Disruption of Autolysis in Bacillus subtilis using TiO2 Nanoparticles

In contrast to many nanotoxicity studies where nanoparticles (NPs) are observed to be toxic or reduce viable cells in a population of bacteria, we observed that increasing concentration of TiO2 NPs increased the cell survival of Bacillus subtilis in autolysis-inducing buffer by 0.5 to 5 orders of magnitude over an 8 hour exposure. Molecular investigations revealed that TiO2 NPs prevent or delay cell autolysis, an important survival and growth-regulating process in bacterial populations. Overall, the results suggest two potential mechanisms for the disruption of autolysis by TiO2 NPs in a concentration dependent manner: (i) directly, through TiO2 NP deposition on the cell wall, delaying the collapse of the protonmotive-force and preventing the onset of autolysis; and (ii) indirectly, through adsorption of autolysins on TiO2 NP, limiting the activity of released autolysins and preventing further lytic activity. Enhanced darkfield microscopy coupled to hyperspectral analysis was used to map TiO2 deposition on B. subtilis cell walls and released enzymes, supporting both mechanisms of autolysis interference. The disruption of autolysis in B. subtilis cultures by TiO2 NPs suggests the mechanisms and kinetics of cell death may be influenced by nano-scale metal oxide materials, which are abundant in natural systems.

the orderly cell wall proton and charge distribution is lost, which could lead to the uncontrolled confirmation of teichoic acids, giving rise to unlimited autolytic activity.
The interaction between wall-associated teichoic acids and materials with high specific surface areas, such as nanoparticles, has been highlighted by Jiang et al. 18 , who demonstrated that metal oxide NPs (Al 2 O 3 , TiO 2 , and ZnO) adsorbed and altered the structure of cell wall biomolecules, including teichoic acids in vitro 18 . It was suggest that the resulting teichoic acid structural changes that occur when gram-positive bacterium and metal-oxide nanomaterials interact might be responsible for bacterial toxicity 18 . Through solid-state NMR studies, Wickman & Rice further concluded that, when lipoteichoic acids were simultaneously adhered to peptidoglycan, the positively charged alanine group binds to the surface of negatively charged TiO 2 19 . This suggests that through their interactions, metal oxide NPs may influence cellular processes that are mediated through structural changes in cell-wall proteins, such as autolysis. However, to the best of our knowledge, no research has highlighted the influence of metal oxide surfaces on autolysis. We chose titanium dioxide nanoparticles (TiO 2 NPs) as a model nanomaterial for its high production rate, consumer and industrial usage, and projected release into environmental compartments [20][21][22] . Furthermore, while TiO 2 NPs are a semiconducting material that generate reactive oxygen species when irradiated with ultraviolet light 23,24 , in the absence of a UV source TiO 2 NPs are not known to be particularly toxic to microorganisms [24][25][26] . Using B. subtilis as a model organism, the concentration dependence and kinetics of autolytic disruption, the impact of TiO 2 NPs on the PMF and activity of released lytic enzymes, and the hyperspectral mapping of TiO 2 deposition on the cell wall are described.

Methods and Materials
TiO 2 nanoparticle preparation and characterization. Degussa P-25 TiO 2 (70% anatase/30% rutile) NPs were suspended (0.3% wt.) in 15 mL of sterile lysis buffer (5 mM NaHCO 3 buffer pH 7.0), via probe sonication (550 Sonic Disembrator, Fischer Scientific) with an acoustic power input of 4.11 W (determined using calorimetric method described by Taurozzi et al. 27 , see Supplementary Fig. S1) for 10 minutes in a glass beaker. To serve as a stock solution, the suspension was further diluted to 3 g/L and stored at room temperature in a glass bottle wrapped in aluminum foil. Prior to experimentation, the stock solution was diluted in the lysis buffer to 1000 ppm and probe sonicated (under the same power level as described above) for 10 min and then diluted to the desired TiO 2 NP concentration.
