and structured properties give BC its uniqueness for use as a scaffold in material science.

Due to its highly packed nature, bacterial cellulose is typically functionalized through chemical treatment or physical absorption. Chemical approaches to modifying the hydroxyl groups include NaOH-Urea/thiourea, ionic liquids, TEMPO mediated oxidation, LiOH, N,N-dimethylacetamide/LiCl and N-methylmorpholine-N-oxides (NMMO)21,22,23,24,25,26,27. Due to the strong intermolecular hydrogen bonds in bacterial cellulose, its processing relies on a limited selection of solvents, which are usually toxic, and harsh chemicals that pose environmental risks impacting recyclability of the products and disposal28,29. Modification techniques based on physical absorption can aid in minor modification of bacterial cellulose with low reaction efficiency. Weak physical interactions between the adsorbed functional moieties and the bacterial cellulose matrix also limit the utility of this approach29,30.

Bacterial cellulose is produced by bacterial workhorses, which continuously convert organic carbon feedstock into a highly packed and dense extracellular scaffold making it challenging to access the inner part of the matrix. To circumvent this limitation, we propose a bioengineering strategy inspired from nature where the interactions between different microbial ecosystems give rise to a multi-species biofilm31. Bioengineering approaches have been previously employed to produce functional biomaterials32,33,34 and have given rise to the coined term, Engineered Living Materials (ELMs)35. ELMs are defined as engineered materials composed of living cells (usually microbes) that are not only responsible for the formation and assembly of the material but also modulate the performance of the material by providing functional capabilities32,35,36. A recent study reported a co-culture strategy to enhance the mechanical properties of bacterial cellulose. The exopolysaccharide (EPS) secreted by E. coli incorporated well into the cellulose matrix and improved its mechanical properties37. In this work, we developed a co-culture strategy for the production of bacterial cellulose by Komagataeibacter hansenii and its functionalization by engineered E. coli bacteria that secrete amyloid curli protein subunit, CsgA, fused to the M6A peptide (curli-M6A)38. The curli fibers are known to bind to bacterial cellulose39. The M6A peptide is derived from the magnetosome of the magnetotactic bacteria, Magnetospirillum magneticum AMB-140,41. We propose that co-culturing bacterial cellulose scaffold with the engineered E. coli will result in a single-step functionalization of bacterial cellulose membranes capable of magnetite nanoparticle nucleation. The expected final product is magnetic bacterial cellulose films.

A top-down approach employs the application of mechanical or strong chemicals to break down the hydrogen bonds present in bacterial cellulose films to form micro/nanofibrils. In contrast, we present a bottom-up dual bacteria functionalization method that circumvents the limitations posed by the existing physical and chemical methods. Because the curli-M6A-producing E. coli become trapped in the cellulose matrix, they enable magnetite formation in the otherwise inaccessible portions of the bacterial cellulose inner matrix.

Results

Bioengineered BC-protein hybrid films

Bacterial cellulose pellicles are produced at the air-water interface of the culture medium using K. hansenii ATCC 53582 and glucose as carbon source. The bacterial cellulose pellicles are introduced into growing E. coli cultures producing curli-M6A. We chose to form the bacterial cellulose pellicles first because co-culturing K. hansenii with E. coli resulted in reduced cellulose yield and compromised the structure of the scaffold. The observation is consistent with the incompatible growth rates of K. hansenii (Td = 8 h) and E. coli (Td = 20 min). Introducing bacterial cellulose films into E. coli cultures, instead of co-culturing them, resulted in retained functionalities of both components of the hybrid system, the cellulose scaffold and the curli-M6A protein. A schematic of the process is presented in Fig. 1.

Fig. 1: Schematic of functional hybrid cellulose film fabrication.
Fig. 1: Schematic of functional hybrid cellulose film fabrication.The alternative text for this image may have been generated using AI.
Full size image

K. hansenii 53582 is used to produce cellulose films. The cellulose films are transferred to cultures containing recombinant E. coli producing curli-M6A to functionalize the cellulose films. Hybrid cellulose films thus formed are subjected to magnetite nanoparticle formation. Particles with uniform size and shape are produced in the presence of the hybrid BC(curli-M6A) films.

