The antibacterial effect of AgNPs depends on their structure and the bacteria with which they interact.
To determine the appropriate proportion of bacteria:AgNPs to be used in the experiments, bacterial viability studies were carried out using E. coli as a reference against various proportions of AgNPs following the procedure described above in the methods section. The results showed that an increase in the concentration of AgNPs had a progressive but limited effect on bacterial viability up to an AgNP:bacteria ratio of 10− 2:1, and then greatly increasing from this point and reaching mortality values around 93–95% at 1:1 ratios, with less pronounced increases at higher ratios. Since the comparison of the toxic effect of the different AgNPs could be best made at concentrations that did not involve total mortality, the 1:1 ratio was chosen as the most suitable for subsequent experiments (data not shown).
To study the toxic effect of the AgNPs on different bacterial species, viable microorganisms were quantified after a variable contact time of between 5 and 30 minutes. In all cases, a high biocidal effect was detected, although the patterns observed were variable depending on the microorganism tested and the type of AgNP used (Fig. 1).
When Gram positive microorganisms were studied, in all cases AgNSs caused a bacterial mortality at least equal to, and in most cases higher than, that observed using the other structures. AgNWs and AgNRs produced very similar effects except in the case of S. pyogenes, where the toxicity of the AgNRs was higher. In general, AgNSs were able to produce their toxic effect after only 5 minutes of incubation, with no significant further alterations being observed when using longer incubation periods. AgNWs and AgNRs, however, produced an effect that was, in several cases, dependent on incubation time in that it progressively increased with increasing incubation time: an effect that was very evident in S. epidermidis and B. globisporus and, to a lesser extent, in S. pneumoniae (Fig. 1a). These results led to a significant negative correlation between bacterial survival values and incubation time for both AgNWs and AgNRs (R = -082 and R = -0.59 respectively) but not for AgNSs.
When the tests were conducted using species of Gram-negative bacteria, the results, in part, coincided with those of the Gram-positives; in most cases the AgNSs produced a greater observable effect at short times little subsequent change at longer times; likewise, AgNWs and AgNRs produced toxicity, although in many cases it was lower than that produced by AgNSs and, in addition, in the cases of H. influenzae, K. pneumoniae, S. typhimurium and S. marcescens, the effect of AgNWs and AgNRs was dependent on the incubation time (R = -0.67 and R = -0.77 respectively). The one exception to the trend was E. coli, where the effects observed were very similar for all three nanostructures and were independent of the incubation time. Another point of note is that for P. aeruginosa and N. gonorrhoeae, the observed toxicity was lower than for the other Gram negatives, even at longer incubation periods. In addition, in N. gonorrhoeae the largest toxic effect was produced by AgNRs (Fig. 1b).
A statistical analysis of the results obtained was carried out, comparing the effect produced by each specific type of nanostructure on the Gram positive bacteria with the those of the Gram negative for each of the different incubation times. In Gram positive bacteria, the existence of a statistically significant difference for AgNSs in comparison with both AgNWs and AgNRs was detected (p < 0.01 in both cases); Although the toxicity values obtained for the different nanostructures became progressively more similar over time, this statistically significant difference was detected throughout the entire study (Fig. 1c). In contrast, in Gram negative bacteria, despite the greater effect produced by the AgNSs, no statistically significant differences were detected compared to the other AgNPs, which could be explained by the great variability in the effects of the AgNSs on the different bacterial species (Fig. 1c). In no case were any statistically significant differences detected for a specific nanostructure when comparing the results of the Gram positive and the Gram negative bacteria (Fig. 1c).
Influence of AgNPs on bacterial growth depends on their morphology and the pathogen involved.
The effect of AgNPs on bacterial growth was also analyzed using long incubation times and different ratios of bacteria:AgNPs, to allow their growth curves to be compared. The results showed that AgNP influence was sometimes dependent on the specific bacterial species involved although general patterns related to the type of nanostructure used were discernable (Figs. 2 and 3). The AgNSs did not show a high capacity for bacterial growth inhibition in most of the microorganisms analyzed (5 out of 6 Gram positives, and 6 out of 7 Gram negatives). The considerable inhibitory capacity of AgNWs, on the other hand, was evident in all the cases analyzed and AgNRs strongly inhibited bacterial growth in the long term in all cases (Figs. 2 and 3).
The antibacterial effect of AgNPs on silver-resistant S. enterica shows distinctive strain-dependent patterns.
The availability of certain silver resistant strains of Salmonella enterica (S.enterica serovar typhimurium 389/97 and S. enterica serovar typhimurium 207/07) allowed us to determine the effect on them of different types of AgNPs as compared to the sensitive phenotype. The viability patterns obtained for these resistant strains after contact with AgNPs were different from the sensitive strain and, interestingly, also differed between the two strains. S. enterica 207/07 showed high resistance to AgNSs, while S. enterica 389/97 showed sensitivity to AgNSs but resistance to AgNWs and AgNRs. In addition, resistance increased with the time of exposure to AgNPs (Fig. 4a).
