Cloning of Genomic DNA of Lactococcus lactis That Restores Phage Sensitivity to an Unusual Bacteriophage sk1-Resistant Mutant

Abstract

An unusual, spontaneous, phage sk1-resistant mutant (RMSK1/1) of Lactococcus lactis C2 apparently blocks phage DNA entry into the host. Although no visible plaques formed on RMSK1/1, this host propagated phage at a reduced efficiency. This was evident from center-of-infection experiments, which showed that 21% of infected RMSK1/1 formed plaques when plated on its phage-sensitive parental strain, C2. Moreover, viable cell counts 0 and 4 h after infection were not significantly different from those of an uninfected culture. Further characterization showed that phage adsorption was normal, but burst size was reduced fivefold and the latent period was increased from 28.5 to 36 min. RMSK1/1 was resistant to other, but not all, similar phages. Phage sensitivity was restored to RMSK1/1 by transformation with a cloned DNA fragment from a genomic library of a phage-sensitive strain. Characterization of the DNA that restored phage sensitivity revealed an open reading frame with similarity to sequences encoding lysozymes (β-1,4-N-acetylmuramidase) and lysins from various bacteria, a fungus, and phages of Lactobacillus and Streptococcus and also revealed DNA homologous to noncoding sequences of temperate phage of L. lactis, DNA similar to a region of phage sk1, a gene with similarity to tRNA genes, a prophage attachment site, and open reading frames with similarities to sun and to sequences encoding phosphoprotein phosphatases and protein kinases. Mutational analyses of the cloned DNA showed that the region of homology with lactococcal temperate phage was responsible for restoring the phage-sensitive phenotype. The region of homology with DNA of lactococcal temperate phage was similar to DNA from a previously characterized lactococcal phage that suppresses an abortive infection mechanism of phage resistance. The region of homology with lactococcal temperate phage was deleted from a phage-sensitive strain, but the strain was not phage resistant. The results suggest that the cloned DNA with homology to lactococcal temperate phage was not mutated in the phage-resistant strain. The cloned DNA apparently suppressed the mechanism of resistance, and it may do so by mimicking a region of phage DNA that interacts with components of the resistance mechanism.

Bacteriophage infection of lactic acid-producing starter cultures is the most persistent problem in manufacturing fermented milk products. Although naturally occurring bacteriophage resistance mechanisms have been applied successfully in the development of new bacteriophage-resistant starter cultures, novel bacteriophages have arisen that overcome these resistances (49). Starter strains bearing latent prophage may serve as a source of genes in the continuous evolution of bacteriophages (7, 24, 60) and their hosts.

To expand the range of possibilities for development of phage-resistant starter strains, we continue to identify host genes that affect phage replication. These genes are a potential source of novel phage resistance. A prototypical example is pip, a gene carried by many strains of Lactococcus lactis (4, 43), which encodes a cell surface protein required for infection by phages of the c2 species (34). Strains of L. lactis have been developed with site-specific mutations in pip that render the resultant strain completely resistant to phages of the c2 species (30, 43).

Bacteriophages of the 936 species (38) are the most frequently problematic phage in buttermilk and cheddar cheese plants (10, 48). Phage sk1 is a small, isometric bacteriophage of the 936 species (38) that infects several strains of L. lactis, a bacterium used in culturing cheddar cheese, sour cream, and buttermilk. The entire genome of phage sk1 has been sequenced (11). Previously, phage sk1-resistant mutants of L. lactis C2 that have cell wall compositions indistinguishable from that of L. lactis C2 and that adsorb phage sk1 particles normally were isolated (55). In this study, one of these phage sk1-resistant strains, RMSK1/1, was restored to a phage-sensitive phenotype by transformation with a library of L. lactis genomic DNA. An analysis of the cloned DNA revealed similarities to specific regions of other lactococcal phages, which suggests a mechanism for suppressing phage resistance in RMSK1/1.

MATERIALS AND METHODS

Bacterial strains, phages, media, and growth conditions.

Lactococcus lactis subsp. lactis strain C2, its phage sk1-resistant derivative, RMSK1/1 (55), plasmid-free derivatives LM2301 and LM0230 (61), and MM210 (Table 1) were propagated in M17 (53) supplemented with 0.5% glucose (M17G) at 30°C. Where necessary, erythromycin was added at 5 μg/ml (E5). Calcium chloride was added at 10 mM in M17G medium (M17GC) for propagation of lactococcal phages. Phages sk1 and 64 are phages of the 936 species from the collection of T. Klaenhammer (North Carolina State University). Phages p2, 712, and jj50 are phages of the 936 species from the collection of S. Moineau (Université Laval) (25, 39). Escherichia coli strains DH5α and HB101 were grown in Luria-Bertani broth at 37°C with shaking or in Luria-Bertani broth supplemented with 1.5% agar and, when required, ampicillin at 100 μg/ml, chloramphenicol at 50 μg/ml, or tetracycline at 20 μg/ml.

