- Email: [email protected]
Biocontrol of Staphylococcus aureus in curd manufacturing processes using bacteriophages
ARTICLE IN PRESS
International Dairy Journal 17 (2007) 1232–1239 www.elsevier.com/locate/idairyj
Biocontrol of Staphylococcus aureus in curd manufacturing processes using bacteriophages Pilar Garcı´ a, Carmen Madera, Beatriz Martı´ nez, Ana Rodrı´ guez Instituto de Productos La´cteos de Asturias (IPLA-CSIC). Apdo. 85, 33300 Villaviciosa, Asturias, Spain Received 25 October 2006; accepted 27 March 2007
Abstract The ability of specific bacteriophages to inhibit Staphylococcus aureus growth in curd manufacturing processes was determined. Two lytic bacteriophages specific against S. aureus were obtained by DNA random deletion from the milk-isolated temperate phages, FH5 and FA72. A cocktail of these lytic phages, F88 and F35, at multiplicity of infection (MOI) of 100, produced a complete elimination of 3 106 cfu mL1 of the pathogen in ultra-high-temperature (UHT) whole milk at 37 1C. Furthermore, the frequency of emergence of bacteriophage-insensitive mutants was reduced up to 200-fold in the presence of the two lytic phages compared with that detected with the combination of the temperate counterparts. The lytic phage derivatives, added to milk, were able to decrease rapidly the viable counts of S. aureus during curd manufacture. In acid curd, the pathogen was not detected after 4 h of incubation at 25 1C, whereas pathogen clearance was achieved within 1 h of incubation at 30 1C in renneted curd. These results indicate that lytic bacteriophages could be used as biopreservatives in the manufacture of particular dairy products. r 2007 Elsevier Ltd. All rights reserved. Keywords: Lytic phage; Biocontrol; Dairy products; Staphylococcus aureus
1. Introduction The current technologies employed to inactivate bacterial pathogens in foods are not always foolproof and, therefore, new approaches for improving food safety are necessary. Bacteriophages provide an attractive alternative since phages are ubiquitous in different environments, unable to infect human cells and, consequently, they have great potential for use as biocontrol agents in foods (Hudson, Billington, Carey-Smith, & Greening, 2005). The exploitation of bacteriophages has already become an interesting tool to fight the emergence of antibioticresistant bacteria (Kutter & Sulakvelidze, 2005). Staphylococcus aureus phages have been used in the treatment of staphylococcal infections in humans and animals (O’Flaherty et al., 2005; Wills, Kerrigan, & Soothill, 2005). In addition, phage components such as endolysins have also been tested for their anti-S. aureus activities Corresponding author. Tel.: +34 985 89 21 31; fax: +34 985 89 22 33.
E-mail address: [email protected] (P. Garcı´ a). 0958-6946/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.idairyj.2007.03.014
(Donovan, Lardeo, & Foster-Frey, 2006; O’Flaherty, Coffey, Meaney, Fitzgerald, & Ross, 2005a). The role of bacteriophages in S. aureus biofilm differentiation and maturation has also been studied (Resch, Fehrenbacher, Eisele, Schaller, & Go, 2005) as well as the application of endolysins to inhibit biofilm formation (Sass & Bierbaum, 2007). The complete genomes and proteomes of 27 S. aureus bacteriophages have recently been obtained (Kwan, Liu, DuBow, Gros, & Pelletier, 2005) and will likely assist in the identification of other proteins involved in host growth inhibition (Liu et al., 2004). S. aureus is one of the pathogenic bacteria considered as a major threat to food safety (de Buyser, Dufour, Maire, & Lafarge, 2001), and was responsible for the 1–9% outbreaks associated with milk and dairy products consumption during the period 1993–1998 in Europe (Tirado & Schmidt, 2000). In Spain, S. aureus was the causative agent in 13.9% and 11.1% of the foodborne outbreaks associated with cheeses and milk, respectively (Anonymous, 2003). The manufacture of cheese from raw milk, particularly in cases of slow or insufficient acidification of curd, has led
ARTICLE IN PRESS P. Garcı´a et al. / International Dairy Journal 17 (2007) 1232–1239
to staphylococcal outbreaks associated with this product (Le Loir, Baron, & Gautier, 2003). S. aureus may also contaminate heat-treated milk or curd if the hygienic conditions are inadequate. Therefore, S. aureus may be found in cheeses made either from raw or pasteurized milk (Coveney, Fitzgerald, & Daly, 1994). Furthermore, an initial population of 103 cfu mL1 of S. aureus in milk may be sufficient for the production of enterotoxin A in cheese at detectable levels (Meyrand et al., 1998). Thus, the risk of enterotoxin production in cheese and the subsequent human intoxication indicates a need for new procedures to control S. aureus in curd and cheese. In this regard, we propose the use of phages in curd bearing in mind that limited data have been published on the effect of phages on S. aureus survival in milk (Gill, Sabour, Leslie, & Griffiths, 2006; O’Flaherty, Coffey, Meaney, Fitzgerald, & Ross, 2005b). We have isolated from milk samples two phages, FH5 and FA72, that were able to infect several S. aureus also isolated from milk. This phage cocktail hampered the development of S. aureus in ultra-high-temperature (UHT) and pasteurized milk (Garcı´ a, P., unpublished data). However, a complete clearance of the pathogen was not achieved and S. aureus-resistant variants were easily generated. This prompted us to select lytic phages from their temperate counterparts, FH5 and FA72. The bactericidal effect of the cocktail of lytic phages on S. aureus during the manufacture of acid and enzymatic curd was investigated. 2. Materials and methods 2.1. Bacterial strains and growth conditions S. aureus Sa9, isolated from a mastitic milk sample was used as the indicator strain of phages. For host range determination, 21 S. aureus strains of bovine origin (13 from mastitic milk samples and eight from non-mastitic milk samples) were used (Table 1). All strains were grown in Luria–Bertani broth (LB) at 37 1C for 18 h. LB plates contained 2% (w/v) bacteriological agar. Baird–Parker agar supplemented with egg yolk tellurite (Scharlau Chemie, S.A. Barcelona, Spain) was used to allow differential counting. Lactococcus lactis Table 1 Host range of the lytic phages F88 and F35 Phage Host rangea
F88 F35 a
Bovine strains (n ¼ 14)
Milk strains (n ¼ 8)
Sa1, Sa9 Sa1, Sa2, Sa3, Sa4, Sa9, Sa11
AC9, AFG1, AC11, GDC6, GRA16, JFL2, AC9, AFG1, AC11, GDC3, GDC6, GRA16, JFL2, JFL6
1233
subsp. lactis IPLA 947 was used as a starter culture (Ca´rcoba, Delgado, & Rodrı´ guez, 2000). Lactococcal cultures were grown at 32 1C in M17 broth (Biokar, Beauvais, France) supplemented with 0.5% lactose. 2.2. Bacteriophages Bacteriophages FH5 and FA72 were isolated from raw milk. All phages were routinely propagated on S. aureus Sa9 according to the following procedure: LB broth supplemented with CaCl2 (10 mM) and MgSO4 (10 mM) was inoculated with 1% (v/v) inoculum of an overnight culture of the strain S. aureus Sa9 and incubated at 37 1C for 2 h with shaking. Phage was added to the bacterial culture and incubation proceeded for a further 4 h at 37 1C with shaking. Concentrated phage preparations were obtained by ultracentrifugation (100,000 g for 90 min) of culture supernatants followed by CsCl gradient centrifugation (Sambrook, Maniatis, & Fritsch, 1989). Plaque assays were performed by using 100 mL of a S. aureus Sa9 overnight culture and 100 mL of the appropriate phage dilution. This mixture was added to 5 mL of molten LB overlay (0.7% agar) and poured onto LB plates and incubated at 37 1C for 18 h. 2.3. Isolation of lytic mutant phages Lytic mutants of the temperate phages FH5 and FA72 were obtained by DNA random deletion as previously described (Ladero et al., 1998) with some modifications. Concentrated phage preparations (105 pfu) were treated with 100 mM sodium pyrophosphate (pH 7.4) for 40 min at 37 1C (phage FH5), and for 30 min at 42 1C (phage FA72). The surviving phages were plaqued and further suspended in SM buffer (20 mM Tris HCl, 10 mM MgSO4, 10 mM CaCl2, 100 mM NaCl, pH 7.5). Several rounds were carried out to obtain lytic mutants that displayed a clear plaque phenotype. The inability of phages to generate lysogenic strains was determined by treatment of mid-exponential phase cultures of surviving cells by mitomycin C (0.5 mg mL1, final concentration). 2.4. Host range The host range of the phages was determined by mixing 0.1 mL aliquots of stationary-phase host cultures with appropriate dilutions of individual phage suspensions in 5 mL of molten LB overlay (0.7% agar) and the mixture was poured on LB agar plates and incubated at 37 1C for 18 h. 2.5. Electron microscope examination
Bovine strains were isolated from mastitic milk samples. Milk strains were isolated from conventional and organic milk.
