Exploring multiple effects of Zn0.15Mg0.85O nanoparticles on Bacillus subtilis and macrophages

The increasing number of multidrug resistant bacteria raises a serious public-health concern, which is exacerbated by the lack of new antibiotics. Metal oxide nanoparticles are already applied as an antibacterial additive in various products used in everyday life but their modes of action have remained unclear. Moreover, their potential negative effects to human health are still under evaluation. We explored effects of mixed metal oxide Zn0.15Mg0.85O on Bacillus subtilis, as a model bacterial organism, and on murine macrophages. Zn0.15Mg0.85O killed planktonic bacterial cells and prevented biofilm formation by causing membrane damages, oxidative stress and metal ions release. When exposed to a sub-inhibitory amount of Zn0.15Mg0.85O, B. subtilis up-regulates proteins involved in metal ions export, oxidative stress response and maintain of redox homeostasis. Moreover, expression profiles of proteins associated with information processing, metabolism, cell envelope and cell division were prominently changed. Multimode of action of Zn0.15Mg0.85O suggests that no single strategy may provide bacterial resistance. Macrophages tolerated Zn0.15Mg0.85O to some extend by both the primary phagocytosis of nanoparticles and the secondary phagocytosis of damaged cells. Bacterial co-treatment with ciprofloxacin and non-toxic amount of Zn0.15Mg0.85O increased antibiotic activity towards B. subtilis and E. coli.

Providing an efficient and safe treatment for bacterial multi-drug resistant strains is a major health challenge worldwide 1 . Some bacterial strains have the potential to adhere on any surfaces and form slimy layer known as a biofilm. The formation of biofilms enhances the bacterial resistance to current treatments by slowing penetration of the antibiotic into the biofilm, altering chemical microenvironment of bacterial cells and by enabling cell differentiation similar to spore formation 2 . There is an urgent need to develop novel pharmacological approaches to fight multidrug-resistance pathogenic bacteria and to destroy or prevent their biofilm formation or sporulation. Metal oxide nanoparticles (NP), such as ZnO, CuO and TiO 2 , have already been proven as a good candidate to fight various bacteria [3][4][5][6][7][8][9] . However, their therapeutic applications as antibacterial agents are still limited as these metal oxides at nanoscale may exhibit high cytotoxicity on mammalian cells 10 . Thus, new insights into the complex tri-part interactions, bacterial cells-metal oxide nanoparticles-mammalian cells, are required to rationally design novel biocompatible antibacterial agents. We hypothesis here that mixed metal oxide NPs with synergic effects of two oxides may provide a new solution for an infectious disease treatment.
MgO NP is a commonly used model system for studying surface reactions at nanoscale, mainly due to its simple rock-salt crystal structure and purely inorganic nature. Unlike other NPs with antibacterial activity, such as colloidal silver NPs, which release cytotoxic Ag + ions, or photocatalytic nanoparticles that demand intense irradiation to be efficient, nanostructured MgO are low cost, easy to manipulate and show intrinsic biocompatibility. Two strategies have been proposed to improve the antibacterial activity of MgO NPs 11,12 . First, antibacterial efficiency of MgO can be significantly enhanced by decreasing the size of MgO nanocubes to less than 10 nm 13 .

Results
Zn 0.15 Mg 0.85 O nanoparticles characterization in solution. Zn 0.15 Mg 0.85 O were produced as monophasic nanocubes with a 4 nm average size ( Fig. S-1). The low Zn/Mg ratio was chosen to assure single crystal structure of doped MgO nanocubes and to prevent enhanced cytotoxicity. Indeed, ZnO nanomaterials are instable in aqueous solutions and have tendency to release highly toxic Zn 2+ ions 5,20 .
