The Genetic Switch Regulating Activity of Early Promoters of the Temperate Lactococcal Bacteriophage TP901-1

Recombinant plasmid DNA from E. coli was isolated by the alkaline lysis technique. Plasmids were isolated from Lactococcus lactis subsp. cremorisby the alkaline lysis technique after incubation with lysozyme (20 mg/ml) in 20 min at 37°C. Preparative portions were further purified by passage through Qiagen columns as recommended by the supplier (Qiagen, Hilden, Germany). Pharmacia Biotech supplied restriction endonuclease enzymes, DNA polymerase Klenow fragment, T4 DNA ligase, and buffer systems. All enzymes were used as recommended by the supplier. PCRs were performed in a DNA thermal cycler (Perkin-Elmer Cetus) using Amplitaq polymerase and buffer supplied by Perkin-Elmer. DNA sequencing was performed by the method of Sanger and coworkers (24) as instructed by the protocol for the Sequenase version 2.0 DNA sequencing kit (U.S. Biochemical Corp., Cleveland, Ohio).

The plasmids used in this study are listed in Table 1. Plasmid pPM92 was constructed by cloning a 2,787-bp XhoI fragment of pG5f4 (4) into the XhoI site of the E. colivector pIC19R (20). This plasmid contains orf4 toorf9 of TP901-1. The following plasmids were constructed by cloning PCR products obtained by using pPM92 as a template, a vector primer (1211; see below), and primers located either in the region between orf4 and orf5 (24BamHI, P5A, and erm2ABamHI), within orf5 (pextPstI), or downstream oforf5 (orf4orf5PstI and porf7PstI). All clonings were performed in L. lactis subsp. cremoris MG1363. Plasmids pAJ93 and pAJ94 were constructed by cloning aBamHI-digested PCR product (primers 24BamHI and erm2ABamHI) into pAK80 (13) also digested with BamHI. The resulting plasmids contain p_L (pAJ93) andp_R (pAJ94) fused to the lacLM genes. PCR products obtained by using primer 1211 and pextPstI were cloned into pAK80 after digestion with PstI, giving rise to plasmids pPM126 and pPM137 carrying orf4 as well asp_L and p_R fused to thelacLM genes, respectively. Plasmids pPM127 and pPM132 were constructed by cloning a HindIII-digested PCR product (primers 1211 and porf7PstI) into pAK80. The plasmids thus containorf4 and orf5 as well asp_L (pPM127) or p_R(pPM132) fused to the lacLM genes. A frameshift mutation was introduced in the beginning of the orf4 gene by filling in the SpeI site of pPM92 with Klenow polymerase. After ligation and redigestion with SpeI, two PCRs were performed on the ligation mixture. In the first PCR, primers 1211 and pextPstI were used, and a PCR product containing the mutated orf4gene was cloned into pAK80 after digestion with PstI, giving rise to plasmids pPM138 and pPM136, carrying only the mutatedorf4 gene as well as p_L andp_R fused to the lacLM genes, respectively. In the second PCR, primers 1211 and orf7PstI were used. After digestion with HindIII, a PCR product containingorf5 in addition to the mutated orf4 was cloned in pAK80, and plasmids pPM131 and pPM139 were constructed. Plasmids pPM131 and pPM139 thus contain p_L andp_R, respectively, fused to the lacLMgenes. In all cases, several independent clones giving rise to the same β-galactosidase specific activity were isolated. Furthermore, the inserts of plasmid pAJ93, pAJ94, pPM127, pPM126, and pPM132 were sequenced, as well as the mutation within the orf4 gene in plasmids pPM131, pPM136, pPM138, and pPM139. The PCR product generated by use of primer 1211 and erm2ABamHI was digested with BamHI and cloned into pCI372 (10) also digested withBamHI, resulting in plasmid pAJ80, carryingp_L, p_R, andorf4. The PCR product produced by using primer 1211 and P5A was cloned into pNZ8010 (7) after digestion withBamHI, giving rise to pAJ115 containing orf4 andp_L fused to the gusA gene. PCR products produced by using primer 1211 and orf4orf5PstI or Orf5A and Orf5B were cloned into pNZ8010 (7) after digestion withPstI and BamHI-PstI, respectively. Thereby plasmids pAJ98 and pAJ142 were constructed so as to containorf4 and orf5 transcribed fromp_R and p_L on a high-copy-number plasmid and orf5 under the control of the nisin promoter, respectively. The inserts of pAJ80, pAJ98, pAJ115, and pAJ142 were sequenced.

