Size-Dependent Antimicrobial Effects of Novel Palladium Nanoparticles

Investigating the interactions between nanoscale materials and microorganisms is crucial to provide a comprehensive, proactive understanding of nanomaterial toxicity and explore the potential for novel applications. It is well known that nanomaterial behavior is governed by the size and composition of the particles, though the effects of small differences in size toward biological cells have not been well investigated. Palladium nanoparticles (Pd NPs) have gained significant interest as catalysts for important carbon-carbon and carbon-heteroatom reactions and are increasingly used in the chemical industry, however, few other applications of Pd NPs have been investigated. In the present study, we examined the antimicrobial capacity of Pd NPs, which provides both an indication of their usefulness as target antimicrobial compounds, as well as their potency as potential environmental pollutants. We synthesized Pd NPs of three different well-constrained sizes, 2.0±0.1 nm, 2.5±0.2 nm and 3.1±0.2 nm. We examined the inhibitory effects of the Pd NPs and Pd2+ ions toward gram negative Escherichia coli (E. coli) and gram positive Staphylococcus aureus (S. aureus) bacterial cultures throughout a 24 hour period. Inhibitory growth effects of six concentrations of Pd NPs and Pd2+ ions (2.5×10−4, 10−5, 10−6, 10−7, 10−8, and 10−9 M) were examined. Our results indicate that Pd NPs are generally much more inhibitory toward S. aureus than toward E. coli, though all sizes are toxic at ≥10−5 M to both organisms. We observed a significant difference in size-dependence of antimicrobial activity, which differed based on the microorganism tested. Our work shows that Pd NPs are highly antimicrobial, and that fine-scale (<1 nm) differences in size can alter antimicrobial activity.


Introduction
The unique chemical and physical properties of nanoparticles (NPs) relative to their bulk counterparts continues to be the driving force behind the discovery of novel applications in several technological areas. Nanoparticles are classified as particles with one dimension within the 1-100 nm size range and behave as a single unit with respect to transport and reactivity. [1] It has been demonstrated that NPs have a wide variety of unique applications, including cell targeting, [2,3] intravenous nucleic acid delivery, [4][5][6] environmental remediation, [7][8][9] catalysis, [10][11][12] and bactericidal effects. [13][14][15] In particular, silver (Ag) NPs are of pharmaceutical interest because of their potential to replace more traditional synthetic antimicrobial drugs and avoid unintended selection for drug-resistant pathogenic bacterial strains. [16] Multidrug resistant pathogenic outbreaks are one of the most important global health issues we face today (e.g. extensively drug-resistant tuberculosis, [17] multi-drug resistant cholera [18] and methicillinresistant Staphylococcus aureus [19]) yet have also led to a decrease in investment by pharmaceutical companies into antimicrobial drug development efforts to avoid economic losses when drug-resistant strains develop. [20,21] As NP research continues, this class of chemicals may provide unique possibilities to address this complex issue. However, as new NP applications continue to emerge, the amount of NPs present in consumer products will inevitably rise, likely leading to large quantities of NPs interacting with humans and the natural environment. [22] In order to conscientiously increase the use of NPs to benefit technological and medical advances, it is essential to understand their implications for human safety and ecosystem health [23][24][25].
While some NPs have been found to have antimicrobial properties (i.e. Ag NPs), a first step toward understanding the broader implications of the impact of novel NPs for any application is to characterize how novel NPs interact with simple biological organisms, such as bacteria. [25][26][27][28][29][30][31][32][33][34][35][36][37][38] Studies with singlecelled microorganisms can provide a framework for examining relative toxicity of NPs to more complex organisms, multi-trophic level interactions and within complex environmental conditions. They can also provide insight as to whether novel NPs could be useful antimicrobial substances that warrant further study for drug development.
In the present study, we report the first investigation of the activity of well-defined size-controlled palladium (Pd) NPs against bacterial growth. Palladium is one of the most widely used transition metals for carbon-carbon and carbon-heteroatom cross-coupling reactions [39] such as the Suzuki-Miyaura reaction, [40] Heck reaction, [41][42][43][44][45] Kumada reaction, [46] Sonogashira reaction, [47] Negishi reaction, [48] Stille reaction, [49] Buchwald-Hartwig reaction, [50] and hydrogenation reactions. [51] Palladium has been used as a catalyst to manufacture pharmaceuticals, [39] degrade harmful environmental pollutants, [52] and as sensors for the detection of various analytes. [53][54][55][56] Additionally, Pd and Pd 2+ ions also play a fundamental role in several biotechnological processes. [56] For example, Baccar et al. developed a non-enzymatic biosensor using various sizes of Pd NPs to detect hydrogen peroxide in milk. [56] While the uses of Pd are extensive, advances are yet to be uncovered as the metals are reduced to the nanoscale. Therefore, it is important to determine a baseline of toxicity for Pd NPs as well as examine their potential for antimicrobial applications.
Here, we examine the inhibitory effects of Pd NPs toward gram negative E. coli and gram positive S. aureus over a 24 h test period.
We synthesized three sizes of NPs, each with a narrow size distribution: 2.060.1 nm, 2.560.2 nm and 3.160.2 nm. This work is unique because it examines the effect of NPs synthesized within a very narrow size distribution and it also distinguishes the effects of NPs that vary within a size rage of 0.5 nm. We investigated the influence of Pd NP size, and examined the stability of these NPs throughout the toxicity tests. We also compared the toxicity of Pd NPs relative to Pd 2+ ions since in some cases Pd 2+ ions are being replaced by Pd NPs for catalytic applications. This data provide the first assessment of size-dependent antimicrobial effects of novel Pd NPs.

