DNA Encodoing Phage Abortive Infection Protein From Lactococcus Lactis and Method of Use Thereof

(1) Field of the Invention

The present invention relates to DNA isolated in a plasmid or in a bacterium encoding a phage resistance protein (designated as AbiE) which aborts infection of the bacterium by the phage. In particular the present invention relates to methods of using the DNA to provide the phage resistance.

(2) Description of Related Art

Lactococcus lactis is widely used in mesophilic milk fermentations to produce cheese, buttermilk, cottage cheese and sour cream. Due to the expanding activities in these industries, there is pressure on L. lactis starter cultures to perform at industrial standards of consistency and efficiency. Phages are the leading cause of fermentation failures during the manufacture of these cultured dairy products (Jarvis A. W., et al., Intervirology 32:2-9 (1991)). They can ruin a fermentation by inactivating the inoculated, sensitive cultures. Since identification of the causal agent in the mid-30s, the dairy industry has learned to manage with this natural phenomenon by developing various solutions such as better sanitation, process modifications and use of phage-resistant cultures (Hill, C., FEMS Microbiol. Rev. 12:87-108 (1993)).

Extensive studies have been carried out on the innate phage resistance mechanisms of L. lactis strains (for review see Hill, C., FEMS Microbiol. Rev. 12:87-108 (1993)). Native barriers against phages have been found and most of them are encoded on plasmids. More than 40 plasmids encoding a variety of phage resistance mechanisms have been identified (Hill, C., FEMS Microbiol. Rev. 12:87-108 (1993)). These anti-phage systems are currently classified in three groups based on their mode of action: blocking of phage adsorption, restriction/modification (R/M), and abortive infection (Abi).

Amongst the natural L. lactis phage resistance mechanisms, the Abi systems are believed to be the most powerful due to their overall effects (Sing, W. D., et al., J. Dairy Sci. 73:2239-2251 (1990)). In a classical abortive infection, the phage lytic cycle is inhibited but after adsorption, DNA injection and early phage gene expression. The end result is the typical Abi.sup.+ phenotype of reduced burst size and plaque size (McLandsborough, L. A., et al., Appl. Environ. Microbiol. 61:2023-2026 (1995); Molineux, I. J., New Biol. 3:230-236 (1991); Sing, W. D., et al., J. Dairy Sci. 73:2239-2251 (1990); Snyder, L., Mol. Microbiol. 15:415-420 (1995)). Generally, the host is also killed in the abortive process. This suicidal outcome traps the phages within the infected cell and limits their dissemination (Sing, W. D., et al., J. Dairy Sci. 73:2239-2251 (1990)).

Many Abi systems have been identified in other bacterial genera (for review see Molineux, I. J., New Biol. 3:230-236 (1991); and Snyder, L., Mol. Microbiol. 15:415-420 (1995)). Some of these systems have been studied extensively but the molecular basis remains somewhat unclear (Snyder, L., Mol. Microbiol. 15:415-420 (1995)). Recent evidence are now only leading to more specific models for their action (Molineux, I. J., New Biol. 3:230-236 (1991); and Snyder, L., Mol. Microbiol. 15:415-420 (1995)). Likewise, Abi systems are very poorly understood in L. lactis (Sing, W. D., et al., J. Dairy Sci. 73:2239-2251 (1990)). To date, five Abi systems have been characterized to molecular level; AbiA (Hill, C., et al., Appl. Environ. Microbiol. 56:2255-2258 (1990), AbiB (Cluzel, P. J., et al., Appl. Environ. Microbiol. 57:3547-3551 (1991), AbiC (Durmaz, E., et al., J. Bacteriol. 174:7463-7469 (1992), AbiD (McLandsborough, L. A., et al., Appl. Environ. Microbiol. 61:2023-2026 (1995) and AbiD1 (Anba, J., et al., J. Bacteriol. 177:3818-3823 (1995). These systems were isolated from different L. lactis strains. There is no homology between the Abi proteins except for AbiD and AbiD1 which shared 28% identity (52% similarity). This was the first indication on the existence of a possible family of Abi proteins in L. lactis (Anba, J., et al., J. Bacteriol. 177:3818-3823). The absence of homology between the other L. lactis Abi mechanisms suggests different mode of action and/or phage targets. Very limited information is available on the molecular mechanisms of L. lactis Abi systems. AbiA is believed to somehow interfere with DNA replication of small isometric phages (Hill, C., et al., Appl. Environ. Microbiol. 57:283-288 (1991); and Moineau, S., et al., Appl. Environ. Microbiol. 59:208-212 (1993)). AbiC does not prevent phage DNA replication but reduces the synthesis of structural phage proteins in infected cells (Durmaz, E., et al., J. Bacteriol. 174:7463-7469 (1992); and Moineau, S., et al., Appl. Environ. Microbiol. 59:208-212 (1993)). Recently, Bidnenko et al. (Bidnenko, E., et al., J. Bacteriol. 177:3824-3829 (1995)) reported that AbiD1 interacted with a small isometric phage operon which contained 4 open reading frames (ORFS). It was proposed that AbiD1 and the orf1 gene product interacted to prevent translation of orf3 RNA (Bidnenko, E., et al., J. Bacteriol. 177:3824-3829 (1995)).