TABLE 1.

Strains and plasmids

Strain or plasmid	Description	Source or reference
L. lactis
C2	Cheese starter strain	41
LM0230	Plasmid-free derivative of C2	61
LM2301	Streptomycin-resistant derivative of LM0230	61
RMSK1/1	Phage sk1-resistant derivative of C2	55
MM210	Cheese starter strain not sensitive to phage sk1	43; Rhodia, Madison, Wisc.
Plasmids
pSA3	Cloning vector used to construct genomic library	17
p17	Genomic fragment in pSA23 that restored phage sensitivity	This work
p17ΔXbaI	p17 deletion downstream of XbaI site	This work
p17ΔAfl-Swa	p17 deletion between AflIII and SwaI sites	This work
pSA3ΔHindIII	pSA3 deletion that lacks origin of replication associated with gram-positive organisms	This work

The isolation of RMSK1/1 was previously described (55). Briefly, L. lactis subsp. lactis C2 was mixed with phage sk1 at a multiplicity of infection (MOI) of 100 and plated on M17 top agar with 10 mM CaCl₂. Spontaneous phage-resistant mutants were collected by washing the top agar and plating the cell suspension on plates of M17C plus 10⁷ PFU of phage sk1 per ml. The carbohydrate composition of the cell wall of RMSK1/1 and the ability of phage sk1 to adsorb to it were identical to those of strain C2 (55).

DNA manipulations.

Restriction endonuclease digestion, agarose gel electrophoresis, and DNA ligation were done by standard procedures described by Sambrook et al. (52), according to the recommendations of the enzyme suppliers (New England Biolabs, Beverly, Mass.). Ligation products were transformed into competent cells of E. coli DH5α (Gibco-BRL, Gaithersburg, Md.). Plasmids were isolated from E. coli with Qiagen (Chatsworth, Calif.) kits. Plasmids were isolated from L. lactis with Qiagen miniprep kits after the pelleted cells had been treated with 30 mg of lysozyme per ml for 30 min at 30°C.

Genomic library construction, transformation, and screening of transformants.

Partially MboI-digested LM2301 DNA was ligated to BamHI-digested pSA3 (17) as described previously (34). Ligation products were transformed into MAX efficiency-competent cells of E. coli DH5α (Gibco-BRL, Rockville, Md.). Plasmid DNA was purified from pooled chloramphenicol-resistant, tetracycline-sensitive colonies to form the LM2301 library. The library was transferred by electroporation of RMSK1/1 as described previously (37). RMSK1/1 colonies carrying library plasmids were screened for phage sk1 sensitivity by patching colonies to M17GCE5 plates and M17GCE5 plates spread (using a sterile glass rod) with 10⁷ PFU of phage sk1. Plasmids were isolated from phage sk1-sensitive isolates and transformed into E. coli DH5α for characterization. Plasmids were transferred by electroporation back into RMSK1/1 and tested for phage sk1 sensitivity. One of two genomic fragments that restored phage sensitivity to RMSK1/1 was named p17. Subclones were constructed in pSA3 by standard methods (52) and electroporated into RMSK1/1. Transformants were evaluated for phage sk1 sensitivity.

Phage sensitivity assay.

Strains were cultured overnight in M17GE5 broth. Lawns were prepared with 0.1 ml of culture and 3 ml of M17G top agar (0.4% agar) either with or without 30 μl of 1 M CaCl₂ (final concentration of 10 mM in the top agar). One-microliter portions of serial 10-fold dilutions of stocks of phages sk1, p2, 64, 712, and jj50 were spotted onto the lawns.

DNA sequence analysis.

Subclones of the original plasmids were constructed for sequencing as follows. The pGB305 portion of pSA3 was deleted from the library plasmids by digestion with AvaI followed by transformation of E. coli to select for recircularized plasmids bearing the insert DNA and pACYC184 as a vector. HindIII and EcoRI fragments from the pACYC184 subclones were shotgun cloned into pUC19. The pACYC184 and pUC19 subclones were sequenced with fluorescent dideoxy termination chemistry on an ABI 377 automated DNA sequencer (Applied Biosystems Incorporated, Foster City, Calif.) at the Center for Gene Research and Biotechnology at Oregon State University, Corvallis, Oreg. Primers included pUC universal primers, tetracycline resistance gene primers (5′ TACTTGGAGCCACTATCGACTACGCGATCA 3′ and 5′ ATGCGTCCGGCGTAGA 3′), and primers designed from previous sequence determinations. Primers were synthesized at the Center for Gene Research and Biotechnology on an ABI 380B DNA synthesizer using phosphoramidite chemistry (2). Sequences were assembled with Staden software (19) and analyzed with the Genetics Computer Group package (20). Similarity searches were done with the BLAST (basic local alignment search tool) programs (1) at the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/BLAST).

