AgAu NPs are a promising approach as alternative antibacterial substance due to the combination of antibacterial and biocompatible elements. Synthesis by LAL allows for ultrapure, spherical AgAu NPs with homogenous elemental distribution and reproducible particle sizes. However, so far, the antibacterial effect of these particles has not been tested on physiologically relevant biofilms, which are known for their inherent resistance to antimicrobial agents. In this study, the effect on S. aureus biofilms, one of the major pathogens in implant-associated infections, was investigated. Albeit biofilms are considerably more tolerant than planktonic bacteria, an antibacterial effect throughout the entire thickness of the biofilms was observed. The AgAu NPs showed their strongest effect on metabolic activity, which could be additionally verified by targets on the molecular level.
Initial characterization of the AgAu NPs generated by LAL in this study showed a clear absorbance peak in UV–Vis-extinction spectroscopy associated to the localized surface plasmon resonance (SPR) of a colloid consisting of 80 mol% silver and 20 mol% gold. As only one peak is observable, large deviations in composition as well as formation of larger fractions of pure Ag and Au nanoparticles could be excluded. Previous studies found the elemental distribution within these nanoparticles to be homogenous34,35. A more detailed recent analysis by XPS revealed that there is a slight enrichment of gold in the near-surface volume for the sample containing 80 mol% silver and 20 mol% gold used in this study38. The SPR maximum at ~ 415 nm is in good accordance with previous findings and can be associated with a gold molar fraction of ~ 0.2, which is close to the nominal composition of the target34,35,38. The results are further confirmed by XRF analysis, where a gold molar fraction of 0.2 was determined for the bulk colloid which is in line with previous characterization approaches based on XRD35 as well as single particle TEM-EDX34. Number-weighted particle size distributions derived from TEM analysis indicate a monodisperse and monomodal particle size distribution with a mean diameter of 8 ± 3.5 nm, with errors derived from the variance (width) of the log-normal distribution curves. Therefore, it can be concluded that the nanoparticle colloids used in this study had a monodisperse particle size distribution and a gold molar fraction of 0.2, closely matching that of the target used during ablation. In a previous study, the nanoparticles were found to be single crystals with a face centered cubic (fcc) structure35. Al-Zubeidi et al.37 analyzed the silver release kinetics of AgAu NPs on a single particle level. They found that alloying with gold reduced the silver leaching rate and identified two different leaching stages. The first one being dependent on silver content, while the second one depended on the oxidation potential of the particles and diffusion of silver ions through the lattice. The second leaching phase is slower and occurs once the relative gold content in the lattice reaches a certain point.
The elemental distribution of 80 mol% silver and 20 mol% gold was chosen for the synergistic properties of both metals. While pure silver results in the highest antibacterial activity of nanoparticles, only alloying it with gold allowed for cytocompatibility27,40. However, adding gold to the nanoparticles also changes the magnitude of the antibacterial effect. Grade et al.27 could show that by introducing 20 mol% gold, the minimum inhibitory concentration (MIC) of the nanoparticles doubled. By using nanoparticles consisting of 50 mol% silver and 50 mol% gold, the MIC increased by 800%. This is hypothesized to be due to a reduced silver ion release caused by the gold31. For the treatment of biofilms, it can be assumed that a much higher dose of nanoparticles is required than for planktonic bacteria. With that in mind, the 80 mol% silver and 20 mol% gold distribution was chosen for its stronger antibacterial effect.
When these particles were used for antibacterial testing in this study, all assays were done in 0.2 mM NaCl and in binding buffer. NaCl was the solution in which the nanoparticles were synthesized and in which they were most stable, based on the influence of ionic strength on stability of ligand-free particles31. However, 0.2 mM NaCl can be considered a hypoosmolaric system for bacteria. In contrast, binding buffer is much more comparable to physiological buffers like phosphate buffered saline based on its ion concentration. Furthermore, it enables functionality of possible further target specific entities, like species-specific aptamers33,41. It also has to be mentioned that nanoparticle concentrations were measured in surface area per volume instead of mass per volume in this study. This was due to the fact that silver nanoparticles do not dissolve in aqueous liquids. Instead, a small amount of silver ions is released from the particle surface and is responsible for biological effects. Hence, particle mass, which is not located at the surface, is not relevant for functionality of these particles, only the active surface contributes silver ions. This is the reason for different effects between nanoparticles and bulk materials and results in surface area being a more meaningful parameter for nanoparticle functionality than total mass33,42. Furthermore, surface weighted concentrations are more meaningful than mass weighted ones when comparing different nanoparticle types that may vary in particle size and density.
