Bacteriophages (phages) are viruses that infect bacteria. Bacteriophagic infection of starter cultures can result in failure of the fermentation and loss of product. Whitehead and Cox (1935) first recognized bacteriophagic infection as a cause of failure of single-strain starter cultures used for Cheddar cheese production. Excellent conditions for development of bacteriophages were created in the 1950s when cheese production increased, resulting in more intensive use of facilities and preparation of larger amounts of lactic cultures (Huggins, 1984). Despite implementation of control measures, bacteriophagic infection still causes production problems in the modern dairy fermentation industry. Adoption of control strategies based on the use of lactic acid bacteria genetically engineered for bacte-riophagic resistance should provide substantial improvements in dependability of starter cultures (Dinsmore and Klaenhammer, 1995).
Bacteriophages that infect lactic acid bacteria usually consist of a head and tail section. The head can be either isometric or prolate (Fig. 2). An isometric head consists of 20 equal-size proteins that form an icosohedron. A prolate head has elongated side units. Phage DNA is enclosed by head proteins. Phages attach to the host by their tail sections, through which DNA passes into bacteria. Tail sections are of variable length and may have collars, sheaths, and base plates. Base plates can be seen at the end of the tail of the phage illustrated in Fig. 3.
Bacteriophages of lactic acid bacteria can be classified by morphology, serology, and DNA-DNA homology. These classification criteria generally produce consistent groupings (Lodics and Steenson, 1993). Six morphological types of lactic phages are commonly encountered. These include small isometric, collared small isometric, short-tailed small isometric, long-tailed small isometric, large isometric, and prolate (Lodics and Steenson, 1993). Each morphological type may include several distinct genotypes of which there are 12 (Neve, 1996).
Figure 3 Electron micrograph of isometric phage of Lactococcus lactis. (From Moineau et al., 1994).
Bacteriophages of S. thermophilus form one homologous grouping as opposed to bacteriophages of mesophilic lactococci and Lb. delbrueckii, which are genetically diverse (Jarvis, 1989; Brussow et al., 1994).
and is illustrated in Fig. 4. Phagic infection is initiated by adsorption of the virion onto the surface of the host cell. Only bacteria with specific adsorption sites serve as hosts for the bacteriophage; the presence of these sites determines to a great extent the host range of a particular phage. Recognition of an appropriate site and adsorption to it are mediated by the base plate, spikes, or fibers at the end of the phage tail. Many phages require Ca2+ for adsorption.
After adsorption, the phage injects its DNA into the host. The DNA passes from the head through the tail into the bacterial cell while the ''empty'' virion remains outside. Normal metabolism of the infected cell then ceases as the host first replicates phage DNA and then phage proteins. This process, called maturation, ends with self-assembly of virions within the host cell. Initially, heads form around viral DNA followed by attachment of tails. Finally, the lytic cycle is completed when a lytic enzyme (lysin), encoded on viral DNA, is produced, resulting in cell lysis and release of infective phagic particles into the surrounding environment. Lysin released from infected cells can also lyse noninfected cells. The time from initial adsorption to release of phages is called the latent period. For lactococcal phages, this period ranges from 10 to 140 mins. The number of virulent particles released per infected cell is called the burst size. This ranges from less than 10 to more than 300 for lactococcal phages (Klaenhammer and Fitzgerald, 1994).
c. Temperate Cycle Infection with a temperate phage does not necessarily lead to immediate production of new virions. DNA of a temperate phage may instead be incorporated into the chromosome of the host cell or maintained as a plasmid within the cell (Cogan and Accolas, 1990). This DNA, referred to as a prophage, replicates with the bacterium without affecting its metabolism. The resulting condition, lysogeny, is common in lactococci (Davidson et al., 1990) and lactobacilli (Sechaud et al., 1988), rare in S. thermophilus (Brussow et al., 1994), and unreported in Pediococcus, Leuconostoc, and Propionibacterium spp. (Davidson et al., 1990). Lysogeny can be maintained indefinitely. Lysogenous bacteria are immune to the infecting and other closely related phages. They maintain the potential to produce virulent phages and can spontaneously realize this potential. Phage production can also be induced by exposing cells to ultraviolet (UV) light or mitomycin C to inactivate the repressor protein that blocks expression of growth genes (Lodics and Steenson, 1993).
