- Research article
- Open Access
Characterization and genome sequencing of two Propionibacterium acnes phages displaying pseudolysogeny
BMC Genomics volume 12, Article number: 198 (2011)
Propionibacterium acnes is a Gram positive rod inhabiting the human skin that also infects orthopaedic implants and is associated with acne vulgaris. Previously, one lytic bacteriophage, PA6, from P. acnes has been sequenced and partially characterized. We recently isolated several inducible phages from P. acnes classified as Siphoviruses based on morphology and partial genome sequencing.
In this study we sequenced the inducible P. acnes phages PAD20 and PAS50, isolated from deep infection and from skin, respectively. The genomes of PAD20 and PAS50 are 29,074 and 29,017 bp, respectively, compared with the 29,739 bp of PA6. The phage genomes have 87.3-88.7% nucleotide sequence identity. The genes are divided into clusters with different levels of similarity between the phages. PAD20 and PAS50 share four genes encoding identical amino acid sequences. Some deletions and insertions in the genomes have occurred, resulting in lack of genes, frame shifts, and possible regulatory differences. No obvious virulence factor gene candidates were found. The phages are inducible, but bacteria can be cured of phages by serial colony isolations and lose their phages during stationary phase, but are still sensitive to new phage infections. Construction of a phylogenetic tree based on more than 459 phage genomes, suggested that P. acnes phages represent a new lineage of Siphoviruses.
The investigated P. acnes Siphovirus genomes share a high degree of homology to other P. acnes phages sequenced, but not to genomes of other phages isolated from Propionibacteria. The phage genomes are not integrated in the bacterial genome, but instead, most likely have a pseudolysogenic life cycle.
Propionibacterium acnes is a Gram positive rod normally found on human skin , but has also been isolated from infections of orthopaedic implants [2, 3], and has been suggested to be involved in the development of acne vulgaris. Recently, the first genome of a P. acnes strain was sequenced  and several putative virulence factors were identified . The sequenced P. acnes genome only contains one cryptic prophage, but it has earlier been shown that some P. acnes strains are carriers of inducible phages . Other groups have also studied Propionibacterium phages both in order to establish a tool to differentiate between serotypes, but also to further characterize phages mainly infecting bacteria important for the dairy industry [7–16].
Recently, Farrar et al, sequenced the first genome of a P. acnes phage, PA6 (GenBank accession no. DQ431235) . The genome consists of 29,739 bp dsDNA with 48 putative coding genes. The phage was determined to be a lytic phage since no integrase or repressor could be detected in the genome. Eighteen of the genes were assigned predicted functions based on their similarities to other known phage genes. Recently, we isolated and characterized inducible phages from P. acnes isolated from both deep infections and from skin of healthy individuals . By comparing the genomes of these different phages, a better understanding of how the phage influences its host and an insight in the great genetic diversity seen among phages, as shown by other comparative studies on phage genomes can be achieved [18, 19].
Phages can enter two principally different life cycles, e.g. a lytic life cycle or a lysogenic cycle. In the lysogenic cycle most phages integrate their genome into the host genome. This is most often regulated by a phage-borne integrase, but there is also a recent report of a filamentous phage, CTXϕ from Vibrio cholerae, that uses the host recombinases XerC and XerD to integrate into the host genome [20, 21]. Furthermore, there are a few phages known to form stable linear or circular plasmids as a part of their integration mechanism [22–24], but most known temperate phages encode their own integrases.
Other than those two well-defined phage life cycles, phages are also known to be able to interact with their host in other life cycles, termed as persistent infections, pseudolysogeny, carrier-state, and chronic infections [25, 26]. Those phage life cycles, even though common, are less defined compared to the lytic and lysogenic life cycles. Common for all states, except the chronic infection, is that the bacteria carry the phage DNA, but the DNA is not integrated into the bacterial chromosome, but rather exists as a plasmid. The chronic infection is characterized by the release of phages from the bacteria without causing lysis .
In this study, we have sequenced the genome of one phage induced from a P. acnes strain isolated from deep infection (PAD20), and the genome of one phage induced from a P. acnes strain isolated from skin (PAS50), and we have compared the two genomes with the reported lytic phage PA6 . The phages have earlier been shown to have dissimilarities in the genes encoding a putative major head protein and an amidase , but no other genes have been examined. This offers an opportunity to investigate if phages isolated from P. acnes contribute to the virulence seen in different strains of P. acnes, and to examine the genetic diversity seen among phages.
Results and discussion
Genomic organization and homology
The genomic organization of the three examined phages are very similar, even though some genes differ (see Figure 1). Both phages have a 13 bp ssDNA extension at both ends of the genome, identical to that of phage PA6. These extensions enable the formation of potential concatamers and circularization of the phage genome. The genome of PAD20 is 29,074 bp and has a GC content of 54.1%, compared to the genome of PAS50 that is 29,017 bp and has a GC content of 53.9%. PAD20 has two larger deletions in the genome compared to PA6. The first is a 226 bp deletion resulting in the deletion of the main part of gp22 and thereby in a frame shift giving rise to a combination of the gene products Gp22 and Gp23 (Gp22.23, see additional file 1). The second larger deletion is a 378 bp deletion of the entire gp42 gene. PAS50 has two larger deletions and one insertion. The first deletion in PAS50 is a 207 bp deletion resulting in the deletion of the gp45 gene, and a second deletion of almost 500 bp in a non-coding region, corresponding to bp 27,932-28,419 in PA6. The insert in PAS50 is an 84 bp insertion in gp37 resulting in a 28 amino acid larger exonuclease.