Industrial L. lactis strains with enhanced phage resistance have been constructed by introducing natural plasmids containing Abi systems into phage-sensitive strains (Sanders, M. E., et al., Appl. Environ. Microbiol. 52:1001-1007 (1986)). These improved strains have already been successfully employed for large scale dairy fermentations. However new phages capable of overcoming the introduced Abi defense system have emerged (Alatossava, T., et al., Appl. Environ. Microbiol. 57:1346-1353 (1991); and Moineau, S., et al., Appl. Environ. Microbiol. 59:197-202 (1993)). Thus, the search for novel phage resistance mechanisms is an ongoing objective for culture suppliers (Moineau, S., et al., Appl. Environ. Microbiol. 61:2193-2202 (1995)).

Lactococcal phages are classified in 12 different species based on morphology and DNA homology (Jarvis, A. W., et al., Intervirology 32:2-9 (1991)). Only three have been studied for genetic details because they are commonly encountered worldwide in dairy plants (Jarvis, A. W., et al., Intervirology 32:2-9 (1991)). Members of the species 936 (small isometric heads), c2 (prolate heads) and P335 (small isometric heads) have been, by far, the most disturbing lactococcal phages (Jarvis, A. W., et al., Intervirology 32:2-9 (1991); and Moineau, S., et al., Can. J. Microbiol. 38:875-882 (1992)). DNA--DNA hybridization studies have revealed the absence of significant DNA homology between the three species (Jarvis, A. W., et al., Intervirology 32:2-9 (1991)). The current consensus is that the 936, P335 and c2 species are genetically distinct and L. lactis starter cultures should be resistant against these phages (Moineau, S., et al., Appl. Environ. Microbiol. 61:2193-2202 (1995)).

It is an object of the present invention to provide DNA encoding a sixth and novel abortive infection mechanism (Abi) from L. lactis which acts prior to or at the phage DNA replication and shares no homology with the previously isolated Abi from L. lactis. Further, it is an object of the present invention to provide a novel Abi system which is efficient against 936, c2 and P335 phages.

Further, it is an object of the present invention to provide a method and bacteria which prevent phage inactivation of Lactococcus lactis strains. Further still, it is an object of the present invention to provide recombinant bacteria which are economical to prepare and effective in phage inhibition. These and other objects will become increasingly apparent by reference to the following description and the drawings.

FIG. 1 is a photograph of an electrophoresis gel showing a plasmid analysis of Lactococcus lactis strains. Lanes: 1 and 5, supercoiled DNA ladder (Gibco/BRL, Grand Island, N.Y.); 2, L. lactis SMQ-16; 3, L. lactis SMQ-20; 4, L. lactis W-1 (containing plasmid pSRQ800).

FIG. 2 is a circular restriction map of lactococcal plasmid pSRQ800. Site positions are indicated in kilobases.

FIG. 3 is a linear restriction map of plasmid pSRQ800 and subclones used to localized the abiE gene. The DNA regions represented by the horizonal lines were subcloned into pSA3. The lines are bordered with vertical bars which indicates the site of cloning. The clones were electroporated in L. lactis LM0230 and the transformants tested for resistance against .phi.p2. For sequencing purposes, the 4.5-kb EcoRI fragment was cut in three segments and subcloned into pBS.

FIGS. 4A to 4D are a nucleotide sequence of the 4.5-kb EcoRI fragment from pSRQ800 as shown in SEQ ID NO:1. The deduced amino acid sequence of the only significant open reading frame is presented. The putative promoter and RBS are underlined.

FIGS. 5A and 5B are electrophoresis gels showing .phi.ul36 DNA replication followed at time intervals during infection of L. lactis SMQ-86 (left side) and SMQ-88 (right side). 1-ml samples were taken at time intervals and the total DNA isolated according to the Hill et al. procedure (Hill, C., et al., Appl. Environ. Microbiol. 57:283-288 (1991)). Total DNA was cut with EcoRV and a sample run on a 0.7% agarose gel. Lane M, 1-kb ladder (Gibco/BRL); Lane C, DNA sample prior to phage infection; Lane 0, 15, 30, 45, 65, 75 and 90 indicates time intervals (min.).

