Flow cytometry and growth‐based analysis of the effects of fruit sanitation on the physiology of Escherichia coli in orange juice

Abstract Chlorine‐based solutions are commonly used to sanitize orange fruits prior to juice extraction. We used flow cytometry (FCM) to investigate the physiology of Escherichia coli following its subjection to chlorine‐based solutions and alternative sanitizing agents (H2O2 and organic acids). Green fluorescent protein (GFP)‐generating E. coli K‐12 were washed with 50–200 ppm available chlorine (AC), 1%–5% H2O2, 2%–4% citric acid, 4% acetic acid, or 4% lactic acid, after which they were added to 1.2 μm‐filtered orange juice (OJ). Cell physiology was investigated with FCM during storage at 4°C, and culturability was determined using plate counting. Analysis of GFP fluorescence allowed estimation of intracellular pH (pH i). FCM results demonstrated an inverse relationship between the concentration of AC or H2O2 and cellular health in OJ. Higher concentrations of sanitizer also resulted in a significantly greater number of viable but nonculturable (VBNC) cells. Real‐time FCM showed that supplementation of AC with 2% citric acid, but not with 100 ppm of Tween‐80, led to a significant reduction in pH i of the cells incubated in OJ, and that the majority of the reduction in pH i occurred during the first 2 min of incubation in OJ. Organic acids were found to be more effective than both AC and H2O2 in reducing the pH i, viability, and culturability of the cells in OJ. The results confirmed the hypothesis that consecutive subjection of E. coli to maximum legally permitted concentrations of sanitizers and OJ induces the VBNC state. Furthermore, we demonstrate successful application of FCM for monitoring the efficacy of washing procedures.

of fruit surfaces is crucial for reducing the risks posed by this pathogen, especially as. E. coli O157:H7 has been shown to survive in OJ (Eblen et al., 2004).
Available chlorine (AC) is the most frequently used sanitizer in the food industry and is used for the purpose of washing the surface of various fruits, processing work surfaces, and decontaminating the washing water. AC concentrations of 50-200 ppm with a contact time of 1-2 min are commonly used for washing fresh produce including oranges (Parish et al., 2003;Suslow, 2000). Consequently, numerous studies have investigated the effectiveness of AC for eliminating E. coli from the surface of orange fruits (Bagci & Temiz, 2011;Martinez-Gonzales, Martinez-Chavez, Martinez-Cardenas, & Castillo, 2011;Pao, Davis, & Kelsey, 2000).
However, these studies have only measured elimination of bacteria using plate counts; the physiological state of the bacteria has not been determined using nongrowth-based methods. Considering that an AC concentration as low as 0.4 ppm can cause sublethal injury and induce a viable but nonculturable (VBNC) state in E. coli whereby bacteria are viable but do not grow on agar plates (Kolling & Matthews, 2001;Singh, Yeager, & McFeters, 1986), and measurement of viability using plate counts could grossly underestimate the number of viable bacteria. VBNC E. coli including AC-injured bacteria have also been previously shown to be capable of revival, growth, and pathogenicity in vivo when the stress conditions were removed (Palmer, Baya, Grimes, & Colwell, 1984;Pao, Davis, Kelsey, & Petracek, 1999).
In addition to AC, H 2 O 2 and organic acids such as citric, lactic, and acetic acids have also been suggested as suitable alternatives to AC for sanitation of fresh produce (Parish et al., 2003). However, H 2 O 2 , lactic acid, and acetic acid are also known to induce the VBNC state in E. coli (Li, Ahn, & Mustapha, 2005;Oliver, 2005). E. coli K-12 strain SSC1 used here constitutively expresses green fluorescent protein (GFP) (Miao, Ratnasingam, Pu, Desai, & Sze, 2009); It has previously been demonstrated that decreases in intracellular pH (pH i ) result in GFP deactivation (Kneen, Farinas, Li, & Verkman, 1998). The primary aim of this study was to use flow cytometry (FCM) to investigate the effects of washing E. coli K-12 SCC1 with the sanitizers AC, H 2 O 2 , and organic acids on the viability, physiological state, and pH i of the cells as well as their culturability as measured using plate counts, before and after their inoculation in OJ. E. coli K-12 has previously been used as a model for E. coli O157:H7 (Anvarian, Smith, & Overton, 2016;Valdramidis, Geeraerd, & Van Impe, 2007). We reveal the presence of VBNC bacteria in OJ following washing with both AC and H 2 O 2 , use real-time FCM to determine the effects of acid and surfactant on available chlorine washing, and demonstrate the effects of organic acids on cells subsequently incubated in OJ.