Materials for Nanoparticle Preparation
Palladium acetate [Pd 3 (OAc) 6 ] (OAc = acetate) was purchased from Strem Chemicals (Newbury Port, MA). Dodecyl sulfide was purchased from the Sigma Aldrich Company (Minneapolis, MN). Absolute ethanol (200 proof -ACS grade) was purchased from PHARMCO-AAPER (Shelbyville, KY). We purchased all chemicals and used them as received, without further purification.

Instrumentation
Transmission Electron Microscopy (TEM). The particle size and size distribution of the Pd nanoparticles in suspension were examined using a Model JEM-1230 JEOL TEM. We dried a 1 mL aliquot of Pd nanoparticles on a 400 mesh Formvar-coated copper grid and placed it in a vacuum desiccator for 5 h.
X-ray Diffraction (XRD). Powder x-ray diffraction (XRD) data were collected using a Scintag XDS Model 2000 diffractometer to analyze the crystalline properties of the Pd NPs. For XRD measurements, we dried and mixed Pd NP samples with 325 mesh Si powder and placed them on a Si wafer sample holder.

High-resolution Transmission Electron Microscopy
(HRTEM). We obtained high resolution TEM images and selected area electron diffraction patterns of the Pd NPs using a JEOL 3011 HRTEM at the Center for Microscopy at Michigan State University (East Lansing, MI).

Synthesis of Pd Nanoparticles (NP) from [Pd 3 (OAc) 6 ] by Pyrolysis
To synthesize Pd NPs, we added Pd 3 (OAc) 6 (0.05 g, 0.075 mmol) and n-dodecyl sulfide (0.14 g, 0.37 mmols) to 30 mL of ethanol. We heated the reaction mixture at 90uC for 3 h. We carried out similar reactions in ethanol at 90uC for 1 h and 2 h. The color of the solution turned from orange to dark brown. We removed a 1 mL aliquot of the resulting solutions for TEM imaging. Next, we diluted 10 mL of each of the reaction solutions with 100 mL of milli-Q water (18 MV?cm 21 ). We removed a 1 mL aliquot of the diluted solutions for TEM imaging as well. We prepared a solution of 2.5610 24 M Pd 2+ ions by dissolving a known amount of Pd 3 (OAc) 6 (0.0084 g, 0.25 mmol L 21 ) in 50 mL of milli-Q water as a control.

