Abstract

In literature, the microbial diversity of hot spring waters has been the focus of extensive research, while there is a paucity of studies on thermal water distribution network of spa centres and, as yet, no studies have been carried out on the bacterial population of thermal muds. In this context, the aim of our study is to characterize the microbial community of sulphurous-bromine-iodine thermal water and mud within an Italian spa complex using Next Generation Sequencing (NGS) technologies. This is the first report of microbiome characterization along a water supply network from the spring to points of use within a spa. According to the 16S rRNA gene sequences analysis, our data highlight the presence of a typical microbial community, mainly composed of sulphur-cycling bacteria belonging to Desulfomonile, Thermodesulfovibrio, Geothermobacterium, Thermus, Thiofaba and Syntrophomonas genera. In addition, the characterization and evolution of the bacterial community in thermal muds during the maturation process is investigated for the first time. Interestingly, the microbiome of mature mud resulted dominated by bacteria capable of lipid biosynthesis, suggesting that these bacteria may play a role in the anti-rheumatic properties of thermal mud.

INTRODUCTION

Thermal waters are used for therapeutic purposes or recreational activities, in the form of baths, inhalation, irrigation and mud therapy (Carretero 2002; Costantino et al. 2006; Giampaoli et al. 2012; Forestier et al. 2016; Trabelsi et al. 2016). The specific therapeutic properties of thermal waters depend on their chemical and physical characteristics and the microbial diversity of these waters can indirectly contribute to their curative effects (Trabelsi et al. 2016). The peculiar properties of thermal waters, i.e. temperature, pH and concentration of salts, can create habitats suitable for the survival and multiplication of opportunistic pathogens such as legionellae (Hsu et al. 2006). In literature, the microbial diversity of hot spring waters has been extensively investigated (Miller et al. 2009; Everroad et al. 2012; Wemheuer et al. 2013; Headd & Engel 2014; Panosyan & Birkeland 2014; Panda et al. 2016; Chaudhuri et al. 2017), while there is a paucity of studies on water distribution networks within spas. Thermal muds are produced by mixing clayey materials with thermo-mineral waters, which are enriched with organic materials formed by the metabolic activity of microorganisms growing during the maturation process (Centini et al. 2015). Thermal muds have peculiar healing properties depending on the kind of clay minerals, the physical and chemical properties of the thermal water they contain and the growth and colonization of microorganisms (Carretero 2002; Veniale et al. 2007).

Over the last few years, some scientific studies have focused on the characterization of muds and evaluation of the maturation process, and in particular on the changes in the physicochemical properties and organic substances induced by maturation (Galzigna et al. 1995; Bruno et al. 2005; Centini et al. 2015). As regards the microflora analysis, only the presence of microalgae and diatoms has been investigated (Andreoli & Rascio 1975; Tolomio et al. 1999), therefore microbial community studies are still required.

Moreover, it is difficult to cultivate most environmental microorganisms under laboratory conditions, thus limiting our knowledge of the whole bacterial community (Badhai et al. 2015). In the last decade, the widespread use of Next Generation Sequencing (NGS) and bioinformatic tools has offered us a more extensive approach for examining the culturable and unculturable members of microbial communities (Everroad et al. 2012; Kittelmann et al. 2013; Wemheuer et al. 2013; Paul et al. 2016; Amin et al. 2017).

The Working Group ‘Movement Sciences for Health’ of the Italian Society of Hygiene has recently started a research project, Microflora Thermarum Atlas (MTA), aimed at characterizing the microbiome of Italian thermal springs with various therapeutic properties. We drew inspiration for our study from this project with the aim of characterizing the microbial community of water, flowing from springs to points of use, and of muds during the various maturation stages, within the thermal spa in Sirmione (Lombardia region, Northern Italy) using NGS technologies. Indeed, the thermal water and mud of Sirmione have a long history of exploitation, but their typical microflora has never been investigated.

The results of the pilot study are presented here.

MATERIAL AND METHODS

Sampling site

The current study was carried out in Sirmione, a lovely historical town located on a thin peninsula that protrudes into Lake Garda for approximately 4 km (Figure 1).

Figure 1

Geographic location of sampling sites in Sirmione area. Numbers indicate thermal water and mud samples collected (see legend). Legend. Water samples: 1=Drilling well V (t = 65.7 °C); 2 = Boiola spring at ‘Lido delle Bionde’ beach (t = 55.0 °C); 3 = Drilling well C (t = 63.8 °C); 4 = Drilling well A (t = 68.4 °C); 5 = Water inlet to Spa A (t = 63.3 °C). Water samples in thermal centre B: 6 = Water inlet (t = 44.1 °C); W1 = cold tank (t = 36.1 °C); W2 = hot tank (t = 51.3 °C); W3 = water for mud maturation (t = 44.7 °C); W4 = mixed tank (t = 39.0 °C); W5 = station for cold inhalation (t = 33.5 °C); W6 = station for hot inhalation (t = 49.7 °C). M1 = young mud; M2 = intermediate mud; M3 = mature mud. Please refer to the online version of this paper to see this figure in colour: http://dx.doi.org/10.1166/ws.2017.209.