Biofilms are affected by AgNPs in a way that depends on both the bacterial species involved and the concentration and shape of the AgNP
AgNPs were able to produce a biocidal effect on biofilms produced by different bacteria, although the results showed different patterns to those obtained for planktonic bacteria. In addition, the effect was variable and depended on various factors.
First, general patterns were not observed that depended on the Gram positive or Gram negative nature of the microorganism, resulting in no statistically significant differences between the data for the two microbial groups. However, notable differences between the bacterial species analyzed were detected: while some microorganism, such as B. globisporus, did not experience any sharp decrease in viability induced by the treatments, the bactericidal effect was strong in other such as S. pneumoniae, E. faecalis, S. marcescens and N. gonorrhoeae (Fig. 5a and 5b).
Both the type of nanostructure and the bacteria:AgNP relationship had a decisive influence on the effects observed. In general, and in contrast to the results obtained for planktonic bacteria, AgNRs were more effective (S. aureus, S. epidermidis, B. globisporus, E. faecalis, S. pyogenes, H. influenzae, S. marcescens, P. aeruginosa, N. gonorrhoeae), although in some cases the AgNWs were able to produce similar effects (S.pneumoniae, E. coli, K. pneumoniae) (Fig. 5a and 5b).
The analysis of the pooled results for all the microorganisms studied did not show any significant differences between the different types of AgNPs at bacteria:AgNP ratios of 1:1, although AgNRs showed a greater, and statistically significant, effect at bacteria:AgNR ratios of 1:10 ( p < 0.05 and p < 0.01 compared to AgNWs and AgNSs, respectively), and 1:100 (p < 0.01 and p < 0.001 compared to AgNWs and AgNSs, respectively). AgNWs, meanwhile, showed toxic effects superior to those of AgNSs, although the difference was only statistically significant at the 1:100 ratio (p < 0.01) (Fig. 5c).
Increases in the bacteria:AgNP ratio also caused a decrease in viability in a general way. The results showed statistically significant differences between AgNWs at ratios of 1:100 compared to 1:10 and 1:1 (p < 0.05 and p < 0.001, respectively), although the comparison between the values obtained at ratios of 1:1 with 1:10 did not reach statistical significance. On the other hand, AgNRs showed statistically significant increases in their antibacterial effect as their ratio to bacteria increased in all the cases studied (Fig. 5c). The differential effect of AgNSs, AgNWs and AgNRs on the viability of bacteria in biofilms at different concentrations also had an impact on the correlation between increasing AgNP:bacteria ratio and the antibacterial activity observed: it was low for AgNSs (r = 0.28), higher for AgNWs (r = 0.56) and higher still for AgNRs (r = 0.7) (Fig. 5c).
Culture medium conditioned by the presence of AgNPs exhibits antibacterial activity, but the effect is not dependent on the release of Ag ions.
BHI was maintained for 5 minutes under stirring with different concentrations of AgNPs, after which the nanostructures were removed by centrifugation, and the antimicrobial activity of the conditioned medium was evaluated using E. coli as a test organism. The results showed the existence of concentration-dependent antibacterial activity (Fig. 6a).
To analyze the possibility that the observed antimicrobial effect could be due to the release of Ag ions by the nanostructures, the viability of E. coli was first analyzed when maintained for 30 minutes in a wide range of AgNO3 concentrations; the results showed that viability was reduced by 50% in these conditions at concentrations of around 1 µM (Fig. 6b). Next, the presence of Ag ions in AgNP-free conditioned medium was analyzed through ICP-AES. The tests were carried out using different combinations of BHI and deionized H2O with the different nanostructures, both in the presence and absence of E. coli. In all cases, the concentration of Ag was below 5 ppb (Table 1), ruling out the possibility of attributing the observed antibacterial effect to the release of Ag ions.
Based on these prior data, the possibility that the antibacterial effect was due to the generation of toxic chemical species in the conditioned media, caused by the presence of AgNPs, was investigated. Since this type of effect would be expected to be dependent on contact time, both BHI and deionized water were conditioned using different contact times with AgNPs, followed immediately by contact with E. coli. The results showed the existence of a time dependent toxic conditioning of both media tested for all types of AgNPs analyzed (Fig. 6c)
Another important aspect of this work was to analyze how long the antibacterial effect of conditioned media lasted. To do so, the bactericidal effect on E. coli was determined in BHI and deionized water, both previously conditioned by contact with AgNPs at various intervals after the removal of the nanostructures. The results showed a progressive decrease over time in toxicity, including its total disappearance after prolonged periods (Fig. 6d)
To compare the effect of contact with AgNPs with that produced by the conditioned medium, BHI was conditioned for 30 minutes with the different AgNPs, and its effect on the viability of E. coli was compared with incubations of the bacterium with each AgNP at the same concentration (incubation of 30 minutes in both cases). The results showed that in all cases the viability reduction was markedly greater when there was physical contact with the AgNPs (Fig. 6e).