Decreasing Pasteurization Treatment Efficiency against Amoeba-Grown Legionella pneumophila—Recognized Public Health Risk Factor

Legionellae are gram-negative bacteria most commonly found in freshwater ecosystems and purpose-built water systems. In humans, the bacterium causes Legionnaires’ disease (LD) or a Pontiac fever. In this study, the different waters (drinking water, pool water, cooling towers) in which Legionella pneumophila has been isolated were studied to assess the possible risk of bacterial spreading in the population. The influence of physical and chemical parameters, and interactions with Acanthamoeba castellanii on L. pneumophila, were analyzed by Heterotrophic Plate Count, the Colony-forming units (CFU) methods, transmission electron microscopy (TEM), and Sequence-Based Typing (SBT) analysis. During the study period (2013–2019), a total of 1932 water samples were analyzed, with the average annual rate of Legionella-positive water samples of 8.9%, showing an increasing trend. The largest proportion of Legionella-positive samples was found in cooling towers and rehabilitation centers (33.9% and 33.3%, respectively). Among the isolates, L. pneumophila SGs 2–14 was the most commonly identified strain (76%). The survival of Legionella was enhanced in the samples with higher pH values, while higher electrical conductivity, nitrate, and free residual chlorine concentration significantly reduced the survival of Legionella. Our results show that growth in amoeba does not affect the allelic profile, phenotype, and morphology of the bacterium but environmental L. pneumophila becomes more resistant to pasteurization treatment.


Introduction
Legionellae are gram-negative bacteria commonly found in freshwater ecosystems such as rivers, lakes, ponds, and streams [1,2]. The entry of bacteria into purpose-built water systems and drinking-water distribution systems make a serious health concern. After the growth and replication of bacteria in drinking-water distribution systems, people can County, Croatia. A second aim of the study was to identify SBT types and phenotypes of environmental isolates and investigate whether a change would occur in the allelic profile, phenotype, or resistance of bacteria after cultivation in amoeba. The influence of physical-chemical parameters and interactions of L. pneumophila with amoebae were analyzed to identify possible factors that can contribute to reduction in human infection.

Samples and Amoeba Culture
A total of 1932 water samples were taken from swimming pools (public and spa), drinking-water distribution systems (hotels, apartments, camping sites, rehabilitation centers, ferries), and cooling towers in Primorje-Gorski Kotar County. Public water supply in Primorje-Gorski Kotar County is based on groundwater and surface water resources. Groundwater was used in cooling towers and swimming pools are filled with water from the groundwater system. All genotyped Legionellae came from groundwater or surface water samples (see paragraph 3.4. for details). The samples were collected as pre-flush samples, without flushing and without flaming the outlet (in order to define the bacterial colonization at a specific outlet point). The samples were placed in 1-litre sterile bottles, with 10% sodium thiosulphate (1 mL/L) in order to neutralize the chlorine or other halogen-based biocides. The samples were protected and transported to the laboratory in appropriate cooled containers, at a temperature of 2-8 • C, and tested for the presence of Legionella pneumophila within 24 h. After detection in environmental water samples, Legionella pneumophila was grown on Buffered Charcoal Yeast Extract agar (BCYE). Legionella pneumophila AA100 was used as a positive control.
Acanthamoeba castellanii was obtained from the American Type Culture Collection, 30234, and cultivated in peptone-yeast extract-glucose medium (PYG = ATCC medium 712) at 25 • C.