Figure 1

Geographic location of sampling sites in Sirmione area. Numbers indicate thermal water and mud samples collected (see legend). Legend. Water samples: 1=Drilling well V (t = 65.7 °C); 2 = Boiola spring at ‘Lido delle Bionde’ beach (t = 55.0 °C); 3 = Drilling well C (t = 63.8 °C); 4 = Drilling well A (t = 68.4 °C); 5 = Water inlet to Spa A (t = 63.3 °C). Water samples in thermal centre B: 6 = Water inlet (t = 44.1 °C); W1 = cold tank (t = 36.1 °C); W2 = hot tank (t = 51.3 °C); W3 = water for mud maturation (t = 44.7 °C); W4 = mixed tank (t = 39.0 °C); W5 = station for cold inhalation (t = 33.5 °C); W6 = station for hot inhalation (t = 49.7 °C). M1 = young mud; M2 = intermediate mud; M3 = mature mud. Please refer to the online version of this paper to see this figure in colour: http://dx.doi.org/10.1166/ws.2017.209.

Sirmione thermal water is of meteoric origin and it springs in the watershed basin of Baldo Mount at an altitude of 800 m above sea level; it flowed underground for more than 2,100 m below sea level for almost 20 years and then rose up from the Boiola spring, located into the lake at a depth of 20 m, on the eastern side of the Sirmione peninsula (Figure 1). Three hundred metres of metal pipe (stainless steel AISI 316 L) transport the thermal water from the Boiola spring (Figure 1, red ball) to the peninsula (Figure 1, number 2). Besides the Boiola spring, thermal water has also been obtained by drilling three thermal water wells (Figure 1, numbers 1, 3 and 4). Thermal waters from spring and drilling wells are transported through stainless steel pipelines to a mixing plant located at ‘Lido delle Bionde’ beach (Figure 1, blue ball) where they are blended together and then distributed to hotels and spas through an aqueduct. Spas A (Figure 1, black ball) and B (Figure 1, green ball) are situated at the ends of the aqueduct: the former is approximately 500 m and the latter 5 km from the mixing plant at ‘Lido delle Bionde’ beach. The pipes transporting thermal water from the ‘Lido delle Bionde’ mixing plant to Spa A are made of stainless steel with a nominal diameter (DN) of 150 mm (water speed 0.97 m/s) and 200 mm (water speed 0.47 m/s), while the pipes transporting thermal water to the Spa B are mainly made of polypropylene with an outside diameter (OD) of 125 mm (water speed from 0.26 to 1.00 m/s, average speed 0.65 m/s). Water samples were collected at the inlet pipes to Spas A (Figure 1, number 5) and B (Figure 1, number 6) before treatment or storage. In addition, in the case of Spa B, which was the farthest point supplied by the thermal aqueduct, a further six samples were taken along the thermal water distribution pathway (Figure 1, numbers W1-W6). In the Spa B, indeed, the thermal water from the aqueduct is then stored in an outdoor tank (sample W2) from which the water is directed towards the inhalation treatment department inside the spa (sample W6) and to the outdoor pool where it undergoes the mud maturation process (sample W3). Part of the hot water stored cools down due to heat exchange with lake water. The cooled thermal water is then stored in a second tank (sample W1). Cold and hot thermal waters are mixed together in a third tank (sample W4) which is then used for cold inhalation treatments such as nasal irrigation (sample W5).

Water (n.12) and mud (n.3) samples were collected in the summer (June for water samples from n.1 to n.6 and July for water samples from W1 to W6 and muds) in the Sirmione area.

Sirmione thermal water is a hyper thermal water because of its temperature of 69 °C at the spring, with a high mineral content and a fixed residue at 180 °C >1,000 mg/L. Due to its chemical composition, it is defined by Marotta & Sica (1929) as sulphurous-bromine-iodine thermal water (Table 1), which is used for treating and preventing ear, nose and throat diseases, as well as bronchopneumologic, rheumatologic, orthopedic, dermatological and gynecological disorders.

Table 1

Chemical analysis of thermal water from spring and drilling wells (data from annual monitoring by environmental protection agency of Emilia-Romagna region – ARPAE – year 2015)

