English (pdf)
Article in xml format
Article references
Automatic translation
Send this article by e-mail
Cited by SciELO 
Cited by Google 
Similars in
SciELO 
Similars in Google 
Permalink
Introduction
Several food products have been associated with Salmonella enterica serovar Enteritidis (SE) outbreaks worldwide (EFSA and ECDC, 2015). Eggs and egg derivatives are the second foodstuff most frequently involved in human salmonellosis. A total of 54 and 112 outbreaks linked to consumption of egg or egg derivatives, particularly homemade mayonnaise, have been reported between 2002 and 2014 in the UK and USA, respectively (Chousalkar et al., 2017). Thus, several control methods have been implemented in the egg production process, but they have not been able to reduce outbreaks. A promising alternative tool to reduce Salmonella contamination in food production and processing is the use of lytic bacteriophages.
Bacteriophages (or phages) are viruses that infect and lyse prokaryotic cells, are target- specific, self-replicating, with rapid bactericidal activity, and do not alter the original properties of food (Jorquera et al., 2015a). Therefore, they could be useful as biocontrol agents in foodstuffs. Bacteriophages have been isolated from various foods, including vegetables, seafood, dairy products, meat, and fish (Hudson et al., 2005). Bacteriophages are classified into lysogenic and lytic. Lysogenic bacteriophages are recommended neither as therapeutic agents nor as bio-controllers in food because they can carry genes encoding virulence factors.
Although phage-based biocontrol studies in a wide variety of foods (i.e., raw meats, ready- to-eat food, and fresh products) have been conducted (Bigwood et al., 2008; Guenther et al., 2012), there are limited studies in eggs and none in mayonnaise. Guenther et al. (2012) showed that application of a bacteriophage reduced by 2.6 log10 S. Typhimurium counts in pasteurized egg yolk after 48 hours of incubation at 15 ºC. Spricigo et al. (2013) registered a reduction of S. Typhimurium and S. Enteritidis concentrations (0.9 log10 CFU/cm2 each) in fresh eggs after being sprayed bacteriophages and incubated at 25 °C for 2 hours.
Our previous experiments demonstrated the SE-reducing effectiveness of a cocktail of five lytic bacteriophages in chicken breast, goat cheese (Jorquera et al., 2015b), salmon (Galarce et al., 2014), and sausages (Galarce et al., 2016). However, to the best of our knowledge, there is no information on the use of bacteriophages as biocontrol agents in raw-egg derivative products. Therefore, the present study aimed to preliminary determine the effectiveness of a phage cocktail to reduce SE counts in a homemade mayonnaise-like matrix maintained at 21 °C, simulating a cold chain loss.
Materials and methods
Bacterial strain
A spontaneous mutant SE phage type 4 (PT 4) strain resistant to nalidixic acid and rifampicin (SE nal r rif r ), originally isolated from a laying hen in Chile, was used to inoculate the mayonnaise samples. This strain was grown in Luria-Bertani broth (Difco®, Franklin Lakes, NJ, USA) and incubated at 37 °C for 18 h. The SE nal r rif r culture was then adjusted to 0.6−0.8 OD625 (Spectroquant Pharo 300, Merck, Darmstadt, Germany) to achieve 108 CFU/ml.
This suspension was serially diluted in buffered peptone water (BPW, Difco®, Franklin Lakes, NJ, USA) to obtain the bacterial concentrations used to contaminate the samples (103 CFU/mL). These concentrations were confirmed by viable counts on xylose-lysine-deoxycholate (XLD, Difco®, Franklin Lakes, NJ, USA) agar plates, supplemented with rifampicin and nalidixic acid (50 μg/ml each, Sigma®, Franklin Lakes, NJ, USA), and incubated at 37 °C for 24 h.
Bacteriophages
Five Salmonella specific phages were selected from our collection. The selection was based on their lytic properties against the challenge bacterial strain and nine different SE strains, their stability during the required conditions (4 to 18 °C during 10 days), the host range against 13 serovars of Salmonella enterica (serotypes that usually infect poultry), and tolerance to pH (4.0 to 9.0) and temperature (4 to 45 °C). The bacteriophage- insensitive mutants (BIMs) isolation range was 1.3x10-6 to 2.9x10-6 CFU (Robeson et al., 2014; Galarce et al., 2016). Due to its morphological characteristics, all phages belonged to the Siphoviridae family and were originally isolated from sewage samples (fSE7, fSE8, and fSE12) and foodstuff (fSE1C and fSE4S) (Robeson et al. 2014). High titer lysates were individually prepared using a spontaneous mutant SE PT4 strain (VAL 222) resistant to both nalidixic acid and rifampicin as the host bacterial strain. The original wild type SE PT4 strain was provided by Dr. Roy Curtiss III (The Biodesign Institute, Arizona State University). Phages were suspended equitably in modified SM buffer (50 mM Tris- HCl, 8 mM MgSO4 heptahydrated, pH 7.5) to reach the required concentration (108 PFU/mL each). A multiplicity of infection (MOI) of 105 was used. The phage titer was determined by plating adequate dilutions onto lawns of the target strain (Robeson et al., 2008).