To initially test antibacterial activity, planktonic S. aureus was incubated with different concentrations of AgAu NPs for 3 h and metabolic activity was determined using a resazurin assay. Resazurin (7-hydroxy-10-oxidophenoxazin-10-ium-3-one) is a membrane permeable, blue, non-fluorescent dye. Once inside the cell, it can be irreversibly reduced by the reducing equivalents FMNH2, FADH2, NADH, or NADPH to the pink, fluorescent resorufin43. By measuring the increasing fluorescent signal, metabolic activity can be quantified. This assay is a well-established method to analyze antibacterial effects44. This initial test was done as a basis for later biofilm experiments and for comparability to other studies quantifying antibacterial activity. As metabolic activity assays are more reliable for biofilms than CFU44, this method was chosen for planktonic bacteria for comparability. The S. aureus type strain DSM 20231 was used for its reproducibility and comparability, as it has also been used in other studies researching AgAu NPs27,45. During all antibacterial testing, untreated bacteria were used as controls. Dead controls were not included, as, to the authors’ knowledge, there is no standard substance available that could kill mature biofilms with a similar mode of action.
The results showed that an AgAu NP concentration of 1.25 cm2/mL was sufficient to reduce metabolic activity by 81% to 93%, depending on the medium, which could be considered a strong antibacterial effect. Silver ions are known to affect, amongst others, the enzymes PdhB, which is responsible for a reduction of NAD+ to NADH during glycolysis, and 6PGDH, which is part of the reduction of NADP+ to NADPH/H+ in the pentose phosphate pathway21. Consequently, less of these molecules are left to reduce resazurin after treatment with AgAu NPs. Other enzymes relevant for these intracellular mechanisms are affected by silver ions as well, for example PfkA in glycolysis and Pgl in the pentose phosphate pathway21. This further decreases the amount of available NADH and NADPH and finally disrupts the cellular metabolism. The slight difference between binding buffer and 0.2 mM NaCl might be explained by the absolute fluorescence readout values of the resazurin assay (not shown). In 0.2 mM NaCl, the values were roughly 33% lower than in binding buffer, which means metabolic activity is generally reduced in this medium. This could most probably be attributed to the fact that 0.2 mM is a very low concentration of NaCl and the pH of the solution was between 4.5 and 5. These conditions deviate to a high extent from physiological growth conditions of bacteria. Consequently, this could mask or impair the effect of AgAu NPs by influencing the uptake of silver ions or how much metabolic activity could be further reduced by an antibacterial effect. It is also conceivable that the dissolution equilibrium of the AgAu NPs was affected by pH and particularly the chloride concentration. A strong antibacterial activity on planktonic bacteria is typical for other types of nanoparticles containing silver as well. Salunke et al. found that their biologically synthesized Ag NPs, Au NPs, and AgAu NPs exhibited a minimum inhibitory concentration (MIC) of 8 µg/mL against S. aureus, each46. Gurunathan et al. also synthesized their Ag NPs biologically and reported an MIC of 0.75 µg/mL against S. aureus (colony forming units (CFU) reduced by roughly 99%)47. Bhatia and Banerjee reported an MIC of 10 µg/mL against S. aureus for their chemically synthesized AgAu NPs48. For the AgAu NPs used in this study, with an average size of 8 nm and a molar silver-to-gold ratio of 4:1, the effective concentration of 1.25 cm2/mL AgAu NPs corresponds to roughly 1.75 µg/mL of silver. This is in the same magnitude as the MICs in literature, with a relatively strong antibacterial effect. The slight differences in MICs can most probably be attributed to the nanoparticle size, as it was on average 35 nm, 5 nm, and 40 nm in their studies, respectively46,47,48. Since silver from nanoparticles does not dissolve completely but rather nanoparticles release only a small amount of their silver as ions, size and, thus, active surface area affect how many nanoparticles are required for the same effect.