Dynamic light scattering (DLS) measurements were collected on an ALV CGS-3 goniometer with ALV/ LSE-5004 Light Scattering Electronics and Multiple Tau Digital Correlator. All samples were prepared in 5 mM NaHCO 3 and adjusted to pH 7.7 using HCl or NaOH. Two-minute duration measurements were taken in triplicate on 1 mL of 10 ppm TiO 2 NPs. Analysis of the autocorrelation function was performed using a constrained regularization algorithm in the ALV-7004 Correlator Software. Diffusion coefficients, D, were converted to hydrodynamic radius, d H values using the Stokes-Einstein equation (1): around 250 nm (standard deviation: 8 nm) and the zeta potential was found to be − 41.3 mV (standard deviation: 5 mV). TEM imaging was used to observe the primary and aggregate dried TiO 2 NPs size distribution. All TEM images were produced after B. subtilis cells were exposed to TiO 2 NPs. TEM samples were fixed with 1% formaldehyde and 1% glutaraldehyde in 5 mM NaHCO 3 for 24 hours at 4 °C. An aliquot of 1 μ L was then placed on a carbon coated copper grid for examination. TEM images showed that TiO 2 NPs were visible aggregates consisting of primary particles ~20 nm in diameter and distinct from cell components with higher electronic density ( Supplementary Fig. S2). Some TiO 2 NPs appeared to be closely associated with the cell wall of B. subtilis.
Cell culture selection. Bacillus subtilis ATCC 6051 is a well-studied model organism for autolysis research 12,[28][29][30] . It is ubiquitous in the environment and known to have several survival strategies, e.g., biofilm formation, spore formation, and motility 31,32 . It is a robust bacterium used in biotech applications, such as enzyme production 33 , and laboratory toxicity studies as a model gram-positive organism 34 .
Preparation of cell cultures. Experimental cultures of B. subtilis were grown in 20 ml of LB broth-Miller (Fisher Scientific) at 37 °C, aerobically on a shaker plate, for 10 hours, to reach a concentration of ~10 8 CFU/mL. Cells were harvested via centrifugation (4,000 × g, for 7 minutes at 4 °C) followed by resuspension of the pellet in lysis buffer to rapidly deplete the culture of a nutrient source. Washing was repeated twice.
B. subtilis-nanoparticle exposures. The effect of TiO 2 NPs on B. subtilis in a bicarbonate lysing buffer (5 mM NaHCO 3 , pH 7.7) was studied under five different NP exposure conditions: 0, 1, 10, 50, and 100 ppm. All trials were done in the dark to prevent reactive oxygen species generation. A volume of 1 mL washed B. subtilis culture (~10 8 CFU/mL) was added to 9 mL of TiO 2 NP suspension to reach desired TiO 2 NP concentrations. The exposed B. subtilis were incubated at 25 °C in 20 mL glass culture tubes while being shaken. After 1, 2, 4, and 8 h of incubation, 0.3 mL of each suspension was removed for analysis. Viable cell concentrations were determined by plating and counting colony-forming units (CFU) on LB agar. Each treatment was performed in triplicate.
Subsequently, using the same conditions described above, TiO 2 was added at three different time points after B. subtilis was washed and suspended in lysing buffer. First, 1 mL of washed B. subtilis culture (10 8 CFU/mL) was added to three different reactors containing 9 mL of lysing buffer with no TiO 2 NPs. Next, TiO 2 NPs were added to the 3 independent culture tubes (final concentration of 50 ppm TiO 2 ) at 0, 30, and 60 minutes, respectively. Aliquots of 0.3 mL were sampled and viable cell concentrations were determined via CFU plating as described above. Each treatment was performed in triplicate.
Membrane potential (ΔΨ) and proton gradient (ΔpH). The membrane potential (Δ Ψ ) and proton gradient (Δ pH) of B. subtilis cells was determined by the distribution of 3,3'-dipropylthiadicarbocyanine iodide (DiSC 3 (5)) between cells and the suspending medium as described earlier 35 . Due to the compensatory interaction between the transmembrane Δ pH and Δ Ψ , a decrease in fluorescence is an indication of Δ pH dissipation 36 . The interaction between Δ Ψ and Δ pH that governs the PMF is described by equation 2: where R is the gas constant, F is the Faraday constant, and T is temperature. An overnight culture of B. subtilis was grown and harvested as described above being resuspended in 30 mL of lysing buffer with either 0 or 50 ppm of TiO 2 NPs. Immediately after the final resuspension, 3 mL of cell suspension was added to a quartz cuvette. 1 μ L of 3 mM DiSC 3 (5) dissolved in DMSO was added to the cuvette, capped, and shaken by hand. The suspension was then monitored for fluorescence (Horiba Fluoromax-4 using FluorEssence software) at excitation wavelength 643 nm, slit width 4 nm, and emission wavelength 666 nm, slit width 4 nm. The fluorescence was monitored for 1 hour at intervals of 0.5 seconds with an integration time of 0.1 seconds. The emission signal from the photomultiplier tube was corrected by dividing by the reference signal from photodiode detector, which measures the output of the xenon lamp. Counts were kept below 2 million counts per second, as the detector is not linear over this count rate (Personal communication with Horiba technicians).
