Model system for quantification of encapsidated bacterial genes
Staphylococcus aureus COL (pT181, pUSA300-HOUMR-like) was susceptible to the selected transducing phages classified into the serological group B (φ80, φ80α, φ11). To compare transducing abilities of the phages, phage φ81 of the serological group A that was previously considered a non-transducing phage (Dowell and Rosenblum, 1962) was selected and was modified to grow on the donor strain. Using multiplex PCR targeting the integrase, structural capsid and amidase bacteriophage genes (Kahánková et al., 2010), the presence of a serogroup A prophage from the COL strain in the transducing lysates was excluded.
In order to detect and quantify bacterial genes inside the phage particles, it was essential to test the integrity of phage particles and exclude the presence of exogenous bacterial DNA in the individual samples (Appendix S1). Presence of the added exogenous DNA of Bifidobacterium animalis was not detected in the steps following the enzymatic treatment by RNase A and DNase I, which confirmed that the samples did not contain any free unpackaged bacterial DNA and that the obtained results describe only detection of bacterial, plasmid and phage genes localized inside the phage particles. Analogous approach has been used in one of the first studies demonstrating encapsidation of DNA of the host strain by the cyanomyovirus S-PM2 performed in the marine cyanobacterium Synechococcus, where it has been proven that is able to package a region of kanamycin resistance cassette into its capsids (Clokie et al., 2003).
The abilities of phages to package and transduce bacterial DNA from the donor strain were first tested by transfer of pT181 (4.4 kb) and pUSA300-HOUMR-like (27 kb) plasmids into the recipient restriction-deficient strain RN4220. Bacteriophages of serogroup B transduced the pT181 plasmid with a frequency of 1.0 × 10−8 to 1.1 × 10−7 and the pUSA300-HOUMR-like plasmid with a frequency of 4.1 × 10−9 to 4.2 × 10−8. With the phage φ81, no transductants were obtained with pT181 and the penicillinase plasmid was transferred at a very low frequency of 2.1 × 10−10. The ability of the serological group B bacteriophages to transduce plasmid DNA with markedly higher frequency than does bacteriophage φ81 of serogroup A confirms that the selected bacteriophages and the strain S. aureus COL comprise a suitable model for comparative analysis of packaged genes by the real-time quantitative PCR (qPCR) method.
Efficient packaging of bacterial genes by serogroup B bacteriophages demonstrated by qPCR
Quantification of bacterial DNA of the 12 monitored bacterial genes in 4 DNA samples isolated from phage particles by qPCR is summarized as log10 values for the number of gene copies per ng (GC ng−1) in Fig. 1, where the differences between the quantities of the monitored genes in different samples of phage particles are easily seen. The absolute values for the number of gene copies are available in Table S1. The marked difference in the ability to package bacterial DNA between serogroup B bacteriophages φ80, φ80α and φ11 and serogroup A bacteriophage φ81 is apparent here. In all phage DNA samples, the ratios between mean GC ng−1 of the monitored bacterial gene and the number of infectious particles derived from mean GC ng−1 of gene for putative tail fibre protein of the serological group B phages or based on a conserved phage gene for a structural tail protein in non-transducing phage φ81, described previously (Doškař et al., 2000), were determined to assess frequency of transducing particles in phage lysate (Table 1).
Figure 1. Quantification of bacterial DNA in samples isolated from transducing phage lysates propagated on donor strain S. aureus COL (mean values of three measurements transformed as log10 bacterial gene copies per 1 ng phage DNA).
