Synthetic antibacterial minerals: harnessing a natural geochemical reaction to combat antibiotic resistance

The overuse of antibiotics in clinical and livestock settings is accelerating the selection of multidrug resistant bacterial pathogens. Antibiotic resistant bacteria result in increased mortality and financial strain on the health care and livestock industry. The development of new antibiotics has stalled, and novel strategies are needed as we enter the age of antibiotic resistance. Certain naturally occurring clays have been shown to have antimicrobial properties and kill antibiotic resistant bacteria. Harnessing the activity of compounds within these clays that harbor antibiotic properties offers new therapeutic opportunities for fighting the potentially devastating effects of the post antibiotic era. However, natural samples are highly heterogenous and exhibit variable antibacterial effectiveness, therefore synthesizing minerals of high purity with reproducible antibacterial activity is needed. Here we describe for the first time synthetic smectite clay minerals and Fe-sulfide microspheres that reproduce the geochemical antibacterial properties observed in natural occurring clays. We show that these mineral formulations are effective at killing the ESKAPE pathogens (Enterococcus sp., Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter sp., Pseudomonas aeruginosa and Enterobacter sp.) by maintaining Fe2+ solubility and reactive oxygen species (ROS) production while buffering solution pH, unlike the application of metals alone. Our results represent the first step in utilizing a geochemical process to treat antibiotic resistant topical or gastrointestinal infections in the age of antibiotic resistance.


Supplementary Information Text
Publication rate for research on antibacterial clays. The study of natural antibacterial clays started with the seminal paper by Williams et al., 2004 1 , revealing that natural clay minerals with no pretreatments or alterations had intrinsic antibacterial properties. An exponential increase in publications on this research topic occurred over the next decade ( Figure S1.). However, to date no study has shown that this process can be reproduced with samples that exhibit consistent reactivity, high chemical purity, and the absence of any unwanted elements. Figure S1. Number of publications per year on antibacterial clays showing an exponential increase over the last 15 years, obtained from a PubMed (https://pubmed.ncbi.nlm.nih.gov/) keyword search for antibacterial clays.
Smectite synthesis reaction conditions. Typical smectite synthesis methods require long reaction times, many times requiring 3-6 months to produce a layered 2:1 clay. Additionally, many secondary minerals (zeolites) can form during synthesis depending on the SiO2 source, cation content and pH 2,3 . When samples were reacted at 180°C for 3 days, a portion of the fumed SiO2 particles remained ( Figure  S2A). The amorphous XRD peak from the fumed SiO2 particles was not present when samples were reacted for 5 days at 200°C ( Figure 1A, main text). The basal spacing of the synthetic F-hectorite interlayer spaces (001 hkl peak) is 12.8 Å, which is common for smectite clay minerals ( Figure 2C). Fluorine greatly increased the crystallization rate of the smectites and samples reacted without F had weak 001 (hkl) peaks after reacting for 5 days at 200°C, and no additional peaks associated with 2:1 expandable smectite clay mineral were observed ( Figure S2B-D). Upon saturation with ethylene glycol, the interlayer space expands to 17.2 Å and hkl reflections, characteristic of expandable smectite clays, appear ( Figure S2B). The random powder mounts show a peak at 1.52 Å that is characteristic of trioctahedral hectorite clay minerals (Figure 2A  Hectorite smectite clays have trioctahedral site occupancy with Li + and Mg 2+ occupying all potential octahedral sites 4,5 . The FTIR spectra of the synthetic F-hectorites are indicative of natural hectorite clay minerals 4,5 . The FTIR spectra show octahedral stretching from Mg and Li in the octahedral sheets at 3616 cm -1 and 3687 cm -1 respectively ( Figure S3, Table S1). Stretching and deformation bands from tetrahedral Si-O bonds are present at 955, 778, 700 and 412 cm -1 ( Figure S3, Table S1) 4,5 . The amorphous fumed SiO2 nano-particles used in the synthesis reaction have Si-O stretching and deformation bands at 1190, 1067 and 800 cm -1 (Figure 2). Combination Gaussian-Lorentzian curve deconvolution of FTIR bands at 1190 and 1067 cm -1 was used to determine the percentage of unreacted fumed SiO2 remaining in the synthetic clays under different reaction conditions. Samples reacted at 180°C for 3 days have 24% unreacted SiO2 nano-particles present ( Figure S3). After reacting for 5 days at 200°C samples contain 4% unreacted SiO2 nano-particles.  