| Bacterial strain and preparation of the culture
Escherichia coli K-12 SCC1 (MG1655 P A1/04/03 -gfpmut3*) (Miao et al., 2009) was used in this study. Escherichia coli were taken from frozen glycerol stocks, grown for 24 hr on a nutrient agar plate at 37°C, then restreaked and grown for a further 24 hr at 37°C. A single colony of E. coli was inoculated in 20 ml of 2× LB (Lysogeny broth; 20 g/L tryptone, 10 g/L yeast extract (both Difco) and 10 g/L NaCl (Sigma, UK)) and allowed to grow in a shaking incubator (New Brunswick Scientific Innova 4000) for 18 hr at 37°C with agitation (150 rpm).
The overnight culture was subsequently diluted 1:1,000 in 50 ml of 2× LB medium and grown for another 24 hr under the same conditions to obtain late-stationary-phase culture.

| Orange juice
Orange juice was obtained from a local retailer and centrifuged at 17,696 g for 40 min (Beckman J2-21, Beckman Coulter) to remove pulp. The pulp-free supernatant was then filtered through a sterile filter paper of 1.2 μm pore size (Whatman, Maidstone, UK) to prevent the interference of OJ cloud particles with bacterial detection using FCM, as previously described (Anvarian, Smith, & Overton, 2018;Anvarian et al., 2016).

| Sanitizing solutions
Sodium hypochlorite (NaOCl; Sigma) containing 10%-15% AC (Sigma) was utilized for preparing chlorine-based sanitizers. The estimated midrange value of 12.5% AC was used for calculating the concentration of NaOCl needed for preparation of the sanitizers. The ACcontaining solutions (50, 100, 200, and 250 ppm) were prepared by diluting NaOCl in distilled water (dH 2 O) just before use. In order to prepare the acid-supplemented AC (ASAC) and surfactant-supplemented AC (SSAC) solutions, citric acid (BDH Ltd., UK) or Tween-80 (Sigma) was used as acidulant and surfactant, respectively. With regard to ASAC (200 ppm AC + 2%, i.e., 20,000 ppm, citric acid), the solution with 250 ppm AC was gently added to a 10% (w/v) citric acid solution in the ratio of 4:1 (v/v). For SSAC, 250 ppm AC solution was mixed with 500 ppm Tween-80 in the ratio of 4:1 (v/v)

| Washing procedure
Two ml of the culture was transferred to a 2.5 ml centrifuge tube (Sarstedt, UK). Cells were harvested by centrifugation (5 min at 16,873 g), the supernatant was disposed, and the pellet was dispersed in 50 μl of sterile Dulbecco's Phosphate-Buffered Saline (PBS). Next, 1 ml of sanitizer at 30°C prepared 10 min prior to the experiment was added to the cell suspension and gently mixed for 2 min by rotating the tubes in a test tube rotator (20 rpm). This temperature (30°C) was chosen in an attempt to mimic the recommended temperature for a sanitizer during the washing stage of the fruits with core temperature of around 22.5°C (room temperature). The positive temperature difference has been shown to prevent infiltration of pathogens into the fruit pulp (Zhuang, Beuchat, & Angulo, 1995). After washing, cells were harvested by centrifugation (16,873 g), washed with PBS, and dispersed in 1 ml of fresh PBS. Subsequently, 3 × 10 9 cells were added to 15 ml of filtered OJ in glass universal bottles. Samples were then incubated at 4°C for 13 days with no shaking.
F I G U R E 1 The effects of washing stationary-phase Escherichia coli K-12 SCC1 with different concentrations of available chlorine (AC) on viability, health, and culturability during incubation in orange juice (OJ) at 4°C. Stationary-phase E. coli were washed with (a) dH 2 O (0 ppm AC), (b) 50 ppm AC, (c) 100 ppm AC, or (d) 200 ppm AC for 2 min. Cells were washed with Phosphate-Buffered Saline (PBS) and 3 × 10 9 cells dispersed in 15 ml of OJ (filtered with 1.2 μm filter paper). Samples were then incubated at 4°C for 13 days. At each time point, samples were diluted in PBS, stained with PI, and analyzed by flow cytometry, and aerobic viable count was determined. The experiment was repeated twice each with a duplicate. The reported values are the mean values of duplicate samples for a representative experiment. "Washed" refers to cells postwashing with AC but before addition of OJ, while "+OJ" refers to time 0 hr postaddition of OJ   previously been described (Anvarian et al., 2016). Cells that retained GFP fluorescence and did not stain with PI (which stains dead cells red) were considered healthy with near neutral pH i (GFP + /PI − ) while GFP − /PI − cells were considered to be stressed but viable bacteria with a pH i of <5 (Anvarian et al., 2016;Kneen et al., 1998 Nonetheless, cells that remained culturable after the initial treatment with 200 ppm AC were more successful in retaining their culturability than those washed with 100 ppm AC, indicating a possible induction of resistance to components of OJ.