Inhibitory Effects of Pd NPs to Bacteria
Stock cultures of Escherichia coli (ATCC strain number 11303) and Staphylococcus aureus (ATCC strain number 12600) were obtained from Dr. Silvia Rossbach, Western Michigan University. E. coli is a rod-shaped, flagellated microorganism with a gram negative cell wall; S. aureus is a coccoid-shaped microorganism with a gram positive cell wall. We grew cultures of E. coli and S. aureus on sterile Tryptic Soy (TS, Fisher Scientific) plates for 24 h at 37uC. We removed a single colony of each microorganism from a plate and re-cultured it in 10 mL of sterile Bacto Tryptic Soy broth (Fisher Scientific) under the same growth conditions with constant agitation at 200 rpm in a shaking incubator. We then diluted pure cultures from the broth in 200 mL of sterile deionized water to an optical density of 0.889 for E. coli and 0.795 for S. aureus. We used a BioPhotometer (Eppendorf) to determine the optical density at 600 nm wavelength.
The Pd NPs were prepared and characterized within 24 h of the start of each antimicrobial test as described above. Within 2 h of the start of each antimicrobial test, we prepared a series dilution for each of the different sizes of Pd NPs. The starting concentration of all Pd NPs and Pd 2+ ions was 2.5610 24 M. The dilution series we used to test for antimicrobial activity included the 2.5610 24 M stock, a 1:10 dilution of the stock in sterile deionized water, and a series of five 1:10 dilutions in sterile deionized water. The lowest concentration we tested was 2.5610 29 M. All dilutions were prepared under a sterile hood. We prepared 10 mL of each dilution concentration, plus 10 mL of sterile deionized water as a control; all dilutions and treatments were performed in triplicate. To begin the antimicrobial test, we added 1 mL of the diluted bacterial culture to each dilution tube. Immediately after adding the culture to the dilution tube we vortexed the tube, and plated 75 mL of the treatment onto a TS agar using a spread plating technique. This was the initial (0 h) time point of the experiment. We kept the experimental tubes at 25uC during the course of the 24 h  Toxicity of Palladium Nanoparticles PLOS ONE | www.plosone.org experiment. We repeated the spread plate procedure at 4 h, 8 h and 24 h after initial exposure of the bacteria to the test chemical. We incubated the plates at 37uC for 20 h to grow the bacterial cultures and used a standard darkfield Quebec colony counter (Fisher) to manually count the number of colony forming units (CFU) present on each plate after growth. We chose the absorbance and dilution levels for the two microorganisms based on preliminary data indicating that this procedure would result in comparable and countable colonies (250-300 per plate) when 75 mL of the final dilution was grown using a spread plate procedure. Of the 1440 plates used in this experiment, we classified 48 plates as ''uncountable'' due to colony spreading. These plates were not included in our statistical analyses of the results. We observed consistent growth in the positive controls for both assays, indicating that comparisons of percent mortality over time were experimentally appropriate. For the S. aureus tests, the average CFUs growing in the deionized water controls was 256610 at 0 h, 22265 at

Statistical Analyses
We conducted Repeated Measures ANOVAs, One-Way ANOVAs and Tukey's HSD Pairwise Comparisons using Systat Version 13 (SPSS, Inc.) and evaluated significance at a = 0.05. We determined significant differences in colony growth from the control (as presented in the text) and significant differences among the four test chemicals.

