Diatom Microbubbler for Active Biofilm Removal in Confined Spaces

Abstract

Bacterial biofilms form on and within many living tissues, medical devices, and engineered materials, threatening human health and sustainability. Removing biofilms remains a grand challenge despite tremendous efforts made so far, particularly when they are formed in confined spaces. One primary cause is the limited transport of antibacterial agents into extracellular polymeric substances (EPS) of the biofilm. In this study, we hypothesized that a microparticle engineered to be self-locomotive with microbubbles would clean a structure fouled by biofilm by fracturing the EPS and subsequently improving transports of the antiseptic reagent. We examined this hypothesis by doping a hollow cylinder-shaped diatom biosilica with manganese oxide (MnO₂) nanosheets. In an antiseptic H₂O₂ solution, the diatoms doped by MnO₂ nanosheets, denoted as diatom bubbler, discharged oxygen gas bubbles continuously and became self-motile. Subsequently, the diatoms infiltrated the bacterial biofilm formed on either flat or microgrooved silicon substrates and continued to generate microbubbles. The resulting microbubbles merged and converted surface energy to mechanical energy high enough to fracture the matrix of biofilm. Consequently, H₂O₂ molecules diffused into the biofilm and killed most bacterial cells. Overall, this study provides a unique and powerful tool that can significantly impact current efforts to clean a wide array of biofouled products and devices.

Graphical Abstract

1. INTRODUCTION

Biofilms are communities of bacterial cells that are prevalent in many materials, such as medical devices, injured/diseased living tissues, household products, and infrastructure.^1–11 Multiple microbial cells build the biofilm by accumulating on solid–liquid interfaces and producing a protective matrix of extracellular polymeric substances (EPS), which includes various biomolecules, such as oligopeptides, lipopolysaccharides, proteins, lipids, and DNA.^12–16 The resulting biofilms impact public health, product function, and infrastructural sustainability significantly.^3–10 About 80% of all medical infections originate from biofilms of pathogens. Therefore, tremendous efforts have been conducted to remove biofilms from the substrates using various antibiotics and disinfectants.^17–24 However, bacterial cells residing in biofilms are deemed 100–1000 times more resistant to antibiotics and disinfecting agents than planktonic cells because the EPS matrix limits the transport of the antimicrobial agents and neutralizes them chemically.^{12–14,24,25} Furthermore, the biofilms formed in confined spaces are even more challenging to eliminate using conventional methods than that on the open surface because the biofilm matrix becomes even sturdier and denser.^26,27

To this end, we hypothesized that microparticles assembled to generate microbubbles and self-propel in the antibacterial H₂O₂ solution would act as an active cleaning agent. Due to the small size and active movement, the particles would be able to penetrate the EPS matrix of biofilm and continue to generate microbubbles inside the biofilm. The resulting microbubbles would give rise to a wave of mechanical energy high enough to deform the biofilm. In turn, H₂O₂ molecules would diffuse into the biofilm matrix and eradicate bacterial cells. We examined this hypothesis by doping manganese oxide (MnO₂) nanosheets on the diatom biosilica, which is a skeleton of dead algae. MnO₂ nanosheets can generate oxygen (O₂) bubbles by decomposing H₂O₂.^28,29 The porous and cylindrical geometry of diatom is advantageous to facilitate the diffusion of H₂O₂ into MnO₂ nanosheets.

2. EXPERIMENTAL SECTION

2.1. Fabrication of the MnO₂-Diatom Bubbler

The MnO₂-diatom bubbler was fabricated using diatom particles doped with MnO₂ nanosheets. Amine-substituted diatom particles were prepared by the reaction with (3-aminopropyl)triethoxysilane (APTES, Sigma-Aldrich). First, 2 g of diatom particles was added to 60 mL of toluene in a three-necked round-bottom flask fitted with a thermometer, a reflux condenser, and an N₂ gas tube. Distilled water (0.6 mL) was added to the mixture and stirred for 2 h at room temperature. Then, 3.4 mL of APTES was added to the mixture, which was taken to reflux for 6 h at 60 °C. Secondly, the mixture was cooled and washed with toluene, 2-propanol, and distilled water three times. The obtained sample was dried in a vacuum desiccator for 2 days. Finally, 0.1 g of amine-substituted diatoms was added to 1 mL of 50 mM potassium permanganate (KMnO₄, Sigma-Aldrich) solution and sonicated for 30 min at room temperature. Then, the samples were filtered and washed with distilled water and ethanol three times. The obtained sample was dried in an oven for 1 day at 60 °C.

