Temporal dynamics of Legionella (Proteobacteria, Legionellaceae) in two Pampean shallow lakes from Argentina

doi:10.21203/rs.3.rs-4298158/v1

Download PDF

Research Article

Temporal dynamics of Legionella (Proteobacteria, Legionellaceae) in two Pampean shallow lakes from Argentina

https://doi.org/10.21203/rs.3.rs-4298158/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Aquatic systems have traditionally played a key role in the development of human life, providing multiple ecosystem services to society and being a reservoir for a wide biodiversity of organisms. Among them, bacteria belonging to Legionella stand out, mainly because they are of great interest both in the field of microbial ecology and public health, since some of them turn out to be pathogenic for humans. The aim of this work was to study the monthly temporal dynamics of Legionella spp. and its relationship with the environmental variables measured in two Pampean shallow lakes (Gómez and Carpincho, Buenos Aires Province, Argentina). The analysis was carried out using a quantitative approach by real-time Polymerase Chain Reaction (qPCR) and a non-quantitative approach using bacterial diversity data obtained by Next Generation Sequencing (NGS), using the Illumina MiSeq platform. Our results showed that the overall Legionella abundance was very high in the studied Pampean shallow lakes. Notably, fluctuations in dissolved organic carbon and temperature influenced the dynamics shifts in Legionella abundances. Correlation analyses between Legionella reads from NGS and copy numbers obtained through qPCR revealed positive relationships, unveiling distinctions attributable to the diverse sequence processing algorithms employed in the analysis of NGS data.

Legionella spp.

Pampean shallow lakes

temporal dynamics

Next Generation Sequencing

qPCR

environmental microbiology

The Legionella genus comprises gram-negative, aerobic bacilli that are naturally found in both freshwater (Rowbotham 1980) and saltwater environments (Gast et al. 2011), as well as artificial water systems (cooling towers, building pipes and whirlpools) (Lee et al. 2011). These species are of great public health concern mainly because among the 60 identified species, 24 are known to cause human diseases (Lösch and Merino 2016). Legionella spp. infections comprise one of the leading causes of pneumonia in humans, whether they are in-hospital or community-acquired (Stout and Yu 1997). Given the non-specific nature of its symptoms, diagnosis is often complicated, resulting in two distinct clinical outcomes: Pontiac fever, an influenza-like illness characterized by fever, headache and myalgia without signs of pneumonia; and Legionnaries’ disease, a potentially fatal pneumonia (Murdoch 2003). This pulmonary and multisystemic disease can be potentially fatal, with 3 to 33% of infections leading to death (National Academies of Sciences 2020). The 2021 annual epidemiological report from the European Centre for Disease Prevention and Control (ECDC) has informed the highest annual notification rate, with 29 countries reporting 10,004 of 10723 (93%) confirmed cases of Legionnaires’ disease (ECDC, 2023). In particular, in Argentina during September 2022, a cluster of 11 cases of bilateral pneumonia was detected in the province of Tucumán in which Legionella was finally detected as the causative agent, resulting in the death of 4 patients (WHO, 2022).

The recurrence of epidemiological outbreaks associated with Legionella is primarily attributed to their transmission routes, with the most significant source being the inhalation of aerosols containing these microorganisms. This is particularly noteworthy because aerosols can travel significant distances from their point of origin (Caicedo et al. 2016). The majority of outbreaks occur in treated recreational waters (swimming pools, spas and recreational parks). Conversely, the risk associated with natural waters, such as rivers, shallow lakes, beaches, among others, is considered less important (Doménech-Sánchez et al. 2008, Leoni et al. 2018).

However, natural surface water and groundwater have been identified as potential sources of Legionella outbreaks. This is supported by studies that have linked the incidence of legionellosis to the use of natural water sources (van Heijnsbergen et al. 2015). In addition to evidencing the ubiquity of Legionella pneumophila in aquatic ecosystems, research carried out by Fliermans et al. (1981) demonstrated that this bacterium can also survive in a wide range of physical and chemical conditions. Additionally, Legionella displays resilience in adverse conditions through the evolution of diverse adaptive mechanisms, including its ability to inhabit biofilms and eukaryotic microorganisms (Grabowska-Grucza and Kiersztyn 2023). The parasitization of diverse protozoa contributes to Legionella transmission, enabling the bacteria to endure, proliferate and migrate in challenging conditions (William et al. 2022). This adaptive capacity enables Legionella to thrive across diverse environments with distinct environmental characteristics. Prior research has identified Legionella in various natural water bodies such as the Antarctic Peninsula (Carvalho et al. 2008), hydrothermal waters (Verissimo et al. 1991), glacial-origin eutrophic lakes (Grabowska et al. 2020), natural European aquatic habitats (lakes, groundwaters, rivers; Kanarek et al. 2022) and coastal shallow lakes (Selak et al. 2022).

Despite this widespread occurrence, to the best of our knowledge, no previous studies have assessed the occurrence of Legionella in Pampean shallow lakes (Argentina). Moreover, there have been no reported outbreaks of Legionella-related disease in the Pampa Region. It is noteworthy that Legionella-induced pneumonia is often underreported on a global scale, potentially by a magnitude of eight to ten times (National Academies of Sciences 2020), due to its symptoms closely resembling pneumonia caused by other pathogens (Walser et al. 2014).

Accordingly, investigating potential reservoirs where Legionella is present is considered crucial for both sanitary and ecological reasons. Exploring the habitat of Legionella may be helpful in the effective control of this pathogen (Kanarek et al. 2022). Consequently, the aim of this work was mainly focused on assessing the presence and temporal dynamics of Legionella spp. by molecular techniques such as qPCR and NGS, while analyzing their relationship with environmental variables in two Pampean shallow lakes (El Carpincho and Gómez, Buenos Aires Province, Argentina). We hypothesized that Legionella is present in these natural shallow lakes and that variations in environmental factors, including both physical and chemical water parameters, drive fluctuations in the abundance of these bacteria.