Zn 0.15 Mg 0.85 O NPs tended to aggregate and precipitate in water. To check whether the solubility of Zn 0.15 Mg 0.85 O NPs increased when ions, detergent or proteins were added to the water, DLS analysis was performed to measure the mean hydrodynamic radius (R H ) of NPs in aqueous solutions (Fig. 1A). The R H value of Zn 0.15 Mg 0.85 O were found to shift from about 1 µm in water to 0.1 µm in PBS containing 150 mM NaCl. Such sizes were much higher than the primary nanoparticle size indicating aggregation of particles in both water and PBS. However, the possibility that NPs were dissolved to some extend cannot be roll out. DLS cannot detect small particles in the presence of large ones as the scattered light at smaller particle sizes is extremely reduced compared with that at the larger particles 21 . The apparent smaller particles with R H ~ 10 nm were found when Zn 0.15 Mg 0.85 O was dissolved in water solution containing BSA, Tween20 or NP40 indicating that protein and non-ionic detergents increased solubility of Zn 0.15 Mg 0.85 O NPs (Fig. 1A). This is in line with the well-established finding that surfactants adsorb on metal oxide NPs and stabilize them in solutions 11,22 . Two main populations of NP aggregates with R H of about 50 and 500 nm were found for Zn 0. 15 Figure 1C shows the increase in XTT absorption peak intensity obtained in various solutions containing Zn 0. 15  *− in all solutions tested while no absorption peak was observed in the absence of NPs (Fig. 1D) in LB medium and PBS, respectively over one hour. Interestingly, in both solutions, no further increase in H 2 O 2 production was observed when NP concentration was increased from 0.5 mg/mL to 1 mg/mL (Fig. 1F).
Antibacterial activity of Zn 0.15 Mg 0.85 O on B. subtilis planktonic cells. The antibacterial activity of Zn 0.15 Mg 0.85 O was examined against B. subtilis, a representative of Gram (+) bacteria. Growth curves of B. subtilis exposed to 0.1 or 1 mg/mL of NPs were determined by monitoring the optical density (OD) at 600 nm over time.
We observed that growth of B. subtilis decreased significantly with the increasing concentration of NPs ( Fig. 2A). Cell viability evaluated by the colony counting method was about 100-fold reduced as compared to the initial concentration of cells (10 8 CFU/ml) upon incubation with 1 mg/mL Zn 0. 15  B. subtilis cultured in the absence of NPs showed intact rod-shaped and round-shaped cells protected with smooth and well-structured membrane layer. All cells were viable and no membrane damage could be observed. After exposing to sub-inhibitory concentration of Zn 0.15 Mg 0.85 O (0.1 mg/mL) bacterial cells still maintained their integrity. However, upon exposure to 0.5 mg/mL Zn 0.15 Mg 0.85 O NPs the attachment of nanoparticles to the bacteria was observed (Fig. 3C). The shape of bacterial cells changed to more irregular suggesting that Zn 0.15 Mg 0.85 O NPs injured the membrane of B. subtilis. Membrane damage further caused membrane leakage and increased cell permeability leading ultimately to bacterial death, as illustrated in Fig. 3C. To preserve cell integrity, the bacterial culture were treated with low concentration of NPs, 0.05 mg/mL ( Fig. S-3). The optimized analyses by LC-MS/MS of three biological replicates resolved more than 1550 membrane proteins or proteins attached to the membrane (Table S1). Among them, 62 proteins showed significant abundance variations (p < 0.05; Kruskal-Wallis test and one way ANOVA) in response to Zn 0.15 Mg 0.85 O NPs (Fig. 3A). Identified proteins mainly belong to 7 large functional categories according to their annotation in the SubtiWiki database 24 (Fig. 3B).