For construction of plasmids, the following primers were used: 1211 (5′-GTAAAACGACGGCCAGT-3′), 24BamHI (5′-GCGCGGATCCTATAGCGCATCTTGAAC-3′), P5A (5′-GCGCGGATCCTAATCATAACTCATTTATGT-3′), erm2ABamHI (5′-GCGCGGATCCCACGTTTCATGAACTTT-3′), orf4orf5PstI (5′-GCGCCTGCAGCCCAAGCTTCATCAGTTC-3′), Orf5A (5′-GCGCGGATCCAAGAAAGGAGAAACATAAAT-3′), Orf5B (5′-GCGCCTGCAGTTAATGAACTTTTGCATTAA-3′), pextPstI (5′-AAAACTGCAGAGACACAGTTCTCTCTG-3′), and porf7PstI (5′-AAAAAACTGCAGATCCCCCATGTGCTTTCC-3′). For primer extensions, primer PE (5′-AGACACAGTTCTCTCTGAAAGCCCC-3′) was used for mapping of the p_L promoter, while primer 23 (5′-GGTCAGGAGATTGAACG-3′) was used forp_R.

E. coli XL1-Blue was made competent with CaCl₂ and was transformed as described by Sambrook et al. (23). Selection was performed with 100 μg of ampicillin per ml. L. lactis subsp.cremoris MG1363 was transformed by electroporation according to the method described by Holo and Nes (12), using 0.03 to 0.5 μg of DNA per electroporation. When L. lactis subsp.cremoris 3107 was made competent, the growth medium (12) contained only 0.2 M sucrose, and 25 μg DNA was used per transformation. Transformants were selected on plates containing 5 μg of erythromycin per ml. Screening on 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) plates was performed at a concentration of 90 μg/ml.

Bacterial strains used in this study are listed in Table 2. The temperate bacteriophage TP901-1 was induced from L. lactis subsp.cremoris 901-1 (1) by the use of UV light as previously described (4).

E. coli strains were grown with agitation at 37°C in Luria-Bertani broth (23) (Difco Laboratories, Detroit, Michigan). Bacto Agar (Difco Laboratories) was used at 1.5% (wt/vol) in solid media. L. lactis strains were propagated at 30°C without shaking in GM17 (M17 broth [Oxoid Limited, Basingstoke, Hampshire, United Kingdom] containing 0.5% [wt/vol] glucose) (26). For determination of β-galactosidase activity, exponential growing cells were diluted in fresh medium to an optical density at 600 nm (OD₆₀₀) of 0.05. At an OD₆₀₀ of 0.8, 70 ml of culture was harvested and resuspended in 7 ml of Z-buffer, and cell extracts were obtained with a French press. The activity of β-galactosidase was determined according to Miller (21), and specific activity was calculated in Miller units.

Exponentially growing cells of strains LKH169 and LKH206 were diluted in fresh medium (GM17 containing 5 μg of erythromycin per ml and 5 μg of chloroamphenicol per ml) to an OD₆₀₀ of 0.05. At an OD₆₀₀ of 0.2, nisin was added to the culture to a final concentration of 1 ng/ml. Samples were withdrawn before and after addition of nisin, diluted, and plated on plates containing X-Gal but no nisin. After overnight incubation at 30°C, blue and white colonies were counted.

Strains were grown in GM17 containing 5 μg of erythromycin per ml at 28°C, due to the temperature sensitivity of the recA strain. Exponentially growing overnight cultures were diluted to an OD₆₀₀ of 0.05 in fresh medium, and growth was monitored. At an OD₆₀₀ of 0.4, each culture was diluted into two separate cultures to an OD₆₀₀ of 0.1. Subsequently, mitomycin C was added to a final concentration of 2.5 μg/ml to one of the cultures, while the other served as a control. Samples were removed at different time points, harvested, extracted, and assayed as described above for β-galactosidase activity.