Hyperspectral microscopy. The NP interactions with B. subtilis were assessed using an enhanced resolution dark-field microscope system (BX51, Olympus, USA) equipped with CytoViva Hyperspectral Imaging System (HSI, Auburn, AL). 20 μ L of sample was deposited on a clean glass slide and covered with a coverslip for imaging. Hyperspectral images were acquired using 100% light source intensity and 0.6 s acquisition time per line. Each pixel of the hyperspectral image contains a light reflectance spectrum, ranging from 400 to 1000 nm with a spectral step of 1.5 nm. Each pixel thus has a spectral signature modulated by the nature of the material it contains 37 .
A spectral pre-library of TiO 2 NPs was build using hyperspectral images of TiO 2 NPs in lysing buffer (100 ppm). The pixels comprising endmembers hyperspectral signal were identified and grouped into the TiO 2 NPs pre-library, following steps previously described by Badireddy and collaborators 37 . The specificity of the pre-library was assessed by mapping it on negative control images (abiotic lysing buffer and non-exposed cells in lysing buffer). Spectra in the pre-library that matched spectra of pixels in the negative control hyperspectral images were considered as unspecific false positives and removed from the pre-library. The remaining spectra built the final TiO 2 NP library, containing specific hyperspectral TiO 2 NP signature (see Supplementary Fig. S3 for specificity tests). This NP library was mapped on hyperspectral images (B. subtilis exposed to 0, 1 or 100 ppm during 1 hour in lysing buffer as described in the "Preparation of cell cultures" section above) to assess NPs-cells Scientific RepoRts | 7:44308 | DOI: 10.1038/srep44308 interactions. The mappings of the NPs pre-library and library was processed using a Spectral Angular Mapping algorithm (SAM, ENVI 5.2 software, in short, an algorithm comparing angles between vectors), whereby two vectors (i.e. spectra bands) with angles ≤ 0.09 rad were considered as similar. Pixels containing the NPs spectral signature were labeled with a chosen color of red. This method has already been tested and validated for bacteria-NP interaction 38 .
Preparation of autolytic cell wall extracts. Cell wall associated autolysins and associated enzymes were extracted according to the method developed by Brown 39 . Briefly, B. subtilis strain was inoculated and maintained in LB Miller medium overnight before 2 liters of fresh LB medium was added and incubated for 12 hours with aeration by sterile air until the cell concentration reached 10 8 CFU/ml. The cell suspension was further concentrated by centrifuging at 6000 × g for 10 min to a final concentration of 5% (wet weight) in 0.75% NaCl. The concentrated cell suspension was centrifuged once more at 6000 × g for 10 min. The supernatant was discarded and the pellet was resuspended in 2 L of 2 M LiCl solution. After one-hour of incubation at 4 °C, the suspension was centrifuged at 16,000 g for 15 min and the pellet was discarded. The cell-associated autolysin extracts were resuspended in the supernatant and stored at 4 °C for further experimentation.

Adsorption of autolysins and peptidoglycan degradation. An autolysin extract isotherm on TiO 2
NPs was measured by the solution depletion method, in which any protein concentration change, ΔC w , in the bulk solution prior to and after exposure to TiO 2 NPs was attributed to adsorption to the surface, according to the following equation: Γ = ΔC w V/A, where Γ is the mass of protein adsorbed per surface area unit, V is the volume of the bulk solution, and A is the TiO 2 NP surface area available for adsorption, calculated using the hydrodynamic radius and assumed spherical shape 40 .