FIG. 5C is a graph showing accumulation of the major capsid protein of .phi.uL36 followed at time intervals during infection of L. lactis SMQ-86 and SMQ-88 using an ELISA detection system described previously (Moineau, S., et al., Appl. Environ. Microbiol. 59:208-212 (1993)). An O.D. of 0.389.+-.0.132 was obtained at time 105 with infected SMQ-88 cells. Symbols: .largecircle., SMQ-86 (pSA3); .circle-solid. SMQ-88 (pSRQ802). Error bar indicates standard deviation for 5 trials. The background level of non-infected cells was 0.296.+-.0.085. Substantial accumulation of protein was produced in .phi.uL36 infected SMQ-86 cells.

FIG. 5D is a graph showing one-step growth curves for .phi.ul36 on L. lactis SMQ-86 and SMQ-88. The phages were allowed to adsorb to the cells for 5 min., cells were washed twice to remove non-adsorbed phages and then resuspended in LM17 and incubated at 30.degree. C. Time 15 min corresponded to the first phage count after resuspension in LM17. There was much more progeny phage released from SMQ-86.

The present invention relates to an isolated DNA encoding a protein designated as AbiE which in a Lactococcus lactis increases resistance to phages by aborting infection of Lactococcus lactis by phages, particularly as contained in Lactococcus lactis SMQ20 deposited as NRRL-B-21443 and as set forth in SEQ ID NO:1.

The present invention also relates to a recombinant plasmid containing DNA encoding a protein designated as AbiE which in a Lactococcus lactis increases resistance to phages by aborting infection of Lactococcus lactis by phages, particularly containing the DNA set forth previously.

The present invention also relates to a bacterium harboring a recombinant plasmid containing DNA encoding a protein designated as AbiE which in a Lactococcus lactis increases resistance to phages by aborting infection of Lactococcus lactis by phages, particularly a bacterium harboring DNA as set forth previously. The bacterium are preferably Lactococcus, Lactobacillus, Leuconostocs, Pediococcus, Streptococcus thermophilus, Enterococcus, and Bifidobacterium.

The present invention further relates to a method of imparting phage resistance to a bacterium which is phage sensitive which comprises transferring DNA encoding a protein designated as AbiE which increases resistance to phages by aborting infection of Lactococcus lactis by phages into the bacterium to impart the phage resistance. Further the present invention relates to a method for fermenting a dairy product, the improvement which comprises using a culture of Lactococcus lactis for the fermenting containing transferred DNA encoding a protein designated as AbiE which increases resistance to phages by aborting infection of Lactococcus lactis by phages to impart the phage resistance and to produce the dairy product.

Most particularly the present invention relates to Lactococcus lactis naturally lacking in phage resistance and containing transferred DNA encoding a protein designated as AbiE which increases resistance to phages by aborting infection of Lactococcus lactis by phages, wherein the DNA has a sequence essentially as set forth in SEQ ID NO:1 to impart phage resistance to the Lactococcus lactis.

The natural plasmid pSRQ800 was isolated from Lactococcus lactis subsp. lactis W1. When introduced into a phage-sensitive L. lactis strain, pSRQ800 conferred strong phage resistance against small isometric phages of the 936 and P335 species. It had very limited effect on prolate phages of the c2 species. The phage resistance mechanism encoded on PSRQ800 is a temperature-sensitive abortive infection mechanism (Abi). Plasmid pSRQ800 was mapped and the Abi genetic determinant localized. Cloning and sequencing of the Abi system allowed the identification of a single open reading frame. This ORF coded for a predicted protein of 599 amino acids with an estimated molecular weight of 71.4 kDa and a pI of 8.0. No significant DNA or protein homology was observed with databases. This novel phage resistance mechanism was named AbiE. No phage replication or production of phage major capsid proteins were detected in infected AbiE.sup.+ L. lactis cells. This system is believed to act at or prior to phage DNA replication. When cloned into a high copy vector, AbiE became effective against the c2 species. Thus when delivered in an appropriate vector, AbiE system was efficient against the three most commonly found lactococcal phage species.

The plasmid pSRQ800 is contained in a deposit of Lactococcus lactis SMQ-20 deposited under the Budapest Treaty with the Northern Regional Research Laboratory in Peoria, Ill. on May 17, 1995 as NRRL-B-21443. DNA plasmid pSRQ700 encoding a restriction or modification system (LlaDCHI) is deposited under the Budapest Treaty as NRRL-B-21337 on Sep. 29, 1994 and is used with pSRQ800 to produce a synergistic result. This DNA is described in U.S. application Ser. No. 08/366,480, filed Dec. 30, 1994, which is incorporated by reference herein. The DNA sequence is deposited with GenBank (V16027).

