Evaluation of Ultrasound-Induced Damage to Escherichia coli and Staphylococcus aureus by Flow Cytometry and Transmission Electron Microscopy

Abstract

As a nonthermal sterilization technique, ultrasound has attracted great interest in the field of food preservation. In this study, flow cytometry and transmission electron microscopy were employed to investigate ultrasound-induced damage to Escherichia coli and Staphylococcus aureus. For flow cytometry studies, single staining with propidium iodide (PI) or carboxyfluorescein diacetate (cFDA) revealed that ultrasound treatment caused cell death by compromising membrane integrity, inactivating intracellular esterases, and inhibiting metabolic performance. The results showed that ultrasound damage was independent of initial bacterial concentrations, while the mechanism of cellular damage differed according to the bacterial species. For the Gram-negative bacterium E. coli, ultrasound worked first on the outer membrane rather than the cytoplasmic membrane. Based on the double-staining results, we inferred that ultrasound treatment might be an all-or-nothing process: cells ruptured and disintegrated by ultrasound cannot be revived, which can be considered an advantage of ultrasound over other nonthermal techniques. Transmission electron microscopy studies revealed that the mechanism of ultrasound-induced damage was multitarget inactivation, involving the cell wall, cytoplasmic membrane, and inner structure. Understanding of the irreversible antibacterial action of ultrasound has great significance for its further utilization in the food industry.

INTRODUCTION

With increasing demands for safe, nutritious, and minimally processed products, ultrasound has attracted great interest as an alternative nonthermal technology for microbial inactivation in food preservation (1, 2). Since the 1960s, many studies have been conducted to investigate the bactericidal effects and mechanisms of ultrasound (3–5). At present, acoustic cavitation is most widely accepted as the mechanism of sterilization by high-power ultrasound (frequencies between 18 kHz and 100 kHz). The formation, growth, and collapse of cavitation bubbles in liquid media cause mechanical effects (microstreaming, high shear force, shock waves) and sonochemical reactions (free radicals, hydrogen peroxide), eventually resulting in the impairment or disruption of bacterial cells (6–8). The inhibitory effects of ultrasound on microbial cells are multifactorial, including pore formation, cell wall thinning, cell membrane disruption, release of cytoplasm contents, and damage to DNA structure (9, 10). However, there is still no consensus about the primary effect of ultrasound that causes cell death. Most researchers have argued that the primary target of ultrasound is the cytoplasmic membrane, which consists of lipoprotein layers (11, 12). Ananta et al., on the other hand, suggested that ultrasound-induced cell death might not be related to cytoplasmic membrane damage (13).

There have been numerous studies on the antimicrobial efficacy of ultrasound in the food industry, which have been reviewed by Awad et al. (2), Bilek et al. (14), and de São José et al. (3). Unfortunately, in order to determine the impact of ultrasound treatment, most of these studies employed the classical plate count method which is a retrospective method because of the long incubation time. More importantly, some bacteria present in the environment cannot be cultured but are nevertheless metabolically active; they are known as viable but nonculturable (VBNC) cells (15–17). Pathogens in a VBNC state may remain virulent and produce enterotoxins (18). Therefore, culture-independent techniques are superior to the plate culture method for real-time, quantitative assessment of cell viability and functionality.

Flow cytometry (FCM) offers a powerful tool for real-time data acquisition and multiparameter analysis of cell populations at the single-cell level (19–21). Parameters such as bacterial size and cell complexity can be obtained by means of FCM scattering signals, corresponding to forward scatter (FS) and side scatter (SS), respectively (22, 23). Currently, there is a trend in FCM to combine various fluorescent probes that are targeted to specific cellular components, such as nucleic acids and intracellular esterase (20, 24). For example, propidium iodide (PI) is a membrane-impermeant nucleic acid dye but can enter into cells with compromised membranes and bind to the DNA and RNA, emitting red fluorescence (25, 26). In contrast, carboxyfluorescein diacetate (cFDA), a lipophilic nonfluorescent precursor that diffuses readily across cell membranes, is widely used for the assessment of nonspecific enzymatic activity in cells. Once inside the cell, cFDA is converted by nonspecific esterases into a polar, membrane-impermeant green fluorescent compound, carboxyfluorescein (cF) (27, 28). Therefore, we can gain a better understanding of process-induced changes, such as compromised membranes and inactivated esterase, by using both PI and cFDA. More importantly, cell population heterogeneities may be determined by the use of combined FCM probes (29).

