Escherichia coli ATCC 25922, obtained from the American Type Culture Collection, and methicillin-resistant Staphylococcus aureus (MRSA), obtained from Sonora Quest Laboratories (Tempe, AZ, USA), were used for all studies as previously described [7]. E. coli was grown on Luria-Bertani (LB) agar or in LB broth, and MRSA was grown on trypticase soy agar (TSA) or in trypticase soy broth (TSB). Both bacterial strains were grown at 37°C with gentle rotary mixing.

E. coli and MRSA exponential phase cultures were prepared by diluting overnight cultures into fresh growth medium to a concentration of 10⁷ CFU/mL and continuing growth at 37°C with gentle rotary mixing until the cultures reached mid-logarithmic phase of growth. Bacterial cells were collected by centrifugation, washed once in 0.85% sterile saline, and suspended in the appropriate clay mixture leachate solution, clay mixture (at 1% or 10% w/v), or sterile UV-irradiated ultrapure H₂O (dH₂O) at an initial concentration of 10⁷ CFU/mL. Samples were incubated at 37°C with gentle rotary mixing for 24 h, and cell survival was determined by plating duplicate 10-fold serial dilutions for each sample at appropriate time points and enumerating colonies on plates after overnight incubation at 37°C. Due to sample processing, the indicated 0 h experimental exposure times were equivalent to approximately 3 min exposures.

Clay mixtures were autoclaved for 1 h at 121°C before experimental use. A 10% CB suspension refers to 0.1 g of clay mixtures mixed in 1 mL sterile dH₂O. CB-derived leachates (CB-L) were obtained by continuously stirring clay mixtures (1 g/20 mL) in sterile dH₂O for 18–24 h. Subsequently, the hydrated clay mixture suspensions were centrifuged (31,000×g) for 3 h at 4°C to separate insoluble and soluble fractions. The aqueous supernatant (leachate) was collected and sterilized by passage through a 0.22 µm filter.

Qualitative mineralogical profiles of CB07, CB08, CB09, and CB10 samples were analyzed as oriented mounts on glass slides from 2° to 35° two-theta with the standard treatments of K-saturation followed by heating to 350°C and 550°C, Mg-saturation, and Mg-saturation/glycerol solvation [16]. X-ray diffraction analyses were conducted at the University of Arizona Center for Environmental Physics and Mineralogy using a PANalytical X'Pert PRO-MPD X-ray diffraction system (PANalytical, Almelo, AA, The Netherlands) producing Cu–Kα radiation at an accelerating potential of 45 kV and current of 40 mA and fitted with a graphite monochromator and sealed Xenon detector.

Quantitative analyses of the four CB clay mixture samples were run as random powder mounts and measured from 5° to 65° two-theta. Samples were prepared by pulverizing in a McCrone micronizing mill (The McCrone Group, Westmont, IL) after the addition of a 20% (w/w) corundum internal standard. After measurement, diffractograms were imported into RockJock [17], a program for determining quantitative mineralogy from powder X-ray diffraction data, and analyzed with Reitveld refinement. Measurement conditions were as follows: 0.02° step size, 3 sec dwell time per step, spinning at 1 revolution per second, 1/4° divergent anti-scatter slit, 15 mm divergent mask, 1/4° incident anti-scatter slit, and 0.6 mm fixed receiving slit.

Measurements of the elemental concentrations in the leachate samples were determined using inductively coupled plasma mass spectrometry (ICP-MS) (ThermoQuest/Finnigan Element 2), and the clay mixture samples were analyzed by ICP-MS and inductively coupled plasma – optical emission spectrometry (ICP-OES) (Perkin Elmer Optima 3000). For clay mixture sample analysis, the procedures used were similar to those described by Janney et al. [18] and Solidum et al. [19] with some modifications. Prior to analysis, 0.025–0.030 g of sample powder was digested using ∼1 mL of a 2∶1 mixture of double-distilled and concentrated hydrofluoric (HF) and nitric (HNO₃) acids and either placed in an ultrasonic bath for 60 minutes or heated at 60°C overnight. After the powder was fully dissolved, the sample-acid solution was evaporated to dryness over a hotplate at 60°C and under a heat lamp in a teflon evaporating unit. The digested residue was then treated twice with ∼0.5 mL of concentrated HNO₃ and evaporated to dryness. For ICP-MS analyses, the digested sample powder was first diluted 100× (by weight) with a 2.5% HNO₃ acid solution containing 1 ppb indium (In) as an internal standard, and then an aliquot of the digested sample solution was diluted ∼4000× with the 2.5% HNO₃ solution containing indium. Calibration was achieved using one blank and a series of three synthetic standard solutions with elemental concentrations ranging from 0.05 to 10 ppb. For the ICP-OES analyses, an aliquot of the 100× diluted sample solution described above was diluted 2000× the sample weight using a 1% HNO₃ acid solution. Instruments were calibrated with one blank solution and a series of four synthetic standard solutions with element concentrations ranging from 20 to 500 ppb in 1% HNO₃ acid medium. The accuracy and precision of the analytical method were monitored by analysis of natural rock standards, AGV-1 and BHVO-1, prepared in the same manner as the unknown samples.