The obtained bacterial cellulose-protein hybrid film is imaged with confocal laser scanning microscopy (CLSM) to check for the incorporation of the amyloid protein in the hybrid films. All bacterial cellulose films – pristine BC, BC(curli) and BC(curli-M6A) – are soaked in a solution containing the amyloid protein specific dye, curcumin (25 µM), for 15 min. The hybrid films, BC(curli) and BC(curli-M6A) displayed green fluorescence due to the binding of the dye with the amyloid protein (Supplementary Video 1a and 1b). The absence of green fluorescence in pristine BC films, further confirms that the green fluorescence observed in hybrid films is a result of curli protein incorporation (Fig. 2).

Fig. 2: Incorporation of curli protein in the cellulose network.
Fig. 2: Incorporation of curli protein in the cellulose network.The alternative text for this image may have been generated using AI.
Full size image

(Top row) pristine BC; (Middle row) BC(curli); (Bottom row) BC(curli-M6A). Left, middle, and right columns are different angles of the same matrices presented in each row. All films were soaked in 25 µM of amyloid protein specific dye, curcumin, for 15 min. Images of the BC films are observed with CLSM green channel. The scale bar corresponds to 100 µm.

Magnetite nanoparticle nucleation in hybrid films and their surface morphologies

Upon 20 h incubation with E. coli induced to secrete curli-M6A, bacterial cellulose is harvested and incubated in 30 mM iron (ii) heptahydrate for 20 min. The iron-soaked bacterial cellulose hybrid is subsequently freeze-dried to form dry bacterial cellulose films. The nucleation of iron oxide nanoparticles is observed in all bacterial cellulose films (Fig. 3a). It is evident that BC(curli-M6A) films resulted in more integration of iron oxide nanoparticles as indicated by the uniform intense black color compared to pristine BC and BC(curli). The pristine BC films show minimal nanoparticle formation, due to the limited distribution of pores and the availability of functional groups for nucleation and growth. Observation of bacterial cellulose morphologies using scanning electron microscopy (SEM) confirms the higher density of magnetite nanoparticles present in the BC(curli-M6A) films compared to pristine BC and BC(curli) films (Fig. 3b). The structural morphology of bacterial cellulose fibers in the hybrid films is not affected by the introduction of curli proteins. The nanofiber thickness is 30–40 nm, which is similar to that of pristine BC films. We observe aggregation of nanoparticles and/or large particles reflected by the wide particle size distribution with a mean of 82.4 nm on BC and BC(curli) films (Supplementary Fig. 1). The presence of M6A peptide domains in the hybrid BC films appears to control the magnetite nucleation resulting in nanoparticles with a narrow size distribution of 47.0 nm (Supplementary Fig. 1). The presence of PEG as a surfactant and capping agent is important in preventing the magnetite particles from aggregating while maintaining an optimum size and shape (Supplementary Fig. 2). The results confirm that the bioengineering approach is successful in functionalizing bacterial cellulose by uniform nucleation of magnetite nanoparticles across the matrix.

Fig. 3: Images of cellulose films and their morphologies under scanning electron microscopy.
Fig. 3: Images of cellulose films and their morphologies under scanning electron microscopy.The alternative text for this image may have been generated using AI.
Full size image

a (left to right) Singapore 10-cent mint as a reference, pristine BC film, BC film loaded with 30 mM of iron salt, hybrid BC-curli without M6A peptide film displaying spotty nucleation of magnetite nanoparticles, and dense black hybrid BC(curli-M6A) film incorporating magnetite nanoparticles; b Characteristic fibrous network present in pristine BC film, BCFe and BC(curli)Fe films displayed large aggregates of iron particles of various shapes and sizes, BC(curli-M6A)Fe film displayed magnetite nanoparticles with controlled size and shape. The scale bars on the SEM images correspond to 100 nm.

Physical characterizations of functional BC films

FT-IR spectroscopy is performed to identify the chemical incorporation of the magnetite nanoparticles in BC (Fig. 4a). The absorption bands at 1429, 1160 and 1050 cm−1 correspond to the H-C-H plane bending, -OH wagging vibration and C-O-C pyranose backbone ring vibration42,43. The IR spectrum of PEG coated magnetite nanoparticles contain absorption bands at 580, 870 and 1100 cm−1 assigned to Fe-O vibration in magnetite, C-H rocking, and C-O stretch in ether groups of PEG chains, respectively44,45. A small absorption band at 630 cm−1, represents the Fe-O bond vibrations found in maghemite, caused by partial oxidation of magnetite46.