Food matrix
Fresh eggs were acquired from a local supermarket to prepare homemade mayonnaise. Before breaking the eggs, the eggshells were disinfected with 70% v/v alcohol and air-dried. According to the traditional recipe (yolk, albumen, vegetable oil, and salt), mayonnaise was prepared inside of a biosafety cabinet. Only Salmonella spp. negative mayonnaise samples by culture (ISO 6579 2002) and by genus-specific PCR analysis targeting invA gene (Malorny et al., 2003) were included in this study.
Trial
Two groups (control and experimental) of 48 mayonnaise samples at 25 ml per sample were individualized in sterile Whirl-Pack bags (VWR, Lutterworth, United Kingdom). Samples were experimentally contaminated with SE nal r rif r (103 CFU/ml) in a biological safety cabinet (Heal Force safe 1200®, Shanghai, China) by homogenization with an inoculum volume corresponding to 50% of the sample to improve homogenization of bacteria and phages in the samples. The contaminated samples were kept at room temperature for 2 h to allow bacterial adaptation. In the experimental group, 2.5 ml of the phage cocktail was added to each contaminated sample. Samples of the control group were contaminated with SE nal r rif r and added with 2.5 ml of modified SM buffer. Both groups, control and experimental, were incubated at 21 ºC, and SE enumeration was performed at 24 h of incubation in 24 samples, and then at 72 h in the remaining 24 samples. Experimental and control groups were kept and processed separately, and the bacteria enumerated on the same days as the experimental samples.
Five mayonnaise samples not inoculated with SE or phages were kept in the laboratory throughout the experiment to confirm the absence of intra-laboratory contamination.
After incubation (24 and 72 h), and for bacterial count, 225 ml of BPW were added to each bag and homogenized (Stomacher 400 circulator, Seward LTD®, Worthing, UK) for 1 min. Dilutions with BPW were made and 100 µl of each dilution was plated on XLD agar supplemented with rifampicin and nalidixic acid and incubated at 37 ± 1 °C for 24−48 h. Bacterial counts were performed in duplicate. Samples without evident bacterial growth were subjected to enrichment following ISO 6579:2002, and results recorded as positive or negative.
For bacteriophage mixture detection from treated samples, the aqueous phase of homogenized samples was recovered and treated with chloroform to inactivate bacterial cells, which were discarded by centrifugation. Serial dilutions were carried out in SM buffer, followed by duplicate plating using the soft-agar overlay technique, with the challenge SE nal r rif r strain as an indicator. Thus, 100 μL of each dilution and 100 μL of the challenge strain were mixed in a tube with 4 mL of molten soft LB agar (Difco®, Franklin Lakes, NJ, USA). The suspension was poured onto solid LB agar plates and incubated at 37 °C for 18-24 h before plaque detection (Galarce et al., 2016). Detected bacteriophages were not individually identified nor quantified.
Statistical analysis
Mean bacterial counts (CFU/ml) were expressed in logarithmic units (log10). ANOVA was used to compare mean differences between phage-treated and untreated samples at p ≤ 0.05 significance level. Samples negative to bacterial count but positive to enrichment were assumed as 103 CFU/ml, corresponding to the minimum detection limit. Data were analyzed using InfoStat® 2008 software.
Results
The application of bacteriophage cocktail in the food matrix significantly reduced bacterial counts, ranging from 2.14 to 3.75 log10 CFU/ml, at 24 and 72 h, respectively. Table 1 shows the bacterial reductions and mean SE counts observed.
Lytic activity was stable throughout the incubation time and phages were detected in all treated samples (data not shown). No phages were isolated from any of the control samples. Mayonnaise samples that were not inoculated with SE and were not treated with phages also showed negative results to SE nal r rif r growth and the presence of phages.
Table 1 Mean counts and reductions of Salmonella Enteritidis (SE) in a homemade mayonnaise-like matrix with and without phage treatment, incubated for 24 and 72 hours at 21 ºC.
| Incubationtime (h) | Group | SE initialdose(log10 CFU/ml) | Mean SEcounts(log10 CFU/ml) ± SD | SE reduction(log10 CFU/ml) |
| 24 | ExperimentalControl | 2.62a | 3.23a ± 0.3 | 2.14 |
| 5.37b ± 0.2 | ||||
| 72 | ExperimentalControl | 2.98a | 5.49a ± 0.4 | 3.75 |
| 9.24b ± 0.2 |
Different superscript letters (a, b) within columns indicate significant differences (p≤0.0001).