2.2. Characterization of the MnO₂-Diatom bubbler

The optical images and movies of diatom bubblers were obtained with the optical microscope (Leica DMIL). The scanning electron microscopy (SEM) images and elemental mapping of microparticles were obtained with HITACHI S-4700 microscope operating at 2.0 kV. The transmission electron microscopy (TEM) images were obtained using JEOL 2100 TEM with an accelerating voltage of 200 kV. The elemental analysis of manganese within the diatom bubbler was conducted by the inductively coupled plasma-atomic emission spectroscopy (ICP/AES). The nitrogen adsorption–desorption isotherms were measured at 77 K with the Quantachrome NOVA 2200e surface area and pore size analyzer. The specific surface area was calculated with the Brunauer–Emmett–Teller equation from the adsorption and desorption curves in the relative pressure (P/P₀) range of 0.05–0.30. The pore volume and pore size distribution were calculated from the desorption curve by the Barrett–Joyner–Halenda method in the relative pressure P/P₀ range of 0.05–0.99. The powder X-ray diffraction data were obtained using the Rigaku Miniflex 600. The size distribution of microparticles was obtained by the image analysis with the ImageJ software. The speed of particles was measured by using the Fiji software. The mass of O₂ gas generated from the decomposition of H₂O₂ was quantified by adding 2–4 mg of diatom bubbler to 2 mL of 3% H₂O₂ solution in a gas-sampling bottle and measuring the partial pressure of O₂ in the gas phase with the sensor (Vernier).

2.3. Preparation of Microgrooved Poly(dimethylsiloxane) (PDMS) Substrates

First, the silicon master was fabricated using the photolithography technique. To begin with, the flat silicon wafer was washed and dried. Then, a photoresist (SU-8 2050, MicroChem Corporation) was spin-coated (100 μm thickness) onto the wafer to reach desired thickness. Next, the substrate underwent soft bake, UV exposure, and postexposure bake. During the UV exposure step, a photomask with the designed pattern was placed on the wafer to cure only the area of interests. After the curing process, the excess uncured photoresist was washed out. The master was silanized by treatment with a vapor of trichloro (1H,1H,2H,2H-perfluorooctyl)silane, (Sigma-Aldrich) to create a hydrophobic surface. Secondly, PDMS prepolymer (SYLGARD 184 Silicon Elastomer Kit, Dow Corning) and its curing agent were thoroughly mixed in the 1:10 mass ratio. After degassing, the prepolymer solution was cast onto the silicon master with the desired feature on it. Then, it was cured at 60 °C for 2 h. The cured PDMS with the desired pattern was gently peeled off from the master. The fabrication of the flat PDMS followed the same procedure. The only difference was that the prepolymer solution was poured on the flat silicon wafer.

2.4. Biofilm Growth on the PDMS Substrate

Microbial broths were prepared using Luria-Bertani (LB) Broth, Lennox media (Difco). Escherichia coli (ATCC No. 25922) was grown in the LB Broth at 37 °C under shaking (100 rpm) until they reached the midlogarithmic growth phase. The concentration of E. coli was adjusted to obtain an optical density of 0.07 at the wavelength of 600 nm on a microplate reader (TECAN, Switzerland). This optical density corresponds to the concentration of McFarland 1 solution (3 × 10⁸ CFU mL⁻¹) E. coli (250 μL, 3 × 10⁸ CFU mL⁻¹). Flat or microgrooved PDMS substrates were placed on the cell culture media in 24-well plates, while immersing only the groove parts in solution. After 24 h, the PDMS substrates were rinsed once with sterile phosphate-buffered saline (PBS). Then, 250 μL of 3 × 10⁸ CFU mL⁻¹ fresh E. coli suspension was additionally added to the PDMS substrates. This process was repeated every day for 9 days to allow for the biofilm formation. After 9 days, the PDMS substrates were rinsed with sterile PBS before being used for further analysis.

