Résumé
Staphylococcus aureus est l'une des bactéries pathogènes les plus fréquemment
retrouvées dans les produits laitiers. Dans un but de réduire les problèmes de sécurité alimentaire, les phages virulents sont étudiés comme agents antibactériens pour le contrôle des pathogènes d'origine alimentaire. L’objectif de cette étude était de sélectionner un ensemble de phages virulents et les utiliser dans un cocktail pour contrôler les staphylocoques dans le fromage. Les phages sélectionnés, appartenant aux trois familles de
Caudovirales (Myoviridae, Siphoviridae, Podoviridae), étaient virulents et ne possèdent
pas de gènes encodant des facteurs de virulence dans leur génome. De plus, ils ont un large spectre d'hôte et sont stables depuis leur production jusqu'à leur application de biocontrôle. Leur utilisation n'a pas déclenché la surproduction de l’entérotoxine C de S. aureus. Deux cocktails de phages anti-S. aureus, contenant chacun trois phages et un phage par famille, ont été confirmés comme efficaces suite à l'élimination de la population de S. aureus après 14 jours de maturation fromagère à 4 ºC. L'utilisation de cocktails de phages en rotation devrait empêcher l'émergence de souches bactériennes résistantes aux phages.
Avant-propos
Contribution des auteurs
Marie-Josée LeMay a participé aux résultats de l’encapsulation des phages en plus d’avoir testé la résistance du phage Team1, conservé pendant 3 mois à 4 ºC et à -20 ºC sous forme micro-encapsulé et libre. J’ai réalisé le reste des expériences. Claude P. Champagne et Steve Labrie ont supervisé les travaux. J’ai rédigé l’article au complet. Le professeur Sylvain Moineau a co-rédigé le manuscrit en plus d’avoir supervisé le projet.
90
Publication
Lynn El Haddad1, Marie-Josée LeMay2, Claude P. Champagne2, Steve Labrie3, and Sylvain Moineau1. Efficacy of two Staphylococcus aureus phage cocktails in a small-scale laboratory-based cheese production. In preparation for submission.
Abstract
Staphylococcus aureus is one of the most prevalent pathogenic bacteria found in dairy
products. In an effort to reduce food safety concerns, virulent phages are investigated as antibacterial agents to control foodborne pathogens. The aim of this study was to select a set of virulent phages and use them in a cocktail to control pathogenic staphylococci in cheese. Selected phages belonging to the three Caudovirales families (Myoviridae,
Siphoviridae, Podoviridae) were virulent and did not carry genes coding for virulence traits
in their genomes. In addition, they had a broad host range and were stable from the time of production to biocontrol application. Their use did not trigger over-production of S. aureus enterotoxin C. At MOIs levels between 15 and 150, two anti-S. aureus phage cocktails, each containing three phages, one from each of the three phage families, efficiently eradicated a 106 CFU/g S. aureus population after 14 days of cheese ripening at 4ºC. The use of phage cocktails in rotation should prevent the emergence of phage resistant bacterial strains. In the aim of commercial distribution of the phages to the cheese industry, assays on phage concentration and storage stability were also carried out. Microencapsulation in alginate gel beads enabled a 6 fold concentration of phages. Microencapsulation improved at least 100-fold the stability of phages during storage at 4ºC, and stability was better at -20 ºC than at 4ºC.
Introduction
Bacteriophages (or phages) are the most abundant and diversified biological entities on Earth. These bacterial viruses are found everywhere their bacterial hosts are present (Breitbart & Rohwer, 2005; Clokie et al., 2011; Sulakvelidze, 2011). Because of their ubiquitous nature, phages are regularly consumed when eating and drinking (Mahony et al.,
91 2011). Phages can also lyse specific bacterial strains without affecting the remaining microflora (García et al., 2010). Accordingly, some virulent phages have been successfully tested in different medical and food settings to prevent and control the proliferation of bacterial pathogens (Goodridge & Bisha, 2011; Sulakvelidze, 2013). These phage-based strategies improve the safety of processed foods, preserve ready-to-eat foods, and sanitize medical and factory equipment as well as environments.