The leachate samples were analyzed on an ELAN DRC-II ICP-MS (Perkin Elmer, Shelton, CT). Calibration standards were prepared from multi-element stock solutions (except for Hg, which is a single element standard) purchased from AccuStandard (New Haven, CT). The stock solutions were diluted in 1% nitric acid to provide a working calibration curve of at least five measurements. Samples were also diluted with 1% nitric acid until their response was determined to be within the calibration range. Internal standards (Rh, In, and Ga) were added to both standards and samples prior to analysis.

For the ion supplementation experiments, CB09-L was prepared as described above and then subsequently supplemented with chloride salts of Fe, Co, Ni, Cu, and Zn to final ion concentrations equivalent to values present in natural CB07-L. Where indicated, the pH of the solutions was adjusted with 1 M HCl or 1 M NaOH to a pH of 3.4 or 5.0 to mimic the native pH values of CB07-L and CB09-L, respectively. Finally, the leachates were sterilized by passage through a 0.22 µm filter prior to further analyses.

Half-cell oxidation-reduction potential (ORP) measurements were made using an Accumet Basic AB15 Plus pH meter equipped with a single junction Ag/AgCl reference electrode and a platinum ORP-indicating electrode. Final E_h values were calculated by adding the half-cell potential measurement for a Ag/AgCl electrode at 25°C (+199 mV) to the measured value. The electrode was calibrated with Zobell’s solution, and the ORP value was normalized to this value accordingly.

Aqueous speciation and solubility of Fe, Co, Ni, Cu, and Zn were calculated using Visual MINTEQ (version 3.0) modeling software. The ICP-MS/OES-determined ion concentrations and pH of the CB leachates were used to calculate ion speciation and solubility. Experimentally determined E_h values used for each leachate (CB07-L, 567.1 mV; CB08-L, 514.2 mV; CB09-L, 524.0 mV; CB10-L, 567.5 mV) were included in the calculation of redox active ion pairs present in the leachates (Fe²⁺/Fe³⁺, Co²⁺/Co³⁺, and Cu¹⁺/Cu²⁺).

Reduction in bacterial concentration was measured as the log₁₀ decrease in bacterial viability following a 24 h exposure to the experimental conditions and tested using one-sample t-tests. Relative differences in antibacterial activity between two leachates were tested using two-sample t-tests with Satterthwaite approximation. Associations between ion species concentrations and antibacterial activity were assessed using linear regression of the log₁₀ reduction in bacterial concentration against log₁₀ ion species concentration. Multivariate regression models were constructed using stepwise regression and the Bayes Information Criterion. Analyses were conducted using R 2.15.3 statistical software.

In vitro antimicrobial susceptibility experiments with clay-water suspensions and leachates were performed to assess the effect of the CB clay mixtures on the growth of E. coli and MRSA. Ten percent suspensions of CB07, CB08, CB09, and CB10 completely killed E. coli and MRSA within 24 h (Figure 1). Exposure of E. coli and MRSA to 1% suspensions of CB07 and CB10 resulted in complete killing and nearly complete killing, respectively (Figure 2). However, when E. coli was exposed to 1% suspensions of CB08 and CB09, there was no decrease in viability (Figure 2A). Exposure of MRSA to 1% suspensions of CB08 and CB09 minerals resulted in bactericidal activity (4.8- and 4.7-log₁₀ unit reductions, respectively), but did not completely kill the cells (Figure 2B). The effects of clay mixture leachate exposures to E. coli and MRSA are shown in Figure 3. CB07-L completely killed E. coli and MRSA, while CB10-L resulted in 3.1- and 2.5-log₁₀ decreases in E. coli and MRSA viability, respectively (Figure 3). E. coli and MRSA viability was minimally decreased or affected following 24 h exposures to CB08-L and CB09-L, respectively (Figure 3). Overall, the CB07 clay mixture and CB07-derived leachate, followed by CB10 clay mixture suspensions and CB10-L, have the greatest antibacterial efficacy, compared to the other CB clay mixture and leachate samples.