Detection and Quantification of Legionella spp. in Water Samples
The isolation of Legionellae and the estimation of their number in different types of water samples was carried out using the culture method, according to ISO 11731:2017 [25]. Over the period from 2013 to 2016, L. pneumophila was analyzed in agreement with ISO 11731:1998, which is the previous version of the current standard, albeit, the procedure remained unmodified. Filtration was used to concentrate Legionella in water samples. One liter of water was filtered through a polycarbonate membrane filter (0.2 µm pore-size, diameter 47 mm, Pall Corporation, Port Washington, NY, USA). Then, the membrane filter was transferred to a sterile container filled with 10 mL of distilled water. In order to enhance elution, the membrane was fragmented into smaller pieces using sterile scissors. It was placed in an ultrasonic water bath at 35 Hz/2 min (to detach the bacteria from the surface of the membrane filter).
Aliquots of 0.1 mL of the water samples subjected to heat treatment were spread onto culture-selective media, namely, Glycine Vancomycin Polymyxin B Cycloheximide (GVPC) agar (OXOID, UK). Inoculated plates were inverted and incubated in humidified air with 2.5% CO 2 at 36 ± 1 • C for 7-10 days. The plates were inspected every 2-3 days. Presumptive colonies were taken and sub-cultured on plates of the following media: (1) BCYE, and (2) Buffered Charcoal Yeast Extract without L-cysteine (BCYE-Cys) (OXOID, UK), or other suitable culture media (for example, sheep blood agar-sheep BA). The media were incubated at 36 ± 1 • C for more than 2 days. Growth of morphologically characteristic bacterial colonies on the BCYE medium, and no growth on the BCYE-Cys, were presumed to be Legionella. This commercial test enables identification of different serogroups of L. pneumophila, serogroup 1, serogroups 2-14, as well as seven additional Legionella species considered important for human infection. The limit of detection of the Legionella quantification method was 100 CFU/L.  [26], which is routinely used in water microbiology with Yeast Extract Agar (YEA, Biolife, Italy), incubated at 36 • C ± 2 • C for 44 ± 4 h and at 22 • C ± 2 • C for 68 ± 4 h, and (2) by means of sheep Blood Agar-BA (Liofilchem, Italy), incubated at 36 • C ± 2 • C during 44 ± 4 h and 96 ± 4 h.
In both cases, 1 mL (or appropriate dilution) of a water sample was taken and spread using the poured plate method. The results were calculated as the number of colonies in mL (CFU/mL). Separate counts were made for two culture medium types, each of them under two different incubation conditions.

Physical-Chemical Parameters
Water temperature was measured using a calibrated alcohol-filled thermometer, temperature range between 1 and 90 • C, 0.1 • C/graduation, according to the Standard Method 2550 B [27]. The concentration of residual free chlorine was determined according to ISO 7393-2:2018 [28], using a portable Colorimeter™ II (Hach, CO, USA), with a quantification limit of 0.02 mg Cl 2 /L.
The S47 instrument (SevenExcellence, Mettler Toledo, Germany) was used to record electrical conductivity, with a quantification limit of 9 µS/cm, and pH value of the water according to ISO 7888:1985 [29] and ISO 10523:2008 [30], respectively.

Infection of Amoeba with L. pneumophila
The number of A. castellanii cells (10 6 cells per mL) was determined using a Neubauer chamber [31]. The amoeba cells were transferred in six-well plates and incubated overnight at 27 • C in 3 mL of media. On the following day, the amoeba cells were infected [32] with L. pneumophila strains isolated from various water samples and control bacteria, L. pneumophila AA100, at a multiplicity of infection (MOI) of 100, centrifuged at 240× g for 3 min at room temperature to synchronize the infection, and then incubated at 27 • C for 6 h. After incubation, the amoeba cells were washed three times with ATCC glucose-free media to remove extracellular bacteria. This was considered as the initial incubation time (time zero). At certain points in time, after infection (0, 2, 4, 6, 8, and 10 h), the amoeba cells were treated with 0.1% Triton X-100 (Sigma, St. Louis, MO, USA [33]) for 10 min to lyse the amoeba cells. The number of bacteria cultured in amoeba was determined by the CFU method, by adding serially diluted suspensions of Legionella on BCYE agar plates, and incubation for 72 h at 37 • C. The bacteria cultured in amoeba for 4 days were used in pasteurization trials.

Preparation of Bacterial Suspensions
For the preparation of bacterial suspensions, six representative L. pneumophila samples were selected (autocamp (cold water), autocamp (warm water), hotel, apartment, ferry, and cooling tower). L. pneumophila AA100 was used as a control. To prepare the agar-grown Legionella samples, L. pneumophila were grown on BCYE agar plates at 35 ± 2 • C for 72 h. For the preparation of the A. castellani-grown Legionella samples, L. pneumophila isolates were incubated in amoeba for 96 h at 27 • C, as described above, and treated with 0.1% Triton X-100 for 10 min to lyse the amoeba cells. Bacterial suspensions from amoeba-grown samples were prepared immediately after being isolated from the amoeba. The agar and amoeba-grown bacterial suspensions were prepared in sterile water, and then measured by spectrophotometry to yield a standard suspension containing approximately 10 9 Legionellae per mL.

Pasteurization Treatments
The agar-grown or A. castellanii-grown L. pneumophila were used in a pasteurization treatment. A 5 µL measure of standard suspension was added to tap water samples (5 mL) to yield inoculated water samples containing 10 6 CFU of L. pneumophila per mL for use in pasteurization trials. Water samples were pasteurized at the laboratory as follows. Aliquots (1 mL) of inoculated tap water samples containing 10 6 CFU/mL of L. pneumophila were heated in thermoblock operating at different temperatures and time frames (55 • C/10 min, 53 • C/10 min, 50 • C/10 min, 50 • C/30 min). After the appropriate holding time, the tubes were placed on ice for rapid cooling. The number of bacteria after the pasteurization treatment was determined by the CFU method.