Download figure to PowerPoint
Table 1. Frequencies of phage transducing particles carrying the targeted gene calculated as ratio of gene copy number per ng to the number of infectious particles.
|Plasmid borne genes||tetK||2.7 × 10−9||4.6 × 10−7||7.4 × 10−6||6.3 × 10−7|
|blaZ||7.6 × 10−8||1.3 × 10−5||1.2 × 10−4||2.8 × 10−5|
| || || || || || |
|Parts of SCCmec||mecA||1.8 × 10−8||1.6 × 10−4||7.0 × 10−4||1.7 × 10−4|
|ccrA1||1.9 × 10−8||1.6 × 10−4||6.0 × 10−4||1.9 × 10−4|
| || || || || || |
|Parts of SaPI1||Sa1int||1.5 × 10−7||2.3 × 10−2||3.2 × 10−3||8.0 × 10−3|
|seb||1.2 × 10−7||1.8 × 10−2||1.9 × 10−3||3.9 × 10−3|
| || || || || || |
|Genomic islands||set5||1.3 × 10−7||1.7 × 10−3||6.1 × 10−3||1.1 × 10−2|
|lukD-lukE||2.5 × 10−7||8.7 × 10−4||2.9 × 10−3||3.7 × 10−3|
| || || || || || |
|Bacterial genes on chromosome||Sa2052||5.8 × 10−8||7.1 × 10−4||2.8 × 10−3||4.9 × 10−3|
|clfB||4.8 × 10−7||1.7 × 10−4||7.5 × 10−4||2.1 × 10−4|
|sspA||1.7 × 10−7||8.6 × 10−3||2.7 × 10−4||2.5 × 10−4|
|nuc||3.6 × 10−7||5.4 × 10−3||8.8 × 10−4||4.2 × 10−3|
As plasmids occur in several copies per cell, in order to quantify the transducing particles carrying the plasmid genes blaZ and tetK in comparison to the standard consisting of a chromosomal DNA, it was necessary to determine plasmid copy number (PCN) in the donor strain. Plasmid pT181 was represented in 65 copies in the strain S. aureus COL (pT181) and in 52 copies in the donor strain S. aureus COL (pT181, pUSA300-HOUMR-like). The difference in the number of copies is probably due to the fact that the donor strain S. aureus COL (pT181, pUSA300-HOUMR-like) also contains four copies of the penicillinase plasmid in its genome. The determined numbers of plasmid copies correspond to the theoretical copy number of such plasmids in the S. aureus species (Novick, 1990). The gene tetK and blaZ were found in the largest GC ng−1 in the bacteriophage φ80α (Table S1). Surprisingly, presence of these resistance genes was detected also in phage particles φ81 (Fig. 1). Antimicrobial resistance genes have been also demonstrated in other studies describing HGT mediated by phages isolated from activated sludge, urban sewage, river water and animal faecal wastes (Parsley et al., 2010; Colomer-Lluch et al., 2011a,b).
The most important new finding was the evidence as to the presence of genes of the SCCmec mobile element and their quantification in phage particles. Even though SCCmec horizontal transfer is assumed, evidence for the transfer of the entire cassette or its parts and the exact mechanism has not been described heretofore. The genes mecA and ccrA1 were represented in phages of the serogroup B in significantly higher numbers of copies than in non-transducing phage φ81 (Fig. 1, Table S1). The frequencies (Table 1) of transducing particles harbouring parts of SCCmec indicate that one out of 104 in φ11, φ80 and φ80α or 108 in φ81 bacteriophage particles package both monitored SCCmec-borne genes. Moreover, the comparable quantity of the two monitored genes inside the phage particles suggests the possibility of encapsidation for larger segments of this MGE into one particle, which is also supported by the results from long PCR detecting an 8.7 kb region bordered by the genes mecA and ccrA1 (Fig. 2).
Figure 2. Analysis of presence of 8.7 kb SCCmec region bordered by the genes mecA and ccrA1 in phage particles φ80, φ80α, φ11 and φ81 by long PCR. S. aureus COL DNA was used as a positive control. 2-Log DNA Ladder (New England Biolabs) was used as a molecular weight marker.