| Effect of hydrogen peroxide
Compared to AC, H 2 O 2 has been suggested as a more effective sanitizer against E. coli (Sapers, Miller, & Mattrazzo, 1999

| Real-time FCM-based investigation of the cellular physiology
During the previous experiments, it was noted that the significant drop in the GFP + population occurred very rapidly following addition of OJ, within the time taken to disperse bacteria in OJ (around 10 s).
Therefore, in order to study the effects of AC-based sanitizers on the

| Simultaneous use of GFP, BOX and PI
The physiology of GFP − /PI − cells was further investigated by the use of the lipophilic dye Bis-(1,3-Dibutylbarbituric Acid) Trimethine Oxonol (BOX) which enters cells with a collapsed membrane potential, thereby permitting identification of injured cells. As GFP and BOX both emit green fluorescence, exploratory experiments were performed. Figure S1 shows an example of the simultaneous use of GFP, BOX, and PI for determining the number of healthy, stressed,

| Effect of organic acids
As was shown above, washing the cells with ASAC had a significant effect on the subsequent stress response of the cells in OJ. However, these results raised the question of whether this effect was due to the combined antimicrobial effects of the citric acid and AC or simply due to lowering the pH of the alkaline AC solution. Therefore, it was decided to investigate the effects of 2% or 4% citric acid (2CA and The effects of washing Escherichia coli K-12 SCC1 with organic acid on its subsequent viability in orange juice (OJ) during 48hr incubation at 4°C compared to ASAC-washed cells. The method used for this experiment was similar to that described in Figure 1, except cells were washed with either (a) ASAC (pH 6.0), (b) 2% citric acid (2CA, pH 2.1), (c) 4% citric acid (4CA, pH 1.9), (d) 4% lactic acid (4LA, pH 2.1), or (e) 4% acetic acid (4AA, pH 2.5). Samples were incubated at 4°C for 2 days. "Washed" and "+OJ" refer to the before and after addition of OJ at time 0 hr. Error bars are the ±SD of the mean value obtained at each time point for each type of population (e.g., healthy GFP + , injured) and not the total percentage of the cells. (f) Culturable cells were determined by plate count pre-and postwashing and after addition of OJ (+OJ) and after 48 hr

| Effects of available chlorine
Numerous studies have investigated the effectiveness of AC-based sanitizers in reducing the microbial load on the surface of fruits; however, the results have been conflicting (Parish et al., 2003). In light of this, we decided to model washing using the methods in the present manuscript. It was difficult to compare the results presented here to those reported in the literature, primarily because of the differences in the experimental methods used, typically with whole fruits or other produce being washed, and differences in the temperature and pH of the sanitizing solutions as well as the duration of the treatment (Bagci & Temiz, 2011;Neo et al., 2013;Pao et al., 2000;Sapers et al., 1999). These parameters could significantly influence the antimicrobial efficacy of the AC. For instance, the reduction in pH from alkaline to neutral could change the equilibrium between OCl − and HOCl resulting in greater concentration of the latter which has been shown to exhibit stronger antimicrobial effects (Fukuzaki, 2006;McGlynn, 2004). In addition, unlike the current study in which the cells were suspended in PBS and washed with AC before their inoculation in OJ, in these studies cells were artificially inoculated on the surface of the produce before being immersed in or sprayed with AC-based solutions. Moreover, in some studies the terms AC, chlorine, hypochlorite, NaOCl, and bleach have been used interchangeably, making it difficult to determine the exact concentration of free chlorine used (Chung, Bang, & Drake, 2006;McGlynn, 2004;Narciso & Plotto, 2005).
Chung and coworkers reported that the increase in concentration of NaOCl from 50 to 200 ppm resulted in a greater decrease in the total microbial, coliform, and E. coli counts (Chung, Huang, Yu, Shen, & Chen, 2011). The observed level of decrease in the plate count of E. coli was within the range of 1-2 log reduction reported in the literature for antimicrobial efficacy of 200 ppm AC (Parish et al., 2003;Sapers, 2001). Washing the cells with 200 ppm AC also resulted in a significant increase in the number of VBNC cells. The induction of VBNC in AC-stressed cells has previously been reported (Dukan, Levi, & Touati, 1997;Kolling & Matthews, 2001;Oliver, Dagher, & Linden, 2005;Singh et al., 1986).

