https://orcid.org/0000-0003-3203-7046
Figures
Abstract
Given the recent global surge in Legionnaires’ disease cases, the monitoring of Legionella pneumophila becomes increasingly crucial. Epidemiological cases often stem from local outbreaks rather than widespread dissemination, emphasizing the need to study the characteristics of this pathogen at a local level. This study focuses on isolates of L. pneumophila in the Italian region of Friuli Venezia Giulia to assess specific genotype and phenotype distribution over time and space. To this end, a total of 127 L. pneumophila strains isolated between 2005 and 2017 within national surveillance programs were analysed. Rep-PCR, RAPD, and Sau-PCR were used for genotypic characterization, while phenotypic characterization was conducted through fatty acids analysis. RAPD and Sau-PCR effectively assessed genetic characteristics, identifying different profiles for the isolates and excluding the presence of clones. Although Sau-PCR is rarely used to analyse this pathogen, it emerged as the most discriminatory technique. Phenotypically, hierarchical cluster analysis categorized strains into three groups based on varying membrane fatty acid percentages. However, both phenotypic and genotypic analyses revealed a ubiquitous profile distribution at a regional level. These results suggest an absence of correlations between strain profiles, geographical location, and isolation time, indicating instead high variability and strain dissemination within this region.
Citation: Colautti A, Civilini M, Bortolomeazzi R, Franchi M, Felice A, De Martin S, et al. (2024) Genotypic and phenotypic profiling of 127 Legionella pneumophila strains: Insights into regional spread. PLoS ONE 19(7): e0307646. https://doi.org/10.1371/journal.pone.0307646
Editor: Marta Palusinska-Szysz, Maria Curie-Sklodowska University, POLAND
Received: April 15, 2024; Accepted: July 10, 2024; Published: July 19, 2024
Copyright: © 2024 Colautti et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: This work was supported by the University of Udine, Italy Biomolecular phenotypic characterization and incidence of Legionella pneumophila in Friuli Venezia Giulia (PRID2018-DI4A004) grant. MC received the grant MC was the funder. Other than funding acquisition, he contributed to Conceptualization, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Reviewing the manuscript.
Competing interests: The authors have declared that no competing interests exist.
1 Introduction
In recent years, there has been a steady increase in the incidence of legionellosis cases, a water-borne disease caused by exposure to contaminated water, with Legionella pneumophila (Lp) [1] being the main etiological agent, observed both in Europe [2] and the United States [3]. Various factors may correlate with this surge, including climatic fluctuations such as elevated temperatures and changes in precipitation patterns, which have been reported by several authors as impacting the proliferation of Lp in aquatic environments. These conditions increase the risk of human exposure and, consequently, a rise in legionellosis cases [4–9]. Furthermore, poor water management and changes in distribution systems can enhance Lp growth in water facilities, raising the risk of human exposure. This was exemplified during the COVID-19 pandemic, wherein the discontinuation of several water facilities’ utilization fostered water stagnation [10]. Additionally, the pandemic has increased the number of individuals potentially at risk, as recovered patients often have compromised respiratory systems. The aging population is an additional contributory factor to the increased susceptibility to this infection, as elderly or immunocompromised individuals are at a higher risk of developing severe forms of the disease [11]. Moreover, international travel and globalization can facilitate the spread of local outbreaks of legionellosis through the movement of infected individuals, with numerous cases reported among travellers who have stayed in accommodation facilities such as hotels or cruise ships [12]. It should also be considered that statistics on the incidence of legionellosis may be influenced by advancements in diagnostic techniques and increased awareness among medical personnel, leading to a more accurate detection of legionellosis cases. This may seemingly indicate an increase in cases when, in reality, the disease is simply being diagnosed and identified more effectively [13].
Regarding its dissemination, it is essential to note that legionellosis is typically associated with localized outbreaks [14] rather than a generalized increase on a global scale [3]. Episodes of legionellosis are frequently linked to specific sources of exposure in particular locations or structures, as environmental conditions and risk factors can vary significantly among different places [15]. Therefore, local monitoring enables the improvement of prevention and control strategies tailored to the specific conditions of each area, facilitating timely outbreak detection. The swift identification and response to local outbreaks contribute to safeguarding public health, reducing the risk of disease spread, and limiting the number of cases. For these reasons, it is also important to characterize the isolates at both genotypic and phenotypic level, to identify their characteristics and more effectively counteract their presence at both environmental and medical level in the case of infected patients.
In this regard, besides Pulsed Field Gel Electrophoresis (PFGE), other techniques such as RAPD, rep-PCR, and Sau-PCR have proven effective in Lp genotypic characterization. These methodologies, previously employed in other contexts to differentiate between various Legionella spp., were deemed advantageous for their rapidity and cost-effectiveness [16]. In addition to species differentiation, some studies have highlighted their effectiveness in discriminating among similar strains of Lp, even distinguishing their serotype [17–19]. Regarding phenotypic characterization, one of the most effective methods reported is the typing of membrane fatty acids. Initial studies on the cellular fatty acid profile of Lp were conducted on pathogenic strains isolated from patients involved in the 1976 pneumonia outbreak [20]. The primary phospholipid constituting the membrane of Lp is phosphatidylcholine (PC), a phospholipid typical of the cellular membranes of eukaryotic organisms. The presence of PC in the bacterial membrane can influence the interaction with host cells, as this phospholipid is commonly associated with the cellular membranes of host organisms. Furthermore, Lp is recognized for its ability to modify the lipidic composition of its membrane in response to the surrounding environment, including the biological fluids of the host [21]. These alterations can modulate the interaction with host cells. Collectively, the unique lipid composition of Lp membrane influences how the bacterium interacts with host cells. This interaction is crucial for Lp ability to adapt to the hostile host environment and evade or circumvent the host immune response, thereby contributing to its capacity to cause Legionnaires’ disease [22]. Furthermore, the phenotypic plasticity of the Lp membrane, modulated by fatty acids, and its ability to utilize exogenous choline allow it to adapt to environmental variations, including the intracellular environment of amoebae in which it thrives. Additionally, the atypical structure of lipid A contributes to its immunomodulatory properties. The lipid A structure of Lp often features a long and complex side chain, which can vary depending on the serotype. This atypical structure is considered one of the reasons why Legionella spp. can partially evade the host immune response, modulate the inflammatory response, and protect the bacterium from esterases within amoebae [23].