Construction of site-directed mutants.

The region of p17 downstream of the XbaI site was deleted by cutting p17 with XbaI. The large XbaI fragment was isolated from an agarose gel, ligated, and transformed into E. coli, forming p17ΔXbaI.

An allele of the lysL gene was constructed that contained nonsense codons in all three frames and contained an NheI restriction site as a distinguishing marker. An AvaI deletion of p17ΔXbaI created p17ΔXbaIΔAvaI, which has a single HindIII site internal to the lysL gene. p17ΔXbaIΔAvaI was digested with HindIII, treated with Klenow fragment (Gibco-BRL) in the presence of all four deoxynucleoside triphosphates, and then ligated. This created a frameshift mutation and an NheI site at the former HindIII site. The XbaI-SalI fragment (the SalI site is in the vector pACYC184) containing the lysL-nhe allele was then cloned into pGhost6 (6), an allele replacement vector with temperature-sensitive replication. L. lactis strain LM2301 was transformed by electroporation with the pGhost-lysL-nhe plasmid. Homologous integration of pGhost-lysL-nhe was selected by shifting cultures to 37°C while maintaining selection for the erythromycin resistance encoded by pGhost6. The integrated vector subsequently was excised by growing integrants without erythromycin at 30°C. Temperature-shifted colonies were screened for erythromycin sensitivity and the presence of the lysL-nhe allele by PCR with primers 5′ TTAAAACTGTTAAAGAGGTT 3′ and 5′ TACTATATTATTCATACCTTGCT 3′.

To delete nucleotides 818 to 2049, p17ΔAvaI was digested with AflIII and SwaI, treated with Klenow fragment (Gibco-BRL) in the presence of all four deoxynucleoside triphosphates, purified from a gel, ligated, and then transformed into DH5α. The XbaI-SalI fragment containing the p17Δ(Afl-Swa) allele was cloned into pGhost6 and used to replace the wild-type allele as described above. The presence of the Δ(Afl-Swa) allele was detected by analytical PCR using primer pair 5′ GGCCACAGAGGGAACTAACTATATA 3′ and 5′ CTGCAGCACATATCTTGGTTATTA 3′ (Pacific Oligos, Lismore, Australia).

Growth and cell viability.

An exponential-phase culture (optical density at 600 nm [OD₆₀₀] ≅ 0.1) of RMSK1/1 or C2 in M17G–10 mM CaCl₂ was infected with 3 × 10⁸ PFU of phage sk1 (MOI ≅ 10) and incubated at either 20 or 30°C. Identical cultures were prepared without phage. The OD₆₀₀ was measured at approximately 30- or 60-min intervals for 4.5 to 6 h. Samples of RMSK1/1 were taken immediately and 2 to 4 h after addition of phage to the infected culture. The samples were diluted and plated in triplicate on M17G plates. The number of viable cells was determined by counting the colonies on the plates after an overnight incubation at 30°C. The numbers of viable cells are expressed as averages of two separate experiments ± standard deviations (SD). Growth rates were calculated using linear regression over the interval from 0 to 240 min after addition of phage. Each curve is an average of two separate experiments (see Fig. 1).

FIG. 1.

Growth curve. An exponential-phase culture of RMSK1/1 was diluted and calcium was added. Half of the culture was infected (○) at time zero with phage sk1 at an MOI of 30 and the other half was left uninfected (▵). A culture of strain C2 was prepared in an identical manner (■, infected; ×, uninfected). Growth at 30°C was monitored by spectroscopy at 600 nm for 270 min. Each point is the mean of two separate experiments. The error bars indicate SD.

Adsorption of phage to plasma membranes in vitro.

Plasma membranes were prepared from L. lactis (50) and E. coli (40) as described earlier. Phage adsorption to membranes was measured in vitro as follows. Membranes (2 μg) (as measured by phospholipid) were mixed with approximately 2.5 × 10² PFU in 25 μl of 25 mM bis-Tris (pH 6.8) and incubated at 0°C for 1 to 2 h. Duplicate 10-μl aliquots were removed and the titers were determined (53) on strain LM2301 or LM0230. The amount of inactivation was calculated by subtracting the average titer of a reaction mixture with membranes from the average titer of a control reaction mixture without membranes, dividing by the latter, subtracting from 1.00, and multiplying by 100.