Bacterial strains, plasmids, and media. Strains and plasmids used in this application are listed in Tables 1 and 2. Escherichia coli was grown at 37.degree. C. in LB broth. Lactococcus lactis strains were grown at 30.degree. C. in M17 (Terzaghi, B. E., et al., Appl. Environ. Microbiol. 29:807-813 (1975)) supplemented with 0.5% glucose (GM17) or 0.5% lactose (LM17). When appropriate, antibiotics were added as follows: for E. coli, 50 .mu.g of ampicillin per ml, 10 .mu.g of tetracycline per ml, and 20 .mu.g of chloramphenicol per ml; for L. lactis, 5 .mu.g of chloramphenicol per ml, and 5 .mu.g of erythromycin per ml. Bacteriophage propagation and assays. Bacteriophages used in this study are listed in Table 1.

Bacteriophages c2, p2, sk1, and jj50 were a gift from T. R. Klaenhammer (North Carolina State University). All phages were isolated from a single plaque with a 1-ml sterile pipette (Moineau, S., et al., Appl. Environ. Microbiol. 60:1832-1841 (1994)). They were transferred to GM17 containing 10 mM CaCl.sub.2 previously inoculated (1%) with an overnight culture of the host strain. They were incubated at 30.degree. C. until lysis and filtered through a 0.45 .mu.m filter (Acrodics, Gelman Sciences, Ann Arbor, Mich.). High phage titers were then obtained by the method of Jarvis (Jarvis, A. W., Appl. Environ. Microbiol. 36:785-789 (1978)). Efficiency of plaquing (EOP) and adsorption assays were performed as described by Sanders and Klaenhammer (Sanders, M. E., et al., Appl. Environ. Microbiol. 40:500-506 (1980)). Cell survival was assayed by the method of Behnke and Malke (Behnke, D., et al., Virology, 85:118-128 (1978)) using a MOI of 3. One-step growth curves, center of infection (COI) assays, burst size and latent period determination were performed as described previously (Moineau, S., et al., Appl. Environ. Microbiol. 59:208-212 (1993)). The efficiency at which COI formed (ECOI) was obtained by dividing the number of COI from the resistant strain by the number of COI from the sensitive strain. The intracellular replication of .phi.ul36 DNA was also followed at time intervals after phage infection using the Hill et al. procedure (Hill, C., et al., Appl. Environ. Microbiol. 57:283-288 (1991)). The production of .phi.ul36 major capsid protein (MCP) was followed at time intervals after phage infection using monoclonal antibodies and a sandwich enzyme-linked immunosorbent assay (ELISA) as described previously (Moineau, S., et al., Appl. Environ. Microbiol. 59:2034-2040 (1993); and Moineau, S., et al., Appl. Environ. Microbiol. 59:208-212 (1993)).

DNA isolation and manipulation. Plasmid DNA from E. coli was isolated as described previously (Moineau, S., et al., Appl. Environ. Microbiol. 60:1832-1841 (1994)). Large quantities of E. coli plasmid DNA were isolated with the QIAGEN (Chatsworth, Calif.) plasmid MIDI or MAXI kit. Plasmid DNA from L. lactis was isolated using the method of O'Sullivan and Klaenhammer (O'Sullivan, D. J., et al., Appl. Environ. Microbiol. 59:2730-2733 (1993)). Large quantities of lactococcal plasmid DNA were obtained as described by Gonzalez and Kunka (Gonzalez, C. F., et al., Appl. Environ. Microbiol. 46:81-89 (1983)) . Restriction endonucleases (Gibco/BRL, Grand Island, N.Y.) and T4 DNA ligase (Boehringer Mannheim, Indianapolis, Ind.) were used as described by the manufacturer's instructions.

Electroporation. E. coli was electroporated as previously described (Moineau, S., et al., Appl. Environ. Microbiol. 60:1832-1841 (1994)). L. lactis was grown in GM17 supplemented with 0.5M sucrose and 1% glycine and electroporated according to the Holo and Nes procedure (Holo, H., et al., Appl. Environ. Microbiol. 55:3119-3123 (1989)) as modified previously (Moineau, S., et al., Appl. Environ. Microbiol. 60:1832-1841 (1994)).

Sequencing. The entire abi gene (4.5-kb EcoRI fragment from pSRQ800) could not be cloned in E. coli pBluescript KS+. The 4.5-kb insert was segmented into 3 small fragments (FIG. 3), subcloned into pBS and transformed into E. coli DH5.alpha. with blue and white color selection using IPTG (isopropyl-.beta.-D-thiogalactopyranoside) and X-Gal (5-bromo-4-chloro-3-indoly-.beta.-D-galactopyranoside). The resulting clones were named pSRQ810, pSRQ812 and pSRQ814 (FIG. 3). Nested deletions were made in both orientation from pSRQ810, pSRQ812 and pSRQ814 with an ERASE-A-BASE kit (Promega, Madison, Wis.). Plasmid DNA was extracted from the nested clones with QIAGEN and used directly for sequencing. The sequencing reactions were performed with the DYEDEOXY TERMINATOR TAQ sequencing kit for use on the model 373A automated DNA sequencing system (Applied Biosystems, Foster City, Calif.). The T7 and T3 primers were used for annealing.