Transmission Electron Microscopy (TEM)
TEM analysis was performed to assess the morphological and structural changes in bacteria after the pasteurization treatment or growth in amoeba. The samples were prepared for TEM on carbon-coated Cooper grids (SPI Supplies, USA) using 2% phosphotungstic acid for staining. The L. pneumophila suspension (10 µL) was applied to the grid for 2 min and then removed from the grid using filter paper. Subsequently, 10 µL of acid was added to the grid for 1 min, and then drained using filter paper. The grids were air-dried before use, and 10 fields of each grid were randomly observed under a TEM device (Zeiss 902A).

Sequence-Based Typing (SBT) Analysis
A Wizard Genomic DNA Purification Kit (Promega, USA) was used for L. pneumophila DNA extraction, according to the instructions provided by the manufacturer. The EWGLI standard protocol [34,35] was used for amplification of the seven gene loci (flaA, mompS, proA, asd, mip, pilE, neuA).
A BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, USA) was used for Sanger sequencing in an ABI 3500 Genetic Analyzer (Applied Biosystems, Foster City, CA, USA) according to the instructions of the manufacturer. Sequencing PCR reaction clean-up was performed using a BigDye XTerminator Purification Kit (Applied Biosystems, USA), according to the manufacturer's instructions.
Sequencing data obtained with both primers were aligned using the CLC Main Workbench version 7.9.1 (QIAGEN Aarhus S/A, Germany), and consensus sequences were produced. The consensus sequences were then entered into the EWGLI Legionella SBT database (Public Health England) in order to determine respective allele numbers. The combination of all seven allele numbers was queried in the database for each isolate, and returned as the Sequence-Based Type (ST) number.

Phenotypic Characterization
L. pneumophila serogroup specificity is conferred by the lipopolysaccharide structure. Using monoclonal antibodies (MAbs), some serogroups can be further subdivided into MAb subgroups of the Dresden panel by IFA test. The analysis divided L. pneumophila SG 1 into the following subgroups: Knoxville, Philadelphia, France/Allentown, Benidorm, OLDA, Oxford, Heysham, Camperdown, and Bellingham [36]. A positive reaction was detected by green fluorescence revealed by an Eclipse E400 (Nikon, Tokyo, Japan) fluorescence microscope.

Statistical Analysis
Descriptive statistics were used to present the following data: relative frequency, mean value (MV) and median, standard deviation (SD), standard error (SE), interquartile range (IQR), and data range (MIN-MAX) as measures of data dispersion; the data were also presented graphically. The normality of data distribution was tested using the Kolmogorov-Smirinov test. Logarithmic transformation of microbiological data was carried out to normalize non-normal distributions. When obtained, the data did not follow normal distribution; nonparametric tests-Spearman's Correlation Coefficient, Mann-Whitney U test (M-W U)-were performed using the TIBCO Statistica 13.5.0 software package (TIBCO Software Inc., Palo Alto, CA, USA). Statistical significance was set at 0.05. Student's t-test, * p < 0.05, ** p < 0.01, *** p < 0.001 were considered significantly different in comparison to the control sample.

Legionella Prevalence
The prevalence of Legionella spp. in water distribution systems was examined during the 2013-2019 period in Primorje-Gorski Kotar County, Croatia [37]. The number of analyzed samples ranged from 93 in 2014 to 511 in 2019. In the first three years of the study period (2013-2015), Legionella spp. was not detected. From 2016 to 2019 the average percentage of Legionella-positive samples was 12.0%, with the largest value of 14.6% recorded in 2018 (Table 1). Of the total 1932 water samples analyzed, 171 (8.9%) were L. pneumophila positive. Samples were taken primarily from the following facilities: pools (N = 1091), hotels (N = 333), camping sites, (N = 331), ferries (N = 68), cooling towers (N = 59), apartments (N = 26), and rehabilitation centers (N = 24). Considering the share of positive samples per site, Legionella was most frequently found in cooling towers and rehabilitation centers (33.9% and 33.3%, respectively), in a similar percentage in camping sites, ferries, and apartments (between 19.2 and 21.8%), less frequently in hotels (11.7%), and the least frequent in pools (1.2%) ( Figure 1).