Membrane proteins involved in toxic metal export were among the most up-regulated proteins after bacterial exposure to Zn 0.15 Mg 0.85 O NPs (Fig. 3C, Table 1). In particular, up-regulation of CzcD and CadA, that protect the cell against elevated levels of Zn 2+ -ions 25   AhpC or bacilliredoxins BrxA and BrxB, which perform redox switch in response to oxidative stress. By contrast, we observed down-regulation of the mini-ferritins Dps and MrgA, which act as internal iron metal chelators. Proteomic analysis additionally revealed a wide and prominent effect of Zn 0.15 Mg 0.85 O NPs on proteins associated with information processing, metabolism, cell envelope dynamics or cell division ( Fig. 3C; Table 2).  TEM analysis was done on thin sections of treated macrophages to analyze effects of Zn 0.15 Mg 0.85 O NP on cellular and subcellular morphology. Untreated macrophages served as control (Fig. 5C). Ultrastructural analysis of macrophages incubated with a non-toxic dose of NP (0.1 mg/mL) revealed that cellular and organelle architecture of most treated cells changed (Fig. 5D). The electron dense areas were observed within cells suggesting that Zn 0.15 Mg 0.85 O NPs were internalized. The localization of NPs inside cells was rather dispersed showing a different degree of aggregation. Typically, electron dense aggregates were in a vicinity of membrane-rich regions. Some treated macrophages displayed features of cell death: loss of cell membrane specialization like pseudopodia, scroll-like arrangement of a lipid bilayer called myelin bodies, ballooning degeneration, swelling of mitochondria, shrunken or fragmented nucleus. Those dead cells were often in contact to pseudopodia of neighbor healthy macrophages. This suggests that dead cells were phagocytosed. The healthy macrophages also displayed intracytoplasmic vacuoles with debris suggesting phagocytosis of extracellular debris or autophagocytosis. Autophagy  potentiation effect on ciprofloxacin activity. However, Zn 0.15 Mg 0.85 O (up to 10 µg/disk) had no effect on the activity of penicillin and vancomycin towards B. subtilis and E. coli.

Discussion
The innovative nanomaterials that kill multidrug-resistant bacteria and disturb antibiotic resistant biofilm are needed for industry, agriculture and healthcare. Here, we showed that (i) Zn 0. 15  Zn 0.15 Mg 0.85 O NPs aggregated in water and PBS but tended to dissolve in biological media when coated with proteins and non-ionic surfactants. MgO NPs were shown to mainly produced ROS when their electrons localized in crystal structure defects and holes, that have high oxidation and reduction energies, reduce molecular oxygen to superoxide ion (O 2 *− ) 6 . Subsequently O 2 *− can become a precursor of highly cytotoxic species as hydroxyl radicals (*OH) or singlet oxygen (1O 2 ). In contrast, ZnO NPs were shown to generate mainly H 2 O 2 and OH 34 . Hydrogen peroxide is usually generated upon water oxidation by photo-generated holes forming hydroxyl radicals. Interestingly, we observed that Zn 0.15 Mg 0.85 O NPs produced both O 2 *− and H 2 O 2 when admixed in aqueous solutions. The production of both ROSs increased with increasing NP concentration but saturated at 1 mg/mL Zn 0.15 Mg 0.85 O. Likely, NPs aggregated at high concentrations, which reduced their surface reactivity, and thus the production of ROS.
Oxidative stress in bacteria induced by ROS is considered to play a key role in molecular mechanism of metal oxide NP antibacterial activity. Our proteomic data highlight that ROS generation in combination with Zn 2+ -and Mg 2+ -ion release from Zn 0.15 Mg 0.85 O NPs triggered a broad oxidative stress response (Table 1). For instance, Spx and YraA proteins related to thiol oxidative stress, which interferes with zinc metabolism 35 were up-regulated. Synthesis of both proteins are under the control of the regulator CzrA, which is an indicator of Zn-ions excess. This response suggests an intracellular dissolution of up-taken Zn 0.15 Mg 0.85 O. Remarkably, upregulated GcsH is involved in lipoic acid biosynthesis but also acts as an antioxidant and free-radical scavenger 36 . The exposure to NPs induced other stress response-related proteins, such as RsbW, which is involved in control of the general stress sigma factor SigB activity, and DnaK, a chaperone protein activated in response to heat shock. Several of up-regulated proteins have a function related to cation-dependent cellular processes. For instance, the phosphatase PrpC requires divalent metal cations such as Mg 2+ or Mn 2+ to be active. The SepF protein is a part of the divisome and recruits FtsZ to the membrane. It has been shown that Mg 2+ impact cell division of bacilli due to its involvement in FtsZ assembly 37 .
