INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Lactococcus lactis is a mesophilic gram-positive bacterium widely used in the manufacture of a variety of fermented dairy products. This lactic acid-producing bacterium contributes to the development of flavor and texture, while protecting the product from spoilage microorganisms (14). Because a diverse group of lytic lactococcal bacteriophages may be present in nonsterile dairy environments, industrial milk fermentation is prone to bacteriophage attacks (23, 26). Upon phage infection, the starter culture is rapidly lysed, causing a severe reduction in the rate of milk acidification and, consequently, an important economical loss for the industry (50).

Three genetically unrelated groups of lytic lactococcal phages (936, P335, and c2) are responsible for most large-scale fermentation failures (33, 39). The three groups share a few characteristics, including a double-stranded DNA genome and a long, noncontractile tail (Siphoviridae family). Both the 936- and P335-like phages have small isometric heads (morphotype B1) and are morphologically similar to phage  (Escherichia coli) and SPP1 (Bacillus subtilis). The c2-like phages have prolate heads (morphotype B2) and are unique among bacterial viruses. The c2 taxon is the only group of lactic acid bacteriophages ranked at the genus level by the International Committee on Taxonomy of Viruses (53).

To survive infection from a diverse phage population, some L. lactis strains possess plasmid-encoded antiphage barriers defined as phage resistance mechanisms. The antiviral systems can be transferred into phage-sensitive commercial strains to increase their resistance to these bacterial viruses. The defense systems are currently classified, based on their mode of action, into four groups (26). The first two systems act at the cell surface by blocking phage adsorption or DNA ejection. The other two mechanisms are intracellular processes that include restriction-modification systems and the abortive-infection proteins (Abi).

Typically, Abi proteins irreversibly shut down the phage lytic cycle, and the infected host is killed either by the Abi protein itself or by the upheaval caused by the infection (20, 43, 58, 59). This outcome limits phage dissemination and is seen by the absence or reduction of plaque size. In recent years, several lactococcal Abi have been characterized to the molecular level (AbiA [29]; AbiB [11]; AbiC [21]; AbiD [37]; AbiD1 [2]; AbiE and AbiF [25]; AbiG [47]; AbiH [51]; AbiI [61]; AbiJ [13]; AbiK [22]; AbiL, GenBank U94520; AbiN [52]; and AbiP, GenBank U90222). Most Abi systems are plasmid encoded. Few studies have contributed to unveiling the molecular basis of Abi systems, and our understanding is limited to their mode of action on small isometric phages (936 and P335 species). For instance, AbiA interferes with the DNA replication of small isometric phages (17, 18). In cells carrying AbiB, it appears that an early phage product induces the synthesis or stimulates the activity of an RNase (49). AbiC reduces the synthesis of structural phage proteins (21, 40). Finally, AbiD1 protein interacts with an isometric phage gene product (ORF1) to prevent the translation of the phage ORF3 RNA (5).

Natural Abi systems have already been introduced into industrial phage-sensitive L. lactis strains and exploited successfully in large-scale milk fermentations (56). The extensive use of the improved phage-resistant strains led to the emergence of mutant phages capable of bypassing the anti-phage systems (1). As for the antibiotics in the medical field, the search for novel anti-phage barriers is still a priority for the dairy sector. New systems should be tested for efficiency against members of the three phage groups (936, c2, and P335).

Presented here is a novel Abi from L. lactis that is efficient against prolate c2-like and small isometric 936-like phages but not against small isometric P335-like phages. The impact of this Abi on the lytic cycle of these phage groups was evaluated at the microbiological and molecular levels.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains, bacteriophages, plasmids, and media. Bacterial strains, bacteriophages, and plasmids used in this study are listed in Table 1. E. coli was grown at 37°C in Luria broth (54). L. lactis was grown at 30°C in M17 (62) supplemented with 0.5% glucose (GM17), except for strain W-37, which was grown in M17 supplemented with 0.5% lactose (LM17). Calcium chloride was added to a final concentration of 10 mM in M17 media for propagation of lactococcal phages. When needed, antibiotics were added to the media for selection and plasmid maintenance as indicated for E. coli (50 µg of ampicillin, 10 µg of tetracycline, and 20 µg chloramphenicol per ml) and for L. lactis (5 µg of chloramphenicol and 5 µg of erythromycin per ml).