Center-of-infection experiments.

Overnight cultures of C2 and RMSK1/1 were diluted 2 × 10⁻² in M17G and grown to an OD₆₀₀ of approximately 0.1. An aliquot of each culture was diluted and plated in duplicate on M17GC plates for viable cell counts. Calcium was added to a final concentration of 10 mM and phage sk1 was added to an MOI of 3.6 × 10⁻³. A no-cell control was included by substituting M17G for bacteria and treating it the same as the infected cultures. The infected cultures were incubated at 30°C for 10 min and immediately diluted 10⁻¹ into ice-cold M17G. Half of each diluted and cooled culture was centrifuged at 12,000 × g for 3 min at 4°C, and the supernatant was removed. The titers of samples were determined in duplicate on M17GC top agar and plates using either strain C2 or C2(pSA3) as the indicator. Erythromycin (5 μg/ml) was included in the top agar and plates when C2(pSA3) was used as the indicator.

One-step growth experiment.

Cultures were prepared as described above for center-of-infection experiments, except incubation at 30°C was continued for 120 min and aliquots were taken every 10 min and diluted 10⁻². Strain C2 was used as the indicator for the plaque assay.

Electroporation with phage DNA.

DNA was isolated from phage sk1 and phage c2 as described previously (60), except that no DNase or RNase was added. Phage DNA and pSA3 DNA was electroporated into strains LM2301, C2, RMSK1/1, and MM210 as described previously (37). After 1.5 h in the recovery medium, those electroporated with phage DNA were centrifuged and titers of the supernatants were determined on LM2301. Those electroporated with pSA3 were diluted and plated on M17GE5. PFU or CFU per milliliter of recovery media were calculated.

Testing cloned DNA for replication-origin activity.

The 1.3-kb HindIII-XbaI fragment of p17ΔXbaI was subcloned into pSA3ΔHindIII. pSA3ΔHindIII lacks the origin of replication of pSA3 (associated with gram-positive organisms) and was constructed by digesting pSA3 with HindIII, removing the small HindIII fragment, and ligating the large HindIII fragment. The ligation reaction product was transformed into E. coli DH5α and selected on chloramphenicol. L. lactis LM2301 was electroporated with pSA3ΔHindIII, pSA3ΔHindIII containing the 1.3-kb HindIII-XbaI fragment, and pSA3. The electroporation frequencies were compared.

Nucleotide sequence accession numbers.

The EMBL data library accession numbers for gene sequences determined in this study are AJ132604 and LLA132604.

RESULTS

Characterization of phage sk1-resistant strain RMSK1/1.

No plaques were detected on RMSK1/1 with 936 species phage sk1, p2, jj50, or 64. However, RMSK1/1 was fully sensitive (efficiency of plating [EOP] = 1) to another 936 species phage, phage 712.

Although no plaques of phage sk1 were visible on RMSK1/1, when transformed with the cloning vector pSA3, the transformant RMSK1/1(pSA3) formed tiny (0.1-mm) plaques with an EOP of 0.1 at 20°C but formed no plaques at 30°C.

Plaques of phage sk1 formed at 20°C on RMSK1/1(pSA3) were harvested. When the phage was plated on RMSK1/1, no plaques formed. When it was plated on RMSK1/1(pSA3) at 20°C, the plaques were again much smaller than those formed on strain C2 (the parental strain of RMSK1/1), and the EOP was approximately 0.1.

The growth rate at 30°C and viable cell count of RMSK1/1 infected with phage sk1 at an MOI of 10 were not significantly different from those of an uninfected culture (Fig. 1). The growth rates (means ± SD, n = 2) were 0.75 ± 0.02 h and 0.77 ± 0.03 h for infected and uninfected cultures, respectively. The numbers (means ± SD, n = 2) of viable cells from infected and uninfected cultures at time zero were 1.14 × 10⁷ ± 0.046 × 10⁷ CFU/ml and 1.06 × 10⁷ ± 0.23 × 10⁷ CFU/ml, respectively. Two hours after infection the numbers of viable cells in the infected and uninfected cultures were 1.04 × 10⁸ ± 0.14 × 10⁸ CFU/ml and 1.10 × 10⁸ ± 0.25 × 10⁸ CFU/ml, respectively.

At 20°C, phage sk1-infected cultures (MOI = 10) of RMSK1/1 lysed in a manner identical to that of strain C2. Phage-infected cultures of RMSK1/1(pSA3) at either 20 or 30°C lysed in a manner identical to that of strain C2(pSA3) (data not shown).