Phage particles were negatively stained with 2% uranyl acetate, and electron micrographs were taken using a JEOL 12.000 EXII transmission electron microscope (JEDL USA Inc, Peabody, MA, USA).
ARTICLE IN PRESS 1234
P. Garcı´a et al. / International Dairy Journal 17 (2007) 1232–1239
2.6. Bacteriophage DNA isolation and restriction Phage DNA was extracted by treatment of pure stocks as previously described (Sua´rez & Chater, 1981). DNA was digested with restriction enzymes (Takara Bio Inc., Japan) according to the supplier instructions. 2.7. In vitro bacterial–phage challenge tests Challenge experiments were performed in LB medium supplemented with CaCl2 (10 mM) and MgSO4 (10 mM) and inoculated with S. aureus Sa9 to obtain 1 106 cfu mL1 and a mixture of the generated lytic phages F88 and F35 (the FH5 and FA72 lytic mutants, respectively) at different multiplicities of infection (MOI of 0.01, 1 and 100). Cultures were incubated at 37 1C and optical density at 600 nm (OD600 nm) monitored over time. Challenge tests were also performed in commercial UHT whole milk. S. aureus Sa9 was inoculated to obtain 3 106 cfu mL1 and phages (F35+F88 or FH5+FA72) were added at MOI of 100. Cultures were incubated at 37 1C and samples removed and plated at regular intervals to determine viable cells counts and phage titre. Assays were performed in triplicate. 2.8. Determination of bacteriophage-insensitive mutant frequency The frequency of emergence of bacteriophage-insensitive mutants (BIMs) was carried out by mixing the appropriate volume of an overnight culture of S. aureus Sa9 (109 cfu mL1) and phage suspension in order to obtain a MOI of 100. CaCl2 (10 mM) and MgSO4 (10 mM) were added to the bacterium–phage mixture and incubated for 10 min at 37 1C. Serial dilutions of the mixtures were plated and incubated overnight at 37 1C. Resulting colonies were counted, and the BIM frequency (surviving viable counts divided by the initial viable counts) was determined. All the experiments were performed in triplicate. 2.9. Phage stability in curd The effect of curd formation on the stability of phages was tested. Pasteurized (72 1C for 15 s) whole milk, cooled to 25 1C, was inoculated with phages (1.2 104 pfu mL1) and the starter strain Lactococcus lactis subsp. lactis IPLA 947 (1%, v/v) to obtain an acid curd. Incubation proceeded for 12 h at 25 1C. For enzymatic curd formation, the starter strain was substituted for calf rennet (0.3 g L1, activity 1:10,000, Laboratorios Arroyo, Santander, Spain), and incubation was undertaken at 30 1C for 6 h. Curd samples (1 g) were drawn and homogenised in 9 mL of quarterstrength Ringer solution. Decimal dilutions were assayed for phage titre determination.
2.10. Bacterial–phage challenge test during curd manufacture Pasteurized whole milk was inoculated at 1% (v/v) with an overnight culture of the starter strain Lactococcus lactis subsp. lactis IPLA 947 (about 107 cfu mL1). S. aureus Sa9 (1 106 cfu mL1) and a mixture of phages F88+F35 (2.5 108 pfu mL1; MOI of 250) were also added. Acid coagulation proceeded for about 24 h at 25 1C. The curd was filled into moulds and the whey was allowed to drain at room temperature. Samples of milk and curd were analyzed for S. aureus counts and phage titre. In a similar way, pasteurized whole milk, cooled to 30 1C, and supplemented with 0.02 g L1 CaCl2, was inoculated with S. aureus Sa9 (1 106 cfu mL1) and a mixture of phages F88+F35 (3.5 108 pfu mL1; MOI of 350). Calf rennet (0.3 g L1, activity 1:10,000, Laboratorios Arroyo) was added 60 min after inoculation, and milk coagulation was performed for 45 min at 30 1C. Curd was cut into cubes of ca. 5 mm and stirred into the whey for 30 min. The whey was drained off and the curd filled into moulds and stored at 4 1C. Samples of milk, curd and whey were taken at different times. Decimal dilutions of milk and homogenates of curd were made in Ringer solution. Baird–Parker plates were used to determine S. aureus counts. Phage titre was obtained by plaque assays on Luria–Bertani medium. All the curd manufacturing assays (acid and rennet) were performed in triplicate. 3. Results 3.1. Characterization of phages FH5 and FA72 Phages FH5 and FA72 were isolated from raw bovine milk by enrichment in the presence of S. aureus Sa9. The virions of both phages were observed under electron microscopy. FH5 has an isometric capsid (5471 nm diameter) and a long non-contractile flexible tail (1827 4 nm) (Fig. 1A). FA72 presents an elongated head (927 2 nm long for 4371 nm wide) and a long non-contractile tail (29574 nm) (Fig. 1B). Both phages present a baseplate with spikes at the end of the tail. Their morphology allows the classification of these two phages into the family Shiphoviridae of the order Caudovirales. The DNA of FH5 and FA72 was purified and digested with several restriction enzymes and the fragments were visualized after agarose gel electrophoresis (Fig. 1C). The deduced genome size is ca. 40 kb for both phages, a value that falls within the normal range for viruses of the Shiphoviridae family (Ackermann, 2003). 3.2. Generation of lytic mutants from the temperate phages FH5 and FA72 Deletion mutants of the temperate phages FH5 and FA72 were obtained by sodium pyrophosphate treatment. The presence of this chelating agent destabilizes the capsid
ARTICLE IN PRESS P. Garcı´a et al. / International Dairy Journal 17 (2007) 1232–1239
1235
Fig. 1. Electron micrographs of phages FH5 (A) and FA72 (B). Phages were negatively stained with 2% uranyl acetate. (C) Phage DNAs digested with restriction endonucleases. Lane identification is as follows: Lambda PstI ladder (lane 1). DNA of phage FH5 digested with EcoRI (lane 2) and HindIII (lane 3). DNA of phage FA72 digested with EcoRI (lane 4) and HindIII (lane 5).
3.3. Bactericidal activity of the lytic mutants F88 and F35 in milk To verify the effectiveness of the lytic mutants in milk, bacterial–phage challenge tests were carried out and their performance was compared to that of the temperate phages. UHT whole milk was inoculated with S. aureus
1.6 1.4 1.2 OD600
and only phages which have lost some DNA are viable. In a first round, the treatment resulted in a 0.1% survival of phages in both cases. After six treatment rounds, the survival percentage increased up to 10%. At this point, phages from plaques that displayed a clear plaque phenotype were picked individually and suspended in SM buffer. Volumes (5 mL) were spotted on plates containing the host strain and incubated for 18 h at 37 1C to further purify the virulent variants and confirm the clear plaque phenotype (data not shown). One variant from each phage FH5 and FA72 was randomly selected and named F88 and F35, respectively. The bacteriophage host ranges of these lytic phages were determined against a panel of 21 S. aureus strains (laboratory collection) isolated from milk samples (Table 1). The development of BIMs of S. aureus Sa9 was tested using the combination of either temperate phages or their lytic counterparts. The frequency obtained was 2.57 0.12 104 for FH5+FA72 and 1.370.24 106 for F88+F35. As expected, the temperate phage mixture gave rise to a higher proportion of BIM, likely due to the acquisition of the lysogenic state by the host cells. The lytic activity of the mixture of phages F88+F35 was examined in detail in LB broth inoculated with S. aureus Sa9 (1 106 cfu mL1) and infected with the two phages (MOI 0.01, 1 and 100). As shown in Fig. 2, both MOI of 100 and 1 prevented S. aureus growth, and no changes in OD were recorded. By contrast, MOI of 0.01 allowed S. aureus growth throughout 3 h of incubation at 37 1C in a similar fashion as occurred in the absence of phages. Later on, a reduction in growth rate was observed and OD declined after 4 h of incubation, indicating lysis of the culture.
1 0.8 0.6 0.4 0.2 0
0
100
200 Time (min)
300
Fig. 2. Challenge in LB medium of S. aureus Sa9 with a mixture of the lytic phages F88 and F35 at 37 1C at different multiplicities of infection (MOI). n S. aureus Sa9, ’ S. aureus Sa9+F88+F35 (MOI 0.01), m S. aureus Sa9+F88+F35 (MOI 1), & S. aureus Sa9+F88+F35 (MOI 100). Assays were performed in triplicate. Data reported are means7standard deviations.