To further investigate the genomic organization of the different P. acnes phages, the genomes were screened for putative sigma70-dependent promoters and transcriptional terminators (see Figure 1 and Table 1). Several putative promoters were detected in all phages upstream of gp7, gp11 and gp46, as well as upstream of gp45 and gp30, potentially enabling the phage to transcribe different parts of the genome at different stages of infection. One of the promoters with the highest score was upstream of a hitherto unannotated ORF (3693-3301, gp50). However, the predicted gene product does not show any similarity to known sequences or domains.
Except for the six identified promoters in each genome, one putative transcriptional terminator was also identified (see Figure 1). The putative terminators are placed immediately downstream of the amidase and the holin in the lysis gene cassette.
Even though most of the genes are very similar in the examined phages, a certain pattern of similarity between gene clusters can be observed. Packaging genes and putative head genes have a higher similarity between PAD20 and PAS50 than between the other examined P. acnes phage groups. Other than the clusters mentioned above, gp27-gp48 also shows the highest similarity between PAD20 and PAS50, except for gp42 and gp43. This exception is mainly due to the deletion of gp42 in PAD20. Furthermore, most tail genes and genes involved in the lysis of the host have the highest similarity between PA6 and PAS50. The overall genome nucleotide similarity is very high between all phages examined (PA6/PAS50 - 88.7%, PAD20/PAS50 - 87.8%, and PA6/PAD20 - 87.3%).
By searching for homologues corresponding to PAD20 and PAS50 proteins using BLASTp , most hits were in the Actinobacteria phylum, with Streptomyces, Mycobacterium and phages infecting Mycobacterium, Propionibacterium, Nocardia, and Bifidobacterium (see Table 2). This is not surprising since these species are closely related and thereby also might have similarities among their phages. Most hits were on bacterial proteins, most likely indicating that these bacterial strains have prophages that have not yet been examined. More interestingly, several phage proteins showed similarity to phages isolated from low G+C bacteria as Streptococcus and Lactococcus (see Table 2). These species are not very closely related to P. acnes, but the results might indicate that their phages have had several possibilities to exchange genes. Surprisingly, there were no hits on any other Propionibacterium phages or prophages. If this merely reflects the lack of sequenced phages from Propionibacteria and bacterial genomes in the Propionibacterium genus, or indicate that phages isolated from P. acnes are distinct from those infecting other Propionibacteria is unknown. Furthermore, the putative cryptic prophage in P. acnes strain KPA171202  does not show any similarity to PAD20, PAS50 or PA6. However, most recently, Machuca et al isolated a phage infecting Fusobacterium nucleatum. This phage was partially sequenced (390 bp), and showed 90% sequence identity to parts of gp3 (portal protein) and gp4 in PA6. However, the whole phage genome of the Fusobacterium phage needs to be sequenced to conclude how related this phage is to Propionibacterium phages.
Genes encoding packaging proteins and structural proteins
The packaging and structural proteins (Gp1-Gp19) are generally very well conserved, with high sequence similarities between all three phages, ranging from 89.7% to 99.3% amino acid identity (see Figure 1 and additional file 2) with one exception - gp19. Gp19 has 67.1% to 77.4% amino acid identity and is predicted to function as a tail protein, with high similarity to collagen. This protein is likely to be involved in host specificity, so a more diverged gene would imply that the phages might have differences in host specificity. The difference in host specificity between PAD20 and PAS50 is however not very pronounced, since both PAD20 and PAS50 are equally potent to initiate lysis in most examined P. acnes isolates, even though some isolates can be used to differentiate the phages . This does not rule out that a larger screen would detect more differences in host specificity and more clearly point out the genetic varieties that determine the host specificity.
Another gene of particular interest is gp17 encoding a protein that has not yet been assigned a putative function. Gp17 has a H-type lectin domain with E-values ranging from 0.001-0.04 in the phages examined. Even though the E-values are rather low, they might still give an indication of the putative function. H-type lectins are known to interact with carbohydrates, and the genomic position of gp17, close to genes encoding tail proteins, makes it possible that this protein is also a tail protein interacting with carbohydrates in order to bind to certain structures and potentially mediate a first binding site for the phage.
