The biosensor constructed in the environmental isolate P. putida KT2440 effectively and rapidly, within minutes, demonstrated dose-dependent toxicity of NP of Ag, CuO and ZnO. These findings illustrate that the toxicity was not restricted to bacteria with pathogenic potential. Rather an environmental isolate, studied because of its bioremediation potential, was affected. The NP of Ag, CuO and ZnO were more toxic, causing loss of Lux activity in the biosensor, than their equivalent bulk materials indicating that the nano-size of the material was important. The findings that nano-Ag, nano-CuO and nano-ZnO reduced Lux activity were consistent with the observations by other groups that these NP caused bacterial membrane damage [6, 10, 13]. We speculate that such damage altered the membrane potential of the cell and, we presume, the availability of the FMNH2 required for the Lux activity. Consequently, Lux activity declined in the biosensor cells.
With Ag, the toxic doses of the NP and the ion were similar (~0.2 mg Ag/L) in the KT2440 cells. For Cu, complete loss of light output required exposure to 10 mg Cu/L from nano-CuO compared with 1.0 mg Cu/L of the Cu ions. Similarly, 7–10 mg Zn/L was required for toxicity of nano-ZnO compared to about a ten fold lower dose of Zn ions. Using nano-CuO and nano-ZnO from sources different from our own, nano-ZnO was more toxic than nano-CuO for Vibrio fischeri , compared with similar toxicity with for KT2440. Combinations of nano-Ag and nano-ZnO or nano-CuO and nano-ZnO were not interactive. However, the combination of nano-Ag plus nano-CuO was more inhibitory than their effects alone and the decrease in Lux correlated with reduction in culturability. These findings suggest that the target sites for nano-Ag and nano-CuO differed.
Toxicity as assessed with the pseudomonad biosensor was at lower NP levels than observed in other assays where culturability on solid or liquid media was the bioassay. For instance, in assays in rich medium, nano-ZnO toxicity required 126 mg Zn/L with S. aureus  and for E. coli and B. subtilis 70 mg/L for nano-Ag  compared with 7–10 mg Zn/L from nano-ZnO and 0.3 mg Ag/L for the pseudomonad. The KT2440 bioassays were performed under conditions with no other added metal ions, thus, limiting possible competition with the heavy metal for bacterial binding sites. Likewise, the inorganic and organic materials that compose most bacterial growth media were not present. Such materials might otherwise complex the metals and change bioavailability.
Size and, thus, aggregation of the NP are important in nanotoxicity. For nano-ZnO, particles of 8 nm in size were more toxic to S. aureus than those that were reported to be larger (50–70 nm); these latter products were from the same Sigma-Aldrich source that we used . Thus, it is interesting that we observed by FlFFF that 5 nm NP were present in the nano-CuO and-ZnO preparations. Exposing the biosensor to filtrates of nano-CuO and ZnO that would contain such particles showed dose dependent effects on light output and cell culturability. The FlFFF fractograms also showed that the aqueous NP suspensions prepared from manufactured NP powders were aggregated into poly-dispersed particulates ranging in size range from 70 nm to larger than 300 nm, with the majority of the Cu and Zn mass being associated with the larger particles.
Unlike the treatments with Cu or Ag, nonlethal doses of zinc from bulk, nano-ZnO and the ion increased light output above the control in the bioassays. To explore whether this was due to Zn activation of the promoter of the PP_0588 locus, we added zinc to a biosensor prepared with the fusion of the same luxAB-npt cassette to the promoter of the pseudomonad catalase gene. No increase in light output was observed with addition of Zn in this construct where the promoter region lacked a metal-sensitive motif (data not shown). These findings suggest that increased Lux activity with the KT2440 biosensor by Zn was promoter-driven, in agreement with the existence of a heavy metal-sensitive element in the promoter of the PP_0588 used in biosensor construct. Also, in the biosensor KT2240 strain we observed zinc caused bacteriostasis. Two other studies report that nano-ZnO was bacteriostatic to Streptococcus and Staphylococcus isolates in both broth medium or on solid agar plates [18, 42]. Additionally the antimicrobial effect of nano-ZnO was reported to be sensitive to activation by the UV-radiation from laboratory lighting , conditions under which our assays were performed. Other studies on toxicity of nano-ZnO to mammalian cells found that solubilization of nano-ZnO as well as release of Zn ions from the NP contributed to activity .
Our observations confirmed that the biosensor generated with Lux as the output signal was a sentinel for cellular toxicity. Similar bacterially-based biosensors have been used previously to examine the toxicity of Cu and Zn in sludges . Collectively, our findings show that NP preparations containing the heavy metals Ag, Cu and Zn were toxic to the beneficial environmental microbe, P. putida KT2440, suggesting that the NP at certain concentrations (≤ 1 mg Ag/L, ≈ 10 mg Cu, Zn/L) can be an environmental risk. The impact of the nano-metal oxides on cell culturability was dependent on the chemistry of the particles, with Zn causing bacteriostasis whereas Cu and Ag were bactericidal. FlFFF of the aqueous suspensions of the nano-metal oxides showed most of the mass was in aggregates greater than 300 nm although these ranged downward with another peak at 5 nm. Our findings suggest that further studies on determining the factors that affect aggregation of commercial NP in the environment are required. It is likely that such aggregation would reduce the deleterious effect of as-made NP on nontarget microbes. Implementing conditions promoting NP aggregation could alleviate point-source contamination.