The silver nanoparticles used in this study were provided by the Center for the Environmental Implications and Nanotechnology (CEINT) at Duke University. Both Ag- NPs and Ag₂S-NPs were synthesized using a modified polyol process [27] [28] . The reported purities of both nanoparticles were 99%. Nanoparticles were provided as concentrated stock dispersions in water with concentrations of 1.8 g/L Ag for Ag-NPs and 1.1 g/L Ag for Ag₂S-NPs. Before each experiment, the stock dispersions were sonicated for 10 mins and diluted with ultra-pure milli-Q water to target concentrations. The size and morphology of nanoparticles were analyzed with a transmission electron microscope (TEM). The hydrodynamic diameters and surface charge of nanoparticles in solution were determined by Nanoseries zeta-sizer (Malvern, MA).

Pseudomonas aeruginosa GFP (ATCC® 10145GFP™) served as model organism in this study. The fluorescence and high attachment abilities of this model strain enhanced good characterization and analysis. The ability of the strain to emit green fluorescence protein was verified before experiments. Luria-Bertani (LB) broth, supplemented with 300 µg/mL of ampicillin as per manufacturer’s instructions, served as culture medium for all experiments. Prior to each experimental assay, P. aeruginosa was grown overnight from frozen stocks on LB agar plates at 37˚C. Then a few colonies were grown in suspension at 37˚C for 4 - 5 hours, harvested, and diluted in LB to a final optical density (OD 600) of 0.005, which was equivalent to 10⁸ cells/mL. This diluted subculture was then used to initiate biofilm cultures.

Biofilms were grown in a MBEC™ biofilm inoculator (Innovotech Inc. Technologies Canada) as previously described by Harrison et al. [29] with minor modifications. This device is commonly used for high throughput determination of the minimum biofilm eradication concentrations (MBECs) in laboratory settings. The device involves specialized MBEC™ polystyrene culture plates with protruding pegs (on the lids of the plates) that support surface-attached bacterial growth (i.e., biofilms).

In this study, each well of the MBEC™ 96-well plates was filled with 200 μL of P. aeruginosa suspension in LB at OD 600 of 0.005. The plates were then incubated in an environmental rotary shaker at a speed of 180 rpm and ambient temperature of 37˚C (Benchmark Scientific Incu-shaker) for 72 hours. Fresh LB medium was added at 24- hour intervals to replenish depleted nutrients. This protocol resulted in the formation of mature P. aeruginosa biofilms attached to the surfaces of the pegs on the lids of the MBEC™ plates.

Where applicable, optical density and cell concentration measurements were done with a Genesys spectrophotometer (Thermo Fisher Scientific, Waltham, MA) and a Multisizer™ 4 Coulter Counter (Beckman Coulter, Inc., Fullerton, CA) respectively. Biofilm growth was monitored with confocal laser scanning microscope (CLSM) and scanning electron microscopy (SEM) as described below. All experiments were performed under aseptic conditions, and all media and glassware used were autoclaved and sterile.

After 72 hours of incubation, MBEC™ pegs harboring mature P. aeruginosa biofilms were transferred to a new 96 well plate, filled with various concentrations of silver nanoparticles (in LB broth). Environmentally-relevant test concentrations of 0.3125 mg/L, 0.625 mg/L, 1.25 mg/L, 2.5 mg/L and 5 mg/L were used to challenge established biofilms [30] [31] [32] . Samples without nanoparticles and samples without biofilms both served as controls. Each experimental condition was replicated in four or more wells per assay plate. The plate was then incubated for another 24 hours at 37˚C with constant rotation of 180 rpm. After 24 h of incubation with nanoparticles, the biofilms were gently transferred into 200 µL of sterile phosphate-buffer saline (PBS) solution for 1 min to remove non-adherent biofilm cells. Thereafter, the attached biofilms and/or the residue LB broth in the culture wells were further analyzed as described in the following sections.

To determine the biomass in the biofilms, MBEC™ plate pegs were carefully removed and transferred to 1.5 mL centrifuge vials containing 200 µL of PBS. Thereafter, a water-bath sonicator (Quantrex L & R Ultrasonic, NJ, USA) set at 60 Hz for 10 minutes was used to disrupt biofilms into suspension. Recovered biofilm cells in suspension were pipetted into 96-well clear bottom black micro-well plates and examined with a multi-plate reader (Infinite M200 Pro Tecan microplate reader). The total biomass was determined by optical density (OD 600) measurement, and live population was determined by monitoring the green fluorescent protein (GFP) intensity (GFP Excitation: 475 nm; Emission: 570 nm) expressed by live P. aeruginosa GFP cells.