Nanoparticle Fabrication, Size Control, and Characterization
We prepared Pd NPs from Pd 3 (OAc) 6 in the presence of ndodecyl sulfide (M:L = 1:5) in ethanol by a modified pyrolysis reaction. [57] The NP sizes were controlled by varying the length of time of the reaction. After 15 minutes of heating, the reaction solution (for all reactions: 1 h, 2 h, and 3 h) changed color from orange to dark brown, indicating NP formation. We placed a 10 mL aliquot of each reaction mixture in separate 125 mL Erlenmeyer flasks and diluted it with 100 mL of milli-Q water. We removed a 1 mL aliquot of each reaction mixture to prepare a TEM grid. Based on Pd content the concentration of the solutions was 2.5610 24 M. We imaged the NPs to characterize the size and distribution using TEM. Figure 1 shows TEM images of the assynthesized Pd NPs and their corresponding size histograms. The images show a homogeneous dispersion of the Pd NPs with an adopted spherical shape indicating isotropic growth. We determined the following sizes of the NPs from the TEM images: (a) 2.0 nm, (b) 2.5 nm, (c) 3.1 nm. The TEM images also depict nonaggregated Pd NPs with well controlled sizes.
The HRTEM image (Figure 2a) shows the twinning feature which is typical for most metal nanoparticles. The selected area electron diffraction (SAED) pattern shown in Figure 2b of a single Pd crystal exhibits 5 diffused rings which are assigned to the (111), (200), (220), (311) and (222) reflections that are characteristic of a face-centered cubic (fcc) structure. [57] The powder x-ray diffraction (XRD) (Figure 3) results of the as-synthesized Pd NPs exhibited (111), (200), (220), and (311) diffraction peaks at 40u, 47u, 68u, 82u and120u respectively, and were consistent with the fcc structure of Pd. We calculated the average particle size based on the line broadening of (111) by using the Scherrer formula. Our calculations show that the size is 2.5 nm, which is consistent with the value of the particle measured by TEM (Figure 1), indicating that TEM provides an accurate assessment of Pd NP size.
We captured additional TEM images to depict the exposure of Pd NPs toward S. aureus and E. coli at all concentrations (2.5 610 24 , 10 25 , 10 26 , 10 27 , 10 28 , and 10 29 M) for each size of nanomaterials after 24 h. Figures 4-6 show the NP exposure to bacteria for the three higher concentrations (2.5610 24 , 10 25 , 10 26 M). It is illustrated that no significant change in size and/or shape can be observed after the Pd NPs have been exposed to the bacteria. Throughout the remainder of the paper, we will refer to the molar concentrations of Pd NPs by the order of magnitude only.  Figure 8A), fewer colonies grew in the 2.5 nm-sized Pd NP treatment as compared to the control (Tukey's HSD, p = 0.018). In the 10 29 M test with E. coli ( Figure 8U), a slightly larger number of colonies grew in the 2.0 nm Pd NP test as compared to the control (Tukey's HSD, p = 0.025). In general, differences at 0 h between the four Pd NP treatments and the control were minor, indicating no immediate effects of the Pd NPs on the growth of either bacterium.

Toxicity of Pd NPs to Two Bacterial Populations Over Time
For both microorganisms, we observed a significant decrease, as compared to the control, in the amount of colonies mL 21 Figures 7B & 8B). We observed growth of E. coli   Figures 7C, 7D & Figures 8C, 8D). At 10 25 M concentration, all four test chemicals induced near 100% mortality in the S. aureus tests, as compared to the control (p,0.001; Figures 7F, 7G, 7H). We observed growth of S. aureus colonies in all three Pd NP treatments at 4 h ( Figure 7F) and in the 2.0 nm Pd NP test at 24 h ( Figure 7H). However, the average colony growth did not differ significantly from 0 CFU mL 21 Figures 8F, 8G, 8H). However, the magnitude of inhibition by the three Pd NPs at 10 25 M concentrations was far less in the E. coli tests than in the S. aureus tests. For example, after 4 h of exposure, the 2.0 nm and 2.5 nm Pd NMs were 85% more inhibitory to S. aureus than to E. coli, the 3.1 nm Pd NM was 81% more inhibitory and the Pd 2+ ions were 100% more inhibitory ( Figure 7F & Figure 8F). By 8 h and 24 h, all three Pd NMs remained less inhibitory to E. coli than to S. aureus, but the Pd 2+ ions completely inhibited growth of both bacteria ( Figure 7G, 7H & Figure 8G, 8H).
We continued to observe a trend of nearly complete inhibition by all four test chemicals to S. aureus colonies in the 10 26 Figure 8L), though we observed nearly complete inhibition by the 2.0 nm and 2.5 nm Pd NPs, and only 7% inhibition by the 3.1 nm Pd NP and 28% inhibition by the Pd 2+ ions ( Figure 8L).
The first sub-lethal effects we saw of the test chemicals toward S. aureus were in the 10 28 M concentration test at 4 h ( Figure 7R). Statistically, all four chemicals tested were equivalently inhibitory toward S. aureus (F 3,8 = 2.990, p = 0.096). As exposure time continued, we observed a similar trend of inhibition toward S. aureus at 8 h and 24 h in the 10 28 M concentration tests ( Figures 7S, 7T). Though some growth did occur at 8 h and 24 h, average colony growth in these treatments did not differ significantly from 0 CFU mL 21 . Conversely, none of the treatments at 10 28 M concentrations significantly reduced the growth of E. coli, as compared to the control (One-way ANOVA:  Figure 7V). However, the 2.5 nm Pd NP significantly inhibited S. aureus growth after 4 h of exposure, as compared to the control (Tukey's HSD: p = 0.006). By 8 h of exposure in the 10 29 M concentration tests, both the 2.5 nm and 3.1 nm Pd NPs significantly inhibited the growth of S. aureus (Tukey's HSD: p,0.001, Figure 7W), but the 2.0 nm Pd NP and the Pd 2+ ions did not reduce S. aureus growth, as compared to the control. After 24 h of exposure, all four test chemicals inhibited S. aureus growth significantly ( Figure 7X). While we did observe S. aureus colonies in the 2.0 nm and 3.1 nm Pd NP tests at 24 h, the average growth in these treatments did not differ significantly from 0 CFU mL 21 . Thus, extremely low (10 29 M concentrations) of all Pd NPs and Pd 2+ ions inhibited growth of S. aureus after 24 h of exposure ( Figure 7X), but did not inhibit growth of E. coli ( Figure 8X).