Chemical parametersMeasure UnitThermal water samples
Well V sample n.1Boiola spring sample n.2Well C sample n.3Well A sample n.4
pH 7.2 7.2 7.3 7.1 
Conductivity at 20 °C μS/cm 3,780 3,730 3,660 3,740 
Fixed residue at 180 °C mg/L 2,472 2,495 2,388 2,537 
Oxidability mg/L O2 19.6 18.8 17.2 18.8 
Silica (SiO2mg/L 58.8 58.8 55.3 61.8 
Bicarbonate (HCO3mg/L 316 318 317 313 
Chloride (Clmg/L 1,168 1,158 1,108 1,140 
Sulphate (SO4−−mg/L 145.3 152.5 141.9 159.7 
Sodium (Na+mg/L 643.0 653.4 638.6 670.0 
Potassium (K+mg/L 57.3 57.4 55.2 58.5 
Calcium (Ca++mg/L 175.4 177.7 170.5 177.0 
Magnesium (Mg++mg/L 34.0 34.4 33.8 34.1 
Ammonium (NH4+mg/L 1.85 1.82 1.79 1.88 
Hydrogen sulphide (H2S) mg/L 13.1 13.0 13.9 13.1 
Bromide (Brmg/L 5.2 5.0 4.7 5.0 
Iodide (Img/L <0.5 <0.5 <0.5 <0.5 
Fluoride (Fmg/L 4.25 4.25 4.10 4.78 
Nitrate (NO3mg/L 0.5 <0.1 0.5 0.7 
Nitrite (NO2mg/L <0.002 <0.002 <0.002 <0.002 
Chemical parametersMeasure UnitThermal water samples
Well V sample n.1Boiola spring sample n.2Well C sample n.3Well A sample n.4
pH 7.2 7.2 7.3 7.1 
Conductivity at 20 °C μS/cm 3,780 3,730 3,660 3,740 
Fixed residue at 180 °C mg/L 2,472 2,495 2,388 2,537 
Oxidability mg/L O2 19.6 18.8 17.2 18.8 
Silica (SiO2mg/L 58.8 58.8 55.3 61.8 
Bicarbonate (HCO3mg/L 316 318 317 313 
Chloride (Clmg/L 1,168 1,158 1,108 1,140 
Sulphate (SO4−−mg/L 145.3 152.5 141.9 159.7 
Sodium (Na+mg/L 643.0 653.4 638.6 670.0 
Potassium (K+mg/L 57.3 57.4 55.2 58.5 
Calcium (Ca++mg/L 175.4 177.7 170.5 177.0 
Magnesium (Mg++mg/L 34.0 34.4 33.8 34.1 
Ammonium (NH4+mg/L 1.85 1.82 1.79 1.88 
Hydrogen sulphide (H2S) mg/L 13.1 13.0 13.9 13.1 
Bromide (Brmg/L 5.2 5.0 4.7 5.0 
Iodide (Img/L <0.5 <0.5 <0.5 <0.5 
Fluoride (Fmg/L 4.25 4.25 4.10 4.78 
Nitrate (NO3mg/L 0.5 <0.1 0.5 0.7 
Nitrite (NO2mg/L <0.002 <0.002 <0.002 <0.002 

Water samples (2 L) were collected after flushing water for 1 min, filling the glass bottles up to the top and closing them with caps and under-caps in order to prevent the dispersion of H2S. At the same time as sampling, the water temperature was measured with a digital thermometer.

In the Sirmione spas, the muds are prepared in situ by maturation of clayey virgin materials mixed with sulphurous-bromine-iodine thermal water. Before use, the muds are matured in thermal water in open air pools located near Spa B for almost 6 months. The main clay mineral is smectite, a silicoaluminate, with the presence of illite and kaolinite. The mineral and chemical composition of this clay is reported in Table 2. Muds are used for treating rheumatic and dermatological diseases.

Table 2

Main mineralogical and chemical composition of virgin clay

ConstituentsPercentage (%)
Minerals 
 Smectite 60 
 Kaolinite 10 
 Illite 
 Calcite 10 
 Quartz, 
 Feldspars 
Chemical compounds 
 SiO2 50.46 
 Al2O3 17.56 
 Fe2O3 6.59 
 CaO 3.36 
 Na23.00 
 MgO 2.11 
ConstituentsPercentage (%)
Minerals 
 Smectite 60 
 Kaolinite 10 
 Illite 
 Calcite 10 
 Quartz, 
 Feldspars 
Chemical compounds 
 SiO2 50.46 
 Al2O3 17.56 
 Fe2O3 6.59 
 CaO 3.36 
 Na23.00 
 MgO 2.11 

Mud samples were collected at three stages of maturation: young (2 months old), intermediate (4 months old) and mature (6 months old). Mud (50 mL) was collected in a high-density polyethylene container with a cap and under-cap in order to avoid water contamination and evaporation; after mixing the mud on the surface and at the bottom of the maturation pool, the container was immersed and completely filled with mud. A sample of the thermal water used for the maturation process (sample W3) was taken simultaneously with the mud sample.

Characterization of microbial community

DNA extraction

Water (2 L) was filtered with a 0.47 μm polycarbonate membrane (Minerva BioLabs, Germany). The filter membrane was turned upside down onto an incubation dish filled with 2 mL of Lysis buffer (Minerva BioLabs, Germany) and then incubated at 37 °C for 30 min. Subsequently, the lysis solution was transferred into an incubation tube with 0.1 mg of glass beads (Sigma Aldrich, USA) and incubated at 56 °C for 15 min after vortexing for 1 min. DNA purification was carried out using Aqua screen Fast Extract Kit according to the manufacturer's protocol (Minerva BioLabs, Germany). For the mud samples (0.5 g), DNA was extracted according to the FastDNA®SPIN Kit (MP Biomedical, USA) protocol.