Confocal laser scanning microscopy (CLSM) imaging was conducted to visualize biofilm cells in-situ to determine the live/dead population distribution. After biofilms were harvested, sample pegs were then stained with propidium iodide for 30 minutes in the dark at room temperature and immediately imaged with CLSM (Zeiss LSM 710 Confocal Microscope). All sample pegs mounted for confocal microscopy were sequentially scanned frame-by-frame with a 63 x water immersion objective in fluorescence mode using FITC (green) and Rhodamine (red) filters. Confocal image visualization, analysis and two-dimensional (2D) projections of acquired Z-stacks were performed using Carl Zeiss Microscopy software systems (ZEN black edition).

Scanning electron microscopy (SEM) was used to visually examine the surface morphology of biofilms. All biofilm samples were rinsed with sterile PBS and fixed with a solution of 2.5% glutaraldehyde in 0.1 M cacodylate buffer at 25˚C for 2 hours. Thereafter, biofilm samples were air-dried for 7 days before SEM imaging. SEM measurements were performed after samples were sputter-coated with gold nanoparticles to enhance electron conductivity on sample surfaces. SEM images of fixed biofilms were taken with a JOEL JSM-7600F SEM at an accelerating voltage of 20 kV.

To further examine the effects of biofilm culture on silver ion release from NPs, silver dissolution experiments were conducted with nanoparticle concentrations of 0.625 mg/ L, 1.25 mg/L, 2.5 mg/L and 5 mg/L. These experiments were setup similarly to the aforementioned nanoparticle challenge experiments, with the following exception: residual media in the plate well were collected in lieu of the pegs on the lid. The contents from each well were transferred into 2 mL centrifuge vials individually. Therefore, the silver ions were separated from silver nanoparticles by a centrifuging method by Wirth et al. [14] . The samples were centrifuged (Eppendorf ultracentrifuge 5418) for 45 mins at a rotor speed of 10,000 rpm (4˚C). Subsequently, supernatants containing only dissolved silver were gently removed and diluted with 5% nitric acid and tested with Graphite Furnace Atomic Absorption Spectrometry (PerkinElmer Waltham, MA). The dissolved silver concentration in water was also tested for comparison.

The Ag-NPs and Ag₂S-NPs were characterized in terms of their morphologies, size distributions and surface charge. As shown in Figure 1, Ag-NPs and Ag₂S-NPs had a particle size distribution from 30 to 50 nm. Particle size distribution and surface charge may sometimes be influenced by organic matter present in LB media. However, the PVP surface coating stabilized the nanoparticles, preventing aggregation, even in the presence of organic material (Table 1). In addition, the zeta potentials of the nanoparticles were negatively charged, with electrophoretic mobilities (μ) of −11.10 ± 0.600 mV and −9.51 ± 0.636 mV for Ag-NPs and Ag₂S-NPs, respectively.

The antimicrobial activities of Ag-NPs against environmental and clinical bacteria have

been established [33] [34] [35] . A comparative evaluation of the respective activities of Ag-NP and its sulfidized counterpart, Ag₂S-NP, against P. aeruginosa biofilms was performed using fluorescence spectrophotometry, scanning electron microscopy (SEM) and confocal laser scanning microscopy (CLSM). Fluorescence spectrophotometry confirmed the ability of both pristine and sulfidized Ag-NPs to disrupt biofilms (Figure 2) and kill bacterial cells (Figure 3). Exposures in the experimental ranges tested (0.3 - 5 mg/L)

were not capable of fully eradicating the P. aeruginosa biofilms. Biofilm-growing bacteria such as P. aeruginosa are capable of withstanding up to 1000 times the concentrations of antimicrobial agents that would inhibit the growth of free-floating bacteria [36] . However, at all exposure concentrations, reduction in total biofilm biomass (determined by absorbance measurements at OD 600; Figure 2) corroborated well with the ability of NPs to kill bacterial cells (evaluated by monitoring the fluorescence activity of live bacterial cells; Figure 3).