Size-Dependence of Toxicity
Given the high levels of growth inhibition toward S. aureus, we only observed variances in the effects of the different-sized NPs in the 10 29 M concentration test at the 4 h and 8 h time points (Figures 7V, 7W). At 4 h, the 2.5 nm-sized Pd NPs were more inhibitory to S. aureus colony growth than any of the other test chemicals (F 3,8 = 8.070, p,0.008, Figure 7V). The effects of 10 29 M concentrations of 2.0 nm, 3.1 nm Pd NPs and the Pd 2+ ions did not differ at the 4 h time point. These results indicate that the mid-sized (2.5 nm) particles are the most toxic to S. aureus, while smaller (2.0 nm) and larger (3.1 nm) NPs and Pd 2+ ions are less toxic. This basic trend was also observed in the 10 29 M test at 8 h ( Figure 7W) and in the 10 28 M test at 4 h and 8 h (Figures 7R,  7S), though it was not a significant effect in the 10 28 M tests. By 24 h of exposure, we observed no difference in the effect of any of the NP sizes on S. aureus growth, indicating that the size-dependent toxicity effect is also time-dependent.
The trend in size-dependent toxicity differed between the S. aureus and E. coli tests. In the E. coli test, we observed differences in the effects of the four test chemicals most obviously in the 10 26 M concentration test (Figures 8J, 8K