The lysis mechanism
All three phages encode an amidase, Gp20, with very high similarity to a P. acnes autolysin. They also encode a putative holin (Gp21) with more pronounced sequence divergence. Phage PA6 and PAS50 have 94% amino acid identity in the holin protein, while PA6 and PAD20 only have 84.3% amino acid identity. Interestingly, phage PAD20 has a deletion in gp22 resulting in a mainly deleted gp22 and a new open reading frame encoding a protein with similarities to both Gp22 and Gp23 in PA6 (see additional file 1). gp22 will not be fully translated in PAS50 due to two internal stop codons.
Identical proteins and putative virulence factors
Several genes (gp39, gp40, gp44 and gp47) encode identical amino acid sequences in PAD20 and PAS50, but none of these proteins have predicted functions, nor any significant BLASTp hits, so these genes seem to be specific for P. acnes phages, and might, due to their sequence similarity be a good tool to use in phylogenetic analyses of phages isolated from P. acnes. gp29, a gene in PA6, is absent in both PAD20 and PAS50, but another ORF in the same region (gp29.1, 19,490-20,056 in PA6) exists in all three phages and might thereby be more likely to be a true coding gene. However, none of the putative gp29.1 genes have similarities to other genes, neither has gp29 in PA6.
Some proteins did get BLAST hits indicating that the proteins could be putative virulence factors. Among those are Gp34 that is annotated as a DNA helicase due to a DnaB helicase domain (E-value 2e-04 for PAD20). More interesting, this protein also has a Multidrug Resistance Protein-like transporter domain (E-value 7e-06 for PAD20). This domain is less pronounced in PAS50 (E-value 0.15) and PA6 (E-value 0.001). PAS50 also has a weak hit (E-value 0.21) in Gp30 on a hyaluronidase. However, due to the low similarities to other proteins, and thereby the low E-values, we can not assign a function to the phage proteins based on the similarities to other domains, only determine which domains they are most similar to.
The phages might also be able to circularize and form concatamers due to the ssDNA extensions on each side of the genome. When circularized, a new putative gene (gp49) can be transcribed. This gene encodes a protein with some similarity to different nucleoproteins from Ebola virus, and has a topoisomerase II-associated protein PAT-1 domain (E-value 0.005 for PAS50). Thus, this gene can only be transcribed when the phage has circularized and potentially integrated, or formed concatamers.
Identification of phage proteins
To verify that some of the genes encoded proteins, PAD20 was precipitated with PEG 6000, the proteins were separated on a polyacrylamide gel, and then analyzed with MS/MS. All visible bands but band F were identified (see Figure 2 and Table 3). One of the bands, band E, was identified as a P. acnes succinate dehydrogenase flavoprotein subunit, thus indicating that this protein exists as a large complex, thereby being eluted in the same fractions as the whole phage particles during the FPLC gel filtration. Other than the P. acnes flavoprotein, all proteins were identified as phage proteins. Bands A-D were all identified as the major head protein of PAD20 (Gp6), which has a calculated molecular weight of approximately 33 kDa. However, the bands A-D have a much higher molecular weight, starting from approximately 160 kDa for band D, reaching over 500 kDa for band A. Interestingly, we did not get any other hits on the major head protein with lower molecular mass proteins, indicating that these proteins tend to form complexes of 6-8 monomers, based on the apparent molecular weight, that are not disrupted even when boiling the samples with SDS and reducing agents. A similar phenomenon has been described before with the Enterobacteria phage HK97, which has complexes of covalent linked major head proteins, and not a single major head protein exist as a monomer . Both the N-and C-terminal part of the major head protein were covered, showing that this protein is not processed after its translation, as some major head proteins require in order to establish their final conformation.
Other proteins that were identified were Gp14 (tape measure protein), Gp3 (portal protein), Gp16, Gp15 (minor tail protein) and Gp11 (major tail protein. Furthermore, band K did not result in a significant hit. However, the peptide with the highest score was similar to Gp17 from P. acnes phage PA6. This is further strengthen by that the apparent molecular weight of the protein band (29 kDa) is close to the theoretical molecular weight (28,835 Da). Thus, all phage proteins identified seem to be structural proteins, dominated by Gp6 (major head protein) and Gp11 (major tail protein).
Phage life cycle
The phages used in this study were originally isolated by mitomycin C treatment of P. acnes isolates. However, those phages were able to infect their host strains, and thus did not confer superinfection immunity . This data made us conclude that the phages were lysogenic. A similar phenomenon has been reported for Leuconostoc oenos phages, where more than 65% of the bacterial isolates could be induced with mitomycin C and where the phages were able to infect their host strains .
To further investigate how stable the interaction between the P. acnes phages and the bacterium was, we performed ten serial single-colony isolations and then screened for phages using a PCR assay. All examined colonies had been cured from phages (data not shown), indicating that the phages are not true lysogens. This ability to cure the bacterial isolates from phages suggests that the interaction between the phage and bacterium rather might be pseudolysogenic [31–33]. The inability to cure bacteria from phages was one of the criteria used to determine that phages from the P. acnes related bacterium Nocardia erythropolis were lysogenic, due to that the spontaneous curing of phages was less than 0.4% after 5 serial single-colony isolations . These phages were also induced with mitomycin C, but they all showed superinfection immunity. It has also been shown in several experiments that the environment may influence the virus-host interaction. This is evident in the case of phages infecting the archeal species Halobacterium, where the NaCl concentration determines if the phages enter a lytic cycle, or cause a persisting infection. At low NaCl concentration, the phages effectively kill the host, while at higher concentrations of NaCl the lysis is delayed for several bacterial generations [35–37], leading to a persistent infection.