Discussion
In the control samples, SE increased up to 2.75 log10 CFU/ml at 24 h, and up to 6.26 log CFU/ml at 72 h, showing the effects that cold losses can produce in the bacterial load present in foods. According to our previous results, the lytic bacteriophage cocktail application significantly reduced SE counts in 2.14 log10 CFU/ml at 24 and 3.75 log10 CFU/ml 72 h at the incubation temperature. These reductions can significantly enhance food safety, considering that the infective dose for humans for various serovars, including S. Enteritidis, is 105 CFU, and that high concentration doses are associated with higher rates of illness and shorter incubation periods (Kothary and Babu, 2001). This is the first study assessing the bio-controlling properties of a bacteriophage cocktail in a matrix similar to homemade mayonnaise, demonstrating its preliminary lytic activity in an egg-derived raw food preparation.
Similar to our results, Guenther et al. (2012) observed a reduction of 2.6 log10 CFU/g in pasteurized egg yolk contaminated with S Typhimurium (103 CFU/g) and treated with FO1-E2 phage (3 x 108 PFU/g) incubated at 15 °C for 2 days. Towards the fifth day, they observed increased bacterial counts, even exceeding the control group, explained by phage immobilization on food surfaces and matrix, possibly in combination with the appearance of phage-insensitive bacteria. Our previous findings suggest that the frequency of emergence of bacteriophage insensitive mutants to the phage cocktail was as low as 2.9 x 10-6 CFU (Galarce et al., 2016), hindering the loss of effectiveness due to phage-resistance. In the same context, Spricigo et al. (2013) reported reductions lower than 1 log10 CFU/cm2 in fresh eggs contaminated with S. Enteritidis (107 CFU/ ml) treated with a bacteriophage cocktail (1010 PFU/ml) and incubated at room temperature for 2 h. Compared to other matrices, this minor reduction was explained by a non-homogeneous Salmonella contamination of the eggshells or by the translocation of microorganisms from the shell surface to its external and internal membranes, which could have affected the phage effectiveness.
Mayonnaise is an edible oil emulsion with additives, forming a semi-liquid matrix. It is well-established that the matrix strongly influences the phage activity. In liquid matrices, phages can spread almost unrestricted and easily collide with the bacterium. Hence, bacterial reduction should be higher compared to solid matrices (Guenther et al., 2009). Guenther et al. (2012) indicated that, due to its viscosity, bacteriophages distribution in yolk is non-homogeneous and has limited diffusion. For this reason, in the present study an inoculum volume equivalent to 50% of the total sample was used to improve bacterial inoculum homogenization and phage distribution in the mayonnaise. Thus, higher bacterial reduction was expected in the mayonnaise-like matrix than observed in our previous studies in solid matrices, with reductions of 3.19 log in salmon fillets (Galarce et al., 2014) and 3.92 log in turkey breasts (Jorquera et al., 2015b). It is known that phage binding to ligand on the bacterial surface is influenced by intrinsic factors, such as ionic strength, pH, and substances that may interfere with the bactericidal process. Moreover, target bacteria may be embedded within this matrix and could be partially shielded from diffusing liquid, and therefore are also protected from the phage (Gunther et al., 2009; Jorquera et al., 2015a). Our results suggest that the mayonnaise- like matrix viscosity and its relatively complex composition of fat, proteins and carbohydrates may have interfered with bactericidal viral activity.
Matrix dilution (50%) decreased the viscosity and changed the characteristics of homemade mayonnaise, which prevents inferring the behavior of these phages; however, it suggests that it could be used in sauces containing mayonnaise. More studies are needed to determine the reducing effect of these phages on mayonnaise.
The lytic activity showed similar levels of bacterial reduction at 24 and 72 h of incubation. Although we have demonstrated the stability of these bacteriophages in 10 different food matrices, where they remain stable over 10 days of incubation at 18 and 4 °C (Robeson et al., 2014), their stability in mayonnaise or other egg- derived raw food has not yet been determined; thus, the influence of this factor on the current results should be considered in future studies.
Incubation time can also affect phage activity, with longer incubation time corresponding to higher bacterial reductions (Bigwood et al., 2008). As incubation time (24 and 72 h) did not influence bacterial reduction (p≤0.05), it was excluded from further analysis. Thus, a one-way ANOVA model was adjusted, using treatment (phage vs. no phage) as a factor. The coefficient of determination (R2) of the ANOVA ranged between 0.96−0.99, indicating that the adjusted model explained more than 95% of the bacterial count's total variability in each case and, consequently, by the treatment with bacteriophages.
In conclusion, our results suggest that under experimental conditions, this bacteriophage cocktail significantly reduces SE counts in a homemade mayonnaise-like matrix, regardless of incubation time. Therefore, bacteriophages could be used to control SE in egg derivatives. Further studies are needed to improve bacterial control in undiluted mayonnaise focusing on increased MOI, characterization of lytic activity of phages in whole egg and its components by separate, and determining their effectiveness in conjunction with conventional methods.