However, not all phages can be used as efficient biocontrol agents. Several desirable properties are needed. First, the phage must be strictly virulent without containing a lysogeny module and the ability to integrate into the host genome. Second, the phage genome sequence must be known and be free of any known virulence genes. Third, the phage should have a broad host range, thereby infecting several strains of the target bacterial species. In addition, the efficacy of the phage should be maintained over long periods of storage or product shelf-life (Goodridge & Bisha, 2011; Hagens & Loessner, 2010). All of these features may not necessarily be found in a single phage, and phage cocktails (multiple phages) can be designed to compensate. Moreover, the use of virulent phage cocktails was shown to limit the emergence of phage resistant bacterial strains (Garcia et al., 2007; O'Flynn et al., 2004).
Some phage products have been approved and commercialized to prevent the growth of
Listeria monocytogenes (Leverentz et al., 2003; Soni & Nannapaneni, 2010), E. coli (Carter et al., 2012), and Salmonella (Carol, 2013) in foods. Milk products have been among the
preferred foods for testing the efficacy of virulent phages as biocontrol agents, likely because this industry is highly familiar with the destructive effect of virulent phages (Samson & Moineau, 2013). Several phages have been tested in dairy products to reduce the occurrence of the above foodborne pathogens (Guenther & Loessner, 2011; McLean et
al., 2013; Modi et al., 2001), as well as S. aureus (Bueno et al., 2012).
An anti-S. aureus cocktail containing two phages (phiIPLA35 and phiIPLA88) of the
Siphoviridae family (non-contractile tail and double-stranded DNA genome) was recently
found to synergically reduce the S. aureus concentration during fresh and hard-type cheese productions (Bueno et al., 2012; Garcia et al., 2007; Garcia et al., 2009). However, they did not completely eliminate the bacterial population in hard-type cheeses. Phage cocktails have also been designed to target biofilms and chronic infections caused by S. aureus
92
(Markoishvili et al., 2002; Merabishvili et al., 2009; Rhoads et al., 2009; Drilling et al., 2014).
Commercialization of phages for the cheese industry, presumably by specialized suppliers, would require that technologies for their concentration and storage be developed. Microencapsulation (ME) has been used for these purposes with lactic starters and probiotic cultures (Champagne, 2006) but no study has been done on the benefits of ME for phages of S. aureus.
In a previous study, we characterized and compared three very broad host range staphylococcal phages belonging to the Myoviridae family (contractile tail and dsDNA genome) and found one S. aureus strain resistant to all three polyvalent phages. This strain was, however, sensitive to phages of the Podoviridae family (short tail and dsDNA genome) (El Haddad et al., 2014).
Here, we tested the efficacy of two cocktails of three staphylococcal phages in reducing
S. aureus cells in cheeses. Phages were selected according to the different criteria cited
above and tested during a small-scale, laboratory-based Cheddar-like cheese production. Moreover, their stability in various production and storage conditions was determined as well as their safety using several means.
Materials and Methods
Bacterial strains and culture media
Fifty-seven strains of S. aureus were used to determine the host ranges of staphylococcal phages. These 57 strains were identified, genotyped and clustered into 14 groups according to their clonal complexes, sequence type numbers, and sources of isolation (El Haddad et al., 2014). Strains were obtained from various sources including the Félix d’Hérelle Reference Center for Bacterial Viruses (www.phage.ulaval.ca), the Ministère de l’Agriculture, des Pêcheries et de l’Alimentation du Québec, and the Canadian Bovine Mastitis Research Network. Strains were grown in Tryptic Soy Broth (TSB) medium at 37ºC. Baird-Parker medium supplemented with egg yolk-tellurite (BD™) was used for counting. The bacterial strains were maintained as frozen stocks in TSB broth containing 150 g/L glycerol. The presence of the five enterotoxin genes (sea, seb, sec, sed, and see) in these S. aureus strains was tested using the PCR primers developed elsewhere
93 (Hait et al., 2012). S. aureus SMQ1320, the strain added in the cheese assays (see below) was isolated from a mastitis infection in Ontario (Canada), while the starter culture used is a Lactococcus lactis subsp. cremoris strain CUC-222 (Cargill, Dairyland®) that was grown in 120 g/L sterile skimmed milk at 21ºC for 16 hours.