Compared to planktonic bacteria, biofilms are much more resistant to antibacterial action11,13,14. Therefore, antibacterial results cannot be simply transferred but need to be tested individually. For this purpose, AgAu NPs were first tested on early-stage biofilms, which were grown for 5 h only. Preliminary experiments had shown that, for a similarly distinct effect compared to planktonic bacteria, increasing nanoparticle concentration alone had not been sufficient but also longer treatment times were needed. Therefore, 5 h old biofilms were incubated with 10 cm2/mL or 25 cm2/mL AgAu NPs for 18 h and metabolic activity was determined using the resazurin assay. Additionally, biofilms incubated with or without 25 cm2/mL AgAu NPs were stained with two fluorescent dyes. Red propidium iodide can only stain cells with damaged membranes whereas green SYTO9 is able to enter and stain cells with intact membranes. CLSM was then used to acquire 3D images which could be analyzed for biofilm volume and distribution of membrane integrity. During treatment, agitation was provided with a plate shaker to ensure distribution of nanoparticles. In this study, we focused on situations comparable to inserted orthopedic implants in synovial fluid. There, no considerable flow, but a comparable gentle liquid movement is present.
The results showed an effect of AgAu NPs on metabolic activity and membrane integrity of S. aureus early-stage biofilms. Although a 20-times higher nanoparticle concentration was required for a similar effect than on planktonic bacteria, metabolic activity of early-stage biofilms was drastically reduced by AgAu NPs. A considerably higher resilience towards antimicrobial substances is typical for biofilms, while the here reported decrease in susceptibility is at the low end of the 10–1,000-fold decrease found in literature11,36. This resilience is most likely explained by reduced diffusion of the antibacterial substance through the biofilm matrix49,50 and a large amount of bacteria being in the stationary phase and thus being affected less by anti-metabolic action51,52. Another aspect, why a higher concentration of nanoparticles is needed might be the larger number of bacteria being present. The effect of AgAu NPs on membrane integrity was considerably weaker than on metabolic activity. This difference between different viability parameters can most probably be explained by silver’s mechanism of action. As described above, silver damages various metabolic pathways in S. aureus, strongly affecting metabolic activity21. However, disruption of membrane to the point of the membrane being permeable for propidium iodide seems to be a minor silver target compared to metabolic pathways. Even though other studies also reported disruption of membranes of S. aureus and E. coli by silver, they did not compare it to metabolic activity53,54,55. Therefore, future studies should address this observation in more detail by also analyzing whether reduction of membrane integrity is due to direct damage from AgAu NPs or a second-level result of cell death from damage to metabolic pathways. In contrast to metabolic activity and membrane integrity, biofilm volume was not affected by AgAu NP treatment, suggesting no eradicative properties. There are studies reporting eradicative properties for silver56,57. Most probably, whether eradication occurs depends on the properties of the applied silver, e.g. concentration, size, elemental composition, ligands, or other supporting substances. However, the detailed correlation to eradication still needs to be researched.
For all viability parameters of early-stage biofilms treated with AgAu NPs, there are also significant differences between the media used. In binding buffer, biofilms present with less than half of the volume compared to NaCl. This can most probably be explained by the polysorbate-type nonionic surfactant Tween 20 being an ingredient of binding buffer. As surfactant, it increases the potential of the solution to remove entities from solid surfaces. This could possibly have reduced the amount of adherent bacteria. Furthermore, membrane integrity is significantly reduced in NaCl compared to binding buffer. This is most likely due to the low pH and low osmolality of 0.2 mM NaCl being non-physiological to the bacteria. Finally, the smaller effect of AgAu NPs on metabolic activity of early-stage biofilms in NaCl can be explained likewise as for planktonic bacteria, with less reduction possible when the metabolic activity is already low. This might be enhanced by the fact that, due to biofilm volume in NaCl being higher, the ratio of nanoparticle/bacterial cell is lower. Nevertheless, even with differences in magnitude, AgAu NPs were able to exert anti-biofilm properties in both media analyzed. 0.2 mM NaCl is the ideal medium for AgAu NP stability and therefore favorable for effectiveness31. However, it might distort results by being non-physiological for bacteria. Binding buffer is closer to physiological conditions. Since application in 0.2 mM NaCl may not be feasible, binding buffer could be more relevant for possible future applications. Additionally, it would be especially relevant if target specific entities, like species-specific aptamers, were to be attached to the nanoparticles33,41.
After analyzing early-stage biofilms, the antibacterial effect of AgAu NPs was assessed for 24 h-old, mature biofilms using similar methods. As for early-stage biofilms, an effect of AgAu NPs on metabolic activity and membrane integrity could be observed for mature biofilms, while biofilm volume was unaffected. These results demonstrate that AgAu NPs can significantly damage even mature biofilms. However, the effect on metabolic activity was considerably weaker than against early-stage biofilms. This can be explained by the increased number of bacteria and thicker biofilm matrix, which roughly doubled in both media compared to the previous experiment and increase the natural biofilm resilience. The effect on membrane integrity was also reduced, but only slightly. However, membrane integrity is still affected significantly less than metabolic activity. This again strengthens the hypothesis of silver’s mechanism of action primarily targeting metabolic pathways21. As for early-stage biofilms, the different media influenced the magnitude of AgAu NPs’ antibacterial effect, however, differences were less pronounced. This is most likely due to the mature biofilm being more resilient towards the unphysiological properties of 0.2 mM NaCl, while the influence of the surfactant Tween 20 remains similar.