To further investigate if P. acnes might harbour pseudolysogenic phages, single colonies from the P. acnes isolates AD20 and AS50 were inoculated into BHI broth and different parameters such as OD600 and pfu were measured two times a day (see Figure 3 and Table 4). No free infectious phages were detected at any time point, indicating that if any phages were spontaneously induced, their number have not exceeded the detection limit of 100 pfu/ml. However, at time points corresponding to mid-log phase and late-log phase, both AD20 and AS50 had 10-60% colonies harbouring phage DNA, as judged by PCR amplification of the major head gene (see Table 4). However, this carriage of phages is lost during the stationary growth phase, thus indicating that the phage DNA is not being replicated at this time point. Furthermore, the cells at the various time points were sensitive to PAD20 and PAS50 infection, indicating that the loss of phages were not due to the development of resistance (data not shown). This implies that the phages are not strictly lytic, neither strictly lysogenic.
Much of the data indicates that the P. acnes phages might have a pseudolysogenic cycle, which is also supported by the genetic information. There are no genes on the phage genomes with similarities to known integrases or repressors. It is, however, known that certain filamentous phages can integrate their genomes using host recombinases [20, 21]. However, when extracting P. acnes chromosomal DNA, phage DNA is lost, as judged by the absence of a positive PCR reaction for the phage major head gene (data not shown), indicating that the phage indeed does not integrate. Thus, it seems like P. acnes phages can persist as extra-chromosomal plasmids, however when in this state, they are not replicating, since they easily are cured after 10 serial single-colony isolations.
Phage phylogenetic analysis
Bacteriophages can be divided into different groups based on different criteria as host range, morphology and genetic material. Earlier attempts to classify phages according to their proteome have been conducted by Rohwer et al using 105 phage genomes . We aligned 453 bacteriophage genomes, including PAD20 and PAS50, using the progressive Mauve algorithm  in order to investigate how the P. acnes phages were related to other phages. Mauve recognizes conserved nucleotide segments in the genomes, but has the advantage that it can rearrange the order of these segments, thereby also identifying genome rearrangements . However, Mauve has several limitations when it comes to comparing several diverged phage genomes, due to that Mauve works best with related organisms, and can not identify conserved regions shared only by a fraction of the genomes. Thus, aligning the genomes using Mauve will not generate a reliable phylogenetic tree. Instead, we used the progressive Mauve algorithm, which is developed to be able to align more diverged genomes with less than 50% nucleotide identity. Our results are in good agreement with the results obtained by Rohwer et al, but some of the phages are placed differentially. The Sk1-like Siphoviruses with Lactococcus phage sk1 and phage c2 form two closely related lineages, and also the lambda-like Siphoviruses align in a similar way as Rohwer et al. However, Lactococcus phage r1t is more closely related to Lactococcus phage Tuc2009 and TPP901-1 in our alignment than was showed previously. Furthermore, Acinetobacter phage AP205 is not as diverged from the other Leviphages in our model as Rohwer shows. Other than that, most phages cluster in the same groups as was shown by Rohwer et al, even though we have several more phages and subgroups of phages, resulting in not as distinct groupings as in Rohwer's model. We found that the P. acnes phage genomes sequenced so far form a new phylogenetic lineage, well separated from other phages, with no close relatives (see Figure 4 and additional files 3 and 4). Furthermore, the different families of phages are widely distributed over all classes of bacteria, with Siphoviruses being the most abundant family reported in the database. Remarkably, even though PAD20 and PAS50 have a morphology that classifies them as Siphoviruses, some of their closest related phages are Propionibacterium freudenreichii phage B5 (inovirus), Acinetobacter phage AP205 (Levivirus) and Pseudomonas phage phi6 segment L (Cystovirus). Furthermore, a group of podoviruses including Yersinia phage Berlin seem to be closely related to a group of myoviruses with Vibrio, Aeromonas and Enterobacteria phages. Also of interest is that most phages infecting Mycobacterium and Streptomyces were rather closely related to phages from Propionibacterium. This is not surprising, since the bacterial hosts also are related. Of more interest is that Proteobacteria such as Pseudomonas, Burkholderia and Ralstonia phages also were related to phages infecting the Actinobacteria phylum, while other Proteobacteria as Vibrio and Salmonella were not. Other than that, bacterial families tend to have closely related phages, even though the classification of the phage families can differ in a bacterial species, as the very closely related Mycobacterium phage Rizal and Rosebush, that are classified as a Myovirus and a Siphovirus, respectively. Thus, this clearly shows that classification of phages solely based on morphology is not sufficient to determine to which family a phage should be assigned, but that genomic data is also necessary.