In addition, B. subtilis recruited proteins that participate in translation and transcription cellular processes that depend on Zn and Mg availability ( Table 2). The levels of at least 10 ribosomal proteins were affected by the NPs. In B. subtilis, composition of ribosomal sub-units can be modified in response to zinc availability 38 . Moreover, up-regulated RpoZ protein is part of the RNA polymerase. Structure and assembly of RNA polymerase multisubunits require Zn 2+ while its catalytic activity is assisted by Mg 2+ 39 . Similarly, an addition of zinc markedly increased yields of active RNA polymerase in Escherichia coli 40 . Zn 0.15 Mg 0.85 O also impacted metabolic pathways. Both up-regulated proteins BglH and BglP (also named Ptv3b, Table 2) are involved in the specific carbon source utilization. In Streptococcus pyogene, shifts in metabolic pathways occured in response to zinc excess 41 . This further suggests that Zn 0. 15   seems to involve the production of the adhesive flagellum protein flagellin, which binds and extracellularly aggregates NPs. We also observed the accumulation of NPs at the external surface of B. subtils cells (Fig. 2C). Since Zn 0.15 Mg 0.85 O NPs were of negative surface potential, their accumulation on the negatively charged bacterial surface suggests that B. subtils made efforts to sequestrate extracellularly NPs, probably to prevent their entry into the cell.
Macrophages are a canonical model of immune-competent cells that are likely to afford the first-line of defense responsible for clearing, processing and degrading foreign materials from circulation. As expected, macrophages phagocytized Zn 0.15 Mg 0.85 O NPs. Upon 24 h of incubation with macrophages, NPs were observed segregated into membrane rich region or dispersed within electron dense area. Such localization suggests that macrophages preceded their transformation as previously observed with Fe 3 O 4 NPs 43 . Interestingly, many treated macrophage cells that showed loss of pseudopodia, swelling mitochondria or fragmented nucleus where linked to pseudopodia of neighbor healthy macrophages suggesting that damaged cells were eliminated by the secondary phagocytose. However, increasing concentration of Zn 0.15 Mg 0.85 O NPs to 1000 mg/L impeded biodegradation mechanism and led to macrophage death.
We show that sub-inhibitory amounts of Zn 0.15 Mg 0.85 O applied with ciprofloxacin had higher antibacterial efficiency compared to ciprofloxacin alone towards E. coli and B. subtilis. This finding suggests a synergistic bacterial killing that may result from the additive bactericidal activity of ROS generated by Zn 0.15 Mg 0.85 O NPs with that of ciprofloxacin, which inhibits bacterial DNA gyrase and cell division. Our proteomics data suggest that Zn 0.15 Mg 0.85 O affected bacterial physiological state, which may also increase bacterial susceptibility to antibiotics. Previously was shown that nano-ZnO enhanced activity of ciprofloxacin and ceftazidime against A. baumannii by modifying bacterial morphology from rod to cocci forms and by inducing bacterial filamentation 44 . Similarly, we observed that Zn 0.15 Mg 0.85 O NPs modified bacterial morphology and damaged cell membrane. The increased permeability of bacterial cell membrane facilitates ciprofloxacin uptake, which is expected to enhance its efficiency. Nevertheless, metal oxide NPs as well as divalent metal ions were reported to complex antibiotics and improve their antibacterial affinity [45][46][47] . For instance, protonated nitrogen atoms of ciprofloxacin quinolone ring may directly bind hydroxylated Zn 0.15 Mg 0.85 O NPs by ionic bonds as evidenced for some divalent metal ions by spectroscopic and X-ray analyses 48 . In addition, the oxygen from the carbonyl group of the ciprofloxacin ring was shown to bind Mg 2+ -ions forming stable complexes 49 . Such interactions between ciprofloxacin and divalent metal ions were shown to facilitate ciprofloxacin interaction with bacterial DNA 50 . Moreover, the efficiency of the Zn 2+ -ciprofloxacin complex was shown to be additionally increased by addition of H 2 O 2 51 . To elucidate the exact mechanism of Zn 0.15 Mg 0.85 O enhancing effect on ciprofloxacin activity a deep structural-functional study remains to be done.
