Tolerance to stress conditions associated with food safety in Campylobacter jejuni strains isolated from retail raw chicken

Campylobacter jejuni is a microaerophilic foodborne pathogen that is sensitive to stress conditions. However, it is not yet understood how this stress-sensitive pathogen may cause a significant number of cases of human gastroenteritis worldwide. In this study, we examined stress tolerance in 70 C. jejuni strains isolated from retail chicken under several stress conditions related to food safety. Compared to oxygen-sensitive (OS) strains of C. jejuni, C. jejuni strains with increased aerotolerance, such as hyper-aerotolerant (HAT) and aerotolerant (AT) strains, were more tolerant to peracetic acid, refrigeration and freeze-thaw stresses. However, the levels of thermotolerance and hyper-osmotolerance were not associated with the aerotolerance level of C. jejuni. The HAT and AT strains of C. jejuni exhibited significantly increased activities of catalase and superoxide dismutase (SOD), compared to the OS strains. Consistently, the HAT and AT strains were highly tolerant to oxidants, such as hydrogen peroxide, cumene hydroperoxide and menadione, compared to the OS strains. The AT and HAT strains that were tolerant to stresses, particularly peracetic acid and refrigeration, predominantly belonged to multilocus sequence typing (MLST) clonal complex (CC)-21. This study shows that oxidative stress resistance plays a role in determining the differential level of aerotolerance in C. jejuni and that AT and HAT strains of C. jejuni are more tolerant to oxidants and low temperatures than OS strains.


Results
Disinfectant resistance in C. jejuni isolates from retail chicken. Since chicken carcasses are usually treated with disinfectants during processing, we measured the tolerance of 70 C. jejuni strains to PAA, a common disinfectant used by the poultry industry. The PAA exposure significantly reduced the CFU levels of C. jejuni strains. More significant reduction (ca. 3.7 log CFU/g) was observed in the OS strains compared to the AT and HAT strains (ca. 2.1 and 2.5 log CFU/g, respectively) ( Fig. 1). Although most of the tested strains of C. jejuni exhibited a certain level of PAA tolerance, the results showed that AT and HAT strains of C. jejuni were more tolerant to PAA than OS strains.
Increased capability of ROS detoxification in AT and HAT strains of C. jejuni. Since oxidative stress defense plays an important role in aerotolerance by detoxifying ROS 23,24 and AT and HAT strains of C. jejuni were highly resistant to PAA (Fig. 1), an organic peroxide, we hypothesized that AT and HAT strains of C. jejuni might be more capable of detoxifying ROS than OS strains. To examine this hypothesis, we measured the activities of superoxide dismutase (SOD) and catalase, two important oxidative stress defense enzymes detoxifying the superoxide anion and H 2 O 2 , respectively. Interestingly, the SOD activity was significantly higher in the AT and HAT strains than the OS strains (Fig. 2a,b). The catalase activity was determined by measuring the intracellular level of H 2 O 2 in C. jejuni; thus, a lower H 2 O 2 level indicates a higher catalase activity. The AT and HAT strains of C. jejuni accumulated less H 2 O 2 than the OS strains (Fig. 2c,d), suggesting that the catalase activity is higher in the AT and HAT strains than the OS strains. These findings clearly demonstrated that AT and HAT strains of C. jejuni are more resistant to oxidative stress than OS strains; this may contribute to the enhanced PAA tolerance of AT and HAT strains. Tolerance to peracetic acid (PAA) in 70 C. jejuni isolates from retail chicken. Solid black bar indicates the mean CFU. The experiment was repeated three times and produced similar results. Two-way ANOVA was used to compare the results in the different aerotolerance groups. ***P < 0.005, ****P < 0.0001.