Bacterial community 16S profiling and bioinformatic analysis

The libraries were prepared in accordance with the ‘16S Metagenomic Sequencing Library Preparation’ guide (Part # 15044223 rev. A; Illumina, USA). The amplicon was obtained using primers with overhang adapters (Kittelmann et al. 2013). The libraries were quantified using the PicoGreen dsDNA quantitation assay (Thermo Fisher Scientific, USA) and validated on a Bioanalyzer DNA 1000 chip (Agilent, USA). Sequencing was performed on a MiSeq desktop sequencer (Illumina), following manufacturer's protocol. The sequence reads were analysed in the BaseSpace cloud environment using the 16S Metagenomics app (version 1.0.1; Illumina®): the taxonomic database used was the Illumina-curated version (May 2013 release of the Greengenes Consortium Database; Wang et al. 2013). All data sets were systematically screened to remove low quality reads (short reads >200 nt, zero-ambiguous sequences). Low abundance operational taxonomic units (OTUs) (<0.005%) were removed from the analysis. Relative abundances of community members were determined with rarefied data and summarized at each taxonomic level. OTU-based Shannon index (H′) was calculated with EstimateS software at a level of 97% sequence similarity (Colwell et al. 2012). Moreover, water samples were clustered into three temperature categories according to tertile distribution (group 1: T > 55 °C, group 2: 44 °C < T ≤ 55 °C, group 3: T ≤ 44 °C) in order to evaluate their phylogenetic distribution. Based on the Bray–Curtis index, the dissimilarity between groups was visualized using Principal Coordinates Analysis (PCoA) (METAGENassist platform; Arndt et al. 2012). The normalized abundance of the OTUs was square root-transformed before the analysis.

RESULTS AND DISCUSSION

Water samples

A total of 3,552,267 sequence reads were generated using the NGS approach, thus obtaining the reads for each sample after passing through filters ranging between 78,955 and 738,397 (Table 3). Rarefaction curves were calculated for each sample (Figure S1, available with the online version of this paper) and showed an adequate and reliable sampling and sequencing effort for describing the bacterial community. All of the curves reached a stable plateau, thus confirming that a larger number of sequences would not increase the number of OTUs obtained (Wen et al. 2017). The total number of sequences for each sample led to the identification of OTUs ranging between 373 to 655 defined at 97% identity. As determined by the Shannon index, the diversity of bacterial communities varied along the thermal water aqueduct (Table 3). This index variation followed the temperature trend, showing higher biodiversity in cooler samples, i.e. in sample 6 (H’ = 2.255; temperature 44 °C). Similarly, in thermal Spa B, samples W1, W4 and W5 with the lowest temperatures (36.1 °C, 39.0 °C and 33.5 °C respectively) showed the highest bacterial diversity and genus abundance. Other authors have observed that community complexity of hot springs increased inversely with temperature (Skirnisdottir et al. 2000; Miller et al. 2009; Everroad et al. 2012).

Table 3

Summary of NGS analysis after quality assessment step of sequences: the number of obtained reads after passing filter (PF) and the index of biodiversity, Shannon α-Diversity, are indicated for each sample

Sample IDNumber reads PFNumber OTUsShannon index
500,102 584 2.201 
216,857 491 1.258 
282,515 558 1.646 
739,397 655 1.950 
178,885 373 0.674 
327,139 599 2.255 
W1 410,980 491 1.672 
W2 151,193 355 1.083 
W3 256,982 393 1.147 
W4 240,793 478 1.789 
W5 78,955 385 1.859 
W6 168,469 321 1.088 
Sample IDNumber reads PFNumber OTUsShannon index
500,102 584 2.201 
216,857 491 1.258 
282,515 558 1.646 
739,397 655 1.950 
178,885 373 0.674 
327,139 599 2.255 
W1 410,980 491 1.672 
W2 151,193 355 1.083 
W3 256,982 393 1.147 
W4 240,793 478 1.789 
W5 78,955 385 1.859 
W6 168,469 321 1.088 

Based on 16S rRNA gene sequences analysis, a total of 12 main bacterial phyla were identified and their relative abundance was determined for each sample. Along the water network from the spring to the inlet points of the spas, the four dominant phyla were Proteobacteria (33.8–85.7%), Thermi (0.8–32.3%), Thermodesulfobacteria (1.8–40.1%) and Firmicutes (3.2–29.8%), as reported for hot spring waters by other authors (Wemheuer et al. 2013; Panda et al. 2016; Paul et al. 2016; Chaudhuri et al. 2017). There were only two dominant phyla in the samples collected in Spa B: Proteobacteria (53.0–83.5%) and Firmicutes (15.7–45.0%).

