The Cell Membrane as a Major Site of Damage during Aerosolization of Escherichia coli

Bacterial strains and culture.E. coli MRE162 was obtained from the in-house culture collection at the Defense Science & Technology Laboratories (Dstl, Porton Down) and routinely cultured on Luria-Bertani (LB) agar plates at 37°C for 24 h for the isolation of single colonies. Assessment of viability was made on LB agar plates containing final concentrations of 0.1% (wt/vol) sodium deoxycholate, 4% (wt/vol) sodium chloride, or 10 μg ml⁻¹ polymyxin B sulfate or reactive oxygen species (ROS) scavengers (500 units ml⁻¹ catalase, 10 units ml⁻¹ superoxide dismutase [SOD], and 50 mM sodium pyruvate). All chemicals were purchased from Sigma-Aldrich Ltd. (United Kingdom). For aerosol exposures, LB broth cultures were shaken at 120 revolutions min⁻¹ for 24 h at 37°C, producing a concentration of 1.0 × 10⁹ ± 0.36 × 10⁹ (mean ± standard error) CFU ml⁻¹ in stationary phase. Stock cultures were maintained at −80°C in LB broth containing 10% (vol/vol) glycerol.

Aerosol generation and sampling.Five experiments were performed with samples collected in triplicate for analysis. Aerosols were generated within a sealed Perspex aerosolization chamber with dimensions of 900 mm (length) by 545 mm (height) by 500 mm (depth) and collected on fresh Whatman glass microfiber filters, grade GF/A, 47 mm in diameter (Sigma-Aldrich Ltd., United Kingdom), held in stainless steel filter holders, by drawing compressed air through at a flow rate of 66 liters min⁻¹. The relative humidity and temperature were 49.0% ± 2.1% and 19.3% ± 0.4°C, respectively. E. coli cells were aerosolized at a concentration of 10⁹ ml⁻¹ from either a 3-jet Collison nebulizer or an FFAG (75-μm-diameter orifice; Ingeniatrics Technologías, Seville, Spain) as previously described (40). The 3-jet Collison nebulizer was operated at a pressure of 26 lb/in² (179,000 Pa). The FFAG was operated at a pressure of 16 lb/in² (110,000 Pa) derived from a filtered compressed air supply, and the liquid supplied via a syringe pump (pump 11 plus advanced; Harvard Apparatus, United Kingdom) at a flow rate of 50 μl min⁻¹. The particle size distribution was determined by the helium-neon-laser-optical-system (HELOS/BFS; Sympatec-GmbH Ltd., Germany) utilizing laser diffraction within a parallel laser beam (632.8 nm) over a range of 0.1 to 875 μm. The WINDOX 5 sensor control program recorded and analyzed the particle size distribution over a 30-s period at a rate of 2,000 distributions per s (Sympatec-GmbH Ltd., Germany). The Collison nebulizer and the FFAG produced aerosol particle distributions with mass median aerodynamic diameters (MMAD) of 1 to 3 μm and 12 μm, respectively (40).

Determination of the effect of aerosolization on bacterial viability.E. coli MRE162 was cultured to a density of 1.6 × 10⁹ ± 0.04 × 10⁹ CFU ml⁻¹ in nutrient broth and split into equal 10-ml volumes for aerosolization using the Collison nebulizer or the FFAG for 1, 5, and 10 min. The aerosolized bacteria were collected onto filters as described above. Immediately after collection, the filters were placed into 5 ml of phosphate-buffered saline (PBS) and vigorously shaken for 30 s to remove the deposited bacteria. An extraction efficiency of >98% using Bacillus atrophaeus endospores was demonstrated. The PBS was immediately subjected to downstream processing via spread plating onto nonsupplemented LB agar and supplemented LB agars and by incubation with a range of viability stains. Determination of the numbers of culturable cells was by serial dilution and spread plating of 100-μl aliquots in triplicate onto LB agar or LB agar supplemented with 0.1% (wt/vol/) sodium deoxycholate, 4% (wt/vol) sodium chloride, 10 μg ml⁻¹ polymyxin, or ROS scavengers (500 units ml⁻¹ catalase, 10 units ml⁻¹ superoxide dismutase, and 50 mM sodium pyruvate). The plates were incubated for 24 h at 37°C. The percentage of injury was determined by dividing the viable count obtained on the supplemented LB agar by that obtained on nonsupplemented LB agar prior to multiplication by 100%.