We have found that phages isolated from P. acnes are highly similar and show high nucleotide sequence identity. Several of the phage structural proteins were identified using MALDI-TOF MS/MS. The sequenced phages do not seem to encode any putative virulence factors. Finally we could show that the phages seem to be able to cause pseudolysogeny in their hosts.
Bacteriophages and bacterial isolates
Bacteriophages PAD20 and PAS50 were isolated as previously described . Stock solutions were stored at 4°C in SM-buffer (20 mM Tris-HCl pH 7.5, 100 mM NaCl, 60 mM MgSO4 (7H2O), 0.01% gelatin). Bacterial isolates KPA171202, AD12, AD20, AD21, AD33 and AS50 were plated from frozen stocks on Tryptic Soy Broth supplemented with 1.5% agar (Bacto, Mt Pritchard, NSW, Australia) and incubated at 37°C under anaerobic conditions until colonies were clearly visible.
DNA sequencing and analysis
Since we had problems getting high concentrations of pure phage DNA, we adapted a PCR based approach. Phage lysate (1010 PFU/ml) was boiled for 5 min and 1 μl added as a DNA template to Fermentas High Fidelity PCR Enzyme Mix (Vilnius, Lithuania). Primers were designed to generate 4-6 kbp long overlapping products based on the already sequenced P. acnes phage PA6 genome, covering the whole genome (see Table 5). The overlapping regions were in most cases more than 200 bp, and confirmatory PCRs were run over the whole genome to exclude that we by this approach missed any inserts. The PCR was run under the following conditions: 94°C 2 min, 10 cycles of 94°C 30 sec, annealing temperature 30 sec, 72°C 6 min. This was followed by 35 repeats of 94°C 30 sec, annealing temperature 30 sec and 72°C 6 min (+20 sec/repeat). PCR products were cleaned up with SpinPrep PCR Clean-up Kit (Novagen, Madison, WI, USA). Products were sequenced using BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) and analyzed using an ABI 3100 Genetic Analyzer (Applied Biosystems). For each sequencing, a new PCR generated DNA template was used, minimising the impact of the PCR. When assembled, a >3-fold coverage with identical nucleotides was achieved. Sequences were assembled and aligned using the Clustal W algorithm . Analyses were performed using the MacVector v9.5.2 software package (Cary, NC, USA). The sequence of the ssDNA ends was determined by similarity to phage PA6 after a PCR over the ends of the phage genome.
Identification of promoter regions and terminator structures
The genome of PAD20, PAS50 and PA6 were screened for putative sigma70-promoters using SAK and BPROM (Softberry, Inc.). Using SAK , a value higher than 1 was regarded as positive. Terminator structures were identified using FindTerm (Softberry, Inc.) and EMBOSS Explorer . Palindromes followed by a high content of U and with a stable secondary structure (less than -20.0 in minimum free energy) were considered to be transcriptional terminators.
Phage protein identification
The phages were purified according to Boulanger , with some modifications. Briefly, bacterial lysates were treated with 0.5 M NaCl for 1 h at 4°C to separate phage particles from bacteria. The solution was centrifugated (10 min, 8,000 g, 4°C) and the supernatant moved to a new tube. PEG 6000 was added to a final concentration of 10%, and incubated for 1 h at 4°C. The phages were pelleted by centrifugation at 10,000 g for 15 min at 4°C. The supernatant was discarded and the pellet resuspended in PBS. Whole phage particles were then separated from bacterial proteins using gelfiltration (Superose 6 10/300 GL, GE Healthcare, Sweden, Uppsala) operated on a ÄKTAprime plus (GE Healthcare). Protein fractions containing phage particles were eluted early during the run due to their high molecular mass. Fractions with phage particles were pooled and precipitated with 10% PEG 6000 for 1 h at 4°C. The pellet was resuspended in PBS, mixed with SDS loading buffer, boiled for 3 min and loaded on a 10% SDS-PAGE gel . The gels were run at 170 V for 1 h and stained with 0.25% Coomassie Brilliant Blue. All visible bands were cut out using a sterile scalpel. The gels spots were destained with 50 μ1 50%.acetonitrile (ACN) in 25 mM NH4HCO3 and dried using a speed vacuum for 10-15 min. To the dried spots, 10 mM DTT dissolved in 100 mM NH4HCO3 was added and incubated for 1 hour at 56°C. The DTT was removed and 55 mM iodoacetamide dissolved in NH4HCO3 was added and incubated for 45 min at room temperature in the dark. The gel pieces were washed several times with NH4HCO3, followed by dehydration with ACN and were dried using a speed vacuum. The gel pieces were incubated with 12.5 ng/μl trypsin (Promega) in 50 mM NH4HCO3 at 4°C for 45 minutes. The digestion continued over night at 37°C, and was terminated by the addition of 5% trifluoroacetic acid in 75% ACN. Samples were subjected to HPLC-MS/MS, by analyzing them on a Qtof Ultima API (ESI-MS/MS from Waters, Manchester, UK) coupled to a CapLC (Waters). The peptides were injected and trapped on a pre-column (C18, 300 μm × 5 mm, 5 μm, 100Å, LC-Packings), and separated on a reversed phase analytical column (Atlantis, C18, 75 μm × 150 mm, 3 μm, 100Å, Waters). The flow rate through the column was 250 nl/min. Solvent A was 2% ACN in water with 0.1% formic acid, and solvent B was 90% ACN, 10% water and 0.1% formic acid. The HPLC started with 5% solution B for 5 min, increased the concentration of B to 60% in 40 min, and to 80% B in 5 min. The concentration of B was kept at 80% for 25 min and finally at 5% B for 15 min. The mass spectrometer analysis was made by Data Dependent Acquisition (DDA). The mass range m/z was from 50-1800 for MS/MS. Only spectra from ions with charge states 2 and 3 were acquired. Generated files were used to search in MASCOT MS/MS Ions search using a peptide tolerance of 0.6 Da and a MS/MS tolerance of 0.8 Da. Hits on peptides were regarded positive if they had a p-value below 0.05.