DNA and protein analysis. The DNA sequence was analyzed with the Genetics Computer Group (Madison, Wis.) sequence analysis software. The GenBank (release 90, 15 August 95) and SwissProt (release 31.0, March 95) databases were searched for homology with the deduced abiE gene and protein.

Nualeotide sequence accession number. The complete sequence of 4,467 bp was deposited in the GenBank database and is available under the accession number U35629.

Isolation of PSRQ800. The objective of this study was to identify phage resistance mechanism(s) present in Lactococcus lactis subsp. lactis W1. The total plasmid DNA of W1 was co-electroporated with the shuttle vector pSA3 into the laboratory strain L. lactis LM0230 which is phage-sensitive and plasmid-free. LM0230 was used because of its sensitivity to phages of the 936 and c2 species. The W1/pSA3 DNA ratio used for electroporation was 10:1. Erythromycin-resistant colonies were obtained (due to pSA3) and tested for phage resistance by spot assay (10.sup.4 .phi.p2/spot). Some phage-resistant transformants were obtained, analyzed and found to contain pSA3 and a 8.0-kb plasmid which was named pSRQ800 (FIG. 1). One phage-resistant transformant was selected and named L. lactis SMQ-20.

Effectiveness of pSRQ800 on three lactococcal phage species.

L. lactis SMQ-20 was tested for resistance against three small isometric-headed phages of the 936 species as well as three prolate-headed phages of the c2 species (Table 1). These 6 phages were individually tested on SMQ-20 and their EOPs are presented in Table 3. EOPs ranging from 10.sup.-6 to 10.sup.-7 were obtained for the 936-type phages .phi.p2, .phi.sk1 and .phi.jj50 whereas EOPs of 10.sup.-1 were obtained with prolate .phi.c2, .phi.ml3 and .phi.eb1 (Table 3).

The activity of pSRQ800 was also tested against P335 phages. Since LM0230 cannot replicate P335 phages, pSRQ800 was introduced into an appropriate host. L. lactis UL8 was electroporated with pSRQ802, a derivative of pSRQ800 (see below), and the transformant named SMQ-88. The EOPs of three P335 phages (.phi.ul36, .phi.Q30 and .phi.Q33) on SMQ-88 are presented in Table 3. The EOPs were variable ranging from 10.sup.-5 to 10.sup.-8. These results showed that the phage resistance mechanism encoded on pSRQ800 was effective against small isometric phages of the 936 and P335 species. However, pSRQ800 had very limited effect on prolate phages (c2 species).

Effect of temperature on pSRQ800. The effect of temperature on the activity of pSRQ800 was tested using .phi.p2. Similar EOPs were obtained at 21.degree. C. and at 30.degree. C. However, the EOP of .phi.p2 increased from 4.6.times.10.sup.-6 to 6.7.times.10.sup.-2 at 38.degree. C. indicating that the phage resistance mechanism encoded on pSRQ800 is heat-sensitive. Identification of the phage resistance mechanism on pSRQ800.

The type of phage defense mechanism located on PSRQ800 was examined. Adsorption experiments indicated that phages adsorbed to the same level (approximately between 90 to 95%) on the phage-sensitive L. lactis LM0230 and on the phage-resistant L. lactis SMQ-20, ruling out the adsorption blocking mechanism.

To determine if the mechanism was a R/M system, L. lactis SMQ-20 was challenged with high concentration of .phi.p2. Phages capable of overcoming the pSRQ800 defense system and forming plaques on SMQ-20 were isolated. When these phages were tested back on SMQ-20, we observed that some phages were still inhibited by pSRQ800 while others were unaffected. These former phages were still untouched by pSRQ800 even after propagation on the sensitive host LM0230, indicating a permanent modification. These mutants or .phi.p2 derivatives are under investigation. Nevertheless, these results ruled out the presence of host-controlled modifications such as R/M systems.

Based on the current classification of lactococcal phage resistance mechanisms, this system was considered an abortive infection mechanism. Furthermore when challenged with phages, L. lactis SMQ-20 exhibited the typical Abi.sup.+ phenotype of reduced plaque size.