In this context, this study aimed to deepen the understanding of the spread of Lp in the Italian region of Friuli Venezia Giulia in northeastern Italy. To this end, the study investigated the presence of correlations between the spatio-temporal distribution of various genotypic and phenotypic profiles of 127 L. pneumophila strains. These strains were isolated from tap water in different types of buildings identified as sources of legionellosis outbreaks between 2005 and 2017. The investigation covered various municipalities and was part of the national surveillance plans conducted by the Regional Environmental Protection Agency of Friuli Venezia Giulia (ARPA FVG). For the analysis, two approaches were used, focusing on conventional and well-established techniques for studying this pathogen. The first characterization method concentrated on the genetic aspect, assessing the differences between the strains using RAPD-PCR, rep-PCR, and Sau-PCR techniques. The second approach involved the phenotypic typing of membrane fatty acids in different strains. The results obtained from these methods provided an initial characterization and overview of the differentiation of L. pneumophila strains isolated in this region, thus laying the foundation for more in-depth studies using Whole Genome Sequencing (WGS) techniques.
2 Materials and methods
2.1 Regional strains isolation
The strains examined in this study were collected as part of national surveillance plans in the Friuli Venezia Giulia, a northeastern region of Italy, between 2005 and 2017. This analysis, conducted by the Regional Agency for Environmental Protection of Friuli Venezia Giulia (ARPA FVG), aimed at identifying the sources of legionellosis outbreaks by sampling tap water systems of buildings presumed to be sources of infection. Samples were obtained following the ISO 11731:2017 protocol. Briefly, 1-Liter water samples were membrane filtered (Pall Corporation, USA). The filtering membrane was re-suspended in a sterile tube with 10 mL of water from the same sample to resuspend the microorganisms. Then, 0.1 mL of the final sample was inoculated on Legionella Agar (Biolife, Italy) growth media. Colonies considered positive underwent presumptive identification through sub-culturing each colony on both Legionella Agar plates (Biolife, Italy) and Legionella Agar without Cysteine plates (Biolife, Italy). Inoculated plates were incubated in a jar under microaerophilic conditions using CampyGen™ 2.5 L in a thermostat at 36°C for up to 10 days. Colonies grown exclusively on the former were considered as Legionella spp.
Species and serogroup identification were conducted based on antigenic reactions using latex agglutination serological tests with monoclonal antibodies (Legionella rapid latex test kit, Mascia Brunelli S.p.a., Italy) and real-time PCR following Annex 6 of the Italian national Guidelines [24]. If the definitive result was positive, L. pneumophila isolates were preserved at -80°C in Cryobanks (Mast House, UK) until further analyses were performed.
2.2 Reference strains
For the comparison of the tested strains, four strains of different Legionella pneumophila serogroups and other strains of Legionella spp. were employed. The following strains were purchased in lyophilized form from the Leibniz-Institut DSMZ-Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (DSMZ) collection: Legionella pneumophila DSM 7513 serogroup 1 (STD1), Legionella pneumophila DSM 25071 serogroup 2 (STD2), Legionella pneumophila DSM 25182 serogroup 6 (STD6), Legionella pneumophila DSM 25184 serogroup 8 (STD8), Legionella longbeachae DSM 10572, Legionella bozemanii DSM 16523, Legionella micdadei DSM 16640, Legionella dumoffii DSM 17625, and Legionella gormanii DSM 25296.
2.3 DNA extraction
For all strains, DNA extraction was performed using the modified Querol method [25]. Cryopreserved colonies, cultured for 5 days on Legionella Agar, were collected using a sterile 10 μL loop and transferred to a sterile tube containing 500 μL of Solution A (50 mg/mL Lysozyme, 1 M sorbitol, 0.1 M EDTA, in MilliQ water, final pH = 7.4) and incubated at 37°C for 2 h. Subsequently, the tubes were centrifuged at 8000 × g for 10 minutes, the supernatant was discarded, and the obtained pellet was resuspended in 500 μL of Solution B (50 mM Tris-HCl, 20 mM EDTA, in MilliQ water, final pH = 7.4). Samples were then incubated in a water bath at 65°C for 30 minutes, with the addition of 50 μL of 10% SDS Solution. After the incubation period, 200 μL of 5% KAc solution were added, and the samples were kept on ice for 30 minutes. Subsequently, the tubes were centrifuged at 14000 × g for 5 minutes. The supernatant, containing DNA, was transferred to new sterile tubes and 1 mL of absolute ethanol (Carlo Erba, Milan, Italy) was added. The tubes were then centrifuged at 14000 × g for 10 minutes, the supernatant was discarded, and 500 μL of cold 70% ethanol were further added. The tubes were centrifuged again at 14000 × g for 5 minutes. After removing the supernatant ethanol, the DNA pellet was air-dried overnight at 37°C. The pellet was then resuspended in 50 μL of sterile MilliQ water, and 1 μL of RNase (Sigma-Aldrich, USA) was added for a 1-hour incubation at 37°C to digest co-extracted RNA. DNA quantification was performed using Nanodrop One (Thermo Scientific, USA) and standardized to 50 ng/μL for the following analyses.
2.4 Rep-PCR, RAPD, and Sau-PCR analyses
The molecular characterization was performed according to Iacumin et al. 2020 [26]. Rep-PCR analysis was performed using (GTG)5 primer (5’-GTGGTGGTGGTGGTG-3’) [27], using the following reaction mix: 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2, 0.2 mM each dNTP, 1 μM primer (GTG)5, 1.25 U Taq-polymerase (Applied Biosystems, USA), and 100 ng of DNA, for a total volume of 25 μL. The reactions were carried out using a Euroclone Thermal Cycler (Celbio, Milan, Italy) and the amplification protocol consisted of 31 cycles of denaturation at 94°C for 3 s followed by one step at 92°C for 30 s, annealing at 40°C for 1 min and extension at 65° C for 8 min. The initial denaturation was at 95° C for 2 min and the final extension at 65° C for 8 min.