The extent to which lysogenic bacteria in starter cultures pose a threat to industrial fermentations is still uncertain (Jarvis, 1989; Davidson et al., 1990). Temperate phages can mutate to become virulent, resulting in fermentation failure (Shimizu-Kodata et al., 1983), although spontaneous induction of virulent phages from lysogenic strains appears to be rare (Teuber and Lembke, 1983). Surveys of lactococcal phage DNA homology indicate that, although some lytic phages appear to be variants of temperate phages, this is generally not true (Davidson et al., 1990).
Phage resistance in lactic acid bacteria is based on at least four different natural mechanisms (Hill, 1993; Dinsmore and Klaenhammer, 1995; Allison and Klaen-hammer, 1998): adsorption inhibition, DNA injection inhibition, DNA restriction and modification systems, and abortive infection. Stages in the lytic cycle where these mechanisms are active are illustrated in Figure 4. Many lactococci used in starter cultures exhibit one or more of these resistance mechanisms. Adsorption inhibition is the failure of phage to attach to the bacterial surface. This can result from spontaneous mutation modifying the attachment site or from a plasmid-linked factor (Dinsmore and Klaenhammer, 1995). Plasmids can encode for production of polymers that coat attachment sites, preventing phage adsorption.
DNA injection inhibition occurs when phage adsorbs to the cell surface but phage DNA stays inside the head section and fails to enter the host cell cytoplasm. This resistance mechanism appears to be rare (Dinsmore and Klaenhammer, 1995). A plasmid-encoded injection-blocking system in Lactococcus was first described by Garvey et al. (1996). They concluded that DNA injection inhibition resulted from an alteration in plasma membrane components of the host cell.
Phage resistance based on DNA restriction and modification enzymes (R/ M) is common in lactococci. The restriction enzyme hydrolyzes phage DNA at a specific site. Host DNA is modified by methylation at this site and is therefore unaffected by the restriction enzyme. Restriction and modification enzymes are linked to the same plasmid. It is possible, but rare, for phage DNA to be methylated by the host modification system before it is hydrolyzed by the restriction enzyme. When this happens, the phage is able to cause a normal infection. Phages whose DNA does not contain the targeted restriction site are also unaffected by this resistance mechanism. Four groups of R/M can be distinguished based on their enzyme structures and cleavage characteristics (Forde and Fitzgerald, 1999).
Abortive infection is a type of phage resistance resulting in decreased production of virulent phages by infected cells but not involving restriction or modification. Abortive infection results in cell death, but because phage replication is much reduced, the phage population does not increase sufficiently to affect culture activity. Abortive infection does not induce genetic changes in the infecting phage. Numerous (at least seven) nonhomologous plasmids encode for abortive infection resistance, indicating that many different types exist (Dinsmore and Klaenhammer, 1995; Neve, 1996).
When a host cell with phage resistance is exposed to sufficiently high numbers of phages, it is possible for the phage to mutate to overcome the resistance mechanism. Also, if phage inhibition is not complete, resistant phages are selected (Hill, 1993). If phage DNA is modified by the host enzyme to become resistant to the restriction enzyme, resulting resistance is lost when the phage infects a cell that lacks the methylase enzyme. More lasting insensitivity occurs when phages mutate at the hydrolysis site of the restriction enzyme. Some lacto-coccal bacteriophages have evolved to have very few sites available for restriction endonuclease hydrolysis (Dinsmore and Klaenhammer, 1995). Phages also develop insensitivity to abortive injection mechanisms, apparently through point mutations.
Many bacteriophages have good survival characteristics. Some can survive high-temperature, short-time pasteurization, so media for starter preparation are usually heated to at least 85°C for 30 mins to ensure inactivation of the phage (Neve, 1996). Phages can also survive spray drying and storage of milk powder (Chopin, 1980). Phagic particles on surfaces are readily inactivated by chlorine but not by iodine or acid sanitizers (Anonymous, 1990). Sanitizer inactivation depends on elimination of organic matter through effective cleaning.