Sa9 and a cocktail of phages (MOI of 100) was added (Fig. 3). The presence of the temperate phages inhibited the growth of S. aureus since a difference of 3.5 log units was observed between the control culture without phages and the test culture after 8 h of incubation. By contrast, a complete clearance of the pathogen was obtained in the presence of the virulent phages after 2 h of incubation (Fig. 3). Thus, the ability of the lytic phage variants to kill S. aureus was by far more efficient than that of the temperate phages. A comparable number of phages (pfu mL1) were detected in the milk cultures after infection with either lytic or temperate phages. In both cases, phage titre increased throughout the sampling period in the assays (data not shown). 3.4. Survival of phages during curd manufacture To determine if the production of lactic acid by starter cultures and the concomitant reduction of pH could
ARTICLE IN PRESS P. Garcı´a et al. / International Dairy Journal 17 (2007) 1232–1239
1236 10
6
7
6
4
5 6 4
5 4
3 3
2
0
pH
Phage titre (log pfu mL-1)
8 Cell counts (log cfu mL-1)
8
2 0
2
4
6
8
2 0
2
4
Time (h)
Fig. 3. Bactericidal ability of mixtures of the temperate phages FH5+FA72 and their lytic variants F88+F35 against S. aureus Sa9. The assay was performed in UHT whole milk at 37 1C. Phages were added at multiplicity of infection (MOI) 100. n S. aureus Sa9, ’ S. aureus Sa9+F35+F88, & S. aureus Sa9+FH5+FA72. Assays were performed in triplicate. Data reported are means7standard deviations.
6 Time (h)
8
10
12
Fig. 4. Evolution of phage titre and pH during acid curd manufacturing. pH in enzymatic curd, pH in acid curd, m phage titre in enzimatic curd, E phage titre in acid curd. Assays were performed in triplicate. Data reported are means7standard deviations.
7
10
8 7
6
6 5
pH
negatively affect the viability of phages, the stability of phage cocktail (phages F88 and F35) in acid and rennetcoagulated curd manufacturing processes was tested. Pasteurized milk was infected with the phage cocktail (1.2 104 pfu mL1) in the presence of the starter strain L. lactis subsp. lactis IPLA 947. The pH was monitored throughout the assays. The phage titre declined by approximately 2 log units throughout a 12 h period. The decrease was more acute between 4 and 6 h, when the pH dropped from 6.19 to 5.38 (Fig. 4). By contrast, phage titre kept stable during enzymatic curd manufacturing.
Phage titre (log pfu mL-1) and cell counts (log cfu mL-1)
9
4 5
3 2 1 0
4 0
3.5. Inhibition of S. aureus by phages in curd manufacturing processes The ability of the lytic phages F88+F35 to inhibit the development of S. aureus Sa9 in acid curd manufacturing process was investigated. The evolution of microbial counts of the staphylococcal strain throughout the milk coagulation process driven by the starter strain L. lactis subsp. lactis IPLA 947 is shown in Fig. 5. As a control, the growth of S. aureus in pasteurized milk in the absence of the starter culture was also tested. In this case, S. aureus was able to grow up to 108 cfu mL1 at the end of the incubation period. However, when the starter was present, partial inhibition of S. aureus growth was observed in the first 8 h, and a difference of 1.6 log units in staphylococcal counts was detected between cultures with and without the starter strain. When the phage cocktail was also present, viable counts were reduced dramatically, being not detected after 4 h of incubation. As expected, the phage titre increased
5
10 15 Time (h)
20
25
Fig. 5. Evolution of S. aureus viable counts (cfu mL1) and phage titre (pfu mL1) in acid curd treated with a mixture of specific phages. pH, E S. aureus Sa9, ’ S. aureus Sa9+L. lactis subsp. lactis IPLA 947, & S. aureus Sa9+L. lactis subsp. lactis IPLA 947+F35+F88, n phages F35+F88 titre. Assays were performed in triplicate. Data reported are means7standard deviations.
during the same period due to the release of new phages, decreasing after 16 h of incubation. A very similar strong bactericidal effect of the lytic phages on the S. aureus strain was observed during enzymatic curd manufacturing (Fig. 6). In the absence of phages, continuous bacterial growth was detected throughout the coagulation process, viable counts being lower in whey than in curd. By contrast, in the presence of phages, a fast decrease of the S. aureus viable counts was observed. The pathogen was not detected either in curd or whey after 1 h of incubation. In parallel, a slight increase of phage titre occurred and remained constant until storage.
ARTICLE IN PRESS 10
10
8
8
6
6
4
4
2
2
0
0
50
100 150 Time (min)
200
Phage titre (log pfu mL-1)
Cell counts (log cfu mL-1)
P. Garcı´a et al. / International Dairy Journal 17 (2007) 1232–1239
0 250
Fig. 6. Evolution of S. aureus viable counts (cfu mL1) and phage titre (pfu mL1) in enzyme-produced curd manufacture. n phages F35+F88 titre, ’ S. aureus Sa9 in curd and & S. aureus Sa9+phages F35+F88 in curd, E S. aureus Sa9 in whey and X S. aureus Sa9+phages F35+F88 in whey. Assays were performed in triplicate. Data reported are means7 standard deviations.
4. Discussion We have evaluated the suitability of S. aureus lytic phages for the biocontrol of this foodborne pathogen in some dairy products. S. aureus is one of the most frequent agents of bovine mastitis that contribute to milk contamination. Of particular relevance to the food processing industry is the ability of some strains to produce heat stable enterotoxins that cause staphylococcal food poisoning (Dinges, Orwin, & Schlievert, 2000). Therefore, new approaches to fight against this pathogen are necessary. Bacteriophages possess attributes that appear to be attractive to inhibit foodborne pathogens and spoilage organisms (Greer, 2005). They are antibacterial agents since they kill their host bacteria at the end of the lytic cycle. Indeed, phage therapy has been used successfully (Kutter & Sulakvelidze, 2005; Sulakvelidze, Alavidze, & Morris, 2001). All attempts to isolate lytic phages from dairy environment were unsuccessful. Thus, we obtained two S. aureus lytic phages, F88 and F35, by DNA random deletion of their temperate counterparts, FH5 and FA72, isolated from raw milk. A mixture of both lytic variants infected six bovine strains out of a panel of 13 and all milk isolated strains available in our laboratory collection. These phages proved to be very efficient in the inhibition of the pathogen in UHT milk and in both acid and enzymatically produced curds. We have taken into account that the use of virulent phages holds several advantages in relation to temperate variants for use in phage-based biocontrol approach in food safety. First of all, the frequency of development of BIM, which could compromise the efficacy of a phage treatment, is often associated with point mutations in genes encoding receptor molecules on the bacterial cell surface (Forde & Fitzgerald, 1999), and commonly these mutants
1237
revert to phage sensitivity rapidly (O’Flynn, Ross, Fitzgerald, & Coffey, 2004). However, in temperate bacteriophages, higher BIM frequencies are found due to the acquisition of a lysogenic state that renders the cells resistant to infection. Consequently, our lytic variants showed a lower rate of BIM, due to their inability to lysogenize. Such properties are crucial for preparing phage mixtures for the control of unwanted bacteria in food. On the other hand, temperate phages are one of the leading causes of dissemination of antibiotic resistance and virulence factors (e.g., enterotoxin production) (reviewed by Brussow, Canchaya, & Hardt, 2004) and their deliberated spread in nature should be avoided. Considering the temperate origin of our lytic-derived phages, confirmation of the lack of any virulence trait in their genome should be obtained. Absence of several enterotoxins (enterotoxin A, D, E, J and leukotoxin lukM-lukF-PV) has been preliminary confirmed by PCR (data not shown). A mixture of the lytic phages (F88+F35) was able to withstand the stresses found in acid curd manufacturing processes. Even though the phage cocktail was partially inactivated by low pH, it was able to completely eradicate viable S. aureus cells in curd made of heavily contaminated milk. Obviously, the presence of host cells, in which the phage is able to replicate, was enough to counteract pH inactivation. Further work is needed to determine the minimum host density which ensures phage replication in these conditions. Preliminary results in pasteurized milk indicate that lower contamination levels (102 cfu mL1) are still enough and the pathogen can be eliminated (Garcı´ a, unpublished observation). Previous reports have shown that milk proteins could inhibit phage adsorption to the cell surface (Gill et al., 2006). However, according to our results, the phages were very stable and active during enzymatic curd formation, implying that pH is the most crucial inactivation factor. The activity of these phages in milk, in contrast to the inactivity of the bacteriophage K in raw milk (Gill et al., 2006), could be related to their milk origin. It is known that phages rapidly evolved along with their host and their environment (Brussow et al., 2004). The use of a mixture of phages to control undesirable bacteria in food has been reported in several food systems (Carlton, Noordman, Biswas, de Meester, & Loessner, 2005; Hudson et al., 2005; Modi, Hirvi, Hill, & Griffiths, 2001). Furthermore, the use of an anti-Listeria phage preparation has been recently approved by the FDA (FDA, 2006). Studies on the application of phages to animals reported no adverse or unexpected effects (Biswas et al., 2002; Bruttin & Bru¨ssow, 2005; Cerveny, DePaola, Duckworth, & Gulig, 2002). In addition, their specificity and ubiquitous presence in nature makes a disturbance in the intestinal microbiota unlikely. Hence, the intake of pathogen-specific phages along with food may be harmless to humans. Similarly, the food microbiota, particularly relevant in the production of fermented products, would not be disturbed. Data obtained concerning the S. aureus