Study area

The studied shallow lakes are situated in the Pampa Plain (Argentina) and are characterized for being polymictic, turbid phytoplanktonic and hypertrophic (Quirós and Drago 1999). Due to their shallow depths – averaging 1.1 meters in Gómez (area 36.6 km²) and 1.2 meters in Carpincho (area 4.4 km²) – and the influence of persistent winds, these lakes exhibit significant homogeneity in most physical and chemical parameters throughout the water column (Quirós 2004, Schiaffino et al. 2019). Gómez (34.6650°S; 61.0287°W) and Carpincho (34.5769°S; 60.8988°W) shallow lakes are located near the city of Junín (34.5972°S; 60.9436°W) and are connected upstream and downstream to the Salado River, serving as natural reservoirs along its main course. Both lakes present floodgates that regulate their water levels, and are also mainly used for recreational purposes (fishing, canoeing, windsurfing, etc.). Additionally, Gómez shallow lake is used for bathing and swimming during spring and summer seasons. A comprehensive and detailed characterization, including principal physical, chemical and biological parameters, has been conducted by Quirós and Drago(Quirós and Drago 1999) and Schiaffino et al. (Schiaffino et al. 2019, 2020).

Sampling and measurement of environmental variables

Integrated water samples (5 liters each) were collected from Balneario de Gómez, a bathing area of Gomez shallow lake, and from Carpincho shallow lake throughout January to December 2016 (N = 12 for each lake). All samples were obtained from the pelagic zone of the lakes at a depth of 30–40 cm below the surface and transported to the laboratory at 4ºC under dark conditions. It should be noted that the sampling year was characterized by an excess of water due to increased rainfall in summer and autumn, with minimums observed during the winter and spring (Quiroga 2020). Several in situ environmental variables were registered: conductivity, pH, temperature (HANNA HI991301), dissolved oxygen (DO) (HACH HQ30d multisensor) and water level (Echosounder Fish Finder).

Determinations for total nutrients, including Kjeldahl’s total nitrogen (KTN), total phosphorus (TP) and dissolved nutrients such as ammoniacal nitrogen (N-NH₃) and soluble reactive phosphorus (SRP), along with dissolved organic carbon (DOC) were performed according to Schiaffino et al. (2019).

Molecular approaches

To obtain the environmental DNA, samples were collected in duplicates, and 100 mL of water were pre-filtered through 51 µm nets to remove large particles. Subsequently, samples were filtered again using 0.22 µm pore polycarbonate membranes (Millipore). Two sets of 12 DNA filters were generated: one set was utilized in this study for performing qPCR to examine the dynamics of Legionella spp. abundance. The second set had been previously used for NGS of 16 rRNA gene by Illumina MiSeq platform to investigate the composition of planktonic bacteria (Seone Rocha 2018; Quiroga 2020; Schiaffino et al. 2020; Nuozzi et al. 2022). In the current study, only the Legionella reads obtained from NGS data were utilized. Both sets were preserved in cryovials, snap-frozen in liquid nitrogen, and subsequently stored at -80°C until further processing. The hexadecyltrimethylammonium bromide (CTAB)/chloroform/isoamyl alcohol-based protocol (Fernández Zenoff et al. 2006) was employed to extract the DNA retained on the filters. Extracted DNA was purified by filtration (Amicon® Ultra-4, Merck) and then the integrity and concentration of the purified DNA extracts were assessed using agarose gel electrophoresis and fluorometric techniques (Qubit 2.0 Fluorometer, Life Technologies). Results were expressed as nanograms of DNA per microliter of sample (ng DNA / µl sample).

Real-time PCR (qPCR)

The extracted and purified DNA corresponding to the second set of 12 filters was utilized for the quantification of Legionella spp. through qPCR. Each sample underwent analysis using the Mericon Quant Legionella spp. kit (QIAGEN, ID: 290085). All reactions were run simultaneously on the Bio-Rad CFX96 Touch thermocycler (Bio-Rad, Hercules, CA), following the instructions provided by the manufacturer. The kit has a detection limit of 25,000 copy numbers per PCR reaction, therefore a 20-fold dilution of samples in ultrapure water was required to conform to the standard curve.

Following qPCR analysis, Legionella copy numbers per PCR reaction were standardized to copy numbers per nanogram of DNA. Subsequently, conversion to Colony Forming Units (CFU) was performed using the equation described by AbuOdeh et al. (2017), where CFU Legionella / mL sample = [(Legionella copy number / mL sample) / 20]. This transformation was implemented as a proxy to facilitate a comparison of Legionella levels in the studied shallow lakes with those reported in natural surface freshwater environments (Schwake et al. 2021 and citations therein) and artificial water systems (AbuOdeh et al. 2017)

Comparative analysis of Legionella abundance obtained by qPCR versus Legionella reads obtained by NGS

As previously mentioned, the second set of replicates collected from both shallow lakes underwent NGS. The partial 16S rRNA gene was sequenced by Illumina MiSeq 2 x 300 paired-end sequencing (Macrogen Inc., Corea), using the primers 341F/805R (Herlemann et al. 2011), which cover the hypervariable regions V3-V4. After quality control, reads were analyzed using three different sequence processing algorithms: the UPARSE algorithm (Edgar 2013) that defined Operational Taxonomic Units clustered to a 97% similarity threshold (97% OTUs) as depicted in Seone Rocha (2018), the UNOISE algorithm (Edgar 2016) defining zero-radius OTUs (zOTUs) as shown in Quiroga et al. (2020) and Schiaffino et al. (2020) and clustering of sequences using the DADA2 methodology (Callahan et al. 2016) to obtain Amplicon Sequence Variants (ASVs), as shown in Nuozzi et al. (2022).

Regardless of the type of data processing used, in all resulting matrices, sequences belonging to archaea, mitochondria and chloroplasts were filtered and removed. Subsequently, data were rarefied using R v3.4.4 software along with the foreign (R Core Team 2018), permute (Simpson 2019), lattice (Sarkar 2008), packfor (Dray et al. 2007) and vegan (Oksanen et al. 2015) packages. From all the resulting reads, a subset of data containing only sequences belonging to Legionella was utilized in the present work. This information, obtained from previous studies conducted in the same shallow lakes and sampling dates, was compared with the Legionella copy numbers obtained through qPCR.