Download figure to PowerPoint
Although the excision and integration mechanism of SCCmec into the bacterial chromosome has been described (Wang and Archer, 2010), the exact way for transfer of this element into bacterial cells is unknown. Transfer of the SCCmec element via bacteriophages has not yet been proven, it is hypothesized that in particular the elements of size smaller than the S. aureus siphophage genome (45 kb), such as the SCCmec type I or type IV elements, can be transferred via generalized transduction. Experimental evidence for phage-mediated transduction of methicillin resistance by bacteriophages φ80 and φ53 from different naturally occurring MRSA strains to strain 8325 variants with low frequencies, which depended on the selected transducing systems had been described by Cohen and Sweeney (1970). Later, it was shown that this transduction is efficient only in recipient strains with a penicillinase plasmid (Cohen and Sweeney, 1973; Stewart and Rosenblum, 1980). The possible horizontal transfer of parts of SCCmec element is suggested by the existence of a large number of SCCmec variants described in nucleotide sequence databases, including an SCCmec mosaic containing two recombinase loci and different mec complexes (Heusser et al., 2007) or non-mecA carrying cassettes (Katayama et al., 2003; Garcia-Alvarez et al., 2011). The findings of Hanssen and colleagues (2004) also indicate that the mecA gene may be transferred independently in the presence of the ccr gene complex. On the other hand, Chlebowicz and colleagues (2010) observed a conversion from methicillin-resistant to methicillin-susceptible strain in vivo by means of recombination between ccrC genes in a type V (5C2&5) SCCmec of S. aureus ST398.
In contrast to SCCmec, the mobilization of SaPIs by bacteriophages and the high transduction frequency are well documented (Ruzin et al., 2001). In the DNA samples of phages from the serogroup B, we quantified the genes Sa1int and seb carried by the mobile element SaPI1 with mutually similar quantity. The number of GC ng−1 localized at SaPI1 was the highest from the monitored genes. Ratios between the monitored bacterial genes and conserved phage gene indicate that one out of 102 to 103 particles of serogroup B bacteriophage packages SaPI-borne genes. The high probability of packaging of SaPI-borne genes corresponds with the earlier described specialized transduction of SaPIs (Ruzin et al., 2001; Chen and Novick, 2009). Surprisingly, the genes located at genomic islands νSaα and νSaβ were packaged with frequency comparable to that for genes at SaPI1 (Fig. 1, Table S1). The cause for high packaging frequency of genes localized at SaPIs and genomic islands can be a common evolutionary origin with the bacteriophages. Partial homology of SaPI to phage DNA identified in a conserved regulatory sequence, integrase genes and the non-coding regions suggests a mobile element of phage origin (Yarwood et al., 2002). Novick and colleagues (1986) had observed up to 105-fold elevation of transduction frequency for plasmids with cloned φ11 DNA fragments, which supports the hypothesis that the sequences of phage origin are packaged with higher frequency than are other parts of the genome.
Bacteriophages also package all those monitored bacterial genes that are localized at the core genome of S. aureus COL (Fig. 1). In the group of these genes, we observed large differences in representation among individual DNA samples from phage particles. The clfB gene was detected in a lower total amount in phages of serogroup B in comparison to sspA, nuc and/or Sa2052 locus described by Štěpán and colleagues (2001). Bacteriophage φ81, however, packaged clfB gene with relatively high frequency. Based on large differences in GC ng−1 in transducing phages, we assume that these chromosomal genes are packaged randomly and may be transferred by the mechanism of generalized transduction.