Bacteriophages are primarily a problem in cheese manufacture. This is probably because cheese milk (as compared to cultured milks) is given only a mild heat treatment and because cheese milk and whey are often exposed to a phage-con-taminated environment. Bacteriophages do not proliferate in cheese curd, because virions cannot move through the protein matrix. However, cells infected with phage before coagulation become inactivated during cheese manufacturing. Because latency periods are normally approximately 30 min (but may be much longer), a culture may initially show normal growth in cheese milk but then reduce or stop acid production during manufacture. If one culture preparation is used to inoculate a series of vats of milk, increasing numbers of phages active against this culture may develop within the manufacturing plant. The result is that acid production proceeds normally in the vats of milk inoculated initially but is delayed later in the production day.
Preventing inhibition of acid production resulting from phagic infection requires implementation of control measures throughout the manufacturing process. These should include selection, preparation, and maintenance of cultures free of virulent phage, controlling entry of phages into the processing facility, and controlling spread of phages within the facility.
Growth of phages during production of bulk starter can be controlled by using phage-inhibitory media. These media rely on the ability of phosphate and citrate salts to bind ionic calcium, thus inhibiting phagic absorption (Reiter, 1956). The chelating agents can slow growth of the starter culture. Phage-control media often contain deionized whey, protein hydrolysates, ammonium and sodium phosphate, citrate salts, and other growth stimulants such as yeast extract (Whitehead, 1993). Commercial phage-inhibitory media vary widely in their ability to prevent phage proliferation; the most effective being those that contain sufficient nutrients to overcome the inhibitory nature of the media and contain citrate buffers (Gulstrum et al., 1979). Not all bacteriophages are inhibited by the absence of calcium (Sozzi, 1972; Quiberoni and Reinheimer, 1998), so, to be effective, phage-inhibi-tory media should be used as only one part of an overall phage-control strategy. Proliferation of phages during starter preparation can also be avoided by using cell concentrates designed to be added directly to cheese milk in the vat or by preparing cultures under strict aseptic conditions.
Lactic acid bacteria vary widely in their susceptibility to bacteriophagic infection, so the use of resistant strains is an important aspect of phage control. Phage-resistant strains have been isolated from mixed-culture systems that maintain activity while carrying low levels of phages (Lodics and Steenson, 1993). Strains can also be genetically altered to contain plasmids coding for phage resistance (Klaenhammer, 1991). Phage-resistant variants can be selected by exposure to factory whey containing phages that have developed during cheese manufacture (Sandine, 1989). Resistant variants are tested for rapid acid production and added back to the starter in use in that factory. The use of such a system requires daily monitoring of whey for phages, but it allows the use of a single mixture of five or six defined strains over a long time. This approach to phage control is often used in North America and elsewhere.
Protease-negative strains of lactococci are resistant to phagic infection because of their slow growth rates (Richardson, 1984). Although more cells must be used to compensate for lack of growth during cheese manufacture, these variants offer other advantages, including lowered sensitivity to antibiotics, lowered heat sensitivity (allowing the use of higher cook temperatures), greater yield because of lowered casein solublization, and decreased risk of bitter flavor development in cheese.
Exopolysaccharide-producing strains are more resistant to phage (Moineau et al., 1996). Phages that infect and lyse strains producing exopolysaccharide possess a polysaccharide depolymerase enzyme specific for this particular exopolysaccharide (Hughes et al. 1998).