In the past, it was claimed that physical sterilization technologies, such as pulsed electric fields and UV radiation, were all-or-nothing processes. This meant that cells either were entirely killed or survived intact, and no sublethally damaged cells were generated (30). However, some recent studies (31–33) have utilized FCM combining two probes to demonstrate the existence of a sublethal state under the stress of such physical treatments. As we know, pathogens in the sublethal state show an extended lag time due to physiological and biochemical changes, but they still have the ability to recover and restart growth under suitable conditions, posing a risk to public health. Whether ultrasound treatment causes such sublethal damage to cells remains questionable.

Therefore, the aim of this study was to gain a better understanding of the antibacterial mode of action of ultrasound against Escherichia coli and Staphylococcus aureus by use of the FCM technique. We evaluated the ultrasound-induced damage to the membrane integrity, intracellular enzyme activity, and metabolic activity of E. coli and S. aureus at different initial concentrations in a dynamic bactericidal process. In addition, FCM combined with a double-staining method was used for real-time, quantitative analysis of sublethally injured E. coli and S. aureus cells in order to determine whether the ultrasound treatment was an all-or-nothing process. Also, transmission electron microscopy (TEM) was used to obtain distinct images with which to investigate the effects of ultrasound on morphological changes in E. coli and S. aureus.

MATERIALS AND METHODS

Bacterial strains and preparation of bacterial suspensions.

Experiments were performed using the Gram-negative bacterium Escherichia coli ATCC 25922 and the Gram-positive bacterium Staphylococcus aureus ATCC 25923, which were purchased from Hope Bio-Technology Co., Ltd., Qingdao, Shandong, China. Freeze-dried cells were activated according to the supplier's guidelines. Each strain was maintained on a nutrient agar (NA) (Base Bio-Tech Co., Hangzhou, China) slant. The stock cultures were transferred to 100 ml nutrient broth (NB) (Base Bio-Tech Co., Hangzhou, China) and were grown in an air bath incubator with a reciprocal shaker (TS-2102C; Tensuc, Shanghai, China) for 24 h at 150 rpm and 37°C until stationary phase. Cells were harvested by centrifugation at 2,320 × g and 4°C for 10 min (TGL-20M centrifuge; Kaida Scientific Instruments Co., Ltd., Changsha, Hunan, China) and were washed twice by resuspension in a 0.85% sterile saline solution. In this study, the initial bacterial concentration in each culture was approximately 10⁹ CFU/ml.

Ultrasound treatments.

An ultrasound processor for stationary operation (Scientz-II D; Ningbo Scientz, Zhejiang, China) was used as a treatment system in this study. For low-frequency ultrasound (20 kHz), 30 ml of a cell suspension (dilution, 10⁶ or 10⁸ CFU/ml) was placed in a reaction vessel (an 85-ml cylindrical beaker) and was sonicated by submerging a 10-mm-diameter probe (operating immersion depth, 2.0 cm) in the suspension. Accordingly, the maximum output sonic power density of the sonotrode is 300 W · cm⁻². The ultrasonic power and irradiation time were set at 60 W · cm⁻² and 0 to 20 min, respectively. To prevent a lethal thermal effect, the temperature of the suspension in treatment was maintained at (20 ± 1)°C in an ice water bath.

VPC.

The cultivability of E. coli and S. aureus cells was evaluated immediately after each treatment by the viable plate count (VPC) method. Untreated and ultrasound-treated suspensions were serially diluted in a 0.85% sterile saline solution. From selected dilutions, 1-ml portions were pour plated onto plate count agar (PCA) (Hope Bio-Technology Co., Ltd., Qingdao, Shandong, China). The plates were then incubated at 37°C for 24 h. The experiments for each sample were carried out in triplicate.

Staining procedure and flow cytometric analysis. (i) PI staining.

