Surface-bound reactive oxygen species generating nanozymes for selective antibacterial action

Acting by producing reactive oxygen species (ROS) in situ, nanozymes are promising as antimicrobials. ROS’ intrinsic inability to distinguish bacteria from mammalian cells, however, deprives nanozymes of the selectivity necessary for an ideal antimicrobial. Here we report that nanozymes that generate surface-bound ROS selectively kill bacteria over mammalian cells. This result is robust across three distinct nanozymes that universally generate surface-bound ROS, with an oxidase-like silver-palladium bimetallic alloy nanocage, AgPd0.38, being the lead model. The selectivity is attributable to both the surface-bound nature of ROS these nanozymes generate and an unexpected antidote role of endocytosis. Though surface-bound, the ROS on AgPd0.38 efficiently eliminated antibiotic-resistant bacteria and effectively delayed the onset of bacterial resistance emergence. When used as coating additives, AgPd0.38 enabled an inert substrate to inhibit biofilm formation and suppress infection-related immune responses in mouse models. This work opens an avenue toward biocompatible nanozymes and may have implication in our fight against antimicrobial resistance.


S2
ADDITIONAL RESULTS AND DISCUSSION

Environmental pH imposes negligible effects on AgPd0.38's ROS Production.
To examine whether environmental pH affects AgPd0.38's ROS production, we firstly need to select an ROS probe that functions at different pH. Ascorbic acid (AA) is an antioxidant 1 whose strong absorption at 266 nm disappears upon oxidation 2 . 9,10anthracenediyl-bis(methylene) dimalonic acid (ABDA) is a fluorescent molecule but upon capturing 1 O2 becomes non-fluorescent 3 . Singlet oxygen sensor green (SOSG) is weakly blue fluorescent but upon capturing 1 O2 becomes brightly green fluorescent 4 .
When environmental solution pH was varied from 1 to 12, AA's absorbance (at 266 nm) and SOSG's fluorescence emission spectrum both changed significantly ( Supplementary Fig. 8a-c). Indeed, AA's absorbance (at 266 nm) and SOSG's fluorescence (520-600 nm) disappeared when pH ≥ 11 and ≤ 3, respectively. In similar assays, ABDA's fluorescence was affected by change in environmental pH but remained brightly fluorescent across the whole examined pH range (Supplementary Figure 8b). Therefore, we used ABDA as the ROS probe for evaluating the effect of pH on AgPd0.38's ROS production.
Briefly, we incubated AgPd0.38 in ABDA-containing PBS buffer at 37 ºC for 3 h and then recorded ABDA's fluorescence spectrum (λex/λem = 380 nm / 400-600 nm) with a fluorimeter ( Supplementary Fig. 8d-o). For AgPd0.38-treated ABDA at each pH, we included ABDA treated similarly but with PBS at a same pH as a reference, in efforts to exclude the intrinsic influence of environmental pH on ABDA's fluorescence. We found that the relative fluorescence intensity of AgPd0.38-treated ABDA to its S3 corresponding PBS-treated counterpart was barely affected by change in environmental pH ( Supplementary Fig. 8p), suggesting negligible effects of pH on AgPd0.38's ROS production.
To examine whether environmental temperature affects AgPd0.38's ROS production, we firstly need to select an ROS probe that functions at different temperature. As environmental temperature was changed (4-55 ºC), AA's absorbance was influenced significantly ( Supplementary Fig. 9a) while the fluorescence emission spectra of ABDA and SOSG were almost unaffected ( Supplementary Fig. 9b-c). Therefore, we selected ABDA and SOSG as the ROS probes for evaluating the effects of environmental temperature on AgPd0.38's ROS production.
Briefly, we incubated AgPd0.38 in ABDA-containing PBS buffer for 3 h at a specified temperature (4,15,25,37,45, or 55 ºC) and then recorded the fluorescence emission spectrum of ABDA (λex/λem = 380 nm / 400-600 nm) with a fluorimeter ( Supplementary   Fig. 9d-i). For AgPd0.38-treated ABDA at each temperature, ABDA treated similarly but with PBS at a same temperature was included as a reference, to exclude the intrinsic (though slight) influrence of temperature on ABDA's fluorescence. We found that the relative fluorescence intensity (at 433 nm) of AgPd0.38-treated ABDA to its corresponding PBS-treated counterpart was barely affected by change in environmental temperature ( Supplementary Fig. 9j). Similar results were observed when replacing ABDA with SOSG ( Supplementary Fig.10). Collectively, these results suggests S4 negligible effects of environmental temperature on AgPd0.38's ROS production.

Buffer agent imposes negligible effects on AgPd0.38's ROS Production.
To examine whether buffer agent affects AgPd0.38's ROS production, we used AA as an ROS probe and incubated AgPd0.38 with AA (at 37 ºC, for 3 h) in different buffers (at pH = 7.4), followed by recording AA's absorption spectrum with an UV-vis spectrometer ( Supplementary Fig. 11a-d). For AgPd0.38-treated AA in each buffer, AA treated similarly but with the same buffer (i.e., without AgPd0.38) was included as a reference, to exclude potential influence of buffer agent on AA's absorbance. We found that the relative absorbance of AgPd0.38-treated AA to its corresponding buffer-treated counterpart was negligibly affected by the change in buffer agent (Supplementary Fig.   11a-d), suggesting negligible effects of buffer agent on AgPd0.38's ROS production.

A lipid bilayer is permeable to free 1 O2.
Chlorin e6 (Ce6) is an organic photosensitizer that generates free 1 O2 upon light irradiation (λ ~ 660 nm) 5,6 . To examine whether a lipid bilayer is permeable to free 1 O2, we coated Ce6-preloaded PLGA (poly(lactic-co-glycolic acid)) nanoparticle (Ce6/PLGA) with a lipid bilayer (DOPC:DSPE-PEG = 0.90:0.10) (Fig. 1f) and used the resulting Ce6/PLGA@lipid particle ( Fig. 1g and Supplementary Fig. 17) as a model S5 for nanoparticles that produce free 1 O2. Specifically, dispersion of Ce6/PLGA@lipid in SOSG-containing PBS was irradiated with a solar simulator (at 0.1 W/m 2 , 5-min) and then submitted to recording on the fluorescence emission spectrum of SOSG (λex/λem = 504 nm/510-700 nm), with that of Ce6/PLGA in SOSG-containing PBS included for comparison. Control is nanoparticle-absent PBS that contains same dose of SOSG. Our results (Fig. 1h) show that the fluorescence intensity of SOSG in light-irradiated Ce6/PLGA@lipid dispersion was significantly higher than that of SOSG in lightirradiated PBS, indicative of SOSG oxidation to appreciable extent in response to 1 O2 generated by Ce6 upon light irradiation. Of note, the fluorescence emission spectrum of SOSG in Ce6/PLGA@lipid dispersion was almost identical to that of SOSG in Ce6/PLGA dispersion (Fig.1h), indicative of comparable extent of SOSG oxidation by 1 O2, suggesting negligible retardance on the efflux of free 1 O2 and/or consumption of free 1 O2 by the lipid bilayer coating in Ce6/PLGA@lipid. In short, a lipid bilayer is permeable to free 1 O2.