Chlorine and thermal treatments are the most commonly used procedures to control and prevent Legionella proliferation in drinking water systems of large buildings. However, cases of legionellosis still occur in facilities with treated water. The purpose of this work was to model the effect of temperature and free chlorine applied in similar exposure conditions as in drinking water systems on five Legionella spp. strains and two amoebal strains of the genera Acanthamoeba. Inactivation models obtained were used to determine the effectiveness of the treatments applied which resulted more effective against Legionella than Acanthamoeba, especially those in cystic stages. Furthermore, to determine the influence of the relationship between L. pneumophila and Acanthamoeba spp. on the treatment effectiveness, inactivation models of the bacteria-associated amoeba were also constructed and compared to the models obtained for the free living bacteria state. The Legionella-amoeba association did not change the inactivation models, but it reduced the effectiveness of the treatments applied. Remarkably, at the lowest free chlorine concentration, 0.5 mg L^-1, as well as at the lowest temperatures, 50°C and 55°C, the influence of the Legionella-amoeba associate state was the strongest in reducing the effectiveness of the treatments compared to the free Legionella state. Therefore, the association established between L. pneumophila and amoebae in the water systems indicate an increased health risk in proximal areas of the system (close to the tap) where lower free chlorine concentrations and lower temperatures are commonly observed.

Legionella infections are among the most important waterborne infections with constantly increasing numbers of cases in industrialized countries, as a result of aging populations, rising numbers of immunocompromised individuals and increased need for conditioned water due to climate change. Surveillance of water systems is based on microbiological culture-based techniques; however, it has been shown that high percentages of the Legionella populations in water systems are not culturable. In the past two decades, the relevance of such viable but non-culturable (VBNC) legionellae has been controversially discussed, and whether VBNC legionellae can directly infect human macrophages, the primary targets of Legionella infections, remains unclear. In this study, it was demonstrated for the first time that several starved VBNC Legionella strains (four L. pneumophila serogroup 1 strains, a serogroup 6 strain and a L. micdadei strain) can directly infect different types of human macrophages and amoebae even after one year of starvation in ultrapure water. However, under these conditions, the strains caused infection with reduced efficacy, as represented by the lower percentages of infected cells, prolonged time in co-culture and higher multiplicities of infection required. Interestingly, the VBNC cells remained mostly non-culturable even after multiplication within the host cells. Amoebal infection by starved VBNC Legionella, which likely occurs in oligotrophic biofilms, would result in an increase in the bacterial concentration in drinking-water systems. If cells remain in the VBNC state, the real number of active legionellae will be underestimated by the use of culture-based standard techniques. Thus, further quantitative research is needed in order to determine, whether and how many starved VBNC Legionella cells are able to cause disease in humans.

Legionellae are among the most important waterborne pathogens in industrialized countries. Monitoring and surveillance of Legionella in engineered water systems is usually performed with culture-based methods. Since the advent of culture-independent techniques, it has become clear that Legionella concentrations are often several orders of magnitude higher than those measured by culture-based techniques and that a variable proportion of these non-culturable cells are viable. In engineered water systems, the formation of these viable but non-culturable (VBNC) cells can be caused by different kinds of stress, such as, and most importantly, nutrient starvation, oxidative stress and heat. In this study, the formation of VBNC cells of six Legionella strains under conditions of starvation was monitored in mono-species microcosms for up to one year using a combination of different viability indicators. Depending on the strain, complete loss of culturability was observed from 11 days to 8 weeks. During the starvation process, three distinct phases and different sub-populations of VBNC cells were identified. Until complete loss of culturability, the number of membrane-intact cells decreased rapidly to 5.5-69% of the initial cell concentration. The concentration of the sub-population with low esterase activity dropped to 0.03-55%, and the concentration of the highly esterase-active sub-population dropped to 0.01-1.2% of the initial concentration; these sub-populations remained stable for several weeks to months. Only after approximately 200 days of starvation, the number of VBNC cells started to decrease below detection limits. The most abundant VBNC sub-populations were characterized by partially damaged membranes and low esterase-activity. With this study, we showed that upon starvation, a stable VBNC Legionella community may be present over several months in a strain-dependent manner even under harsh conditions. Even after one year of starvation, a small proportion of L. pneumophila cells with high esterase-activity was detected. We speculate that this highly active VBNC subpopulation is able to infect amoebae and human macrophages.