Heterotrophic Plate Counts (HPCs) in Relation to the Presence of Legionella
The suitability of Heterotrophic Plate Count (HPC) as a warning sign of the presence of L. pneumophila was also assessed. This parameter is usually used as an indicator of general water quality as well as for the assessment of the efficiency of the disinfection treatment. Routine microbiological testing is determined using YEA as a nutrient media. YEA is a nutrient-rich medium that allows recovery of a wide range of bacteria. Besides YEA, sheep BA is also used for HPC determination. BA is a nutrient-rich medium that simulates

Heterotrophic Plate Counts (HPCs) in Relation to the Presence of Legionella
The suitability of Heterotrophic Plate Count (HPC) as a warning sign of the presence of L. pneumophila was also assessed. This parameter is usually used as an indicator of general water quality as well as for the assessment of the efficiency of the disinfection treatment. Routine microbiological testing is determined using YEA as a nutrient media. YEA is a nutrient-rich medium that allows recovery of a wide range of bacteria. Besides YEA, sheep BA is also used for HPC determination. BA is a nutrient-rich medium that simulates the physiological human condition. It supports the growth of all clinically important bacteria that possess water-borne virulence factors [38,39].

Influence of Physical and Chemical Factors on Legionella Survival
According to our results, Legionella is found in systems with cold water within a temperature range of 8.0 to 35.1 °C. At the same time, it was isolated from warm water systems ranging from 22.0 to 61.0 °C. An M-W U test confirmed a statistically significant difference in the median temperature among the Legionella-positive and -negative samples, for both cold and warm water environments (M-W U, Z = 3.30, p < 0.0001 and Z = −5.31, p < 0.0001, respectively). The median of cold-water temperature of the samples that tested positive for Legionella was 1.8 °C higher compared to the negative samples. Likewise, the median of warm water temperature of the Legionella-positive samples was 9.8 °C lower than the negative samples (Figure 4a).

Influence of Physical and Chemical Factors on Legionella Survival
According to our results, Legionella is found in systems with cold water within a temperature range of 8.0 to 35.1 • C. At the same time, it was isolated from warm water systems ranging from 22.0 to 61.0 • C. An M-W U test confirmed a statistically significant difference in the median temperature among the Legionella-positive and -negative samples, for both cold and warm water environments (M-W U, Z = 3.30, p < 0.0001 and Z = −5.31, p < 0.0001, respectively). The median of cold-water temperature of the samples that tested positive for Legionella was 1.8 • C higher compared to the negative samples. Likewise, the median of warm water temperature of the Legionella-positive samples was 9.8 • C lower than the negative samples ( Figure 4a).
Considering the prevalence of L. pneumophila serogroups in relation to temperature, it appears that, in samples with a confirmed strain of L. pneumophila serogroup 1, the median temperature values were 8. perature range of 8.0 to 35.1 °C. At the same time, it was isolated from warm water systems ranging from 22.0 to 61.0 °C. An M-W U test confirmed a statistically significant difference in the median temperature among the Legionella-positive and -negative samples, for both cold and warm water environments (M-W U, Z = 3.30, p < 0.0001 and Z = −5.31, p < 0.0001, respectively). The median of cold-water temperature of the samples that tested positive for Legionella was 1.8 °C higher compared to the negative samples. Likewise, the median of warm water temperature of the Legionella-positive samples was 9.8 °C lower than the negative samples (Figure 4a).  Our results show significantly higher pH values for Legionella-positive samples (M-W U, Z = 3.41, p = 0.0006), ranging from 5.2 to 8.4, though for negative samples the range was wider, from 3.1 to 9 (Figure 5a). Correlation analyses revealed a positive association between the Legionella count and pH values. The results of our study indicate that lower electrical conductivity supports the survival of Legionella (Figure 5b).
A lower median value of free-chlorine concentration was shown in L. pneumophilapositive samples (Figure 5c). In one-third (48/144) of the samples that tested positive for Legionella, the concentration of free residual chlorine was below or at the detection limit of the method (≤0.02 mg/L). The highest proportion of samples (63.2%, 91/144) measured a concentration of chlorine at lower levels (0.03-0.1 mg/L). Only 4.9% (7/144) of the samples showed a chlorine concentration of 0.2-0.5 mg/L, while 7.6% (11/144) of the samples measured a concentration that was above 0.5 mg/L. Hence, in two-thirds of the samples, Legionella was present simultaneously with chlorine. However, the correlation between free residual chlorine and the concentration of Legionella was significantly negative (rs = −0.24, p < 0.05).
The median values of the nitrate concentration were lower in the L. pneumophilapositive samples (Figure 5d).