ARTICLE IN PRESS 1238
P. Garcı´a et al. / International Dairy Journal 17 (2007) 1232–1239
phages F88 and F35 suggest that their use as an additive for biopreservation of dairy products would be efficient and safe provided that no virulence traits are encoded in their genome. Challenge studies are in progress to determine the most advantageous conditions (e.g., host density, temperature) in which these phages can efficiently inhibit S. aureus in milk and other dairy products. These studies are necessary to implement phage biocontrol in dairy processes. 5. Conclusions The results demonstrate that the lytic bacteriophages, F35 and F88, generate a reduced number of bacteriophageinsensitive mutants in relation to their temperate counterparts. Consequently, their ability to inhibit the development of S. aureus Sa9 in milk is enhanced. The addition of a mixture of these lytic phages to milk prior to acid and enzymatic curd manufacture eliminates up to 6 log units of S. aureus. Therefore, phage biocontrol is a feasible additional parameter to enhance the safety of dairy products. Acknowledgement This research study was supported by Grant AGL200603659/ALI from the Ministry of Education and Science of Spain. Pilar Garcı´ a is a fellow of the Spanish Ministry of Education Ramo´n y Cajal Research Programme. References Ackermann, H. W. (2003). Bacteriophage observations and evolution. Research Microbiology, 154, 245–251. Anonymous. (2003). European Commission, Health & Consumer Protection. Directorate-General, 26–27 March. Opinion of the scientific committee on veterinary measures relating to public health on staphylococcal enterotoxins in milk products, particularly cheeses. Biswas, B., Adhya, S., Washart, P., Paul, B., Trostel, A. N., Powell, B., et al. (2002). Bacteriophage therapy rescues mice bacteremic from a clinical isolate of vancomycin-resistant Enterococcus faecium. Infection and Immunology, 70, 204–210. Brussow, H., Canchaya, C., & Hardt, W. D. (2004). Phages and the evolution of bacterial pathogens: From genomic rearrangements to lysogenic conversion. Microbiology and Molecular Biology Reviews, 68, 560–602. Bruttin, A., & Bru¨ssow, H. (2005). Human volunteers receiving Escherichia coli phage T4 orally: A safety test of phage therapy. Antimicrobial Agents and Chemotherapy, 49, 2874–2878. Ca´rcoba, R., Delgado, T., & Rodrı´ guez, A. (2000). Comparative performance of a mixed strain starter in cow’s milk, ewe’s milk and mixtures of these milks. European Food Research Technology, 211, 141–146. Carlton, R. M., Noordman, W. H., Biswas, B., de Meester, E. D., & Loessner, M. J. (2005). Bacteriophage P100 for control of Listeria monocytogenes in foods: Genome sequence, bioinformatic analyses, oral toxicity study, and application. Regulatory Toxicology and Pharmacology, 43, 301–312. Cerveny, K. E., DePaola, A., Duckworth, H., & Gulig, P. A. (2002). Phage therapy of local and systemic disease caused by Vibrio vulnificus in iron–dextran-treated mice. Infection and Immunology, 70, 6251–6262.
Coveney, H. M., Fitzgerald, G. F., & Daly, C. (1994). A study of the microbiological status of Irish farmhouse cheeses with emphasis on selected pathogenic and spoilage microorganisms. Journal of Applied Bacteriology, 77, 621–630. de Buyser, M. L., Dufour, B., Maire, M., & Lafarge, V. (2001). Implication of milk and milk products in foodborne diseases in France and in different industrialised countries. International Journal of Food Microbiology, 67, 1–17. Dinges, M. M., Orwin, P. M., & Schlievert, P. M. (2000). Exotoxins of Staphylococcus aureus. Clinical Microbiology Reviews, 13, 16–64. Donovan, D. M., Lardeo, M., & Foster-Frey, J. (2006). Lysis of staphylococcal mastitis pathogens by bacteriophage phi11 endolysin. FEMS Microbiology Letters, 265, 133–139. FDA. (2006). Food Additives Permitted for Direct Addition to Food for Human Consumption; Bacteriophage Preparation. Federal Register: August 18, 2006., 71(160), 47729–47732. Forde, A., & Fitzgerald, G. F. (1999). Bacteriophage defence systems in lactic acid bacteria. Antonie Van Leeuwenhoek, 76, 89–113. Gill, J. J., Sabour, P. M., Leslie, K. E., & Griffiths, M. W. (2006). Bovine whey proteins inhibit the interaction of Staphylococcus aureus and bacteriophage K. Journal of Applied Microbiology, 101, 377–386. Greer, G. G. (2005). Bacteriophage control of foodborne bacteria. Journal of Food Protection, 68, 1102–1111. Hudson, J. A., Billington, C., Carey-Smith, C., & Greening, G. (2005). Bacteriophages as biocontrol agents in food. Journal of Food Protection, 68, 426–437. Kutter, E., & Sulakvelidze, A. (2005). Bacteriophages Biology and Applications. Boca Raton, FL, USA: CRC Press. Kwan, T., Liu, J., DuBow, M., Gros, P., & Pelletier, J. (2005). The complete genomes and proteomes of 27 Staphylococcus aureus bacteriophages. Proceedings of the National Academy of Sciences of the United States of America, 102, 5174–5179. Ladero, V., Garcı´ a, P., Bascara´n, V., Herrero, M., Alvarez, M. A., & Sua´rez, J. E. (1998). Identification of the repressor-encoding gene of the Lactobacillus bacteriophage A2. Journal of Bacteriology, 180, 3474–3476. Le Loir, Y., Baron, F., & Gautier, M. (2003). Staphylococcus aureus and food poisoning. Genetics and Molecular Research, 2, 63–76. Liu, J., Dehbi, M., Moeck, G., Arhin, F., Bauda, P., Bergeron, D., et al. (2004). Antimicrobial drug discovery through bacteriophage genomics. Nature Biotechnology, 22, 185–191. Meyrand, A., Boutrand-Loei, S., Ray-Gueniot, S., Mazuy, C., Gaspard, C. E., Jaubert, G., et al. (1998). Growth and enterotoxin production of Staphylococcus aureus during the manufacture and ripening of Camembert-type cheeses from goats’ milk. Journal of Applied Microbiology, 85, 537–544. Modi, R., Hirvi, Y., Hill, A., & Griffiths, M. W. (2001). Effect of phage on survival of Salmonella Enteritidis during manufacture and storage of Cheddar cheese made from raw and pasteurized milk. Journal of Food Protection, 64, 927–933. O’Flaherty, S., Coffey, A., Meaney, W., Fitzgerald, G. F., & Ross, R. P. (2005a). The recombinant phage lysin LysK has a broad spectrum of lytic activity against clinically relevant staphylococci, including methicillin-resistant Staphylococcus aureus. Journal of Bacteriology, 187, 7161–7164. O’Flaherty, S., Coffey, A., Meaney, W. J., Fitzgerald, G. F., & Ross, R. P. (2005b). Inhibition of bacteriophage K proliferation on Staphylococcus aureus in raw bovine milk. Letters in Applied Microbiology, 41, 274–279. O’Flaherty, S., Ross, R. P., Meaney, W., Fitzgerald, G. F., Elbreki, M. F., & Coffey, A. (2005). Potential of the polyvalent anti-Staphylococcus bacteriophage K for control of antibiotic-resistant staphylococci from hospitals. Applied and Environmental Microbiology, 71, 1836–1842. O’Flynn, G., Ross, R. P., Fitzgerald, G. F., & Coffey, A. (2004). Evaluation of a cocktail of three bacteriophages for biocontrol of E. coli O157:H7. Applied and Environmental Microbiology, 70, 3417–3424.
ARTICLE IN PRESS P. Garcı´a et al. / International Dairy Journal 17 (2007) 1232–1239 Resch, A., Fehrenbacher, B., Eisele, K., Schaller, M., & Go, F. (2005). Phage release from biofilm and planktonic Staphylococcus aureus cells. FEMS Microbiology Letters, 252, 89–96. Sambrook, J., Maniatis, T., & Fritsch, E. F. (1989). Molecular cloning: A laboratory manual (2nd ed.). Cold Spring Harbor, NY, USA: Cold Spring Harbor Laboratory Press. Sass, P., & Bierbaum, G. (2007). Lytic activity of recombinant bacteriophage F11 and F12 endolysins on whole cells and biofilms of Staphylococcus aureus. Applied and Environmental Microbiology, 73, 347–352.
1239
Sua´rez, J. E., & Chater, K. F. (1981). Development of a DNA cloning in Streptomyces using phage vectors. Ciencia Biologica, 6, 99–110. Sulakvelidze, A., Alavidze, Z., & Morris, J. G., Jr. (2001). Bacteriophage therapy. Antimicrobial Agents and Chemotherapy, 45, 649–659 Review. Tirado, C. & Schmidt, K. (2000). WHO Surveillance Programme for Control of Foodborne Infections and Intoxications in Europe. Seventh Report /http://www.bfr.bund.de/internet/7threport/7threp_fr.htmS. Wills, Q. F., Kerrigan, C., & Soothill, J. S. (2005). Experimental bacteriophage protection against Staphylococcus aureus abscesses in a rabbit model. Antimicrobial Agents and Chemotherapy, 49, 1220–1221.