Statistical analyses

The standardized Legionella abundance data, along with environmental variables, were analyzed using Principal Component Analysis (PCA) and Multiple Factor Analysis (MFA), in order to evaluate the relevance of these variables and their overall trends, and to simplify multidimensional data structures. A multiple regression model was performed with Legionella abundance obtained by qPCR as the dependent variable (response) and environmental variables as independent (explanatory) variables. The relevant predictors were selected using a stepwise procedure. The dependent variable (Legionella copy number per ng DNA) was transformed with natural logarithm (ln) to model a continuous response. The Shapiro-Wilks test was performed to validate the normal distribution of residuals, and the Durbin-Watson test was used to rule out the presence of autocorrelation between the residuals corresponding to each value and the previous one, ensuring the independence of the observations. Finally, comparisons between Legionella abundance obtained by qPCR and Legionella reads obtained by NGS, were conducted by Spearman's paired correlations. The results were then visualized through plots created with GraphPad Prism 8 software. All analyses were performed with the R v.4.2.2 software (R Core Team 2022) and the tidyverse package (Wickham et al. 2019). PCA and MFA were performed using the package FactoMineR (Lê et al. 2008) and their results plotted with the help of factoextra (Kassambara and Mundt 2020). The Durbin-Watson test was performed using the package lmtest (Zeileis and Hothorn 2002).

Physical and chemical variables

Environmental variables of Carpincho and Gómez shallow lakes over the 12-month sampling period are presented in Table 1. The results indicated alkaline conditions (pH>8.56) and oligohaline or brackish characteristics (salinity between 1.66 and 3.04 g/L) in the water, with similar physical and chemical values in both shallow lakes (Table 1). Additionally, Carpincho and Gómez shallow lakes exhibited similar temporal dynamics in the measured physical and chemical variables, such as lower water temperature (Supplementary Fig. 1a,b), pH (Supplementary Fig. 1c,d), salinity (Supplementary Fig. 1e,f) and DOC levels (Supplementary 1g,h) during winter months. Moreover, lower DO values (Supplementary Fig. 1i,j) were observed during the summer period.

Temporal dynamics of Legionella

For the qPCR experiment, the reaction efficiency was E= 93.4% and the regression coefficient of the standard curve R²=0.986 (slope=-3.491; intercept=37.853). The qPCR results exhibited a notably consistent fluctuation of the Legionella bacteria copy numbers in both lakes throughout the 12-month sampling period. Higher values were registered in May, July and October, while lower numbers were observed in January, February and December (Figure 1a and 1b). The overall Legionella abundance was very high in the studied Pampean shallow lakes. Mean values were 3878 copies/ng DNA for Carpincho and 2418 copies/ng DNA for Gómez shallow lakes. Transformed values of Legionella to CFU/mL and copy number from qPCR experiment are shown in Fig. 1a and 1b for Carpincho and Gómez, respectively. CFU numbers were also elevated, ranging from 14 to 32672 UFC/mL (average 10343 UFC/mL) in Carpincho and from 76 to 50462 UFC/mL (average 11976 UFC/mL) in Gómez shallow lakes.

Environmental variables analyses related to Legionella copy numbers

The first three principal components (PC) of the PCA analysis were selected for both shallow lakes, representing 84.2% of the total variability in Carpincho (PC1: 52.0%, PC2: 19.8% and PC3: 12.4%), and 85.2% in Gómez (PC1: 57.8%, PC2: 16.3% and PC3: 11.1%). Examining the correlations of each component with the original variables in Carpincho, the first component demonstrated strong negative correlations with temperature (r=-0.95), pH (r=-0.89), salinity (r=-0.85) and DOC (r=-0.76), while displaying positive correlations with DO (r=0.77), N-NH₃ (r=0.72) and Legionella abundance (r=0.66). The second component showed positive correlations with KTN (r=0.88) and SRP (r=0.63) (Fig. 2a). The third PC was strongly and negatively associated with water level (r=-0.79). In Gómez shallow lake, the first component presented positive correlations with temperature (r=0.92), salinity (r=0.92), pH (r=0.83) and DOC (r=0.82), and negative correlations with DO (r=-0.87), water level (r=-0.84), N-NH₃ (r=-0.72) and Legionella copy numbers (r=-0.72). The second component showed positive correlations with KTN (r=0.80) and SRP (r=0.71) (Fig. 2b). The third PC was moderately correlated with SRP (r=0.55), KTN (r=-0.51) and DOC (r=0.50).

Overall, samples from both shallow lakes demonstrated a temporal ordination primarily influenced by temperature, water nutrients (KTN, SRP and N-NH₃), chemical factors (pH, salinity, DO) and biological variables (Legionella abundance). In both lakes, samples corresponding to May, June and July clustered together with higher concentrations of N-NH₃ and Legionella abundances; concurrently, these periods had the lowest temperature, pH, salinity and DOC values. Conversely, samples from January, February and March showed the highest values of temperature, pH, salinity and DOC, along with the lowest Legionella abundance and N-NH₃ concentrations. Samples collected in September and October also presented elevated Legionella numbers (Figure 1), but their salinity and KTN values surpassed the averages (Table 1), placing them generally near the center of the biplot between the first and second components.

A combined analysis of Carpincho and Gomez shallow lakes was performed using MFA. The two first dimensions explain 71.3% of variability (53.7% the first component and 17.6% the second). Information from both lakes contribute strongly to the first dimension of MFA (normalized coordinates of 0.98 and 0.98), and moderately to the second dimension (normalized coordinates of 0.37 for Carpincho and 0.28 for Gomez). The close proximity of the lakes in the factorial plane indicates that they share a similar structure in the behavior of the environmental variables and of Legionella abundance. The analysis of the consensus structure shows that the first dimension of MFA is strongly and positively correlated with temperature (r=0.95), pH (r=0.86), salinity (r=0.89) and DOC (r=0.81), and negatively with water level (r=-0.65), DO (r=-0.81), N-NH3 (r=-0.73) and Legionella abundance (r=-0.69). The second dimension is strongly and positively correlated with SRP (r=0.73) and KTN (r=0.87). On the other hand, comparison of partial and consensus structures shows that, in Carpincho, environmental variables DO, KTN and SRP are strongly correlated, contrary to Gomez’s correlation structure.

