Retail liver juices enhance the survivability of Campylobacter jejuni and Campylobacter coli at low temperatures

The high prevalence of Campylobacter spp. in retail liver products was previously reported and has been linked to several outbreaks of campylobacteriosis. The main objective of this study was to investigate the influence of retail liver juices on the survivability of several strains of C. jejuni and C. coli, which were previously isolated from various retail meats at 4 °C. All tested Campylobacter strains showed higher survival in beef liver juice (BLJ) and chicken liver juice (CLJ) as compared to beef and chicken juices (BJ and CJ) or Mueller Hinton broth (MHB) at 4 °C. Overall, C. jejuni strains showed greater survival in retail liver and meat juices as compared to C. coli. CLJ enhanced biofilm formation of most C. coli strains and supported growth in favorable conditions. When diluted, retail liver and meat juices enhanced survival of Campylobacter strains at low temperatures and increased aerotolerance. In conclusion, beef and chicken liver juices enhanced the survival of C. jejuni and C. coli strains at low temperatures, which helps explain the high prevalence of Campylobacter spp. in retail liver products.

www.nature.com/scientificreports www.nature.com/scientificreports/ The survival of Campylobacter strains in diluted retail meat and liver juices (5% v/v in MHB) was investigated at 4 °C. With the exception of MHCJ, all diluted retail meat and liver juices significantly enhanced Campylobacter survival relative to MHB (P < 0.0001) (Fig. 2). Campylobacter strains failed to grow in MHCJ after 5 days. The addition of 5% laked horse blood and defibrinated horse blood to MHB (MHBB and MHFB, respectively) also enhanced Campylobacter survival, with most strains showing higher survival in MHBB vs. MHFB (Fig. 2). For C. jejuni strains, survival was highest in MHBLJ, MHBB and MHCLJ than unamended MHB (P < 0.0001). Survival rates in MHBB, MHBLJ, and MHCLJ were comparable to 100% BJ for several C. jejuni strains (Figs 1a, 2a). Among C. coli strains, MHBLJ promoted survival throughout the 35-day experiment. HC2-48 was the only C. coli strain producing visible colonies on MHA plates at 11 days after inoculation in MHCLJ, MHBJ, MHBB, and MHFB.

survival of C. jejuni and C. coli in various food matrices at 4 °C. A comparative analysis of viability
among Campylobacter strains was conducted using log bacterial cell count reduction rates. A higher reduction in bacterial counts (log N/No) was generally observed for C. coli strains as compared to C. jejuni in variable food matrices. Survival patterns were prolonged for all Campylobacter strains incubated in CLJ, BLJ, and MHBLJ (Fig. 3). Interestingly, the survival rate of C. coli HC2-48 was significantly higher than other C. coli strains in CJ, www.nature.com/scientificreports www.nature.com/scientificreports/ BJ, MHBJ, MHCLJ, MHBB, and MHFB but comparatively lower than most of the C. jejuni strains. No strains survived more than five days in MHB and MHCJ. Two tailed ANOVA for species level differences in individual juices showed significant species level differences of survival rate in retail meat juices (CJ and BJ) and all diluted juices (except MHB and MHCJ) (P < 0.0001). However, MANOVA for interaction including all experimental results for the interaction of Campylobacter species, media and time showed significant strains level differences (P < 0.0001) but not species-level differences (P = 0.38) (see Supplementary Table 1). survival at −20 °C. Reference strain C. jejuni NCTC11168 was used to investigate the influence of retail meat juices on the survival of Campylobacter at freezing temperature (−20 °C). We included C. jejuni NCTC11168 www.nature.com/scientificreports www.nature.com/scientificreports/ to compare results with a previous report of survival assay in CJ at freezing temperature (−18 °C) 38 . A rapid decrease in viable cell counts was observed within the first week of incubation (~3.7 logs for MHB, ~2.76 logs for CJ, ~2.17 logs for BLJ, ~1.7 logs for CLJ and ~1.67 logs for BJ); cell counts then remained relatively constant until the end of experiment in most juice matrices (Fig. 4). Diluted retail juices (MHBLJ, MHCLJ, MHBJ, and MHCJ) and horse blood (MHBB, MHFB) increased survival of C. jejuni NCTC11168 relative to the unamended MHB media at −20 °C; however, no significant differences were observed in NCTC11168 survival rates among different retail meat and liver juices.