Culture rotations control bacteriophagic infection by limiting the length of time that a specific strain or mixture of strains is used. Cultures following each other in the series are susceptible to different phage types and are therefore unaffected by phages that may have infected the previous culture. Cultures can be rotated on a daily basis or after each vat of milk is inoculated. Short rotations over 23 days using 6-12 strains and long (5-10 days) rotations of up to 30 strains are used (Huggins, 1984). However, the use of a limited number of cultures at any one time is recommended to reduce exposure to prophages and maintain product uniformity. Culture rotation does not eliminate phage growth in cheese milk in vats, but if phage numbers are kept to less than 10,000 pfu/mL of cheese whey, acid production is not affected (Huggins, 1984). Success of a culture rotation is limited by availability of phage-unrelated strains with acceptable fermentation properties. In addition, using many different cultures can result in lack of product uniformity.
A new type of culture rotation system has been developed by Sing and Klaenhammer (1993) and Durmaz and Klaenhammer (1995). This system uses genetic derivatives of a single strain, each with a different phage-resistance mechanism. When used in rotation or as mixtures, resistant phages fail to develop, because they cannot overcome the multiple resistance mechanisms. This type of rotation avoids the lack of product uniformity associated with conventional culture rotations and allows continuous use of strains with special properties.
O' Sullivan et al. (1998) stacked three plasmids encoding distinct phage resistance mechanisms (adsorption inhibition, R/M, and Abi) in addition to the lactose proteinase plasmid to generate a host with phage resistance and acceptable fermentation characteristics. This isogenic single-strain starter rotation system in which complementary defenses are rotated within one starter limits exposure of phages to any single defense mechanism.
Since phage-resistance plasmids are transferrable by conjugation, application of genetic engineering technology can introduce industrially significant phage-resistance starter strains (Coakley et al., 1997; Allison and Klaenhammer, 1998; O'Sullivan et al., 1998). However, the evolutionary capacity of phages which allows their genetic modules to be exchanged in addition to the presence of lyso-
genic starter cultures show the need for continuous development of novel phage-insensitive mechanisms and strains (Forde and Fitzgerald, 1999).
Bacteriophages in the dairy plant probably are of farm origin, although, as discussed previously, lysogenic bacteria may also be a source. Although the major means by which a phage enters the plant is in raw milk; trucks and personnel having had contact with the farm environment could also be carriers. After monitoring a mozzarella cheese factory for 2 years, Bruttin et al. (1997) postulated a single phage invasion event and diversification of the phage during its residence in the factory. They then introduced a defined starter system that could not propagate the resident factory phage population. It is not practical to eliminate entry of phages into the dairy plant, because raw milk continually enters the facility. However, growth of phages within the plant and dissemination of phages to milk in the cheese vat can be controlled. Bovine colostrum may have antibodies that could protect Lc. lactis strains from phage attack (Geller et al., 1998). The main growth niches for bacteriophages in a cheese plant are raw milk, whey, spilled product, pools of water, stagnant floor drains, equipment, and soiled walls (Anonymous, 1990). Phage development in these growth niches is controlled by effective sanitation. Phages are disseminated throughout the dairy plant by aerosol and human carriers. Air entering cheese manufacturing rooms should be under positive pressure of high-efficiency particulate air (HEPA) filtered air. When preparing bulk starter, air drawn into the tank when the culture medium cools should be filter sterilized. Milk in cheese vats is most susceptible to phage contamination during ripening and setting, so these processes should be accomplished in closed systems. Whey should be removed to a physically separate facility, because whey processing produces aerosols that can carry phage particles. Plant personnel with exposure to whey should not be allowed access to the milk-ripening or bulk starter facilities.
VI. OTHER CULTURE INHIBITORS A. Raw Milk-Associated Inhibitors
Lactic starter cultures grow more slowly in raw than in heated milk; a phenomenon caused by the presence of natural inhibitors. The lactoperoxidase system is the most significant microbial inhibitor in raw milk, but the presence of aggluti-nins is an important problem in acid-coagulated cheeses. Other naturally occurring microbial inhibitors in milk include lysozyme and lactoferrin. Mastitic milk has increased levels of microbial inhibitors and increased phagocytic activity that are part of the cow's response to infection. However, mastitic milk is also higher in protease activity, and the resulting casein fragments can counteract inhibitor effects and even stimulate growth of weakly proteolytic lactics such as S. thermophilus (Marshall and Bramley, 1984; Okello-Uma and Marshall, 1986).
