The Antimicrobial Properties of Silver Nanoparticles in Bacillus subtilis Are Mediated by Released Ag+ Ions

The superior antimicrobial properties of silver nanoparticles (Ag NPs) are well-documented, but the exact mechanisms underlying Ag-NP microbial toxicity remain the subject of intense debate. Here, we show that Ag-NP concentrations as low as 10 ppm exert significant toxicity against Bacillus subtilis, a beneficial bacterium ubiquitous in the soil. Growth arrest and chromosomal DNA degradation were observed, and flow cytometric quantification of propidium iodide (PI) staining also revealed that Ag-NP concentrations of 25 ppm and above increased membrane permeability. RedoxSensor content analysis and Phag-GFP expression analysis further indicated that reductase activity and cytosolic protein expression decreased in B. subtilis cells treated with 10–50 ppm of Ag NPs. We conducted X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) analyses to directly clarify the valence and fine structure of Ag atoms in B. subtilis cells placed in contact with Ag NPs. The results confirmed the Ag species in Ag NP-treated B. subtilis cells as Ag2O, indicating that Ag-NP toxicity is likely mediated by released Ag+ ions from Ag NPs, which penetrate bacterial cells and are subsequently oxidized intracellularly to Ag2O. These findings provide conclusive evidence for the role of Ag+ ions in Ag-NP microbial toxicity, and suggest that the impact of inappropriately disposed Ag NPs to soil and water ecosystems may warrant further investigation.


Introduction
Silver nanoparticles (Ag NPs) are the most widely used nanomaterial in healthcare today, with total annual worldwide production estimated to be in the range of 500 tons [1]. The superior antimicrobial, antifungal, and antiviral properties of Ag NPs mean that they are frequently present in coatings for bone prostheses, surgical devices, infusion systems, and dental composites [2][3][4], as well as air/water filters, food containers, textiles, and many other consumer X-ray absorption near edge structure (XANES) spectroscopy can provide information on the electronic configuration, stereochemistry, and oxidation state of Ag atoms [46][47], and extended X-ray absorption fine structure (EXAFS) spectroscopy can offer additional information regarding the atomic arrangement of Ag atoms, in terms of bond distance, coordination number, near-neighbor type, and thermal/static disorder [45,48]. Therefore, XANES and EXAFS analysis can elucidate the oxidation state and fine structure of Ag atoms in B. subtilis cells placed in contact with nanosilver, which would help to clarify Ag-NP microbial toxicity mechanisms.

Ag-NP synthesis and characterization
Silver nitrate (AgNO 3 , 0.85 g) and polyvinylpyrrolidone (PVP, 1.0 g) were dissolved in ethylene glycol (EG, 100 mL), with continuous stirring until a homogeneous solution was obtained. PVP was used here to stabilize the Ag suspension. After solution mixing, pulse microwaveassisted synthesis was conducted as follows: the Ag-containing solution was placed in the center of a household microwave oven (Tatung Co., 900 W, 2.45 GHz, Taiwan) in which one thermocouple had been equipped to detect reaction temperature. Pulse microwave-assisted synthesis was carried out at 40°C under a direct current power supply (8 V, 90 A, 720 W).
Microwave-assisted synthesis has been reported to produce consistently small Ag NPs that are uniform and stable at room temperature [26]. We proceeded to characterize our prepared Ag-NP samples in terms of structure, morphology, and elemental composition, using field emission scanning electron microscopy (FE-SEM), transmission electron microscopy (TEM), and X-ray diffraction (XRD). FE-SEM was conducted with a JEOL JSM-6701F field emission scanning electron microscope, to determine the morphology, microstructure, and particle size distribution of synthesized Ag NPs. TEM images were collected with a JEOL TEM-2010 scanning electron microscope. A single 20-μL drop of the silver nanoparticle suspension was placed on 400-mesh TEM grids with a carbon support film (Agar Scientific, Essex, UK). The dried preparations were rinsed with ethanol and dried again, before being mounted on the appropriate holder and placed in the microscope. Transmitted electron images were collected in bright field mode at an accelerating voltage of 200 keV, with the specimen set at an 8 mm working distance. Samples were analyzed within 1 hr to avoid shape degradation. We subsequently used Gatan DigitalMicrograph software to analyze representative TEM images, in order to determine the size of synthesized Ag NPs. A total of 100 nanoparticles were randomly selected for size assessment. The solid phase microstructure and crystallinity of Ag NPs was then characterized by XRD at a scanning range of 20-80°(2θ) and a scan rate of 4°(2θ)/min on a Rigaku RU-H3R diffractometer, using monochromatic Cu K α radiation with a wavelength of 1.5405 Å at 30 kV and 20 mA.