Growth and survival at 37 °C. The influence of retail meat and liver juices on growth and survival of Campylobacter strains (Table 1) at favorable growth temperatures was evaluated at 37 °C in microaerobic conditions. As expected, MHB supported higher growth and survival of all strains up to the end of the experiment (Fig. 5). CJ was the best matrix for growth and survival among juices. Chicken liver juice (CLJ) was more  www.nature.com/scientificreports www.nature.com/scientificreports/ favorable for growth at 37 °C than BLJ and BJ for all C. coli strains and C. jejuni 11168 and CG1-109. None of the strains survived after four days in BLJ. C. jejuni NCTC11168 and CG1-109 and all C. coli strains failed to survive more than two days in BJ; however, C. jejuni T1-21, OD2-67, and WP2-202 survived until the end of the experiment. Interestingly, cell counts of C. jejuni T1-21 and WP2-202 decreased at 48 h and then increased, which might be caused by metabolic adaptation to the available nutrients after the initial incubation period. The lower bacterial counts in CLJ, BLJ, and BJ indicated that these matrices were less suitable food sources for growth than CJ in favorable environmental conditions (e.g. 37 °C). MANOVA showed significant species-level differences (P = 0.0035). Differences in growth on the basis of strain, juice, and origin of juice were also significant (P < 0.0001) (see Supplementary Table 2).
Influence on biofilm formation. On polystyrene surfaces, incubation in CLJ significantly enhanced biofilm formation all C. coli strains except YV1-223 (Fig. 6b). Biofilm formation was higher for C. jejuni T1-21 and C. coli HC2-48, CF2-75, YV1-223 and ZV1-224 in CJ, whereas other strains did not produce significant biofilms in CJ (Fig. 6a,b). All the juices failed to promote biofilm formation in C. jejuni OD2-67, CG1-109, and WP2-202. Biofilm formation on borosilicate glass was highest when C. jejuni NCTC11168 was incubated in CJ, whereas the other juices did not promote biofilm formation (Fig. 6c). Statistical model analysis showed significant effects of strain-media (P < 0.0001) and species-media interactions on biofilm formation (P < 0.0001) (see Supplementary  Table 3). www.nature.com/scientificreports www.nature.com/scientificreports/ Influence on aerotolerance. In comparison to non-amended MHB, C. jejuni NCTC11168 and OD2-67 and C. coli WA3-33 and CF2-75 showed enhanced aerotolerance in MHB amended with 10% beef and chicken liver juice (MHBLJ, MHCLJ) at 24 h of aerobic incubation (Fig. 7). The addition of laked horse blood to MHB (MHBB) enhanced survival in all strains exposed to aerobic conditions but bacterial counts were lower than MHBLJ, MHCLJ and MHBJ; this indicates the presence of an additional factor in retail liver and meat juices that enhances aerotolerance. In general, the addition of BJ to MHB also enhanced aerotolerance, but CJ and non-amended MHB did not.

Discussion
Campylobacter strains routinely encounter low temperatures, and a few studies have speculated that cold tolerance and the acquisition of cryoprotectant molecules are survival mechanisms 43,55 . The prolonged survival of Campylobacter in cold storage conditions has been reported in media and food models [38][39][40][41]48 , which are influenced by the food matrix composition. Retail meat juices from chicken 38,39,45 , pork 46 and beef 56 have been used as models to represent the surface of retail meat products. Analogous to meat juice models, retail liver juice models (BLJ and CLJ) represent the retail liver environment that foodborne pathogens would encounter after contamination. Retail liver possesses choline and other nutrients which are beneficial for human health [18][19][20] . Heme and nonheme proteins and other nutrients and minerals are also highly abundant in retail meat and liver products 57,58 . The elevated survival of Campylobacter in retail liver juices in this study indicates that liver juices provide a nutritious and favorable environment for acquisition of cryoprotectant molecules and nutrients for metabolism 59 during survival at low temperatures.