RAPD analysis was performed using the M13 primer (5’-GAG GGT GGC GGT TCT-3’) [28], using the following reaction mix: 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2, 0.2 mM each dNTP, 1 μM primer M13, 1.25 U Taq-polymerase (Applied Biosystems, USA) and 100 ng of the extracted DNA for a total volume of 25 μL. The reactions were carried out in a Euroclone Thermal Cycler (Celbio, Italy) and the amplification protocol consisted of 35 cycles of 94° C for 1 min, 38° C for 1 min, ramp to 72° C at 0.6° C/s, and 72° C for 2 min. At the beginning of the reaction, an initial denaturation step at 94° C for 5 min was performed, followed by a final extension at 72° C for 5 min.
Sau-PCR analysis was performed by digesting 200 ng of DNA overnight at 37°C with 1 μL of Sau3AI restriction endonuclease (10 U/μL) in a final volume of 20 μL. After enzymatic restriction, the amplification was performed using the SAG1 primer (5’-CCGCCGCGATCAG-3’) [29] with the following reaction mix: 10 mM Tris–HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2, 0.2 mM each dNTPs, 2 μM primer SAG1, 1.25 U Taq-polymerase (Applied Biosystems, USA), and 1 μL of the digested DNA for a total volume of 50 μL. The reactions were carried out using a Euroclone Thermal Cycler (Celbio, Italy) with the following protocol: 25° C for 5 min, ramp to 60°C at 0.1°C/s, 60°C for 30 s, 2 cycles of 95°C for 1 min, 50°C for 15 s, ramp to 25°C at 0.1°C/s, ramp to 50°C at 0.1°C/s, 50°C for 30 s, 35 cycles of 94°C for 15 s, 46°C for 1 min, 65°C for 2 min, and a final extension at 65°C for 2 min.
PCR products obtained from these techniques were separated in a 1.5% (w/v) agarose gel in TBE 0.5X at 120 V, for 4, 6, and 3 h for RAPD, rep-PCR, and Sau-PCR, respectively. Staining was performed for 30 min at the end of the electrophoretic run in TBE 0.5X buffer containing ethidium bromide 0.25 μL/mL (Sigma- Aldrich, St. Louis, USA). Digital images of the gels were acquired using the BioImaging System GeneGenius imaging software (Syngene, Italy).
2.5 Cellular fatty acid analysis
After revitalization and further streaking of the cryopreserved strains on Legionella Agar (36°C for 5 days), strains were cultured on Legionella Agar for 96 h at 36°C for the analysis of cellular fatty acids. As reported in the literature, the fatty acid composition of L. pneumophila is influenced by both incubation time [30, 31] and growth phase [32]. To ensure consistency and avoid growth-induced variations, all strains were carefully cultivated under identical conditions and harvested after 96 hours of growth, in accordance with previous studies that investigated fatty acid composition in Legionella spp. [31, 32]. For the analysis, aliquots of about 80 mg (wet weight) of cells were removed from the surfaces of the agar plates and placed in test tubes equipped with Teflon-lined screw caps. Fatty acid methyl esters (FAMEs) were prepared by saponification, methylation, and extraction, following the method reported by Sasser (1990) [33]. After the final washing step, the extract containing the FAMEs was dried with anhydrous sodium sulphate, transferred to an autosampler vial and the volume reduced to about 0.5 mL under a nitrogen stream.
2.6 Gas chromatography-flame ionization detection (GC-FID)
A Trace 1300 gas chromatograph, equipped with a flame ionization detector, a split/splitless injector, and an autosampler AL 1310 (Thermo Fisher Scientific, Italiy), was used. The separation was performed using a fused silica capillary column HP-5MS UI, 30 m x 0.25 mm, 0.25 μm film thickness (Agilent Technologies, Italy). The initial oven temperature was set at 140°C and increased to 300°C at 5°C/min. The injector and detector temperatures were 280 and 300°C, respectively. Helium was used as the carrier gas at a flow rate of 1.0 mL/min. The injection volume was 1 μL, and the split ratio was 1:10.
2.7 Gas chromatography-mass spectrometry (GC–MS)
A Agilent Technologies 7890B gas chromatograph, coupled to a quadrupolar mass detector (Agilent Technologies, Italy), was used. The separation was performed using a fused silica capillary column HP-5MS UI, 30 m x 0.25 mm, 0.25 μm film thickness (Agilent Technologies Italia, Italy). The initial oven temperature was set at 140°C and increased to 280°C at 5°C/min. The injector, transfer line, source, and quadrupole temperatures were 250, 280, 175, and 150°C, respectively. Helium was used as carrier gas at a flow rate of 1.0 mL/min. The injection volume was 1 μL, and the split ratio was 1:10. The mass spectra were recorded under electron impact (EI) at 70 eV.
2.8 Identification of fatty acids
FAMEs identification from GC-FID chromatograms was achieved by comparing retention times and mass spectra with those of authentic standards Bacterial Acid Methyl Ester (BAME) Mix (Supelco, Italy). For the compounds not present in the BAME mix, tentative identification was obtained by comparing the linear retention index, calculated by injecting a mixture of n-alkanes C7-C30 (Supelco, Sigma-Aldrich, Milan, Italy) under the same conditions, with literature data and the mass spectra with the mass spectral library NIST 14 of the instrument. Quantitative analysis was carried out by the percentage peak area method, considering all the compounds to have the same response to the flame ionization detector.
2.9 Statistical analysis
The fingerprint analysis of RAPD, rep-PCR, and Sau-PCR was carried out with the Gel Compare II software Version 4.1 (Applied Maths, Sint-Martens-Latem, Belgium), calculating their similarity using Sørensen–Dice correlation coefficient. Dendrograms were obtained using the Unweighted Pair Group Method with Arithmetic Average (UPGMA) clustering algorithm [34]. All further statistical analyses were conducted using R v4.1.2 software. For the creation and plotting of the different images, the ape [35], phangorn [36], plotly [37], dendextend [38], and ggplot2 [39] libraries were used.
