The human pathogen Legionella pneumophila causes a severe pneumonia so-called legionellosis or Legionnaires’ disease (LD); this Gram negative bacterium was identified after the 1976 Philadelphia outbreak during the American Legion convention where 29 people succumbed . Further outbreaks were associated with aerosol-producing devices like showers, cooling towers, whirlpools and fountains, but Rowbotham was the first to show a link between Legionella ecology and LD [2, 3]. Actually, L. pneumophila is ubiquitous in aquatic environment and is able to multiply intracellularly in fresh water protozoa.
L. pneumophila displays 15 serogroups but the majority of human cases are due to the serogroup1 (Lp1) (84% worldwide, 95% in Europe) [4, 5]. Lp1 is frequently found in the environment and accounts for 28% of environmental isolates in France. Other Legionella species, as L. anisa, L. dumoffii and L. feeleii that frequently colonize the water distribution systems, are rarely involved in human disease . These data suggest that the high frequence of LD involving Lp1 is not due to its predominance in the environment but rather linked to a higher virulence than other species or serogroups of Legionella. The only exception is Legionella longbeachae accounting for 30% of human cases in Australia and New-Zealand, and even 50% of cases in South Australia . In contrast to L. pneumophila, L. longbeachae is found predominantly in potting soil and transmitted by inhalation of dust of contaminated soils.
A lot of attention has been paid to the identification of Lp1virulence factors. It is now recognized that the co-evolution between eukaryotic hosts and L. pneumophila had led to the selection of a set of virulence factors which allow this bacterium to exploit host cellular processes; among these factors, eukaryotic-like proteins, encoded by genes identified on the basis of genome sequence analysis, are involved in different steps of the Legionella intracellular cycle [5, 7–10]. Recently, comparison of Legionella genome sequences has shown that some genes encoding the lipopolysaccharide biosynthesis were specific of Lp1 and constitute specific markers for the molecular typing .
We focused our attention on the identification and virulence capacities of different serogroups of L. pneumophila strains present in the French thermal spa where five cases of legionellosis were diagnosed in 1986, following by two cases in 1994 and 1997 [12, 13]. In order to determine the source of infection, water samples had been collected throughout the water distribution system as well as the three natural springs (S, sulphur; A, alum and P, cold) and two bore holes feeding the system. Eighty one L. pneumophila strains belonging to five serogroups (27 Lp1, 1Lp2, 62 Lp3, 3 Lp6 and 9 Lp13) had been identified from water samples collected over a two-year period (1997–1998); thus this water system appeared mainly contaminated by Lp1 and Lp3, also present in two natural spring (S and A). Nevertheless, comparative analysis of genomic DNA, by PFGE (“Pulse Field Gel Electrophoresis”), of both clinical Lp1 isolated from patients and environmental Lp1 isolates did not allow identifying the source of infection.
In this study, our goal was to identify legionellae directly virulent towards protozoa and as a consequence with the ability to survive in a specific environment, like the spring S characterized by a temperature of 37°C and a high level of sulphates and thiosulphates as the calcium and sodium salts . Thus, we isolated legionellae from natural biofilms developed on glass slides immersed in this contaminated spring. After typing by different approaches, the DNA genome diversity of these environmental Lp strains was analyzed, and their virulence and cytotoxicity towards the amoeba Acanthamoeba castellanii were compared to those of well-known French clinical isolates (Lp1 strains Lens, Paris and Lorraine).