A stepwise multiple regression analysis was also conducted between the measured environmental variables and the ln of Legionella abundance. The Shapiro-Wilks test confirmed the normal distribution of residuals (W=0.9653, p=0.55) and the Durbin-Watson test (d=2.43, p=0.78) ensured compliance with the assumption of independence of observations. The lake was introduced in the model as a potential predictor, but resulted not significant (p>0.05). The model selected only the variables DOC and temperature as predictors of variations in Legionella abundances in both shallow lakes. An adjusted R² of 0.72 was obtained, leaving the model equation as follows (1):

Log_eLegionella abundance = 9.02 – 0.19 * DOC + 0.57 * temperature – 0.02 * temperature²

The model indicates that the effect of DOC and temperature on the abundance of Legionella is the same in both shallow lakes. For a constant temperature, an increase in DOC produces a decrease in the mean level of the ln of Legionella abundance. The effect of temperature on the ln abundance of Legionella, however, is not linear: for a constant DOC level, an elevation in temperature can lead to either an increase or decrease in the ln abundance of Legionella, depending on the original temperature range. Up to 15°C, Legionella abundance increases as temperature rises. Conversely, in warmer conditions surpassing 15°C, the abundance of Legionella decreases as temperature increases (Fig. 3). This suggests that the response of Legionella to temperature differs depending on the time of the year, and possibly in relation to other environmental variables. For the lowest observed DOC levels (during cold months), increase in temperature produce a large effect in the mean of Legionella abundance. Conversely, at higher values of DOC (warm months), a rise in temperature induces a small decrease in mean Legionella abundance (Fig. 3).

NGS analysis and comparison with Legionella abundances obtained by qPCR

Data derived from NGS, processed using the 97% OTUs, zOTUs and ASVs approaches, revealed divergent outcomes, both in the overall richness of reads and in those belonging specifically to Legionella spp. In both shallow lakes, the NGS analysis of the 16S rRNA gene, employing a 97% similarity threshold of OTU grouping, unveiled a total bacterial richness of 3605 97% OTUs, with 22 belonging to Legionella. As for the cleaning using the zero-radius OTUs, it yielded a total bacterial richness of 7528 zOTUs, including 16 linked to Legionella. It should be noted that in both cleaning approaches, no particular Legionella species were identified. The sequence processing using the ASVs approach ultimately yielded the highest bacterial richness, generating a total of 7806 ASVs, with 44 corresponding to Legionella spp. Notably, this approach allowed for species differentiation, including three previously identified species (L. adelaidensis, L. tunisiensis and L. quateirensis) and two still unclassified species (L. KR14 and L. D2863).

In general, we observed comparable patterns in the dynamics of qPCR abundance and the number of Legionella reads obtained through NGS (Fig. 4 a-f). However, correlations between the number of Legionella reads obtained from different sequence processing algorithms (NGS) and the Legionella copy numbers obtained by qPCR revealed distinct outcomes (Table 2). The zOTU bioinformatic approach demonstrated the strongest correlation with Legionella qPCR abundances in both lakes (Carpincho r=0.93; Gómez r=0.91; Table 2), with peak abundances observed in the months of May, July and October, aligning with the dynamics of Legionella determined by qPCR (Fig. 4c, d). Cleaning by grouping 97% OTUs showed correlation coefficients of 0.68 for Carpincho and 0.58 for Gómez shallow lakes (Table 2). Maximum values of NGS Legionella reads, except for the month of January, coincided with those recorded for the abundance of Legionella by qPCR (Fig. 4a,b). Finally, processing by ASVs exhibited the lowest correlation coefficient for Carpincho shallow lake (r=0.26, Table 2). Although a similar trend was observed between both abundances, disparities were noted during the months of January, October and December (Fig. 4e). On the other hand, for Gómez shallow lake, the ASV bioinformatic approach yielded a positive correlation (r=0.63, Table 2), with coincident peaks in April, May, July and October for qPCR copy number and ASVs reads (Fig. 4f).

Influence of environmental variables on Legionella abundance

In this study, we investigated the dynamics of aquatic bacteria belonging to the Legionella genus, and explored the environmental variables influencing their abundances. Our results support the initially proposed hypothesis, confirming the presence of Legionella in the studied Pampean shallow lakes and revealing significant temporal fluctuations and abundance peaks of Legionella, notably occurring during the months of May, July and October. Furthermore, our study underscores a substantial influence of environmental factors on the Legionella abundance in water samples from both lakes. Temperature and DOC mainly modulated Legionella spp. abundance in the studied Pampean shallow lakes. Both variables emerged as the most important included in the multiple regression model, emphasizing their role in shaping the dynamics of Legionella in the studied lakes. We found that the effect of temperature on the ln abundance of Legionella is quadratic. For a constant DOC level, at temperatures up to 15°C the abundance of Legionella tends to increase with rising temperatures. However, in warmer conditions exceeding 15°C, the abundance of Legionella diminishes as temperatures continue to rise. This implies that Legionella’s response to temperature varies temporally or seasonally, and may also be influenced by other environmental factors, such as DOC. When temperature remains constant, an increase DOC levels results in a decrease in the average level of the Legionella abundance. However, it is important to remark that the strong correlation structure between temperature and other environmental factors induces a collinearity issue during the model selection process. Adding more variables to a model that already includes temperature as a predictor, does not provide significant new insights, as many of these variables are closely correlated to temperature. This highlights the complex interplay of environmental factors on Legionella populations (Schwake et al. 2021).

Previous studies found changes in the survival, viability and metabolism of L. pneumophila in culture, responding to variations in temperature in marine, thermal and drinking waters (e.g., Heller et al. 1998; Ohno et al. 2003; Bennett and Bentham 2014). In addition, Legionella has the ability to survive and grow as parasites within free-living protozoa and within biofilms that develop in water systems (WHO 2022), having the potential to survive in a wide range of environmental factors. Legionella lives and grows in water systems at an ideal temperature range from 20 to 50ºC, when nutrients are available (WHO 2022). However, it was found that cold temperature freshwater sources (< 20ºC) have the potential to contain Legionella (Schwake et al. 2021). In the Netherlands, 100% of river water with temperatures of 3 to 4ºC were found to have L. pneumophila, and relatively high concentrations up to 2.5 x 10³ cells/mL were recorded (Wullings and van der Kooij. 2006). These authors found that a large diversity of yet-uncultured Legionellae are common members of the microbial communities in surface waters and groundwaters treated at water temperatures of below 15°C. Accordingly, Bennett and Bentham (2014), performed laboratory microcosms and reported that both increasing salinity and temperature reduced survival of Legionella. They suggested that the protective role of amoebae and biofilm in Legionella survival and multiplication cannot be disregarded in this context. A survey of five lakes and rivers in differing climate zones of Southern Korea demonstrated the ubiquity of L. pneumophila using qPCR results (Bahk et al. 2020). All the sites had detectable L. pneumophila in at least one sample, with higher positivity observed during winter season. This seasonal pattern throughout a variety of climates shows that other environmental factors beyond temperature play a key role in Legionella populations (Wullings and van der Kooij 2006).