3 Results
3.1 Dataset description
The dataset comprised a total of 127 strains whose isolation time, location, and serogroup were detailed in S1 Table. As shown in Fig 1A, these strains were isolated across different years, spanning from 2005 to 2017, distributed throughout all the months of the year, with maximum values (22 strains) in August, and minimum values (1 strain) in the month of November. Concerning geographical distribution, as depicted in Fig 1B, the strains were classified into four macro zones within the Friuli Venezia Giulia region based on climatic conditions. The northern area, sparsely populated, is characterized by mountainous terrain and a cooler climate. The plain area, where the majority of the population resides, can be further divided into western and eastern areas, separated by the Tagliamento river, the primary watercourse of the region. Lastly, a southern zone, distinguished by a milder climate and lower precipitation near the coast. Fig 1C illustrates the various municipalities where the different strains were isolated; notably, out of the 127 total strains, 48 strains were isolated in the western zone, 35 in the eastern zone, and 44 in the southern zone.
(A) Number of strains analysed for each study year and month; (B) Climatic subdivision of the Friuli Venezia Giulia region; (C) Municipalities included in each subzone, and the number of L. pneumophila strains isolated in them.
3.2 Rep-PCR, RAPD, and Sau-PCR profiles analysis
From the genotypic characterization, only Sau-PCR and RAPD techniques resulted effective in providing genetic fingerprints suitable for strain differentiation. Conversely, the primer used for the rep-PCR technique, in many instances, failed to yield genetic characterization profiles deemed adequate.
Therefore, considering the Sau-PCR and RAPD profiles of the standard strains, significant differences in similarities among the various species analysed emerged in relation to the two different techniques (Fig 2). It was possible to observe that, using both methodologies, L. micdadei resulted the most dissimilar strain compared to others, followed by L. dumoffii and L. longbeachae. Conversely, considering the results obtained from Sau-PCR, it was observed that the serotype reference strains of Lp did not strictly cluster together. In fact, strains STD1 and STD2 were separated from strain STD8, which showed a higher similarity (44.45%) with L. gormanii. In the case of strain STD6, a similarity of 60.01% was observed with the L. bozemanii reference strain. In contrast, from the results obtained from RAPD technique, it was noticeable that all reference strains from different serogroups of Lp separated from other Legionella spp. with a similarity threshold above 43.83%. Within this cluster, strains STD1 and STD2 showed an identical profile, while differentiating from strains STD8 and STD6 by a dissimilarity percentage above 58.58%. In turn, these latter two strains differentiated at a similarity percentage of 66.67%. Analysing the results of regional query strains in comparison with the reference strains of the four different serogroups of L. pneumophila, the Sau-PCR demonstrated a higher discriminatory capacity than the RAPD technique. In the obtained dendrograms (Fig 3), by applying a dissimilarity cutoff level of 35%, chosen to be more restrictive than the discrimination threshold observed previously among different species, and capable of segregating the various standard strains of different serotypes of L. pneumophila, Sau-PCR technique distinguished strains into 47 branches with 32 clusters and 15 single strains, whereas the RAPD technique yielded only 20 branches with 17 clusters and 3 single strains. Notably, the Sau-PCR technique identified several strains as clones, forming clusters with higher numerosity. Given these differences in clusters composition, and the fact that in few cases the same strains were clustered together by both techniques, it can be noted how the two techniques assessed the presence of distinct genetic traits and clustered the different strains differently.
Sau-PCR (left) and RAPD (right) dendrograms, represented as Dice’s dissimilarity matrix, of Legionella spp. reference strains.
Dice’s dissimilarity matrix calculated for Sau-PCR (left) and RAPD (right) profiles.
3.3 Fatty acids composition
FAMEs identification from GC-FID chromatograms (Fig 4A) achieved by comparing retention times and mass spectra with those of BAME standards Mix or tentative identified (Fig 4B), identified 21 profiles.
(A) Example of GC-FID chromatogram of fatty acid methyl esters from L. pneumophila strain SG1; (B) FAMEs identification from GC-FID chromatograms achieved by comparing retention times and mass spectra with those of BAME Mix, and tentative identification obtained by comparing the linear retention index calculated by injecting a mixture of n-alkanes C7-C30, with literature data and the mass spectra with the mass spectral library NIST 14.
From the subsequent analyses, comparing the fatty acid compositions of reference strains, it was possible to observe the differences between the four L. pneumophila reference strains and those of others Legionella spp. strains (Table 1).
The mean and standard deviation values are reported, followed by letters indicating significantly different groups, evaluated using the Tukey test (p < 0.01).
The prevalence of α-15:0 fatty acid was notable in L. bozemanii (27.4 ± 1.6%), L. micdadei (36.8 ± 0.44%), L. gormanii (20.36 ± 0.38%), and L. dumoffii (30.61 ± 2.26%). Conversely, in accordance with prior observations, L. longbeachae exhibited a higher percentage of the unsaturated fatty acid 16:1cisΔ9 (26.62 ± 0.96%) [40, 41].
L. micdadei was characterized by a higher percentage of branched α-17:0 fatty acid (22.63 ± 0.68%) and iso-17:1 fatty acid (3.25 ± 0.02%), while displaying a lower percentage of 16:0 fatty acid (5.94 ± 0.16%).
Notably, L. pneumophila strains were characterized by significantly higher presence of iso-16:0 fatty acid in comparison to the others considered Legionella spp. strains. A better visualization of the correlation between the strains and the fatty acids percentages could be observed in the PCA analysis reported in Fig 5A. Furthermore, the hierarchical cluster analysis reported in Fig 5B provided a detailed view, revealing that the four different serotypes of L. pneumophila exhibited a similarity height >95% among themselves, differing by more than 6.5% from L. longbeachae that resulted the most similar strain. The other species showed greater disparities, with L. bozemanii and L. gormanii sharing a similarity height of approximately 97% but differing by just under 9% from L. micdadei and L. dumoffii, which, in turn, shared a similarity height of about 6.5%.
(A) PCA analysis of the strains based on their fatty acid composition; (B) Hierarchical Cluster Analysis of the reference strains measured using Ward’s minimum variance method. The height of the fusion provided on the horizontal axis indicates the dissimilarity between two strains.