Sequence-Based Typing (SBT) and Phenotypic Characterization
One of the aims of our study was to identify the presence of SBT types most commonly circulating in the Primorje-Gorski Kotar County. Nine bacterial isolates from different water samples were typed by SBT analysis. The allelic profile of the seven alleles (flaA, pilE, asd, mip, mompS, proS, and neuA) was obtained for each isolate. Six different allelic profiles were identified, and six ST groups were defined. Prevalent ST 1317 exhibited the profile "16, 21, 12, 19, 31, 21, 215" ( Table 2). The allelic profiles obtained from our samples were already present in the EWGLI SBT database. The same nine isolates of L. pneumophila were used to determine phenotypic characterization. The isolates presented different phenotypic characteristics, but the most common phenotypes were Chicago 8 and Oxford/Olda ( Table 2).
The median values of electrical conductivity (M-W U, Z = −4.36, p < 0.0001), nitrate concentration (M-W U, Z = −5.74, p < 0.0001), and the concentration of free residual chlorine (M-W U, Z = −10.00, p < 0.0001) were lower in samples positive for Legionella compared to the negative ones. This was found to be in contrast with the pH values (M-W U, Z = 3.42, p = 0.0006) (Figure 5a-d). Our results show significantly higher pH values for Legionella-positive samples (M-W U, Z = 3.41, p = 0.0006), ranging from 5.2 to 8.4, though for negative samples the range was wider, from 3.1 to 9 (Figure 5a). Correlation analyses revealed a positive association between the Legionella count and pH values. The results of our study indicate that lower electrical conductivity supports the survival of Legionella (Figure 5b).
A lower median value of free-chlorine concentration was shown in L. pneumophilapositive samples (Figure 5c). In one-third (48/144) of the samples that tested positive for Legionella, the concentration of free residual chlorine was below or at the detection limit In our study, we also investigated whether a change would occur in the allelic profile and/or phenotype of the isolate after a pasteurization treatment at 53 • C for 10 min or after cultivation in A. castellanii. Legionella samples with an identified allelic profile underwent a pasteurization treatment as well as cultivation in amoebae, after which their phenotype was re-determined. To confirm the genetic changes, the isolates from the hotel (grown in amoeba or pasteurized) were selected for gene sequencing. Our results show that the pasteurization treatment and growth in amoeba do not affect the allelic profile and phenotype of the bacterium ( Table 2).

The Growth of Legionella in Amoeba A. castellanii
Representative samples of L. pneumophila isolated from different water systems were used to determine growth kinetics in amoeba cells. A. castellanii cells were infected with L. pneumophila (MOI 100) and incubated at 27 • C for 10 days. At 0, 2, 4, 6, 8, and 10 days after infection, the number of intracellular bacteria was determined by the CFU method.

Pasteurization Treatment of Agar-Grown and A. castellanii-Grown Legionella
L. pneumophila displays an ability to survive in purpose-built water systems, despite routine disinfection treatment. In our study, we determined the highest temperature at which Legionella in the samples taken in Primorje-Gorski Kotar County can survive. Tap water samples were inoculated with six Legionella isolates in a concentration of 10 6 CFU/mL for use in pasteurization trials. Pasteurization treatment at 63.5 and 55 °C for 10 min resulted in bacterial death. The highest temperature at which environmental Legionella isolates from this study can survive was 53 °C when treated for 10 min. All the bacteria survived heat treatment at 50 °C for 10 min, while only two out of the six tested L. pneumophila (autocamp cold and warm water) bacteria were killed after treatment at 50 °C for 30 min. L. pneumophila AA100 was used as control. Student's t-test, * p < 0.05, ** p < 0.01, *** p < 0.001 were accepted as significantly different from control sample.

Pasteurization Treatment of Agar-Grown and A. castellanii-Grown Legionella
L. pneumophila displays an ability to survive in purpose-built water systems, despite routine disinfection treatment. In our study, we determined the highest temperature at which Legionella in the samples taken in Primorje-Gorski Kotar County can survive. Tap water samples were inoculated with six Legionella isolates in a concentration of 10 6 CFU/mL for use in pasteurization trials. Pasteurization treatment at 63.5 and 55 • C for 10 min resulted in bacterial death. The highest temperature at which environmental Legionella isolates from this study can survive was 53 • C when treated for 10 min. All the bacteria survived heat treatment at 50 • C for 10 min, while only two out of the six tested L. pneumophila (autocamp cold and warm water) bacteria were killed after treatment at 50 • C for 30 min.
to the aforementioned results, L. pneumophila AA100 grown in amoeba were more susceptible to destruction by pasteurization treatment compared to agar-grown bacteria. The number AA100 amoeba-grown bacteria after pasteurization treatment was 4.3 × 10 5 CFU/mL, while the number of agar-grown bacteria was 1.4 × 10 6 CFU/mL (Figure 7, Student's t-test, p < 0.05). We can conclude that only the bacteria from environmental sources, after adaptation to the aquatic environment, and A. castellanii become more resistant to destruction by high temperature.