Fig. 4: Physical characterizations of BC films.
Fig. 4: Physical characterizations of BC films.The alternative text for this image may have been generated using AI.
Full size image

a FTIR spectra of BC (black), BCFe (red), BC(curli)Fe (blue) and BC(curli-M6A)Fe (magenta), displaying a structural change. New peaks are observed at 580 cm−1 indicating the incorporation of magnetite nanoparticle in hybrid BC films. b XRD spectra of BC (black), BCFe (red), BC(curli)Fe (blue) and BC(curli-M6A)Fe (magenta) display change in crystallinity due to the introduction of nanoparticles. c TG and d DTG analysis of the various BC films. The lowest endothermic peaks are 280°C for BCFe (red), BC(curli-M6A)Fe (magenta) films and 275°C for BC(curli)Fe (blue) films. For pristine BC it is observed at 300°C. The crystallization temperature of the hybrid films is lower than pristine BC films by 20–25°C due to the incorporation of the proteins and nanoparticles. e XPS analysis of Fe 2p scan shows higher intensity of magnetite nanoparticles for BC(curli-M6A)Fe (magenta) films. f VSM curves of various BC films. Pristine BC (black) is diamagnetic, BCFe (red) and BC(curli)Fe (blue) are paramagnetic while BC(curli-M6A)Fe (magenta) is found to be superparamagnetic and displays higher saturation magnetization.

The X-ray diffraction (XRD) patterns of pure BC and magnetic BC films are shown in Fig. 4b. The pristine BC shows all the characteristic crystalline phases corresponding to the three main diffraction peaks at 14.7° (100), 16.8° (110) and 22.7° (010), respectively47,48. All characteristic Bragg’s peaks of magnetite (Fe3O4) nanoparticles are present in the hybrid bacterial cellulose films whereas, for BC films loaded with nanoparticles, either some peaks are absent (37.7° (222), 58.0° (511) and 63.7° (440)), or the intensity of the peaks is too low to detect.

The thermal stability and thermal decomposition of the bacterial cellulose -protein hybrid films are analyzed with TGA. Thermal decomposition begins at 270–280°C for BCFe (red), BC(curli)Fe (blue) and BC(curli-M6A)Fe (magenta) films and 300°C for pristine BC film (black) (Fig. 4c). The onset of degradation at 240°C in the magnetic membranes corresponds to the desorption and evaporation of the PEG coating with loss of 30–40 wt%. The differential thermogravimetric analysis (DTG) showed that the lowest endothermic peaks of BCFe (red), BC(curli)Fe (blue) and BC(curli-M6A)Fe (magenta) films are 275°C, and 350°C for BC (black) (Fig. 4d). The maximum weight loss observed is due to the complete degradation of β−1,4-glycosidic bond. The two sharp endothermic peaks correspond to the degradation of the crystalline regions present in pure bacterial cellulose and the transformation of iron oxides during combustion process (Fig. 4d).

X-ray photoelectron spectroscopy (XPS) was used to analyze the surface chemistry of the BC films. The Fe 2p core level spectra of BC (black), BCFe (red), BC(curli)Fe (blue) and BC(curli-M6A)Fe (magenta) are shown in Fig. 4e. The Fe 2p peaks are located at the same binding energy level for all the films, but with a higher intensity for BC(curli-M6A)Fe (magenta) films, confirming the higher nucleation and growth of magnetite nanoparticles49.

The magnetic moment (M) is measured using VSM at various field strengths (H) at room temperature for all the films (Fig. 4f). The M vs. H plot clearly shows that the pristine BC film does not demonstrate any magnetic behavior as it is a diamagnetic material. BCFe (red) and BC(curli)Fe (blue) films loaded with magnetite nanoparticles exhibit paramagnetic behavior. BC(curli-M6A)Fe (magenta) film is superparamagnetic due to the high residual saturation magnetization (Supplementary Video 2a and 2b). Superparamagnetic behavior is characterized by the lack of hysteresis loop and low/no coercivity. The saturation magnetization (Ms) of BC(curli-M6A)Fe (magenta) film is 3-4 times greater than other films. The observed high Ms of 40 emu g−1 is likely due to the high concentration of magnetite nanoparticles homogeneously dispersed across the matrix, as observed in SEM (Fig. 3b). For reference, the Ms of ferrofluids is between 60 and 70 emu g−1 due to uniform particle distributions with a size of a few nanometers50. The difference in Ms for our films corresponds to nanoparticle size distribution and surface modification methods (in this case, PEG), which is instrumental in maintaining the size and shape during the particle formation.