Analysis at genus level showed that the Boiola spring (sample 2) had a unique genus profile dominated by sulphate-reducing microorganisms such as Desulfomonile (75%) and Thermodesulfovibrio (3.8%). The large amount of sulphide present in this spring (13.0 mg/L) may be due to the presence of sulphate-reducing bacteria. Indeed, these bacteria reduce sulphate (SO42−) to hydrogen sulphide (H2S), further supporting the natural sulphur-related properties of these waters, including a possible natural antibacterial activity (Giampaoli et al. 2013). Each drilling well had a characteristic bacterial profile compared to the others, with a different distribution of some thermophilic genera such as Geothermobacterium (44.2% in sample 3), Thermus (33.5% in sample 1) and Thermodesulfovibrio (32.4% in sample 4). Geothermobacterium and Thermodesulfovibrio genera comprehend species with sulphate-reducing activity, while the oxidation of sulphur compounds has been reported as a common trait in members of the genus Thermus. Further analyses revealed that T. scotoductus was the only species present in all wells, which was previously isolated in a hot spring with a high sulphide concentration (Skirnisdottir et al. 2001). Thiobacter genus, a sulphur-oxidizing bacterium, was only found in drilling well 1 (27.9%). All genera described above comprise bacteria involved in sulphur metabolism which are expected to abound in these sulphide-rich habitats. Indeed, a correlation between the composition of microbial community and geochemical properties of hot springs has already been reported as a bidirectional phenomenon (Inskeep et al. 2013). A slight transformation in microbial community from spring and wells to distal points was observed at genus level (Figure 2). The water inlet to Spa A (sample 5) had a microbiome dominated by Geothermobacterium (32.8%), due to its temperature >60 °C. This genus comprehends thermophilic and sulphate-reducing bacteria. The other predominant genera were Thiofaba (23.7%) and Syntrophomonas (11.2%) including sulphur-oxidizing species. The drop in inlet water temperature to Spa B (sample 6) resulted in the growth of a large variety of genera, as confirmed by the Shannon index. The most dominant genera were Caloramator (20.1%), which had previously been isolated from hot springs (Seyfried et al. 2002) and geothermal waters (Ogg & Patel 2011) by other authors, and Thermodesulfovibrio (13.4%).

Figure 2

Distribution of genera determined by metagenomic NGS (cut off 2%) in all water samples.

Figure 2

Distribution of genera determined by metagenomic NGS (cut off 2%) in all water samples.

By analysing the dendrogram of samples collected in Spa B (Figure 3), the similarities observed between samples matched the water distribution within the centre. In more detail, the similarities observed between the hot water samples W2, W3 and W6 was related to the common origin of W3 and W6 from the tank (sample W2), while the similarities between the other three samples was attributed to the cooling process of thermal water. Thiofaba and Syntrophomonas were detected in all samples (Figures 2 and 3). Sulphate-reducing Desulfobacca was found in 4 out of 6 sections of the water distribution system. Sulphur-oxidizing bacteria belonging to Arcobacter genus were found in samples W1, W3 and W5. This genus has been already identified in an Armenian geothermal spring (Panosyan & Birkeland 2014) and in the Tully Valley sulphidic springs in New York (Headd & Engel 2014). Interestingly, genera such as Tepidimonas, described in sulphate bicarbonate waters containing alkaline earth metals were not detected in these sulphurous-bromine-iodine thermal waters, further supporting the association between the microflora and the chemical-physical conditions of the thermal spring (Coman et al. 2013; Valeriani et al. 2016).

Figure 3

Hierarchical clustering dendrogram of samples from thermal centre B based on genus-level classification. The bar chart shows the relative abundance of genera for each sample.

Figure 3

Hierarchical clustering dendrogram of samples from thermal centre B based on genus-level classification. The bar chart shows the relative abundance of genera for each sample.

As well as in the spring and wells, bacteria involved in sulphur metabolism remained dominant in the thermal water at distal points of the aqueduct. The recurrent finding of specific microbial groups may indicate a kind of biological signature for thermal waters, exhibiting a characteristic bio-fingerprint compared to other aquatic environments (Kemp & Aller 2004).

No genera enclosing opportunistic pathogens such as Legionella were detected in any of the water samples by the sequencing platform and protocol used. The cultural analysis carried out within a routine program for Legionella spp monitoring in Sirmione spas confirmed the negative results obtained with the metagenomics approach (data not shown).

Figure 4 describes the dissimilarity of water samples clustered for temperature. The PCoA plot shows that the range of PC1 mainly differed for the water samples of group 1 (from −31 to − 2.5) compared with those of group 2 (from − 24 to +28.2) and group 3 (from +14 to +26.3), with the exception of sample 6. Indeed, this sample differed from the others in group 2 and showed higher similarity with the samples in group 1. Since all of the water circulating in the Sirmione spas was obtained by mixing the water from the four sources (spring and drilling wells of the group 1), this result can be explained by the thermal water pathway: sample 6 was collected in Spa B before the hot storage tank, while the other samples in group 2 were collected after the tank. Similarly, for PC2-values, group 1 (from − 49.9 to +14.1) differed from groups 2 (from − 0.28 to +4.0) and 3 (from − 2.9 to +3.8). These observations support the effectiveness of the approach in identifying and tracing the different waters within the same plant, based on the water source and other parameters such as temperature. Within group 1, indeed, sample 2 differed from all of the others due to the drop in temperature that occurs when thermal water moves through the pipelines under Lake Garda from the Boiola spring to the point of sampling on ‘Lido delle Bionde’ beach. The lower temperature of sample 2 compared to the temperature of the samples collected from wells may explain its microbial profile dominated by only one genus accounting for the 75% of the genera found.