Smectite cation exchange capacity, zetapotential and particle size. The synthetic F-hectorites have a cation exchange capacity (CEC) of 98.9 ± 3.1 meq/100g (milli-equivalents of charge per 100 g of sample) ( Figure S4A). The CEC values of the synthetic F-hectorites are in the range of natural 2:1 smectite clay minerals that have CEC values ranging from 60 to 120 meq/100g 6 . The fumed SiO2 used in the synthesis reaction had almost no cation exchange capacity with a value of 0.3 meq/100g measured. Zetapotential measurements of synthetic smectites with varying pH show the clays maintain a net negative surface charge from pH 3-9. This is due to the negatively charged basal surfaces of the smectite that dominate the surface charge of the clay. Pyrite synthesis reaction conditions and removal of elemental sulfur. Synthesis reactions using only S 0 as the sulfur source formed pyrite, however the reaction was not complete and weight % quantities of S 0 remained in the final product ( Figure S5A). Additionally, the pyrite particles that formed did not have uniform size and shape ( Figure S6A-B). When polysulfides were used in the pyrite synthesis reaction only minor S 0 impurities were observed in the XRD pattern ( Figure S5B). Rinses with xylene effectively removed these impurities and produced pure pyrite powders ( Figure S5C and S6). The SEM images of the purified pyrite samples indicate that 1-2 µm spheres form ( Figure S6), similar to the size of pyrite particles observed in the natural antibacterial minerals 7 .  Mineral mixtures and antibacterial activity. Initial antibacterial susceptibility tests were performed on mixtures of synthetic smectite and pyrite powders. Mixtures of smectite with 5 and 10 wt. % pyrite were prepared in a mortar in pestle and transferred into 15 mL centrifuges tubes and autoclaved at 120°C for 30 minutes prior to antibacterial susceptibility testing. These percentages of pyrite were similar to those observed in natural antibacterial samples 8 . The F-hectorite-pyrite mixtures alone did not produce any antibacterial activity and levels of soluble Fe 2+ and H2O2 were below detection limits (1 µM and 5 µM, respectively) ( Figures S7 and S8). The pH of the smectite-pyrite mixtures was 5.5 when tested against E. coli ( Figure S7). The pH of the solution rose to 7.1 after 24 hours as a result of bacterial growth. Antibacterial mineral mixtures observed in nature instantaneously released Fe 2+ and generated H2O2 upon hydration. The cation exchange capacity of the expandable smectite clay minerals was hypothesized to be a source of Fe 2+ that would rapidly release in solution and initiate the redox cycling reactions with pyrite 9 . Therefore, the cation exchange of synthetic smectites may result in the rapid release of Fe 2+ which starts the redox cycling of Fe 2+ /Fe 3+ with pyrite surfaces while generating H2O2.
A series of experiments using Fe 2+ exchanged smectites and pyrite were conducted to determine if smectite interlayer Fe 2+ would initiate antibacterial activity. All Fe 2+ exchange reactions were conducted in an anerobic glove box to prevent the oxidation of Fe 2+ during sample processing. The concentration of minerals in all Fe 2+ exchange reactions was 10 mg/mL. Samples were exchanged with 120 mM FeSO4 (nitrogen purged) 3 times and centrifuged at 3500 rpm for 10 minutes between each exchange. After the Fe 2+ exchange, the samples were rinsed with nitrogen purged DIW in the glove box 3 times, with a centrifugation step between each rinse. The samples were then resuspended in 95% ethanol and centrifuged at 3500 rpm for 10 minutes and dried under a nitrogen purge. After the samples had dried, they were autoclaved at 120°C for 30 minutes.
