Clonal population structure of Legionella pneumophila inferred from allelic profiling

The population structure of Legionella pneumophila was investigated by analysing nucleotide sequences from six loci (flaA, pilE, asd, mip, mompS and proA) of 335 globally distributed isolates from clinical and environmental sources over a 29-year period (1977-2006). Data were obtained from unrelated isolates from Europe (n=270), Japan (n=31), Canada (n=7), the USA (n=24) and Australia (n=1). The country of origin of two strains was unknown. Analysis of these isolates indicated significant linkage disequilibrium between the six loci. Application of six sequence-based recombination detection tests did not reveal evidence of recombination, but estimates of rates of recombination and mutation made by a seventh test suggested that recombination could have occurred at a rate similar to, but probably lower than, that of mutation. Genealogies inferred under models with and without recombination were congruent with each other, providing no definitive evidence regarding recombination, and were in agreement with sequence clusters identified by graph methods. Further evidence supporting the distinct nature of two of the three subspecies of L. pneumophila, subsp. fraseri and subsp. pascullei, was also found. The ratios of non-synonymous to synonymous nucleotide polymorphisms for each of the allele sets were examined and revealed that the putative virulence loci mompS and pilE are under diversifying pressure, while the allelic regions of three other loci linked to virulence (flaA, proA and mip) do not appear to be.