1 mL Reactors were prepared by adding varying concentrations of TiO 2 NPs in the range of 0-10,000 ppm to 55 mg/L of autolysin extract in lysing buffer. The suspensions were immediately capped and rotated for 24 h in the dark at 25 °C. The samples were then centrifuged at 16,000 × g for 15 min (Eppendorf Microcentrifuge 5415 D) to remove TiO 2 NPs and any TiO 2 -adsorbed protein from the bulk solution. A portion of the supernatant was pipetted off and analyzed for protein concentration using the Pierce BCA assay Kit with BSA as the protein standard. The remaining portion of the supernatant (0.5 mL) was incubated with 1.5 mL of purified B. subtilis peptidoglycan (purchased from Sigma-Aldrich) to study the digestion process of peptidoglycan. The change in peptidoglycan concentration was monitored by measuring absorbance at 450 nm over time via a Cary 300 Bio UV-Visible Spectrometer/Cary WinUV software at room temperature.
DNA and L-alanine concentration. Extracellular double-stranded DNA (dsDNA) was monitored using a PicoGreen dsDNA Kit (Molecular Probes, Invitrogen) coupled with a fluorometer (VersaFluor Fluorometer, Bio-Rad) following the supplier's protocol. Two B. subtilis growth conditions were used to create standard curves: a positive control using liquid minimal Davis (MD) medium, and a negative control using lysing buffer. Cells were grown in MD medium because of its ability to keep B. subtilis cells vegetative. 0.2-0.3 mL samples were taken from each reactor at 0, 1, 2, 4, 8, and 12 hrs. Each sample point was diluted, with a portion being plated for CFUs, and 0.1 mL was used to determine dsDNA concentration. Cell suspensions were diluted to within the detection concentration of the PicoGreen dsDNA Kit (25 pg to 1 μ g of dsDNA/mL). 0.1 mL of the kit reagent was added to 0.1 mL diluted sample and incubated in a glass cuvette at 25 °C for 3-4 min before being placed in a fluorescence reader with excitation at 480 nm and emission at 520 nm. L-alanine was monitored using an abcam L-Alanine Assay Kit (ab83394) according to the manufacturers protocol.

Gene expression.
To capture the influence of TiO 2 NPs on B. subtilis gene expression, reverse transcription quantitative polymerase chain reaction (RT-qPCR) was used to detect RNA fold changes. Three genes were chosen for analysis: skfA, sdpC, and lytC; rpoB was used as an internal reference for normalization. Cells were grown, washed, and exposed as described in the "Preparation of cell cultures" section above. The washed cells were then exposed to either 100 ppm TiO 2 NPs or no TiO 2 NPs. Samples for RNA extraction were taken both right before exposure, t = 0, and one hour after exposure, t = 1 hr.
Total cellular RNA was extracted using a Qiagen RNeasy Mini Kit according to the manufacturer's instructions. Purified RNA was verified by quantification using a Qubit 2.0 Fluorometer according to the Qubit RNA HS Assay Kit manufacturer protocol. Extracted RNA was then used as a template to synthesize first-strand complimentary DNA (cDNA) using an iScript cDNA Synthesis Kit according to the manufacturer's instructions and using a BIO RAD MyCycler thermal cycler set at 25 °C for 5 minutes, 42 °C for 30 minutes, 85 °C for 5 minutes, and a holding temperature of 4 °C. Real-time PCR amplification of the cDNA was done using a Applied Biosystems 7500 Real Time PCR System. Primers used for PCR were designed using Integrated DNA Technologies PrimerQuest (http://www.idtdna.com/Primerquest) (Supplementary Table S1). Each 20 μ L PCR mixture contained H 2 O (6.5 μ L), forward primer (1 μ L), reverse primer (1 μ L), probe (0.5 μ L), Taqman master mix (10 μ L), and DNA sample (1 μ L). The amplification program was set to 50 °C for 2 minutes (1 rep.), 95 °C for 10 minutes (1 rep.), 95 °C for 15 seconds (40 reps.), and 56 °C for 1 minute. The rpoB gene was used as a reference for data normalization. All the samples were analyzed in triplicate.