Discussion
While many previous studies have successfully demonstrated the toxic effects of different size ranges of NPs on bacteria, most have not been able to test the for the potential effects of NPs with narrow size distributions (i.e. 60.2 nm) and these studies do not distinguish between particle sizes that vary ,1 nm size changes on bacterial toxicity. This is primarily because the commercially 24 h (horizontal). The concentration range of each of the Pd NPs and Pd 2+ tests were from 10 24 -10 29 M (vertical). Significant differences shown in the figure represent comparisons between the four test chemicals using One-Way ANOVA and do not include comparison to the control (white bar); significance is represented by lowercase letters. Significance was assessed at a = 0.05. Capitalized letters refer to individual sub-figures A-X, as corresponds to exposure time and concentration. doi:10.1371/journal.pone.0085981.g007 Toxicity of Palladium Nanoparticles PLOS ONE | www.plosone.org available particles used in most studies lack monodispersity. Our ability to control the size of the NPs we synthesize and obtain them with narrow size distributions gives the added advantage that we can determine how specific sizes (i.e. 60.2 nm) affect bacterial toxicity. In general, traditional toxicological studies to determine lethal and effective doses using NPs as test chemicals are highly challenging, because nanomaterial size and shape is often inconsistent from test to test and manufactured NPs consist of a range of sizes. For example, commercially purchased TiO 2 NPs that were described as having a particle size of 5 nm by the manufacturer were shown to have an actual particle size of 3.561.0 nm (size range of 2-5 nm). [58] Similarly, manufactured Al NPs, Al 2 O 3 NPs and Al 2 O 3 nanowhiskers purchased from a supplier were shown to be highly polydispersed and batchdependent. [59] Thus, it was crucial for us to perform toxicity assays to new NPs only with well-characterized test particles.
In this study, we examined how ,1 nm differences in Pd NP sizes influence microbial growth to determine if Pd NPs are viable antimicrobial materials for future study and if their antimicrobial activity can be fine-tuned to target specific types of microorganisms. The as-synthesized Pd NPs used in our study were monodisperse with a narrow size distribution (2.060.1 nm, 2.560.2 nm, 3.160.2 nm). This unique feature is produced by controlling the size and shape through an effective synthetic pathway. The NP size was directly proportional to the reaction time. Thus, as the reaction time increased so did the particle size which is consistent with the Oswald ripening mechanism; as the reflux or heating time is prolonged the NP diameter increases. [60] The size of the NP can easily be tuned by varying the ratio of the reactants and or reaction conditions (metal precursor, stabilizer, reducing agent, time, temperature). [57] Aggregation has been shown to play a role in reducing the surface area of the particle which ultimately affects the surface properties. [61][62][63] In turn, NP aggregation can result in dissolution whereby metal atoms or ions are generated that may have toxic effects. [64] The small size of the Pd NPs is an important aspect of this study because previous work has shown that the electronic structure of the surface of the nanoparticles will change as a function of the size, and ultimately enhance the surface reactivity. [65,66] In particular, nanoparticles with sizes ,10 nm can attach to the surface of the cell membrane of the bacteria and disrupt the normal function (e.g. respiration and permeability). [65].
The mechanism by which Pd NPs are more toxic to S. aureus than to E. coli or why the 2.5 nm Pd NP was most toxic to S. aureus at sub-lethal concentrations is not clear from our current work and warrants further study. One likely explanation for these results is that the size of highly constrained Pd NPs used in our tests ranged from 1-3 nm, whereas studies of polydispersed Ag NPs used a higher size limit of 5-50 nm NPs. [26][27][28]65,[68][69][70]85] S. aureus cells are 1 mm diameter spheres that have a surface area of approximately 3 mm 2 , whereas E. coli cells, which are 1 mm wide by 3 mm long rods, have a surface area of approximately 11 mm 2 . If size is the only factor considered, .10 nm NPs would interact more easily with E. coli cells than S. aureus cells, simply given size and steric constraints. However, using smaller 2-3 nm NPs used in this study, the reaction sites for Pd on S. aureus cells would become saturated much more quickly and at much lower concentrations than E. coli cells.
Another possibility is that Pd NPs become ionized during the course of the toxicity assays. It is well known that gram negative bacteria, such as E. coli, can utilize efflux complexes to remove toxic compounds from the cell. For example, the CusCBA efflux system is responsible for removing biocidal copper and silver ions from E. coli cells. [86] Efflux pumps contribute significantly to the problem of acquired bacterial antibiotic resistance because of the broad variety of substrates they can recognize. [87] While it is possible that the E. coli cells in our tests effectively removed Pd NPs after they had entered the cell, our time-series TEM images of Pd NPs (Figures 4-6) suggest that Pd NPs do not change in size and are not oxidized over the test period.
