**3. Discussion**

This work explores the permissiveness of three amoeba strains regarding the intracellular multiplication of three pathogenic *L. pneumophila* strains under two temperature conditions (22 ◦C and

37 ◦C) that correspond to temperatures found in cooling towers in which *L. pneumophila* are known to replicate within certain strains of amoebae [10,25]. It is important to demonstrate that *W. magna* C2c Maky does not multiply *L. pneumophila* as we aim to propose it as a natural biocide to treat cooling towers.

The three *L. pneumophila* strains are a representative set of *L. pneumophila* serogroup 1 that is responsible for 95% of the legionellosis disease world-wide [5]. Strain Philadelphia is a clinical isolate that is historically responsible for the very first outbreak. It possesses gene traits that allow for multiplication in a number of hosts such as peripheral blood mononuclear cells, peritoneal macrophages, and *A. castellanii*, *A. polyphaga*, or *A. lenticulate* [26–29]. The Philadelphia strain is, according to the EN 13623 European standard, the only strain for which testing is required to validate a disinfectant against *Legionella* in Europe. *L. pneumophila* Lens was chosen because it was responsible for an outbreak in the north of France between November 2003 and January 2004 where 86 confirmed cases resulted in 17 deaths [30]. *L. pneumophila* Paris was chosen because, among the endemic strains of *L. pneumophila* serogroup 1, sequence type 1 (ST1) strains are among the most prevalent, particularly the ST1/Paris pulsotype. This endemic type was responsible for 8.2% of French culture-proven cases of Legionnaire's disease from 1995 through 2006. ST1/Paris pulsotype isolates also have been detected in clinical and environmental samples taken from several other countries around the world, including Switzerland, Italy, Spain, Sweden, the United States, Japan, Senegal, and Canada [21,30].

Our experiments demonstrate differential behaviours among amoeba species infected by the pathogenic bacteria. Compared to *A. castellanii* and *W. magna* Z503, the intracellular *L. pneumophila* are efficiently eliminated by *W. magna* C2c Maky at 22 ◦C and 37 ◦C. Indeed, the experiments report not only a non-replication, but also an elimination of the intracellular strains Lens, Paris and Philadelphia within *W. magna* C2c Maky. Furthermore, the coculture medium used in the survey is not adapted to the survival of the legionella bacteria, and they, therefore, must parasitize the amoebae to facilitate their own growth. Indeed, the experiments demonstrate that the three legionella strains were unable to remain at the inoculation level and began to die after 24 h (Figure 1). Although the medium is not adapted to *L. pneumophila* strains, it was chosen for the co-culture study because an increase of the bacterial number during the co-culture experiment necessarily indicates that the multiplication occurred within amoeba. The bacterial multiplication is observed both in *A. castellanii* and *W. magna* Z503, and it is not observed in *W. magna* C2c Maky. The assays reveal a multiplication of all legionella strains within *A. castellanii* at 37 ◦C and the intracellular multiplication of strain Lens and Paris at 22 ◦C. Indeed, the strain Philadelphia grows at 37 ◦C (Figure 3c) and does not multiply at 22 ◦C (Figure 3a) within *A. castellanii*. Based on this, these results sugges<sup>t</sup> a behaviour that is influenced by the temperature conditions. Several previous studies revealed the effect of temperature on the relationship between *L. pneumophila* and free-living amoeba (FLA) [9,31,32]. *L. pneumophila* serogroup 1, for example, replicated in *A. castellanii* at 25 ◦C but were digested at temperatures below 20 ◦C [25]. Dupuy et al. assessed the ability of 12 amoeba strains of *Naegleria* sp., *Acanthamoeba* sp., and *Vermamoeba* sp. to support the multiplication of *L. pneumophila* Lens at various temperatures (25 ◦C, 30 ◦C and 40 ◦C), and they revealed a more efficient intracellular proliferation with increasing temperatures [33]. Additionally, we did not observe the same behaviour according to the different bacteria and amoeba strains used during our experiments. Indeed, the strain Lens replicates at 37 ◦C within *W. magna* strain Z503, but not in *W. magna* C2c Maky (Figure 3d). The co-culture at 22 ◦C of *W. magna* Z503 with *L. pneumophila* strain Paris and strain Lens reveals a multiplication of the bacteria; however, no replication is observed during co-culture with strain Philadelphia (Figure 3b). The difference in amoeba permissiveness has been highlighted previously, especially in regard to *Naegleria*, *Acanthamoeba*, *Vermamoeba* and *Micriamoeba tesseris* [9,34]. The non-replication of legionella within *W. magna* C2c Maky was previously observed with strain Paris [20]. Our study confirms this result, as the resistance of *W. magna* C2c Maky towards *L. pneumophila* Paris is illustrated by the observed significant decrease in the bacterial concentration after 4 days of co-culture at 22 ◦C and 37 ◦C (Figure 4c,d). Dey et al. [20], however, reported a moderate increase in strains Philadelphia and Lens within *W. magna* C2c at 37 ◦C while in our study the intracellular bacterial concentration significantly decreased in culture with *W. magna* C2c Maky at 22 ◦C and 37 ◦C. These di fferences can be explained by the protocol parameters used in the former study, particularly regarding the culture medium and elimination of extracellular bacteria. The authors used serum casein glucose yeas<sup>t</sup> extract medium (SCGYEM) that was favourable to *L. pneumophila* survival, so bacteria were not forced to multiply into amoeba to survive. Additionally, Dey and co-workers did not eliminate extracellular bacteria by centrifugation, and the observed increase could be due to extracellular bacterial replication, such as that resulting from necrotrophic growth as previously demonstrated [35].