Propidium iodide (PI) (Sigma-Aldrich Co., USA) was dissolved in distilled water to make a 1.5 mM stock solution and was kept in a refrigerator (4°C) in the dark (34). One milliliter of each diluted suspension was incubated with 10 μl of PI for 10 min to allow labeling of membrane-compromised cells. After centrifugation, pelleted cells were resuspended in 1 ml 0.85% sterile saline solution to remove excess PI. Samples were kept in the dark on ice and were used within 1 h for FCM analysis.

(ii) cFDA staining.

A stock solution (1 mM) of carboxyfluorescein diacetate (cFDA) (Sigma-Aldrich Co., USA) was prepared by dissolving that dye in acetone and was stored at −20°C in the dark. Ultrasound-treated cells were incubated with 50 μM cFDA at 37°C for 15 min to allow intracellular enzymatic conversion of cFDA into carboxyfluorescein (cF). The cells were then washed to remove excess cFDA. Stained samples were kept in the dark for no more than 1 h until FCM analysis was performed.

(iii) Metabolic performance of the cell.

Experiments were performed to measure the metabolic performance of ultrasound-treated cells in extruding intracellularly accumulated cF activity, a process that was most likely mediated by an ATP-driven transport system. Cells stained with cF were incubated with 20 mM glucose for 20 min at 37°C. In FCM dot plot analysis, cF extrusion could be monitored by the apparent shift of the population from quadrant 3 (cF-stained cells) to quadrant 4 (unstained cells) upon glucose addition, owing to the loss of fluorescence. Based on the extent of the population shift after a 20-min incubation period, we were able to discern the degree of injury to cellular pump activity. The following equation was used to measure the metabolic performance of ultrasound-treated cells: $\text{cFA}(\%)=[1-({q^3}_{\mathrm{glu}}/q^3)]\hspace{0.17em}\times \hspace{0.17em}100$, where the percentage of cFA is the measure of metabolic performance in extruding cF, q3 is the percentage of the population in quadrant 3 prior to glucose addition, and q3_glu is the percentage of the population in quadrant 3 after a 20-min incubation with glucose.

(iv) Double staining with PI and cFDA.

Ultrasound-treated cells were initially incubated with 50 μM cFDA at 37°C for 15 min to allow the intracellular enzymatic conversion of cFDA to cF. Cells were then centrifuged and were washed with a 0.85% sterile saline solution to remove excess cFDA. Then 30 μM PI in an ice bath was added, and the mixture was incubated for 10 min to allow the labeling of membrane-compromised cells. The cells were then washed to remove excess PI.

(v) FCM measurement.

Analysis was performed with a Gallios flow cytometer (Beckman Coulter Inc., Miami, FL, USA). The forward scatter (FS), side scatter (SS), and green (FL1) and red (FL3) fluorescence of each cell were measured, amplified, and converted into digital signals for further analysis. cF emits green fluorescence at 525 nm following excitation with laser light at 488 nm (FL1 channel), whereas red fluorescence at 620 nm is emitted by PI-stained cells (FL3 channel). All registered signals were logarithmically amplified. A gate named “[A],” created in the dot plot of FS versus SS, was preset to distinguish bacteria from artifacts (see Fig. 3 and 4). Data acquisition was set to 20,000 events at a low flow rate (400 to 600 events/s). The Kaluza software package (Beckman Coulter Inc., Miami, FL, USA) was used to analyze flow cytometry data.

FIG 3.

Fluorescence dot plots of E. coli (a to f) and S. aureus (g to l) in response to staining with cFDA and PI after ultrasound treatment of bacteria at an initial concentration of 10⁶ CFU/ml. The duration of treatment was 0 min (a and g), 4 min (b and h), 8 min (c and i), 12 min (d and j), 16 min (e and k), or 20 min (f and l). Quadrants 1 to 4 are indicated.

FIG 4.

Fluorescence dot plots of E. coli (A to F) and S. aureus (G to L) in response to staining with cFDA and PI after ultrasound treatment of bacteria at an initial concentration of 10⁸ CFU/ml. The duration of treatment was 0 min (A and G), 4 min (B and H), 8 min (C and I), 12 min (D and J), 16 min (E and K), or 20 min (F and L). Quadrants 1 to 4 are indicated.

TEM analysis.