Both NPs revealed a significant increase in antimicrobial activity when their exposure concentrations were doubled from 0.313 to 0.625 mg/L followed by a plateau in activity at higher concentrations (Figure 2 and Figure 3). However, Ag₂S-NPs revealed a significant lower ability to either disrupt established biofilms (Figure 2) or kill P. aeruginosa cells (Figure 3) than Ag-NPs at all concentrations tested. This is consistent with previous findings that the sulfidation of Ag-NPs may reduce nanosilver toxicity [37] [38] . The sole exception to this rule was observed with the least effective exposure concentration (0.313 mg/L) where Ag-NPs and Ag₂S-NPs only resulted in 27.2% and 26.0% reductions in total biofilm biomass (corroborated with 25.8% and 21.5% reductions in live bacteria cells) respectively. This contrasted with 67.9% and 49.2% reductions in total biomass (associated with 84.9% and 70.2% reductions in live cells) for Ag-NPs and Ag₂S-NPs respectively (Figure 2 and Figure 3).

To further assess the nanoparticle antimicrobial properties, pre-established P. aeruginosa biofilms were exposed to identical concentrations (0.625 mg/L) of Ag₂S-NPs and Ag-NPs and monitored with CLSM. SEM investigations of unchallenged P. aeruginosa cultures (Figure 4(A) and Figure 4(D)) validated the ability of the MBEC™ assays to

develop and sustain fully established biofilms on the surfaces of MBEC™ pegs. Again, Ag₂S-NPs were shown (Figure 4(B) and Figure 4(E)) to be less efficient at disrupting pre-established P. aeruginosa biofilms than Ag-NPs (Figure 4(C) and Figure 4(F)), corroborating the results achieved during quantification of total microbial biomass (Figure 2). Similar trends were observed with CLSM analysis. Corroborating previous live/dead assay findings (Figure 3), confocal images of biofilms challenged with NPs demonstrated the enhanced ability of Ag-NPs to kill P. aeruginosa cells as compared to Ag₂S-NPs (Figure 5). Taken together, these data emphasize the fact that sulfidation of Ag-NPs significantly impacts its antimicrobial properties, namely here its ability to kill biofilm-embedded cells and disrupt established P. aeruginosa biofilms. Since sulfidation of is one of the transformation processes that occur naturally in the environment, these data further suggest that the rate and extent of transformation of nanosilver should always be monitored and considered as a major factor when evaluating the potential impact of nanosilver on biological systems [26] .

The bioavailability of nanoparticles is greatly affected by the transport and diffusion of NPs into the biofilms. In this section, we examined the transport of Ag-NPs and Ag₂S NPs into biofilms. Firstly, SEM imaging of nanosilver challenged biofilms confirms the transport activity of both NPs. Significant amount of Ag-NPs and Ag₂S NPs were observed on the biofilms. As shown in Figure 6, the nanoparticles appear to form loose clusters on biofilms, and the majority of the nanoparticles remain unaggregated, most likely due to the hydrophilic PVP coating [38] .

Silver nanoparticles suspensions are a mixture of various forms: dissolved NPs and

dissolved silver ions. A difference in diffusion rate and transport into biofilms, as well as effective bioavailability for antimicrobial challenge, may exist between Ag-NPs and Ag₂S-NPs due to the differences in the transport abilities of dissolved NPs and diffusion of silver ions. Research has shown that silver nanoparticles can effectively transport from culture media to biofilms and that silver ions diffuse much faster than the dissolved NP forms [15] [39] . Additionally, the sulfidation process changes the dissolution rate of silver nanoparticles in water and reduces the release of silver ions. However, the presence of organic matters in culture solution may also alter the toxicity threshold [40] [41] .

Therefore, we investigated the dissolution of Ag-NPs and Ag₂S-NPs in culture in the presence of P. aeruginosa established biofilms. After an exposure period of 24 hours, the residual silver ion concentrations in the culture were tested. As shown in Figure 7, concentrations of dissolved silver were significantly higher in suspensions of Ag-NPs than those of Ag₂S-NPs across the test concentrations after exposure to biofilms. This suggests a higher rate of silver ions transport into P. aeruginosa biofilms. Together, the higher solubility of Ag-NPs and higher amount of transported silver ions into biofilms would correlate with the higher impacts on biomass and live populations in the challenged biofilms by Ag-NPs than Ag₂S-NPs. However, the difference in biofilm inhibition effects between Ag-NPs and Ag₂S-NPs were not proportional to the difference in solubility, this may result from the difference in the transport behaviors of the two NPs. Overall, the bioavailabity of silver nanoparticles is greatly reduced by sulfidation.