During harsh environmental conditions, Campylobacter spp. often enter the VBNC state 9 and undergo a morphological change to coccoid forms 59 . The reported high prevalence in retail liver products [30][31][32] and high survival rate at lower temperatures (this study) indicates that the food matrix provided by retail liver products helps retain culturable conditions of Campylobacter cells for prolonged periods. Our experimental design represents culturable cell counts of Campylobacter on MHA plates after incubation in various media at reduced temperatures. One previous report showed a rise in the C. jejuni population in inoculated chicken livers at 4 °C for a 24-h incubation, but a reduction in bacterial numbers was observed in chicken skin medallions and chicken meat 60 . Higher viability (culturability) of C. coli strains in retail liver juices vs. retail meat juices ( Fig. 1b) correlated with the higher prevalence of reported C. coli in retail liver products as compared to retail meat products 30 . Similarly, the high viability of C. jejuni in retail liver and meat juices with comparable survival rates (Fig. 1a) might explain the reported high prevalence of C. jejuni in retail meats and liver products 30,32,50 . The decline in bacterial counts in our study when using retail liver juices is higher than previous studies, which found no significant reduction in bacterial counts in artificially-inoculated liver samples (liver slices and liver homogenate food models) at 4 °C 48,49,61 . Discrepancies in experimental design (food models used) and differences in the Campylobacter strains used in our study might explain the reduced survivability in our study as compared to previous reports.
During the processing and handling of retail meat and liver products, dilutions of substances naturally occur in the working environment and on the surface of retail meat and liver products. In previous studies, media supplemented with blood was compared with full-strength retail meat juices in survival assays 39 however, the effect of dilutions on survival of Campylobacter at lower temperatures has not been previously studied. In this study, 5% dilutions of juices, laked blood and fresh horse blood in MHB were compared to detect the influence of the blood component and additional factor(s) in juices on the survival of Campylobacter strains at low temperatures. Enhanced survival of Campylobacter in MHB supplemented with 5% laked or fresh horse blood (MHBB or MHFB) (Fig. 2) demonstrates the influence of blood components on survival. The higher survival of selected strains in MHBB vs. MHFB indicates that the lysed blood environment improves survival. The prolonged storage of retail meat and liver products at lower temperatures promotes lysis due to the freeze-thaw process, which creates a more favorable environment for contaminant Campylobacter cells. In our study, the higher survival in MHBLJ than MHBB or MHFB for most Campylobacter strains (Fig. 2) indicates that BLJ contains other essential nutrients in addition to blood components. For C. jejuni strains, all diluents except MHCJ enhanced survival (Fig. 2a); however, only MHBLJ improved the survival of all C. coli strains (Fig. 2b). MHBJ enhanced the higher survival of some C. jejuni strains more than 100% BJ, which might be due to a more balanced www.nature.com/scientificreports www.nature.com/scientificreports/ nutritional composition in MHBJ vs. full-strength BJ. None of the Campylobacter strains survived well in MHCJ at low temperatures (Fig. 2), which shows the potential loss of the protective or favorable environment observed with full-strength CJ (Fig. 1). These results indicate different nutritional or environmental requirements among Campylobacter spp. for survival at low temperatures.
Significantly higher survival of C. jejuni than C. coli in CJ, BJ, and diluted juices (p < 0.0001) (Fig. 3) supports our contention that the two Campylobacter spp. have different nutritional requirements. Our results suggest that C. coli might require more nutritional support than C. jejuni strains. Hence, highly nutrient rich conditions (like BLJ and CLJ) might have supported similar survival rate for both species which differed for other juices with different nutritional contents. A similar inference was made in a previous report showing the higher survival of C. jejuni than C. coli in water samples maintained at 4 °C and 20 °C 53 . MANOVA for interaction of species, strains, media and time by including all results for survival at 4 °C, also showed significant strain level differences (P < 0.0001). C. jejuni strains were previously shown to be more acid tolerant than C. coli 62 . Among C. jejuni strains, used clinical strains survived better than poultry strains at low temperatures (4 °C and 10 °C) in a previous report 63 . Likewise, a waterborne C. jejuni strain showed better survival in defined fresh water media at 4 °C than a foodborne strain 51 . In contrast, a few reports have shown similar survival rates for both species 40,41 and higher survival rate of C. coli at low temperatures in different food models 64 .
At freezing temperatures (−20 °C), all retail liver and meat juices and diluents enhanced survival of C. jejuni NCTC11168 relative to MHB (Fig. 4). Hence, retail liver juices likely function as a protective food matrix composition for Campylobacter spp. at subzero temperatures. Similar inference for the presence of protective materials in CJ for Campylobacter at freezing temperature had been proposed in a previous report 38 . The rapid decrease of bacterial numbers after inoculation into juices maintained at −20 °C (Fig. 4) is similar to previous studies where a rapid decline in bacterial numbers was observed as early as 0.5 h after inoculation 49,54,65 .