The samples were centrifuged at 5,940 × g for 10 min at 4°C to collect bacterial precipitates. Then the precipitates were washed twice with a NaCl solution (0.85%). After that, the washed specimens were first fixed with 2.5% glutaraldehyde (TAAB) for more than 4 h and then washed three times with phosphate buffer (0.1 M; pH 7.0) for 15 min. The cells were postfixed with 1% OsO₄ for 1 to 2 h and were again washed three times with phosphate buffer (0.1 M; pH 7.0) for 15 min. Next, the specimens were dehydrated by a graded ethanol series (30%, 50%, 70%, 80%, 90%, 95%, and 100%) and were then transferred to absolute acetone. The dehydrated specimens were embedded in Spurr's resin, incubated for 4 h at room temperature, and then placed in an oven (65°C) for 24 h to polymerize. Finally, the prepared specimens were sectioned in a Leica EM UC7 ultramicrotome and were stained with uranyl acetate and alkaline lead citrate for 5 to 10 min. The specimens were examined in a JEM-1230 TEM (JEOL, Japan).

RESULTS AND DISCUSSION

Enumeration of surviving cells by plate counts (cultivability).

The rates of killing of 10⁶ CFU/ml E. coli and S. aureus after treatment for 20 min were 98.14% ± 0.89% and 91.68% ± 1.45%, respectively, while the rates of killing of 10⁸ CFU/ml E. coli and S. aureus after 20 min of treatment were 99.24% ± 0.44% and 92.75% ± 1.06%, respectively (Table 1). Although the factors affecting ultrasound sterilization have been explored extensively (2, 35), studies of the relationship between inactivation efficiency and the initial bacterial number have been scarce. Our results indicated that the killing rate was independent of the initial bacterial number. Al Bsoul et al. investigated the effect of initial cell concentrations on the destruction process and also found no significant effect on the disinfection rate (36). It was possible that with a higher bacterial density, the mechanical and chemical energy generated by ultrasonic waves could act on a higher proportion of bacteria. On the other hand, the extent of aggregation might be higher with a larger initial number of bacteria, which might result in more resistance to cell destruction by ultrasound. Our findings and those of Al Bsoul et al. might be due to various effects on the medium through which these mechanisms offset each other. We also observed that the Gram-negative bacterium E. coli was more sensitive to ultrasound treatment than the Gram-positive bacterium S. aureus (Table 1), a finding that was in accordance with previously published studies (37–40). According to the literature, this result might be due to the thicker and more tightly adherent layer of peptidoglycan in Gram-positive cells (41). But some researchers reported no significant relationship between bacterial structure and ultrasonic inactivation (11). We believe that different results with regard to the lethal effects of ultrasound can be associated with processing conditions, such as the intensity applied or the duration of treatment, the position of the ultrasonic probe, and the properties of the medium. Thus, for future research comparing and evaluating ultrasound treatments, it will be important to obtain detailed treatment parameters.

TABLE 1.

Surviving populations of E. coli and S. aureus, at different initial concentrations, after treatment with ultrasound for 20 min^a

Bacterium and initial concn (log CFU/ml)	Surviving population (log CFU/ml) at the following time (min):
	4	8	12	16	20
E. coli
6.02 ± 0.06	5.63 ± 0.05	5.31 ± 0.15	4.86 ± 0.13	4.42 ± 0.05	4.29 ± 0.17
8.04 ± 0.18	7.64 ± 0.08	7.24 ± 0.16	6.58 ± 0.12	6.21 ± 0.12	5.92 ± 0.20
S. aureus
6.09 ± 0.08	5.82 ± 0.05	5.57 ± 0.04	5.39 ± 0.13	5.10 ± 0.03	5.01 ± 0.07
8.12 ± 0.12	7.87 ± 0.06	7.65 ± 0.08	7.40 ± 0.10	7.15 ± 0.05	6.98 ± 0.06

^aValues are means of triplicate measurements ± standard deviations.

Effects of ultrasound on the cell membrane, intracellular enzymes, and metabolic performance.