| Effects of hydrogen peroxide
Available chlorine within the permitted range (maximum 200 ppm) is generally not capable of reducing the microbial load on the surface of the fruit in excess of 2 log, and therefore, it does not ensure the safe elimination of food pathogens from the surface of fresh produce (Parish et al., 2003;Sapers, 2001). This is particularly the case in fresh produce such as orange fruit with porous and hydrophobic surfaces (e.g., due to the presence of wax and essential oils) which can potentially protect the bacteria against AC action by reducing the accessibility of the aqueous AC-based sanitizer to the bacteria (Adams et al., 1989;Beuchat & Ryu, 1997;Martinez-Gonzales et al., 2011). It has been suggested that washing solution containing up to 5% H 2 O 2 could be used as a suitable alternative to AC (Sapers et al., 1999(Sapers et al., , 2002 et al., 1999, 2002). Others found 1%-5% H 2 O 2 to be as effective as 200 ppm AC in reducing the E. coli population (Sapers, 2001;Sapers, Miller, Jantschke, & Mattrazzo, 2000). The discrepancy between our results and those reported in the literature is believed to be due to combined effects of using a different strain of E. coli, incubation of the stationary-phase cells at 4°C for 24 hr before H 2 O 2 treatment as well as using different temperature for the H 2 O 2 solution (50°C instead of 30°C used in the current study) by these researchers.  (Adams et al., 1989).

| Effects of acidified and surfactantsupplemented AC
Moreover, supplementation with Tween-80 decreased the microbial load by 34% compared to unsupplemented AC (Adams et al., 1989

| Effects of organic acids
Although both the dissociated and undissociated forms of organic acids can exert antimicrobial effects against a wide range of microorganisms, the undissociated form has been shown to be more effective. According to the "weak acid preservative" theory, upon entry of the undissociated organic acids into the cell they dissociate, leading to generation of excess protons, hence reducing the pH i of the cells. In order to prevent the adverse effects of low pH i on enzymatic activity and nucleic acids, cells actively employ ATP in order to extrude the excess proton and this eventually leads to cell death (Davidson & Harrison, 2003;Lu, Breidt, Perez-Diaz, & Osborne, 2011;Salmond, Kroll, & Booth, 1984). In order for the undissociated weak acid to lower the pH i and exert antimicrobial effects, it needs to be either small (containing fewer than 3 carbons) or lipophilic in order to be able to diffuse through the plasma membrane (Stratford & Eklund, 2003). Citric acid is not only a large weak acid (six carbons) but also hydrophilic (partition coefficient or P oct of log −0.172).
Lactic acid on the other hand is a small organic acid (3 carbons) which can pass through the plasma membrane. However, because of its high hydrophilicity (P oct log −0.62), its diffusion is very slow.
Therefore, both acids are believed to exert antimicrobial effects by acting as an acidulant reducing the pH of the environment. On the other hand, despite being a hydrophilic acid (P oct log −0.319), acetic acid can pass through the membrane in undissociated form due to its small size (3 carbons) and therefore reducing the pH i and causing intracellular damage (Stratford & Eklund, 2003). The diffusion of undissociated acetic acid and to some extent lactic acid through the membrane could explain the greater antimicrobial efficacy of these acids against E. coli compared to hydrophilic citric acid. Compared to AC, organic acids had significantly greater adverse effects on both the health and culturability of the cells. 4% acetic acid was found to be more effective than 4% citric or lactic acid in reducing the population of healthy (BOX − ) cells. Taken together, the data highlight the need to analyze bacterial viability and physiology using nongrowth-dependent methods in order to quantify VBNC cells.

This work was supported by a UK Biotechnology and Biological Sciences
Research Council PhD studentship to AA. The BD Accuri C6 flow cytometer was awarded to TWO by the BD Accuri Creativity Award.

CO N FLI C T O F I NTE R E S T
TWO and AA were paid speaker expenses by BD for speaking at BD Accuri users' events.

E TH I C A L S TATEM ENT
This study does not involve any human or animal testing.