Cell Morphology
The morphology of agar-grown, amoeba-grown, and pasteurized L. pneumophila was observed using TEM, by monitoring the changes in the structure of the cytoplasm, cell wall, and the existence of bacterial appendages such as flagella and pili (Figure 8).
L. pneumophila is a rod-shaped bacterium. Most of the bacteria had between one and three polar or lateral flagella. On the surface, L. pneumophila expressed pili of variable lengths. In addition, the cells showed the intact cell wall layer and compact and high electron-dense cytoplasm (Figure 8a,d). Interestingly, compared to the agar-grown cells, A. castellanii-grown Legionella cells did not show significant differences in cell morphology (Figure 8b,e).

Cell Morphology
The morphology of agar-grown, amoeba-grown, and pasteurized L. pneumophila was observed using TEM, by monitoring the changes in the structure of the cytoplasm, cell wall, and the existence of bacterial appendages such as flagella and pili (Figure 8).  However, the pasteurization treatment at 53 °C for 10 min caused changes in bacterial morphology (Figure 8c,f). The changes were manifested by cell wall damage and the separation of the cytoplasm from the cell wall in around 70% of the bacteria (Student's t-test, p < 0.001). The morphology of the cytoplasm changed from high electron density to intermediate electron density. The disorganized cytoplasm also showed a tendency to clump in 67% of the pasteurized bacteria, when compared to the untreated bacteria (Student's ttest, p < 0.001). Furthermore, the pili were lacking on the cell surface and the number of flagella was drastically reduced (Figure 8c,f), Student's t-test, p < 0.01).
Examination of L. pneumophila by transmission electron microscopy demonstrated that growing in amoeba did not result in significant differences in cell morphology, while L. pneumophila is a rod-shaped bacterium. Most of the bacteria had between one and three polar or lateral flagella. On the surface, L. pneumophila expressed pili of variable lengths. In addition, the cells showed the intact cell wall layer and compact and high electron-dense cytoplasm (Figure 8a,d). Interestingly, compared to the agar-grown cells, A. castellanii-grown Legionella cells did not show significant differences in cell morphology (Figure 8b,e).
However, the pasteurization treatment at 53 • C for 10 min caused changes in bacterial morphology (Figure 8c,f). The changes were manifested by cell wall damage and the separation of the cytoplasm from the cell wall in around 70% of the bacteria (Student's t-test, p < 0.001). The morphology of the cytoplasm changed from high electron density to intermediate electron density. The disorganized cytoplasm also showed a tendency to clump in 67% of the pasteurized bacteria, when compared to the untreated bacteria (Student's t-test, p < 0.001). Furthermore, the pili were lacking on the cell surface and the number of flagella was drastically reduced (Figure 8c,f), Student's t-test, p < 0.01).
Examination of L. pneumophila by transmission electron microscopy demonstrated that growing in amoeba did not result in significant differences in cell morphology, while the pasteurization treatment showed significant changes compared to the untreated bacterial cells.