TP901-1 phages were diluted in A buffer (0.8% [wt/vol] NaCl, 10 mM Tris-HCl [pH 7.5], 5 mM MgCl₂, 1 mM CaCl₂) and plated on lawns of L. lactissubsp. cremoris 3107 transformed with either pPM126, pPM138, pPM127, or pPM131. As a control, L. lactis subsp.cremoris 3107 was used as an indicator strain. After incubation at 30°C overnight, plaques were counted. For each strain, the efficiency of plaquing (EOP) was calculated as the number of plaques formed on a lawn of the strain of interest divided by the number of plaques formed on L. lactis subsp.cremoris 3107.

RNA was isolated from strains PM93, PM94, and PM80 growing exponentially in defined medium SA (14) (1.0% glucose [wt/vol], 5 μg of erythromycin per ml), using the method described by Johnsen et al. (15) as modified by Madsen and Hammer (19).

Primers PE and 23 were phosphorylated by using T4 polynucleotide kinase (Gibco BRL) and [³²P]ATP (Amersham) as described by Sambrook et al. (23); 25 μg of total RNA isolated from strain PM93, PM94, or PM80 was mixed with 1 pmol of phosphorylated primer in a volume of 10 μl and incubated at 70°C for 10 min. The mixture was allowed to cool to 37°C over a 1-h period. Then 1 μl of Superscript reverse transcriptase (Gibco), 4 μl of 5× first-strand buffer (Gibco), 2 μl of 0.1 M dithiothreitol (Gibco), and dATP, dGTP, dCTP, and dTTP to a final concentration of 500 μM were added, and the mixture was incubated at 37°C for 1 h.

Previously, we identified two divergently orientated promoters (p_L andp_R) located between orf4 andorf5 of the temperate bacteriophage TP901-1 (19). To investigate the expression and regulation of the promoters, we cloned a phage DNA fragment of 837 bp, containing both promoters andorf4, in both orientations in the promoter probe vector pAK80. The lacLM genes contained in this plasmid give rise to β-galactosidase activity when a functional promoter is inserted in the cloning region, and stop codons in all three reading frames ensure that a transcriptional fusion is constructed. Whenp_L (pPM126) and p_R(pPM137) are located upstream of the lacLM genes, low β-galactosidase activities (1.0 and 0.5 Miller units, respectively) can be measured when the plasmids are present in L. lacticsubsp. cremoris MG1363 (Fig.1). These enzyme levels are, however, significantly different from the basal level of 0.02 Miller units measured in MG1363 containing the pAK80 vector, which is in agreement with the existence of the divergently orientated promotersp_L and p_R on this phage DNA fragment. The effect of the presence of orf4 on the activity of the promoters was investigated by introduction of an early frameshift mutation at codon position 6. The β-galactosidase activities of the resulting plasmids, carrying thep_L fusion (pPM138) and thep_R fusion (pPM136) were measured as 62 and 15 Miller units, respectively, in MG1363. These results demonstrate the presence of divergently orientated promoter activities on the cloned phage DNA fragment and clearly show that the presence of a functional ORF4 protein represses the expression of both phage promoters. The levels of repression are 62- and 31-fold for p_Land p_R, respectively (Fig. 1).

Since the spacer between the initiation sites of thep_L and p_R transcripts is only 101 bp, it is possible that binding of the RNA polymerase to one of them affects transcription of the other. To exclude this possibility, a 111-bp PCR fragment carrying thep_L promoter but only the −35 region of thep_R promoter was cloned in both orientations upstream of the lacLM genes in the promoter probe vector pAK80. The specific activity of β-galactosidase of L. lactis subsp. cremoris MG1363 containing either pAJ94 (p_R fusion) or pAJ93 (p_Lfusion) was measured. The presence of pAJ94 (p_Rfusion) gave rise to only a background level of β-galactosidase activity (0.14 Miller units), as expected due to the absence of the −10 region of the p_R promoter. MG1363 containing pAJ93 (p_L fusion), in which thep_R promoter is not active, showed the same specific β-galactosidase activity (66.5 Miller units) as when the larger p_L fusion (pPM138) carrying an activep_R promoter was present. Thus, the activity of the p_L promoter is the same regardless of whether the p_R promoter is present, showing that the activity of the p_R promoter does not affect the efficiency of the p_L promoter.