Additionally, DOC serves as a necessary element for the growth of heterotrophic bacteria, playing an important role in the carbon cycle through the microbial loop, over which it returns to higher levels of the food web. Phytoplankton, particularly cyanobacteria, serve as major contributors to autochthonous organic matter, supplying DOC and other nutrients to the ecosystem. In our study, a negative association was found between Legionella abundance and DOC concentration in the water, for both Carpincho and Gómez shallow lakes. This association was partially explained by Tison et al. (1980), suggesting that a decrease in DOC concentrations may result from increased utilization of algal-derived extracellular products by the heterotrophic microbiota, including bacteria belonging to Legionella.

On the other hand, Gómez and Carpincho shallow lakes exhibited remarkably similar fluctuation in physical, chemical and biological variables (i.e., abundance of Legionella spp.), likely due to their connection to the Salado River and interconnection with each other, as it was previously reported by Schiaffino et al. (2020) and Sagua et al. (2023).

Comparison of Colony Forming Units (CFU) of Legionella with other water environments

As expected, among the 24 collected samples, the months with the highest Legionella counts in CFU/mL coincided with those which presented the maximum Legionella copy numbers. The obtained values were much higher than those reported for other natural and artificial water environments (Schwake et al. 2021, AbuOdeh et al. 2017). In particular, the majority of these samples had Legionella concentrations exceeding those reported in cooling towers from Qatar by AbuOdeh et al. (2017), which ranged from 0.02 to 199.50 CFU/mL. However, fluctuations in these levels across different months were consistent with the broader variability observed in Legionella concentrations across various surface freshwater ecosystems (Schwake et al. 2021, and citations therein). These higher abundances of Legionella may be explained because when compared against other lakes worldwide, Pampean shallow lakes depart from most of them as having higher TP, KTN, and chlorophyll-a concentrations and much lower transparency, and therefore they stand as extremes of the trophic-state continuum (Diovisalvi et al. 2015). These environmental conditions likely support high abundances of Legionella. The elevated levels of Legionella compared to other studies represent a potential health concern despite the fact that no legionellosis outbreaks had yet been registered in the area. Notably, the Committee of Management of Legionella in Water Systems (National Academies of Sciences, 2020) identifies a Legionella concentration exceeding 50 CFU/mL as a critical threshold, indicating a level of concern that demanded prompt remediation efforts.

Despite the potential overestimation of Legionella occurrence by qPCR due to the detection of nonviable organisms, given the high values detected of these potentially pathogenic bacteria, we suggest further monitoring of the studied shallow lakes, as conventional microbiology cultures were not performed. Especially since there is no epidemiological monitoring and scarce environmental microbiological control by the local authorities. We hope that with this study a more exhaustive control can be carried out in the future.

Comparative of NGS and qPCR abundances

As mentioned above, qPCR data was compared with a subset of NGS reads from previous studies, employing different processing methodologies: OTUs with a 97% similarity threshold (Seoane Rocha 2018), zero-radius OTUs (Quiroga 2020; Schiaffino et al. 2020) and ASVs (Nuozzi et al. 2022). It becomes evident that each bioinformatic analysis method places different importance on species distinctiveness, overall diversity and fidelity. This variability contributes to differences in outcomes and interpretations of genetic information across various methodologies. Specifically, for Carpincho and Gómez shallow lakes, the zOTU approach showed the highest correlation between the number of copies obtained by qPCR and the number of reads generated by NGS. This is attributed to the denoising process, which eliminates sequencing errors and enhances read accuracy. Classical processing, grouping OTUs at a 97% similarity threshold (Edgar 2016; Logares 2018), displayed a significant medium correlation coefficient in both shallow lakes. However, none of these processing methodologies (zOTU or 97% OTU) were able to identify any species.

Processing using the DADA2 algorithm (Callahan et al. 2016) allowed for the determination of more ASVs (7806 total, 44 corresponding to Legionella), although it led to a reduction in correlation with the number of copies obtained by qPCR in Carpincho shallow lake. Nevertheless, this methodology facilitated the identification of Legionella species. DADA2 controls errors sufficiently to resolve single nucleotide differences in the sequenced gene region without the need to impose a similarity threshold as required in the case of 97% OTUs (Callahan et al. 2017). The ASVs approach successfully identified Legionella species in both shallow lakes, including previously characterized species such as L. adelaidensis, L. tunisiensis and L. quateirensis, along with unclassified species like L. KR14 and L. D2863. Legionella adelaidensis was first characterized by Benson et al. (1991) from a sample taken from a cooling tower of an air conditioning system and is included in the list of species detected by the Mericon Quant Legionella spp. used in our work. Similarly, L. quateirensis was identified in a water sample from a shower in a hotel in Quarteira, Portugal, and is one of the five species of the Legionellaceae family described by Dennis et al. (1993). None of the aforementioned species has been found to be pathogenic for humans. Legionella tunisiensis was isolated via amoeba co-culturing in Sabkha Lake, a hypersaline lake in Tunisia, North Africa (Campocasso et al. 2012). This novel species is interesting due to its large amount of potential coding sequences and unusually high number of resistance genes (50% more sequences and 37 more genes than L. pneumophila). While not determined, a large number of encoded proteins could be related to mechanisms for overcoming osmotic stress (Schwake et al 2021, and cites therein).

Regarding Legionella KR14 and Legionella D2863, neither of them are described in the literature nor classified in databases (Schoch et al. 2020), leaving their characteristics and pathogenicity unknown to date. Thus, further investigation is needed to fully unravel the composition of Legionella species in the studied shallow lakes.