Results and Discussion
Exposure of B. subtilis to TiO 2 NPs in bicarbonate lysing buffer, under dark conditions, resulted in a negative correlation between NP concentration and loss of viable cell counts with time (Fig. 2). In the absence of NPs, the number of viable cells decreased by over 2.5 orders of magnitude in the first hour followed by 3 orders of magnitude loss over the next 7 hours (over a 5-log drop over the 8 hours monitored). A similar loss in cell viability was observed in the presence of 1 ppm of TiO 2 NP. However, in the presence of 10 ppm TiO 2 NPs, total viable cell counts decreased by less than 2 orders of magnitude over 8 hours. Viable cell counts further increased after exposure to TiO 2 NPs at 50 and 100 ppm. The difference between 50 ppm and 100 ppm exposures is significant at the 8 hour time point and might be explained by a biphasic response common in toxicity response curves, where the 50 ppm range would be in the hormetic zone. Overall, exposure of B. subtilis to a nutrient-limited buffer of bicarbonate induced autolysis of cells, and that cell lysis was blocked by the presence of TiO 2 NP in a dose-dependent manner.
The observation of cell lysis in nutrient-limited buffer agrees with previous findings that such solutions cause the onset of autolysis in B. subtilis 12 . Under nutrient poor conditions, a collapse of the protonmotive-force (PMF) in B. subtilis leads to a conformational change in associated teichoic acids and thus an activation of autolytic enzymes that digest the surrounding peptidoglycan (depicted in Fig. 1) 12,41,42 . Thus, the membrane potential (Δ Ψ ) and Δ pH of B. subtilis in lysing buffer was monitored using DiSC 3 (5) fluorescence. Apparent equilibration of the fluorescence signal between cells and the buffer, as indicated by a flat line in the fluorescence profile, took approximately 40 minutes. Ten minutes after fluorescence stability was reached (50 minutes after suspension in lysing buffer in the absence of TiO 2 NPs), the fluorescence signal decreased before increasing 60 fold (Fig. 3). The slight decline and then sharp rise in fluorescence in the no TiO 2 control indicates the collapse of the membrane potential. The dip in Δ Ψ in the no TiO 2 control is an indication of an increase in the membrane potential, which is likely the result of the dissipation of the membrane pH gradient due to the compensatory relationship between membrane pH and membrane potential (described in equation 2) 35 . When B. subtilis cells were suspended in  lysing buffer in the presence of 50 ppm TiO 2 NPs, minimal change in Δ Ψ was observed over the same time period, indicating that the membrane potential, and pH gradient, remains intact. It is possible that the mode of prevention arises from the direct interaction between TiO 2 NPs and the outer layer of wall teichoic acids that make up the bacteria-nanoparticle interface.
Hyperspectral imaging (HSI) was performed on cultures of B. subtilis exposed TiO 2 NPs to visualize TiO 2 NPs-cell interactions (Fig. 4). HSI confirmed that at higher concentrations of TiO 2 (100 ppm), the TiO 2 spectral signature was consistently found to be associated with the surface of individual cells (Fig. 4e2,f2,g2) and bright dots we suspect to be TiO 2 aggregates. Moreover, at 100 ppm TiO 2 exposure, cells existed exclusively as individual, or short chained, planktonic cells (Fig. 4e1,f1,g1). These in vivo visualizations of TiO 2 deposition on B. subtilis cell walls support previous in vitro findings which observed teichoic acid attachment to TiO 2 sufaces 18,19 . At the experimental pH of 7.7, the surface of TiO 2 is negatively charged (− 41.3 mV), which would support electrostatic  interactions between NP surface and positively charged alanine groups, as described previously 19 . It is possible that the protons on the surface of metal oxide NPs alter the microenvironment of the D-alanyl ester and teichoic acid interaction, affecting the activation of autolysin considering it has been demonstrated that the adsorption of a metal oxide surfaces influence the D-alanyl ester and teichoic acid interaction 18 . Therefore, a potential hypothesis for the disruption of autolysis by TiO 2 NP is that association between NP surfaces and cell surface proteins gives rise to a conformational change in the surface proteins that regulate autolytic activity.
A kinetic assessment of cell lysis was performed in which TiO 2 NPs were added before and after the apparent PMF collapse. 50 ppm of TiO 2 NPs was dosed into cultures at 0, 30 and 60 minutes after the cells were suspended in lysing buffer (Fig. 5). Adding TiO 2 NPs immediately (t = 0) after exposure of cells to lysing buffer led to a ~0.5-log decrease in cell counts after 8 hours. Adding TiO 2 NPs 30 minutes after exposing cells to lysis buffer stabilized cell viability for the following 7.5 hours. When dosed at 60 minutes, after the observed PMF collapse, the following 7 hours saw a log drop in cell viability.