At a favorable growth temperature (37 °C), CJ promoted growth at 37 °C for Campylobacter strains (Fig. 5a,b), which agrees with previous reports documenting the favorable nutrient composition of CJ for enhanced growth 39 . Chicken liver juice also supported growth and higher survival of all Campylobacter strains in comparison to BLJ and BJ (Fig. 5a,b). Although BLJ enhanced survival of Campylobacter strains at low temperatures (Figs 1, 3, 4), it did not support growth at 37 °C. Hence, BLJ might provide cryoprotectant molecules and required nutrients for survival at low temperatures, but these factors are not conducive for growth or potentially toxic at favorable temperatures.
To our knowledge, our study is the first to use retail liver juices to investigate the influence of retail liver environments on biofilm formation and aerotolerance, which was not feasible with previously used food models (liver slices and liver homogenates) 48,49,61 . Chicken juice induced high levels of biofilm formation for C. jejuni strains NCTC11168 (borosilicate glass surface) and T1-21 (polystyrene) and C. coli strains HC2-48, CF2-75, YV1-223 and ZV1-224 (polystyrene) (Fig. 6a-c). In a previous report, full strength as well as 5% dilution of CJ was shown to enhance the attachment of Campylobacter strains to abiotic surfaces and biofilm formation 45 . It has been found that CJ environment enhances biofilm formation of both motile and non-motile variants of Campylobacter. Among tested juices, remarkably high biofilm formation was seen for most C. coli strains in CLJ (Fig. 6b), but C. jejuni biofilms were not significantly different in CLJ vs. MHB (Fig. 6a,c). In general, we observed higher biofilm formation among C. coli than C. jejuni strains in retail juices, which agrees with a previous study 45 . Significant strain-dependent differences in biofilm formation were observed in our study for the various retail liver and meat juices (P < 0.0001). Extracellular DNA (eDNA) has been shown to be a major component in biofilm formation of Campylobacter, where DNase and eDNase (from Campylobacter strains) treatment could rapidly remove or inhibit Campylobacter biofilms 66,67 . DNA components available in retail meat and liver juices after lysis of blood, meat and liver cells might enhance biofilm formation of foodborne pathogens like Campylobacter. Feng et al. previously reported higher biofilm formation of Campylobacter spp. in polymicrobial environments than in monomicrobial conditions 68 . Hence, other bacterial contaminants found in retail liver products 25,69 could contribute to biofilm formation by Campylobacter in the retail liver environment.
Although Campylobacter spp. are microaerophilic, aerotolerant strains show enhanced survival in aerobic conditions 47 . The oxidative stress response is also associated with the mechanistic basis of iron acquisition 70 . Heme-containing proteins in retail meat and liver products 57,58 function as cofactors for important enzymes in the oxidative stress response, including catalase and superoxide dismutase 70 . Iron content in media influences the aerotolerance mechanism of Campylobacter regulated by regulatory proteins (PerR and Fur) and genes like ferroxoxin (fdxA) and alkyl hydroperoxide reductase (ahpC) [71][72][73][74][75] . Hence, the aerotolerance observed for the four Campylobacter strains in our study (Fig. 7) might be related to the iron and nutrient content found in diluted (10%) retail meat and liver juices. Although diluted CJ did not enhance aerotolerance; other diluted retail meat and liver juices enhanced survival in aerobic conditions, possibly because of their higher nutrient composition or iron content. It is also important to note that overly high iron levels can promote the formation of toxic superoxide radicals that may be detrimental to Campylobacter metabolism 70 . We mention this because it could explain the survival data shown in Fig. 5, where full-strength juices resulted in reduced bacterial counts relative to MHB at 37 °C. The iron content of retail meat and liver juices might also play a significant role in survival at lower temperatures, since the oxidative stress response is activated during cold shock 43 .
Functional metabolic activities and genomic expression data have been reported in Campylobacter at lower temperatures (4 or 5 °C) in various growth conditions [35][36][37] . Similarly, differential expression of genes related to quorum sensing and glycosylation of flagellin have been reported in CJ when compared to artificial Brain heart infusion media 36 . A variety of genes were essential for survival at low temperatures in nutrient-rich or nutrient-poor media 55 . All transcriptomic and genome fit analyses of Campylobacter at low temperatures have been conducted with C. jejuni 9,36,43,51,55 . Genomic differences between C. jejuni and C. coli strains 52 might contribute to differences in survival. Further investigations are needed to validate the effect of genomic differences on survival at low temperatures. (2019) 9:2733 | https://doi.org/10.1038/s41598-018-35820-7 www.nature.com/scientificreports www.nature.com/scientificreports/ In conclusion, our results show that retail liver juices enhanced the survival of all Campylobacter strains at low temperatures, whereas other retail meat juices and dilutions had differential effects on survival. This is a highly relevant finding with respect to food safety since retail liver juices represent an environment encountered by foodborne Campylobacter after contamination. Overall, C. jejuni strains showed greater survival at 4 °C in chicken juice, beef juice, and diluted retail meat and liver juices as compared to C. coli. Chicken liver juice enhanced biofilm formation of most C. coli strains and supported growth in favorable growth conditions. Further investigations are needed to explore the mechanisms by which the retail liver environment is enhancing the survival of Campylobacter at 4 °C.

