4. Discussion

The results obtained allow us to conclude that lysogenization by the phage φ24B of diverse E. coli strains producing O antigens was not due to an unusual ability of this virus to penetrate the O antigen shield, but was mediated by spontaneous formation of bacterial rough mutants or of mutants with significantly compromised O antigen biosynthesis. It is not clear why the lysogenization was not effective for some strains. The activity of antiviral systems, such as restriction-modification, avoiding the lysogenization at stages after the viral DNA penetration into the cell46, cannot be excluded. Also, the effect may be due to point mutations present in BamA protein or lower frequency of rough mutants in particular strains.
In the conditions of our experiment, the high concentration of bacteriophage used allowed almost all the cells potentially susceptible to the phage to be infected. However, in vivo the populations ofE. coli are very unlikely to face such a massive viral attack. The fraction of rough mutants in natural habitats is hard to estimate, but we can speculate that it should be lower than in in vitroconditions because such mutants have compromised protection not only from the phage attack but also from immune system agents such as serum bactericidal activity47-49 and from other environmental factors32 and therefore should be counter-selected. Moreover, if stx-phage lysogens were formed by infection of such rough mutants, their expected fitness and/or virulence would be significantly lower than that of the parental strains. These strains, noteworthily, were highly sensitive to SBA of the horse serum to which the parental O-antigen producing strains were completely resistant. Therefore the lysogens for the φ24B phage are expected to have reduced virulence. Thus, the factor of non-specific protection of the bacterial cells by the O antigen should not be neglected during the evaluation of the potential significance of stx-converting phage transmission in nature (as it currently is neglected in many studies25, 26).
We also should note that the lysogenization by φ24B:cat appears to be a simple and efficient procedure for selection for mutants with compromised or completely abolished O antigen synthesis. This procedure may be particularly valuable for the researchers working with field isolates of E. coli for which the genomic sequences are not yet available and/or in which other rapid techniques such as recombination with PCR fragments for genes knockout50 are frequently less effective than in laboratory E. coli.