Vitality staining protocols.The samples were also subjected to a range of staining protocols to determine the physiological status of the bacteria. As controls, bacteria were either unstressed cultures held at 37°C or cells heated at 70°C for 1 h. The following kits were used for staining. (i) The Live/Dead BacLight bacterial viability and counting kit. Bacteria that have intact membranes stain green (SYTO-9), while those with compromised membranes stain red (propidium iodide [PI]). (ii) The BacLight bacterial membrane potential kit. Bacteria that possess a membrane potential stain bright green, while those with a depolarized membrane potential have reduced staining, appearing as dull green. Additional control bacteria were treated with 10 μM carbonyl cyanide 3-chlorophenylhydrazone (CCCP) to depolarize the membrane potential. (iii) The BacLight RedoxSensor green viability kit. Bacteria with active oxidoreductases reduce the stain, producing a bright green coloration, while those with inactivated enzymes have reduced intensity, appearing dull green. Additional control bacteria were treated with 10 μM CCCP or 10 mM sodium azide. (iv) The BacLight RedoxSensor CTC viability kit. Bacteria with active respiratory dehydrogenases reduce 5-cyano-2,3-ditolyl tetrazolium chloride (CTC) to an insoluble red fluorescent formazan. Nonrespiring bacteria are not stained. Bacteria are counterstained with the membrane-permeable nucleic acid stain SYTO-24. All of the kits were purchased from Molecular Probes, Inc. The assays were performed immediately after collection of the sample according to the manufacturer's instructions. Immediately after staining, 4% (vol/vol) formalin was added to prevent the occurrence of further cellular processes. Gelatin-coated microscope slides were prepared according to a previously described methodology (40). A 50-μl drop of the stained bacteria was allowed to dry into the gelatin prior to the addition of coverslips and microscopic analysis.

Microscopic analysis.Slides were analyzed with an Olympus IX70 inverted laser confocal microscope (Olympus, Germany) using a ×60 water immersion lens. The fluorescent samples were excited at 488 nm and detected by using either a 530-nm band-pass filter (green fluorescent dyes) or a 565-nm long-pass filter (red fluorescent dyes). Images were recorded and processed using the FluoView imaging system (Olympus, Germany). On each slide, at least 200 cells were counted from 20 fields of view and scored according to the staining protocol used. The percentage of cells possessing a vital function was calculated by dividing the number of stained cells by the total number of cells counted and multiplying by 100.

Flow cytometric analysis.For each aerosolization experiment, five samples per time point (0, 1, 5, and 10 min) were stained with the Live/Dead BacLight bacterial viability and counting kit before flow cytometric analysis on a BD FACSCanto II system (BD Biosciences, Oxford, United Kingdom). A blue solid-state laser operating at 50 mW excited the dyes at 488 nm. Optical filters were set up so that propidium iodide (red fluorescence) was measured at 620 nm and SYTO-9 (green fluorescence) was measured at 585 nm. The data were analyzed using FACSDiva analysis software (Becton, Dickinson, United Kingdom).

Characterization of bacterial injury within particles.E. coli cells prepared in PBS at a concentration of 10⁹ ml⁻¹ were stained with the Live/Dead BacLight bacterial viability and counting kit. The E. coli suspensions were aerosolized and collected into a seven-stage “ultimate” cascade impactor (23). Individual particles were collected on gelatin-coated microscope slides according to a previously described methodology (40). The slides were analyzed as described above. On each slide, at least 200 particles were counted from 20 fields of view. The number of particulates contained within each particle was recorded. The diameter of the original particles prior to impaction into the gelatin was derived in μm from the scale provided by the imaging software. The actual size of the original particle was obtained by multiplication of the diameter of the impacted particle by a conversion factor of 0.59 (21).

Statistical analysis.All data are expressed as the standard error around the mean from either three or five replicate experiments. The two-sample paired t test was used for comparison between groups using the MiniTab release 14, version 02, statistical package.