International Dairy Journal 17 (2007) 1232–1239 www.elsevier.com/locate/idairyj
Biocontrol of Staphylococcus aureus in curd manufacturing processes using bacteriophages Pilar Garcı´ a, Carmen Madera, Beatriz Martı´ nez, Ana Rodrı´ guez Instituto de Productos La´cteos de Asturias (IPLA-CSIC). Apdo. 85, 33300 Villaviciosa, Asturias, Spain Received 25 October 2006; accepted 27 March 2007
Abstract The ability of specific bacteriophages to inhibit Staphylococcus aureus growth in curd manufacturing processes was determined. Two lytic bacteriophages specific against S. aureus were obtained by DNA random deletion from the milk-isolated temperate phages, FH5 and FA72. A cocktail of these lytic phages, F88 and F35, at multiplicity of infection (MOI) of 100, produced a complete elimination of 3 106 cfu mL1 of the pathogen in ultra-high-temperature (UHT) whole milk at 37 1C. Furthermore, the frequency of emergence of bacteriophage-insensitive mutants was reduced up to 200-fold in the presence of the two lytic phages compared with that detected with the combination of the temperate counterparts. The lytic phage derivatives, added to milk, were able to decrease rapidly the viable counts of S. aureus during curd manufacture. In acid curd, the pathogen was not detected after 4 h of incubation at 25 1C, whereas pathogen clearance was achieved within 1 h of incubation at 30 1C in renneted curd. These results indicate that lytic bacteriophages could be used as biopreservatives in the manufacture of particular dairy products. r 2007 Elsevier Ltd. All rights reserved. Keywords: Lytic phage; Biocontrol; Dairy products; Staphylococcus aureus
1. Introduction The current technologies employed to inactivate bacterial pathogens in foods are not always foolproof and, therefore, new approaches for improving food safety are necessary. Bacteriophages provide an attractive alternative since phages are ubiquitous in different environments, unable to infect human cells and, consequently, they have great potential for use as biocontrol agents in foods (Hudson, Billington, Carey-Smith, & Greening, 2005). The exploitation of bacteriophages has already become an interesting tool to fight the emergence of antibioticresistant bacteria (Kutter & Sulakvelidze, 2005). Staphylococcus aureus phages have been used in the treatment of staphylococcal infections in humans and animals (O’Flaherty et al., 2005; Wills, Kerrigan, & Soothill, 2005). In addition, phage components such as endolysins have also been tested for their anti-S. aureus activities Corresponding author. Tel.: +34 985 89 21 31; fax: +34 985 89 22 33.
E-mail address: [email protected] (P. Garcı´ a). 0958-6946/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.idairyj.2007.03.014
(Donovan, Lardeo, & Foster-Frey, 2006; O’Flaherty, Coffey, Meaney, Fitzgerald, & Ross, 2005a). The role of bacteriophages in S. aureus biofilm differentiation and maturation has also been studied (Resch, Fehrenbacher, Eisele, Schaller, & Go, 2005) as well as the application of endolysins to inhibit biofilm formation (Sass & Bierbaum, 2007). The complete genomes and proteomes of 27 S. aureus bacteriophages have recently been obtained (Kwan, Liu, DuBow, Gros, & Pelletier, 2005) and will likely assist in the identification of other proteins involved in host growth inhibition (Liu et al., 2004). S. aureus is one of the pathogenic bacteria considered as a major threat to food safety (de Buyser, Dufour, Maire, & Lafarge, 2001), and was responsible for the 1–9% outbreaks associated with milk and dairy products consumption during the period 1993–1998 in Europe (Tirado & Schmidt, 2000). In Spain, S. aureus was the causative agent in 13.9% and 11.1% of the foodborne outbreaks associated with cheeses and milk, respectively (Anonymous, 2003). The manufacture of cheese from raw milk, particularly in cases of slow or insufficient acidification of curd, has led
ARTICLE IN PRESS P. Garcı´a et al. / International Dairy Journal 17 (2007) 1232–1239
to staphylococcal outbreaks associated with this product (Le Loir, Baron, & Gautier, 2003). S. aureus may also contaminate heat-treated milk or curd if the hygienic conditions are inadequate. Therefore, S. aureus may be found in cheeses made either from raw or pasteurized milk (Coveney, Fitzgerald, & Daly, 1994). Furthermore, an initial population of 103 cfu mL1 of S. aureus in milk may be sufficient for the production of enterotoxin A in cheese at detectable levels (Meyrand et al., 1998). Thus, the risk of enterotoxin production in cheese and the subsequent human intoxication indicates a need for new procedures to control S. aureus in curd and cheese. In this regard, we propose the use of phages in curd bearing in mind that limited data have been published on the effect of phages on S. aureus survival in milk (Gill, Sabour, Leslie, & Griffiths, 2006; O’Flaherty, Coffey, Meaney, Fitzgerald, & Ross, 2005b). We have isolated from milk samples two phages, FH5 and FA72, that were able to infect several S. aureus also isolated from milk. This phage cocktail hampered the development of S. aureus in ultra-high-temperature (UHT) and pasteurized milk (Garcı´ a, P., unpublished data). However, a complete clearance of the pathogen was not achieved and S. aureus-resistant variants were easily generated. This prompted us to select lytic phages from their temperate counterparts, FH5 and FA72. The bactericidal effect of the cocktail of lytic phages on S. aureus during the manufacture of acid and enzymatic curd was investigated. 2. Materials and methods 2.1. Bacterial strains and growth conditions S. aureus Sa9, isolated from a mastitic milk sample was used as the indicator strain of phages. For host range determination, 21 S. aureus strains of bovine origin (13 from mastitic milk samples and eight from non-mastitic milk samples) were used (Table 1). All strains were grown in Luria–Bertani broth (LB) at 37 1C for 18 h. LB plates contained 2% (w/v) bacteriological agar. Baird–Parker agar supplemented with egg yolk tellurite (Scharlau Chemie, S.A. Barcelona, Spain) was used to allow differential counting. Lactococcus lactis Table 1 Host range of the lytic phages F88 and F35 Phage Host rangea
F88 F35 a
Bovine strains (n ¼ 14)
Milk strains (n ¼ 8)
Sa1, Sa9 Sa1, Sa2, Sa3, Sa4, Sa9, Sa11
AC9, AFG1, AC11, GDC6, GRA16, JFL2, AC9, AFG1, AC11, GDC3, GDC6, GRA16, JFL2, JFL6
1233
subsp. lactis IPLA 947 was used as a starter culture (Ca´rcoba, Delgado, & Rodrı´ guez, 2000). Lactococcal cultures were grown at 32 1C in M17 broth (Biokar, Beauvais, France) supplemented with 0.5% lactose. 2.2. Bacteriophages Bacteriophages FH5 and FA72 were isolated from raw milk. All phages were routinely propagated on S. aureus Sa9 according to the following procedure: LB broth supplemented with CaCl2 (10 mM) and MgSO4 (10 mM) was inoculated with 1% (v/v) inoculum of an overnight culture of the strain S. aureus Sa9 and incubated at 37 1C for 2 h with shaking. Phage was added to the bacterial culture and incubation proceeded for a further 4 h at 37 1C with shaking. Concentrated phage preparations were obtained by ultracentrifugation (100,000 g for 90 min) of culture supernatants followed by CsCl gradient centrifugation (Sambrook, Maniatis, & Fritsch, 1989). Plaque assays were performed by using 100 mL of a S. aureus Sa9 overnight culture and 100 mL of the appropriate phage dilution. This mixture was added to 5 mL of molten LB overlay (0.7% agar) and poured onto LB plates and incubated at 37 1C for 18 h. 2.3. Isolation of lytic mutant phages Lytic mutants of the temperate phages FH5 and FA72 were obtained by DNA random deletion as previously described (Ladero et al., 1998) with some modifications. Concentrated phage preparations (105 pfu) were treated with 100 mM sodium pyrophosphate (pH 7.4) for 40 min at 37 1C (phage FH5), and for 30 min at 42 1C (phage FA72). The surviving phages were plaqued and further suspended in SM buffer (20 mM Tris HCl, 10 mM MgSO4, 10 mM CaCl2, 100 mM NaCl, pH 7.5). Several rounds were carried out to obtain lytic mutants that displayed a clear plaque phenotype. The inability of phages to generate lysogenic strains was determined by treatment of mid-exponential phase cultures of surviving cells by mitomycin C (0.5 mg mL1, final concentration). 2.4. Host range The host range of the phages was determined by mixing 0.1 mL aliquots of stationary-phase host cultures with appropriate dilutions of individual phage suspensions in 5 mL of molten LB overlay (0.7% agar) and the mixture was poured on LB agar plates and incubated at 37 1C for 18 h. 2.5. Electron microscope examination
Bovine strains were isolated from mastitic milk samples. Milk strains were isolated from conventional and organic milk.
Phage particles were negatively stained with 2% uranyl acetate, and electron micrographs were taken using a JEOL 12.000 EXII transmission electron microscope (JEDL USA Inc, Peabody, MA, USA).