Bacteriophage propagation and microbiological assays. Phage propagation was performed as described previously (41). The phage titer was determined by the method of Jarvis (32). Efficiency of plaquing (EOP) and adsorption assays were performed as described by Sanders and Klaenhammer (55). Cell survival was assayed by the method of Behnke and Malke (3) by using a multiplicity of infection (MOI) of 5. One-step growth curves and centre of infection (COI) assays were performed as described previously by using an MOI of 0.2 (40). The burst size was estimated by dividing the average titer when phages reached the higher plateau by the average titer of the lower plateau on the one-step growth curves. The efficiency at which COI formed (ECOI) was calculated as follows: (the number of COI from the resistant strain)/(the number of COI from the sensitive strain) ×100.

DNA isolation, manipulation, and sequencing. Plasmid DNA was isolated from E. coli and L. lactis as described by Sambrook et al. (54) and O'Sullivan and Klaenhammer (48), respectively. Restriction and modification enzymes were used according to the manufacturer's recommendations (Boehringer Mannheim, Laval, Quebec, Canada). DNA manipulations were essentially carried out as described by Sambrook et al. (54). Competent E. coli cells were prepared and transformed with the Gene Pulser II apparatus as described by the manufacturer (Bio-Rad Laboratories, La Jolla, Calif.). The methods for preparing competent cells and for the electrotransformation of L. lactis have been described elsewhere (31). The DNA to be sequenced was cloned in pBluescript II (Stratagene, La Jolla, Calif.), and nested deletions were generated by using the Erase-a-Base kit (Promega, Madison, Wis.). Plasmid DNA of the selected mutants was purified with the Qiagen plasmid kit (Qiagen, Chatsworth, Calif.) and used in the sequencing reactions with the DyeDeoxy Terminator Taq sequencing kit. Both strands were sequenced with T7 and T3 primers. Products were separated on a model 373A automated DNA sequencing system (Applied Biosystems, Foster City, Calif.).

DNA and protein sequence analysis. The Genetics Computer Group sequence analysis software package was used to run standard analysis on DNA and putative proteins deduced from the nucleic acid sequence (15). Searches were performed with GenBank release 97.0 (10/96), EMBL release 48.0 (9/96), PIR-Protein release 50.0 (9/96), SWISS-PROT release 33.0 (3/96), and PROSITE release 13.0 (12/95). The Propsearch Program (30) was used to find putative protein families that failed to show homology with programs based on primary sequence comparisons (BLAST and FASTA). The putative ribosome-binding sites were identified by alignment with the 3' end of L. lactis 16S rRNA (3'-UCUUUCCUCCA [35]), and the free energy was calculated by the method of Freier et al. (24).

Phage DNA replication assay. The method of Hill et al. (28) was used for phage infection (MOI = 1), sampling, and total DNA purification. The DNA was digested with EcoRV and heated at 65°C for 10 min, and the DNA fragments were immediately separated by electrophoresis on 0.8% agarose gel. DNA fragments were stained with ethidium bromide, photographed under UV illumination, and then transferred to Hybond-N nylon membranes (Amersham, Oakville, Ontario, Canada) by capillary blotting (60). Probes were prepared by labeling EcoRV fragments from the phage genomes with the DIG-High Prime kit (Boehringer Mannheim). Prehybridization, hybridization, and posthybridization washes, as well as detection, were performed as suggested in the "DIG system user's guide for filter hybridization" (Boehringer Mannheim). The standard hybridization buffer (50% formamide) and CSPD (Boehringer Mannheim) were used for the hybridization steps and chemiluminescent detection, respectively.