Evaluation of culture media and viability stains.A range of techniques were used to distinguish between the two populations of E. coli cells that physiologically would be defined as live and dead (Table 1). A 16-h culture at 2.3 × 10⁹ ± 0.22 × 10⁹ CFU ml⁻¹ was split into two equal volumes; one was held at 37°C, while the other was heated at 70°C for 60 min. After 60 min of incubation, the culturable counts on LB medium were 2.42 × 10⁹ ± 0.14 × 10⁹ CFU ml⁻¹ and 0 ± 0 for the populations incubated at 37 and 70°C, respectively. The Live/Dead BacLight bacterial viability stain and the BacLight bacterial membrane potential stain differentiated clearly between cells heated at 37 and 70°C, indicating that one of the sites of damage was the cell membrane. The microscopic counts and flow cytometric counts for bacteria treated with the Live/Dead BacLight bacterial viability stain were not significantly different whether the cells were incubated at 37 or 70°C (P = 0.69). Heating at 70°C for 60 min reduced the percentage of stained bacteria, indicating loss of oxidoreductase and dehydrogenase activities.

A variety of supplemented media prepared from an LB agar base were examined for their ability to detect classical bacterial injury upon temperature shift from 37 to 50°C for 60 min (Table 2). Compared to control cells incubated at 37°C, significant injury was observed in cells cultured on LB medium containing polymyxin (P = 0.006), sodium deoxycholate (P = 0.002), or sodium chloride (P = 0.002) after heating at 50°C. Protection from sublethal heat injury was afforded by the inclusion of ROS scavengers in the LB medium, where no significant difference in injury was observed between cells incubated at 37 and 50°C (P = 0.72).

Nebulization exerts a greater stress than flow focusing on E. coli cells during initial aerosol generation.The stress associated with aerosol generation of a bioaerosol can be divided into two parts. First, some bacteria may go through the process of aerosol generation but not actually be released as airborne bacteria, instead falling back into the liquid reservoir. Second, those released as airborne bacteria will experience a second set of stresses associated with release through the nozzle and the subsequent atmospheric conditions. Prior to aerosol release, the culturable count in the reservoirs (10 ml) of both the FFAG and the Collison nebulizer was 1.24 × 10⁹ ± 0.03 × 10⁹ CFU ml⁻¹ on unsupplemented LB agar. No significant difference was observed on any of the supplemented media (P = 0.82). After 10 min of aerosolization, the reservoirs of the FFAG and the Collison nebulizer yielded counts of 9.48 × 10⁸ ± 0.06 × 10⁸ and 4.22 × 10⁸ ± 0.03 × 10⁸ CFU ml⁻¹, respectively, on unsupplemented LB agar. After 10 min of nebulization using the Collison nebulizer, sublethal injury in the population in the reservoir was observed at levels of 13.3% ± 2.6%, 24.0% ± 5.3%, and 26.8% ± 4.1% on selective medium containing polymyxin, sodium deoxycholate, or NaCl, respectively. No significant differences were observed on any of the supplemented media using the FFAG (P = 0.57).

Prior to aerosol generation, 0.47% ± 0.21% of the E. coli MRE162 cells were PI-stained with the Live/Dead BacLight bacterial viability kit. After 10 min of aerosolization, the percentages of PI-stained cells in the reservoirs of the FFAG and the Collison nebulizer increased to 7.53% ± 0.76% and 32.07% ± 1.65%, respectively; these differences between aerosol generators were significant (P = 0.009).

Classical injury increases during extended aerosolization.Airborne particles containing the bacteria were collected onto filters and kept in an air stream for various periods of time prior to being assayed for viability using the above-described techniques. Bacteria aerosolized by either the Collison nebulizer or the FFAG experienced sublethal injury within the first minute of aerosolization, ranging from 96.5 to 98.2% depending on the selective component employed in the medium (Fig. 1). There was no significant difference between the degree of injury experienced during the first minute when the cells were aerosolized using either the Collison nebulizer or the FFAG for medium supplemented with 0.1% (wt/vol) sodium deoxycholate (P = 0.63) or 10 μg ml⁻¹ polymyxin (P = 0.14). In contrast, a significant difference between the degree of sublethal injury experienced during aerosolization by the two spray devices was observed when using medium containing 4% (wt/vol) NaCl (P = 0.011). Increased exposure to the stresses of aerosolization (10 min) experienced when using the Collison nebulizer significantly increased the quantity of sublethally injured cells, from 98.9 to 99.9%, for medium containing deoxycholate, NaCl, or polymyxin (P < 0.03). In contrast, 10 min of exposure to the aerosolization stress imposed by the FFAG did not significantly increase sublethal injury compared to that observed after 1 min of exposure for medium containing deoxycholate, NaCl, or polymyxin (P > 0.5). Irrespective of the length of time of aerosolization and method of aerosol generation, no significant difference was observed for counts observed on medium containing ROS scavengers (P > 0.83).