In bacteriophage λ, the Cro and the cI repressors compete for the λ operator sites and thereby regulate expression of the p_RM,p_R, and p_L promoters. To investigate the influence of both ORF4 and ORF5 on thep_L and p_R promoters of TP901-1, we cloned a 1,279-bp phage DNA fragment carrying bothorf4 and orf5 in both orientations in pAK80, resulting in plasmids pPM127 (p_L fusion) and pPM132 (p_R fusion). The transformants arising from both plasmids exhibited clonal variation when screened on GM17 plates containing X-Gal. Of 200 transformants arising from pPM127 (p_L fusion), strong blue colonies were observed in 191 cases and the remaining 9 colonies were white, giving a frequency of 4.5% of white colonies. This frequency is comparable to the frequency of lysogenization of approximately 1 to 3% by TP901-1 (3). When plasmid pPM132 (p_R fusion) was used, the same type of clonal variation was observed but with opposite frequencies: 95% white colonies and only 5% light blue transformants were observed. The phenotype was stable for both plasmids when restreaked to single colonies more than three consecutive times and also after freezing and restreaking. When plasmid DNA was isolated from either blue or white transformants and used for retransformation of MG1363, segregation into white and blue transformants was observed again with the same frequencies as seen before. The phenotype of a transformant was thus independent of the phenotype of the strain from which the plasmid was prepared.

In light of the observed variation of phenotype when bothorf4 and orf5 were present, we repeated the transformation of MG1363 with plasmids pPM126 (p_L fusion) and pPM137 (p_R fusion), containing only orf4. Screening of the transformants on 90 μg of X-Gal per ml made it possible to visualize the difference in enzyme levels of 0.5, 1.0, and 62 Miller units as light blue, blue, and very blue colonies, respectively, all being distinctively different from the white colonies. Again 200 transformants containing either pPM126 (p_L fusion) or pPM137 (p_Rfusion) were obtained and all were either light blue (pPM137) or blue (pPM126), corresponding to the measured enzyme levels of 0.5 and 1.0 Miller units (Fig. 1). Clonal variation was therefore not observed in the presence of only ORF4, suggesting that ORF5 is necessary for this phenomenon.

To assess the effect of the repression mediated by ORF4 in the presence of ORF5, orf4 mutants were constructed in both plasmids by filling in the SpeI site in the N-terminal region oforf4. β-Galactosidase activities of the transformants containing the clonal variants of pPM127 (p_Lfusion) and of the clonal variants of pPM132 (p_Rfusion) as well as the plasmids carrying the orf4 mutation were measured (Fig. 1). First, let us consider expression from the lytic promoter p_L. In the presence oforf4 and orf5, the expression fromp_L observed in PM94 is about 1,000-fold repressed compared to the activity found in the derepressed state (PM93) (Fig. 1). Furthermore, the p_L promoter activity in PM93 (containing orf4 and orf5) is as high as in the orf4 mutant (PM80); thus, PM93 is phenotypically an orf4 mutant (Fig. 1). Expression fromp_R is also subject to clonal variation. However, in the presence of both orf4 and orf5, the difference between the repressed state of the p_Rpromoter in strain PM81 and the derepressed state in strain PM95 is only fourfold. When a mutation was introduced into orf4, the expression from p_R becomes even more derepressed, 439-fold compared to the expression found in strain PM81 (Fig. 1). Thus, also in the presence of ORF5, ORF4 is needed for repression of both p_R andp_L.

The effect of ORF4 and ORF5 donated from a plasmid in trans to the promoter fusions contained on the compatible pAK80 plasmid was investigated. As observed in Table 3, the presence oforf4 on the donor plasmid results in repression of both ap_L fusion (pPM138) and ap_R fusion (pPM136) in the recipient strain, whereas the vector plasmid itself (pCI372) does not change the constitutive expression monitored as blue colonies. Thus, ORF4 is able to repress both p_L and p_Rwhen given in trans. Subsequently, the effect of donation of a plasmid containing both orf4 and orf5 (pAB223) to either the p_R or thep_L fusion strain was investigated. The result clearly showed clonal variation, with the majority of thep_L fusions turning blue and the majority of thep_R fusions turning white. Thus, the same effect of ORF4 and ORF5 was observed whether orf4 andorf5 were donated in trans or were located on the same plasmid as the promoters (pPM127 and pPM132). Regulation ofp_L and p_R thus seems to be dependent on ORF4 and ORF5, both being able to function intrans.