Phage major capsid protein production assay. Sensitive and resistant hosts were grown in GM17 to an OD₆₀₀ of 0.5. CaCl₂ was then added, followed by phage c21 at an MOI of 1. Infected cells were incubated at 30°C. At various time points, 1 ml of culture was collected in an Eppendorf tube, and cells were harvested by centrifugation. The pellets and supernatants were separated and rapidly frozen by immersing in 80% isopropanol at 80°C. After being thawed on ice, 1 ml of Tris-buffered saline (25 mM Tris, 15 mM NaCl, 3 mM KCl [pH 7.4]) was added to the samples containing pelleted cells. About 1 g of glass beads was added to the samples, and the cells were disrupted by vortexing at maximum setting (three times for 2 min) at 4°C on a multitube vortexer (model VX5000; Troemmer Philadelphia, Pa.). A portion (20 µl) from each sample was mixed with 90 µl of transfer buffer (25 mM Tris, 200 mM glycine [pH 8.4]) and spotted with a Hybridot apparatus (Gibco/BRL, Burlington, Ontario, Canada) onto a Hybond-N nylon membrane previously equilibrated in transfer buffer. After transfer, the membrane was equilibrated in a washing buffer (100 mM maleic acid, 150 mM NaCl, 0.3% Tween [pH 7.5]) and then incubated at room temperature for 2 h in a blocking solution (2% blocking reagent [Boehringer Mannheim], 100 mM maleic acid, 150 mM NaCl [pH 7.5]). The monoclonal antibody 2A5, which is specific for the native major capsid proteins of c2-like phages (10), was added to a final concentration of 4 µg/ml, and the mixture was further incubated at 37°C for 30 min. The membrane was washed three times for 10 min with a washing buffer and blocked as described above. The second antibody (anti-mouse alkaline phosphatase conjugate) was diluted 1:50,000 in a blocking solution and incubated with the membrane at 37°C for 30 min. The membrane was washed, and detection was performed with the chemiluminescent substrate CSPD.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

The phage resistance mechanism in strain W-37 is encoded by an 11-kb plasmid. L. lactis W-37 was previously isolated from raw milk and was found to be resistant to phage infection. This strain carries several plasmids that potentially encode the genetic determinants responsible for the phage resistance phenotype (Fig. 1, lane 2). Cryptic plasmids from W-37 were coelectroporated in the phage-sensitive strain L. lactis LM0230, along with the shuttle vector pSA3 (DNA mass ratio of 10:1). Colonies growing on GM17 plates containing erythromycin were randomly picked and tested for phage resistance. Several Em^r and phage-resistant isolates were identified. Plasmid DNA of these isolates was extracted, digested, and analysed by gel electrophoresis. The EcoRV restriction pattern of one representative showed, in addition to the 10.2-kb pSA3, two additional bands, one of 4.8 kb and another of 6.2 kb (Fig. 1, lane 6), corresponding to an 11-kb plasmid. This plasmid was named pSRQ900.

View larger version (90K): [in this window] [in a new window]   FIG. 1.   Identification of a plasmid encoding a phage resistance mechanism in L. lactis subsp. lactis W-37. Lane 1, supercoiled DNA ladder (Gibco/BRL); lane 2, L. lactis W-37; lane 3, L. lactis SMQ16(pSA3); lane 4, L. lactis SMQ21(pSA3+pSRQ900); lane 5, L. lactis SMQ16(pSA3)/EcoRV; lane 6, L. lactis SMQ21(pSA3+pSRQ900)/EcoRV; lane 7, 1-kb DNA ladder (Gibco/BRL).

The phage resistance system in pSRQ900 is very effective against 936- and c2-like phages. The effectiveness of the phage resistance mechanism carried by pSRQ900 was tested on phages belonging to the three main groups known to impede industrial fermentation, namely, 936, c2, and P335. The EOPs of three representatives of the 936- and the c2-like phages were reduced to <10⁸ on lactococcal cells carrying pSRQ900 (Table 2). Since the host used to test 936 and c2 phages was not sensitive to phages of the P335 species, a functional derivative of pSRQ900 carrying a selectable marker (pSRQ901 [see Table 1 for details]) was introduced into the strain L. lactis UL8. The phage resistance mechanism encoded by pSRQ901 was not effective against three P335 representatives (EOP = 1 [Table 2]).

The phage resistance system in pSRQ900 is heat stable. The effectiveness of pSRQ900 was assayed on phage p2 and c2 at three temperatures commonly used in industrial dairy fermentation. No increase in the EOP (<10⁸) was observed at 21, 30, and 38°C, indicating that the phage defense mechanism encoded by pSRQ900 was effective over this temperature range.