Discussion
During the study period (2013-2019), the average annual share of Legionella-positive samples in Primorje-Gorski Kotar County was 8.9% and was characterized by an upward trend. The largest percentage of Legionella-positive samples was found in cooling towers and rehabilitation centers (around 34%), while the smallest number in swimming-pool waters (1.2%). In a similar study conducted in Italy from 2014 to 2017 [40], Legionella was most frequently found in hotels (57.1%), followed by sport centers (41.2%), retirement homes (28.6%), and camping sites (12.5%). A low share of positive pool water samples was also revealed in an Italian study [41], in which Legionella spp. was found in 4.2% (2/48) of analyzed pool water samples. Therefore, it was concluded that the risk of being infected in such places is higher when showers are used, rather than by recreation in swimming pools.
In this study, L. pneumophila SGs 2-14 was the most commonly identified strain (76%) isolated. That is generally consistent with the results of numerous other studies, in which serogroups 2-14 accounted for over 50% of isolates [42]. The highest proportion of serogroup 1 was found in apartments (100%) and swimming-pool waters (69.2%). L. pneumophila was found in 13 of the 1091 examined pool water samples, of which 12 of them were of whirlpool spa pool type, where higher water temperature and aerosol formation in hydro-massage systems enhance the risk of infection [43]. According to the national regulations for swimming-pool water quality, in swimming pools with a water temperature ≥23 • C and the possibility of aerosolization of water, the concentration of free residual chlorine must be between 0.7 and 1.0 mg/L. Of the 12 L. pneumophila-positive spa water samples, all had a temperature greater than 30 • C, while only three samples (25%) had a free residual chlorine concentration greater than the required minimum of 0.7 mg/L. These results indicate the need for increased disinfection of the spa while meeting the prescribed criteria. Considering the prevalence of L. pneumophila strains, based on the type of facilities, an unequal distribution of serogroups can be perceived. Our results are consistent with those of a French study (2001)(2002), which found a proportion of L. pneumophila SG 1 in samples from environmental sources of 28.2% [44]. The presence of L. pneumophila SG 3 (10.8%) was significant in the environmental samples, while non-L. pneumophila species, L. anisa, accounted for 13.8% of the isolates. In a study conducted in South Korea in 2010 [45], a different ratio of serogroups among L. pneumophila species was found in environmental samples. It showed a higher share of L. pneumophila SG 1 (54.7%) when compared to L. pneumophila SGs 2-15 [45]. On the other hand, research carried out by Sikora, et al. [46]  The Legionella load in L. pneumophila SG 1-positive samples was higher compared to the L. pneumophila SGs 2-14-positive samples. These results are in contrast with those of a study conducted in Rome, Italy (2014-2017), in which the Legionella load was significantly higher in samples testing positive for L. pneumophila SGs 2-14 [40]. In the Italian study, Legionella-positive samples were mostly collected from larger buildings (hotels and sports centers) supplied with groundwater via the public distribution system, whereas in our study samples were collected from smaller distribution systems of camping sites using surface water. Considering the known higher virulence of serogroup 1, this could be taken into account when assessing the risk of Legionella contamination in camping sites [47]. In 35.5% (N = 61) of the positive samples, the acceptable value of 1000 CFU/L for L. pneumophila was exceeded (according to the proposal of the New Water Consumption Directive for water from domestic distribution). That share included more than half (20/38; 52.6%) of L. pneumophila SG 1 strain and less than a third (39/130; 30.0%) of the SG 2-14 isolates.
In our study, the Legionella load was found to be positively associated with the HPCs values determined by the pour plate method using sheep BA at 36 • C for 96 h. Overall, the correlation between HPCs values and the Legionella load is inconsistent throughout the literature. Thus, this relationship requires further research [21,40,48].
L. pneumophila multiply in the range of temperatures between 20 and 48 • C [49,50]. The highest temperature at which Legionella from Primorje-Gorski Kotar County can survive was 53 • C when treated for 10 min. This is consistent with the results of another study conducted in Southern Croatia, which found that water temperature higher than 54 • C is protective, while the temperature range of 44-54 • C positively affects the colonization of bacteria [24]. Our results showed that Legionella-positive samples were frequently collected from cold water distribution systems with higher water temperature and from warm water distribution systems with lower water temperature. The results of our study are in accordance with the results of a study conducted in Greece by Dimitriadi and Velonakis [51]. In their study, they measured the mean temperature of hot water of the Legionella-positive samples and found that it was 4.4 • C lower compared to that of the negative samples. A study conducted in Split-Dalmatian County by Rakić and Štambuk-Giljanović [52] included measurement of the lower median value of temperature in Legionella spp.-positive samples (by 7.0 • C) compared to the negative samples (47 • C vs. 54 • C). In contrast, in research carried out in Iran, Tabatabaei, et al. [53] found that the average temperature of Legionella spp.-positive water samples was 6.6 • C higher compared to the average temperature of the negative samples (44.2 • C vs. 37.6 • C). In other words, both medians fell within the temperature range that is favorable for bacterial growth. In samples with a confirmed strain of L. pneumophila SG 1, the median temperature values were 8.8 • C higher compared to the samples that tested positive for SGs 2-14. A higher incidence of L. pneumophila SG 1 in warm water was also found by a Korean study conducted by Lee et al. [45]: SGs 2-14 that occurred to a greater extent in cold water (61.8%), and SG 1 in warm water (58.8%).
In our study, the survival of Legionella was enhanced in the samples with higher pH values, while nitrate concentration, electrical conductivity, and free residual chlorine significantly reduced Legionella survival. In general, it is believed that higher pH values (6.0-8.0) enhance the survival rate of bacteria [52]. The survival of Legionella at lower electrical conductivity was also found by a study carried out by Lasheras, et al. [54], although in contrast with the results of Karki and Cann [55].
It is known that the correlation between free residual chlorine and Legionella concentration is significantly negative. In our study, however, Legionella was present simultaneously with chlorine in two-thirds of the samples. This indicates the possibility of bacteria being protected within the biofilm, thus reducing the effect of chlorine disinfection [52]. Rakić et al. [24,52], in their work, carried out in the Split-Dalmatia County (Croatia), did not determine the correlation between chlorine concentration and Legionella presence and suggested possible resistance of Legionella spp. to chlorine.
Most of the environmental samples in our study belonged to ST 1317, while the most common phenotypes were Chicago 8 and Oxford/Olda. In a study conducted in Northern Sicily, SBT analysis of 86 L. pneumophila SG 1 isolates, from water distribution systems, showed that most samples belonged to ST1 [56]. The ST1, which we have found in a wellness center, was also the most frequently detected ST among clinical isolates in Germany [57], and is considered as the most common SBT in Europe. In a study conducted in China, L. pneumophila SG 1 isolates from cooling towers, hot springs, and drinking-water systems were analyzed by SBT. The results of the study also show that ST1 was the most prevalent ST in China, accounting for 49.4% of the analyzed strains [58].
In 2010, there was an outbreak of legionellosis in a Slovenian nursing home. The SBT analysis showed an identical sequence type 23 in the water supply system and the isolates from the patients [59]. In a study in Catalonia, 528 L. pneumophila isolates from the environment (cooling towers, hotels, hospitals, water distribution system, spas, sprinklers, tanks, pools) and the clinic were described and compared using SBT and monoclonal antibody subgrouping, and 32% of the clinical isolates belonged to ST1, ST23, and ST37, while 40% of the environmental isolates belonged to ST1 and ST284 [60]. The most abundant subgroups of clinical isolates were Philadelphia (26.61%), Knoxville (19.27%), and OLDA (14.68%), while OLDA (33.1%) were the most frequent environmental isolates [60]. A recent study in the Croatian northern Adriatic presented a case of LD in a patient who visited a spa complex in a hotel, where the isolate from the water belonged to ST82 and the Allentown/France MAb subgroup [61].
L. pneumophila, as well as numerous other bacteria, have shown the ability to enter and multiply in A. castellanii cells [16]. A recent study also showed that amoebae likely promote survival of some bacteria in aquatic environments [62]. After the life cycle in an amoeba, L. pneumophila becomes more resistant to destruction by pasteurization treatment compared to agar-grown bacteria. Our results show that pasteurization treatment and growth in amoeba do not affect the allelic profile and phenotype of the bacterium, while examination by transmission electron microscopy demonstrated significant changes in the morphology of pasteurized bacterial cells. A. castellanii-grown Legionella cells did not show significant differences in cell morphology compared to agar-grown cells. This is in contrast with a recent study on Francisella novicida, in which the morphology of F. novicida changed after growth in A. castellanii; this was manifested through cell wall damage, the reduction in cytoplasm electron density, and leakage of intracellular material [32]. However, our results are consistent with the Legionella counts, which showed that amoeba-grown Legionella becomes more resistant to pasteurization treatment.