The 70 strains of C. jejuni were exposed to three different oxidants, including H 2 O 2 , cumene hydroperoxide (CHP; an organic peroxide), and menadione (MND; a superoxide generator). Despite variations depending on the strain and the oxidant (Fig. 3a,c,e), the viability reduction was significant in the OS strains compared to the AT and HAT strains (Fig. 3b,d,f). The results of the oxidative stress defense enzyme assays (Fig. 2) and the susceptibility tests (Fig. 3) consistently showed that HAT and AT strains of C. jejuni were highly tolerant to oxidative stress. tolerance to refrigeration, freeze-thaw, and heat in C. jejuni strains from retail chicken. After packaging at processing plants, chicken carcasses are refrigerated or frozen for distribution and preservation. Thus, refrigeration and freezing are harsh stress conditions that C. jejuni encounters in the chicken supply system. In addition, C. jejuni will be exposed to high temperatures during cooking. The OS strains exhibited a significant CFU reduction at a refrigeration temperature. Storing at 4 °C for seven days reduced bacterial counts by approximately 4.9, 5.1, and 7.4 CFU/ml in the HAT, AT, and OS strains of C. jejuni, respectively (Fig. 4a). Similarly, the OS strains of C. jejuni were more sensitive to freeze-thaw stress that the AT and HAT strains. More than half (ca. 52%) of HAT and AT strains survived at −20 °C for seven days; however, only two OS strains were detected after seven days (Fig. 4b).
Unlike the results of the refrigeration and freeze-thaw tolerance tests, the aerotolerance level was not associated with thermotolerance as only one HAT strain and one AT strain survived after exposure to 70 °C for 30 sec (Fig. 4c). Based on the findings, C. jejuni strains from retail chicken exhibited different levels of tolerance to cold and heat stresses, and AT and HAT C. jejuni were highly tolerant to refrigeration and freeze-thaw stresses, but not to heat stress. tolerance to hyperosmotic stress. Hyperosmotic stress is a stress condition for Campylobacter in food preservation and cooking, such as marination. When exposed to different concentrations of NaCl (1%, 2%, and 4%), all of the tested strains except for one AT strain survived at 1% NaCl, and approximately 36% of HAT, 48% of AT, and 40% of OS strains survived at 2% NaCl with a wide range of CFU levels (Fig. 5a). Whereas five AT and HAT strains survived at 4% NaCl, none of the OS strains survived; however, the difference was not statistically significant (Fig. 5a). Based on the results of fluorescence microscopic analysis, C. jejuni exhibited heterogeneous morphology with mixed populations of helical rod, elongated, and coccoid cells depending on the strain (Fig. 5b).
MLSt sequence types of stress-tolerant C. jejuni strains. The MLST sequence types of the 70 strains of C. jejuni from raw chicken have been reported in our previous study 20 . CC-21 was predominant in AT and HAT strains that were tolerant to stress conditions, particularly PAA and refrigeration (Fig. 6a,b). Among the OS strains tolerant to PAA, CCs-21 and 45 were dominant with a statistical significance (Fig. 6c). Commonly, CC-21 is the primary MLST sequence type in stress-tolerant C. jejuni strains from raw chicken.

Discussion
Most foodborne pathogenic bacteria of public health concern usually originate from animals 25 . Since poultry is the major reservoir for Campylobacter, poultry carcasses are likely to be contaminated by Campylobacter during processing, particularly defeathering and evisceration 26 . To maintain food quality by reducing spoilage and pathogenic bacteria on poultry carcasses, various intervention methods are used by the poultry industry, including low storage temperature, marination, modified atmospheric gas packaging, and antimicrobial disinfectants 27 . The results show the means and standard deviations of a single experiment with triplicate samples, and the experiment was repeated three times. Two-way ANOVA was used for statistical analysis. ns: not significant, *P < 0.05, ****P < 0.0001.