The anti-phage system carried by pSRQ900 is an abortive infection mechanism. Microbiological and biochemical experiments were conducted to determine the type of defense mechanism encoded by pSRQ900. These experiments were performed with the pSRQ901 derivative (Table 1) in the presence of erythromycin to ensure plasmid maintenance throughout the assays. The plasmid pSRQ900 does not encode an adsorption blocking mechanism since phage p2 reacted on both sensitive (SMQ-16) and resistant (SMQ-42) cells (80% adsorption). Adsorption of phage c21 (c2-like) was also not affected, since phages were adsorbed at a similar level (ca. 66%) on both SMQ-16 and SMQ-42 (Table 3). Thus far we have been unable to obtain phages capable of overcoming the defense mechanism encoded by pSRQ900. It was therefore impossible to exploit the classical approach for testing DNA modifications by restriction-modification systems (42). Alternatively, endonucleolytic activity was directly tested by incubating genomic DNA of phage p2 or L. lactis LM0230 with cell extracts prepared from SMQ-42 (42). Gel electrophoresis of the DNA incubated overnight with the extracts did not show endonucleolytic cleavage (results not shown). Finally, the presence of pSRQ901 had no impact on the survival of phage-infected L. lactis cells, indicating that the host still died upon infection (Table 3). Based on all the above results and the current classification of phage defense mechanisms, the system encoded by pSRQ900 was classified as an abortive infection mechanism and was named AbiQ.

AbiQ dramatically reduces the burst size and the ECOI of phages p2 and c21. The effectiveness of the Abi was assessed by comparing the burst size and the ECOI of the small isometric phage p2 and the prolate phage c21 on SMQ-16 and SMQ-42. When infected by phage p2, SMQ-16 released about 58 progeny phages. The addition of AbiQ to the host bacteria resulted in a dramatic reduction in phage progeny, as the burst size dropped to about 1 and the ECOI was virtually zero on SMQ-42 (Table 3). Similarly, the burst size of phage c21 was reduced about 30-fold when the phage was propagated on SMQ-42 (ca. three progeny phages) compared to SMQ-16 (ca. 89 progeny phages). The ECOI of phage c21 that propagated on SMQ-42 was reduced to 25% of that of SMQ-16.

The locus for AbiQ maps on a 2.2-kb EcoRV/BclI fragment of pSRQ900. In order to localize the genetic determinants encoding the Abi, the restriction map of pSRQ900 and a series of deletion mutants were generated. pSRQ900 was digested with EcoRI, EcoRV, HindIII, BclI, or NcoI, and the fragments were inserted into the shuttle vectors pSA3 or pMIG3. Constructions were introduced into E. coli, confirmed by restriction mapping, and finally introduced into L. lactis. Transformants were tested for their resistance to phages p2 and c21. The results presented in Fig. 2 clearly show that an EcoRV/BclI fragment of about 2.2 kb was necessary in order to confer the resistant phenotype to phages p2 and c21.

View larger version (13K): [in this window] [in a new window]   FIG. 2.   Localization and frameshift mutation of the phage defense mechanism carried by pSRQ900. The linear restriction map of pSRQ900 is shown on the top of the diagram, and the distance between two ticks corresponds to 500 bp. Putative orf, promoter (P), and terminators (T) shown on the restriction map were inferred from the sequence analysis presented in Fig. 3. Thick lines below the linear map represent pSRQ900 restriction fragments cloned into pSA3 or pMIG3. The recombinant plasmids containing those inserts are indicated on the left, and the numbers on the right represent the EOPs of phage p2 and c21 on cells carrying these plasmids. The asterisk indicates the location of the frameshift mutation on the EcoRV/BclI fragment of pSRQ933. The following enzymes did not cut pSRQ900: ApaI, AvaI, BalI, BamHI, HpaI, NruI, PstI, PvuI, SalI, ScaI, SmaI, SphI, SstI, XbaI, and XhoI.