X-ray diffraction (XRD) analyses of the CB minerals (Figure 4) revealed that the minerals are all composed of approximately 52% clays and 48% non-clay minerals (Table 1). The most abundant clay mineral in all of the samples was illite-smectite (36–37%), followed by montmorillonite (9.7–14.2%) and kaolinite (1.4–3.6%) (Table 1). The CB08 and CB09 samples also contained small amounts of chlorite (Table 1). The most abundant non-clay mineral present in the CB samples was quartz (34–37.3%), followed by pyrite (4–5.5%) and jarosite (2.4–4.7%). Overall, the XRD analyses of the CB minerals revealed minor mineralogical differences between the four samples.

Clay mineral surfaces are naturally negatively charged, thus allowing free exchange of positively charged species, such as metal cations. When in a hydrated suspension, these cations can be released from the surface of the minerals into the surrounding medium. Previously, we demonstrated that the physical interactions with the clay mixtures do not mediate antibacterial activity on their own, but rather, the freely exchangeable metal ions bound to the surface of the minerals are responsible for the antibacterial activity of CB07 clay mixtures [10]. ICP analyses of the four CB bulk clay mixture samples revealed that the ion compositions are similar, consistent with the mineralogical XRD data (Table 2). While the four CB samples were mineralogically similar, there were notable differences associated with the composition of ions bound to their surfaces (Table 3). For example, exchangeable Cu concentrations ranged from 0.04 to 3.18 µM, an approximate 80-fold difference in concentration across the four samples (Table 3). Likewise, exchangeable zinc concentrations ranged from 3.05 to 14.52 µM across the four leachate samples (Table 3). These ICP data revealed chemical variability across the samples, suggesting that the differences in antibacterial activity may be due to the variable concentrations of the exchangeable ions available in each of the clay mixture samples. However, while overall ion concentration is commonly assumed to be an accurate indicator of metal ion toxicity, it does not always accurately predict bactericidal activity [20]. Previously published experimental data and mathematical modeling [23] have demonstrated that Cu, Ni, and Zn toxicity increases with increasing pH [20]–[22]. The authors hypothesize that the pH-dependent changes in metal ion toxicity are due to increased metal-ligand interactions and increased concentrations of intracellular ions in the cases of Cu and Ni [20], [21] or changes in speciation over a defined pH range in the case of Zn [22]. With one exception, the concentrations of Fe, Co, Ni, and Zn were lower in CB10-L compared to the non-antibacterial CB08-L and CB09-L leachates, yet CB10-L antibacterial activity is maintained. The four CB leachates have a pH range of 3.4–5.0 (Table 3). Thus, while CB10-L has lower ion concentrations than the non-antibacterial leachates, the variable pH across the samples likely modulates speciation, solubility, and metal-cell interactions to affect antibacterial activity.

To evaluate the influence of pH and ion speciation on the antibacterial activity, we modified the pH of CB07-L (antibacterial) and CB09-L (non-antibacterial) to that of the other respective leachate. The natural pH of CB07-L is 3.4 and was subsequently increased to a pH of 5 prior to antibacterial susceptibility testing. Likewise, the natural pH of CB09-L is 5.0 and was decreased to 3.4 for further testing. Figure 5 shows that increasing the pH of CB07-L rescued MRSA (4.7-log₁₀ unit increase; p = 0.039) from killing. However, decreasing the pH of CB09-L to 3.4 only slightly increased the antibacterial activity of the leachate against E. coli (p = 0.018) and did not alter the antibacterial activity against MRSA.

Previously, we performed antibacterial susceptibility experiments whereby we supplemented the clay mixtures with metal chelators [10]. The addition of desferrioximine, which readily chelates Fe, Co, Cu, Ni, and Zn, rescued E. coli cells from clay-mediated killing [10]. Therefore, we hypothesized that these ions were major contributors to the antibacterial activity of natural clay mixtures. The ICP analyses revealed that the concentrations of these five ions are higher in the CB07 bactericidal leachate compared to the non-antibacterial counterparts, CB08-L and CB09-L. To confirm the role of these ions in the bactericidal activity of CB clay mixtures, we supplemented the non-antibacterial leachate, CB09-L, with Fe, Co, Ni, Cu, and Zn such that the final ion concentration of these elements and pH were equal to that of CB07-L (pH 3.4). Antibacterial susceptibility testing of the ion-supplemented CB09-L (IS CB09-L) resulted in complete killing of E. coli (p<0.001) and MRSA (p<0.001) following a 24 h exposure (Figure 5). When the IS CB09-L pH was subsequently increased from 3.4 to 5, it no longer killed E. coli or MRSA (Figure 5).