Cytocompatibility and simulated wound healing with human dermal fibroblast (HDFα) cells

The bioengineered magnetic cellulose films are evaluated for cytocompatibility in the presence of human dermal fibroblast (HDFa) cells (Fig. 5). This cytocompatibility assay is performed to assess if any components of the engineered membrane leachate led to toxicity. Our results reveal that over the course of 7 days of incubation, the cell viability of hybrid magnetic bacterial cellulose films is comparable to the control group without bacterial cellulose. As expected, pristine BC films alone does not enhance cell proliferation of cells, due to its morphological, chemical, and biocompatible properties. The observations found in cytocompatibility experiments are put to test for wound closure in a simulated environment.

Fig. 5: Cytocompatibility of the bacterial cellulose films with human dermal fibroblasts (HDFα).
Fig. 5: Cytocompatibility of the bacterial cellulose films with human dermal fibroblasts (HDFα).The alternative text for this image may have been generated using AI.
Full size image

Cell viability assay is conducted to evaluate the compatibility of HDFα cells in various BC films for 7 days incubation at 37°C with 5% CO2. Control is HDFα cells grown without any BC film (white), pristine BC, BCFe, BC(curli)Fe and BC(curli-M6A)Fe. Cell viability is normalized against day 1 control (white). The cell viability of the hybrid films is comparable to the control samples without any films. The error bars represent the standard deviation for cell viability from three independent experiments, each performed in triplicates.

A scratch test is performed in this study to simulate a wound site and to evaluate the performance of the magnetic hybrid bacterial cellulose films as wound healing substrates. We evaluated the effects of 10 mT static magnetic field application on cell migration. About 1 mm wide scratch is created and the effect of various bacterial cellulose membranes in the absence and presence of magnetic field is analyzed. The migration of HDFa cells is significant in the presence of functional magnetic membranes compared to the other sets in the absence of static magnetic field (Fig. 6). We observed a 30% improvement in cell migration for functional BC(curli-M6A)Fe films compared to the control group (Fig. 6). In the presence of a 10 mT static magnetic field, there is an obvious improvement in cell migration post 48 h of incubation (Supplementary Fig. 5). The control group devoid of any film did not improve even in the presence of magnetic field due do the direct effect of magnetic field on the cellular activity in the absence of magnetite nanoparticles (Supplementary Fig. 4)51. With the incorporation of magnetite nanoparticles and stimulation of cells with magnetic field, scratch closure improved by 44% with the application of bioengineered BC(curli-M6A)Fe films compared to the control group (Fig. 6).

Fig. 6: Simulated wound closure response of HDFα in the presence of magnetic field.
Fig. 6: Simulated wound closure response of HDFα in the presence of magnetic field.The alternative text for this image may have been generated using AI.
Full size image

A scratch, simulating wounds of 1 mm width are generated with HDFα cells. Response of HDFα cells in scratch closure with and without a magnetic field of 10 mT is monitored in the presence of the BC films. BC(curli-M6A)Fe films accelerated cell migration in the presence of magnetic field. The y-axis represents the scratch gap width after 48 h of incubation. The box represents 25–75% percentile of where the data points lie. *p ≤ 0.01 and **p ≤ 0.005.

Discussion

Conventionally, bacterial cellulose modifications are carried out through physical or chemical processes. Physical processes involve defibrillation of bacterial cellulose films using mechanical force, sputter coating of nanoparticles or physical adsorption of reinforcement materials52. The presence of multi-hydroxyl structure contributes to strong intermolecular H-bonds in bacterial cellulose, resulting in poor solubility in water19. Chemical modifications on bacterial cellulose have several limitations, such as disposal of toxic chemicals, low reaction efficiency and waste of resources. Our work overcomes these drawbacks. The key goal is the use of a microbial fermentation system to produce functional bacterial cellulose hybrids. With this system we can achieve better utilization of the raw materials and improved control over the final product design53. Therefore, we anticipate that a biological functionalization approach can aid in the development of bacterial cellulose-based functional hybrids with tailored properties.

