A 7.275-kb DNA fragment which encodes resistance by abortive infection (Abi+) to bacteriophage was cloned from Lactococcus lactis subsp. cremoris S114. The genetic determinant for abortive infection was subcloned from this fragment. This gene was found to confer a reduction in efficiency of plating and plaque size for prolate-headed bacteriophage φ53 (group I homology) and for small isometric-headed bacteriophage φ59 (group III homology). This new gene, termed abiN, is predicted to encode a polypeptide of 178 amino acid residues with a deduced molecular mass of 20 461 Da and an isoelectric point of 4.63. No homology with any previously described genes was found. A probe was used to determine the presence of this gene only in S114 from 31 strains tested.
Lactococcus lactis strains are often susceptible to phage in dairy fermentations. Many natural lactococcal phage resistance mechanisms occur, classified as interference with phage adsorption, inhibition of phage DNA injection, DNA restriction modification, and abortive infection, as previously reviewed . Several genes conferring abortive phage infection (Abi+), usually associated with plasmid DNA elements, were cloned and sequenced [2–12]. Only one was shown to be located in the chromosome .
Lactococcus lactis subsp. cremoris S114 is a natural phage-resistant strain used in industry . S114 contains at least one self-transmissible plasmid, pPF66, which confers an abortive infection phenotype. In the present study, the cloning and sequencing of a new gene involved in abortive phage infection from S114 is described. Its localization was shown to be in the chromosome, rather than in pPF66. Its distribution among industrial lactococcal strains was investigated.
2Materials and methods
2.1Strains, plasmids, bacteriophages and culture conditions
Escherichia coli TG1 was used . L. lactis subsp. cremoris S77 and S114, L. lactis subsp. lactis S91, MG1363 and S45, L. lactis subsp. lactis biovar diacetylactis S94 and S96 were used [13, 15]. The shuttle plasmid pLDP1 was used . The plasmid pPF107-3 was isolated from S96 and sequenced (Y12675). Bacteriophages φ53 (group I, prolate-headed phage) and φ59 (group III, small isometric-headed phage), active on MG1363 and S45, were used to quantify the phage resistance . Luria broth (LB) and Luria plates were used for routine culturing of E. coli strains. M17  broth and plates were used for lactococcal strains. Erythromycin was used for E. coli and L. lactis strains at 200 and 5 μg ml−1, respectively.
2.2Construction of a S114 DNA bank
Genomic DNA and a DNA bank of L. lactis S114 were prepared as previously described . After 30 h of incubation, 3600 erythromycin-resistant cells (5 μg ml−1) were tested for their resistance to phages φ59 and φ53 by spot tests of phage dilutions .
2.3DNA amplification by PCR
Oligonucleotides D (TACGTGAATTC/TTTTCATAGTCTAGCTATAC) and E (TACGTGAATTC/GTAAAAAGTAAAAACGTTAG) were used to amplify a DNA fragment of 817 bp containing ORF7 by PCR.
2.4Sequencing and subcloning
Plasmid pLAB201 was sequenced directly by Taq cycle sequencing using fluorescence-based chain termination chemistry (Perkin Elmer/Applied Biosystems, Foster City, CA) and an automatic DNA sequencer (Perkin Elmer/Applied Biosystems, model 373 A). The pLAB201 plasmid was digested with SalI, treated with the exonuclease Bal31, and ligated with T4 DNA ligase. After transformation of TG1, plasmids were pooled and used to transform MG1363.
A PvuI restriction site was introduced by PCR into the DNA fragment of 817 bp containing ORF7: the sequence CGAAGT was replaced by the sequence CGATCG without any change in the amino acid sequence. This new site was used to introduce or to remove the DNA sequence TCGTAAATGAATATATAATCAATTACGA upon digestion with PvuI and ligation. This sequence contains codon stops in the six frames.
Analysis was conducted using programs included in the GCG (Genetics Computer Group, Madison, WI) package. The whole sequence of the 10 ORFs from the 7275-bp DNA fragment, and the sequences between each ORF, were compared with the whole set of known nucleic sequences of GENE/EMBL sequence databases, using standard algorithms of biosequence analysis, such as the FastA and the BLAST commands. The deduced amino acid sequences of the ORFs were compared with the whole set of protein sequences stored in SWISSPROT/PIR databases, using the FastA algorithm.
2.6Southern blot hybridization
This was performed as previously described . Genomic DNA from 26 industrial lactococcal strains and from L. lactis subsp. lactis biovar diacetylactis S94 and S96, L. lactis subsp. lactis S45, MG1363 and S91, and L. lactis subsp. cremoris S114 and S77 was digested with HindIII and probed with purified abiN gene. To test plasmid or chromosomal location of abiN, total DNA digested with Sau3AI and CsCl plasmid preparation of S91, S94, S96, S114 and S77 strains were probed with purified abiN gene.
2.7Nucleotide sequence accession numbers
The sequences of the 7275-bp Sau3AI fragment encoding abiN and of the pPF107-3 plasmid have been deposited in GenBank under accession numbers Y11901 and Y12675, respectively.