The ability of RMSK1/1 to replicate phage sk1 DNA and assemble and release mature phage particles was analyzed. Phage sk1 DNA was transfected into RMSK1/1, C2, LM2301 (a plasmid-free derivative of C2), and MM210 (a strain that does not replicate phage sk1). The phage titer of transfected RMSK1/1 was 5 × 10⁸ PFU/ml, whereas the titers of transfected C2, LM2301, and MM210 were 1 × 10⁹, 2 × 10⁸, and 0 PFU/ml, respectively. Control transformations with plasmid pSA3 showed comparable transformation efficiencies for each strain.

Center-of-infection experiments showed that RMSK1/1 was infected by phage sk1, but at a reduced level. RMSK1/1 was mixed with phage sk1, and then titers of the mixture were determined using strain C2 as the indicator. The phage titer was 0.21 ± 0.08 (mean ± SD, n = 2) times the titer of a culture of C2 infected with the same amount of phage sk1. As a control for unadsorbed phage, the bacteria were removed from the mixtures by centrifugation. The phage titers of the supernatants were both 1.6% (±0.8%) of the titers of the mixtures before centrifugation. As an additional control, the titers of the mixtures and cell-free supernatants were determined on C2(pSA3) as the indicator and plated with top agar and plates that contained erythromycin. The titers of the mixtures were the same as those from the corresponding supernatants.

One-step growth experiments using strain C2 as the indicator showed that the burst size of phage sk1 in RMSK1/1 (9 ± 1 phage, n = 4) was significantly reduced and the latent period (36 ± 2 min, n = 4) significantly increased compared to those of C2 (burst size, 48 ± 11 phage; latent period, 28.5 ± 3.5 min; n = 2).

Adsorption of phage sk1 to isolated plasma membranes from either RMSK1/1 or LM0230 was measured in vitro. Membranes from either strain irreversibly adsorbed phage to the same extent (Table 2). Adsorption was not inhibited by rhamnose, which is an inhibitor of phage sk1 adsorption to the cell wall receptor (56). In addition, purified plasma membranes from E. coli did not inactivate phage sk1 (Table 2).

TABLE 2.

Plasma membrane adsorption of phage sk1 in vitro

Plasma membrane	Additive	Mean % adsorption of phage sk1 ± SD (n)
L. lactis
LM0230	None	86 ± 19 (3)
RMSK1/1	None	83 ± 14 (3)
LM0230	Rhamnose	99 ± 0.7 (2)
RMSK1/1	Rhamnose	98 ± 1.4 (2)
E. coli HB101	None	0 ± 0 (2)

Restoration of phage sk1 sensitivity to RMSK1/1.

RMSK1/1 was transformed with a library of wild-type chromosomal DNA and screened for restoration of phage sk1 sensitivity at 30°C. Despite the partial restoration of phage sensitivity at 20°C by the cloning vector pSA3 alone, two library transformants were clearly more phage sensitive than the control transformant RMSK1/1(pSA3) at 30°C. The two library transformants contained overlapping regions of DNA cloned in opposite orientations on the vector. One of the cloned fragments (named p17) is depicted in Fig. 2a. Although the cloned DNA restored the EOP to 1.0, the plaque diameter (∼0.5 mm) was about half of that formed on strain C2 or LM2301. Phage eluted from these small plaques produced larger plaques on strain C2 or LM2301 and no plaques on RMSK1/1.

FIG. 2.

Map of the region that restored phage sensitivity. (a) Restriction map of the region that restored phage sensitivity and adjacent sequences. (b) Plasmids assayed for restoration of phage sensitivity. Thick lines indicate plasmids that restored phage sk1 sensitivity to RMSK1/1. Thin lines indicate plasmids that did not restore phage sk1 sensitivity to RMSK1/1. Dashed lines indicate a deleted region. (c) Database similarities with DNA and predicted protein sequences include a lysozyme gene (lysL), a tRNA gene, and ORFs with similarities to the genes encoding the Sun protein of B. subtilis (sunL), a phosphoprotein phosphatase (pppL), and a protein kinase (prkL). (d) Distal DNA that aligns with temperate bacteriophage of L. lactis. A, AflIII; B, BamHI; H, HindIII; M, MboI; S, SwaI; X, XbaI.

The 5,832-bp insert was sequenced and found to have similarity to tRNA (for example, 73% identity over 60 nucleotides compared with tRNA from Staphylococcus aureus [35]), a phage attachment site (attP) (93% identity with lactococcal phage TP901-1 [12]), and four open reading frames (ORFs) corresponding to the following: lysozyme, the Sun protein of Bacillus subtilis (38% identity and 56% similarity over 444 residues [29]), phosphoprotein phosphatases (for example, 42% identity and 58% similarity over 242 residues compared with a probable phosphoprotein phosphatase from Mycobacterium tuberculosis [13]), and protein kinases (for example, 55% identity and 72% similarity over 293 residues compared with a putative protein kinase from B. subtilis [29]) (Fig. 2c). Subclones containing the region with similarities to tRNA, the phage attachment site, Sun protein, phosphoprotein phosphatase, or protein kinase did not restore phage sk1 sensitivity (Fig. 2b) and were not studied further. The subclone (p17ΔXbaI) that contained the lysozyme gene and the contiguous downstream 736 bp restored sensitivity of RMSK1/1 to phages sk1, p2, jj50, and 64 (Fig. 2c), although the plaque size was about half of that formed on strain C2.