The biological approach employed in this study is synonymous to bottom-up approach wherein, bacteria produce the base scaffold as well as impart functionality to the scaffold36. The bioengineered functional cellulose films comprised of bacterial cellulose and curli-M6A, the amyloid fusion protein curli subunit, CsgA, genetically tagged with nanoparticle nucleating peptide domain, M6A. Upon incubation of the preformed bacterial cellulose, the bacterial cellulose scaffold network traps recombinant E. coli bacteria within the matrices and aids in incorporating the curli-M6A peptides. The M6A peptide domains add functionality to the cellulose matrix devoid of any harsh chemical or mechanical processes. Additionally, the secretion of curli-M6A peptides by the engineered E. coli within the vicinity of the bacterial cellulose fibers provides high local concentration of the peptides with minimum diffusion distance. The strategy allows for dense functionalization of the inner matrix of bacterial cellulose which has not been observed in previous cellulose composite design approaches.

Hybrid BC(curli-M6A) films display an intense black color post magnetite synthesis suggesting a high density of Fe3O4 nanoparticles in the matrix. The absence of the m6A peptide however resulted in formation of non-uniform magnetic particles with inferior magnetic properties. These non-specific nanoparticle formation (around amino acid side chains of curli and hydroxyl groups of BC) lack the nucleation site provided by M6A. SEM confirms the high density of nanoparticles across the matrix, while the absorption band of Fe-O confirms the chemical incorporation of Fe3O4 in BC(curli-M6A). XPS validated the interaction of the cellulose matrix with the magnetic nanoparticles owing to the chemical shift in the binding energy of atom core-levels. The left shift of the preferential orientation of the (200) plane with the introduction of the nanoparticles in the matrix further confirms the successful incorporation of the nanoparticles within the matrix. In contrast to physicochemical method of magnetic films, bioengineered magnetic films are produced at much lower reaction temperatures (i.e., 80°C) and without any structural disruption. Despite its relatively low synthesis temperature, the BC(curli-M6A) films are thermally stable at 200°C. VSM measurements reveal high saturation magnetization of 40 emu g−1 for BC(curli-M6A) films, four times higher than conventional in/ex situ bacterial cellulose doping54,55,56,57. Overall, the M6A peptides present in BC films assist in controlled nucleation of the magnetite particles thereby adding functionality to BC films without the use of harsh chemical or physical processes.

Wound healing is a complex cellular response process. It involves the activation of several cells responsible for wound closure. Fibroblast cells are involved in the repair of structure and function at the wound site. Magnetic membranes have previously been shown to aid in wound healing process especially the proliferation phase, which usually lasts for 4–24 days58. Our results indicate that the presence of magnetite nanoparticles significantly improves the scratch closure provided by the magnetic field. With 72 h of incubation, BC(curli-M6A) films promote HDFα cell proliferation. The hybrid BC film improved cell migration compared to the control BC films in the presence of 10 mT magnetic field. The results show that BC(curli-M6A) has clear potential as a substrate for wound healing applications.

This work reports the successful biological functionalization BC by leveraging on the molecular precision of self-assembly process. The functionalization is achieved by employing microbial systems that combine living with non-living components. With a simple design strategy of using the most amenable microbes for both the production of the scaffold (bacterial cellulose) as well as the functionalizing peptides (curli), functional bacterial cellulose membrane is generated. Also, with the availability of the genetic toolkits for engineering various microbial species, multiple functionalities could be engineered in the same microbe. This platform can be extended further for the sustainable development of new classes of bacterial cellulose-based hybrid materials.

The key tools involved in biofunctionalization approach are (i) selection of suitable microbes, (ii) design of the genetic component containing information and (iii) understanding the metabolism of the selected microbes. Based on these considerations, culturing strategy can be simplified further by having better control over distribution of functionality and improvement in large-scale bacterial cellulose production. An ideal modification process ought to be (i) cost effective, (ii) scalable, (iii) better in providing reaction control and (iv) eco-friendly. In the present situation of energy crises, there is pressing need for sustainable development, clean and green solutions to meet the energy demands. The hybrid film developed through the biofunctionalization strategy requires few chemicals, demonstrating its contribution to sustainability and an environmentally friendly material, further highlighting its potential in a myriad of practical applications.

Methods

Strains and culture medium

Komagataeibacter hansenii ATCC 53582 was used for BC production grown in the standard Hestrin-Schramm (HS) medium containing 5.0 g bacterial yeast extract, 5.0 g bacterial peptone, 2.7 g sodium hydrogen phosphate dibasic, 1.2 g citric acid and 20.0 g D-glucose59. All cultures were incubated at 26°C60 static conditions for 5 days.

