In modern dairy fermentation facilities, plant design, bacterial starter cultures, sanitation regime, equipment and materials are all carefully chosen, monitored and controlled in order to limit production inconsistencies and achieve optimal product output. In such processes, the incoming raw milk is often subjected to thermal treatments, such as pasteurisation to reduce the microbial load followed by the deliberate introduction of selected starter bacteria or mixed cultures which acidify the milk and impart desirable organoleptic attributes upon the product. In contrast to this, traditional and artisan cheese manufacturing processes employ a diverse array of conditions in order to adhere to traditional production practices and/or to acquire “protected designation of origin” (PDO) status. For example, in some cases fermentations may rely solely on the acidifying activity of the autochthonous microbiota of the fermentation environment or wooden vessels for the production of cheeses with particular organoleptic and physical properties. In other cases, defined cultures and/or stainless steel vats may be employed while still retaining certain aspects of the traditional fermentation practices to achieve typical flavours and/or textures [1
]. The variety of conditions under which such artisan cheeses are produced creates opportunities for phage proliferation. Furthermore, the production of certain traditional and regional artisan cheeses involves the application of raw (i.e., unpasteurized) milk and animal-derived rennet to increase the rate of coagulation of the curd (as opposed to industrially produced, recombinant enzymes). The traditional Sicilian cheeses PDO Vastedda della Valle del Belìce (referred to herein as Vastedda), Canestrato Siciliano (Canestrato), PDO Pecorino Siciliano (Pecorino) and Caciocavallo Palermitano (Caciocavallo) may be classified among such artisanal cheeses as they are regional cheeses that employ raw sheep, goat and cow’s milk in their production and often incorporate animal-derived rennets. Vastedda and Caciocavallo are stretched cheeses, while Pecorino and Canestrato cheeses are pressed cheeses and Pecorino additionally undergoes a cooking step after moulding. Additionally, wooden vats are traditionally used in the production of these cheeses, a practice that further contributes to the microbiota of the fermentation [1
]. Beyond the biological hurdles presented in modern production facilities and procedures, the only major limitation to phage proliferation in traditional practices is the availability of a suitable host.
Lactococcal phages are currently classified into ten taxonomic groups based on morphology and nucleotide sequence relatedness [4
]. Among these, several phage isolation studies have reported that members of the 936, c2 and P335 groups are the most frequently encountered in the dairy fermentation environment [5
]. In contrast, members of the remaining seven lactococcal phage groups are much less frequently encountered [4
]. The description of a lactococcal phage with a distinctively long tail in 1951 [12
] marked the first observation of what is now called the 949 group of lactococcal phages. This isolate was shown to exhibit a tail with an estimated length of 560–610 nm, which far exceeds the typical tail length of lactococcal Siphoviridae
phages, which normally ranges in the region of 150–200 nm [4
]. More than 60 years later, the first genome sequence of a member of this group was published [13
], revealing a genome of 114,768 kb with limited homology to other lactococcal phages. Subsequently, two additional members of this group have been genomically characterised, namely phiL47 and WRP3 [11
], which were isolated from grass and a Sicilian cheese whey, respectively. The latent period of 949 (period during which phages replicate and assemble inside the host cell), was calculated to be approximately 70 min, which may in part explain the irregularity of its apparent occurrence [8
]. The P087 group has currently only one described member, its namesake P087 [15
], representing a Siphoviridae
phage with a broad appendage at its distal tail end, called a baseplate, which is reminiscent of those of the sub-group II P335 phages Tuc2009 and TP901-1 [17
]. The genomes of five members of the 1706 group of lactococcal phages have been sequenced recently and significant diversity is observed between the groups namesake 1706 and the more recently sequenced members (P118, P078, P092 and P162) bearing at most 49% nucleotide homology [22
]. These 1706 phage group members were isolated from raw milk (P118, P078, P092 and P162) and cheese (1706). For the remaining four rare lactococcal phage groups (P034, Q54, KSY1 and 1358), only one representative genome sequence is available, impeding in-depth analysis of the diversity of these phage groups [10
The current study reports on the isolation of a variety of lactococcal phages from the 936, P087 and 949 groups from cheese whey and rennet samples associated with the production of artisanal cheeses in Sicily. The genomes of two 936 as well as six P087 and ten 949 group phages were sequenced and compared to previously reported members of their respective groups. This study highlights the significance of hurdle technology in reducing microbial loads in the factory environment and the observed sensitivity of the rare phages to thermal treatments such as pasteurisation. Furthermore, the potential sources of phages in the specific artisanal production environment are discussed.
In traditional cheese fermentations, an array of cheese- and region-specific practices are employed to produce cheeses with particular flavours and textures. While defined cultures may sometimes be employed in the production of artisan cheeses, in many cases the autochthonous bacteria present in the fermentation environment are responsible for the acidification and development of the derived cheese. In such situations, the microbial landscape is continuously in a state of flux as phage populations infect susceptible strains within the complex culture and resistant strains become dominant until the phage population adapts to infect the previously dominant strains in turn. Additionally, microbial biofilms established within the wooden fermentation vats and the incoming raw milk contribute to the microbial ecology of these fermentations [3
]. As the starter cultures are undefined and complex, while also subject to change and factory-specific, it is impossible to precisely define the phage population by culture-dependent methods. An additional variable is the application of pasteurized or unpasteurized milk in the production of artisan cheeses. The current study was aimed at exploring the diversity of lactococcal phages present in a range of samples derived from the production of traditional Sicilian cheeses using a test panel of 25 lactococcal strains that have been employed in previous phage-host interaction studies to gain an insight into the overall phage population [11
]. The samples were taken from whey and rennet used in the production of Canestrato, Caciocavallo, Pecorino and Vastedda cheeses produced in stainless steel (Canestrato only) and wooden vats (Caciocavallo, Vastedda and Pecorino). Through this process, a total of 59 individual plaque isolates were propagated and selected for further characterisation. Using established multiplex PCR approaches [31
], eight isolates were identified as belonging to the 936 group (Table 1
) while the remaining 51 could not be typed according to these methods.