2.5. Biofilm Removal from the PDMS Substrate

After 9 days culture, biofilms grown on PDMS substrates were treated by replacing the cell culture media with 600 μL of 3% H₂O₂ solution or the mixture of 3% H₂O₂ solution and the MnO₂-diatom bubblers. After 30 min, all of the solution removed and gently rinsed once with PBS before the analysis.

2.6. Rheological Analysis of the Biofilm

Rheological properties of the biofilm samples grown on flat PDMS substrates were measured by strain-controlled oscillatory shear experiments using a rotational rheometer (ARES-G2, TA instrument). Oscillatory strain (amplitude)-sweep tests were performed to determine the storage (G′) and loss modulus (G″) with 25 mm diameter parallel plate geometry at a fixed frequency of 1 rad sec⁻¹. The dynamic moduli were measured with the biofilm on the PDMS substrate that is fixed to lower plate of the rheometer. All of the analyses were conducted and averaged from the results of five different measurements with different samples.

2.7. Viability Assay of the Biofilm

The Cell Titer-Blue cell viability assay was performed to quantitatively analyze live E. coli cells attached to the PDMS. E. coli in the LB Broth, Lennox media (Difco) (100 μL, 3 × 10⁸ CFU mL⁻¹) were seeded onto PDMS substrates cured in 24-well plates and cultured at 37 °C for 9 days. After 9 days, 600 μL of 3% H₂O₂ solution or the mixture of 3% H₂O₂ and the MnO₂-diatom bubbler particles was added to the biofilms for 30 min. The surfaces were washed twice with sterile PBS, followed by incubation with 100 μL of PBS and 20 μL of Cell Titer-Blue Reagent at 37 °C for 6 h. The fluorescence intensity of resorufin resulting from reduction of resazurin by dehydrogenases of viable cells was determined at the excitation wavelength of 560 nm and the emission wavelength of 590 nm using the microplate reader. The experiment was conducted in replicates of 3.

2.8. Fluorescence Imaging of the Biofilm

Proteins in the biofilm were stained by incubating them for 1 h at room temperature with 500 μL of 10 mg mL⁻¹ fluorescein isothiocyanate (Sigma-Aldrich) in 0.1 M NaHCO₃ buffer to conjugate the dye onto deprotonated amino groups. Then, the biofilm was washed with PBS before staining polysaccharides in the biofilm. Polysaccharides in the biofilms were stained by incubating the biofilm for 2 h with 500 μL of 250 μg mL⁻¹ fluorescently labeled lectin concanavalin A conjugated with tetramethyl rhodamine (Molecular Probes). The chemical binds to α-glucopyranosyl and α-mannopyranosyl sugar residues. Subsequently, the biofilms were washed with PBS before incubating them for 2 h with 500 μL of 300 μg mL⁻¹ fluorescent brightener 28 (MP Biomedicals, LLC), which binds to β-linked polysaccharides. The biofilms were washed with PBS. Then, the samples were mounted on glass slides and observed with a laser scanning confocal microscope (Zeiss LSM 700, Germany, a 200× objective).

3. RESULTS AND DISCUSSION

3.1. Synthesis and Characterization of MnO₂ Nano-sheet-Doped Diatom Bubbler

The diatom particles used in this study have the hollow cylinder morphology with 10 μm diameter and 18 μm length, whereas the wall exhibits many pores with an average diameter of 500 nm (Figures 1a and S1a). First, the diatom surface was functionalized with (3-aminopropyl)triethoxysilane (APTES) to present amine groups (Figure 1b). Then, MnO₂ nanosheets were doped on the amine-functionalized diatom by reducing potassium permanganate (Figures 1b and S1b). The loading amount of MnO₂ nanosheets was approximately 2.5 wt %, according to the inductively coupled plasma-atomic emission spectroscopy analysis. The elemental mapping of Mn in SEM confirms that the MnO₂ nanosheets are loaded on the diatom particles (Figure 1c). The high-resolution transmission electron microscopic (TEM) images show that MnO₂ nanosheets are located on the porous wall surfaces (Figure 1d). According to the N₂ adsorption analysis, the specific surface area and pore volume of MnO₂ nanosheet-doped diatom (24 m² g⁻¹ and 0.057 cm³ g⁻¹) are comparable to those of bare diatom particles (27 m² g⁻¹ and 0.076 cm³ g⁻¹). This result indicates that most pores of the diatom particle were still open after deposition of MnO₂ nanosheets (Figure 1e). On the other hand, the powder X-ray diffraction pattern of the MnO₂ nanosheet-doped diatom exhibits no difference from bare diatom. This result reveals that the doped MnO₂ nanosheets on diatom particles are in an amorphous state (Figure 1f).