Microbial inhibition by the lactoperoxidase system derives from interaction of three components: lactoperoxidase, an enzyme native to milk; thiocyanate, derived from hydrolysis of cyanogenic glucosides found in certain feeds; and hydrogen peroxide, generated by leukocytes and through oxygen metabolism of lactic acid bacteria (Limsowtin, 1992). The inhibitor, hypothiocyanite, is produced when lactoperoxidase catalyzes oxidation of thiocyanate and simultaneous reduction of hydrogen peroxide. Bovine colostrum and milk contain about 11-45 mg/ L and 13-30 mg/L lactoperoxidase, respectively (Korhonen, 1977). Hydrogen peroxide is usually the limiting component in raw milk, but thiocyanate is also often present in suboptimal concentrations (Limsowtin, 1992). Lactoperoxidase is only partially inactivated by pasteurization (Wolfson and Sumner, 1993). However, more severe pasteurization temperatures (80°C for 15 s) will completely inhibit the lactoperoxidase system. This might explain why sometimes milk pasteurized at 72°C exhibits better keeping quality than that pasteurized at higher temperatures (Barrett et al., 1999). The lactic starter cultures most sensitive to lactoperoxidase inhibition are those that generate hydrogen peroxide. This includes some strains of Lb. delbrueckii subsp. bulgaricus and Lb. acidophilus (Guirguis and Hickey, 1987b). Other lactic acid bacteria, including S. thermophi-lus and some strains of lactococci, are sensitive to lactoperoxidase inhibition when combined with cultures that produce hydrogen peroxide. The inhibitory effects of the lactoperoxidase system can be controlled by limiting aeration of milk, avoiding the use of hydrogen peroxide-generating cultures, using cultures that degrade hydrogen peroxide, and using heat treatments more severe than pasteurization. Lactoperoxidase activity suppresses acid production in yogurt during refrigerated storage and produces product having a softer texture (Nakada, et al., 1996; Hirano et al., 1998).
2. Immunoglobulins (Agglutinins)
Bovine milk contains four types of immunoglobulins: IgG1, IgG2, IgM, and IgA at concentrations of 0.3-0.4, 0.03-0.08, 0.03-0.06, and 0.04-0.06 g/L, respectively (Pakkanen and Aalto, 1997). Lactic starter cultures can interact with immu-noglobulins in milk to form aggregates or clumps. As the cells produce acid, casein coagulates around these clumps and they settle out of the milk forming a sludge (Grandison et al., 1986). Acid production is inhibited, because diffusion of acid out of the sludge is limited, causing acid inhibition of the culture before the milk is properly acidified (Hicks and Ibrahim, 1992). This type of inhibition is of significance when acid coagulation is desired, as for cottage cheese, which exhibits a loss of curd. Culture agglutination can be reduced by selecting agglutination-resistant cultures, using whey-based culture media with agglutinins removed by protease treatment (Ustunol and Hicks, 1994), homogenization of milk before culturing (Hicks and Hamzah, 1992), and homogenization of the starter culture (Hicks et al., 1998). Susceptibility of starter cultures to bind milk immu-noglobulins can be determined by using an enzyme-linked immunosorbent assay (ELISA) (Ustunol and Sypien, 1996).
Lysozyme inactivates bacteria by cleaving the glycosidic bond between N-ace-tylmuramic acid and N-acetylglucoseamine in the peptidoglycan of the cell wall. Gram-positive bacteria are highly susceptible to lysozyme activity because of the high peptidoglycan content of their cell wall and a lack of protective lipopolysac-charide. Bovine milk contains only approximately 0.07-0.6 mg/L (Korhonen, 1977).
Lactoferrin is an iron-binding protein that inhibits bacteria by denying their access to iron. Cow's milk contains only 20-200 |g/mL of lactoferrin (Masson and Heremans, 1971), and its activity is limited because it competes with citrate for binding iron (Batish et al., 1988). Inhibition of starter cultures by lactoferrin is unlikely to be significant.
Was this article helpful?