Serial single-colony isolation
AD20 and AS50 were cultured on Brain Heart Infusion Agar plates (BHIA-plates) for three to four days, until visible single colonies were seen. Ten single colonies from each bacterium were selected, and ten serial streak isolations were made. The presence of phages in the final isolated colonies was detected by a PCR assay screening for the well-conserved major head gene, as described elsewhere .
Growth curve of P. acnes isolates
P. acnes isolates AD20 and AS50 were plated on BHIA-plates for three to four days. Single colonies were inoculated into pre-reduced BHI broth, and samples were analysed every 12th hour for OD600, cfu, infectious phages (pfui), and carriage of phages (pfuc). Colony forming units were measured by dilution of the bacterial culture and plating on BHIA-plates. Pfui was measured by sterile filtration of the undiluted growth medium, using overlay plates as described elsewhere . The detection limit using this method was >100 pfu/ml. To investigate the percentage of bacteria carrying phage DNA (pfuc), single colonies at each time point were subjected to PCR for the identification of the major head gene, using the recA gene as a positive control, as described elsewhere . All experiments were performed in triplicates.
Phage phylogenetic analysis
The software Geneious (Geneious Pro 4.6.6, Biomatters Ltd) was used to download 451 different phage genomes from GenBank, searching for genomes with 'phage' in the title. The genomes for Propionibacterium phage PAD20 and PAS50 were added, and aligned using the software Mauve  with progressive Mauve, using the default settings. The generated guide-tree was then visualized with Geneious. For better visualization, 28 phage genomes were selected based on their localisation and their relationship with P. acnes phages on the complete phylogenetic tree, and a new alignment with the same settings was run with only these phages.
Nucleotide sequence accession number
The GenBank accession numbers for the bacteriophages PAD20 and PAS50 genomes are FJ706171 (protein accession number ACX30796-ACX30840) and FJ706172 (protein accession number ACX30841-ACX30886), respectively.
McGinley KJ, Webster GF, Leyden JJ: Regional variations of cutaneous propionibacteria. Appl Environ Microbiol. 1978, 35 (1): 62-66.
Zeller V, Ghorbani A, Strady C, Leonard P, Mamoudy P, Desplaces N: Propionibacterium acnes: an agent of prosthetic joint infection and colonization. J Infect. 2007, 55 (2): 119-124. 10.1016/j.jinf.2007.02.006.
Tunney MM, Patrick S, Gorman SP, Nixon JR, Anderson N, Davis RI, Hanna D, Ramage G: Improved detection of infection in hip replacements. A currently underestimated problem. J Bone Joint Surg Br. 1998, 80 (4): 568-572. 10.1302/0301-620X.80B4.8473.
Brüggemann H, Henne A, Hoster F, Liesegang H, Wiezer A, Strittmatter A, Hujer S, Dürre P, Gottschalk G: The complete genome sequence of Propionibacterium acnes, a commensal of human skin. Science. 2004, 305 (5684): 671-673. 10.1126/science.1100330.
Brüggemann H: Insights in the pathogenic potential of Propionibacterium acnes from its complete genome. Semin Cutan Med Surg. 2005, 24 (2): 67-72. 10.1016/j.sder.2005.03.001.
Lood R, Mörgelin M, Holmberg A, Rasmussen M, Collin M: Inducible Siphoviruses in superficial and deep tissue isolates of Propionibacterium acnes. BMC Microbiol. 2008, 8 (1): 139-10.1186/1471-2180-8-139.
Webster GF, Cummins CS: Use of bacteriophage typing to distinguish Propionibacterium acne types I and II. J Clin Microbiol. 1978, 7 (1): 84-90.
Pulverer G, Sorgo W, Ko HL: Bakteriophagen von Propionibacterium acnes. Zentralbl Bakteriol Parasitenk Infektionskr Hyg 1 Abt Orig A. 1973, 225: 353-363.
Pulverer G, Ko HL: Fermentative and serological studies on Propionibacterium acnes. Appl Microbiol. 1973, 25 (2): 222-229.