Figure 1.

Fabrication and characterization of the MnO₂ nanosheets-doped diatom. (a) SEM image of diatom. (b) Schematic illustration of fabrication steps for MnO₂ nanosheets-doped diatom. (c) Elemental mapping images showing homogenous distribution of MnO₂ nanosheets on diatom. (d) High-resolution TEM image of MnO₂ nanosheets on diatom. (e, f) Nitrogen adsorption–desorption isotherms (e) and powder X-ray diffraction patterns (f) of diatom and MnO₂ nanosheets-doped diatom.

The MnO₂ nanosheet-doped diatom (MnO₂-diatom) instantly generated O₂ microbubbles in a 3% H₂O₂ solution, whereas the diatom without MnO₂ nanosheets did not show any bubbles (Figure S2). Interestingly, O₂ bubbles are continuously ejected from the hollow hole of the MnO₂-diatom, as shown in the time-lapse images and movie (Figure 2a and Movie S1). The bubbles showed a uniform diameter of 12 μm, which is similar to that of diatom. From this result, we assume that H₂O₂ molecules in the solution diffuse through nanopores of the diatom wall and quickly decompose to generate O₂ gas by reacting with MnO₂ nanosheets (Figure 2b). The generated O₂ gas bubbles start to nucleate and form microbubbles inside the hollow space of a diatom. As the bubbles build up the pressure, the diatom bubbler continuously jets microbubbles through the hollow channel. The ejection of bubble from the hollow channel allows the continuous diffusion of H₂O₂ molecules and generation of the microbubbles. Due to the continuous inward diffusion, no microbubbles were formed on the outer surface of MnO₂-diatoms. This spatial effect on the microbubble formation will be studied systematically in future. Subsequently, the MnO₂-diatom moved at a rate of 60 μm s⁻¹ (Figure 2c and Movie S2). The O₂ gas generation from the diatom bubbler was also confirmed with the oxygen sensor. The initial reaction rates with different MnO₂-diatom concentrations were calculated to determine the order of reaction (Figures 2d and S3). The order of the reaction was approximately 1.2 with respect to the MnO₂-diatom concentration.

Figure 2.

Self-locomotion of MnO₂ nanosheet-doped diatom in H₂O₂ solution. (a) Time-lapse images of the microbubble generated from the diatom in 3% H₂O₂ solution (scale bar = 10 μm). (b) Schematic illustration of the mechanism by which the MnO₂-diatom produces O₂ bubbles and self-propels. (c) Tracking of the self-locomotion of MnO₂-diatom bubbler added to H₂O₂ solution (scale bar = 20 μm). (d) Quantification of O₂ gas generated from 3% H₂O₂ solution mixed with varying amounts of the MnO₂-diatom bubbler. Unmodified diatom was included as a control.

3.2. Active Biofilm Removal with Diatom Bubbler

Next, we examined the extent to which the MnO₂-diatom bubbler cleans the flat polydimethylsiloxane (PDMS) substrate fouled by the Gram-negative E. coli biofilm. First, the fouled PDMS substrate was exposed to 3% H₂O₂ solution for 30 min. This procedure resulted in macrosized bubbles with a few hundred micrometers above the biofilm because the catalase in E. coli decomposed H₂O₂ into O₂ gas (Movie S3). Approximately, 80% of the biofilm remained on the PDMS substrates when they were examined with microscope images (Figures 3a,b, S4, and Movie S3). According to the oscillatory shear test, the storage modulus of the biofilm decreased from 20 to 10 Pa (Figure 3c). However, H₂O₂-treated biofilm showed a high yield strain, indicating that the structure of biofilm matrix was retained even after H₂O₂ treatment, suggesting no significant change to the structure. In contrast, the MnO₂-diatom bubbler mixed with H₂O₂ solution-generated microbubbles on the fouled PDMS substrate and removed most of the biofilm within 30 min (Figures 3a,b, S4, and Movie S3). Because there was no remaining biofilm on the substrate, it was not possible to measure the rheological property of the biofilm after the treatment of MnO₂-diatom bubbler.