Examining in detail the regional L. pneumophila strains, a fatty acid profile comparable to that of the previously described L. pneumophila reference strains was observed (S2 Table). Indeed, a composition characterized by fatty acids with both even and odd carbon chain lengths ranging from 14 to 21 carbon atoms was detected (Fig 6A). The quantitatively most abundant fatty acids were branched, with i-16:0 alone constituting 30.70 ± 4.68%, followed by C16:1 cisΔ9 (15.11 ± 3.61%), a-C15:0 (10.94 ± 1.98%), a-17:0 (4.83 ± 0.95%), together comprising over 60% of the identified fatty acids for each strain. These fatty acids are in fact considered the primary markers for a preliminary differentiation of different Legionella species. In addition to 16:1 cisΔ9 acid, considering the 16-carbon chain fatty acids, 16:0 acid was present at a significant percentage (10.67 ± 3.96%). Also, the presence of 17:0-9cp acid (4.52 ± 2.00%), and 3-OH-i-14:0 hydroxy acid (0.77 ± 0.38%) were identified. Analysing the differences in fatty acid composition in relation to the serotype, no clear differences were observed among standard strains of different serogroups, both among themselves and in relation to the average values of the 127 analysed Lp strains. In particular, observing the PCA plot (Fig 6B), a high similarity was evident among STD1, STD6, and STD8 strains, while the strain STD2 resulted the only strain that clearly differentiated from the others, mainly due to a higher percentage of the fatty acid C16:0.
(A) Mean relative percentage of fatty acids composition the query strains; (B) PCA analysis of the strains based on their fatty acid composition.
From the hierarchical cluster analysis, it was possible to observe the relationships between different strains based on fatty acids (Fig 7). As also detected through the Gap Statistic Method [42], it was possible to identify three main clusters (distinguished with the same colours also in the PCA plot of Fig 6B). At a dissimilarity height exceeding 35%, 40 strains separated (cluster 3, represented in green), which were the most dissimilar from the others, characterized by higher percentages of i-C14:0, C14:1 cisΔ9, a-C15:0, C15:1 cisΔ9, 3-OH-i-C14:0, and i-C16:1 fatty acids. More similar were the other two clusters (cluster 1, highlighted by blue colour, grouping 56 strains, and cluster 2, highlighted by red colour, collecting 31 strains), which separate from each other at a dissimilarity height of about 20%. Among these, all standard strains of different serogroups grouped in cluster 1, while cluster 2 was found to be the most dissimilar from cluster 3 and composed of strains characterized by a higher presence of C16:0, a-C17:0, C17:0, i-C18:0, C18:0, C19:0, and C20:0.
The height of the fusion provided on the horizontal axis indicates the dissimilarity between two strains.
3.4 Obtained cluster analysis
Analysing the different clusters obtained from various techniques (whose branch code number for each of the techniques applied is reported in S1 Table) it was possible to observe how they were subdivided based on year, geographic area, and serogroup. Beginning with the results obtained from the Sau-PCR technique (Fig 8A), it demonstrated to be the technique with the highest discriminatory power, resulting in the highest number of branches obtained. Excluding reference strains, considering only the regional strains were divided into 44 branches, of which 27.3% consisted of single strains, 40.9% of two strains, and only the remaining 31.8% of more than three. Due to the high level of discrimination, it was therefore possible to observe a high correspondence of strains within each branch regarding the serogroup and a good correspondence regarding the year of isolation, while the correlation to the isolation area resulted weaker. Of particular interest in this case were the larger clusters. Particularly, it was observed that in branches No. 1, No. 13, No. 17, No. 18, No. 19, and No. 45, strains from different years, different areas, and even discordant serogroups grouped together. Conversely, branch No. 8 presented strains from the same area and serogroup but from two distinct years, suggesting persistence of this strain over time. Similarly, branch No. 26 was clustered strains of the same serogroup isolated in the same year but from different areas. Lastly, branch No. 45 presented strains isolated in different years, from different areas, but all belonging to serogroup I.
Composition of the clusters obtained from Sau-PCR (A), RAPD (B) and Fatty Acids (C) analyses in relation to the year, serogroup, and geographical area of isolation. In each cell the number of strains (Strains N°) of each cluster is reported as a percentage value. DSMZ reference strains are not considered.
Turning to the results obtained from RAPD analysis, the regional strains were divided into 19 branches. In this case, no clear correlations were found between the branches and the investigated variables. Similar considerations can be made for the clusters obtained from fatty acids analysis. Indeed, it was not possible to observe a clear correlation between the strains and year and place of isolation, while concerning the serogroup, only cluster No. 3 showed a predominance of strains belonging to serogroup 1. It should also be noted that the absence of correlation cannot be attributed to the cluster selection threshold, as observable in Fig 7, this lack of relationships is maintained even at much higher cutoff levels.
4 Discussion
4.1 Characteristics of the strain collection
The strain collection preserved by ARPA FVG considered in this work consisted of 127 strains isolated over a period of 12 years throughout the regional territory of Friuli Venezia Giulia in northeastern Italy. The strains were evenly distributed among the different zones into which the region could be divided, except for the northern zone, where both the colder climate and lower population density contribute to reducing the incidence of Lp presence in water systems [43]. Furthermore, the number of strains in the dataset followed a temporal distribution observed in larger datasets from this region, with periods of maximum risk of Lp presence at the end of summer (August–November) and a decrease in winter and spring months [44]. Therefore, given its size and uniform spatio-temporal distribution, the dataset can be considered adequate for representing the spread of this pathogen in the region.