*W. magna C2c* Maky is demonstrated to possess a high e fficiency for digesting the intracellular *L. pneumophila* cells in all strains used in this survey. The growth of *L. pneumophila* within amoebas is known to enhance the pathogenicity and invasion of *L. pneumophila* [15,36]; however, no intracellular bacterial replication is observed when we infect *W. magna* C2c Maky with *L. pneumophila* strains derived from a first co-culture that was thought to be more virulent (unpublished data).

The action on di fferent *L. pneumophila* strains and the absence of internal proliferation support the fact that *W. magna* C2c Maky could be used as a biocide to combat *L. pneumophila* proliferation in cooling tower water. This observation is consistent with the control of legionella by *W. magna* C2c Maky observed in real conditions during field trials in functioning cooling towers (http://www.amoeba-biocide.com/ sites/default/files/180711\_cp\_amoeba\_us\_positive\_e fficacy\_field\_test\_en\_vedf\_0.pdf). The traditional method to control bacterial growth in cooling tower water is primarily based on the use of chemical biocides [37,38]. Indeed, the oxidizing agen<sup>t</sup> chlorine is the most used product for cooling tower treatment [39]. The chemical biocide is e fficient to prevent *L. pneumophila* proliferation, although some previous studies reported incomplete eradication of legionella from installations and progressive re-colonization within these systems within weeks or months [40,41]. Moreover, these chemical biocides are dangerous to the environment, they degrade the installation systems, and they require the application of other products such as anti-corrosive agents [42,43]. Described by Iervolino, treatment with another oxidizing agen<sup>t</sup> (H2O2/Ag) was inadequate for legionella control, and, instead, it caused a rapid increase of one logarithmic unit [44]. Chemical biocide action also is not completely e fficient against biofilms and amoeba cysts that can provide protection against disinfection treatment [16,17,45]. Finally, chemical biocides used in cooling towers can select *L. pneumophila* populations, and chemical biocides can promote resistance to biocides and to human health antibiotics [46,47].

To conclude, *W. magna* C2c Maky is not associated with any human or animal infection, and this is in agreemen<sup>t</sup> with the lack of pathogenicity demonstrated in vivo and suggested by genomic analysis [24,48]. This organism is likely a safe and e fficient candidate for legionella control in cooling towers and could provide an alternative solution to chemical biocides.