This study, the first of its kind conducted in Pampean surface aquatic systems, revealed a notably high overall Legionella amount in the studied shallow lakes. We found a distinct temporal pattern of Legionella abundances, with maximum values occurring in May, July and October in both lakes, used for recreational purposes. We also found that the Legionella dynamics were predominantly driven by water temperature and DOC. The molecular approaches (NGS and qPCR) employed to investigate the Legionella dynamics demonstrated congruence and consistency, albeit some variations were noted among the different sequence processing algorithms used for NGS data.

The potential for human exposure to Legionella in natural surface aquatic systems can manifest in many ways, ranging from direct contact through recreational water activities to more indirect exposure via aerosolization from water bodies. Consequently, we advocate for continued monitoring of Pampean shallow lakes, including culture-based studies, to better assess and mitigate the risk posed by the presence of Legionella in these natural surface waters.

Funding

This study was financed by the National Agency of Scientific and Technical Promotion (ANPCyT PICT 0891-2017, PICTO 00006-2019) and the National University of North-Western Buenos Aires (UNNOBA SIB 2053-2022), Argentina. We also thank QIAGEN for the 2018 Young Scientist Contest award won by Mara Inés Sagua and National Interuniversity Council (CIN) for providing Encouragement of Scientific Vocations (EVC) fellowships.

Competing Interests

The authors declare no relevant financial or non-financial interests to disclose.

Author Contributions

All authors contributed to the study conception and design. Sample acquisition was performed by María Pía Quiroga and María Romina Schiaffino. Laboratory analyses were executed by Julieta Bianchelli, Mara Sagua and Guillermina Nuozzi. Statistical analyses were performed by Julieta Bianchelli, Julia Fernandez and María Romina Schiaffino. The first draft of the manuscript was written by Julieta Bianchelli. Mara Sagua and Julieta Bianchelli contributed equally to this work. All authors commented on previous versions of the manuscript. All authors read and approved the final version of the manuscript.

Ethical approval Not applicable

Consent to participate All the authors of the article agree to participate in the journal submission

Consent for publication All the author listed have approved the manuscript that is enclosed

Competing interest The authors declare no competing interest

Data availability The data presented in this study are available upon request from the corresponding author.

AbuOdeh R, Aziz H, Moussa H, et al (2017) First study in Qatar to reveal high Legionella counts in cooling towers. E Mediterr Health J 23:703–707. https://doi.org/10.26719/2017.23.10.703

Bahk YY, Kim HS, Rhee O, You K, Bae KS, Lee W, Kim T, Lee S (2020). Long-Term Monitoring of Noxious Bacteria for Construction of Assurance Management System of Water Resources in Natural Status of the Republic of Korea. J Microbiol Biotechnol 30:1516–1524. 10.4014/jmb.2004.04064

Bennett S, Bentham R (2014) Effects of Seawater Concentration and Temperature on the Survival of Legionella pneumophila Serogroup 1. In: Legionella. ASM Press, Washington, DC, USA, pp 420–422. https://doi.org/10.1128/9781555815660.ch99

Benson RF, Thacker WL, Lanser JA, et al (1991) Legionella adelaidensis, a new species isolated from cooling tower water. J Clin Microbiol 29:1004–1006. https://doi.org/10.1128/jcm.29.5.1004-1006.1991

Caicedo C, Beutel S, Scheper T et al (2016) Occurrence of Legionella in wastewater treatment plants linked to wastewater characteristics. Environ Sci Pollut R 23:16873–16881. https://doi.org/10.1007/s11356-016-7090-6

Callahan BJ, McMurdie PJ, Holmes SP (2017) Exact sequence variants should replace operational taxonomic units in marker-gene data analysis. ISME J 11:2639–2643. https://doi.org/10.1038/ismej.2017.119

Callahan BJ, McMurdie PJ, Rosen MJ, et al (2016) DADA2: High-resolution sample inference from Illumina amplicon data. Nat Methods 13:581–583. https://doi.org/10.1038/nmeth.3869

Campocasso A, Boughalmi M, Fournous G, et al (2012) Legionella tunisiensis sp. nov. and Legionella massiliensis sp. nov., isolated from environmental water samples. Int J Syst Evol Microbiol 62:3003–3006. https://doi.org/10.1099/ijs.0.037853-0

Carvalho FR, Nastasi FR, Gamba RC, Foronda AS, Pellizari VH (2008) Occurrence and diversity of Legionellaceae in polar lakes of the Antarctic peninsula. Curr Microbiol 57:294-300. 10.1007/s00284-008-9192-y

Dennis PJ, Brenner DJ, Thacker WL, et al (1993) Five New Legionella Species Isolated from Water. Int J Syst Bacteriol 43:329–337. https://doi.org/10.1099/00207713-43-2-329

Diovisalvi N, Bohn VY, Piccolo MC, Perillo GM, Baigún C, Zagarese HE. (2015). Shallow lakes from the Central Plains of Argentina: an overview and worldwide comparative analysis of their basic limnological features. Hydrobiologia, 752:5-20. https://doi.org/10.1007/s10750-014-1946-x

Doménech-Sánchez A, Olea F, Berrocal CI (2008) Infecciones relacionadas con las aguas de recreo. Enferm Infecc Microbiol Clin 26:32–37. https://doi.org/10.1157/13128778

Dray S, Legendre P, Blanchet G (2007) packfor: Forward Selection with permutation (Canoco p.46)

Edgar RC (2013) UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat Methods 10:996–998. https://doi.org/10.1038/nmeth.2604

Edgar RC (2016) UNOISE2: improved error-correction for Illumina 16S and ITS amplicon sequencing. BioRxiv. https://doi.org/10.1101/081257

European Centre for Disease Prevention and Control. Legionnaires’ disease. In: ECDC. Annual Epidemiological Report for 2021. Stockholm: ECDC; 2023.