Sequence and analysis of the DNA locus encoding the Abi in pSRQ900. The EcoRV/BclI fragment was sequenced on both strands. Of the three open reading frames (ORFs) identified, only one was complete. To confirm that the complete ORF was responsible for the phage resistance phenotype, a frameshift mutation was generated by digesting pSRQ928 with BglII, filling the ends with Klenow, and self-ligating the plasmid. The frameshift mutation completely abolished host resistance to phages p2 and c21 (Fig. 2). This result confirmed that the phage resistance phenotype is attributable to the complete ORF, which was named abiQ. The nucleotide sequence of abiQ was deposited in the GenBank database under the accession number AF001314. The nucleotide sequence presented in Fig. 3 corresponds to nucleotides 3954 to 5573 of the sequence submitted to the GenBank. The gene encoding AbiQ was preceded by a putative promoter (35 box [TTGCAT], 20-bp spacer, 10 box [TATAAT]) and a putative ribosome binding site (AAAG; G = 0.1 kcal/mol) located at a suitable distance from the translation start codon. A tandem repeat (2.5 repeats) located in the promoter region and an inverted repeat encompassing the ribosome binding site were also identified. In addition to the putative promoters shown in Fig. 3, we identified a number of other promoter-like structures (35 and 10 boxes) located within the direct repeats upstream of abiQ. Terminator-like structures were identified upstream (T1) and downstream (T2) of abiQ. We identified a GC rich region in the stem of terminator T1 and a stretch of T downstream of terminator T1 and T2, a finding characteristic of rho-independent terminators. As for other Abi, the GC content of abiQ (28.3%) is below the 37% average observed for L. lactis genes, which could reflect a common origin or function (47).

View larger version (62K): [in this window] [in a new window]   FIG. 3.   DNA sequence of the pSRQ900 fragment coding for AbiQ. The BglII restriction site used to generate the frameshift mutation in abiQ is indicated. Putative terminators, 35 boxes, 10 boxes, and ribosome binding sites are underlined. The free energy of terminator-like structures is given above the corresponding sequences. Tandem repeats and inverted repeats are shown underneath the sequence by thin and thick arrows, respectively.

Phage DNA is replicated in the presence of AbiQ. The DNA replication of the cos-containing prolate phage c21 and small isometric phage p2 was monitored at time intervals in sensitive and resistant cells. For phage c21, DNA appeared 20 min after infection in sensitive cells, peaked at 60 min, and then gradually decreased (Fig. 4). The reduction of the DNA concentration corresponded to cell lysis, as evidenced by a reduction of the optical density of the culture (results not shown). Both the replicative (circular and concatemer) and the encapsidated (mature) forms of phage DNA were observed in sensitive cells. The replicative form was ascertained by the appearance of a 20.6-kb EcoRV fragment, the product of the 13-kb and the 7.6-kb EcoRV fragments linked together through their cos termini. These results indicate that phage c21 genome was encapsidated and cleaved from concatemer DNA at a relatively early stage during the phage lytic cycle in sensitive L. lactis cells. Resistant cells also allowed phage DNA replication with a kinetic similar to that of sensitive cells. Phage DNA only accumulated in its replicative form, however, as shown by the absence of the 7.6- and 13-kb fragments.

View larger version (91K): [in this window] [in a new window]   FIG. 4.   Phage c21 and phage p2 DNA replication in sensitive (SMQ-16) and resistant (SMQ-21) lactococcal cells. Left lanes, control DNA extracted from mature phages (mature form); lanes 2, control DNA extracted from host prior to infection. The subsequent lanes in each panel contain DNA extracted from samples taken at different time points after phage infection. All of the samples were digested with EcoRV. Phages, hosts, and time of sampling are indicated above each lane.

Phage MCPs are produced in the presence of AbiQ. The effect of AbiQ on the production of virion structural proteins was investigated by using a monoclonal antibody specific to the major capsid protein (MCP) of c2-like phages (10). MCP was detected in cell extracts of SMQ-16 infected with phage c21 (Fig. 5). The signal rapidly increased over time, indicating that the MCP was readily produced after infection. Similarly, the signal increased in the supernatant as cells released progeny phages. SMQ-21 cells also produced MCP early after infection and the production increased over time. No signal increase was observed in the supernatant of resistant cells. These results indicated that the production of the MCP of phage c21 is not hindered by AbiQ but that the lytic cycle is blocked at a later stage.