Gautier M, Rouault A, Sommer P, Briandet R, Cassin D: Bacteriophages infecting dairy propionibacteria. Lait. 1995, 75: 427-434. 10.1051/lait:19954-532.
Chopin MC, Rouault A, Ehrlich SD, Gautier M: Filamentous phage active on the gram-positive bacterium Propionibacterium freudenreichii. J Bacteriol. 2002, 184 (7): 2030-2033. 10.1128/JB.184.7.2030-2033.2002.
Cheong JPE, Brooker JD: Isolation of a virulent bacteriophage from a Propionibacterium species in the sheep rumen. Aust J Agric Res. 2000, 51: 119-123. 10.1071/AR99069.
Borisova TG, Baranova NA, Vorobyeva LI, Trusova ZM, Istomina IB, Gavrina TP, Terebkova LS: A phage of propionic bacteria (Russian). Prikl Biokhim Mikrobiol. 1973, 9: 246-249.
Zierdt CH: Properties of Corynebacterium acnes bacteriophage and description of an interference phenomenon. J Virol. 1974, 14 (5): 1268-1267.
Jong EC, Ko HL, Pulverer G: Studies on bacteriophages of Propionibacterium acnes. Med Microbiol Immunol. 1975, 161 (4): 263-271. 10.1007/BF02122714.
Hervé C, Coste A, Rouault A, Fraslin JM, Gautier M: First evidence of lysogeny in Propionibacterium freudenreichii subsp. shermanii. Appl Environ Microbiol. 2001, 67 (1): 231-238. 10.1128/AEM.67.1.231-238.2001.
Farrar MD, Howson KM, Bojar RA, West D, Towler JC, Parry J, Pelton K, Holland KT: Genome sequence and analysis of a Propionibacterium acnes bacteriophage. J Bacteriol. 2007, 189 (11): 4161-4167. 10.1128/JB.00106-07.
Kwan T, Liu J, DuBow M, Gros P, Pelletier J: The complete genomes and proteomes of 27 Staphylococcus aureus bacteriophages. Proc Natl Acad Sci USA. 2005, 102 (14): 5174-5179. 10.1073/pnas.0501140102.
Lee CN, Hu RM, Chow TY, Lin JW, Chen HY, Tseng YH, Weng SF: Comparison of Genomes of Three Xanthomonas oryzae Bacteriophages. BMC Genomics. 2007, 8 (1): 442-10.1186/1471-2164-8-442.
Huber KE, Waldor MK: Filamentous phage integration requires the host recombinases XerC and XerD. Nature. 2002, 417 (6889): 656-659. 10.1038/nature00782.
McLeod SM, Waldor MK: Characterization of XerC- and XerD-dependent CTX phage integration in Vibrio cholerae. Mol Microbiol. 2004, 54 (4): 935-947. 10.1111/j.1365-2958.2004.04309.x.
Inal JM, Karunakaran KV: f20, a temperate bacteriophage isolated from Bacillus anthracis exists as a plasmidial prophage. Curr Microbiol. 1996, 32 (4): 171-175. 10.1007/s002849900030.
Girons IS, Bourhy P, Ottone C, Picardeau M, Yelton D, Hendrix RW, Glaser P, Charon N: The LE1 bacteriophage replicates as a plasmid within Leptospira biflexa: construction of an L. biflexa-Escherichia coli shuttle vector. J Bacteriol. 2000, 182 (20): 5700-5705. 10.1128/JB.182.20.5700-5705.2000.
Eggers CH, Casjens S, Hayes SF, Garon CF, Damman CJ, Oliver DB, Samuels DS: Bacteriophages of spirochetes. J Mol Microbiol Biotechnol. 2000, 2 (4): 365-373.
Weinbauer MG: Ecology of prokaryotic viruses. FEMS Microbiol Rev. 2004, 28 (2): 127-181. 10.1016/j.femsre.2003.08.001.
Porter K, Russ BE, Dyall-Smith ML: Virus-host interactions in salt lakes. Curr Opin Microbiol. 2007, 10 (4): 418-424. 10.1016/j.mib.2007.05.017.
Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ: Basic local alignment search tool. J Mol Biol. 1990, 215 (3): 403-410.
Machuca P, Daille L, Vinés E, Berrocal L, Bittner M: Isolation of a novel bacteriophage specific for the periodontal pathogen Fusobacterium nucleatum. Appl Environ Microbiol. 2010, 76 (21): 7243-7250. 10.1128/AEM.01135-10.
Popa MP, McKelvey TA, Hempel J, Hendrix RW: Bacteriophage HK97 structure: wholesale covalent cross-linking between the major head shell subunits. J Virol. 1991, 65 (6): 3227-3237.
Tenreiro R, Santos R, Brito L, Paveia H, Vieira G, Santos MA: Bacteriophages induced by mitomycin C treatment of Leuconostoc oenos strains from Portuguese wines. Letters in Applied Microbiology. 1993, 16: 207-209. 10.1111/j.1472-765X.1993.tb01398.x.