Bacteriophages
All phages used in this study are stored at the Félix d’Hérelle Reference Center for Bacterial Viruses. Eleven phages belonging to the three families of the Caudovirales order (tailed phages) were used (Table 4.1). The genome of the phages not previously available (El Haddad & Moineau, 2013; Vybiral et al., 2003) were sequenced using the Illumina Technology as described elsewhere (El Haddad et al., 2014). Tryptic-Soy Agar (TSA) and TSB soft agar (7.5 g/L agar) were used for counting phages. Phages used in cocktail during cheesemaking assays were concentrated with polyethylene glycol (PEG) and purified using two CsCl gradients (Sambrook & Russell, 2011). Purified phages were recovered by ultracentrifugation using a Beckman SW41 Ti rotor at 35,000 rpm (210,053 × g) for 3 h, followed by a second ultracentrifugation using a Beckman NVT65 rotor at 60,000 rpm (342,317 × g) for 18 h. The phage preparations were then dialyzed against phage buffer (0.05 M Tris-HCl [pH 7.5], 0.1 M NaCl, 8 mM MgSO4).
Table 4.1. Phages used in this study.
Name Classification Origin of isolation Publication
Pyophage Myoviridae, Twort-like Wound infection This study
Team1 Myoviridae, Twort-like Wound infection El Haddad et al., 2014
Team2 Myoviridae, Twort-like Wound infection This study
Phi812 Myoviridae, Twort-like Clinical infection Pantucek et al., 1998
K Myoviridae, Twort-like Unknown O’Flahery et al., 2005
P68 Podoviridae, 44AHJD-like Unknown Vybiral et al., 2003
44AHJD Podoviridae, 44AHJD-like Unknown Vybiral et al., 2003
LH1 Siphoviridae, B2 Raw milk El Haddad & Moineau, 2013
LH1-MUT Siphoviridae, B2 Raw milk, natural mutant of
LH1
El Haddad & Moineau, 2013
Phi2 Siphoviridae, Raw milk This study
MSA Siphoviridae, B2 Clinical infection, natural
94
Phage resistance to heat treatment
The Pyophage and Team2 phages were discarded due to redundancy whereas the toxin PVL-encoding phage LH1 was substituted with its mutant LH1-MUT (lacking lysogenic and virulence-related genes) (El Haddad & Moineau, 2013). Thus, eight of the 11 phages were tested for their resistance to an extended pasteurization treatment by exposing them to 73ºC for 20 seconds in milk as follows. Ultra-filtered pasteurized milk (2% M.F.) was heated in tubes at 73ºC. Each phage was then added at a concentration of 107 PFU per ml of heated milk and incubated for 20 seconds at 73ºC. After treatment, milk samples were put on ice and the phages enumerated. The titers of the nine phages were compared before and after the treatment to quantify the phage count reduction, expressed as log10 PFU/ml. Phage
titers were obtained on TSA overlaid with TSB soft agar containing 100 μL of the S. aureus host strain and 100 μL of the phage dilution.
Phage stability in a model Cheddar cheese curd
As a semi-solid medium with a low pH, cheese curd can limit the efficiency of phages. Two phages from each family were tested for their resistance in a Cheddar curd model under 39% humidity and 1.77% salt. First, a lyophilized Cheddar-like cheese curd (Lacroix
et al., 2010) was mixed with 1.77% salt and acidified water (pH 2.6 obtained with 85%
lactic acid) in order to obtain a pH of 5.2. Then, the phage being tested was added at a concentration of 107 PFU/ml and mixed manually with the model curd for 5 minutes. Ten grams of reconstituted curd were removed and initial (T=0) phage counts were determined. The remaining model curd was divided into 10-gram aliquots, vacuum-packed at 95%, and stored at 4ºC. These curds were sampled at 4, 7, 14, and 28 days. Each 10-gram curd sample was homogenized in 90 mL of sterile 2% sodium citrate solution pre-warmed to 45 ºC, using a Stomacher Lab-Blender (Seward Stomacher 400 Circulator). Dilutions ranging from 100 to 10-5 were made in 0.1% phage buffer. Phage titers were obtained on TSA overlaid with TSB soft agar containing 100 μL of S. aureus SMQ1320 and 100 μL of the phage dilution.