Localization of the abi gene on pSRQ800. The plasmid pSRQ800 was mapped using single, double and triple endonuclease digestions. The results are presented in FIG. 2. The following endonucleases did not cut pSRQ800: ApaI, AvaI, BalI, BamHI, HpaI, NruI, PstI, SalI, ScaI, SmaI, SphI, SstI, XbaI, XhoI. Since appropriate restriction sites were present on pSA3 and PSRQ800, total plasmid DNA from L. lactis SMQ-20 was digested with EcoRI, EcoRV and/or NcoI, and religated. The ligation mixture was directly electroporated in E. coli DH5.alpha.. After obtaining the appropriate clones, plasmid DNA was electroporated into L. lactis LM0230. The Em.sup.r transformants were tested for phage resistance. The relevant clones are presented in FIG. 3. Most of the clones (PSRQ801, pSRQ804, pSRQ806, pSRQ809) conferred no resistance (EOP of 1.0) against .phi.p2 (FIG. 3). Only the clone pSRQ802 gave phage resistance similar to pSRQ800 (EOP of 10.sup.-6, FIG. 3). Thus, the full Abi.sup.+ phenotype was localized on a 4.5-kb EcoRI fragment.

DNA and protein sequence analysis of the Abi system. The 4.5-kb fragment containing the Abi.sup.+ phenotype was sequenced in both directions and found to contain 4,467 bp (FIGS. 4A to 4D). This fragment had a G+C content of 29.5%. Two significant open reading frames (ORFs) were found on the 4.5-kb fragment. The first ORF was localized in one direction from position 1279 to 3075. The second ORF was localized in the other direction from position 4345 to 3491. The clones pSRQ804 and pSRQ809 which contained the full ORF2 but a disrupted ORF1 exhibited no resistance (EOP of 1.0) against .phi.p2 (FIG. 3). We concluded that only ORF1 had all the necessary information to confer the Abi+ phenotype.

The complete abi gene contained 1,797 bp with a very low GC content of 23.9%. A putative ribosome binding site (AAAGGAG) was found 8 bases preceding the abi gene start codon (FIG. 4A to 4D). Putative promoter regions -10 (TATAAT), 16 bp spacer and -35 (AGCACA), were found upstream of the abi gene (FIGS. 4A to 4D). No region of dyad symmetry was found at the end of the abi gene suggesting the presence of a rho-dependent terminator. No significant DNA homology was found between this abi gene and the GenBank database including the previous sequenced L. lactis abi genes. This novel system was named AbiE.

The predicted AbiE protein contained 599 amino acids with an estimated weight of 71,416 Da and a pI of 8.0. No homology to the AbiE protein was found in the SwissProt database, confirming the novelty of AbiE. No transmembrane helix or secretory signals were found indicating that the AbiE protein is most likely intracellular. The Abi.sup.+ phenotype was totally abolished with the clone pSRQ809 where the last 42 amino acids were missing from the AbiE protein. This result indicated that the C-terminal of the AbiE protein was critical to the Abi.sup.+ phenotype.

Enhancement of AbiE efficiency. The copy number of pSRQ800 was not determined but was estimated to be roughly the same as the low copy vector pSA3 (FIG. 1). The AbiE system was cloned into the high copy plasmid pNZ123 to determine the effect of gene copy number on the AbiE.sup.+ phenotype. Previously, it was estimated that pSA3 was present in 5 to 10 copies (Dinsmore, P. K., et al., Appl. Environ. Microbiol. 60:1129-1136 (1994)) whereas pNZ123 was found between 50 to 100 copies in L. lactis (DeVos, W. M., FEMS Microbiol. Rev. 46:281-295 (1987)). The resulting high copy clone and transformant were named pSRQ815 and SMQ-143, respectively. The EOP of .phi.p2 on SMQ-143 was 10.sup.-8 which is 2 negative logs higher than the EOP on SMQ-20 which contained the natural plasmid pSRQ800. Interestingly, the EOP of prolate phage c2 was also significantly decreased on SMQ-143. An EOP of .sub.10.sup.-4 was observed with pSRQ815 whereas EOPs of 10.sup.-1 were observed with pSRQ800 and pSRQ802 (Table 4).

These results indicated that multiple copies of abiE gene increase phage resistance against small isometric and prolate phages.

Another way of increasing the efficiency of Abi system is to supply in trans an another phage resistance mechanism (Sing, W. D., et al., Appl. Environ. Microbiol. 59:365-372 (1993)). Previously, we have isolated a L. lactis R/M system named LlaDCHI (formerly LlaII), that was encoded on the plasmid pSRQ700 (Moineau, S., et al., Appl. Environ. Microbiol. 61:2193-2202 (1995)). This plasmid was isolated from the industrial strain L. lactis subsp. cremoris DCH-4. The LlaDCHI genes were previously cloned into pSA3 and the Em.sup.r clone named pSRQ701 (Moineau, S., et al., Appl. Environ. Microbiol. 61:2193-2202 (1995)). Phage p2 had an EOP of 10.sup.-6 on L. lactis LM0230 harboring pSRQ701 (SMQ-39, Table 4). This phage had also an EOP of 10.sup.-6 on L. lactis LM0230 harboring pSRQ800 (SMQ-20, Table 4). The abiE gene was cloned into the vector pMIG3 (which contains a Cm.sup.r gene) and the clone was named pSRQ813. The EOP of .phi.p2 on LM0230 containing pSRQ813 was slightly lower (10.sup.-5) (SMQ-130, Table 4). Plasmids pSRQ701 and pSRQ813 were electroporated into L. lactis LM0230 and the resulting Em.sup.r /Cm.sup.r transformant was named SMQ-138. The EOP of .phi.p2 on SMQ-138 was <10.sup.-9 indicating a full phage resistance phenotype. The EOP of the P335 phage ul136 was also <10.sup.-9 on another L. lactis strain harboring both plasmids (data not shown). These results showed that AbiE and LlaDCHI R/M system, when supplied in trans, can complement each other to enhance phage resistance levels.