B. subtilis growth conditions and assessment of Ag-NP antibacterial effects B. subtilis wild-type strain 3610 [49] was maintained at 37°C in Luria-Bertani (LB; 10 g tryptone, 5 g yeast extract, 5 g NaCl per liter) broth or on 1.5% Bacto agar plates. Ag-NPs were added where appropriate to achieve the following final concentrations: 0, 1, 5, 10, 25, or 50 ppm. For growth assays conducted in a minimal medium [50], overnight cultures were diluted 100-fold in the minimal medium and grown at 37°C under shaking at 200 rpm for 12 h, with Ag NPs added where appropriate to achieve the following final concentrations: 0, 0.1, 1, or 10 ppm.
To assay time-dependent growth inhibition caused by Ag NPs, overnight cultures of approximately 1 × 10 9 CFU/mL were diluted 100-fold into 50 mL of LB broth in 250 mL flasks. Ag-NPs were added where appropriate, to achieve final concentrations of 0, 1, 5, 10, 25, or 50 ppm. Cultures were then grown for up to 30 h at 37°C, with shaking at 200 rpm. Bacterial growth was measured by optical density at 600 nm (OD 600 ). To assess the antibacterial effects of Ag NPs against B. subtilis, initial cultures (1 × 10 9 CFU/mL) were prepared from 50-mL LB liquid cultures harvested at exponential growth. Bacterial cells were treated with Ag NPs at increasing concentrations of 0-50 ppm for 5 h, with shaking at 200 rpm. Both treated and untreated cultures were then serially diluted and plated on LB agar plates at the time points indicated. Plates were incubated overnight at 37°C and then subjected to a colony count [51]. All experiments were performed in triplicate and averaged, and statistical analysis indicated the results were of good reliability.

Chromosomal DNA integrity analysis
Chromosomal DNA was isolated from B. subtilis cells grown in LB medium with final concentrations of 0, 1, 5, 10, 25, or 50 ppm of Ag NPs, using a Wizard Genomic DNA Purification kit (Promega, Madison, WI, USA). Each lane was loaded with~1 μg DNA, and gel electrophoresis was performed on a 1.0% agarose gel for 30 min at 120 V.

Transmission electron microscopy (TEM) of bacterial cells
Bacterial cells were fixed in 2% v/v paraformaldehyde/3% v/v glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) for 2 h at 4°C. Samples were then washed in 0.1 M cacodylate buffer three times, and postfixed in 1% osmium tetroxide at 4°C for 1 h. Bacterial cells were then washed three times as described above, dehydrated in a graded series of ethanol, and embedded in Eponate 12 resin (Ted Pella Inc., Redding, CA) for 7 h. Sections of 70-80 nm in thickness were sliced and stained with uranyl acetate and lead citrate, then viewed under a Hitachi transmission electron microscope (H-7500, Tokyo, Japan) at 200 kV. TEM montage images were manually acquired at 75,000× magnification.

Measurement of RedoxSensor activity
RedoxSensor activities of B. subilis 3610 were determined using a BacLight TM RedoxSensor TM Green Vitality Kit (Molecular Probes, Eugene, OR, USA). Overnight cultures of B. subtilis 3610 were treated with the indicated concentrations of Ag-NPs for 3 h at 37°C. Cells were then washed and diluted 10-fold in 1× PBS buffer, then mixed with 1 μL of RedoxSensor TM Green reagent and vortexed. To assess membrane integrity, 1 μL of propidium iodide (PI) was added, and the mixture was incubated in the dark at room temperature for 5 min. Stained cells (10 μL) were spotted onto a clean slide and covered with a poly-L-lysine treated coverslip. Slides were observed under a Leica TCS-SP2 laser-scanning confocal microscope at a magnification of 630×.