PAA is a disinfectant widely used to decontaminate poultry carcasses in process water for washing, rinsing, and chilling due to its strong antimicrobial efficacy 28,29 . Interestingly, AT and HAT strains of C. jejuni exhibited enhanced tolerance to PAA compared to OS strains (Fig. 1). This may be attributed to the increased oxidative stress defense in AT and HAT strains (Fig. 2) as PAA is a mixture of H 2 O 2 and acetic acid. Whereas most other bacteria harbor redundant copies of genes encoding ROS-detoxification enzymes, such as KatA, SodB and AhpC, C. jejuni possesses only a single gene copy encoding the enzymes. Although AhpC is the major enzyme contributing to aerotolerance in C. jejuni 23 , the other two enzymes (i.e., KatA and SodB) also affect the viability of C. jejuni under aerobic conditions 24 . Presumably, the increased aerotolerance would be associated with the augmented activities of oxidtative stress defense enzymes, which also may increase the capability of decomposing PAA. This is also supported by the augmented survivality of AT and HAT strains after exposure to various types of oxidants (Fig. 3).
Interestingly, AT and HAT strains of C. jejuni were more tolerant to refrigeration and freezing temperatures compared to OS strains (Fig. 4a,b). Most bacterial species, such as E. coli, Salmonella, and Bacillus, produce cold shock proteins upon a temperature downshift [30][31][32] . However, C. jejuni does not possess genes encoding cold shock proteins 15 , which suggests that C. jejuni may have other tolerance mechanisms to respond to cold shocks. Studies thus far have shown oxidative stress defense, particularly SodB, plays an important role in the cold stress tolerance of Campylobacter. A knockout mutation of sodB increases the sensitivity of C. jejuni to both superoxide and peroxide stress 33,34 . The level of sodB expression in C. jejuni increases by exposure to cold-shock 35 , and a sodB mutation makes Campylobacter more susceptible to freeze-thaw stress than the wild type 36,37 . However, the survival of a sodB mutant is comparable to that of the wild type in the absence of oxygen 36 , indicating that oxidative stress impacts C. jejuni's ability to survive under freeze-thaw conditions. Notably, the SOD activities were determined to be significantly higher in AT and HAT strains than OS strains (Fig. 2a,b), and the AT and HAT strains were more tolerant to MND, a superoxide generator (Fig. 3c,d). Presumably, the elevated levels of SOD activity may possibly contribute to the enhanced survival of AT and HAT strains under refrigeration and freeze-thaw stress conditions. The association of cold stress with oxidative stress defense has also been reported in some other bacteria. For instance, exposure of Pseudomonas fluorescens MTCC 667, an isolate from Antarctic soil, to low temperature (4 °C) increases the production of ROS and elevates the activity of SOD 38 . Although molecular mechanisms still remain to be explained, our findings and the studies done by others consistently www.nature.com/scientificreports www.nature.com/scientificreports/ suggest oxidative stress defense may affect C. jejuni tolerance to cold and freezing stresses. C. jejuni strains from retail chicken were relatively sensitive to heat stress (Fig. 3c), compared to human clinical strains of C. jejuni in our previous study 22 . The reason for the different levels of thermotolerance between chicken isolates and clinical isolates of C. jejuni remains unknown.
C. jejuni is sensitive to hyperosmotic stress 39 and is easily inactivated at >2% NaCl 40 . Thus, high (1.5% ~ 3%) salt concentrations in marinated poultry meat 41 would be another stress to C. jejuni during foodborne www.nature.com/scientificreports www.nature.com/scientificreports/ transmission. NaCl is a general food preservative and inhibits the growth of foodborne pathogens in foods 42 . Cameron et al. reported that exposure to 1% NaCl modestly upregulates oxidative stress genes, such as katA and sodB, in C. jejuni, suggesting oxidative stress defense may affect osmotic stress response 43 . However, in the current study, the level of hyper-osmotolerance was not related to aerotolerance, although the results showed the level of hyper-osmotolerance is highly variable depending on the strain (Fig. 5). In C. jejuni, capsular polysaccharides (CPS) are involved in hyper-osmotolerance as mutations in the genes involved in genes encoding CPS, such as kpsM, kpsS, and kpsC, significantly (ca. 100-fold) increase C. jejuni susceptibility to 1% NaCl 43 . To maintain the intracellular turgor pressure properly under hyper-osmotic stress, bacteria generally accumulate solutes by increasing the uptake of K + and synthesizing osmolytes, such as trehalose and glutamate 44 . Although molecular mechanisms for osmotolerance have not yet been elucidated in C. jejuni, it has been reported that highly-frequent spontaneous variations in housekeeping genes related to purine biosynthesis, such as purF and apt, is associated with C. jejuni response to hyperosmotic stress 45 .