Analysis of lysL.

The only ORF in subclone p17ΔXbaI was designated lysL. Proteins most similar to LysL were lysozymes (β-1,4-N-acetylmuramidase) of Chalaropsis spp. (31% identical and 44% similar [26]), Clostridium acetobutylicum (26% identical and 45% similar [15]), and Streptomyces globisporus (27% identical and 43% similar [45]) and lysins of virulent phage LL-H of Lactobacillus delbrueckii subsp. lactis (27% identical and 44% similar [47]), of lytic phages Cp-1 and Cp-9 of Streptococcus pneumoniae (30 and 28% identical and 45 and 43% similar, respectively [31, 32]), and of temperate phages mv4 of L. delbrueckii subsp. lactis (27% identical and 43% similar [58]) and adh of Lactobacillus gasseri (27% identical and 42% similar [36]) (Fig. 3).

FIG. 3.

Alignment of LysL of L. lactis with similar proteins. A consensus is calculated from four or more matched amino acids. The regions most similar to LysL are highlighted. Abbreviations: bc1, lysin of lytic phage Cp-1 of S. pneumoniae (P15057); bc9, lysin of lytic phage Cp-9 of S. pneumoniae (P19386); bll, lysin of lytic phage LL-H of L. delbrueckii subsp. lactis (L02496); bmv, lysin of temperate phage mv4 of L. delbrueckii subsp. lactis (Z26590); bph, lysin of phage adh of Lactobacillus gasseri (X78410); cha, muramidase from the fungus Chalaropsis (P00721); sgl, muramidase from S. globisporus (P25310); cac, muramidase from C. acetobutylicum (P3420); lla, lysozyme LysL from L. lactis strain LM2301; con, consensus. Accession numbers in parentheses are for either the GenBank or SwissProt database. Dots indicate gaps created to align regions of similarity. Dashes indicate lack of consensus. Asterisks indicate ends of proteins.

An alignment of the amino acid sequences of the similar lysozymes and lysins (Fig. 3) revealed that all have aspartate and glutamate at positions 5 and 32, respectively, which are thought to be at the active site of the enzymes (14, 28). The similarity of LysL to the aligned proteins ranges from 28 to 38% amino acid identity in the N-terminal regions (nucleotides 4 through 167, according to the numbering in Fig. 3) of the aligned proteins. Similarities among the aligned proteins decrease, however, distal to residue 167 of LysL. C-terminal regions of the pneumococcal bacteriophage lysins have a repeated amino acid motif due to modular recombination between host and bacteriophage lysin genes (31, 32). The autolytic lysozyme from C. acetobutylicum also has a repeating amino acid motif, but with a different periodicity (15). There were no repeating motifs found in the C-terminal region of LysL. Like the products of fungal, clostridial, and phage genes, LysL lacks an amino-terminal signal sequence (59).

A potential promoter upstream of lysL, which deviates from the consensus lactococcal promoter (21) in two positions, was identified at nucleotides 202 through 207 (−35 region) and 225 through 230 (−10 region). A potential ribosome-binding site (21) was identified at nucleotides 447 through 455. A region of imperfect dyad symmetry, which might serve as a rho-independent transcriptional terminator, was identified downstream of lysL at nucleotides 1322 through 1332 and 1337 through 1347.

A frameshift allele (lysL-nhe) was constructed and transformed into RMSK1/1. p17ΔXbaI containing either the mutated lysL-nhe allele or wild-type lysL restored phage sk1 sensitivity to RMSK1/1. The mutated allele was used to replace the wild-type lysL allele in L. lactis LM2301. The resultant lysL mutant was fully sensitive to phage sk1.

Analysis of the lysL-sunL intergenic region.

Starting 309 bp distal to lysL was DNA similar to that of temperate lactococcal bacteriophage of the P335 species (Fig. 2d), including BK5-T (8), rlt (57), lc3 (46), and Tuc2009 (3). These regions of the phage genomes contain sequences involved in transcription termination and homologous recombination and are from 173 to 197 bp distal to the respective lysin genes. An alignment of these similar nucleotide sequences revealed a 266-bp stretch of DNA that interrupts the region of similarity in L. lactis (Fig. 4). The 266-bp intervening sequence is bordered by a set of inverted repeats and contains another set in the exact middle of the sequence.