Assessing Ag-NP impact on P hag -GFP expression Plasmid pHag-gfp was constructed by inserting a DNA fragment containing the gfp sequence transcribed from the hag promoter into pHY300PLK (Takara, Shiga, Japan). B. subtilis 3610 (P hag -GFP) was treated with Ag NPs for 3 h at 37°C. Bacterial cultures (3 mL) were centrifuged and washed with 300 μL 1× T-Base buffer [15 mM (NH 4 ) 2 SO 4 , 80 mM K 2 HPO 4 , 44 mM KH 2 PO 4 , 3.4 mM sodium citrate, and 3 mM MgSO 4 ], and bacterial cells were resuspended in 50 μL 1× T-Base buffer containing 10 μg/mL FM1-43FX and 5 μg/mL DAPI. The mixture was incubated in the dark at room temperature for 15 min, and 4 μL of stained cells were spotted onto a clean slide and covered with a poly-L-lysine treated coverslip. Slides were observed under a Leica TCS-SP2 laser-scanning confocal microscope at a magnification of 3,150×.

Flow cytometry analysis of P hag -GFP expression and RedoxSensor activity
For the P hag -GFP expression assay, bacterial cultures were treated with indicated concentrations of Ag NPs for 3 h at 37°C, after which 3 mL of culture was centrifuged and washed with 300 μL of 1× T-Base buffer. Cells were resuspended in 1 mL of 1× T-Base buffer, and flow cytometry was directly performed on a FACSCalibur flow cytometer (BD Biosciences, San Jose, CA, USA). Fluorescence filters and detectors were all standardized, with green fluorescence collected in the FL1 channel (530 ± 15 nm). All parameters were collected as logarithmic signals. For the RedoxSensor activity assay, bacterial cells were treated with the indicated concentrations of Ag NPs, washed and diluted 10-fold in 1× PBS buffer, then mixed with 1 μL of a 1:10 dilution of RedoxSensor TM Green reagent and vortexed. To assess membrane integrity, 1 μL of a 1:10 dilution of propidium iodide (PI) was added, and the mixture was incubated in the dark at room temperature for 5 min. Samples (1 mL) were assayed by flow cytometry using a FACSCalibur flow cytometer and standardized fluorescence filters and detectors. Green fluorescence was collected in the FL1 channel (530 ± 15 nm), and red fluorescence was collected in the FL3 channel (> 650 nm). All parameters were collected as logarithmic signals. Data were analyzed using CellQuest Pro software. In density plots of light scatter properties, bacterial cells were gated from irrelevant counts for fluorescence analyses. Flow cytometry was calibrated using BD Calibrite beads (BD Biosciences, San Jose, CA, USA). Data are representative of results derived from two separate experiments.