ARTICLE IN PRESS 1234
P. Garcı´a et al. / International Dairy Journal 17 (2007) 1232–1239
2.6. Bacteriophage DNA isolation and restriction Phage DNA was extracted by treatment of pure stocks as previously described (Sua´rez & Chater, 1981). DNA was digested with restriction enzymes (Takara Bio Inc., Japan) according to the supplier instructions. 2.7. In vitro bacterial–phage challenge tests Challenge experiments were performed in LB medium supplemented with CaCl2 (10 mM) and MgSO4 (10 mM) and inoculated with S. aureus Sa9 to obtain 1 106 cfu mL1 and a mixture of the generated lytic phages F88 and F35 (the FH5 and FA72 lytic mutants, respectively) at different multiplicities of infection (MOI of 0.01, 1 and 100). Cultures were incubated at 37 1C and optical density at 600 nm (OD600 nm) monitored over time. Challenge tests were also performed in commercial UHT whole milk. S. aureus Sa9 was inoculated to obtain 3 106 cfu mL1 and phages (F35+F88 or FH5+FA72) were added at MOI of 100. Cultures were incubated at 37 1C and samples removed and plated at regular intervals to determine viable cells counts and phage titre. Assays were performed in triplicate. 2.8. Determination of bacteriophage-insensitive mutant frequency The frequency of emergence of bacteriophage-insensitive mutants (BIMs) was carried out by mixing the appropriate volume of an overnight culture of S. aureus Sa9 (109 cfu mL1) and phage suspension in order to obtain a MOI of 100. CaCl2 (10 mM) and MgSO4 (10 mM) were added to the bacterium–phage mixture and incubated for 10 min at 37 1C. Serial dilutions of the mixtures were plated and incubated overnight at 37 1C. Resulting colonies were counted, and the BIM frequency (surviving viable counts divided by the initial viable counts) was determined. All the experiments were performed in triplicate. 2.9. Phage stability in curd The effect of curd formation on the stability of phages was tested. Pasteurized (72 1C for 15 s) whole milk, cooled to 25 1C, was inoculated with phages (1.2 104 pfu mL1) and the starter strain Lactococcus lactis subsp. lactis IPLA 947 (1%, v/v) to obtain an acid curd. Incubation proceeded for 12 h at 25 1C. For enzymatic curd formation, the starter strain was substituted for calf rennet (0.3 g L1, activity 1:10,000, Laboratorios Arroyo, Santander, Spain), and incubation was undertaken at 30 1C for 6 h. Curd samples (1 g) were drawn and homogenised in 9 mL of quarterstrength Ringer solution. Decimal dilutions were assayed for phage titre determination.
2.10. Bacterial–phage challenge test during curd manufacture Pasteurized whole milk was inoculated at 1% (v/v) with an overnight culture of the starter strain Lactococcus lactis subsp. lactis IPLA 947 (about 107 cfu mL1). S. aureus Sa9 (1 106 cfu mL1) and a mixture of phages F88+F35 (2.5 108 pfu mL1; MOI of 250) were also added. Acid coagulation proceeded for about 24 h at 25 1C. The curd was filled into moulds and the whey was allowed to drain at room temperature. Samples of milk and curd were analyzed for S. aureus counts and phage titre. In a similar way, pasteurized whole milk, cooled to 30 1C, and supplemented with 0.02 g L1 CaCl2, was inoculated with S. aureus Sa9 (1 106 cfu mL1) and a mixture of phages F88+F35 (3.5 108 pfu mL1; MOI of 350). Calf rennet (0.3 g L1, activity 1:10,000, Laboratorios Arroyo) was added 60 min after inoculation, and milk coagulation was performed for 45 min at 30 1C. Curd was cut into cubes of ca. 5 mm and stirred into the whey for 30 min. The whey was drained off and the curd filled into moulds and stored at 4 1C. Samples of milk, curd and whey were taken at different times. Decimal dilutions of milk and homogenates of curd were made in Ringer solution. Baird–Parker plates were used to determine S. aureus counts. Phage titre was obtained by plaque assays on Luria–Bertani medium. All the curd manufacturing assays (acid and rennet) were performed in triplicate. 3. Results 3.1. Characterization of phages FH5 and FA72 Phages FH5 and FA72 were isolated from raw bovine milk by enrichment in the presence of S. aureus Sa9. The virions of both phages were observed under electron microscopy. FH5 has an isometric capsid (5471 nm diameter) and a long non-contractile flexible tail (1827 4 nm) (Fig. 1A). FA72 presents an elongated head (927 2 nm long for 4371 nm wide) and a long non-contractile tail (29574 nm) (Fig. 1B). Both phages present a baseplate with spikes at the end of the tail. Their morphology allows the classification of these two phages into the family Shiphoviridae of the order Caudovirales. The DNA of FH5 and FA72 was purified and digested with several restriction enzymes and the fragments were visualized after agarose gel electrophoresis (Fig. 1C). The deduced genome size is ca. 40 kb for both phages, a value that falls within the normal range for viruses of the Shiphoviridae family (Ackermann, 2003). 3.2. Generation of lytic mutants from the temperate phages FH5 and FA72 Deletion mutants of the temperate phages FH5 and FA72 were obtained by sodium pyrophosphate treatment. The presence of this chelating agent destabilizes the capsid
ARTICLE IN PRESS P. Garcı´a et al. / International Dairy Journal 17 (2007) 1232–1239
1235
Fig. 1. Electron micrographs of phages FH5 (A) and FA72 (B). Phages were negatively stained with 2% uranyl acetate. (C) Phage DNAs digested with restriction endonucleases. Lane identification is as follows: Lambda PstI ladder (lane 1). DNA of phage FH5 digested with EcoRI (lane 2) and HindIII (lane 3). DNA of phage FA72 digested with EcoRI (lane 4) and HindIII (lane 5).
3.3. Bactericidal activity of the lytic mutants F88 and F35 in milk To verify the effectiveness of the lytic mutants in milk, bacterial–phage challenge tests were carried out and their performance was compared to that of the temperate phages. UHT whole milk was inoculated with S. aureus
1.6 1.4 1.2 OD600
and only phages which have lost some DNA are viable. In a first round, the treatment resulted in a 0.1% survival of phages in both cases. After six treatment rounds, the survival percentage increased up to 10%. At this point, phages from plaques that displayed a clear plaque phenotype were picked individually and suspended in SM buffer. Volumes (5 mL) were spotted on plates containing the host strain and incubated for 18 h at 37 1C to further purify the virulent variants and confirm the clear plaque phenotype (data not shown). One variant from each phage FH5 and FA72 was randomly selected and named F88 and F35, respectively. The bacteriophage host ranges of these lytic phages were determined against a panel of 21 S. aureus strains (laboratory collection) isolated from milk samples (Table 1). The development of BIMs of S. aureus Sa9 was tested using the combination of either temperate phages or their lytic counterparts. The frequency obtained was 2.57 0.12 104 for FH5+FA72 and 1.370.24 106 for F88+F35. As expected, the temperate phage mixture gave rise to a higher proportion of BIM, likely due to the acquisition of the lysogenic state by the host cells. The lytic activity of the mixture of phages F88+F35 was examined in detail in LB broth inoculated with S. aureus Sa9 (1 106 cfu mL1) and infected with the two phages (MOI 0.01, 1 and 100). As shown in Fig. 2, both MOI of 100 and 1 prevented S. aureus growth, and no changes in OD were recorded. By contrast, MOI of 0.01 allowed S. aureus growth throughout 3 h of incubation at 37 1C in a similar fashion as occurred in the absence of phages. Later on, a reduction in growth rate was observed and OD declined after 4 h of incubation, indicating lysis of the culture.
1 0.8 0.6 0.4 0.2 0
0
100
200 Time (min)
300
Fig. 2. Challenge in LB medium of S. aureus Sa9 with a mixture of the lytic phages F88 and F35 at 37 1C at different multiplicities of infection (MOI). n S. aureus Sa9, ’ S. aureus Sa9+F88+F35 (MOI 0.01), m S. aureus Sa9+F88+F35 (MOI 1), & S. aureus Sa9+F88+F35 (MOI 100). Assays were performed in triplicate. Data reported are means7standard deviations.