Conclusions
Legionellae are opportunistic pathogens that are widely distributed in aquatic environments. They survive in aqueous media at varying temperatures, electrical conductivity, pH, and nutrient levels, as well as varying concentrations of free residual chlorine in treated waters. The risk of Legionella proliferation is increased in cold water distribution systems with a temperature above 22.3 • C and warm water systems with a temperature below 43.5 • C. The risk of Legionella proliferation is also increased in water with electrical conductivity less than 419 µS/L, nitrate concentration less than 1.7 mg/L, free residual chlorine concentration less than 0.04 mg/L, and pH greater than 7.8. The sources of Legionella transmission to humans are well characterized, and all of the sources examined in this study involve aerosolization of water contaminated with Legionella. Drinking water is considered the most important source of Legionella transmission, especially in complex systems such as camps, where water supply pipes are often upgraded and not laid deep enough in the ground, making it difficult to maintain the required temperatures. In such systems, the risk is additionally increased due to the higher Legionella load of serogroup 1. It is important to monitor serotypes and genotypes from clinical samples in order to link them with the source of infection in environment, especially in tourist-attractive countries. Our results showed that most of the samples from the apartments belong to a highly infectious serogroup 1, indicating the importance of preventive measurements to control the infection. Of all Legionella-positive drinking-water samples, 35.5% were above the acceptable new EU Water Consumption Directive level of 1000 CFU/L for L. pneumophila. Legionellae survive and thrive in vesicles after being ingested by amoebae in water resources. The data from this study contribute to the understanding of the effects of increased resistance of amoeba-cultured L. pneumophila from amoeba cultures to pasteurization treatment.
The fact that amoeba-grown L. pneumophila becomes more resistant to pasteurization treatment suggests that growth in protozoan cells may be responsible for the development of the disease in humans, and future studies should focus on a better understanding of the way in which bacterial growth influences the pathogenesis of the disease. The results of this study will be useful for managing Legionella in water systems of different facilities.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.