95
Host range of the phages
The host range of two podophages (P68, 44AHJD) and three siphophages (LH1-MUT, phi2, and MSA) were determined in this study. The host range of phages LH1, Team1, phi812, and K were previously reported (El Haddad et al., 2014). Phages were propagated on their respective host strains and tested for their polyvalent nature on different strains of
S. aureus. Strains of S. aureus were grown at 37°C to an OD600nm of 0.1 and approximately 105 phages were added to the medium. The cultures were incubated at 37ºC to complete bacterial lysis, and the resulting lysate was filtered using a 0.45 μm syringe filter. The host range, expressed by phage efficiency of plaquing (EOP), were determined by spotting 5 μL of the phage lysate serially diluted (10-1 to 10-8) in phage buffer on TSB soft agar containing 100 to 200 μL of a staphylococcal strain. At least, two biological and two technical repetitions were done per phage-host. The EOP values were calculated by dividing the titer of the phage on the tested strain by the titer of the phage on its host strain.
Use of phage cocktails to control S. aureus during a small-scale Cheddar-like production
Two phage cocktails (Team1/P68/LH1-MUT and phi812/44AHJD/phi2) were investigated separately in a small-scale laboratory-based Cheddar cheese production. For each trial, 200 ml of whole pasteurized ultra-filtered milk was heated to 32ºC. Starter culture (L. lactis) grown overnight was added to the milk at 1% (107 colony forming units per mL (CFU/mL)) along with 106 of S. aureus SMQ1320. Milk was supplemented with CaCl2 (0.4 g/L) and phage cocktails were added to a multiplicity of infection (MOI) of 15,
45, and 150. We thus tested single phage at MOIs of 5, 15, and 50 respectively. A maturation step took place at 32ºC until the pH dropped to 6.5. Then, 0.01% rennet (v/v) was added to the mixture and incubated at 32ºC for 50 minutes to allow coagulation to occur. When the pH dropped to 6.4, the coagulum was cut into cubes and rested for 10 minutes. The curd-in-whey suspension was then heated and stirred at 38ºC and the whey was drained. The resulting curd was vacuum-packed at 95% and ripened at 4ºC for 2 weeks.
Cheeses containing only the phage cocktail were also performed as a control for any potential changes in phage concentration in the absence of S. aureus. The second control condition included cheeses containing S. aureus but without phages to measure changes in
96
S. aureus population without viral addition. Finally, a third control condition contained
only the starter culture to monitor pH and temperature during the normal cheese process. Sampling took place during the initial addition of bacteria and phages, and after the steps of maturation, coagulation, curd formation, and ripening. Samples were homogenized in sterile 20 g/L sodium citrate solution pre-warmed to 45ºC, using a Stomacher Lab-Blender. Phage titers were obtained on TSA overlaid with TSB soft agar containing 100 μL of S.
aureus SMQ1320 and 100 μL of the sample dilution. Baird-Parker medium was used for
staphylococcal counts. Five biological independent trials as well as three technical replicates for each trial were performed.
Enterotoxin production tests
Staphylococcal enterotoxin C (SEC) detection was performed using the RIDASCREEN® SET Total Art. No. R4105 (r-biopharm, Darmstadt, Germany). It is based on an enzyme immunoassay using anti-enterotoxin monoclonal antibodies, with a minimum detectable limit of 0.25 ng/mL in liquid samples. The enterotoxin assay was performed according to the manufacturer’s instructions. S. aureus SMQ1320 grown to early stationary phase was added to i) pasteurized milk supplemented with the starter culture CUC-222 and incubated at 37ºC, ii) raw milk incubated at 37ºC, and iii) Cheddar cheese productions. These media were supplemented with phage cocktails, if applicable, and incubated for 2 hours to perform the ELISA assay. Testing for the presence of SEC during bacterial-phage challenges used 100 µL of the resulting filtrate or centrifuge product. The OD450 values for
each well were obtained using a Synergy 2 plate reader (BioTek Instruments, Inc.) and compared.
Phage encapsulation
An encapsulation procedure was selected as a method of phage concentration and conservation. The myophage Team1 was selected as a representative and was encapsulated in macro-beads (~2 mm) and micro-beads (~0.5 mm) and then frozen (-20 ºC) and freeze- dried. For both series of beads, 109 PFU/mL of Team1 was mixed with 20 g/L sodium alginate. To obtain the macro-beads, the mixture was then pumped (silicone tube #14) at 11 rpm at a rate of ~ 20 mL/min, in a solution of 1 mole/L calcium chloride at room
97 temperature. Formed beads were then dipped in a medium containing 100 g/L maltodextrin, 100 g/L glycerol, 2 g/L tryptone, and 4 g/L sodium ascorbate for 20 minutes. The liquid medium was discarded and the phage count was determined in 1 g of fresh beads. One gram of beads was frozen at -20ºC overnight and tested the next day for phage stability. The remaining beads were freeze-dried for 48 hours (24 hours programmed at 4ºC) and phage counts were compared.