In conclusion, Zn 0.15 Mg 0.85 O NPs are a promising antibacterial agent as exert multiple effects on bacterial cells. The efficiency of Zn 0.15 Mg 0.85 O may be inhibited by particle aggregation in solution that reduce ROS production and metal ion release, and probably by their aggregation at the bacterial surface. The activation of multiple cellular mechanisms by Zn 0.15 Mg 0.85 O, suggests that bacteria need multiple simultaneous gene mutations to acquire resistance to mixed metal oxide NPs. We expect that further sustainable development of antibacterial metal oxide NPs will combine various doping and coating of particles to deliver safe nanomaterials that kill infection agents at high efficiency. Since effects of metal oxide NPs are additive with that of other compounds, the combination of Zn 0.15 Mg 0.85 O NPs with currently used antibiotics could be helpful to prevent new antibiotic resistance crises. In addition, different surfaces can be coated with highly stable and uniform Zn 0.15 Mg 0.85 O NPs to, which opens the way for a wide range of applications in agriculture, industry and medicine.

Methods
Synthesis of Zn 0.15 Mg 0.85 O NPs. Nanoparticles were prepared and characterized as previously described (see Supporting Information S1, for details) 18,19 . Bacterial strains, growth conditions and antibiotics. Bacillus subtilis 168 strain (lab's collection), Escherichia coli TGI strain, and Bacillus subtilis NDmed strain 52 (a kind gift from Roman Briandet) were cultivated in LB medium (10 g/liter tryptone, 5 g/liter yeast extract, 5 g/liter NaCl). Biofilm formation was studied in MSgg medium (5 mM potassium phosphate (pH 7), 100 mM MOPS (pH 7), 2 mM MgCl 2 , 700 μM CaCl 2 , 50 μM MnCl 2 , 50 μM FeCl 3 , 1 μM ZnCl 2 , 2 μM thiamine, 0.5% glycerol, 0.5% glutamate, 50 μg mL −1 tryptophan, 50 μg mL −1 phenylalanine). A Penicillin, ciprofloxacin and vancomycin were from Sigma. 1 mg/ml. Microtiter plates were incubated without agitation at 30 °C. Biofilm amount was measured by discarding the medium, rinsing the wells with phosphate buffered saline (PBS) once, and staining bound cells with a 1% crystal violet solution at room temperature for 20 min. The wells were then washed with PBS buffer three times. The dye was solubilized with acetone:ethanol 20:80, and absorbance at 595 nm was determined using a microtiter plate reader. For each experiment, background staining was corrected by subtracting the crystal violet bound to control wells. To perform pellicle assay 2 µl of bacterial culture grown at 37 °C upon agitation to an OD 600 ~ 0.6 was added to 2 mL of MSgg (alone or with admixed NPs) in a well of 6-well microtiter plate. The plates were incubated without agitation at 30 °C for 72 h and 144 hours. Photographs were required with Samsung Galaxy smartphone. Each assay was performed at least in three independent experiments.