XANES and EXAFS analyses
XANES and EXAFS results were collected at the Wiggler beamline 01C1 in the National Synchrotron Radiation Research Center (NSRRC) of Taiwan. The electron storage ring was operated at an energy level of 1.5 GeV and a current of 100-200 mA. An Si(111) DCM was used for providing highly monochromatized photon beams with an energy of 6-33 keV (BL01C1) and resolving power (E/4E) of up to 7,000. Data were collected in fluorescence or transmission mode with a Lytle ionization detector [48] for Ag (25,514 eV) K-edge experiments at room temperature. Photon energy was calibrated by characteristic pre-edge peaks in the absorption spectra of metallic silver standards. Local structural parameters, such as the bond length (R), coordination number (CN), and Debye-Waller factor (σ) for different coordination shells surrounding the absorbing atoms, were obtained through non-linear least-square fit. Raw absorption data in the region of 50-200 eV below the edge position were also fit to a straight line, using least-square algorithms. XANES analysis was extended to energy levels of the order of 50 eV above the edge. Spectra were measured with a step size equivalent to less than 0.5 eV in the near-edge, and with a count time weighted to be proportional to k 3 at high energy. Data were normalized using the program Athena (VI), with a linear pre-edge and polynomial postedge background subtracted from the raw ln(I t /I 0 ) data, and then analyzed using the Artemis (VI) software, which makes use of the FFEF code-8 [52][53][54]. After calibration, samples were background-corrected, using a linear pre-edge region and a polynomial for the post-edge region, and subsequently normalized. EXAFS energy spectra were then converted to wavevector K space. The resulting scatter curve was weighed by K 3 to enhance dampened scattering oscillations. This curve was followed by Fourier transformation to yield the radial structure function [54]. These data directly reflect the average local environment around the absorption atoms. Spectra were analyzed using the software package IFEFFIT [52][53]. The theoretical paths for Ag-Ag, Ag-S, and Ag-O species used for fitting the first coordination shell of the experimental data were generated using the FEFF-8 program, based on the crystallographic data of individual species [52][53]. The coordination number, interatomic distance, Debye-Waller factor, and inner potential correction were used as variable parameters for the fitting procedures. In the case of XANES spectra, the intensity of the pre-edge peak, centered at 25,514 eV and present in the Ag K-edge spectra of the different samples, was used to estimate the relative amount of Ag 0 or Ag + species.

Results
Ag NPs induce growth inhibition and chromosomal DNA degradation in B. subtilis Ag NPs can vary broadly in terms of size and morphology [32], and as Ag-NP size has been cited as a determining factor in toxicity [23], it is important to assess the diameter and structure of Ag-NPs used in toxicity experiments, in order to allow for meaningful comparison across studies. Ag NPs used in this study were derived via pulse microwave-assisted synthesis and characterized by FE-SEM, XRD, and TEM. FE-SEM micrographs revealed the average particle diameter of Ag NPs to be about 10 nm (Fig 1A and 1D). TEM images corroborated these results (Fig 1B and 1C), and a random count of 100 Ag NPs revealed the mean particle size to be about 10 (Fig 1E), which correspond to 2θ = 38.4, 43.3, 64.6, and 77.7, based on the band for face-centered cubic (FCC) structures of silver (JCPDS Card Number 87-0597). No peaks were observed in any other phases, indicating that single-phase Ag NPs with a cubic structure were obtained. The overall pattern was comparable with XRD spectra previously reported for such Ag-NPs in other studies [55].
B. subtilis is one of the best-characterized model organisms for Gram-positive bacteria, and is ubiquitously present in soil and water ecosystems, where it is known to exert a broad range of beneficial effects [37][38][39][40][41]56]. We therefore sought to assess the effects of Ag NPs upon B. subtilis, in order to elucidate the toxicity mechanisms involved and to gauge the potential impact of Ag-NP exposure. We treated B. subtilis 3610 cultures with 0-50 ppm of Ag NPs, and evaluated bacterial growth over a period of 12 h in a rich LB medium. At concentrations > 5 ppm, Ag NPs inhibited bacterial growth for 12 h (Fig 2A), and lethal effects were observed at Ag-NP concentrations ! 10 ppm. The effect of Ag NPs was also tested on B. subtilis cells grown in a low nutrient medium, and it was found that 0.1 ppm of Ag NPs was sufficient to halt cell growth for over 12 h (Fig 2B). These results are in line with the general understanding that cultures in minimal medium are more sensitive to Ag NPs, compared with cultures in rich medium. We further added Ag NPs to approximately 1 × 10 9 CFU/mL of freshly grown B. subtilis cultures, and incubated the mixtures for 5 h. Treatment with 25 or 50 ppm of Ag NPs led to a 2 log reduction of colony-forming units after 1 h of incubation, while treatment with 10 ppm of Ag NPs achieved a similar result after 3 h of incubation (Fig 2C), thus confirming that Ag NPs can hinder B. subtilis growth and entry to the exponential phase; moreover, significant lethality was observed with Ag-NP concentrations ! 10 ppm (Fig 2C).