View larger version (30K): [in this window] [in a new window]   FIG. 5.   Kinetics of production of phage c21 MCP in sensitive (SMQ-16) and resistant (SMQ-21) lactococcal cells. The host strain used in each assay is indicated on the left. CE, cell extracts; S, culture supernatant; c, host cell prior to phage infection.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Raw milk is a natural environment of L. lactis and its bacteriophages (36). The selective pressure exerted by the various phages on the cells occupying that niche favors the emergence of strains insensitive to phages or those having defense mechanisms. Based on these premises, we previously characterized a plasmid-encoded abortive infection mechanism (AbiK) expressed by L. lactis W-1, a strain isolated from raw milk (22). Here, we showed that another raw milk isolate, L. lactis W-37, also encodes a phage resistance mechanism (AbiQ) carried by the plasmid pSRQ900. Cells carrying pSRQ900 adsorbed phages normally, allowed phage DNA replication, showed no evidence of endonucleolytic activity, and still died upon infection. AbiQ was therefore classified as an abortive infection mechanism. The absence of homology to any sequenced or deduced proteins indicates that AbiQ is a novel antiviral protein. Evidence for possible regulatory elements in the putative promoter region and terminator structures upstream and downstream of abiQ strongly suggests that it is transcribed as a monocistronic mRNA and that its expression can be regulated. Further work is needed to study the regulation of AbiQ expression in relation to phage infection.

AbiQ is effective against two of the three genetically unrelated lytic phage taxons tested in this study. Other Abi are also effective against c2- and 936-like phages (AbiA, AbiD, AbiD1, AbiF, AbiG, AbiH, and AbiN), but none of these appears to be as strong as AbiQ against both groups. In the presence of AbiQ, the number of cells infected by p2 or c21 that release progeny phages is dramatically reduced, and in those instances where phages are released the number of progeny is virtually nil. In practical terms, it is very unlikely that even a high population of 936- or c2-like phages (10⁸/ml) will overpower AbiQ, unless a phage mutant circumvents the defense mechanism. Although numerous phage mutants insensitive to other Abi systems have been isolated in our laboratory, as well as in many others (5, 17, 18, 22), we were unsuccessful in isolating c21 or p2 mutants capable of bypassing AbiQ in spite of using high MOIs. AbiQ appears to interfere with a common and crucial step in the lytic cycle of c2- and 936-like phages.

Mutant phages are very useful for studying the mode of action of Abi proteins. In the absence of such mutants, we investigated the impact of AbiQ on phage DNA replication and MCP production. Our observations showed that, in the presence of AbiQ, immature concatenated phage DNA accumulates and MCPs are normally produced. The concatemeric DNA intermediates may be defective and unable to serve as a substrate for the synthesis of mature phage DNA. Alternatively, mutations affecting the genes involved in the head morphogenesis are known to result in an accumulation of uncleaved phage DNA (6, 7, 44-46). Based on the results presented herein, it is difficult to distinguish between the two possibilities. However, it is tempting to speculate that the morphogenetic pathway is the target of AbiQ. In , mutations affecting the gene of eight proteins involved in head morphogenesis lead to an accumulation of concatemers, namely, Nu1, A, B, C, Nu3, D, E, and FI (7, 44-46). Amino acid homology was found between the  gpA (terminase) and gpB (head-tail connector) analogues of 936-like and c2-like phages (9, 34, 57). No other head morphogenetic proteins (including MCP) of 936- and c2-like phages have significant levels of amino acid identity. The homology between gpA and gpB analogues, although not very extensive (ca. 22% identity), could allow an interaction of AbiQ with one of these peptides, thus accounting for the similar results of AbiQ on phages of both groups. The absence of homology between gpA and gpB analogues of c2- and 936-like phages with corresponding analogues of the P335-like phage r1t (ORF29 and ORF27, respectively) supports our observations on the sensitivity of P335 phages to AbiQ (63).

The kinetics of DNA replication of phage c21 and p2 revealed interesting features of the assembly pathway of these phages. As phage DNA cannot be cleaved in the absence of procapsids (44), the simultaneous presence of both concatemers and cleaved forms of DNA after 20 min following infection implies that proheads were available in the cell and that packaging readily occurred. This observation is supported by the rapid increase in the amount of c21 MCP in the period between 10 and 20 min. Previous studies of the temporal transcription of phage sk1 and c2 indicated that mRNA from the late genes are readily transcribed about 15 min after infection (4, 8). These results are in agreement with our observations on the MCP of phage c21.

AbiQ is very effective against two of the three lactococcal phages groups responsible for fermentation failures in the dairy industry. The introduction of pSRQ900 into industrial strains, as well as a sound use of the improved starter in a performing rotation scheme, should sideline phages for an extended period.