In order to gain a deeper understanding of ultrasound-induced cell injury and death, two fluorescent probes, PI and cFDA, were used to monitor membrane integrity and esterase activity, respectively. E. coli and S. aureus at different initial concentrations were treated with ultrasound for 20 min, and the numbers of cells penetrated by PI and cFDA were determined (Fig. 1 and 2). Similar changes were found with initial concentrations of 10⁶ and 10⁸ CFU/ml, indicating that the initial bacterial concentration is irrelevant to the extent of cell damage. Figure 1 shows that PI-stained E. coli cells were progressively shifted toward larger channel numbers than PI-stained S. aureus cells as the duration of treatment increased; large channel numbers (as shown in Fig. 3 and 4) indicate that the cells were positively stained. As shown in Fig. 2, the fraction of E. coli cells accumulating cF increased in proportion to time within the first 8 min and then decreased gradually with increasing duration of treatment, while the fraction of S. aureus cells accumulating cF decreased gradually during the whole process.

FIG 1.

Percentages of PI-stained cells after E. coli and S. aureus at different initial concentrations were treated with ultrasound for 20 min. Individual results are expressed as means ± standards errors (n = 3).

FIG 2.

Percentages of cF-stained cells after E. coli and S. aureus at different initial concentrations were treated with ultrasound for 20 min. Individual results are expressed as means ± standard errors (n = 3).

Prior to treatment, we observed that most viable E. coli cells did not yield as high cF fluorescence as S. aureus cells. This behavior could be attributed to the presence in Gram-negative bacteria of an outer membrane (which does not allow lipophilic probes, such as cFDA, to diffuse freely across the cytoplasmic membrane because of the lipopolysaccharide [LPS] layer) with a very stable structure, abundant drug efflux, and restrictive porins (42, 43). For E. coli, as the processing time increased, the permeation of cFDA increased significantly, showing high efficiency in damaging the outer membrane. Similar results were observed with the application of ultrasound to enhance the permeability of the outer membranes of Gram-negative bacteria to antibiotics (44). It was suggested that ultrasound created holes in the outer membrane lipid bilayer as well as perturbing the stability of LPS molecules and porins, thus facilitating the penetration of the outer membrane (45). Given the existence of the outer membrane, no considerable changes in the PI-positive population were observed at first for E. coli compared to S. aureus, indicating that the E. coli cytoplasmic membrane was not severely affected under those circumstances. This phenomenon was in agreement with a published study (13) proving that under certain processing conditions, ultrasound could be selective, merely destabilizing the E. coli outer membrane without rupturing the cytoplasmic membrane. Thus, the outer membrane of an E. coli cell is an important target site for ultrasound treatment.

As the processing time increased, E. coli eventually showed a higher percentage of PI-stained cells than S. aureus. This finding was in a good agreement with the plate count results showing that the Gram-negative bacterium E. coli was more sensitive to ultrasound treatment than the Gram-positive bacterium S. aureus. Therefore, ultrasonic inactivation is associated with damage to the cell wall and cell wall structures. After E. coli cells had been treated for 8 min, the proportion of cF-stained cells started to decrease gradually with increased damage to the cytoplasmic membrane. For S. aureus, membrane damage and intracellular esterase inactivation were synchronized throughout the process. Thus, the cytoplasmic membrane and internal cell structure are also targets for ultrasonic inactivation.

The impact of ultrasound treatment on cF extrusion activity (cFA) is shown in Table 2. Difference analysis showed that there were no significant differences in glucose-energized cF extrusion activity with different initial bacterial concentrations. cFA values decreased gradually as the processing time increased, indicating that metabolic performance was inhibited by ultrasound at different levels even when cells still retained high enzymatic activity and sustained insignificant membrane damage. Although esterase activity did not differ much between E. coli and S. aureus after processing for 8 min (Fig. 2), pressures from ultrasound treatment still resulted in increasing perturbation of cF extrusion activity (Table 2). Besides, we noticed that there was no evident decline in the percentage of cF-stained S. aureus cells from 8 to 12 min (Fig. 2), while a significant decline in metabolic performance was observed. Somehow, the inhibition of the metabolic performance of viable cells by ultrasound treatment seemed to be related to culturability (Table 1). According to a published study (31), cells that almost entirely lost the ability to extrude dye were not able to form colonies on media and thus might be considered dead, although they still possessed enzymatic activity.

TABLE 2.