Fernández Zenoff V, Siñeriz F, Farías ME (2006) Diverse Responses to UV-B Radiation and Repair Mechanisms of Bacteria Isolated from High-Altitude Aquatic Environments. Appl Environ Microbiol 72:7857–7863. https://doi.org/10.1128/AEM.01333-06

Fliermans CB, Cherry WB, Orrison LH et al (1981) Ecological distribution of Legionella pneumophila. Appl Environ Microbiol 41:9–16. https://doi.org/10.1128/aem.41.1.9-16.1981

Gast RJ, Moran DM, Dennett MR et al (2011) Amoebae and Legionella pneumophila in saline environments. J Water Health 9:37–52. https://doi.org/10.2166/wh.2010.103

Grabowska-Grucza K, Kiersztyn B (2023) Relationships between Legionella and Aeromonas spp. and associated lake bacterial communities across seasonal changes in an anthropogenic eutrophication gradient. Sci Rep 13:17076. https://doi.org/10.1038/s41598-023-43234-3

Grabowska K, Bukowska A, Kaliński T, Kiersztyn B, et al (2020). Presence and identification of Legionella and Aeromonas spp. in the Great Masurian Lakes system in the context of eutrophication. J. Limnol, 79, 30-42. http://dx.doi.org/10.4081/jlimnol.2019.1924

Heller R, Höller C, Süßmuth R, Gundermann KO (1998) Effect of salt concentration and temperature on survival of Legionella pneumophila. Lett Appl Microbiol 26:64–68. https://doi.org/10.1046/j.1472-765X.1998.00273.x

Herlemann DPR, Labrenz M, Jürgens K, Bertilsson S, Waniek JJ, Andersson AF (2011) Transitions in bacterial communities along the 2000 km salinity gradient of the Baltic Sea. ISME J 5:1571–9. https://doi.org/10.1038/ismej.2011.41

Kanarek P, Bogiel T, Breza-Boruta B (2022) Legionellosis risk—an overview of Legionella spp. habitats in Europe. Environ Sci Pollut Res 29:76532–76542. https://doi.org/10.1007/s11356-022-22950-9

Kassambara A, Mundt F (2020) factoextra: Extract and Visualize the Results of Multivariate Data Analyses

Lê S, Josse J, Husson F (2008) FactoMineR: An R Package for Multivariate Analysis. J Stat Softw 25(1):1-18. https://doi.org/10.18637/jss.v025.i01

Lee JV, Lai S, Exner M et al (2011) An international trial of quantitative PCR for monitoring Legionella in artificial water systems. J Appl Microbiol 110:1032–1044. https://doi.org/10.1111/j.1365-2672.2011.04957.x

Leoni E, Catalani F, Marini S, et al (2018) Legionellosis associated with recreational Waters: a systematic review of cases and outbreaks in swimming pools, spa pools, and similar environments. Int J Env Res Pub He 15(8): 1612. https://doi.org/10.3390/ijerph15081612

Logares R (2018) Workflow for Analysing MiSeq Amplicons based on Uparse http://dx.doi.org/10.5281/zenodo.259579

Lösch LS, Merino LA (2016) Presencia de Legionella spp. en depósitos domiciliarios de agua potable en Resistencia, Chaco, Argentina. Informe preliminar. Rev Argent Microbiol 48:329–332. https://doi.org/10.1016/j.ram.2016.07.002

Murdoch DR (2003) Diagnosis of Legionella infection. Clin Infec Dis 36:64–69. https://doi.org/10.1086/345529

National Academies of Sciences Engineering, and Medicine (2020) Management of Legionella in Water Systems. The National Academies Press, Washington, DC. http://doi.org/10.17226/25474

Nuozzi G, Seoane Rocha C, Sagua M et al (2022) Estructura y dinámica del bacterioplancton en cuatro lagunas pampeanas de la cuenca del río Salado (Buenos Aires, Argentina). Ecología Austral 32:343–360. https://doi.org/10.25260/EA.22.32.2.0.1811

Ohno A, Kato N, Yamada K, Yamaguchi K (2003) Factors Influencing Survival of Legionella pneumophila Serotype 1 in Hot Spring Water and Tap Water. Appl Environ Microbiol 69:2540–2547. https://doi.org/10.1128/AEM.69.5.2540-2547.2003

Oksanen J, Blanchet G, Kidnt R et al (2015) vegan: Community Ecology Package. CRAN R package

Quiroga MP (2020) Diversidad de bacterias planctónicas en lagunas pampeanas utilizadas con fines recreativos. Dissertation, Universidad Nacional de La Plata

Quirós R, Drago E (1999) The environmental state of Argentinean lakes: an overview. Lakes & Reservoirs 4:55–64. https://doi.org/10.1046/j.1440-1770.1999.00076.x

Quirós R (2004) Sobre la morfología de las lagunas pampeanas. Documento de trabajo del área de Sistemas de Producción Acuática (3).

R Core Team. R: a language and environment for statistical computing (2018). Vienna: R Foundation for Statistical Computing, 2018. Retrieved April 1, 2023 from https://www.R-project.org.

R Core Team (2022). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/

Rowbotham TJ (1980) Preliminary report on the pathogenicity of Legionella pneumophila for freshwater and soil amoebae. J Clin Pathol 33:1179–1183. https://doi.org/10.1136/jcp.33.12.1179

Sagua MI, Nuozzi G, Sánchez ML, Huber P, Perdomo S, Schiaffino MR (2023) Unraveling the effect of land use on the bacterioplankton community composition from highly impacted shallow lakes at a regional scale. FEMS Microbiol Ecol 99(10): fiad109. https://doi.org/10.1093/femsec/fiad109

Sánchez ML, Schiaffino MR, Graziano M, et al (2021) Effect of land use on the phytoplankton community of Pampean shallow lakes of the Salado River basin (Buenos Aires Province, Argentina). Aquat Ecol 55:417–435. https://doi.org/10.1007/s10452-021-09835-8

Sarkar D (2008) Lattice: Multivariate data visualization with R. Springer New York, New York, NY. https://doi.org/10.1007/978-0-387-75969-2

Schiaffino MR, Diovisalvi N, Marfetán Molina D et al (2019) Microbial food-web components in two hypertrophic human-impacted Pampean shallow lakes: interactive effects of environmental, hydrological, and temporal drivers. Hydrobiologia 830:255–276. https://doi.org/10.1007/s10750-018-3874-7

Schiaffino MR, Huber P, Sagua M et al (2020) Covariation patterns of phytoplankton and bacterioplankton in hypertrophic shallow lakes. FEMS Microbiol Ecol 96(11). https://doi.org/10.1093/femsec/fiaa161