Previous in vitro research has shown that Ag NPs are capable of inducing strand breaks in isolated plasmid DNA [57]; however, few in vivo studies examining the effect of Ag NPs on chromosomal DNA are available. We therefore used agarose gel electrophoresis to examine the chromosomal DNA integrity of B. subtilis cells treated with 0-50 ppm concentrations of Ag NPs. DNA electrophoresis on a 1.0% agarose gel revealed significant degradations in the chromosomal DNA of B. subtilis cells treated with 10-50 ppm of Ag NPs ( Figure A in S1 File), suggesting that Ag-NP concentrations of 10 ppm or greater are sufficient to compromise chromosomal DNA integrity in bacterial cells. We further employed TEM to observe the morphological changes in B. subtilis cells following Ag-NP treatment, and found that bacterial chromosomal DNA became looser at Ag-NP concentrations greater than 10 ppm ( Figure B in S1 File), suggesting that chromosomal DNA integrity and cell morphology may be adversely affected by Ag-NP concentrations as low as 10 ppm.
Ag NPs reduced B. subtilis RedoxSensor activity and P hag -GFP expression Ag NPs have been said to increase reactive oxygen species (ROS) levels in cells [21], and thus we sought to examine whether reductase activity in B. subtilis was affected by Ag NPs. Cells were grown to early stationary phase at approximately 10 9 CFU/mL, then treated with 0-50 ppm of Ag NPs. After 3 h of treatment, cells were washed, and approximately 10 9 CFUs were subjected to staining with RedoxSensor Green TM , a fluorescent dye that yields green fluorescence (488 nm excitation) when modified by reductases [58], and PI (red), to assess reductase activity and membrane integrity. RedoxSensor activity was quantified using flow cytometry, and results showed that treatment with Ag NPs led to general declines in the geometric mean of reductase activity levels (0 ppm: 390.89; 5 ppm: 576.67; 10 ppm: 564.54; 25 ppm: 63.13; 50 ppm: 66.81) and the percentage of gated cells in the total cell population (0 ppm: 90.38%; 5 ppm: 92.71%; 10 ppm: 76.59%; 25 ppm: 0.03%; 50 ppm: 0.05%), with significant decreases observed following treatment with ! 25 ppm of Ag NPs (Fig 3A). Furthermore, Ag-NP concentrations above 25 ppm were found to affect the integrity of cell membranes after just 3 h of treatment, with dramatic increases seen in the percentage of cells exhibiting PI fluorescence (0 ppm: 8.18%; 5 ppm: 6.22%; 10 ppm: 17.79%; 25 ppm: 89.51%; 50 ppm: 90.52%), indicating increased membrane permeability (Fig 3B). Together, these results show that after treatment with 25 or 50 ppm of Ag NPs, reductase activity decreased significantly in B. subtilis cells, while membrane permeability increased significantly (Fig 3A and 3B).
We next sought to ascertain the effect of Ag NPs on protein expression. In a separate experiment, B. subtilis 3610 (P hag -GFP) was grown until early stationary phase, at approximately 10 9 CFU/mL, and treated with varying concentrations of Ag-NPs for 3 h. P hag -GFP expression was then quantified using flow cytometry. Results showed that both levels of P hag -GFP expression  (Fig 3C). These results were further confirmed with fluorescent micrographs (Fig 3D), which mirrored the flow cytometry findings. This suggests that Ag-NP concentrations of 25 ppm and above can adversely impact P hag -GFP and cytosolic protein expression in B. subtilis cells.