FIG. 4.

DNA distal to lysL aligned with temperate bacteriophage of L. lactis. Abbreviations: 1852, sequence of DNA distal to lysL that begins at nucleotide 1852; 1477, sequence of DNA distal to lysL that begins at nucleotide 1477; LC3, nucleotides 1435 to 1623 of bacteriophage LC3, DNA sequence accession no. L31348; BK5-T, nucleotides 10967 to 12091 of bacteriophage BK5-T, DNA sequence accession no. L44593; TUC, nucleotides 1503 to 1632 of Tuc2009, DNA sequence accession no. L31348. Accession numbers are from GenBank. Dashes indicate lack of consensus. Tildes indicate flanking nucleotides without similarity.

Nucleotides 1169 to 2209 are 42% identical to the phage sk1 genome nucleotides 22109 to 23198. This region of the phage genome includes four ORFs of unknown function, the early promoter E5, and the 3′ end of a putative phage DNA polymerase subunit. Nucleotides 2051 through 2154 of the cloned DNA are 58% identical to phage sk1 DNA from nucleotides 17501 through 17604. This region of the phage DNA includes the early promoter E6 and a partial ORF of unknown function.

Beginning at nucleotides 2209 through 2258 is DNA with similarities to tRNA genes. The predicted secondary structure lacked a consensus anticodon stem structure, however.

Distal to the tRNA gene starting at nucleotide 2313 is 13 bp of DNA that is 85% identical to attP of the temperate P335 species bacteriophage TP901.

A deletion in p17ΔXbaI was constructed that spanned nucleotides 818 to 2049 and is termed p17ΔAfl-Swa. This deletion construct did not restore phage sensitivity to RMSK1/1 in trans. Furthermore, the same region was deleted from the chromosome of LM2301 and did not confer resistance to phage sk1.

To detect the potential origin of replication function, nucleotides 609 to 1903 were cloned into a plasmid (pSA3ΔHindIII) that lacked an origin of replication associated with gram-positive organisms but included a selectable marker in L. lactis. No transformant of strain C2 or RMSK1/1 could be isolated, although the same amount of control plasmid (pSA3) produced between 50 and 4.5 × 10³ transformants/ml.

DISCUSSION

This study shows that the resistance of strain RMSK1/1 to infection by phage sk1 is caused by a blockage of phage DNA entry into the host. This conclusion is supported by evidence that the resistance is not due to lack of adsorption to either the cell wall or plasma membrane (55) (Table 2). The lack of inhibition of adsorption by rhamnose shows that there was no binding to the cell wall in the membrane preparation. Specificity of adsorption of phage sk1 to lactococcal membranes was shown by the lack of adsorption to membranes from E. coli. The inactivation of phage sk1 shows that adsorption to membranes in vitro was irreversible and that the phage had committed the final step leading to ejection of its DNA from the capsid (5). Circumventing by electroporation the blocked entry of phage DNA shows that steps of phage propagation subsequent to entry into the host cytoplasm are normal. Additionally, the lack of cell death upon infection (Fig. 1) eliminates a classical abortive infection mechanism (22, 33). At an MOI of 10, all cells were infected, but only about 21% of them lysed, as indicated by the efficiency of the center-of-infection assay. Moreover, the restriction and/or modification mechanism cannot account for the results (22), because plaques that formed on RMSK1/1(pSA3) did not plaque with an EOP of 1, nor did they form plaques on RMSK1/1. Of the four known mechanisms of resistance in L. lactis (27, 33, 42), only blockage of phage DNA entry is consistent with our results.

The normal growth rate and cell viability of phage sk1-infected cultures of RMSK1/1 (Fig. 1) were somewhat surprising, considering that RMSK1/1 can replicate phage, as indicated by results from both electroporation and the center-of-infection experiments. The absence of centers of infection on the indicator strain C2(pSA3) showed that centers which formed on indicator strain C2 resulted not from a reversible dissociation of phage from the surface of RMSK1/1 but instead from infection and replication of phage sk1 in RMSK1/1.

An explanation for this is that the reduced frequency of infection of RMSK1/1 coupled to the reduced burst size and increased latent period results in arithmetic instead of exponential increase in phage titer. As exponential cell growth occurs, the proportion of infected cells decreases, making it difficult to detect small changes in cell numbers by optical density or viable cell count.

The resistance of RMSK1/1 to four of five 936 species phages shows that many, but not all, 936 species phage are inhibited by the mechanism of resistance in RMSK1/1. We speculate that the genome of phage 712 differs from that of the other phages in a region that may be a target for blocking phage DNA entry into the host. Alternatively, phage 712 may have an inherently larger burst size and shorter latent period.