• Open in new tab
• Download powerpoint

FIG. 1.

Sublethal injury in Escherichia coli MRE162 populations aerosolized using the FFAG and the Collison nebulizer, determined on LB medium supplemented as indicated.

Cell membrane damage occurs during aerosolization.The temporal effect of aerosolization was examined as a function of the inactivation or maintenance of cellular physiology, such as membrane homeostasis and respiratory activity (Fig. 2). An increase in the percentage of PI-stained cells was observed over the first minute of aerosolization for both the FFAG and the Collison nebulizer, to levels of 39.33% ± 4.01% and 58.87% ± 2.7%, respectively (Fig. 2a). Over the next 10 min, the percentage of PI-stained cells remained constant at 31.63% ± 3.32% for cells aerosolized using the FFAG but increased to 70.87% ± 6.81% for cells aerosolized by the Collison nebulizer. At all time points, the differences between the percentages of PI-stained cells generated by aerosolization via the FFAG and the Collison nebulizer were significant (P < 0.003). Loss of membrane integrity of aerosolized E. coli cells was visualized within the individual 12-μm and 1- to 2-μm particles generated by the FFAG and the Collison nebulizer (data not shown). The degrees of loss of membrane integrity after 5 s of aerosol capture onto gelatin-coated slides within a cascade impactor were 11.33% ± 1.94% and 32.61% ± 4.35% for the FFAG and the Collison nebulizer, respectively.

• Open in new tab
• Download powerpoint

FIG. 2.

Temporal effects on the maintenance of physiological activities of airborne Escherichia coli MRE162 when using the FFAG and the Collison nebulizer for aerosolization. (a) Membrane homeostasis. PI-stained cells indicating loss of membrane integrity by aerosolization using the FFAG (○) or the Collison nebulizer (•) and cells possessing depolarized membranes resulting from aerosolization using the FFAG (□) or the Collison nebulizer (▪) were determined at 0 min and after 1, 5, and 10 min of nebulization. (b) Respiratory enzymatic activities. Loss of respiratory oxidoreductase activity by aerosolization using the FFAG (○) or the Collison nebulizer (•) and loss of respiratory dehydrogenase activity from aerosolization using the FFAG (□) or the Collison nebulizer (▪) were determined at 0 min and after 1, 5, and 10 min of nebulization.

Aerosolization by either the FFAG or the Collison nebulizer produced increased numbers of cells with depolarized membranes (Fig. 2a). Cells aerosolized by the Collison nebulizer exhibited depolarized membranes at a higher rate than cells aerosolized by the FFAG, reaching levels of 68.13% ± 5.24% and 34.23% ± 6.72%, respectively, after 10 min of aerosolization. At each time point, the difference between the percentages of cells with depolarized membranes after aerosolization by the FFAG and the Collison nebulizer was significant (P < 0.0065).

Bacterial respiratory oxidoreductase and dehydrogenase activities decreased over the 10-min period of aerosolization for both the FFAG and the Collison nebulizer (Fig. 2b). There was no significant difference between oxidoreductase or dehydrogenase enzyme activity for either spray device after 1 min of aerosolization (P = 0.73). Thereafter, the rates of loss of oxidoreductase and dehydrogenase activities decreased more sharply, reaching respective levels of 36.97% ± 6.91% and 28.69% ± 7.13% after 10 min of aerosolization for the Collison nebulizer. In contrast, the rates of decline of respiratory enzyme activities in cells aerosolized using the FFAG were lower, reaching levels of 59.23% ± 6.9% and 61.38% ± 7.91% after 10 min. After 10 min of aerosolization, significant differences were observed between the populations of cells aerosolized by the FFAG and the Collison nebulizer with respect to the activities of oxidoreductase (P = 0.0062) and dehydrogenase (P = 0.009) respiratory enzymes.