Glucose-energized cF extrusion activities of E. coli and S. aureus, at different initial concentrations, after treatment with ultrasound for 20 min

Bacterium and initial concn (log CFU/ml)	cFA (%)a at the following time (min):
	0	4	8	12	16	20
E. coli
6	98.09 ± 1.42	81.80 ± 3.69	58.43 ± 2.30	42.76 ± 3.75	20.61 ± 1.16	14.92 ± 2.27
8	96.77 ± 2.45	87.29 ± 2.78	60.71 ± 3.27	37.48 ± 3.88	20.40 ± 2.97	11.40 ± 2.55
S. aureus
6	98.32 ± 1.21	80.01 ± 1.57	76.08 ± 3.87	52.49 ± 2.96	36.78 ± 2.59	25.32 ± 3.45
8	95.59 ± 2.44	82.63 ± 1.72	71.74 ± 4.04	48.57 ± 3.64	34.98 ± 3.89	28.16 ± 2.65

^aValues are means of triplicate measurements ± standard deviations.

We conclude that the primary target of ultrasound with regard to its induction of lethal effects depends on the bacterial species. For a Gram-negative bacterium, the primary target appears to be the outer membrane, while in the case of a Gram-positive bacterium, it is likely to be the cytoplasmic membrane. Besides the commonly known cytoplasmic membrane damage, the present work discovered that alterations in esterase activity and metabolic performance were also involved in the process of inactivation of E. coli and S. aureus under ultrasound stress.

Comparison between viable plate counts and flow cytometry.

In this research, a comparison of VPC and FCM results was used to gain a better understanding of the effects of ultrasound on E. coli and S. aureus. We discovered that the number of bacteria determined by the conventional plate count method, which counted only culturable colonies on agar plates, was significantly lower than the number of viable bacteria determined by fluorescent staining methods. It will be necessary to consider the possible existence of VBNC forms, because bacteria are able to enter that state to counter detrimental environments (15). Further attention needs to be paid to this possibility, because the potential toxin-producing activity of VBNC bacteria can lead to food-borne diseases. Assessment of cell viability and functionality is much more critical than assessment of culturability by the viable plate count method.

Analysis of the sublethal status of cells caused by ultrasound.

FCM in combination with the double-staining method was used to determine whether any cells were in a sublethal state caused by ultrasound treatment. As shown in the dual-parameter dot plots (Fig. 3 and 4), the subpopulations were identified based on their differential staining characteristics with PI and cF: cF-stained cells (cF positive, PI negative) (lower right quadrant) had high esterase activity and intact membranes; PI-stained cells (cF negative, PI positive) (upper left quadrant) had inactivated esterase and damaged membranes; PI and cF double-stained cells (cF positive, PI positive) (upper right quadrant) had sublethal injury with residual esterase activity and compromised membranes; the unstained area (cF negative, PI negative) (lower left quadrant) most likely corresponded to debris or lysed cells attributable to ultrasound treatment. The results showed dynamic changes in the four quadrants as the processing time increased. It turned out that the fluctuation caused by the initial bacterial concentration was insignificant during the course of exposure to stress. The cF-negative, PI-positive population (dead cells) made up 0.02, 1.53, 2.86, 12.36, 20.12, and 25.17% of E. coli cells and 0.04, 3.24, 7.94, 14.09, 20.69, and 20.97% of S. aureus cells at an initial concentration of 10⁶ CFU/ml after exposure to ultrasound for 0, 4, 8, 12, 16, and 20 min, respectively. Under those circumstances, the cF-positive, PI-negative population (viable cells) of S. aureus showed a gradual decrease from 98.59 to 37.92%, but for E. coli, this population increased from 14.49% to 53.08% within the first 8 min, after which it decreased gradually from 53.08% to 7.95% after 20 min of exposure (Fig. 3). With initial E. coli and S. aureus concentrations of 10⁸ CFU/ml, the percentages of dead cells were 0.02, 1.56, 3.01, 11.12, 22.58, and 23.59% and 0.00, 4.46, 8.70, 14.88, 20.43, and 21.87% after exposure for 0, 4, 8, 12, 16, and 20 min, respectively. Meanwhile, the percentage of cF-positive, PI-negative (viable) cells dropped from 99.98 to 38.39% for S. aureus after 20 min of ultrasound treatment, and for E. coli, the percentages of this population were 12.51, 42.63, 50.27, 32.16, 18.90, and 5.64% after exposure for 0, 4, 8, 12, 16, and 20 min, respectively (Fig. 4). The cF-positive, PI-positive category, which was described as comprising sublethal cells, is so small that it can be neglected. Since almost no sublethal cells were generated, this indicated that intracellular esterase inactivation and membrane damage did occur simultaneously. The effect we observed was in agreement with the results obtained for single staining with cFDA or PI, proving that cells either were entirely killed or survived intact, while no sublethally damaged cells were generated. Based on the results, we concluded that ultrasound treatment might be an all-or-nothing process. The discovery of this phenomenon would have great significance for the application of ultrasound in the food industry. A series of studies (31–33) have found that some other nonthermal food-processing methods inactivate bacteria via sublethal injury. Sublethal cells can recover if they encounter appropriate environmental conditions (temperature, moisture, pH, and nutrients), posing a risk to public health. A recent publication (39) also reported that the advantage of ultrasound over other nonthermal techniques is that cells that have been ruptured and disintegrated by ultrasound cannot be revived. Thus, multiparameter FCM combined with the double-staining method proved to be a powerful technique for real-time, quantitative detection of the sublethal status of E. coli and S. aureus and contributed to revealing the sterilization mechanism of ultrasound.