Intracellular effects of AbiE system on the phage life cycle. The effects of the AbiE system was tested on the phage ul36 life cycle using the phage-sensitive L. lactis SMQ-86 (pSA3) and phage-resistant L. lactis SMQ-88 (pSRQ802). Phage ul36 was selected because it was previously used to characterize two other Abi systems, AbiA and AbiC (Moineau, S., et al., Appl. Environ. Microbiol. 59:208-212 (1993)). The replication of .phi.ul36 was severely inhibited on SMQ-88 as seen by the EOP of 10.sup.-6 (Table 5 and FIG. 5C). Plaques were reduced in size from 1 mm to pinpoint and difficult to enumerate (data not shown). The average number of surviving cells following phage infection (MOI of 3) was 0.06% on SMQ-86 and 2.65% on SMQ-88 (Table 5).

AbiE improved cell survival by approximately 50 fold but still a majority of infected cells died.

The ECOI for .phi.ul36 on SMQ-88 was 0.0051 indicating that only 5 out of 1000 infected cells successfully release viable phages (Table 5). The burst size of these productive infections was also reduced from an average of 504 phages on the phage-sensitive SMQ-86 to 36 phages on SMQ-88 (Table 5). Furthermore, the phage latent period was longer on SMQ-88 by approximately 15 min. (FIG. 5C). These results indicated that .phi.ul36 took more time to complete its life cycle on an AbiE.sup.+ host and that the number of released phages was reduced by almost 15 fold. The cumulative effects of AbiE on the cell survival, ECOI, burst size and latent period were responsible for the severely reduced EOP of .phi.ul36 on SMQ-88 (Table 5).

The intracellular replication of .phi.ul36 DNA was followed on SMQ-86 and SMQ-88 (FIG. 5A and 5B). No replication of .phi.ul36 could be detected in L. lactis SMQ-88 whereas significant replication occurred in SMQ-86. The intracellular production of .phi.ul36 major capsid protein was also followed on SMQ-86 and SMQ-88 (FIG. 5C). The MCP is the most abundant structural protein of .phi.ul36 (Moineau, S., et al., Appl. Environ. Microbiol. 59:2034-2040 (1993)). Using an ELISA detection system, barely no MCP protein was detected in infected SMQ-88 cells (O.D.=0.389) whereas massive amount of MCP was detected in infected SMQ-86 cells (corrected O.D.=22.618). These results indicated that AbiE might act at or prior to phage DNA replication.

As can be seen from Examples 1 and 2, Lactococcus lactis subsp. lactis strain W1 harbors a 8.0-kb plasmid (pSRQ800) coding for a temperature-sensitive abortive phage infection mechanism. The genetic element responsible for the Abi+ phenotype was cloned and sequenced from pSRQ800. One gene was necessary for the Abi.sup.+ phenotype.

AbiE displayed the classical abortive infection phenotype (Molineux, I. J., New Biol. 3:230-236 (1991); Sing, W. D., et al., J. Dairy Sci. 73:2239-2251 (1990); and Snyder, L., Mol. Microbiol. 15:415-420 (1995)). The presence of pSRQ800 increased the survival of infected cells by almost 50-fold demonstrating that AbiE was successful in aborting phage infection. Substantial cell death (98%) still occurred upon phage abortion (Table 5) suggesting that cell damage was too considerable for recovery. Very few infected AbiE.sup.+ cells released progeny phages (0.5%) and when these rare productive infections occurred, the latent period was longer (by 15 min) and the burst size reduced by 15-fold. The overall effect of AbiE can be visualized as a reduction in number and size of phage plaques. Thus, AbiE acted internally to interfere with the phage lytic development.