Figure 3.

Effects of the MnO₂-diatom bubbler on the biofilm formed on flat or microgrooved PDMS substrate. (a, b) Optical (a) and SEM (b) images of E. coli biofilm grown on flat PDMS substrates before and after treatments of 3% H₂O₂ solution without and with MnO₂-diatom. (c) Storage modulus (G′) and loss modulus (G″) of the biofilm as a function of the oscillatory strain amplitude. (d) Optical image of the biofilm formed on microgrooved PDMS substrates (50 μm width and 100 μm depth) before and after treatments of 3% H₂O₂ solution without and with MnO₂-diatom. (e) Quantified biofilm area on microgrooved PDMS substrate. (f) The viability of E. coli in the biofilm on the PDMS substrate. The secondary treatment with diatom bubbler led to the death of 99.9% of cells, which is marked with an asterisk (*).

We further examined the capability of the MnO₂-diatom bubbler to invade and remove the biofilm formed in a confined space. To simulate the biofilm in confined spaces, we used the microgrooved PDMS substrates with controlled widths of 50, 100, 200, and 500 μm and depth of 100 μm. The biofilms were formed exclusively within microgrooves regardless of width. The confinement effect of biofilm formation could be explained by cell-to-cell signaling and interaction among bacterial cell clusters.^26,27,30 Most of the biofilms in microgrooves of PDMS substrates were removed 30 min after the mixture of 3% H₂O₂ and the MnO₂-diatom was added (Figures 3d and S5). In contrast, the biofilms still lingered on the microgrooves after 3% H₂O₂ solution treatment (Figures 3d and S5). The quantitative analysis of the images revealed that the MnO₂-diatom treatment cleaned 8-fold larger surface area than H₂O₂ solution treatment alone (Figure 3e). The cell viability gradually decreased with the time at the concentration of the MnO₂-diatom bubbler being 1.0 mg mL⁻¹ in 3% H₂O₂ solution (Figure S6a). On the other hand, increasing the concentration of the MnO₂-diatom bubblers from 1.0 to 3.0 mg mL⁻¹ decreased the cell viability a lot within 5 min (Figure S6b). As a result, 95% of E. coli cells in the biofilm lost viability after the MnO₂-diatom bubbler treatment, whereas most of the cells in the biofilm treated only with the H₂O₂ solution remained viable (Figure 3f). Moreover, the secondary treatment with a fresh mixture of 3% H₂O₂ and the MnO₂-diatom bubblers led 99.9% of E. coli cells to lose viability. In addition, the MnO₂-diatom bubbler was effective to reduce the viability of E. coli in the detached biofilm (Figure S6c).

The structure of biofilms in microgrooves before and after the treatment was analyzed with fluorescence images (Figures 4a and S7). The EPS matrix of untreated biofilm consists of β-linked polysaccharides (blue-colored) and α-glucopyranosyl and α-mannopyranosyl sugar residues (red-colored). Extracellular proteins (green-colored) were localized mostly on E. coli cell walls. Even after 3% H₂O₂ solution treatment, a large amount of EPS and E. coli remained in the microgrooves, as quantified with the mean fluorescence intensity across the microgrooves. In contrast, the mixture of H₂O₂ and MnO₂-diatom bubblers removed all EPS and E. coli cells, as confirmed with no fluorescence yield from the substrate. This result indicates that MnO₂-diatom bubbler treatment could prevent biofilm from recovering because there was no residual amount of EPS and bacterial cells at all after the treatment.

Figure 4.

Mechanistic study of the active biofilm removal using MnO₂-diatom bubbler. (a) Fluorescent images of the β-linked polysaccharides (blue), α-glucopyranoayl/α-mannopyranosyl polysaccharides (red), and extracellular proteins (green) in extracellular polymer substances (scale bar = 10 μm). (b) Optical image of the intermediate stage of biofilm removal using MnO₂-diatom. The red and yellow arrows indicate the cleaned area in microgrooves by microbubbles and the invaded MnO₂-diatom particles in biofilm matrix, respectively. (c) Optical images of the biofilm captured after 5 min of exposure to 3% H₂O₂ solution with and without MnO₂-diatom. (d) Time-lapse image of microbubbles within the grooves. Smaller microbubbles (dot circles) merge into a big bubble (yellow arrows), which collapses eventually. (e) Optical images of the biofilm formed in the PDMS substrate with complex micropattern (i), the biofilm treated by 3% H₂O₂ solution without (ii) and with MnO₂-diatom (iii). (f) Schematic illustration of the active cleaning mechanism.