Mineral mixtures exchanged with Fe 2+ containing F-hectorite and 5 wt. % pyrite immediately released 2 mM Fe 2+ during reaction with E. coli, and maintained 0.5 mM concentrations of Fe 2+ over 24 hours ( Figure S7). These samples resulted in the immediate generation of 319.2 µM H2O2 and maintained 67.6 µM H2O2 over 24 hours ( Figure S7). The initial and 24 hour pH values for the Fe 2+ exchanged Fhectorite 5 wt. % pyrite mixtures were 4.2 and 3.7, respectively. Bactericidal antibacterial activity was observed after 4 hours ( Figure S8). The Fe 2+ exchanged synthetic smectites with no added pyrite immediately released 0.7 mM Fe 2+ and maintained 75 µM Fe 2+ over 24 hours ( Figure S7). The concentration of H2O2 immediately released was 79.2 µM and 7.6 µM H2O2 over 24 hours, with initial and 24 hour pH values of 4.9 and 4.4, respectively ( Figure S7). The Fe-hectorite Fe 2+ exchanged samples were not bactericidal, however they did inhibit bacterial growth over 24 hours ( Figure S8). A mixture of Fe 2+ exchanged F-hectorite containing 20 wt. % pyrite was also tested for antibacterial activity, Fe 2+ and H2O2 generation to determine if increased pyrite concentrations could extend the duration of the ROS generating reactions. These samples resulted in the instant release of 1.5 mM Fe 2+ ( Figure S7). After 4 hours a drop in Fe 2+ concentrations was observed, similar to the Fe 2+ exchanged F-hectorite with or without 5 wt. % pyrite. However, after 24 hours the levels of Fe 2+ increased to 3.6 mM in the samples containing 20 wt. % pyrite. The H2O2 concentrations followed a similar trend with initial concentrations of 319.2 µM H2O2 measured, increasing to 464.8 µM H2O2 after 24 hours ( Figure S7). The pH values for the 20 wt. % pyrite samples were lower, with values of 3.5 and 2.8 measured initially and after 24 hours, respectively. These samples were bactericidal against E. coli after 4 hours, however the pH decreased to 2.8 in these samples which is bactericidal to E. coli and S. epidermidis without the presence of Fe 2+ and H2O2 (Figures S7 and S8). The differences in Fe 2+ release, H2O2 generation and pH are discussed in the next section. Figure S7. The influence of pyrite and smectite interlayer Fe 2+ on pH, Fe 2+ , H2O2 release and antibacterial activity. All samples were tested for antibacterial activity against E. coli (ATCC 25922) growing in TSB broth using 50 mg/mL mineral suspensions.
The samples containing 20 wt. % pyrite initially released 25 % less Fe 2+ when compared to the samples with 5 wt. % pyrite ( Figure S7). The immediate release of Fe 2+ is predominately from the smectite interlayer spaces as the samples containing no pyrite release Fe 2+ at 0.7 mM concentrations upon hydration ( Figure S7). However, the overall extent of Fe 2+ release in the pyrite exchanged samples is greater when compared to the pure Fe 2+ exchanged smectites, despite the pyrite exchanged samples having lower wt. % clay content. The difference in Fe 2+ release is a result of redox reactions driving pyrite oxidation Fe 2+ release and H2O2 generation 8,10 . This initial set of experiments reveal that redox cycling and cation exchange reactions taking place between the smectite clay minerals. The pyrite concentration can be used to alter the intensity and duration of Fe 2+ release and H2O2 generation over 24 hours. However, acidity is a consequence of pyrite oxidation and hydrolysis of Fe 3+ cations during the oxidation of antibacterial clays 9 and pH values decreased below 3, which may limit medicinal applications of these antibacterial clays. Limited Fe 2+ exchange and DIW rinsing of the samples prevented the drop in pH and allowed the synthetic antibacterial mineral assemblages to buffer pH in the range of 5.5-3.7 over 24 hours, depending on the mineral concentration ( Figure 2C, main text).
The initial antibacterial testing experiments above reveal that pyrite concentration and Fe 2+ cation exchange of smectites can influence the antibacterial activity of synthetic minerals mixtures. We varied these two factors to test the degree of control they have on antibacterial activity. All samples were tested for antibacterial activity, Fe 2+ , H2O2 and pH as previously described (using E. coli and TSB broth), with dilution spot plates measured after 4 and 24 hours (Figures 9-11). Mineral mixtures with varying pyrite (1.5 to 20 wt. %) and a fixed Fe 2+ exchange solution concentration (20 mM) reveal that increases in pyrite loading directly correlate with increased concentrations of Fe 2+ and H2O2 measured in solution ( Figure   Figure S8. Antibacterial susceptibility testing (E. coli ATCC 25922) dilution spot plates. Mixtures of Fhectorite and pyrite alone were not antibacterial. Exchanging samples with Fe 2+ resulted in antibacterial activity. The Fe 2+ exchanged samples containing pyrite were bactericidal, while the Fe 2+ exchanged F-hectorites with no pyrite inhibited bacterial growth after 24 hours. 9a). The concentrations of Fe 2+ released after 1 hour ranged from 0.76 mM in samples containing 1.5 wt. % pyrite, to 7.9 mM in samples containing 20 wt. % pyrite ( Figure 1F). The concentrations of H2O2 ranged from 77 to 300 µM in samples with 1.5 and 20 wt. % pyrite, respectively ( Figure 1F). The pH values in these samples were higher with values ranging from 5.45 to 4.7 with pyrite ranging from 1.5 to 20 wt. %, respectively ( Figure 1F). The mineral mixtures became bactericidal when pyrite concentrations were ≥ 10 wt. % and Fe 2+ , H2O2 and pH concentrations were ≥ 4 mM, 209.5 µM and 4.96, respectively. These results indicate that acidity from metal hydrolysis and pyrite oxidation occurred when samples were processed with multiple Fe 2+ exchanges in a glove box. The presence of Fe 3+ in solution can initiate pyrite oxidation even in the absence of oxygen and generate acidity through hydrolysis reactions and pyrite oxidation. Limiting the Fe 2+ exchange and rinsing steps prevent excess acid from accumulating.