It is most likely that surface area interactions between cell walls and Pd NPs are confounded by cell wall type. The outer membrane of gram negative bacteria, composed of a lipopolysaccharide-phospholipid asymmetric bilayer, bacteria provides a significant barrier to antimicrobial compounds. [88] For example, antibiotics can only penetrate into gram negative cells through lipid-mediated pathways and general diffusion porins. [89] The composition of the outer membrane of gram negative bacteria versus the primarily peptidoglycan-based cell wall of gram positive bacteria can significantly influence the ability of metals to bind to the bacterial surfaces. [90] Gram positive cell walls have been shown to act as metal chelators, whereas different transition elements (II) have been shown to vary widely in binding efficiency 24 h (horizontal). The concentration range of each of the Pd NPs and Pd 2+ tests were from 10 24 -10 29 M (vertical). Significant differences shown in the figure represent comparisons between the four test chemicals using One-Way ANOVA and do not include comparison to the control (white bar); significance is represented by lowercase letters. Significance was assessed at a = 0.05. Capitalized letters refer to individual sub-figures A-X, as corresponds to exposure time and concentration. doi:10.1371/journal.pone.0085981.g008 Toxicity of Palladium Nanoparticles PLOS ONE | www.plosone.org to E. coli AB264 cell envelopes. [91,92] For example, Pd metal binds with relatively low efficiency to E. coli cell envelopes (0.010 mmol mg 21 ) as compared to iron (III) (0.200 mmol mg 21 ) and cobalt (0.178 mmol mg 21 ). [92] In comparison, little is known about the binding efficiency of metals in the nano form to gram positive and gram negative bacteria. Morones et al. showed that small-scale Ag NPs (5 nm 62 nm ) interacted directly with bacterial cells by attaching to the membrane and penetrating the cell. [65] This is because Ag NPs create changes in membrane morphology which increase membrane permeability and disrupt proper transport of molecules to the cytoplasm. [28,65] When examined using TEM following treatment with Ag NPs, E. coli cells exhibit significant damage to membranes, in the form of pits on their surfaces. [28].
Surprisingly, the 2.0 nm Pd NPs were less toxic to E. coli at 10 25 M concentrations than at 10 26 M concentrations ( Figures 8G, 8K). In both cases, all three of the replicate treatment tubes and plates were used to generate the calculated averages and the error surrounding the averages was small, indicating that these results were not obtained by chance alone. A similar result was seen in the S. aureus tests, where 2.0 nm NPs allowed more colony growth at 10 25 M than at 10 26 M concentrations at the 4 h time point (Figures 7F, 7J). However, in the S. aureus test, the error surrounding the average colony growth was large enough that the colony growth we observed at 10 25 M at 4 h was not significantly different from 0 CFU mL 21 . Further investigation is required to determine the mechanism by which Pd NPs are toxic to both E. coli and S. aureus to explain this result. An obvious explanation of this pattern would be that the NPs formed aggregates at the 10 25 M concentration, but did not form aggregates at lower concentrations. However, we tested for agglomeration of the Pd NPs at 10 25 M with both bacterial cultures and observed no notable agglomeration occurring (Figures 4 and 6).
While the mode of action of Pd NPs remains unknown, Pd 2+ activity as an enzyme inhibitor is well studied. Pd 2+ is known to inhibit creatine kinase, succinate dehydrogenase, and many other common enzymatic processes in both prokaryotic and eukaryotic cells. [93] If entry into the cytoplasm in high quantities instead of membrane disruption is required to elicit Pd toxic effects, then smaller cell size and fundamentally different cell wall properties of gram positive S. aureus may render it more susceptible to Pd NPs than larger gram negative cells. By examining the toxicity of wellconstrained particle sizes with ,1 nm differences in size, we have demonstrated that fine-tuning the sizes of nanomaterials can have beneficial results for targeting particular bacterial cell types. In this case, 2.5 nm Pd NPs are the ideal size to target S. aureus, and will cause nearly 100% mortality of an S. aureus culture within 4 h of exposure to a mere 10 29 M concentration of that particular NP. The results we obtained are closely related to work done with other nanoparticles, however, the size of the NPs we used have a much narrower distribution relative to other types of nanoparticles reported in the literature. [13][14][15]26,34] Such particles with narrow size distributions enable is to understand the effect of a specific particle size on a bacterial strain, and also how changing the size by a size as small as 0.5 nm affects the interaction with microorganisms.
Using either test organism, the antimicrobial activity of Pd NPs of all sizes was high at low concentrations, indicating that Pd NPs are useful targets for future study of novel antimicrobial compounds. It is inevitable that Pd NPs will make their way into the market and that increased use and release of Pd NPs into wastewater streams and into the natural environment should be carefully regulated. When studied under more complex test conditions, determining the impacts of NPs becomes far from straightforward. For example, Ag NPs are four times less inhibitory to E. coli cells in a biofilm than to planktonic E. coli cells, likely due to aggregation of Ag NPs which inhibited diffusion throughout the biofilm. [67] Further studies investigating the mechanism of toxicity of Pd NPs and the effects of all NPs to ecological systems are a crucial target research area for these high functioning novel materials.

Conclusions
Investigating the interactions of metal NPs with bacterial cells is an important step toward realizing the toxicity of NPs but also in identifying potential biological applications. Pd NPs were found to be more toxic than Pd 2+ ions when exposed to gram positive S. aureus bacteria. Nanoparticles with smaller diameter, i.e. 2.060.1 nm were found to exhibit higher toxicity relative to the 2.560.2 nm and the 3.160.2 nm. The inhibitory growth effect was evident upon contact with the bacteria at concentrations as low as 10 29 M of Pd NPs. Relative to gram positive S. aureus, the gram negative E. coli required higher concentrations of Pd NPs and longer exposure times before an inhibitory growth effect became evident. This observation is apparent that Pd NPs could be useful antimicrobial agents for gram positive bacteria.