Schoch CL, Ciufo S, Domrachev M et al (2020) NCBI Taxonomy: a comprehensive update on curation, resources and tools. Database 2020: https://doi.org/10.1093/database/baaa062

Schwake DO, Alum A, Abbaszadegan M (2021) Legionella Occurrence beyond Cooling Towers and Premise Plumbing. Microorganisms 9:2543. https://doi.org/10.3390/microorganisms9122543

Selak L, Marković T, Pjevac P, Orlić S (2022). Microbial marker for seawater intrusion in a coastal Mediterranean shallow Lake, Lake Vrana, Croatia. Sci Tot Environ 849:157859. https://doi.org/10.1016/j.scitotenv.2022.157859

Seoane Rocha C (2018) Estructura y dinámica del bacterioplancton en la cuenca superior e inferior del río Salado. Dissertation, Universidad Nacional del Noroeste de la Provincia de Buenos Aires

Simpson GL (2019) permute: Functions for Generating Restricted Permutations of Data

Stout JE, Yu VL (1997) Legionellosis. New Engl J Med 337:682–687. https://doi.org/10.1056/NEJM199709043371006

Tison DL, Pope DH, Cherry WB, Fliermans CB (1980) Growth of Legionella pneumophila in association with blue-green algae (cyanobacteria). Appl Environ Microbiol 39:456–459. https://doi.org/10.1128/aem.39.2.456-459.1980

van Heijnsbergen E, Schalk JAC, Euser SM et al (2015) Confirmed and Potential Sources of Legionella Reviewed. Environ Sci Technol 49:4797–4815. https://doi.org/10.1021/acs.est.5b00142

Verissimo A, Marrao G, da Silva FG, Da Costa MS (1991) Distribution of Legionella spp. in hydrothermal areas in continental Portugal and the island of Sao Miguel, Azores. Appl Environ Microbiology 57(10):2921-2927. 10.1128/aem.57.10.2921-2927.1991

Walser SM, Gerstner DG, Brenner B, Höller C, Liebl B, Herr CEW (2014) Assessing the environmental health relevance of cooling towers – A systematic review of legionellosis out- breaks. Int J Hyg Environ Health 217(2–3):145–54. https://doi.org/10.1016/j.ijheh.2013.08.002

World Health Organization (2022) Legionellosis Fact Sheet. Geneva, Switzerland. https://www.who.int/news-room/fact-sheets/detail/legionellosis

World Health Organization (2022) Disease Outbreak News Legionellosis – Argentina. https://www.who.int/emergencies/disease-outbreak-news/item/2022-DON407

Wickham H, Averick M, Bryan J, Chang W, McGowan LD, François R, Grolemund G, Hayes A, Henry L, Hester J, Kuhn M, Pedersen TL, Miller E, Bache SM, Müller K, Ooms J, Robinson D, Seidel DP, Spinu V, Takahashi K, Vaughan D, Wilke C, Woo K, Yutani H (2019) Welcome to the tidyverse. J Open Source Softw 4(43): 1686. doi:10.21105/joss.01686

William HM, Heslin K, Kram JJ, Toberna CP, Baumgardner DJ (2022) Association of natural waterways and Legionella pneumophila infection in eastern Wisconsin: A case-control study. J Patient Cent Res Re 9:128–13. 10.17294/2330-0698.1872

Wullings BA, van der Kooij D (2006) Occurrence and genetic diversity of uncultured Legionella spp. in drinking water treated at temperatures below 15°C. Appl Environ Microbiol 72: 157–166. 10.1128/AEM.72.1.157-166.2006

Zeileis A, Hothorn T (2002). Diagnostic Checking in Regression Relationships. R News 2(3):7-10. https://CRAN.R-project.org/doc/Rnews/

Table 1 Ranges (Min and Max), means and standard deviation (SD) values of environmental variables measured in Carpincho and Gómez shallow lakes from January to December 2016 (N=12).

	Carpincho			Gómez
Variable	Min	Max	Mean±SD	Min	Max	Mean±SD
Water level (m)	1.35	1.52	1.40±0.05	1.35	1.55	1.43±0.06
Temperature (ºC)	8.70	26.30	16.93±6.27	8.7	26.5	17.32±6.37
pH	8.56	8.93	8.77±0.11	8.61	8.91	8.76±0.11
Dissolved oxygen (mg/L)	7.11	11.96	9.41±1.31	8.49	11.06	9.74±0.96
Total phosphorus (mg/L)	0.60	1.48	0.96±0.27	0.59	1.08	0.88±0.14
Soluble reactive phosphorus (mg/L)	0.20	1.18	0.59±0.28	0.26	1.10	0.63±0.24
Kjedahl’s total Nitrogen (mg/L)	2.85	4.89	3.86±0.62	2.84	7.39	4.35±1.35
Ammoniacal nitrogen (mg/L)	0	0.28	0.05±0.09	0	0.29	0.06±0.09
Salinity (g/L)	1.66	2.88	2.39±0.43	1.67	3.04	2.41±0.46
Dissolved organic carbon (mg/L)	22.77	33.08	27.05±3.23	22.16	32.46	26.44±2.65

Table 2 Spearman's correlation coefficients (r) for 97% similarity OTU (97% OTU), zOTU and ASV bioinformatic approaches versus Legionella abundances obtained by qPCR in both shallow lakes. ns = non-significant, * = p < 0.05, ** = p < 0.01, *** = p < 0.001. N=12

	Carpincho	Gómez
	r	r
97% OTU	0.68^*	0.58^*
zOTU	0.93^***	0.91^***
ASV	0.26^ns	0.63^*

SupplementarymaterialBianchellietal2024.pdf

Download PDF

Reviewers agreed at journal
11 Aug, 2024
Reviewers invited by journal
01 Aug, 2024
Editor invited by journal
01 Aug, 2024
First submitted to journal
31 Jul, 2024

You are reading this latest preprint version

Temporal dynamics of Legionella (Proteobacteria, Legionellaceae) in two Pampean shallow lakes from Argentina

Status:

Version 1

Abstract

Figures

Introduction

Materials and methods

Study area

Sampling and measurement of environmental variables

Molecular approaches

Real-time PCR (qPCR)

Statistical analyses

Results

Discussion

Comparative of NGS and qPCR abundances

Conclusions

Declarations

References

Tables

Supplementary Files

Status:

Version 1