The observation that cell viability was maintained with higher efficiency when TiO 2 was added before the observed PMF dissipation suggests that the release of active autolysins from cells that have already undergone autolysis may contribute to lysis of neighboring cells, and led us to consider another potential mechanism by which TiO 2 NPs may interfere with a population undergoing autolysis: by adsorbing released autolysins from cells that have already lysed, limiting cell lysis proliferation. To test this, an adsorption isotherm was first produced for autolysins on TiO 2 NPs (pH 7.7 and 25 °C, Fig. 6a). The autolysin extracts showed a binding affinity to TiO 2 NPs, reaching a surface concentration of 1 mg/m 2 (or 0.18 mg/mg TiO 2 ). In a follow-up experiment to assess the functional activity of lytic enzymes after incubation with TiO 2 NPs, autolysin extracts were exposed to varying concentrations of TiO 2 NPs before being assayed for their ability to degrade peptidoglycan, the natural substrate of autolysins (Fig. 6b). Peptidoglycan degradation was inversely related to the concentration of TiO 2 Prior to exposure to peptidoglycan, cell wall associated enzymes were incubated with varying concentrations of TiO 2 NPs, which are denoted with distinct markers. The Negative control peptidoglycan was not exposed to any cell wall associated enzymes. The Positive control peptidoglycan was exposed to autolysins that were not incubated with TiO 2 NPs. The dotted line represents the baseline of initial peptidoglycan concentration. Connecting lines are provided to guide the eye.
Scientific RepoRts | 7:44308 | DOI: 10.1038/srep44308 NPs present. After 5 hours, concentrations of peptidoglycan decreased by 29% and 57% when autolysins were incubated with 50 and 100 ppm TiO 2 NPs, respectively. In the negative control (no autolysin), peptidoglycan decreased by 0.3 mg/L over 5 hours, whereas in the positive control (autolysins without TiO 2 NP incubation), peptidoglycan was degraded by more than 1.2 mg/L over the same time period.
Generally, the degree of protein adsorption to NPs, and subsequent functionality, depends on both the NP surface and the protein's folding structure, charge, and polydispersivity 43 . Protein adsorption to, and unfolding on, NPs can occur rapidly, within minutes of incubation [44][45][46] . The observed adsorption capacity of cell wall autolysin extracts to TiO 2 NPs (Fig. 6a), and subsequent loss of peptidoglycan-degrading ability (Fig. 6b), could be contributed to the hydrophobicity of the autolysin extracts. It has been suggested that extracted autolysins may be linked to other membrane bound lipids of acids, implying hydrophobic properties 47 . Although protein structure analysis was not part of this work, and thus coming to a conclusion on the forces governing adsorption was not appropriate, the loss of functionality to digest peptidoglycan was correlated to the concentration of TiO 2 incubated with the autolysins. These results are similar to a previous study by Xu et al. 48 which demonstrated a change in structure and loss of functionality in lysozymes after incubation with TiO 2 NPs 48 .
Supplementing the adsorption and peptidoglycan digestion assay results was the observation that the detectable TiO 2 deposition at 1 ppm TiO 2 exposure was associated with the large sticky aggregations of extracellular matter (Fig. 4c1-d2), where it is reasonable to believe starved cells are exuding general stress proteins and cytoplasmic material due to autolysis, leading to an accumulation of autolysins. At lower concentrations of NP exposure (1 ppm TiO 2 ) there was no observable TiO 2 deposition on cell walls (Fig. 4c2,d2). Of the very few observed individual planktonic cells in the 1 ppm TiO 2 exposure, TiO 2 was not mapped on any of them (contrary to HIS visualizations at 100 ppm TiO 2 ).