XANES/EXAFS analysis of Ag-NP fine structure in B. subtilis cells
The question of whether Ag NPs exert antimicrobial effects through inherent toxicity or the release of Ag + ions is yet to be resolved. We therefore wished to characterize the oxidation state and fine structure of Ag atoms in B. subtilis cells treated with Ag NPs, using XANES and EXAFS. Near-edge structure spectra can provide information on Ag valence states, and can be fitted to spectra derived from established standards to determine the oxidation state of Ag atoms in B. subtilis cells treated with Ag NPs. The normalized silver K-edge spectra of B. subtilis cells treated with 100 ppm of Ag NPs, as well as spectra from silver standards (Ag, AgO, Ag 2 O, and Ag 2 S) are presented in Fig 4A, while derivative spectra are shown in Fig 4B. Several sharp absorption peaks in the range between 25,500 and 25,650 eV were observed in the silver standards. At 25,528 eV, spectra derived from B. subtilis cells treated with 100 ppm of Ag NPs overlapped with Ag, AgO, Ag 2 O, and Ag 2 S standards. However, at 25,538 and 25,569 eV, patterns only fit the Ag 2 O and Ag 2 S standards (Fig 4A and 4B). The pattern at 25,590 eV only fit the Ag 2 O standard, indicating that the Ag species in Ag NP-treated B. subtilis cells is very likely to be Ag 2 O (Fig 4B).
EXAFS analysis has elemental specificity and is highly sensitive, and therefore can be used to determine the chemical states of target species that are present at very low concentrations. We used EXAFS to analyze the types of neighbors, bond length (R), and coordination number (CN) of silver atoms in B. subtilis cells treated with 100 ppm of Ag NPs. Fig 5A-5D shows the k 3 -weighted and least-square fitted Ag K-edge EXAFS inverse Fourier transformation oscillation spectra from silver standards (Ag, AgO, Ag 2 O, and Ag 2 S) and B. subtilis cells treated with 100 ppm of Ag NPs. Contributions from the electronic scatter effect between Ag atoms in the first electron shell were ascertained through inverse Fourier transformation of EXAFS spectra between 1 to 2 Å, and the results showed that spectra from the silver standards, Ag 2 O (Fig 5A), Ag 2 S (Fig 5B), and AgO (Fig 5C), closely fitted spectra derived from Ag NP-treated B. subtilis cells. Further analysis revealed that the bond distance spectra of Ag NP-treated B. subtilis cells (black) clearly overlapped with Ag 2 O (red) at 1.55 Å (Fig 5E), thus confirming the coordination shell as Ag(I)-O. Moreover, an elemental analysis of Ag NP-treated B. subtilis cells indicated that only limited levels of sulfur (S) were present (Table C in S1 File), thereby excluding the possibility of the silver species being primarily present as Ag 2 S. Quantitative analysis showed  that the bond length and coordination number of Ag 2 O in B. subtilis cells treated with Ag NPs was 2.06 Å and 1.00, respectively; and R-factor analysis further confirmed the precision of EXAFS model fitting (Ag 2 O < AgO < Ag 2 S < Ag, with lower R-factors indicating tighter fits) ( Table 1). These results show that Ag-NP microbial toxicity likely stems from Ag + ions released from Ag NPs (Ag 0 ), which can enter B. subtilis cells to exert toxic effects, and subsequently become oxidized intracellularly to become Ag 2 O.