The concentration of total ions in the CB leachates did not directly correlate to antibacterial activity. Thus, we hypothesized that specific, toxic ionic species may be at higher concentrations in the antibacterial leachates, CB07-L and CB10-L, compared to the CB08-L and CB09-L non-antibacterial leachates. To test this hypothesis, we used Visual MINTEQ to model the ion speciation and solubility patterns in the CB leachates. Figure 6A shows the concentration and soluble ion species of Zn, Ni, Co, Cu, and Fe predicted to be present in the CB clay mixture leachates. CB07-L has the highest concentration of all measured species (with the exception of Fe³⁺ and Co³⁺), followed by CB08-L, CB09-L, and CB10-L (Figure 6A), which is consistent with the pattern of the total concentration of these ions in the leachates. Metals are more soluble in lower pH environments, thus the larger concentration of metals in CB07-L is likely due to the lower pH of this leachate. Of all the species tested, Cu²⁺ is the only species that is higher in both CB07-L and CB10-L than in CB08-L and CB09-L, making it a possible candidate for contributing to antibacterial activity. Since antibacterial activity is diminished with increasing pH, we modeled speciation changes that occur between pH 3 to 6 (Figure 6B). While the concentrations of Co²⁺, Cu¹⁺, Cu²⁺, Ni²⁺, and Zn²⁺ did not change over this pH range, Fe³⁺ and Co³⁺ concentrations decreased, and the concentration of Fe²⁺ increased when pH decreased (Figure 6B). The increase of Fe²⁺ in a lower pH environment suggests its involvement in the antibacterial activity of the leachates.

Single variable regression analyses revealed that Zn²⁺, Ni²⁺, Co²⁺, and Cu²⁺ concentrations are positively associated with increased antibacterial activity against E. coli, while Zn²⁺, Co²⁺, and Cu²⁺ are positively associated with antibacterial activity against MRSA (p<0.01). These four species, Zn²⁺, Ni²⁺, Co²⁺, and Cu²⁺, were highly correlated (R²>0.93 for each pair of ion species), thus it remains unknown which one (or more) of these species is a key contributor to antibacterial activity. We constructed multivariate regression models to examine the combined role of different ion species and pH on antibacterial activity. The multivariate regression statistical models revealed that pH, Ni²⁺, Fe²⁺, and Fe³⁺ best explained antibacterial activity against E. coli (R² = 0.904), while pH, Co²⁺, Co³⁺, Fe²⁺, and Fe³⁺ best explained antibacterial activity against MRSA (R² = 0.786). We should note that a limitation of the speciation modeling and ICP analyses is that neither includes information about anions, such as sulfur anions, potentially present in the leachates. Sulfite, for example, increases in toxicity with decreasing pH, consistent with the toxicity pattern seen in the leachates [24]. The pyrite and jarosite minerals present in the CB clay mixtures are both sulfur composite minerals offering a potential source of sulfur to the clay mixture suspensions and leachate solutions. Thus, further analyses evaluating the presence and toxicity of anions must be performed to address anions as possible sources of toxicity.

Overall, ICP analyses demonstrated that the concentrations of Fe, Co, Cu, Ni, and Zn vary greatly among the four CB mineral leachates. Non-antibacterial CB leachates supplemented with these additional ions and adjusted to a lower pH exhibited increased antibacterial activity. However, when the pH of the ion-supplemented leachate is increased, the solution loses antibacterial activity. Univariate regression analyses revealed that Zn²⁺, Ni²⁺, Co²⁺, and Cu²⁺ concentrations were positively correlated with antibacterial activity in E. coli and that Zn²⁺, Co²⁺, and Cu²⁺ concentrations were positively correlated with antibacterial activity in MRSA. While pH is an important factor in the speciation of metal ions, we demonstrated that killing activity is not solely attributed to pH. Together, these analyses further indicate that the concentration and speciation of exchangeable metal ions are responsible for antibacterial activity and that pH alone is not responsible for killing. Because of the natural variability across natural clay mixture samples and the correlated variability in antibacterial activity, efforts must be taken to standardize the composition of these minerals to ensure consistent antibacterial efficacy.