Sa9 and a cocktail of phages (MOI of 100) was added (Fig. 3). The presence of the temperate phages inhibited the growth of S. aureus since a difference of 3.5 log units was observed between the control culture without phages and the test culture after 8 h of incubation. By contrast, a complete clearance of the pathogen was obtained in the presence of the virulent phages after 2 h of incubation (Fig. 3). Thus, the ability of the lytic phage variants to kill S. aureus was by far more efficient than that of the temperate phages. A comparable number of phages (pfu mL1) were detected in the milk cultures after infection with either lytic or temperate phages. In both cases, phage titre increased throughout the sampling period in the assays (data not shown). 3.4. Survival of phages during curd manufacture To determine if the production of lactic acid by starter cultures and the concomitant reduction of pH could
ARTICLE IN PRESS P. Garcı´a et al. / International Dairy Journal 17 (2007) 1232–1239
1236 10
6
7
6
4
5 6 4
5 4
3 3
2
0
pH
Phage titre (log pfu mL-1)
8 Cell counts (log cfu mL-1)
8
2 0
2
4
6
8
2 0
2
4
Time (h)
Fig. 3. Bactericidal ability of mixtures of the temperate phages FH5+FA72 and their lytic variants F88+F35 against S. aureus Sa9. The assay was performed in UHT whole milk at 37 1C. Phages were added at multiplicity of infection (MOI) 100. n S. aureus Sa9, ’ S. aureus Sa9+F35+F88, & S. aureus Sa9+FH5+FA72. Assays were performed in triplicate. Data reported are means7standard deviations.
6 Time (h)
8
10
12
Fig. 4. Evolution of phage titre and pH during acid curd manufacturing. pH in enzymatic curd, pH in acid curd, m phage titre in enzimatic curd, E phage titre in acid curd. Assays were performed in triplicate. Data reported are means7standard deviations.
7
10
8 7
6
6 5
pH
negatively affect the viability of phages, the stability of phage cocktail (phages F88 and F35) in acid and rennetcoagulated curd manufacturing processes was tested. Pasteurized milk was infected with the phage cocktail (1.2 104 pfu mL1) in the presence of the starter strain L. lactis subsp. lactis IPLA 947. The pH was monitored throughout the assays. The phage titre declined by approximately 2 log units throughout a 12 h period. The decrease was more acute between 4 and 6 h, when the pH dropped from 6.19 to 5.38 (Fig. 4). By contrast, phage titre kept stable during enzymatic curd manufacturing.
Phage titre (log pfu mL-1) and cell counts (log cfu mL-1)
9
4 5
3 2 1 0
4 0
3.5. Inhibition of S. aureus by phages in curd manufacturing processes The ability of the lytic phages F88+F35 to inhibit the development of S. aureus Sa9 in acid curd manufacturing process was investigated. The evolution of microbial counts of the staphylococcal strain throughout the milk coagulation process driven by the starter strain L. lactis subsp. lactis IPLA 947 is shown in Fig. 5. As a control, the growth of S. aureus in pasteurized milk in the absence of the starter culture was also tested. In this case, S. aureus was able to grow up to 108 cfu mL1 at the end of the incubation period. However, when the starter was present, partial inhibition of S. aureus growth was observed in the first 8 h, and a difference of 1.6 log units in staphylococcal counts was detected between cultures with and without the starter strain. When the phage cocktail was also present, viable counts were reduced dramatically, being not detected after 4 h of incubation. As expected, the phage titre increased
5
10 15 Time (h)
20
25
Fig. 5. Evolution of S. aureus viable counts (cfu mL1) and phage titre (pfu mL1) in acid curd treated with a mixture of specific phages. pH, E S. aureus Sa9, ’ S. aureus Sa9+L. lactis subsp. lactis IPLA 947, & S. aureus Sa9+L. lactis subsp. lactis IPLA 947+F35+F88, n phages F35+F88 titre. Assays were performed in triplicate. Data reported are means7standard deviations.
during the same period due to the release of new phages, decreasing after 16 h of incubation. A very similar strong bactericidal effect of the lytic phages on the S. aureus strain was observed during enzymatic curd manufacturing (Fig. 6). In the absence of phages, continuous bacterial growth was detected throughout the coagulation process, viable counts being lower in whey than in curd. By contrast, in the presence of phages, a fast decrease of the S. aureus viable counts was observed. The pathogen was not detected either in curd or whey after 1 h of incubation. In parallel, a slight increase of phage titre occurred and remained constant until storage.
ARTICLE IN PRESS 10
10
8
8
6
6
4
4
2
2
0
0
50
100 150 Time (min)
200
Phage titre (log pfu mL-1)
Cell counts (log cfu mL-1)
P. Garcı´a et al. / International Dairy Journal 17 (2007) 1232–1239
0 250
Fig. 6. Evolution of S. aureus viable counts (cfu mL1) and phage titre (pfu mL1) in enzyme-produced curd manufacture. n phages F35+F88 titre, ’ S. aureus Sa9 in curd and & S. aureus Sa9+phages F35+F88 in curd, E S. aureus Sa9 in whey and X S. aureus Sa9+phages F35+F88 in whey. Assays were performed in triplicate. Data reported are means7 standard deviations.
4. Discussion We have evaluated the suitability of S. aureus lytic phages for the biocontrol of this foodborne pathogen in some dairy products. S. aureus is one of the most frequent agents of bovine mastitis that contribute to milk contamination. Of particular relevance to the food processing industry is the ability of some strains to produce heat stable enterotoxins that cause staphylococcal food poisoning (Dinges, Orwin, & Schlievert, 2000). Therefore, new approaches to fight against this pathogen are necessary. Bacteriophages possess attributes that appear to be attractive to inhibit foodborne pathogens and spoilage organisms (Greer, 2005). They are antibacterial agents since they kill their host bacteria at the end of the lytic cycle. Indeed, phage therapy has been used successfully (Kutter & Sulakvelidze, 2005; Sulakvelidze, Alavidze, & Morris, 2001). All attempts to isolate lytic phages from dairy environment were unsuccessful. Thus, we obtained two S. aureus lytic phages, F88 and F35, by DNA random deletion of their temperate counterparts, FH5 and FA72, isolated from raw milk. A mixture of both lytic variants infected six bovine strains out of a panel of 13 and all milk isolated strains available in our laboratory collection. These phages proved to be very efficient in the inhibition of the pathogen in UHT milk and in both acid and enzymatically produced curds. We have taken into account that the use of virulent phages holds several advantages in relation to temperate variants for use in phage-based biocontrol approach in food safety. First of all, the frequency of development of BIM, which could compromise the efficacy of a phage treatment, is often associated with point mutations in genes encoding receptor molecules on the bacterial cell surface (Forde & Fitzgerald, 1999), and commonly these mutants
1237
revert to phage sensitivity rapidly (O’Flynn, Ross, Fitzgerald, & Coffey, 2004). However, in temperate bacteriophages, higher BIM frequencies are found due to the acquisition of a lysogenic state that renders the cells resistant to infection. Consequently, our lytic variants showed a lower rate of BIM, due to their inability to lysogenize. Such properties are crucial for preparing phage mixtures for the control of unwanted bacteria in food. On the other hand, temperate phages are one of the leading causes of dissemination of antibiotic resistance and virulence factors (e.g., enterotoxin production) (reviewed by Brussow, Canchaya, & Hardt, 2004) and their deliberated spread in nature should be avoided. Considering the temperate origin of our lytic-derived phages, confirmation of the lack of any virulence trait in their genome should be obtained. Absence of several enterotoxins (enterotoxin A, D, E, J and leukotoxin lukM-lukF-PV) has been preliminary confirmed by PCR (data not shown). A mixture of the lytic phages (F88+F35) was able to withstand the stresses found in acid curd manufacturing processes. Even though the phage cocktail was partially inactivated by low pH, it was able to completely eradicate viable S. aureus cells in curd made of heavily contaminated milk. Obviously, the presence of host cells, in which the phage is able to replicate, was enough to counteract pH inactivation. Further work is needed to determine the minimum host density which ensures phage replication in these conditions. Preliminary results in pasteurized milk indicate that lower contamination levels (102 cfu mL1) are still enough and the pathogen can be eliminated (Garcı´ a, unpublished observation). Previous reports have shown that milk proteins could inhibit phage adsorption to the cell surface (Gill et al., 2006). However, according to our results, the phages were very stable and active during enzymatic curd formation, implying that pH is the most crucial inactivation factor. The activity of these phages in milk, in contrast to the inactivity of the bacteriophage K in raw milk (Gill et al., 2006), could be related to their milk origin. It is known that phages rapidly evolved along with their host and their environment (Brussow et al., 2004). The use of a mixture of phages to control undesirable bacteria in food has been reported in several food systems (Carlton, Noordman, Biswas, de Meester, & Loessner, 2005; Hudson et al., 2005; Modi, Hirvi, Hill, & Griffiths, 2001). Furthermore, the use of an anti-Listeria phage preparation has been recently approved by the FDA (FDA, 2006). Studies on the application of phages to animals reported no adverse or unexpected effects (Biswas et al., 2002; Bruttin & Bru¨ssow, 2005; Cerveny, DePaola, Duckworth, & Gulig, 2002). In addition, their specificity and ubiquitous presence in nature makes a disturbance in the intestinal microbiota unlikely. Hence, the intake of pathogen-specific phages along with food may be harmless to humans. Similarly, the food microbiota, particularly relevant in the production of fermented products, would not be disturbed. Data obtained concerning the S. aureus