Statistical analysis of the number of bacterial gene copies in phage particles
We used geometric mean, median, maximum and minimum to describe the differences between the four bacteriophages φ80, φ80α, φ11 and φ81 in the total amounts of encapsidated bacterial DNA inside phage particles (Fig. 3A) and to describe amount of DNA of the individual monitored genes represented in phage particles of transducing bacteriophages (Fig. 3B). The comparison as to the amount of bacterial DNA from 12 bacterial genes of S. aureus COL between bacteriophage strains was tested using the non-parametric Mann–Whitney U-test. A significant difference was proven in the amount of packaged bacterial DNA between phage φ81 and all bacteriophages of serogroup B: φ80 (P = 0.001), φ80α (P < 0.001) and φ11 (P < 0.001). No significant difference in the amount of bacterial DNA was proven between the bacteriophages of serogroup B. The results indicate that bacteriophage φ81 packages a significantly lesser amount of bacterial DNA than do bacteriophages of serological group B. Exceptions are the numbers of GC ng−1 for the plasmid-borne genes tetK and blaZ, which did not differ significantly between the monitored phage samples. The mutual comparison as to the number of copies of phage conserved genes between bacteriophages φ80α, φ11 and φ80 (from 3.2 × 107 GC ng−1 to 4.1 × 107 GC ng−1) and bacteriophage φ81 (3.7 × 109 GC ng−1) showed that the quantity of infectious particles in the sample of phage φ81 is by two orders of magnitude higher than in the DNA samples from phages with transducing ability. The bacteriophage φ81 does, therefore, package bacterial genes into its capsids, but that transduction is less probable than in bacteriophages of serogroup B due to the low proportion of transducing virions in the phage lysate.
Figure 3. A. Total amount of bacterial DNA in bacteriophage strains φ80, φ80α, φ11 and φ81 described by log10 of mean gene copies per 1 ng phage DNA values from 12 bacterial genes tetK, blaZ, mecA, ccrA1, Sa1int, seb, set5, lukE-lukD, Sa2052 locus, clfB, sspA and nuc.
B. Amounts of different types of bacterial DNA in the three comparable transducing bacteriophage strains φ80, φ80α and φ11 described by values for amount of specific genes transformed as log10 bacterial gene copies per 1 ng phage DNA.
Download figure to PowerPoint
Non-parametric comparison as to the quantities of the different groups of bacterial genes in phage particles φ80, φ80α and φ11 was performed (Table 2). Statistical comparison showed no significant difference between the plasmid resistance genes tetK and blaZ (P = 0.100); no significant differences were determined in the copy number of genes on SaPI1 and the genomic islands genes (P = 0.240) or between the genomic islands genes and a group of four genes on the bacterial chromosome.
Table 2. Non-parametric comparison of differences between amounts of various types of bacterial genes in three bacteriophage strains (φ80, φ80α and φ11; values in log10 GC ng−1 phage DNA) using Mann–Whitney U-test.
|P-value||blaZ||Parts of SCCmec|| ||Genomic islands genes||Other chromosomal genes|
|Parts of SCCmec||–||–||0.002||0.002||0.018|
|Parts of SaPIs||–||–||–||0.240||0.032|
|Genomic islands genes||–||–||–||–||0.180|
Transduction efficiency of pUSA300-HOUMR-like penicillinase plasmid
The obtained absolute values for the number of GC ng−1 of the monitored plasmid gene blaZ, the numbers of transductants selected on cadmium plates from 1 ml of phage lysate and the total concentration of phage DNA enabled determining the number of virions containing the penicillinase plasmid and calculating the efficiency of transduction, E, for the individual bacteriophages according to the relationship:
where T is the number of transductants from 1 ml of phage lysate, c is the concentration of phage DNA from phage lysate, and V is the volume of phage lysate. The frequencies of transductions (transductants/PFU) of the penicillinase plasmid obtained with φ80, φ80α and φ11 were 4.20 × 10−8, 8.40 × 10−9 and 4.14 × 10−9 respectively. The transduction efficiencies calculated from the absolute numbers of virions containing the blaZ gene in 1 ml of lysate and the numbers of acquired transductants for the aforementioned phages were 0.33%, 0.0072% and 0.015% respectively. The bacteriophage φ81 transduced the penicillinase plasmid with lower frequency of 2.09 × 10−10. This confirms the ability of φ81 to package and transfer the plasmid genes with transduction efficiency of just 0.0025%. Comparing the values for transduction efficiency and calculated frequency of transducing particles in samples (Table 1) indicates that the high quantity of GC ng−1 phage DNA is not sufficient condition ensuring high transduction efficiency.
In pT181, it was impossible to determine the transduction efficiency, as it is presumed to be packaged in the form of concatemers (Novick et al., 1986). This does not allow for determining the precise number of transducing virions.