Plasmid pBbB8k containing the curli operon csgBACEFG was used as a template for incorporating M6A peptide. HiFi DNA assembly kit (NEB) was used to perform Gibson assembly. M6A peptide (DIESAQSDEEVE) with linker GSGGSG on either side was fused to the C-terminus of csgA and successful construct upon sequencing was used for expression.

Lysogeny broth (LB) with chloramphenicol (50 μg/ml, Cm) and kanamycin (50 μg/ml, Km) as antibiotics, was used for growth as well as protein production.

Microbial fermentation

For BC production, all cultures were maintained at 26°C, static conditions for 5 days. For curli gene expression, E. coli PQN4, a curli operon deletion mutant was used for expression. Overnight cultures were prepared in LB containing 50 μg/ml Km and 2% glucose. Overnight culture was diluted 50-fold for expression in LB containing Km, Cm and 2% glucose and cell were grown to OD 0.8 at 180 rpm, 37°C. Cells were harvested (4000 × g, 30 min) and transferred to fresh induction medium of LB with Km, Cm, 0.5% arabinose and cellulose pellicles and incubated at 180 rpm, 30°C for 16 h.

Biofunctionalization of bacterial cellulose

Cellulose films collected post 5 days of incubation, were transferred to the induction medium. Induction medium containing BC films was left overnight at 30°C and 180 rpm for CsgA fusion protein expression. Hybrid BC films were harvested from the induction medium the following day. They were washed thrice with deionized water and heated in 1%(w/v) NaOH at 100°C for 10 min, to remove the entrapped the cellulose-producing bacteria. The pellicles were rinsed again thrice with deionized water.

Magnetic BC films design

BC from Gluconacetobacter bacteria provided the backbone scaffold for the membrane design and facilitated in entrapping the recombinant E. coli bacteria. The recombinant E. coli bacteria contained the genetic circuit for functional display. The functional peptide, M6A, assisted in the localization and growth of magnetite nanoparticles as shown in the Fig. 1. The cellulose films were transferred to a culture containing recombinant E. coli producing curli-M6A to functionalize the cellulose films. Hybrid cellulose films thus formed were subjected to magnetite nanoparticle formation. Particles with uniform size and shape were produced in the presence of the hybrid BC(curli-M6A) films.

Nanoparticle nucleation in hybrid BC films

The hydrated hybrid BC films were solvent exchanged overnight in 20%(v/v) ethylene alcohol to allow better penetrability for nanoparticle synthesis61. The solvent exchanged hybrid films were dropped in a 20 ml reaction volume containing 400 mM potassium nitrate (KNO3), 100 mM potassium hydroxide (KOH) and 1%(w/v) poly(ethylene glycol) (PEG) used as a surfactant. The entire mixture was capped and bubbled with N2 gas for 5 min to remove any oxygen present. Iron (ii) heptahydrate (30 mM) solution was introduced dropwise into the mixture using a syringe pump at 100 µL/min for 20 min41. The nanoparticle synthesis was carried out in a silicone oil bath set to 80°C and the mixture was stirred using a magnetic stirrer to ensure there is even distribution of the magnetite nanoparticle across the matrix. Post nanoparticle synthesis, the magnetic membranes were washed with deionized water multiple times to ensure that there was no residual chemicals or excess nanoparticles absorbed on the films. Pristine BC films loaded with nanoparticle were referred to as BCFe, BC films fused with curli protein without the nanoparticle nucleating peptide were referred to as BC(curli)Fe and the BC film fused with the curli protein tagged with the short peptide domain, M6A were referred to as BC(curli-M6A)Fe.

Confocal microscopy

CLSM (Zeiss LSM 800, Germany) was used to confirm the presence of amyloid fusion protein in the hybrid films. Samples were soaked in 25 µM of amyloid protein specific dye, curcumin, for 15 mins. Images were recorded at 20× magnification, in the green channel and were processed using Zen software (blue edition).

Scanning electron microscopy (SEM)

The morphologies of pristine BC, BCFe, BC(curli)Fe and BC(curli-M6A)Fe films were viewed using JSM-6700F (JEOL, Japan) operating at an accelerating voltage of 1.5 V. The samples were mounted on aluminum stubs and coated by platinum sputter.