ARTICLE IN PRESS 1238
P. Garcı´a et al. / International Dairy Journal 17 (2007) 1232–1239
phages F88 and F35 suggest that their use as an additive for biopreservation of dairy products would be efficient and safe provided that no virulence traits are encoded in their genome. Challenge studies are in progress to determine the most advantageous conditions (e.g., host density, temperature) in which these phages can efficiently inhibit S. aureus in milk and other dairy products. These studies are necessary to implement phage biocontrol in dairy processes. 5. Conclusions The results demonstrate that the lytic bacteriophages, F35 and F88, generate a reduced number of bacteriophageinsensitive mutants in relation to their temperate counterparts. Consequently, their ability to inhibit the development of S. aureus Sa9 in milk is enhanced. The addition of a mixture of these lytic phages to milk prior to acid and enzymatic curd manufacture eliminates up to 6 log units of S. aureus. Therefore, phage biocontrol is a feasible additional parameter to enhance the safety of dairy products. Acknowledgement This research study was supported by Grant AGL200603659/ALI from the Ministry of Education and Science of Spain. Pilar Garcı´ a is a fellow of the Spanish Ministry of Education Ramo´n y Cajal Research Programme. References Ackermann, H. W. (2003). Bacteriophage observations and evolution. Research Microbiology, 154, 245–251. Anonymous. (2003). European Commission, Health & Consumer Protection. Directorate-General, 26–27 March. Opinion of the scientific committee on veterinary measures relating to public health on staphylococcal enterotoxins in milk products, particularly cheeses. Biswas, B., Adhya, S., Washart, P., Paul, B., Trostel, A. N., Powell, B., et al. (2002). Bacteriophage therapy rescues mice bacteremic from a clinical isolate of vancomycin-resistant Enterococcus faecium. Infection and Immunology, 70, 204–210. Brussow, H., Canchaya, C., & Hardt, W. D. (2004). Phages and the evolution of bacterial pathogens: From genomic rearrangements to lysogenic conversion. Microbiology and Molecular Biology Reviews, 68, 560–602. Bruttin, A., & Bru¨ssow, H. (2005). Human volunteers receiving Escherichia coli phage T4 orally: A safety test of phage therapy. Antimicrobial Agents and Chemotherapy, 49, 2874–2878. Ca´rcoba, R., Delgado, T., & Rodrı´ guez, A. (2000). Comparative performance of a mixed strain starter in cow’s milk, ewe’s milk and mixtures of these milks. European Food Research Technology, 211, 141–146. Carlton, R. M., Noordman, W. H., Biswas, B., de Meester, E. D., & Loessner, M. J. (2005). Bacteriophage P100 for control of Listeria monocytogenes in foods: Genome sequence, bioinformatic analyses, oral toxicity study, and application. Regulatory Toxicology and Pharmacology, 43, 301–312. Cerveny, K. E., DePaola, A., Duckworth, H., & Gulig, P. A. (2002). Phage therapy of local and systemic disease caused by Vibrio vulnificus in iron–dextran-treated mice. Infection and Immunology, 70, 6251–6262.
Coveney, H. M., Fitzgerald, G. F., & Daly, C. (1994). A study of the microbiological status of Irish farmhouse cheeses with emphasis on selected pathogenic and spoilage microorganisms. Journal of Applied Bacteriology, 77, 621–630. de Buyser, M. L., Dufour, B., Maire, M., & Lafarge, V. (2001). Implication of milk and milk products in foodborne diseases in France and in different industrialised countries. International Journal of Food Microbiology, 67, 1–17. Dinges, M. M., Orwin, P. M., & Schlievert, P. M. (2000). Exotoxins of Staphylococcus aureus. Clinical Microbiology Reviews, 13, 16–64. Donovan, D. M., Lardeo, M., & Foster-Frey, J. (2006). Lysis of staphylococcal mastitis pathogens by bacteriophage phi11 endolysin. FEMS Microbiology Letters, 265, 133–139. FDA. (2006). Food Additives Permitted for Direct Addition to Food for Human Consumption; Bacteriophage Preparation. Federal Register: August 18, 2006., 71(160), 47729–47732. Forde, A., & Fitzgerald, G. F. (1999). Bacteriophage defence systems in lactic acid bacteria. Antonie Van Leeuwenhoek, 76, 89–113. Gill, J. J., Sabour, P. M., Leslie, K. E., & Griffiths, M. W. (2006). Bovine whey proteins inhibit the interaction of Staphylococcus aureus and bacteriophage K. Journal of Applied Microbiology, 101, 377–386. Greer, G. G. (2005). Bacteriophage control of foodborne bacteria. Journal of Food Protection, 68, 1102–1111. Hudson, J. A., Billington, C., Carey-Smith, C., & Greening, G. (2005). Bacteriophages as biocontrol agents in food. Journal of Food Protection, 68, 426–437. Kutter, E., & Sulakvelidze, A. (2005). Bacteriophages Biology and Applications. Boca Raton, FL, USA: CRC Press. Kwan, T., Liu, J., DuBow, M., Gros, P., & Pelletier, J. (2005). The complete genomes and proteomes of 27 Staphylococcus aureus bacteriophages. Proceedings of the National Academy of Sciences of the United States of America, 102, 5174–5179. Ladero, V., Garcı´ a, P., Bascara´n, V., Herrero, M., Alvarez, M. A., & Sua´rez, J. E. (1998). Identification of the repressor-encoding gene of the Lactobacillus bacteriophage A2. Journal of Bacteriology, 180, 3474–3476. Le Loir, Y., Baron, F., & Gautier, M. (2003). Staphylococcus aureus and food poisoning. Genetics and Molecular Research, 2, 63–76. Liu, J., Dehbi, M., Moeck, G., Arhin, F., Bauda, P., Bergeron, D., et al. (2004). Antimicrobial drug discovery through bacteriophage genomics. Nature Biotechnology, 22, 185–191. Meyrand, A., Boutrand-Loei, S., Ray-Gueniot, S., Mazuy, C., Gaspard, C. E., Jaubert, G., et al. (1998). Growth and enterotoxin production of Staphylococcus aureus during the manufacture and ripening of Camembert-type cheeses from goats’ milk. Journal of Applied Microbiology, 85, 537–544. Modi, R., Hirvi, Y., Hill, A., & Griffiths, M. W. (2001). Effect of phage on survival of Salmonella Enteritidis during manufacture and storage of Cheddar cheese made from raw and pasteurized milk. Journal of Food Protection, 64, 927–933. O’Flaherty, S., Coffey, A., Meaney, W., Fitzgerald, G. F., & Ross, R. P. (2005a). The recombinant phage lysin LysK has a broad spectrum of lytic activity against clinically relevant staphylococci, including methicillin-resistant Staphylococcus aureus. Journal of Bacteriology, 187, 7161–7164. O’Flaherty, S., Coffey, A., Meaney, W. J., Fitzgerald, G. F., & Ross, R. P. (2005b). Inhibition of bacteriophage K proliferation on Staphylococcus aureus in raw bovine milk. Letters in Applied Microbiology, 41, 274–279. O’Flaherty, S., Ross, R. P., Meaney, W., Fitzgerald, G. F., Elbreki, M. F., & Coffey, A. (2005). Potential of the polyvalent anti-Staphylococcus bacteriophage K for control of antibiotic-resistant staphylococci from hospitals. Applied and Environmental Microbiology, 71, 1836–1842. O’Flynn, G., Ross, R. P., Fitzgerald, G. F., & Coffey, A. (2004). Evaluation of a cocktail of three bacteriophages for biocontrol of E. coli O157:H7. Applied and Environmental Microbiology, 70, 3417–3424.
ARTICLE IN PRESS P. Garcı´a et al. / International Dairy Journal 17 (2007) 1232–1239 Resch, A., Fehrenbacher, B., Eisele, K., Schaller, M., & Go, F. (2005). Phage release from biofilm and planktonic Staphylococcus aureus cells. FEMS Microbiology Letters, 252, 89–96. Sambrook, J., Maniatis, T., & Fritsch, E. F. (1989). Molecular cloning: A laboratory manual (2nd ed.). Cold Spring Harbor, NY, USA: Cold Spring Harbor Laboratory Press. Sass, P., & Bierbaum, G. (2007). Lytic activity of recombinant bacteriophage F11 and F12 endolysins on whole cells and biofilms of Staphylococcus aureus. Applied and Environmental Microbiology, 73, 347–352.
1239
Sua´rez, J. E., & Chater, K. F. (1981). Development of a DNA cloning in Streptomyces using phage vectors. Ciencia Biologica, 6, 99–110. Sulakvelidze, A., Alavidze, Z., & Morris, J. G., Jr. (2001). Bacteriophage therapy. Antimicrobial Agents and Chemotherapy, 45, 649–659 Review. Tirado, C. & Schmidt, K. (2000). WHO Surveillance Programme for Control of Foodborne Infections and Intoxications in Europe. Seventh Report /http://www.bfr.bund.de/internet/7threport/7threp_fr.htmS. Wills, Q. F., Kerrigan, C., & Soothill, J. S. (2005). Experimental bacteriophage protection against Staphylococcus aureus abscesses in a rabbit model. Antimicrobial Agents and Chemotherapy, 49, 1220–1221.