Figure 4

Principal coordinates analysis (PCoA) scatterplot of the normalized relative abundance of all samples, clustering for temperature level (group 1: T > 55 °C, group 2: 44 °C < T ≤ 55 °C, group 3: T ≤ 44 °C). Data are plotted following the genus-level classification.

Figure 4

Principal coordinates analysis (PCoA) scatterplot of the normalized relative abundance of all samples, clustering for temperature level (group 1: T > 55 °C, group 2: 44 °C < T ≤ 55 °C, group 3: T ≤ 44 °C). Data are plotted following the genus-level classification.

Mud samples

The microbial community of muds was characterized by three predominant genera: Pelobacter, Desulfomonile and Thiobacillus (Figure 5). The abundance of these genera varied during the maturation process, showing an increase of Pelobacter (from 14.7% in M1 to 19.2% in M3) and a decrease of both Desulfomonile (from 29.0% in M1 to 9.6% in M3) and Thiobacillus (from 9.7% in M1 to 6.3% in M3). This trend was confirmed by the dendrogram, which shows higher similarity between young and intermediate compared to mature mud. The pelobacter genus is involved in the biosynthesis of lipids (Sun et al. 2010). Galzigna et al. (1996) reported an increase in the phospholipid and sulpho-glycolipid fraction of mature mud. Moreover, Centini et al. (2015) demonstrated that the number of fatty acids increased during mud maturation. Interestingly, these and other studies have highlighted the fact that the lipid fraction of thermal mud has anti-inflammatory properties which may be the main reason for its anti-rheumatic effects (Galzigna et al. 1995; Tolomio et al. 1999; Bruno et al. 2005). Desulfomonile and Thiobacillus genera comprehend sulphate-reducing and sulphur-oxidizing bacteria, respectively. These genera were highly represented in the water but reduced during mud maturation, probably due to changes in physicochemical parameters and in the sulphur-driven biochemical pathways. Further studies are required in order to gain a better understanding of the mechanisms involved in the microflora modification during the mud maturation process. Even if these and other bacterial genera may play a major role in the water and mud microenvironments, a limit of this study is related to the focus only on bacteria taxa by the analysis of 16S rDNA, excluding other eukaryotic species belonging to the microflora such as protozoa or metazoans, which are known to play an important ecological role in biotic balance (Sanli et al. 2015). However, metagenomic database on these species are still less populated respect to the bacterial 16S rDNA database and their impact on the mature mud composition and therapeutic properties is still unknown. By analysing the data as reported in the dendrogram (Figure 5), the water used for mud maturation shows a microbial profile dissimilar to those of mud samples. Indeed, mud maturation is based on complex biological and biochemical processes involving interactions between virgin clay and thermal water and also depends on the habitats in the open-air pools in which the clay is left to mature (Veniale et al. 2007; Centini et al. 2015). These issues support the need for further studies to monitor and improve the mud maturation process, based on the expected final composition of the microflora and its active components.

Figure 5

Distribution of genera determined by metagenomic NGS in water for mud maturation (W3) and mud samples at different stages of maturation (M1 young; M2 intermediate; M3 mature).

Figure 5

Distribution of genera determined by metagenomic NGS in water for mud maturation (W3) and mud samples at different stages of maturation (M1 young; M2 intermediate; M3 mature).

CONCLUSION

The approach based on the use of NGS technologies for microbiome characterization in thermal water and muds provides us with valuable insight into the relationship between the microbial diversity of a thermal spring and its particular therapeutic properties. Moreover, sequencing-based microbial community analysis can provide us with new information on the unexpected presence of waterborne opportunistic bacteria which may be useful for selecting appropriate control measures aimed at guaranteeing the highest quality of the thermal waters and safeguarding the people using the spa facilities.

To the best of our knowledge, this is the first study on microbiome characterization along a water network from spring to points of use within spas. Several factors such as temperature, biofilms, pipe materials or treatments can interfere with natural water microflora. Despite the slight fluctuations in genera variety and abundance caused by temperature variations along the pipelines, our data highlight the presence of a typical microbial community, mainly composed of sulphur-cycling bacteria. This peculiar microbiome could be considered a kind of biological signature of Sirmione thermal water.

For the first time, the NGS technologies used for analysing thermal mud enabled the characterization and evolution of microbial community during the maturation process. The most important finding of our study is that the microbiome of mature mud is dominated by bacteria that are able to participate in lipid synthesis, which is deemed to be responsible for the beneficial anti-rheumatic properties of thermal mud.

In this pilot study, the analysis of only one sample per location has been carried out, and this aspect can represent a limit of our research. However, the lack of similar studies already published encouraged us to present these preliminary results in order to give a first information on the characterization of microbial populations within a long distribution network of thermal water. In addition, the relative stability of the microbial community in thermal waters from the spring to the points of use, as well as the detection of bacteria involved in the anti-rheumatic properties of the muds may be of interest for public health purposes. Further studies have been scheduled in order to deep our observations amplifying the number of samples and starting a periodic monitoring in order to identify the impact of eventual seasonal variations on the stability of both water microflora and mud maturation processes.

ACKNOWLEDGEMENTS

The authors would like to thank the staff at the Terme di Sirmione S.p.A for their kind support.