#### **4. Materials and Methods**

#### *4.1. Free-Living Amoebae Culture*

*Willaertia magna* C2c Maky (ATCC ® PTA-7824), *Willaertia magna* Z503 (ATCC ® 50035), and *Acanthamoeba castellanii* (ATCC ® 30010) were purchased from ATCC and cultivated according to their recommendation into 10 mL of modified PYNFH medium (ATCC medium 1034) in a T-25 tissue culture flask. Amoebae were then grown in cell factories in serum casein yeas<sup>t</sup> extract medium (SCYEM) at 30 ◦C. SCYEM medium is derived from serum casein glucose yeas<sup>t</sup> extract medium (SCGYEM) medium [49] and contained 10 <sup>g</sup>·L−<sup>1</sup> casein, 5 <sup>g</sup>·L−<sup>1</sup> yeas<sup>t</sup> extract, 10% foetal calf Serum, 1.325 <sup>g</sup>·L−<sup>1</sup> Na2HPO4, and 0.8 <sup>g</sup>·L−<sup>1</sup> KH2PO4. After 72 h (during exponential phase), the cell factories were gently shaken, and the amoeba suspensions were transferred to 50 mL Falcon ® tubes. Amoeba populations were then quantified using a Malassez haemocytometer cell counting chamber method (Thermo Fisher Scientific, France) with Trypan blue by mixing 100 μL of Trypan blue with 100 μL of amoeba sample. According to the results, the amoebae concentration in Falcon ® tubes was then adjusted to 3 × 10<sup>5</sup> cells/mL by the addition of SCYEM. The amoebas were then washed twice in SCYEM using centrifugation at 3000× *g* for 10 min, and the supernatants were then discarded. Amoeba populations were then re-quantified, and the amoeba suspensions were finally adjusted to 3 × 10<sup>5</sup> cells/mL in 100 mL of SCYEM. A final quantification was performed to verify the concentration.

Each final solution of *W. magna* C2c Maky, *W. magna* Z503, and *A. castellanii* corresponded to working suspensions that were named AWSC2C, AWSZ503, and AWSAC, respectively (Table 2).


**Table 2.** Preparation of the co-cultures.

1 AWS: Amoeba Working Solution at 3 × 10<sup>5</sup> cells / mL; 2 BWS: Bacteria Working Solution at 3 × 10<sup>7</sup> CFU / mL.

#### *4.2. Legionella Pneumophila Cultures*

*L. pneumophila* strain Philadelphia (ATCC 33152), *L. pneumophila* strain Lens (CIP 108280), and *L. pneumophila* strain Paris (CIP 107629) were grown on buffered charcoal yeas<sup>t</sup> extract (BCYE) agar plates (Thermo Fisher Scientific, Dardilly, France) at 36 ◦C for 72 hours and then harvested by scraping, suspended in phosphate-buffered saline (PBS), centrifuged at 9500 xg for 10 min, and washed once in PBS. The supernatants were then discarded. The *L. pneumophila* suspensions were then diluted in PBS to obtain 3 × 10<sup>7</sup> bacteria/mL.

The legionella final suspensions represented the bacterial stock working suspensions, and they were identified as BWSPhila, BWSParis, and BWSLens (Table 2).

#### *4.3. Bacterial Survival in the Coculture Medium (Control)*

The three control bacterial conditions were prepared as described in Table 2 by adding 10 mL of SCYEM to the 0.1 mL bacteria working solutions (BWSPhila, BWSParis, or BWSLens) in 25 cm<sup>3</sup> flasks (Dutscher, Brumath, France) and incubated at 22 ◦C or 37 ◦C. This operation corresponded to the T0 time point of the bacterial controls. Occurring at T0, T0 + 24 h, T0 + 48 h, T0 + 72 h, and T0 + 96 h, 1 mL was sampled in each flask and then serially 10-fold diluted in SCYEM and plated on buffered charcoal yeas<sup>t</sup> extract plates (BCYE) in triplicate. BCYE plates were incubated at 36 ◦C, and colony forming units (CFU) were counted after 5 days. Each condition was performed for three independent replicates and repeated three times (n = 9).

#### *4.4. Amoeba Survival in the coculture Medium (Control)*

The three amoeba working solutions (AWSC2C, AWSZ503, or AWSAC) were prepared as described in Table 2 (10 mL of working solutions) and incubated at 22 ◦C or 37 ◦C in 25 cm<sup>3</sup> flasks. Occurring at T0, T0 + 24 h, T0 + 48 h, T0 + 72 h, and T0 + 96 h, the flasks were gently shaken, and the numbers of amoeba cells were quantified using a haemocytometer cell counting chamber method with Trypan blue. Each condition was performed for three independent replicates and repeated three times (n = 9).