3.3. Mechanism Study of Active Biofilm Removal

We monitored the intermediate stage of treatment to address the active cleaning mechanism of the MnO₂-diatom bubbler in more detail. The self-locomotive diatom bubblers infiltrated the biofilm and continuously produced microbubbles within the biofilm (Movie S4). These microbubbles merged and burst to deform and fracture the biofilm until they collapsed. As a result, circular- or ellipsoidal-shaped damages could be found in between the biofilm, whereas the diatom particles are observed inside the biofilm matrix (Figure 4b,c). In contrast, the biofilm exposed to the H₂O₂ solution exhibited macrosized bubbles with a few hundred micrometers stayed outside the substrate due to the diffusion limitation (Figure 4c and Movie S5).

The energy of microbubbles used to rupture the biofilm was estimated by calculating the change in the surface energy of microbubbles.³¹ The coalescence of the microbubbles is considered as a process to minimize surface energy.³² During this coalescence, the surface energy is released in the form of mechanical energy. Since initially formed microbubbles have a nearly spherical shape, the mechanical energy is calculated from surface energy difference between initial and final states as follows,

$$\mathrm{\Delta }E=\sum 4\pi {r_i}^2\mathrm{\Gamma }-\sum 4\pi {r_{\text{f}}}^2\mathrm{\Gamma }$$ (1)

where r_i, r_f, and Γ represent the radius of bubbles at the initial state, the radius of bubbles at the final states, and the surface tension value for the water–air interface, respectively. We calculated the ΔE based on the results observed in optical microscope (Figure 4d and Movie S6). The calculated ΔE was compared to the mechanical work required to deform the biofilm in the observed area. Using the storage modulus (G′) of the untreated biofilm shown in Figure 3c, the required mechanical work per unit volume (E_req) is given as,

$$E_{\text{req}}=\int F\text{d}s={\int }_0^{\gamma }{G}^{\prime }(\gamma )\gamma \text{d}\gamma$$ (2)

where γ is the applied strain. According to the calculation (Supporting Information), the mechanical energy attained from the coalescence of microbubbles (3.02 × 10⁻¹⁰ J) far outweighs E_req (2.03 × 10⁻¹² J). This result indicates that the fusion between microbubbles from the diatom bubbler generates mechanical energy high enough to rupture the biofilm, confirming the experimental observations. We further demonstrated that the MnO₂-diatom bubbler could remove the biofilm formed even in complex confined spaces (Figure 4e). The use of 3% H₂O₂ solution may cause corrosion on the metallic substrates. However, the MnO₂-diatom particles decompose H₂O₂ gradually for the microbubble generation. Consequently, the MnO₂-diatom bubbler not only improve antibacterial activity of H₂O₂ during treatment period but also neutralize H₂O₂ in the end, which decrease the possibility of substrate corrosion.

4. CONCLUSIONS

In summary, this study addresses that diatoms engineered to generate microbubbles and self-propel can invade and damage the biofilm (Figure 4f). In the antiseptic H₂O₂ solution, diatoms doped by MnO₂ nanosheets infiltrate the biofilm by microbubble propulsion. The diatom particles invading the biofilm continue to generate microbubbles that merge to deform and ultimately fracture the EPS matrix of biofilm. Then, H₂O₂ molecules diffuse into the biofilm more actively and induce bacterial cell death more efficiently than those free of the diatom bubbler as observed with the decrease in the cell viability over time. Past studies demonstrated various types of bubble-propelled particles and their applications.^33–37 However, to the best of our knowledge, this is the first study to assess the potential of active bubbling particles for cleaning biofouled surfaces. In addition, this study utilizes a unique particle system created by hybridizing a H₂O₂-decomposing MnO₂ nanosheets with naturally derived diatom skeletons. As the diatoms are abundant in nature, the diatoms doped with MnO₂ nanosheets can readily be produced on a large scale (Figure S8). Our results offer a promising path toward the effective development of active antibiofilm system in biomedical and environmental applications.