4.2 Genetic insights
The initial analyses conducted focused on verifying genetic patterns. For this purpose, rep-PCR, RAPD, and Sau-PCR techniques were used, but only the latter two provided useful profiles for strain comparison. Despite rep-PCR being reported in the literature as effective in discriminating Legionella spp. [45], the primers and conditions used in this study did not yield comparable profiles. On the contrary, RAPD and Sau-PCR proved effective in providing discriminative profiles of the different strains. These techniques are reported to be more suitable for differentiating Lp strains compared to other techniques such as pulsed field gel electrophoresis [16]. By applying a cutoff level of 35% on the dendrograms obtained from both techniques, a much higher discrimination was observed with Sau-PCR compared to the RAPD technique. Despite the high discriminatory capacity observed, very few studies have employed this technique to analyse this pathogen [16]. When comparing how the strains were divided and clustered from these two analytical techniques, there was poor congruence, and in most cases, the strains showed different relationships with each other. However, these techniques target different genetic characteristics and have different levels of stringency. Such discrepancies in clustering have also been observed in other similar studies [46], demonstrating the relevance of comparing different techniques for this type of study. Additionally, no clear clustering pattern was observed regarding serogroup, which could have been influenced by the primers used [47]. However, the primary purpose of this work was not to identify primers suitable for species identification or serogroup division but to identify the presence of potential clones in different locations and periods and to verify the presence of correlation with origins. In this regard, it was possible to observe a ubiquitous geographical distribution of the different clusters, suggesting spread throughout the region rather than specific locations of certain clusters. These observations align with several studies, where a very high genetic variability of Lp strains has been observed [48, 49], with genetically different strains found in the same area [50]. Moreover, as observed in other studies, no correlation was found between geographical origin and genetic characteristics [51].
4.3 Phenotypic insights
These genotypic observations were also supported by phenotypic analysis of fatty acids, where no clear correlations were identified based on geographic area and year of isolation. In this case, as observable in the PCA, regional Lp strains distinguished themselves from other analysed species by a higher association with fatty acids i-C15:0, i-C16:1, 3-OH-i-C14:0, C14:1 cis-9, iC16:0, C14:0, C20:0, i-C14:0, C18:0, C16:1 cis-9, C15:1 cis9. Moreover, these strains showed a fatty acid profile comparable to that previously described for L. pneumophila by earlier authors [31, 32, 52], with a high presence of i-C16:0 [41, 52, 53]. As reported in the literature [52], 17:0-9cp acid, containing a cyclopropane ring in the carbon chain, and 3-OH-i-14:0 hydroxy acid were also identified. This fatty acid, also highlighted by other studies on Lp strains analysed through basic hydrolysis [41], is the only hydroxylated fatty acid identified with the sample preparation method used in this study. Despite its low presence, it proves important in the study of Legionella spp. Hydroxylated fatty acids at positions 2 and 3, both linear and branched, have been primarily detected in studies on the lipopolysaccharides of the cell membrane of this bacterial species. These fatty acids are specifically present in the hydrophobic portion of the lipopolysaccharide, known as Lipid A, to which they are linked by an amide bond and have often been found as distinctive features unique to Lp strains [53–55]. However, it should be noted that different sample preparation methods using basic hydrolysis (as in this study) or acidic hydrolysis release different types of fatty acids. In a study conducted on Legionella lytica, hydroxylated fatty acids of lipopolysaccharides, characterized by a strong amide bond, were determined only after hydrolysis with 4 M HCl at 100°C for six hours, while they were generally not detected by basic hydrolysis [56]. Additionally, other studies report that acidic hydrolysis degrades cyclopropane acids, forming artifacts that may interfere with gas chromatographic analysis [57]. Instead, the regional Lp strains considered in this work could be divided into three main clusters according to their fatty acid composition. The first cluster (cluster 1), which included reference strains of different serogroups, was characterized by strains with intermediate relative percentages of all FAs. Conversely, a second cluster (cluster 2) was characterized by a higher presence of C16:0, C17:0, C18:0, C20:0, C19:0, a-C17:0, i-C18:0, and a third cluster (cluster 3) was characterized by a higher content of C14:1 cis-9, C15:1 cis-9, 3-OH-i-C14:0, a-C15:0, i-C14:0, and i-C16:1.
5 Conclusions
This study has revealed a high genetic differentiation of Lp strains isolated in Friuli Venezia Giulia, showing that the various genotypic and phenotypic traits are not associated with specific locations and times of isolation, but rather there is a widespread distribution across the entire territory and over time. Indeed, in addition to not observing a clear correlation of the strains with the macro areas identified, it was possible to observe a high differentiation within the same municipality, with strains with the same genotypic/phenotypic characteristics present in municipalities very distant from each other. Furthermore, the conducted analysis highlighted the effectiveness of the Sau-PCR technique in characterizing Legionella spp., a technique still scarcely employed for analysing this important pathogen.
Supporting information
S1 Table. Detailed description of the dataset metadata.
https://doi.org/10.1371/journal.pone.0307646.s001
(DOCX)
S2 Table. Fatty acids composition of the different isolated Lp strains.
https://doi.org/10.1371/journal.pone.0307646.s002
(DOCX)
Acknowledgments
The authors would like to thank Arianna Ortali and Veronica Ballus for their technical support in the molecular characterization of the strains and Carlo Liotti and Gabriele Marano for their technical support for the fatty acid characterization.
References
- 1. Cianciotto NP. Pathogenicity of Legionella pneumophila. Int J Med Microbiol. 2001;291: 331–343. pmid:11727817
- 2.
European Centre for Disease Prevention and Control. Legionnaires’ disease. In: ECDC. Annual Epidemiological Report for 2021. Stockholm: ECDC; 2023. Eur Cent Dis Prev Control Legion Dis ECDC Annu Epidemiol Rep 2021 Stock ECDC; 2023. 2023.
- 3. Moffa MA, Rock C, Galiatsatos P, Gamage SD, Schwab KJ, Exum NG. Legionellosis on the rise: A scoping review of sporadic, community-acquired incidence in the United States. Epidemiol Infect. 2023;151. pmid:37503568
- 4. Walker JT. The influence of climate change on waterborne disease and Legionella: a review. Perspect Public Health. 2018;138: 282–286. pmid:30156484
- 5. Han XY. Effects of climate changes and road exposure on the rapidly rising legionellosis incidence rates in the United States. PLoS ONE. 2021. pmid:33886659
- 6. Ana Maria de Roda Husman MAB. Dimensions of Effects of Climate Change on Water-Transmitted Infectious Diseases. Air Water Borne Dis. 2013;02: 1–8.