Methodology
Bacterial strain and growth conditions. Campylobacter isolates (four C. jejuni and six C. coli strains) were used in this study (Table 1) and were previously isolated from retail meat and liver products 30,32,50 . C. jejuni NCTC11168 (clinical isolate) was used as a reference strain. The eleven strains were subcultured from −70 °C stock cultures and grown on Mueller Hinton Agar (MHA) supplemented with 5% laked horse blood at 42 °C for 48 h in microaerobic conditions (6% O 2 , 13% CO 2 , 81% N 2 , Thermo Forma incubator, model 3130). Prior to harvesting bacterial cells, strains were further subcultured for 18 h on a fresh plate of MHA with 5% laked horse blood for survival, biofilm and aerotolerance assays. Bacterial cells were harvested in phosphate buffered saline (PBS) (pH 7.4), and cell suspensions were adjusted to OD 600 = 0.1. In general, bacterial inoculum was prepared similarly for each assay, including survival and growth at variable temperatures, biofilm formation, and aerotolerance.
preparation of retail meat and liver juices. For food models, retail meat and liver juices were prepared.
Chicken juice (CJ) was prepared as described previously 38,39 . Briefly, frozen retail whole chickens without giblets were purchased from various retail meat shops and thawed overnight at room temperature. A similar procedure was used to obtain beef liver juice (BLJ) and chicken liver juice (CLJ) from frozen beef liver slices and chicken livers, respectively. Beef juice (BJ) was collected from retail meat shops after opening packets containing big chunks of beef cuts. Juices were collected aseptically in sterile containers and stored at −20 °C prior to further processing. After thawing overnight at 4 °C, CJ was centrifuged at 10,000 rpm for 15 min, whereas other meat and liver juices (CLJ, BLJ, and BJ) were centrifuged at 15,000 rpm for 30 min to exclude larger particles. Juices were filter-sterilized with a 0.45 µm membrane filter (Nalgene Rapid-Flow) and stored at −20 °C. The absence of any microbial contaminants in filtered retail meat and liver juices was confirmed by culturing (in aerobic, microaerobic and anaerobic conditions at 25 °C, 37 °C and 42 °C) in MHA supplemented with 5% laked horse blood. Microaerobic and anaerobic incubation at variable temperatures were done in gas jars containing microaerobic gas generating kits and anaerobic gas generating kits (Mitsubishi Gas Chemical, New York, NY, USA) respectively. Dilutions of meat and liver juices (5% or 10% v/v) were also prepared with MHB for survival studies at low temperature and aerotolerance assays. Similarly, dilutions of laked horse blood and fresh horse blood (5%) in MHB were also included to study whether the blood in retail meat and liver products influenced the survival of Campylobacter at low temperatures.
Survival at 4 °C. Appropriate volumes of cell suspensions were added to pre-incubated juices and dilutions to create bacterial concentrations of approximately 7 logs CFU/ml. Strains and log CFU/ml were as follows: C. jejuni NCTC11168, 6.63; T1-21, 6.6; OD2-67, 6.98; CG1-109, 6.73; WP2-202, 6.17; and C. coli WA3-33, 7.02; HC2-48, 7.28; CF2-75, 7.17; CO2-160, 7.47; YV1-223, 7.14; and ZV1-224, 6.85. MHB was used as a reference medium. An equal volume of inoculated media and juices were filled to the rim of 5 ml disposable polystyrene test tubes with caps and incubated at 4 °C to ensure microaerobic conditions. At specified time intervals, 10 µl samples were taken and serially diluted with 0.1% peptone in saline solution. Two spots of the 10 µl sample from each dilution were spotted onto MHA and incubated at 42 °C for at least 48 h. Viable cell counts were taken, and data analysis was performed with mean values of triplicate experiments.
survival at −20 °C. C. jejuni NCTC11168 (7.6 log CFU/ml) was used as inoculum in this study for investigating the influence of retail meat and liver juices at freezing temperature. 500 µl of inoculated meat and liver juices (100%), a 5% dilution of meat and liver juices, and MHB were dispensed into 2 ml Eppendorf tubes and maintained at −20 °C. At allocated times, triplicate samples were thawed at room temperature for 10 min and serial dilutions in 0.1% peptone saline were plated. Viable cell counts were taken as described previously.