The clonal variation mediated by the presence of both ORF4 and ORF5 was found to be stably inherited. Particularly, expression fromp_L is affected by the choice of configuration, since an approximately 1,000-fold difference in the specific activities of β-galactosidase of strains PM93 (p_L open) and PM94 (p_L closed) was found (Fig. 1). The ratio between the amount of ORF4 and ORF5 could determine the choice of configuration of p_L, and we therefore examined whether the state (repressed or open) of the p_Lpromoter could be changed by introduction of additional amounts of either ORF4 or ORF5. For this purpose, strain LKH148 carrying pPM127 (orf4 orf5), in which p_L is open, was transformed with plasmids expressing orf4 from thep_R promoter, and the color of the transformants was observed on plates containing X-Gal. In the presence oforf4 on a low-copy-number plasmid (pAJ80), only blue transformants were seen, whereas 92% white and 8% blue colonies were found when orf4 was present on a high-copy-number plasmid (pAJ115) (Table 3). It is thus possible to change an open state of thep_L promoter to a repressed state by introduction of a sufficient amount of additional ORF4.

In the situation where p_L is closed (strain LKH149 carrying pPM127 [orf4 orf5]) we investigated the effect of increasing the expression of orf5 first by introducing a plasmid containing both orf5 andorf4 transcribed from p_L andp_R, respectively (pAB223). However, this did not change the repressed state of the p_L promoter located on pPM127 in LKH149, even when a high-copy-number plasmid was used (Table 3). This was expected since a surplus of ORF4 is able to shut down expression of the p_L promoter andorf5 expression is dependent on p_L in pAB223. To be able to produce a surplus of ORF5, a plasmid containingorf5 under the control of the inducible promoterp_nisA was therefore constructed (pAJ142). Addition of the inducer nisin will lead to a rapid increase inorf5 expression. For this experiment we used strain LKH206, which contains pAJ142 in addition to the p_Lfusion on pPM127 (orf4 orf5). The advantage of this experimental setup is that the influence of a short induction period with nisin in the liquid culture on the final choice of the phenotype of a colony originating from a cell diluted from this culture and plated without nisin can be tested. Thus, after dilution, the color of LKH206 was examined on plates containing X-Gal but no nisin before and after nisin was added to the culture (Fig.2). Before addition of nisin, 4 to 6% blue colonies were seen, probably due to a low frequency of initiation of transcription of the nisin promoter also in the absence of nisin. After 5 min of growth in the presence of nisin, 51% of the colonies were blue if samples were removed and plated. This frequency of blue colonies was increased further to 84% after 10 min, and after 40 min all colonies were blue. The phenotype of the blue colonies obtained in this experiment was stable after restreaking to single colonies. This shows that a repressed state of p_L can be changed into an open state if the amount of ORF5 is increased only momentarily (10 min) and that once established, this configuration can be maintained without further nisin-induced ORF5 production.

Next we wanted to show that overproduction of ORF5 is also able to inhibit the ORF4-mediated repression of the p_Lpromoter in the short fusion pPM126, which contains only 145 bp of thep_L transcript, in contrast to the 587 bp in pPM127. Strains containing pPM126 (orf4 p_Lfusion) and pAJ142 (carrying orf5 under the control of thep_nisA promoter) or pNZ8010 (vector of pAJ142) were grown in the presence of nisin, and the specific activity of β-galactosidase was measured. An 11-fold increase in expression fromp_L in the presence of expressed orf5was found. Thus, ORF5 is also able to inhibit the ORF4-mediated repression of the p_L promoter in the short fusion contained in pPM126.