The results presented above clearly point towards a strong impact of AgAu NPs on bacterial metabolic activity in biofilms. As it has been described that silver nanoparticles affect bacteria by releasing silver ions, which are then able to enter and damage the cells24, this effect would be in line with the known molecular mechanism of silver ions21. In the following, this hypothesis was addressed in more detail.
The first precondition for AgAu NPs to affect the metabolic activity of bacteria in biofilms would be a sufficient diffusion through the biofilm matrix, which is known to act as a diffusion barrier49,50. To investigate this, CLSM images of mature biofilms were sliced into 4 µm thick sections horizontally by digital image analysis and then separately analyzed for membrane integrity. The results not only showed an increase of impaired membranes in the deeper layers of the biofilm independently of the medium, but also a homogeneous increase in all sections upon treatment. This indicates that AgAu NPs were able to diffuse through the entire biofilm and could damage cell membranes—and most probably also metabolic activity—even at the bottom biofilm layer. Whether the AgAu NPs themselves are able to diffuse through the matrix, or only the silver ions they release, needs to be further investigated.
Next, the influence of AgAu NPs on the expression of some bacterial genes was analyzed by RNA isolation and qRT-PCR. The influence of silver ions was also analyzed to be able to compare these two agents to get first insights into the nanoparticle’s mechanism. Two chosen genes are known silver targets: pfkA and ahpD encode phosphofructokinase A of glycolysis and alkylhydroperoxidase D in the oxidative stress response system, respectively21. The third selected gene, icaR, represses the ica operon and thus decreases biofilm formation, which would be interesting regarding future applications58. Additionally, the effect of AgAu NPs on total ROS activity was analyzed.
Treatment with AgAu NPs resulted in a slightly increased expression of pfkA and a decreased expression of ahpD and icaR to 0.04-fold and 0.2-fold, respectively. Silver ion treatment was applied in form of AgNO3 and resulted in a fourfold and 25-fold increased expression of pfkA and ahpD, respectively, and a slightly decreased expression of icaR. By analyses on RNA- and protein-level, Wang et al. could demonstrate that upregulation of a gene was indirectly caused by the inhibition of the encoded enzyme by silver ions21. For downregulation, a likewise promoting effect on the protein level can be assumed, for example by damage to relevant feedback loops. Upregulation of pfkA by AgAu NPs is in line with the results for pure silver ions and also with those from Wang et al.21. This points towards a damage of glycolysis also by AgAu NPs and/or its released silver ions, which would additionally fit to the reduced metabolic activity values. As glycolysis is an important virulence factor of S. aureus, its inhibition might be an important tool in treating S. aureus infections59. In contrast to the effect of pure silver ions and to the observations by Wang et al.21, expression of ahpD was reduced upon AgAu NP treatment, despite ROS activity being increased. It could thus not be concluded that AgAu NPs likewise impair the alkylhydroperoxidase D of the stress response system, but instead possibly promotes its stability by targeting feedback loops. Alternatively, the effect could be explained by the magnitude of antibacterial activity. Since AgAu NPs release a small amount of their silver into the solution as ions but AgNO3 dissolves completely, the latter cause a stronger antibacterial effect (Supplementary Fig. 1). Perhaps there are different damage thresholds at which the cells focus on specific pathways or different amounts of silver ions are required to inhibit different components of the cell. ROS activity also built up over time, so it might take some more time for gene expression to react. However, the exact mechanism would need to be clarified in future studies. In contrast, the biofilm inhibitor gene icaR was downregulated upon both, AgAu NP and silver ion treatment. This is in line with findings from Wang et al., who described a final downregulation of icaR at inhibitory silver concentrations60. This could again indicate a potential counterbalancing due to a direct or indirect promotion of the IcaR protein by silver. However, if silver indeed has the potential to impair S. aureus biofilm formation on the molecular level needs to be addressed in further studies. In this study, first insights could be achieved that the effect of AgAu NPs on S. aureus metabolic activity is based on targeting central molecular pathways like glycolysis with similarity to silver ions, which strengthens silver ions to be the major antibacterial agent of these nanoparticles.