3Results and discussion
In 3600 clones from the S114 DNA bank, six colonies harboring a plasmid conferring phage resistance were found. One clone was tested and showed a partial resistance to φ53 and φ59. This clone has a pLDP1 plasmid containing a 7.5-kb insert, designated pLAB201, which exhibits reduced plaque size and reduced efficiency of plating (EOP) to φ53 and φ59 (Table 1). This resistance is not caused by a restriction and modification mechanism, because plaque size and EOP are reduced for both phages regardless of whether the phage has been propagated on cells with or without pLAB201. Adsorption of phages φ53 and φ59 was unaffected by the presence of pLAB201. These phenotypes indicate that pLAB201 confers an abortive phage resistance phenotype.
Table 1. Phage sensitivity of various strains
plaque size (mm)
plaque size (mm)
MG1363 and MG1363 (pLDP1) were used as controls.
The insert of plasmid pLAB201 was sequenced (Y11901). It shows a size of 7275 bp and contains 10 large ORFs (ORF1–ORF10) (Fig. 1). An unique restriction site, SalI, is present in pLDP1 near ORF1. Analysis showed that removal of ORF1–ORF6 by Bal31 did not affect the phage resistance. When ORF7 was removed, the phage resistance was abolished. DNA fragments of 1272, 1924 and 2634 bp were subcloned by PCR into pLDP1 at BamHI and SalI sites (Fig. 1), giving rise to plasmids pLAB203, pLAB204 and pLAB202. Only pLAB202 was shown to be involved in phage resistance (Table 1). The difference between pLAB202 and pLAB204 is the presence of ORF7 in the latter. A DNA fragment of 817 bp containing ORF7 was subcloned by PCR, resulting in pLAB205. This is the fragment which confers phage resistance. When ORF7 was interrupted by integration of a DNA fragment, no phage resistance was observed. When this segment of DNA was removed, phage resistance was restored.
The DNA fragment of 817 bp containing ORF7 showed no significant homologies with any GenBank-listed DNA sequences, particularly with abiA, abiB, abiC, abiD, abiD1, abiEi, abiEii and abiF, abiG, abiH, abiI, abiJ or abi-859 and abiK. Apparently a new unique lactococcal abi gene has been sequenced which was designated abiN. In this sequence, a putative ribosomal binding site GGCAG was identified 6 bp upstream of the first start codon ATG. However, it is known that L. lactis can also initiate translation at the start codons TTG and GTG . Upstream of the first ATG codon, the start codons TTG and GTG were identified, with putative ribosomal binding sites, GAAGG and GGAGA, 6 and 9 bp upstream of the start codons, respectively. Upstream of the first ribosomal binding site, GGCAG, four potential (two −10 and two −35) Pribnow boxes were observed: TATTAT −10 box and TTGTGA −35 box on one side, and TATAAG −10 box and TTGGAT −35 box on the other side. The first pair differs from the E. coli and L. lactis consensus sequences, TATAAT −10 box and TTGACA −35 box, in three of its bases [18, 19], and the second pair differs in four of its bases. In each case, both hexamers are separated by 16 bp. All these boxes are not located in ORF6. The insert of pLAB205 confers phage resistance in the two orientations meaning that the promoter is included in the insert. No clearly rho-dependent termination site similar to CAATCAA was identified , and no rho-independent terminator was detected . The relatively low A+T content of ORF7 (68%) is different from that already described for the abortive gene (74%). The reason for this is not known. ORF7 encodes a hydrophilic protein of 178 amino acids, assuming that translation of this protein starts at the first ATG codon. The AbiN protein would have a predicted molecular mass of 20 461 Da, which is the smallest size of any lactococcal abortive infection protein yet described.
The protein deduced from ORF1 shows 57.1% similarity and 38.0% identity with the protein deduced from ORF19 of r1t phage . Comparison of both sequences seems to point to a large number of rearrangements in ORF1, at least three deletions and two insertions. The protein deduced from ORF3 shows 98.6% similarity and 96.4% identity with dUTPase deduced from ORF20 of r1t phage. In addition, ORF1 and ORF3 are separated by ORF2. This is not the case for ORF19 and ORF20 of r1t, which are close together. Moreover, the DNA sequence between ORF4 and ORF5 shows 81.8% homology with the r1t genome, between positions 13 034 and 13 156. All this suggests that a DNA prophage homologous to r1t is present in the chromosome of the S114 strain in a highly rearranged form. Supporting this hypothesis, DNA-DNA hybridization with purified abiN gene (Fig. 2) clearly indicates that the abiN gene is located in S114 chromosomal DNA. It would be interesting to know whether there is a relationship between the presence of this rearranged genome and abiN. ORF2, ORF4–ORF6, and ORF8–ORF10 showed no significant homology with any gene in the data banks, at either the protein or the DNA level. No homology with IS sequences was present in the lactococcal insert.
Attempts to clone the whole abiN gene into the high copy number lactococcal plasmid pPF107-3 (Y12675) in MG1363 were unsuccessful. Only deleted forms were cloned, showing the potential toxicity of abiN at a high gene dosage (data not shown).
Southern blot hybridization was performed to determine if industrial lactococcal strains contained this new phage resistance gene. Control strain S114 hybridized with the abiN probe (Fig. 2), but not with S45, MG1363 or any of the other 30 industrial strains tested (data not shown). These results indicate the benefit in introducing this gene into a large number of lactococcal strains used in the dairy industry.
The function of abiN remains to be determined. However, due to the ability of phages to escape single phage resistance mechanism , several genes encoding different mechanisms should preferentially be introduced into the same lactococcal strain to ensure better resistance.
This work was supported by Sanofi and Systems Bio-industries. We thank Ann Husgen for reading the manuscript.