Survival at 37 °C. Inoculated juices and MHB (1.5 ml) were dispensed into a 96-well storage plate, (Square Well, 2.2 ml) and incubated at 37 °C in microaerobic conditions. 40 µl samples were removed at specified time intervals and serial dilutions were prepared as described previously. MHA plates spotted with serial dilutions were incubated at 37 °C in microaerobic conditions for at least 48 h before viable cell counts were determined.
Biofilm formation. Suspensions of all Campylobacter strains were prepared and adjusted to OD 600 = 0.1 in PBS from an 18-h culture. Biofilm assays were prepared as described previously 76 . Briefly, the cell suspension was diluted to 1::10 in meat and liver juice, and MHB was used as a reference media. Biofilm formation on glass surfaces was investigated by incubating 1 ml of C. jejuni NTC11168-inoculated media and juices at 37 °C in 10 ml borosilicate glass tubes. After a 72 h-incubation, bacterial cell suspensions were removed, and wells were washed twice with 1.2 ml of sterile PBS (pH 7.4); plates were agitated gently to dislodge unbound cells. MHB (1.2 ml) supplemented with Triphenyl Tetrazolium Chloride (TTC) (0.05% w/v) was added to each well and incubated for 72 h at 37 °C. The other Campylobacter strains (excluding C. jejuni NCTC11168) were used to evaluate biofilm formation on polystyrene surfaces. Samples (150 µl) were incubated in sterile polystyrene 96-well microtiter plates at 37 °C in microaerobic conditions. After a 72-h incubation, cell suspensions were removed, and wells were www.nature.com/scientificreports www.nature.com/scientificreports/ washed twice with 180 µl sterile PBS (pH 7.4) and agitated gently to remove unbound cells. MHB (180 µl) supplemented with TTC was added to each well and incubated for 72 h at 37 °C. The remaining MHB/TTC solution was then removed, and wells were air-dried. Bound TTC dye was dissolved using a solution containing acetone (20%) and ethanol (80%); absorbance was measured at 492 nm with an Appliskan Multimode Microplate Reader (Thermo Scientific). All experiments were conducted in triplicate and repeated two or more times.

Influence of liver and meat juices on aerotolerance. Aerotolerance assays were conducted using four
Campylobacter strains (C. jejuni NCTC11168, C. jejuni OD2-67, C. coli WA3-33, and C. coli CF2-75) as described previously 47 with minor modifications. Similar approach has been used for aerotolerance assay in previous reports with incubation temperature at 37 °C or 42 °C 71,72,74,77 . Sterilized retail meat and liver juices were mixed with MHB to prepare 10% meat and liver juices. We used 10% dilutions because full-strength (100%) retail meat and liver juices coagulated more quickly, thus hindering assays of viable cell counts. Bacteria were removed from 18-h subcultures on MHA supplemented with 5% laked horse blood. Bacterial suspensions were then diluted to OD 600 = 0.2 in PBS (pH 7.4). Bacterial suspensions (1 ml) were added to 9 ml of preincubated, diluted juices and then incubated aerobically at 42 °C, with agitation at 200 rpm (New Brunswick I2400 Incubator Shaker). 50 µl samples were removed at 0, 12 and 24 h, and viable cell counts were evaluated on MHA as described previously. All experiments were performed in triplicate. The initial data analysis was performed with respect to the origin of the Campylobacter strains (beef, chicken and pork) and the type of retail liver juice added to the growth media (BLJ, CLJ or none). Repeated measures MANOVA was also done with Origin effect, Juice effect, Time effect, Origin × Juice interaction, Origin × Time interaction, Juice × Time interaction, and Origin × Juice × Time interaction. We then examined whether liver juice specifically was significant [retail liver juice (regardless of origin or none)] by replacing juice origin (beef and chicken) in the analysis. For survival and growth at 37 °C, statistical analysis mirrored that of the survival experiment at 4 °C.
For biofilm assays, the study was conducted with mean values of absorbance (OD 492 ). An ANOVA with Strain effect, Growth-Media effect, and Strain × Growth-Media interaction was performed. Student's t-test and two-tailed ANOVA were conducted for each assay of survival and biofilm as needed.