Finally, as B. subtilis may form spores, it was also important to consider that the loss of viable cell counts could be due to cannibalism and spore formation, rather than, or in addition to, autolysis. Aliquots of B. subtilis cultures 12 hours after suspension in lysing buffer were removed and inoculated to fresh LB media where endospore germination could occur 49 . No growth was observed. B. subtilis are not known to form endospores when rapidly deprived of essential nutrients from the system 3 , as was done in this work. Additionally, spore formation takes 8 hours to complete 50 , at which time DNA is released by the mother cell. After 4 hours in the lysing buffer, the supernatant dsDNA concentration was over an order of magnitude higher than that in the maintenance medium (Supplemental Fig. S4a). An inverse relationship was observed between dsDNA concentration and CFUs. CFUs in the MD media had less than a 5% variation over the 12-hour monitored time period. In addition to dsDNA concentration, the release of L-alanine, one of the major components of peptidoglycan, was also monitored (Supplemental Fig. S4b). L-alanine is released into the extracellular environment when B. subtilis undergo autolysis 51 . L-alanine concentrations increased by a factor of three after 24 hours of suspension in lysing buffer. Cultures exposed to the lysing buffer were also examined in TEM micrographs and HSI; spores were not present. Furthermore, spore formation in B. subtilis is preceded by the expression of two main operons, sporulation killing factor (skf) and sporulation delay factor (sdp), in a process dubbed cannibalism 5,52 . Within these operons, there are two genes, skfA and sdpC, that are known to express lytic peptides that are responsible for lysing nonsporulating sister cells. The expression of these genes was monitored to determine if cannibalistic protein production was responsible for the loss of cell viability in cells suspended in lysing buffer, and to identify if TiO 2 NPs had an influence on the genes expressed during cannibalism. After one hour of suspension in lysing buffer, there was less than a 1.5 fold increase of skfA and sdpC (Supplemental Figure S5). Furthermore, there was no significant difference in skfA or sdpC gene expression between cultures suspended in lysing buffer in the presence (50 ppm) or absence of TiO 2 NPs. In addition to skfA and sdpC, the expression of lytC was also monitored. The lytC gene is responsible for producing a major autolysin, LytC, and is located throughout the vegetative cell wall 43 . As with the Scientific RepoRts | 7:44308 | DOI: 10.1038/srep44308 cannibalistic genes, there was no significant difference in lytC expression between B. subtilis suspended in lysing buffer in either the presence or absence of TiO 2 NPs, revealing that TiO 2 NP were not interfering with expression of autolytic genes. Based upon these analyses, complete loss of cell viability was attributed to autolysis.

Conclusions
Results show that exposure of B. subtilis to a nutrient-limited buffer of bicarbonate induced autolysis in the population. Exposure of nutrient limited cultures of B. subtilis to TiO 2 NPs appeared to prevent the propagation of autolysis in the population and enabled the cultures to survive nutrient stress that otherwise results in decimation of population numbers. TiO 2 NPs prevented, or at least delayed, the dissipation of the protonmotive-force. It is possible that the mode of prevention arises from the direct interaction between TiO 2 NPs and the outer layer of wall teichoic acids that make up the bacteria-nanoparticle interface. Hyperspectral imaging supported this hypothesis by revealing that high concentrations of TiO 2 NP (100 ppm) was correlated with a decrease in cell aggregation and an increase in TiO 2 deposition on the cell wall. Additionally, results indicated that released autolysins might lose peptidoglycan-degrading functionality after adsorption to TiO 2 NP surfaces. Overall, the results suggest two potential mechanisms for the disruption of autolysis and cell death by TiO 2 NP in a concentration dependent manner: (i) directly, through TiO 2 NP deposition on the cell wall, delaying the collapse of the PMF and preventing the onset of autolysis; and (ii) indirectly, through adsorption of autolysins on TiO 2 NPs, limiting the activity of released autolysins and preventing further lytic activity. (Depicted in Fig. 7).
The disruption of autolysis in B. subtilis cultures by TiO 2 NPs suggests the mechanisms and kinetics of cell death may be complicated by metal oxide surfaces, which are common in environmental compartments. The association of nanoparticles to B. subtilis cell-wall and autolytic enzymes may not be confined to metal oxides, but rather linked to the surface chemistry of the particle once released in the environment 53 . Therefore, a variety of materials and surfaces are likely to influence autolysis, especially in natural systems such as groundwater where surfaces are abundant. If metal-oxide NPs delay autolysis through cell-wall associated protein interactions and/or limit the proliferation of autolysins via adsorption, as suggested with the results of the present study, the potential for NP to interfere with other cell-wall mediated interactions and extracellular signaling molecules should be explored. For example, disruption of autolytic activity may have an impact on growth-phase activity as autolysins in B. subtilis are involved in more controlled processes such as cell-wall turnover, cellular division, sporulation, and biofilm formation 54 . The ability to fine-tune the interactions between cells and nanoparticles through coatings that alter the electrosteric and electrostatic properties of particles suggests that the interference of cellular process using engineered nanomaterials may also be fine-tuned 55 .