REFERENCES

REFERENCES
Amin
,
A.
,
Ahmed
,
I.
,
Salam
,
N.
,
Kim
,
B. Y.
,
Singh
,
D.
,
Zhi
,
X. Y.
,
Xiao
,
M.
&
Li
,
W. J.
2017
Diversity and distribution of thermophilic bacteria in hot springs of Pakistan
.
Microbial Ecology
74
(
1
),
116
127
.
Andreoli
,
C.
&
Rascio
,
N.
1975
The algal flora in the thermal baths of Montegrotto Terme (Padua). Its distribution over one-year period
.
International Review of Hydrobiology
60
(
6
),
857
871
.
Arndt
,
D.
,
Xia
,
J.
,
Liu
,
Y.
,
Zhou
,
Y.
,
Guo
,
A. C.
,
Cruz
,
J. A.
,
Sinelnikov
,
I.
,
Budwill
,
K.
,
Nesbø
,
C. L.
&
Wishart
,
D. S.
2012
METAGENassist: a comprehensive web server for comparative metagenomics
.
Nucleic Acids Research
40
(
Web Server issue
),
W88
W95
.
Bruno
,
A.
,
Rossi
,
C.
,
Marcolongo
,
G.
,
Di Lena
,
A.
,
Venzo
,
A.
,
Berrie
,
C. P.
&
Corda
,
D.
2005
Selective in vivo anti-inflammatory action of the galactolipid monogalactosyldiacylglycerol
.
European Journal of Pharmacology
524
(
1–3
),
159
168
.
Carretero
,
M. I.
2002
Clay minerals and their beneficial effects upon human health. A review
.
Applied Clay Science
21
(
2002
),
155
163
.
Centini
,
M.
,
Tredici
,
M. R.
,
Biondi
,
N.
,
Buonocore
,
A.
,
Maffei Facino
,
R.
&
Anselmi
,
C.
2015
Thermal mud maturation: organic matter and biological activity
.
International Journal of Cosmetic Science
37
(
3
),
339
347
.
Colwell
,
R. K.
,
Chao
,
A.
,
Gotelli
,
N. J.
,
Lin
,
S. Y.
,
Mao
,
C. X.
,
Chadzon
,
R. L.
&
Longino
,
J. T.
2012
Models and estimators linking individual-based and sample-based rarefaction, extrapolation, and comparison of assemblages
.
Journal of Plant Ecology
5
(
1
),
3
21
.
Coman
,
C.
,
Drugă
,
B.
,
Hegedus
,
A.
,
Sicora
,
C.
&
Dragos
,
N.
2013
Archaeal and bacterial diversity in two hot spring microbial mats from a geothermal region in Romania
.
Extremophiles
17
,
523
534
.
Costantino
,
M.
,
Lampa
,
E.
&
Nappi
,
G.
2006
Effectiveness of sulphur spa therapy with politzer in the treatment of rhinogenic deafness
.
Acta Otorhinolaryngologica Italica
26
(
1
),
7
13
.
Everroad
,
R. C.
,
Otaki
,
H.
,
Matsuura
,
K.
&
Haruta
,
S.
2012
Diversification of bacterial community composition along a temperature gradient at a thermal spring
.
Microbes and Environments
27
(
4
),
374
381
.
Forestier
,
R.
,
Erol Forestier
,
F. B.
&
Francon
,
A.
2016
Spa therapy and knee osteoarthritis: a systematic review
.
Annals of Physical and Rehabilitation Medicine
59
(
3
),
216
226
.
Galzigna
,
L.
,
Lalli
,
A.
,
Moretto
,
C.
&
Bettero
,
A.
1995
Maturation of thermal mud under controlled conditions and identification of an anti-inflammatory fraction
.
Physikalische Medizin Rehabilitationsmedizin Kurortmedizin
5
,
196
199
.
Galzigna
,
L.
,
Moretto
,
C.
&
Lalli
,
A.
1996
Physical and biochemical changes of thermal mud after maturation
.
Biomedicine & Pharmacotherapy
50
(
6–7
),
306
308
.
Giampaoli
,
S.
,
Valeriani
,
F.
&
Romano Spica
,
V.
2012
Thermal water for recreational use: overview of international standards
.
Igiene e Sanità Pubblica
68
(
6
),
863
873
.
Giampaoli
,
S.
,
Valeriani
,
F.
,
Gianfranceschi
,
G.
,
Vitali
,
M.
,
Delfini
,
M.
,
Festa
,
M. R.
,
Bottari
,
E.
&
Romano Spica
,
V.
2013
Hydrogen sulfide in thermal spring waters and its action on bacteria of human origin
.
Microchemical Journal
108
,
210
214
.
Headd
,
B.
&
Engel
,
A. S.
2014
Biogeographic congruency among bacterial communities from terrestrial sulfidic springs
.
Frontiers in Microbiology
8
(
5
),
473
.
Hsu
,
B. M.
,
Chen
,
C. H.
,
Wan
,
M. T.
&
Cheng
,
H. W.
2006