Morphological changes revealed by transmission electron microscopy.

We used TEM to obtain distinct images in order to investigate the morphological changes in bacteria after ultrasound treatment. Micrographs of untreated E. coli cells show that they retained the rod shape with normal cell walls and nuclei (Fig. 5a). However, remarkable morphological changes were observed in E. coli cells after sonication for 20 min (Fig. 5b to d). E. coli cells (Fig. 5b) showed a rough cell wall and a blurry cell membrane, with a decomposing inner structure. Some cells (Fig. 5c) were disrupted, and the cytoplasmic material was released to the extracellular medium. More strikingly, cell wall fragmentation and plasma membrane rupture were also observed (Fig. 5d). On cross section, untreated S. aureus cells showed homogeneous electron density in the cytoplasm and presented continuous and smooth cell membranes and walls (Fig. 6a). In contrast, S. aureus cells treated with ultrasound for 20 min appeared to have decreased electron density in the cytoplasm (Fig. 6b). Other structural changes that we observed included lysed cells with broken walls and membranes, huge vacuoles, and heterogeneity in electron density (Fig. 6c and d).

FIG 5.

TEM photographs of E. coli cells. (a) Untreated bacteria. (b, c, and d) Bacteria treated with ultrasound for 20 min.

FIG 6.

TEM photographs of S. aureus cells. (a) Untreated bacteria. (b, c, and d) Bacteria treated with ultrasound for 20 min.

Although it is hard to compare the efficiencies of ultrasound at sterilizing E. coli and S. aureus through TEM images, we observed that the cells were damaged in different ways. TEM demonstrated that the destructive action of ultrasound had multiple targets. Moreover, the micrographs indicated that the ultrastructural changes caused by ultrasound treatment might be serious enough to be irreversible. Even if the cells could hold their outer structure intact, the inner structure of cells seemed to be strikingly affected.

Conclusion.

The present study showed that FCM analysis in combination with TEM could be used to improve our understanding of the effects of ultrasound on E. coli and S. aureus. We found that impairment of the cell membrane, inactivation of enzymatic activity, and inhibition of metabolic performance were involved in the process of sterilization of E. coli and S. aureus, while the initial concentrations of E. coli and S. aureus had no significant relationship with the extent of ultrasound-induced damage under our experimental conditions. Although no considerable level of germicidal efficacy against E. coli and S. aureus was reached, the destructive mode of action of ultrasound had multiple target sites, including the outer membrane, the cell wall, the cytoplasmic membrane, and the inner structure. In addition, the primary target for destructive action was not necessarily the cytoplasmic membrane. The primary target depends on the bacterial species; it is likely to be the outer membrane for a Gram-negative bacterium and the cytoplasmic membrane for a Gram-positive bacterium. Thus, for E. coli, the outer membrane is important. We also observed that FCM distinguished not only between live and dead cells but also between different physiological states of a stressed population. The analysis of sublethally injured E. coli and S. aureus cells by using the double-staining method suggested that ultrasound treatment might be an all-or-nothing process. The discovery of this phenomenon would have great significance for the application of ultrasound in the food industry.