AbiE encoded on its natural plasmid pSRQ800, inhibited the development of small isometric phages of the 936 and P335 species but had very limited effect on the prolate c2 species (Table 3). However if the abiE gene was cloned into the high copy vector pNZ123, a 4 log increase in resistance was noticed against phage .phi.c2. This indicated that AbiE can be efficient against prolate phages when present in high copy number. A 2 log increase was also observed with .phi.p2 (Table 3). Elevated efficiencies with high copy genes has also been observed with L. lactis AbiA (Casey, J., et al., Appl. Environ. Microbiol. 58:3283-3291 (1992); and Dinsmore, P. K., et al., Appl. Environ. Microbiol. 60:1129-1136 (1994)). Therefore, the natural plasmid pSRQ800 can bestow strong resistance against small isometric phages (936 and P335 species) and if the functional gene is delivered in an appropriate vector, AbiE can be effective against all three common lactococcal phage species (Jarvis, A. W., et al., Intervirology 32:2-9 (1991)).

Five other L. lactis Abi systems have been sequenced (Anba, J., et al., J. Bacteriol. 177:3818-3823 (1995); Cluzel, P. J., et al., Appl. Environ. Microbiol. 57:3547-3551 (1991); Durmaz, E., et al., J. Bacteriol. 174:7463-7469 (1992); Hill, C., et al., Appl. Environ. Microbiol. S6:2255-2258 (1990); and McLandsborough, L. A., et al., Appl. Environ. Microbiol. 61:2023-2026 (1995)). In all cases, only one gene was necessary to confer the Abi.sup.+ phenotype. The five deduced Abi proteins had 628 (AbiA), 251 (AbiB), 344 (AbiC), 366 (AbiD), and 351 (AbiD1) amino acids. Thus, AbiE (599 aa, 71.4-kDa) would be closer to AbiA (628 aa and 73.8 kDa) based on protein size.

Functionally, AbiE was also similar to AbiA. This was evidenced by its efficient inhibition of 936 and P335 species (Casey, J., et al., Appl. Environ. Microbiol. 58:3283-3291 (1992); Dinsmore, P. K., et al., Appl. Environ. Microbiol. 60:1129-1136 (1994); and Hill, C., et al., Appl. Environ. Microbiol. 56:2255-2258 (1990)), its increased activity against prolate phages with high copy (Casey, J., et al., Appl. Environ. Microbiol. 58:3283-3291 (1992); and Dinsmore, P. K., et al., Appl. Environ. Microbiol. 60:1129-1136 (1994)) and is heat sensitivity (Hill, C., et al., Appl. Environ. Microbiol. 56:2255-2258 (1990)). The AbiA phage target is currently unknown but it is believed to act prior to .phi. DNA replication since no DNA replication and MCP production were observed in infected AbiA.sup.+ cells (Moineau, S., et al., Appl. Environ. Microbiol. 59:208-212 (1993)). Similar DNA replication and MCP production results were obtained with AbiE (FIGS. 5C and 5D). Even though AbiE shared phenotypic similarities with AbiA, no significant amino acid homology was found between the two proteins.

When developing phage-resistant cultures, one approach is to stack different phage resistance mechanisms within one strain (Durmaz, E., et al., Appl. Environ. Microbiol. 61:1266-1273 (1995); and Sing, W. D., et al., Appl. Environ. Microbiol. 59:365-372 (1993)). In fact some natural L. lactis plasmids like pTR2030 (Hill, C., et al., Appl. Environ. Microbiol. 56:2255-2258 (1990); and O'Sullivan, D. J., et al., J. Bacteriol. 177:134-143 (1995)) and pTN20 (Durmaz, E., et al., J. Bacteriol. 174:7463-7469 (1992)), already encode two distinct phage resistance mechanisms: an abortive infection mechanism and a R/M system. In these two plasmids, the Abi and R/M systems naturally complement each other to confer very strong phage resistance (EOP <10.sup.-9). AbiE and LlaDCHI are considered strong phage resistance mechanisms since EOPs of 10.sup.-6 were observed with some phages (Moineau, S., et al., Appl. Environ. Microbiol. 61:2193-2202 (1995), Table 3). However on their own, both systems are leaky and a high phage population (>10.sup.6) can overpower them. When AbiE and LlaDCHI were introduced in L. lactis strains, a full resistance phenotype (EOP <10.sup.-9) was observed against 936 and P335 phages (Table 4). These results demonstrated the beneficial effect of providing in trans an Abi and a R/M system. Similar added resistance have been previously observed with other in trans L. lactis anti-phage systems (Durmaz, E., et al., Appl. Environ. Microbiol. 61:1266-1273 (1995); and Sing, W. D., et al., Appl. Environ. Microbiol. 59:365-372 (1993)). This is believed to be the first instance where such a strong resistance is observed with completely characterized Abi and R/M systems presented in trans in L. lactis.

From a practical and current regulatory point of view, the introduction of the natural plasmid PSRQ800 into an industrial L. lactis strain to confer strong resistance against small-isometric phages of the 936 and P335 species, is useful. It is useful against c2 phage species when the AbiE gene is provided in large copy number.

The foregoing description is only illustrative of the present invention and the present invention is limited only by the hereinafter appended claims.