Legionella prevalence in hot spring recreation areas of Taiwan
.
Water Research
40
(
17
),
3267
3273
.
Inskeep
,
W. P.
,
Jay
,
Z. J.
,
Tringe
,
S. G.
,
Herrgård
,
M. J.
&
Rusch
,
D. B.
2013
The YNP metagenome project: environmental parameters responsible for microbial distribution in the Yellowstone geothermal ecosystem
.
Frontiers in Microbiology
4
,
67
.
Kittelmann
,
S.
,
Seedorf
,
H.
,
Walters
,
W. A.
,
Clemente
,
J. C.
,
Knight
,
R.
,
Gordon
,
J. I.
&
Janssen
,
P. H.
2013
Simultaneous amplicon sequencing to explore co-occurrence patterns of bacterial, archaeal and eukaryotic microorganisms in rumen microbial communities
.
PLoS One
8
(
2
),
e47879
.
Marotta
,
D.
&
Sica
,
C.
1929
Classificazione delle acque minerali italiane
.
Giornale di Chimica industriale ed applicata
11
(
6
),
276
.
Panosyan
,
H.
&
Birkeland
,
N. K.
2014
Microbial diversity in an Armenian geothermal spring assessed by molecular and culture-based methods
.
Journal of Basic Microbiology
54
(
11
),
1240
1250
.
Paul
,
S.
,
Cortez
,
Y.
,
Vera
,
N.
,
Villena
,
G. K.
&
Gutiérrez-Correa
,
M.
2016
Metagenomic analysis of microbial community of an Amazonian geothermal spring in Peru
.
Genomics Data
9
,
63
66
.
Sanli
,
K.
,
Bengtsson-Palme
,
J.
,
Nilsson
,
R. H.
,
Kristiansson
,
E.
,
Alm Rosenblad
,
M.
,
Blanck
,
H.
&
Eriksson
,
K. M.
2015
Metagenomic sequencing of marine periphyton: taxonomic and functional insights into biofilm communities
.
Frontiers in Microbiology
6
,
1192
.
Seyfried
,
M.
,
Lyon
,
D.
,
Rainey
,
F. A.
&
Wiegel
,
J.
2002
Caloramator viterbensis sp. nov., a novel thermophilic, glycerol-fermenting bacterium isolated from a hot spring in Italy
.
International Journal of Systematic and Evolutionary Microbiology
52
(
Pt 4
),
1177
1184
.
Skirnisdottir
,
S.
,
Hreggvidsson
,
G. O.
,
Hjörleifsdottir
,
S.
,
Marteinsson
,
V. T.
,
Petursdottir
,
S. K.
,
Holst
,
O.
&
Kristjansson
,
J. K.
2000
Influence of sulfide and temperature on species composition and community structure of hot spring microbial mats
.
Applied and Environmental Microbiology
66
(
7
),
2835
2841
.
Skirnisdottir
,
S.
,
Hreggvidsson
,
G. O.
,
Holst
,
O.
&
Kristjansson
,
J. K.
2001
Isolation and characterization of a mixotrophic sulfur-oxidizing Thermus scotoductus
.
Extremophiles: Life Under Extreme Conditions
5
(
1
),
45
51
.
Sun
,
J.
,
Haveman
,
S. A.
,
Bui
,
O.
,
Fahland
,
T. R.
&
Lovley
,
D. R.
2010
Constraint-based modeling analysis of the metabolism of two pelobacter species
.
BMC Systems Biology
4
,
174
.
Tolomio
,
C.
,
Ceschi-Berrini
,
C.
,
Moschin
,
E.
&
Galzigna
,
L.
1999
Colonization by diatoms and anti-rheumatic activity of thermal mud
.
Cell Biochemistry and Function
17
(
1
),
29
33
.
Valeriani
,
F.
,
Biagini
,
T.
,
Giampaoli
,
S.
,
Crognale
,
S.
,
Santoni
,
D.
&
Romano Spica
,
V.
2016
Draft genome sequence of Tepidimonas taiwanensis strain VT154-175
.
Genome Announcements
4
(
5
),
e00942
16
.
Veniale
,
F.
,
Bettero
,
A.
,
Jobstraibizer
,
P. G.
&
Setti
,
M.
2007
Thermal muds: perspectives of innovations
.
Applied Clay Science
36
,
141
147
.
Wang
,
S.
,
Hou
,
W.
,
Dong
,
H.
,
Jiang
,
H.
,
Huang
,
L.
,
Wu
,
G.
,
Zhang
,
C.
,
Song
,
Z.
,
Zhang
,
Y.
,
Ren
,
H.
,
Zhang
,
J.
&
Zhang
,
L.
2013
Control of temperature on microbial community structure in hot springs of the Tibetan plateau
.
PLoS One
8
(
5
),
e62901
.
Wen
,
C.
,
Wu
,
L.
,
Qin
,
Y.
,
Van Nostrand
,
J. D.
,
Ning
,
D.
,
Sun
,
B.
&
Zhou
,
J.
2017
Evaluation of the reproducibility of amplicon sequencing with Illumina MiSeq platform
.
PLoS One
12
(
4
),
e0176716
.

Supplementary data