- 7. Sakamoto R. Legionnaire’s disease, weather and climate. Bull World Health Organ. 2015;93: 435–436. pmid:26240466
- 8. Heilmann A, Rueda Z, Alexander D, Laupland KB, Keynan Y. Impact of climate change on amoeba and the bacteria they host. J Assoc Med Microbiol Infect Dis Canada. 2024;9: 1–5. pmid:38567368
- 9. Yu AT, Kamali A, Vugia DJ. Legionella Epidemiologic and Environmental Risks. Curr Epidemiol Reports. 2019;6: 310–320.
- 10. Kunz JM, Hannapel E, Vander Kelen P, Hils J, Hoover ER, Edens C. Effects of the COVID-19 Pandemic on Legionella Water Management Program Performance across a United States Lodging Organization. Int J Environ Res Public Health. 2023;20. pmid:37835155
- 11. Bauer M, Mathieu L, Deloge- Abarkan M, Remen T, Tossa P, Hartemann P, et al. Legionella bacteria in shower aerosols increase the risk of pontiac fever among older people in retirement homes. J Epidemiol Community Health. 2008;62: 913–920. pmid:18791050
- 12. Guyard C, Low DE. Legionella infections and travel associated legionellosis. Travel Med Infect Dis. 2011;9: 176–186. pmid:21995862
- 13. Schoonmaker-Bopp D, Nazarian E, Dziewulski D, Clement E, Baker DJ, Dickinson MC, et al. Improvements to the success of outbreak investigations of legionnaires’ disease: 40 years of testing and investigation in new york state. Appl Environ Microbiol. 2021;87: 1–16. pmid:34085864
- 14. Lapierre P, Nazarian E, Zhu Y, Wroblewski D, Saylors A, Passaretti T, et al. Legionnaires’ disease outbreak caused by endemic strain of Legionella pneumophila, New York, New York, USA, 2015. Emerg Infect Dis. 2017;23: 1784–1791. pmid:29047425
- 15. Hamilton KA, Prussin AJ, Ahmed W, Haas CN. Outbreaks of Legionnaires’ Disease and Pontiac Fever 2006–2017. Curr Environ Heal reports. 2018;5: 263–271. pmid:29744757
- 16. De Giglio OD’Ambrosio M, Spagnuolo V, Diella G, Fasano F, Leone CM et al. Legionella anisa or Legionella bozemanii? Traditional and molecular techniques as support in the environmental surveillance of a hospital water network. Environ Monit Assess. 2023;195. pmid:36947259
- 17. Salloum G, Meugnier H, Reyrolle M, Grimont F, Grimont PAD, Etienne J, et al. Identification of Legionella species by ribotyping and other molecular methods. Res Microbiol. 2002;153: 679–686. pmid:12558187
- 18. Qasem JA, Mustafa AS, Khan ZU. Legionella in clinical specimens and hospital water supply facilities: Molecular detection and genotyping of the isolates. Med Princ Pract. 2008;17: 49–55. pmid:18059101
- 19. Bansal NS, McDonell F. Identification and DNA fingerprinting of Legionella strains by randomly amplified polymorphic DNA analysis. J Clin Microbiol. 1997;35: 2310–2314. pmid:9276408
- 20. Moss CW, Weaver RE, Dees SB, Cherry WB. Cellular fatty acid composition of isolates from legionnaires disease. J Clin Microbiol. 1977;6: 140–143. pmid:893658
- 21. Shevchuk O, Jäger J, Steinert M. Virulence properties of the Legionella pneumophila cell envelope. Front Microbiol. 2011;2: 1–12. pmid:21747794
- 22. Kowalczyk B, Chmiel E, Palusinska-Szysz M. The role of lipids in legionella-host interaction. Int J Mol Sci. 2021;22: 1–17. pmid:33540788
- 23. Barker J, Lambert PA, Brown MRW. Influence of intra-amoebic and other growth conditions on the surface properties of Legionella pneumophila. Infect Immun. 1993;61: 3503–3510. pmid:8335382
- 24. Cagarelli R, Caraglia A, La Mura S, Mele G, Ottaviani M, Pompa MG, et al. Linee guida per la prevenzione ed il controllo della legionellosi. Conferenza Stato-Regioni del 07/05/2015, Repertorio Atti n: 79/CSR del 07/05/2015. 2015. Available: https://www.salute.gov.it/imgs/C_17_pubblicazioni_2362_allegato.pdf
- 25. Querol A, Huerta T, Barrio E, Ramon D. Dry Yeast Strain For Use in Fermentation of Alicante Wines: Selection and DNA Patterns. J Food Sci. 1992;57: 183–185.
- 26. Iacumin L, Osualdini M, Bovolenta S, Boscolo D, Chiesa L, Panseri S, et al. Microbial, chemico-physical and volatile aromatic compounds characterization of Pitina PGI, a peculiar sausage-like product of North East Italy. Meat Sci. 2020;163: 108081. pmid:32062526
- 27. Vauterin L., Vauterin P. Computer-aided objective comparison of electrophoretic patterns for grouping and identification of microorganisms. Eur Microbiol. 1992;1: 37–41.
- 28. Huey B, Hall J. Hypervariable DNA fingerprinting in Escherichia coli: Minisatellite probe from bacteriophage M13. J Bacteriol. 1989;171: 2528–2532. pmid:2565332
- 29. Corich V, Mattiazzi A, Soldati E, Carraro A, Giacomini A. Sau-PCR, a novel amplification technique for genetic fingerprinting of microorganisms. Appl Environ Microbiol. 2005;71: 6401–6406. pmid:16204567
- 30. Moss CW, Bibb WF, Karr DE, Guerrant GO, Lambert MA. Cellular fatty acid composition and ubiquinone content of Legionella feeleii sp. nov. J Clin Microbiol. 1983;18: 917–919. pmid:6630470
- 31. Nalik HP, Muller KD, Ansorg R. Rapid identification of Legionella species from a single colony by gas-liquid chromatography with trimethylsulphonium hydroxide for transesterification. J Med Microbiol. 1992;36: 371–376. pmid:1613774
- 32. Verdon J, Labanowski J, Sahr T, Ferreira T, Lacombe C, Buchrieser C, et al. Fatty acid composition modulates sensitivity of Legionella pneumophila to warnericin RK, an antimicrobial peptide. Biochim Biophys Acta—Biomembr. 2011;1808: 1146–1153. pmid:21182824
- 33. Sasser M. Identification of Bacteria by Gas Chromatography of Cellular Fatty Acids. Tech Note. 2001;101: 1–6.