Welsch M: Lysogenicity in streptomycetes. Ann N Y Acad Sci. 1959, 81: 974-993.
Bastias R, Higuera G, Sierralta W, Espejo RT: A new group of cosmopolitan bacteriophages induce a carrier state in the pandemic strain of Vibrio parahaemolyticus. Environ Microbiol. 2010, 12 (4): 990-1000. 10.1111/j.1462-2920.2010.02143.x.
Jarling M, Bartkowiak K, Robenek H, Pape H, Meinhardt F: Isolation of phages infecting Actinoplanes SN223 and characterization of two of these viruses. Appl Microbiol Biotechnol. 2004, 64 (2): 250-254. 10.1007/s00253-003-1473-6.
Crockett JK, Brownell GH: Isolation and characterization of a lysogenic strain of Nocardia erythropolis. J Virol. 1972, 10 (4): 737-745.
Torsvik T, Dundas ID: Persisting phage infection in Halobacterium salinarium str. 1. Journal of general virology. 1980, 47: 29-36. 10.1099/0022-1317-47-1-29.
Daniels LL, Wais AC: Virulence in phage populations infecting Halobacterium cutirubrum. FEMS Microbiology Ecology. 1998, 25: 129-134.
Daniels LL, Wais AC: Ecophysiology of Bacteriophage S5100 Infecting Halobacterium cutirubrum. Appl Environ Microbiol. 1990, 56 (11): 3605-3608.
Rohwer F, Edwards R: The Phage Proteomic Tree: a genome-based taxonomy for phage. J Bacteriol. 2002, 184 (16): 4529-4535. 10.1128/JB.184.16.4529-4535.2002.
Darling AC, Mau B, Blattner FR, Perna NT: Mauve: multiple alignment of conserved genomic sequence with rearrangements. Genome Res. 2004, 14 (7): 1394-1403. 10.1101/gr.2289704.
Introduction to Mauve. [http://asap.ahabs.wisc.edu/mauve-aligner/mauve-user-guide/introduction-to-mauve.html]
Thompson JD, Higgins DG, Gibson TJ: CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 1994, 22 (22): 4673-4680. 10.1093/nar/22.22.4673.
Gordon L, Chervonenkis AY, Gammerman AJ, Shahmuradov IA, Solovyev VV: Sequence alignment kernel for recognition of promoter regions. Bioinformatics. 2003, 19 (15): 1964-1971. 10.1093/bioinformatics/btg265.
Rice P, Longden I, Bleasby A: EMBOSS: the European Molecular Biology Open Software Suite. Trends Genet. 2000, 16 (6): 276-277. 10.1016/S0168-9525(00)02024-2.
Boulanger P: Purification of bacteriophages and SDS-PAGE analysis of phage structural proteins from ghost particles. Methods Mol Biol. 2009, 502: 227-238. 10.1007/978-1-60327-565-1_13.
Laemmli UK: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature. 1970, 227 (5259): 680-685. 10.1038/227680a0.
This work was supported by grants from the Swedish Research Council (project 2005-4791 and 2010-57X-20240), the Foundations of Crafoord, Jeansson, Zoégas, Bergvall, Österlund, Wiberg, Groschinsky, Söderberg, Gustaf V's Memorial Fund, the Scandinavian Society for Antimicrobial Chemotherapy, the Swedish Society for Medicine, and the Royal Physiografic Society. MC is the recipient of an Assistant Professorship from Swedish Research Council. We acknowledge the help with the protein analysis from the SCIBLU Proteomics Resource Centre at Lund University.
RL participated in the design of the study, performed the experiments and drafted the manuscript. MC designed the study and revised the manuscript. Both authors read and approved the final manuscript.
Electronic supplementary material
Additional file 1:Nucleotide sequence alignment for gp22 and gp23 in P. acnes phage PA6, PAD20 and PAS50. Underlined sequences represent different genes, as indicated by the colour of the line. Phage PAS50 has two internal stop codons (red) in the gene encoding gp22 and will not generate a full-length protein. Phage PAD20 has a deletion in a major part of the gene encoding gp22 resulting in a frame shift and a combined gene gp22.23. (TIFF 780 KB)
Additional file 4:Phylogenetic tree of more than 450 bacteriophage genomes. More than 450 bacteriophage genomes were aligned using progressive Mauve. The generated tree-file was then visualized using the software Geneious The different colours represents different families of phages - bright green (Siphovirus), red (Myovirus), blue (Podovirus), pink (Inovirus), light blue (Levivirus), gray (Cystovirus), purple (Tectivirus), brown (Microvirus), pale green (Corticovirus), orange (Plasmavirus) and black (non-assigned). (TIFF 6 MB)
About this article
Cite this article
Lood, R., Collin, M. Characterization and genome sequencing of two Propionibacterium acnes phages displaying pseudolysogeny. BMC Genomics 12, 198 (2011). https://doi.org/10.1186/1471-2164-12-198
- Phage Particle
- Phage Genome
- Phage Protein
- Putative Virulence Factor