A series of synthetic smectite-pyrite mixtures were also prepared with a fixed pyrite concentration (5 wt. %) and varying concentrations of Fe 2+ exchange solution (5 to 60 mM) ( Figure 1G). These samples released 1.3 to 3.5 mM Fe 2+ and 161.1 to 274.1 µM H2O2 after 4 hours ( Figure 1G). The increasing concentrations of Fe 2+ and H2O2 released correlate with increases in the concentration of Fe 2+ exchange solution as expected. The pH in this series of samples was also higher with values ranging from 5.22 to 4.75 in samples exchanged with 5 to 60 mM Fe 2+ ( Figure 1G). Samples exchanged with ≥ 30 mM Fe 2+ were bactericidal when Fe 2+ , H2O2 and pH concentrations were ≥ 2.4 mM, 221.1 µM and 4.91, respectively ( Figure 1G).
The multiple DIW rinsing steps used in the initial exchange reaction, carried out in an anerobic glove box to limit oxidation, resulted in lower overall Fe 2+ concentrations and pH values < 4.2. Limited Fe 2+ exchange reactions and rinsing steps performed in ambient atmospheric conditions still produced minerals with antibacterial properties and resulted in more basic pH values (< 5.5). Furthermore, these results show that the use of an anaerobic glove box and multiple Fe 2+ exchange reactions are not required to produce antibacterial mineral formulations that generate extended release Fe 2+ and ROS. Table S2. ICP-MS elemental analysis of mineral leachates (mg/mL) reacted in TSB media (15 g/L) for 24 hours. All elemental concentrations are reported in micro-molarity (µM), with values below detection limit listed as BDL. The minimum bactericidal concentration of elements are listed along with the corresponding species tested for antibacterial activity.  Agarose hydrogel antibacterial mineral composites. The results from the agarose mineral composites tests against E. coli reveal show antibacterial activity when mineral concentrations reach 100 mg/mL (Figure 9a). At this concentration, the samples were bactericidal after 24 hours. Concentrations below 100 mg/mL did not show any antibacterial activity or growth inhibition for E. coli cultures. The pH of the antibacterial 100 mg/mL sample dropped to 4.7 after 24 hours (Figure 9b). The other mineral concentrations (50 -12.5 mg/mL) had pH values similar to the E. coli control (Figure 9b). The S. epidermidis cells were more sensitive to the mineral agarose composites and bactericidal activity was observed at 50 mg/mL concentrations after 4 hours (Figure 9c). Concentrations of 25 mg/mL resulted in growth inhibition over 24 hours, with pH values measured at 7.1 (Figure 9d). The bactericidal concentrations (≥ 50 mg/mL) had pH values ≤ 4.7 after 24 hours. These results show that antibacterial activity is maintained when minerals are imbedded in agarose hydrogels, and the release of minerals is prevented allowing cell growth to be monitored with UV-Vis spectroscopy. Toxicity of antibacterial mineral mixtures to dermal fibroblasts. The results from the fibroblast viability assay indicate fibroblast cells remain viable after a 24 hour exposure to the antibacterial minerals with approximately 93% of the cells remaining viable. The fibroblast controls with no added minerals maintained 98% viability over 24 hours ( Figure S10). However, when the minerals with no fibroblast cells were measured on the automated cell counter, a viability of 31% was measured ( Figure S10). This indicates that the minerals, to a degree, interfere with the trypan blue and subsequent automated cell counter viability measurement. Subtracting the influence of minerals from the fibroblast cells exposed to the antibacterial minerals indicates that approximately 62% of the fibroblast cells are viable after exposure. These results reveal that fibroblast cells are capable of surviving the toxicity produced from the antibacterial minerals. The application of antibacterial minerals to treat topical bacterial infections will require careful monitoring of the wound environment to determine when the bacterial infection is eradicated so that the minerals can be removed and wound healing can proceed. Figure S10. Fibroblast toxicity assay during 24 hour exposure to antibacterial mineral mixtures using an automated cell counter and trypan blue viability assay. Viability of 3T3 mouse fibroblasts exposed to syn-hectorite 5% pyrite Fe 2+ exchanged minerals at 100 and 25 mg/mL concentrations after 24 hours, compared to fibroblast control with no minerals added and mineral control with no fibroblast cells.