To understand the genetic diversity of the isolated phages and to assign the untyped phage isolates to a lactococcal phage group, eighteen isolates were selected for genome sequencing based on sample source, unique/overlapping host range and/or restriction profile. Among these were two randomly selected 936 group phages isolated from lamb rennet used in Vastedda cheese production (Table 1
), while the remaining 16 isolates were of unknown phage groups. Genome sequencing revealed that the untyped isolates comprised six P087 group isolates and ten 949 group isolates (Table 2
). This finding is in complete contrast to several phage isolation studies that have been performed in recent years in modern global fermentation facilities [5
], which revealed that members of the 936 and P335 phage groups dominate. The apparent relative dominance of 949 and P087 group phages in this study may be biased (i) by the strains naturally present in the starter culture in the fermentation, (ii) the raw milk starting material, (iii) by the application of natural rennets, which we have shown to contain phages, or (iv) by the laboratory strains used in the selection process; however, this remains a remarkable finding among phage isolation studies.
Since it is proposed that the so-called rare lactococcal phages are more often associated with raw milk and that longer tailed phages such as those represented by members of the 949 and P087 groups are susceptible to physical damage [4
], it may thus be expected that these phages may be sensitive to thermal treatments that are employed in dairy fermentation processes. To examine this hypothesis, representative 936, P087 and 949 phage isolates from this study were subjected to thermal treatments. The ability of the phages to resist treatments of 63 °C × 30 min and 83 °C × 10 min was assessed and while the 936 isolates (R31 and R3.4) were relatively stable when treated at 63 °C and exhibited approximately 3-log reduction in infectivity at 83 °C, the 949 isolates showed complete loss of infectivity at the higher temperature in most cases while the P087 isolates exhibit approximately 5-log reductions in titres. The increased thermal sensitivity of these phages may be at least part of the reason that these phage groups are not regularly encountered in modern fermentation facilities where thermal processes are frequently employed.
In this study, phages were isolated from the animal derived rennet samples. This, to our knowledge, is the first description of phages isolated from rennet and it appears to act as an additional reservoir of phages in traditional processes where animal rennets as opposed to purified, recombinant enzymes of microbial origin are employed to expedite the milk coagulation and separation processes. The finding of 936 type phages in the rennet samples also serves to highlight the ability of (some) phages to withstand the relatively low pH (3.5–4) of the abomasum of the young ruminants from which the rennet is extracted. While 936 phage isolates were isolated from the rennet samples in this study, it is likely that other lactococcal phage groups also reside there but were not detected in this study, possibly due to (i) the strain selection applied, (ii) low phage numbers that may require pre-enrichment in order to be detected or (iii) a progressive inactivation due to prolonged exposure to low pH.
From this study, it is clear that the traditional and artisanal cheese production processes permit the proliferation and even an apparent dominance of so-called rare lactococcal phage groups. However, the phages isolated in this study undoubtedly represent only a proportion of the total phage population based on the collection of strains available at the time of processing. The strains that may be present within these “wild” fermentations may be considerably more diverse. Therefore, it is clear that while culture-dependent methods of analyzing such phage populations may provide a “snapshot” of the diversity of phages within such processes, culture-independent methods such as metagenomic analyses would be a useful complement to such studies.
Of the 25 lactococcal strains employed in this study, a small number are particularly sensitive to a number of phages in the collection, such as L. lactis
3107 and SMQ-86 (Table 4
). The sequenced 936 group isolates, R31 and R3.4 infect both of these strains. The RBPs of these phages, which dictate host recognition and binding, bear most significant similarity to those of the RBP subgroup I phages, phi7 and JM2 [29
]. In our previous study, phages exhibiting RBPs of subgroup I were capable of infecting host strains with a CWPS C type only, while in this study both phages infect L. lactis
3107 and SMQ-86, which possess CWPS C and A types, respectively [29
]. This finding highlights the adaptability of the 936 phages to the evolving landscape of available hosts in its environment. Consistent with previous studies of 949 phages, the 949 isolates in this study displayed a relatively broad host range with typically more than five strains infected. The only exceptions to this general observation were phages AM4, AM5 and AM12. However, the panel of strains in this study may not be reflective of their natural host range thus reducing the apparent host range of these isolates. It is noteworthy that the 949 phages infect more CWPS C type strains than those of CWPS A or B types while the P087 isolates infect several strains of CWPS A and B types (Table 4
). These findings indicate that the fermentation environment, from which the phages were derived, contain a wide diversity of lactococcal strains (expressing a variety of CWPS types) and that the phage groups dominantly isolated from particular factories may provide an insight into which lactococcal CWPS type strains are dominant at a given time.
In summary, the artisanal cheese production process incorporates several points at which phages may be introduced, propagated, and distributed to the next production cycle. The application of raw milk, animal-derived rennets, minimal processing combined with the use of wooden vats and back-slopping (if used) introduce the possibility of increasingly complex microbial populations. This is reflected in the identification of a significant collection of 949 and P087 phage isolates in whey samples derived from Sicilian artisanal cheese production facilities in the present study. The heat labile nature of the 949 and P087 isolates in this study relative to the 936 isolates serves to further highlight the adaptability of the 936 phages to modern fermentation facilities while simultaneously explaining the “rarity” of the 949 and P087 phages.