Preparation of B. sutilis spores. B. subtilis 168 cells were induced to sporulate by nutrient exhaustion in
Difco sporulation medium (DSM) 53 . A single colony was picked from a fresh agar plate and used to inoculate 25 ml of DSM and allowed to grow at 37 °C for 48 h. Spores and other cells/cellular debris were collected by centrifugation and washed twice with distilled water. The pellet was then resuspended in 1 ml of distilled water. A heat treatment (20 min at 80 °C) was applied to eliminate vegetative cells. Zn 0.15 Mg 0.85 O NPs were added to spores suspension at final concentration of 1 or 5 mg/ml. After 24 h of incubation, the number of CFU in the presence and in the absence of NPs was determined by plating dilutions on agar plates. Disk diffusion assays. One mL of exponentially growing cells (OD 600 = 0.8-1) of the strain being tested was spread on the Petri plates containing BHI agar medium. The plates were allowed to dry briefly in a laminar flow before 6.6 mm filter paper disks (Whatman) containing the antibiotics and/or nanoparticles (20 µL volume) were placed on the plates (ciprofloxacin 250 µg/mL, penicillin 16 mg/mL, vancomycin 1 µg/mL). The plates were incubated at 37 °C overnight and the zones of inhibition were measured. The values in Fig. 6 are an average of three independent experiments. absorbance values assessed at 560 nm and corrected for a background signal by subtracting the signal measured at 670 nm. Cell survival was expressed as % of cells treated only with water (mock).
Quantitation of ROS. The Amplex red assay was used to quantify H 2 O 2 production. Different amount of Zn 0.15 Mg 0.85 O NPs were admixed with various solutions and incubated during 1 hour in the dark. The suspensions (80 µL) were transferred in a 96-wells plate that contained 20 µl of enzymatic mix (1 µl 10-Acetyl-3,7-dihydroxyphenoxazine (ADHP) reagent, 1 µl horseradish peroxidase and 18 µl assay buffer) in each well. Resorufin fluorescence was measured using a spectrofluorometer (Tecan infinite M200PRO) with excitation and emission wavelengths of 530 and 590 nm, respectively. H 2 O 2 calibration was obtained using H 2 O 2 standard solutions ranging from 100 to 1500 nM. Each experiment was performed in triplicate and repeated at least twice.
The production of superoxide radical ion (O 2 *− ) was evaluated by measuring the adsorption of XTT (2,3-bis (2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide, Sigma). XTT absorbs light at 470 nm when reduced by O 2 *− but not in its oxidized form. XTT dissolved in PBS, pH 7 (0.4 mM). Different concentrations of NPs were mixed with XTT at 0.2 mM and incubated in dark for 5 h. Afterwards, the mixtures were filtered through 0.22 µm syringe filters (Millex) to remove aggregated NPs. The changes in absorbance at 470 nm were monitored using an UV-Vis spectrophotometer.
Cell morphology observation with TEM. TEM analysis were performed to visualize the Zn 0.15 Mg 0.85 O effects on morphology of B. subtilis cells using Hitachi HT7700 electron microscope operated at 80 kV (Elexience, France). Bacterial cells at OD 600 = 1 were incubated with 0.1 mg/ml Zn 0.15 Mg 0.75 O for 1 hour. Cultured cells were fixed with 2% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.2) at room temperature for 1 h. Samples were then contrasted with 0.5% Oolong Tea Extract in sodium cacodylate buffer and post-fixed with 1% osmium tetroxide containing 1.5% potassium cyanoferrate, gradually dehydrated in ethanol (30 to 100%), and substituted gradually in a mixture of propylene oxide-epon and embedded in Epon (Delta Microscopy, Labège, France). Thin sections (70 nm) were collected onto 200-mesh copper grids, and counterstained with lead citrate to allow TEM visualization. Digital images were acquired using a charge-coupled device camera system (AMT).

DLS.
The Z-potential of the NP colloidal solutions was measured using a Zetasizer Nano ZS90 (Malvern, UK).
The results of zeta potential are presented as the average value of three measurements ± SD. Particle size measurements were performed on a Zetasizer Nano-S (Malvern, UK) at 20 °C using a helium-neon laser wavelength of 633 nm and detection angle of 173°. A total of 10 scans with an overall duration of 5 min were performed for each sample. The results were presented as a size distribution.