In a previous study, locations of the divergent promotersp_L and p_R in the region between the divergent reading frames in the early expressed region of TP901-1 were determined by primer extension (19). The results from the transcriptional fusions presented in the present report have also demonstrated the presence as well as the regulation of divergent promoter activity. By primer extension analysis, we were furthermore able to simultaneously monitor expression from thep_R and p_L promoters in the same strain. Therefore, we decided to map the transcriptional start sites from both promoters by primer extension using mRNA prepared from three strains: PM93 and PM94, both containing pPM127 but carrying thep_L promoter in the derepressed state and the repressed state, respectively, as well as PM80, containing the corresponding plasmid with a mutation in orf4 (pPM131). In PM93 (orf4 and orf5 present), expression from thep_L promoter is high, as shown by the β-galactosidase measurements. The mRNA prepared from this strain shows an mRNA start site from the p_L promoter but none from the p_R promoter (Fig.3, lanes 1). On the other hand, when thep_L promoter is in the repressed state as in the clonal variant PM94 containing the same plasmid (pPM127) as PM93, transcription from p_R but not fromp_L is observed (lanes 2). For comparison, we investigated the primer extension results from the strain containing the orf4 mutant plasmid, pPM131. Since the promoter fusions showed that ORF4 represses transcription from both promoters, we would expect to find transcriptional start sites from both promoters in this strain. As shown in lanes 3, this was also observed. The transcripts from p_R are 4 bp longer due to filling in of theSpeI site, which is located between the primer used and the promoter. Therefore, the same transcriptional start site is used for expression from p_R in strains PM94 and PM80, just as in the case of p_L expression in both PM93 and PM80. The results for strains containing pPM127 (orf4 and orf5 present) demonstrate that the two different phenotypes which can be obtained with this plasmid result from a strain where p_L is open andp_R is simultaneously closed, and vice versa: a strain where p_L is closed andp_R is open. However, as judged from the enzyme measurements in Fig. 1, p_R is only partially derepressed in this situation.

The effect of ORF4 and ORF5 on the plating capacity of phage TP901-1 was investigated after transformation of the relevant plasmids into the phage indicator strain L. lactis subsp. cremoris 3107. We tested the presence of ORF4 alone by using pPM126 (p_Lfusion) and, as a control, the corresponding orf4 mutant plasmid pPM138. A strain containing orf4 (LKH51) showed full resistance to TP901-1 and an EOP of less than 10⁻⁶, compared to the EOP of 1 for both the orf4 mutant strain (PM97) and the indicator strain 3107 (Fig.4). The effect of the presence of bothorf4 and orf5 was tested with pPM127 (p_L fusion). Use of this plasmid also gave rise to clonal variation in the indicator strain 3107. LKH48 and LKH49 are representatives of clones containing repressed and derepressedp_L promoters, respectively. No plaques were detected with LKH48, while a few pinpoint plaques were obtained with LKH49. For comparison, we also used PM99 containing the orf4mutant plasmid pPM131, which should produce ORF5 fromp_L since orf5 is located downstream of p_L, which in this strain has a high activity (plasmid pPM131 in strain PM80 [Fig. 1]). As reported in Fig. 4, the presence of this plasmid is inhibitory to the plaque-forming capacity of TP901-1, giving an EOP of 10⁻⁴. This is only about fivefold higher than the EOP on LKH49, indicating that the inhibitory effect in LKH49 is mainly due to the production of ORF5.

For bacteriophage λ, it is known that during the maintenance of lysogeny the cI repressor is bound to the p_R promoter, preventing transcription of the genes involved in lytic development. In the presence of UV light or mitomycin C, which both induce the SOS response, the cI repressor is inactivated by autocleavage induced by the RecA protein (17). To investigate whether transcription can be induced from the clone containing the lyticp_L promoter of TP901-1 by treatment with mitomycin C, β-galactosidase activity was monitored after addition of mitomycin C to strain PM94, carrying both orf4 andorf5. Strain PM94 is the clonal variant carrying thep_L promoter in the repressed state. As seen in Fig. 5, the β-galactosidase activity was induced 28-fold by mitomycin C. Furthermore, mitomycin C induction of the p_L promoter fusion was found to be dependent on a functional recA allele, since no induction was seen in a recA mutant (LKH68 in Fig. 5). However when PM75, the recA⁺ strain containing thep_L fusion plasmid pPM126 with onlyorf4, was exposed to mitomycin C, no induction of β-galactosidase activity was observed (data not shown).