The phage sk1 resistance of strain RMSK1/1 was suppressed in three ways. First, reducing the growth temperature to 20 from 30°C may have inactivated a cold-sensitive host component associated with the mechanism of resistance. Alternatively, it may have had an indirect effect via a temperature-induced change in cell wall or membrane structure or via reduced growth rate or metabolism. RMSK1/1 was also rendered partially phage sk1 sensitive by the plasmid pSA3. This effect could have been caused indirectly by a reduced growth rate or metabolism. Recent results indicate that RMSK1/1 is resistant to phage sk1 at 20°C in growth media other than M17G (B. Geller, unpublished results). This suggests that the mechanism of resistance is not cold sensitive.

The third means of suppression of phage sk1 resistance of RMSK1/1 was the introduction of a region of genomic DNA of L. lactis LM2301. The region responsible for suppression was subcloned and found to contain a gene with similarity to lysozymes plus the contiguous downstream 736 bp of noncoding DNA with similarity to temperate phage of L. lactis.

The predicted LysL protein was similar to lysins of virulent phage LL-H of L. delbrueckii subsp. lactis (47), of lytic phages Cp-1 and Cp-9 of S. pneumoniae (31, 32), and of temperate phages mv4 of L. delbrueckii subsp. lactis (58) and adh of L. gasseri (36). lysL is distinct from acmA (9) and lysA (44), genes encoding autolytic lysozymes in L. lactis strains MG1363 and AM2, respectively. lysL is also distinct from the lysin gene of phage sk1 (11). Although the involvement of host lysozymes in phage release is not without precedent (51), disruption of the lysL gene did not affect phage sk1 sensitivity. LysL is not involved in the mechanism of phage sk1 resistance in RMSK1/1 or in the restoration of sensitivity by the cloned DNA.

The DNA distal to lysL contained noncoding DNA with similarity to DNA distal to lysin genes of temperate lactococcal bacteriophage of the P335 species. In these phage, this region contains terminator sequences and sequences involved in homologous recombination. An alignment of these similar nucleotide sequences revealed a 266-bp stretch of DNA that disrupts the region of similarity and contains two sets of inverted repeats. The symmetry of the inverted repeats in the 266-bp stretch suggests a possible insertion sequence or antitermination machinery.

The similarity of lysL with lysin genes of Lactobacillus and Streptococcus and the similarities of downstream regions with temperate lactococcal phage of the P335 species suggest that this may be the site of previous prophage integration. Lysogeny is the rule rather than the exception for lactococcal strains (16, 18, 54). Indeed, strain C2, the parent of strains RMSK1/1, LM0230, and LM2301, harbors a prophage that produces a small, isometric virion upon induction with UV light (54). The sequences upstream of lysL are not similar to upstream sequences of phage lysin genes, and the regulatory sequences present in lytic and temperate phages of L. lactis (60) were not present in the 5,832 bp sequenced. These differences suggest that this is not a functional prophage but rather a region that had been the site of previous prophage integration activity.

The DNA distal to lysL was responsible for the suppression of the phage sk1 resistance of RMSK1/1. This is evident from the fact that deletion of this region from the cloned plasmid eliminated suppression of phage sk1 resistance in RMSK1/1. Both orientations of the cloned DNA in the cloning vector restored phage sensitivity. This suggests that transcription of the DNA that restored phage sensitivity from a promoter in the vector is not required.

Deletion of the region distal to lysL from the chromosome of LM0230 did not cause phage sk1 resistance. This shows that a deletion in this region of the chromosome is not the cause of phage sk1 resistance in RMSK1/1.

The mechanism by which this noncoding DNA restores phage sk1 sensitivity is unknown. However, the characteristics of this cloned region are similar to those of a region of phage 31 DNA that is involved in its sensitivity to an abortive infection mechanism, AbiA (23). Both regions lack ORFs, contain two sets of inverted repeats, and eliminate phage resistance when cloned in trans. Experimental data suggest that the phage 31 region does not bind directly to AbiA but may require a phage-specific factor for its inhibitory effect. When introduced on a multicopy plasmid, the cloned DNA may exert its effect by mimicking part of the phage genome and adsorbing a protein or RNA that blocks phage propagation. Perhaps in phage-sensitive strains a single, chromosomal copy of the cloned DNA would be insufficient to remove enough of the putative injection-blocking component to cause phage resistance. In any case, the identification of an endogenous DNA sequence that suppresses the phage resistance of RMSK1/1 will be useful in elucidating the molecular details of the mechanism of phage resistance.