This study demonstrated that C. jejuni strains isolated from retail raw chicken were tolerant to several different stress conditions that may affect the survival of C. jejuni in chicken and that aerotolerance is significantly related to C. jejuni tolerance to PAA and refrigeration and freezing temperatures. Further comparative genomics analysis as follow-up studies will help us elucidate the molecular mechanisms underlying stress tolerance in C. jejuni.

Methods
Bacterial strains and culture conditions. Seventy strains of C. jejuni were isolated from retail chicken meat in our previous study 29 . The strains were routinely grown on Muller-Hinton (MH) media at 42 °C under microaerobic condition (5% O 2 , 10% CO 2 , 85% N 2 ).
Stress tolerance testing of C. jejuni. Stress tolerance testing was performed as described in our previous study with slight modifications 22 . C. jejuni strains were grown on MH agar at 42 °C for overnight under microaerobic condition. The strains were harvested in MH broth and resuspended in fresh MH broth to an optical density at 600 nm (OD600) of 0.1 prior to testing.
(1) Tolerance to PAA: Raw chicken skin (0.3 g) was prepared with a sterilized razor, and each piece of chicken skin was inoculated C. jejuni suspension (approximately 10 8 CFU). The chicken skin spiked with C. jejuni suspension was incubated at 4 °C for 1 h under microaerobic conditions and dipped in 750 ppm PAA solution for 15 sec and immediately washed in ultra-pure water. The chicken skin was transferred into a 15 ml tube with 2 ml fresh MH broth and vortexed for 2 min. Bacterial count was determined with a serial dilution and cultivation on Preston Campylobacter-selective agar. The experiment was repeated at least three times. (2) Tolerance to refrigeration and freeze-thaw: C. jejuni suspensions were placed into a 96-well plate and incubated at 4 °C for 3 and 7 days for refrigeration stress and also incubated at −20 °C for 3 and 7 days for freeze-thaw stress. After 3 and 7 days, the incubated C. jejuni suspensions were serially diluted and spread on MH agar for enumeration.     www.nature.com/scientificreports www.nature.com/scientificreports/ Determination of susceptibility to oxidants. Overnight cultures of C. jejuni were harvested from MH agar and diluted with fresh MH broth to an optical density 600 of 0.1. The diluted C. jejuni suspensions were exposed to oxidants, including 2 mM of hydrogen peroxide (H 2 O 2 ), 0.2 mM menadione (MND), and 0.1 mM cumene hydroperoxide (CHP), for 1 h. After washing with fresh MH broth twice, the suspension was serially diluted and spread on MH agar. fluorescence microscope analysis. The morphological changes observed with a fluorescence microscope with SYTO9 and propidium iodine staining. C. jejuni strains were inoculated on MH agar at 42 °C for overnight and harvested with MH broth. The bacterial suspension was serially diluted with fresh MH broth to an OD 600 of 0.07. The C. jejuni suspension was supplemented with sodium chloride to final concentrations of 1, 2 and 4% and incubated at 42 °C for 5 h with shaking (200 rpm) under microaerobic conditions. After washing with 1X PBS buffer twice, C. jejuni was fixed with 4% paraformaldehyde for 1 h at room temperature. The cells were centrifuged at 8,000 × g for 5 min and washed with 1X PBS buffer twice. After staining with SYTO9 and propidium iodine for 20 min at room temperature and washing twice with 1X PBS buffer, C. jejuni was observed with a fluorescence microscope (Carl Zeiss, Germany).

Statistical analysis.
Statistical analysis of the data from stress tolerance tests and catalase and SOD assays was performed with two-way ANOVA (GraphPad Prism Ver. 7, GraphPad Software, USA). Chi-square distribution was performed by SPSS ver. 21 (SPSS Inc., IBM, USA).