Fourier transform infrared (FTIR) spectroscopy

IR spectra were recorded on a Spectrum One FTIR (Perkin Elmer, USA) equipped with potassium bromide (KBr) beam splitter and deuterated triglycine sulfate detector. The spectra were measured in transmission mode utilizing KBr discs. All spectra measurements were obtained at room temperature with a resolution of 4 cm−1, spectral range from 4000 to 400 cm−1 and averaged over 16 scans. The spectra were baseline-corrected and optimized using Spectrum V5.0.1 software1.

X-ray diffraction (XRD)

The lyophilized BC-based films were analyzed using a D2 PHASER (Bruker, USA), benchtop x-ray diffractometer. The wavelength of the Cu/Kα radiation source was at 0.154 nm, operated with an accelerating voltage of 30 kV, filament emission of 10 mA and scanning range between 5 and 70° (2θ).

Thermogravimetric analysis (TGA)

All measurements were done using TGA Q500 (TA instruments, USA). The samples were weighed 2–3 mg and loaded onto aluminum pans. The pans were scanned from 30°C to 800°C with a heating rate of 5°C/min under a nitrogen atmosphere (60 ml/min). DTG curves were obtained using the first derivative, rate of weight loss as a function of temperature. The data were analyzed using the in-built TA Universal Analysis Software.

X-ray photoelectron spectroscopy (XPS)

X-ray photoelectron spectroscopy (XPS) was performed using an AXIS Supra spectrometer (Kratos Analytical, UK) equipped with a hemispherical analyzer and a monochromatic Al K-alpha source (1487 eV) operated at 15 mA and 15 kV. The XPS spectra were acquired from an area of 700 × 300 μm2 with a take-off angle of 90°. Pass energy of 160 eV and 20 eV was used for survey and high-resolution scans, respectively. A 3.1-volt bias was applied to the sample to neutralize charge build up on the sample surface. The binding energies (BEs) were charge-corrected based on the sp3 carbon C 1 s at 284.8 eV.

Vibrating sample magnetometer (VSM)

Magnetic measurements were recorded using 8600 series VSM (Lake Shore Cryotronics, USA). The magnetic property was obtained by measuring the magnetization (M) versus applied magnetic field (Oe) under a maximum field of 20 kOe. The hysteresis curves were analyzed using the 8600 series built-in software.

Cell viability

Resazurin fluorescence assay was conducted to determine the viability of cells as described elsewhere62. Primary human dermal fibroblasts (HDFα) were cultured in DMEM medium supplemented with 10%(v/v) fetal bovine serum, 50 U/mL penicillin and 50 μg/mL streptomycin, in a humidified incubator at 37°C with 5% CO2. Cells (1 × 105 cells well−1) were seeded at the bottom of the 12-well plates (TPP®) and allowed to adhere and grow for 24 h. After 24 h different films were introduced into the wells and tested for cell viability. Readings were taken every alternate day for a period of 7 days. For every reading, 150 μl of 0.15 mg/ml resazurin dye was added to the wells and incubated for two hours after which fluorescence signal was recorded (excitation: 560 nm and emission: 590 nm).

Scratch test

Scratch wound assay was performed according to previously described protocol63. Briefly, primary human dermal fibroblasts (HDFα) were cultured in DMEM medium supplemented with 10%(v/v) fetal bovine serum, 50 UmL−1 penicillin and 50 μgmL−1 streptomycin, in a humidified incubator at 37°C with 5% CO2. Cells (25 × 103 cells well−1) were seeded at the bottom of the 96-well plates (TPP®) and allowed to adhere and to grow for 24 h to form a confluent monolayer. Using Accu Wound 96 (Acea Biosciences Inc.) uniform scratches of 1 mm were created at the center of the well. Each well was washed thrice with prewarmed (37°C) PBS to remove detached cells from the scratch site. The wells were then replenished with fresh DMEM containing 1% FBS and with different BC films placed on the scratch. Wound closure was studied in the absence and presence of 10 mT static magnetic field. The design for the scratch test was represented in Supplementary Fig. 3. Wound closure was captured with a 10× objective using a light microscope. Each wounded area was captured at 0, 24, 36 and 48 h post scratch. Wound closure was quantified using ImageJ software (developed by Wayne Rasband, National Institutes of Health, Bethesda, MD; provided in the public domain at http://rsb.info.nih.gov/ij).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.