Discussion
In this study, we found that Ag-NP concentrations exceeding 5 ppm were capable of inhibiting B. subtilis growth in rich medium (Fig 2A), while concentrations of just 0.1 ppm was sufficient to halt growth in minimal media cultures of B. subtilis (Fig 2B). At Ag-NP concentrations of 10 ppm and above, a 2 log reduction in bacterial colony-forming units was observed (Fig 2C). These results are in line with previous observations made by Kim et al., who found that Ag NPs  averaging 10 nm in diameter were able to inhibit B. subtilis growth, inducing a belated log phase at concentrations of 5 ppm [59]. Interestingly, we found that 0.1 ppm concentrations of Ag NPs was able to halt B. subtilis growth in minimal media, and this may explain the 35% reduction in microbial mass observed for soil samples treated with 0.14 mg Ag/kg soil in a recent field scenario study [43]. In addition, a previous study found that 100 ppm concentrations of 50-nm Ag NPs led to a 2 log reduction in B. subtilis cell numbers, an observation that supports our findings and remains consistent with the understanding that smaller nanoparticles are more toxic to cells and can therefore inhibit cell growth at lower concentrations [35]. However, a recent study showed that 3-nm Ag NPs were only able to inhibit B. subtilis growth at concentrations of 20 ppm [36], which is puzzling as these Ag NPs were clearly smaller than 10 nm, but did not demonstrate greater toxicity or inhibitory effects. This suggests that the shape of Ag NPs may also be an important factor in determining toxicity, as shape and structure could potentially affect the ability of Ag NPs to dissolve in liquids or release Ag + ions [18]. In addition, it is known that the size and shape of Ag NPs can affect their plasmonic properties [29,31], which in turn may have bearing on the physical and biological activity of these nanoparticles. Further research will be needed to assess whether changes in surface plasmon resonance can modulate Ag-NP antimicrobial activity. Furthermore, it has been reported that in biological fluids such as blood, the plasmonic properties and high surface energy of nanoparticles can attract proteins to form a coating, known as the protein corona [60][61]. The protein corona may affect the physical, chemical, and biological properties of nanoparticles in ways that are yet to be understood; moreover, the composition of proteins in the corona varies between nanoparticles and evolves dynamically over time, making it difficult to pinpoint any specific corona effects [61]. Compared to blood or lymphatic fluid, bacterial culture medium may be less complex, but in future experiments, it would be interesting to assess the presence and effect of protein coronas upon Ag NPs in bacterial cultures. More importantly, further research may be needed to understand how protein coronas could affect the impact of Ag NPs inadvertently released into soil or water. Previous research has shown that Ag NPs can cause membrane damage in Gram-negative Escherichia coli [62], but the results were derived from energy filtered TEM analysis, a technique known to induce slight membrane damage during sample preparation and analysis. In light of this, we elected to use in vivo PI staining to confirm if Ag NPs could disrupt bacterial cell membranes in B. subtilis. To our knowledge, this is the first study to examine the effect of Ag-NPs on bacterial membrane integrity using the PI assay. Our findings revealed that at Ag-NP concentrations ! 25 ppm, the number of cells with membrane damage increased (Fig 3B), and further indicated that membrane degradation occurred as a result of increased permeability induced by Ag NPs. It is possible that Ag-NP treatment damaged cellular membranes to increase permeability, or penetrated bacterial cell membranes to cause protein dysfunction. Our results further showed that reductase activity ( Fig 3A) and P hag -GFP expression levels ( Fig  3C and 3D) were lowered by Ag NPs, suggesting that decreased protein expression and/or an enhanced ROS response may also be involved in loss of membrane integrity and cell viability.
There is an ongoing debate over whether Ag NPs inherently possess toxicity, or if microbial toxicity is dependent on leached Ag + ions from nanoparticles. To clarify this issue, we utilized XANES/EXAFS analysis to assess the fine structure of silver particles present within B. subtilis cells treated with 100 ppm of Ag NPs. Following comparison with the spectra of silver standards (Ag, AgO, Ag 2 O, and Ag 2 S), we found that Ag 2 O best fits the spectra derived from Ag NP-treated B. subtilis cells (Figs 4 and 5A-5E, and Table 1). This data strongly suggests that in aqueous solutions, Ag NPs leach off Ag + ions, which subsequently enter bacterial cells and are oxidized to form Ag 2 O. If Ag NPs entered cells directly, it would have been possible to detect spectra corresponding to the Ag standard. To the best of our understanding, this is the first study to directly analyze silver particles present within bacterial cells treated with Ag NPs, and the results indicate that Ag + ions are primarily responsible for Ag-NP microbial toxicity.
In summary, our results support the theory that Ag NPs exert microbial toxicity through the release of Ag + ions that subsequently penetrate into bacterial cells, and to the best of our knowledge, this is the first study to provide direct evidence of this through XANES/EXAFS analyses. We further show that Ag NPs can inhibit the growth of Gram-positive B. subtilis bacteria, and exert toxicity by damaging cellular membranes, degrading chromosomal DNA, lowering reductase activity, and reducing protein expression. These results highlight the risks of engineered nanoparticle contaminants to soil and water ecosystems, and suggest that further research is needed to better understand the potential impact of Ag-NP contamination.