The same procedure and parameters were used for the micro-beads except for the pump employed (Smaller diameter Cole Parmer #14 tube) which was operated at a rate of 6 mL/min into a Var-J1 Open coaxial air flow driven single nozzle unit (Nisco Engineering AG, Zurich, Switzerland). The co-axial air flow settings on Var-J1 unit were at 35 mm and 1.25 bar. Conservation of free phages was also tested as 109 PFU/mL of phage lysate was mixed with the solution containing maltodextrin and glycerol described above. One ml was used for initial phage counts, 1 mL was frozen at -80°C and tested, and the remaining volume was freeze-dried and compared to the other conditions and samples.
For phage titres, beads were first dissolved by homogenization in a sterile 20 g/L trisodium citrate solution, using a Stomacher Lab-Blender for 1 minute, followed by a 15 min incubation at 25°C. Phage counts of Team1 (log10 PFU/mL), were obtained on TSA
overlaid with TSB soft agar containing 300 μL of S. xylosus SMQ121 culture and 100 μL of the phage dilution. Phage stability as determined by plaques assays on samples stored during 24 hours, 1 week, 2 weeks, 1 month, 2 months, and 3 months. Two storage conditions were carried out: 4°c and -20°C. Three biological and three technical repetitions were carried out to obtain the means of the phage titers.
Statistical analyses
Statistical analysis was conducted using STATA version 12. We compared the conditions to each other as well as the change in staphylococcal and phage concentrations during cheese production. For each of the two cocktails, two separate 4 (conditions) × 5 (steps of cheddar cheese production) factorial designs were investigated with respect to (1) staphylococcal concentration (log10 CFU/ml) and (2) phage concentration (log10 PFU/ml),
using repeated-measures analysis of variance (ANOVA). For each relevant statistically significant finding, we reported the contrast term B with a p value and the 95% confidence
98
interval (95% CI). When appropriate, the ANOVA F coefficient was reported with its p value. A p value lower than 0.05 was considered statistically significant.
Results and Discussion
Phage resistance in cheese production
Eight staphylococcal phages were tested for their resistance to a heat treatment mimicking milk pasteurization (Table 4.2).
Table 4.2. Log10 reduction of phage concentration, added at 107 PFU/ml after heat
treatment at 73ºC for 20 seconds.
Most of the phages were sensitive to the heat treatment applied, with phage Team1 as the most sensitive with a 4 log10 reduction. However, two phages were less affected by the
conditions tested (Table 4.2). The titers of siphophages LH1-MUT and phi2 dropped by only 1.4 ±0.5 and 0.6 ±0.1 log10 units. The heat resistance phenotype was not family-
dependent since phages from all three families were sensitive (reduction of more than 3 log10) to the heat treatment. However, it seems related to the source of isolation of the
phages. LH1-MUT and phi2 were the only phages isolated from a dairy environment, and thus may be more adapted to proliferate and be stable in this environment (Brüssow et al., 2004; Koskella & Meaden, 2013). A similar result was previously obtained with S. aureus phages ɸH5 and ɸA72, two phages isolated from raw milk samples. When added to milk, these two phages resisted (98%-99%) to a heat treatment (72ºC for 15 seconds) (Garcia et
al., 2009b).
Two members of each of the three families of phages were also tested for stability during simulated curd ripening conditions (pH 5.2, 39% humidity, 17.7 g/kg salt, and vacuum-packing at 95% N2) at 4ºC.
Myoviridae Podoviridae Siphoviridae
Team1 phi812 K P68 44AHJD LH1-MUT phi2 MSA
99 Figure 4.1. Monitoring phage concentration in a model Cheddar cheese curd stored at 4ºC for 28 days. Error bars represent standard standard deviation.
A statistically significant change in concentration was noted with some of the phages. In particular, a significant 1 log10 unit drop was observed with phages P68, 44AHJD, and
LH1-MUT from day 0 to day 4. However, in all these cases, the phage titers subsequently remained relatively stable over the 28-day period. None of the phage counts dropped below 5 log10 PFU/ml after 28 days of ripening, showing that all phages were resistant to these