- 34. Vauterin L, Vauterin P. Computer-aided objective comparison of electrophoretic patterns for grouping and identification of microorganisms. Eur Microbiol. 1992;1: 37–41. Available: http://hdl.handle.net/1854/LU-232447
- 35. Paradis E, Claude J, Strimmer K. APE: Analyses of phylogenetics and evolution in R language. Bioinformatics. 2004;20: 289–290. pmid:14734327
- 36. Schliep KP. phangorn: Phylogenetic analysis in R. Bioinformatics. 2011;27: 592–593. pmid:21169378
- 37. Carson S. Interactive Web-Based Data Visualization with R, plotly, and shiny. Chapman and Hall/CRC; 2020. Available: https://plotly-r.com
- 38. Galili T. dendextend: An R package for visualizing, adjusting and comparing trees of hierarchical clustering. Bioinformatics. 2015;31: 3718–3720. pmid:26209431
- 39.
Wickham H. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York. Media. 2016. Available: http://link.springer.com/10.1007/978-0-387-98141-3
- 40. Moss CW, Karr DE, Dees SB. Cellular fatty acid composition of Legionella longbeachae sp. nov. J Clin Microbiol. 1981;14: 692–694. pmid:7334078
- 41. Lambert MA, Moss CW. Cellular fatty acid compositions and isoprenoid quinone contents of 23 Legionella species. J Clin Microbiol. 1989;27: 465–473. pmid:2715320
- 42. Tibshirani R, Walther G, Hastie T. Estimating the number of data clusters via the gap statistic. Journal of the Royal Statistical Society: Series B. 2001. pp. 411–423.
- 43. Pampaka D, Gómez-Barroso D, López-Perea N, Carmona R, Portero RC. Meteorological conditions and Legionnaires’ disease sporadic cases-a systematic review. Environ Res. 2022;214. pmid:35964674
- 44. Felice A, Franchi M, De Martin S, Vitacolonna N, Iacumin L, Civilini M. Environmental surveillance and spatio-temporal analysis of Legionella spp. In a region of northeastern Italy (2002–2017). PLoS One. 2019;14: 1–23. pmid:31287819
- 45. Haroon A, Koide M, Higa F, Hibiya K, Tateyama M, Fujita J. Repetitive element-polymerase chain reaction for genotyping of clinical and environmental isolates of Legionella spp. Diagn Microbiol Infect Dis. 2010;68: 7–12. pmid:20727463
- 46. Costa J, Tiago I, Da Costa MS, Veríssimo A. Presence and persistence of Legionella spp. in groundwater. Appl Environ Microbiol. 2005;71: 663–671. pmid:15691915
- 47. Wiese J, Helbig JH, Lück C, Meyer HGW, Jansen B, Dunkelberg H. Evaluation of different primers for DNA fingerprinting of Legionella pneumophila serogroup 1 strains by polymerase chain reaction. Int J Med Microbiol. 2004;294: 401–406. pmid:15595390
- 48. Zeybek Z, Türetgen I, Kimiran Erdem A, Filoǧlu G, Çotuk A. Profiling of environmental Legionella pneumophila strains by randomly amplified polymorphic DNA method isolated from geographically nearby buildings. Environ Monit Assess. 2009;149: 323–327. pmid:18283549
- 49. Sánchez-Busó L, Coscollà M, Palero F, Camaró ML, Gimeno A, Moreno P, et al. Geographical and temporal structures of Legionella pneumophila sequence types in Comunitat Valenciana (Spain), 1998 to 2013. Appl Environ Microbiol. 2015;81: 7106–7113. pmid:26231651
- 50. McAdam PR, Vander Broek CW, Lindsay DSJ, Ward MJ, Hanson MF, Gillies M, et al. Gene flow in environmental Legionella pneumophila leads to genetic and pathogenic heterogeneity within a Legionnaires’ disease outbreak. Genome Biol. 2014;15: 504. pmid:25370747
- 51. Sánchez-Busó L, Coscollá M, Pinto-Carbó M, Catalán V, González-Candelas F. Genetic Characterization of Legionella pneumophila Isolated from a Common Watershed in Comunidad Valenciana, Spain. PLoS One. 2013;8. pmid:23634210
- 52. Diogo A, Veríssimo A, Nobre MF, Da Costa MS. Usefulness of fatty acid composition for differentiation of Legionella species. J Clin Microbiol. 1999;37: 2248–2254. pmid:10364593
- 53. Mayberry WR. Monohydroxy and dihydroxy fatty acid composition of Legionella species. Int J Syst Bacteriol. 1984;34: 321–326.
- 54. Walker JT, Sonesson A, William Keevil C, White DC. Detection of Legionella pneumophila in biofilms containing a complex microbial consortium by gas chromatography‐mass spectrometry analysis of genus‐specific hydroxy fatty acids. FEMS Microbiol Lett. 1993;113: 139–144. pmid:8262363
- 55. Jantzen E, Sonesson A, Tangen T, Eng J. Hydroxy-fatty acid profiles of Legionella species: Diagnostic usefulness assessed by principal component analysis. J Clin Microbiol. 1993;31: 1413–1419. pmid:8314981
- 56. Palusińska-Szysz M, Choma A, Russa R, Droażański WJ. Cellular fatty acid composition from Sarcobium lyticum (Legionella lytica comb. nov.)—An intracellular bacterial pathogen of amoebae. Syst Appl Microbiol. 2001;24: 507–509. pmid:11876357
- 57. Lambert MA, Moss CW. Comparison of the effects of acid and base hydrolyses on hydroxy and cyclopropane fatty acids in bacteria. J Clin Microbiol. 1983;18: 1370–1377. pmid:6418758