Sasso Pisano Geothermal Field Environment Harbours Diverse Ktedonobacteria Representatives and Illustrates Habitat-Specific Adaptations

The hydrothermal steam environment of Sasso Pisano (Italy) was selected to investigate the associated microbial community and its metabolic potential. In this context, 16S and 18S rRNA gene partial sequences of thermophilic prokaryotes and eukaryotes inhabiting hot springs and fumaroles as well as mesophilic microbes colonising soil and water were analysed by high-throughput amplicon sequencing. The eukaryotic and prokaryotic communities from hot environments clearly differ from reference microbial communities of colder soil sites, though Ktedonobacteria showed high abundances in various hot spring samples and a few soil samples. This indicates that the hydrothermal steam environments of Sasso Pisano represent not only a vast reservoir of thermophilic but also mesophilic members of this Chloroflexi class. Metabolic functional profiling revealed that the hot spring microbiome exhibits a higher capability to utilise methane and aromatic compounds and is more diverse in its sulphur and nitrogen metabolism than the mesophilic soil microbial consortium. In addition, heavy metal resistance-conferring genes were significantly more abundant in the hot spring microbiome. The eukaryotic diversity at a fumarole indicated high abundances of primary producers (unicellular red algae: Cyanidiales), consumers (Arthropoda: Collembola sp.), and endoparasite Apicomplexa (Gregarina sp.), which helps to hypothesise a simplified food web at this hot and extremely nutrient-deprived acidic environment.


Introduction
The superheated steams and fumaroles near Sasso Pisano village (Castelnuovo Val di Cecina, Italy) are the central part of the Larderello geothermal field, located in the inner Northern Apennines (Southern Tuscany). The field exhibits extremely high subsurface temperatures (300-350 • C) at a pressure range of 4-7 MPa [1][2][3]. A vapour-dominated reservoir within a metamorphic basement at a depth of 3000 m is connected to a productive 1000 m deep carbonate-evaporite reservoir, which is almost sealed by overlying units. Many artificial steam wells were established for geothermal energy generation. The fumarolic fields of Sasso Pisano between Lagoni del Sasso and Monterotondo Marittimo represent a unique landscape of natural steam vents [3]. The overpressurised hot steam and springs release and/or redeposit CH 4 , CO 2 , sulphur compounds (H 2 S, SO 4 2− ), boric acid (H 3 BO 3 ), ammonia (NH 4 + ), and heavy metals at the surface [4,5], as water vapours flow (50 t/h) along fault zones/fractures through the sedimentary layers [3,6]. The condensation of sulphur dioxide-rich steam at the surface around fumaroles contributes to an acidic environment (pH 0.   [7], which also affects the surrounding area [8,9]. Each sample is labelled "SP" (Sasso Pisano), followed by numbers. The picture inserted in (A) represents the yellowish-orange SP2 sampling site, formed by the surface deposition of the sulphur vapours probably from a solfatara-type vent. Hot spring sites (A,D) were either of clear or whitish liquid discharge. Hot springs (B,C) were splashing muddy greyish/blackish water. The fumarole rocky wall was continuously wrapped in hot steam (H). The greenish biofilm forms a tight layer on the crumbly rock of the fumarole (SP5) and was compared with a cold and pH-neutral water stream (SP4) (I).

Environmental DNA Extraction, Amplicon Generation and Sequencing
The Power soil DNA isolation kit (Qiagen, Hilden, Germany) was used to extract the whole genomic DNA of the stored samples. The samples were subjected to three extractions (triplicates) each, according to the manufacturer's protocol. The extracted gDNA was then concentrated with the Eppendorf Concentrator Plus (Eppendorf, Hamburg, Germany) at 45 or 60 °C for 35 min to obtain the optimal gDNA concentration. Qualitative and quantitative determination of DNA was performed by gel electrophoresis and a Nanodrop instrument (NanoDrop 2000 c Spectrophotometer, ThermoFisher Scientific, Waltham, MA, USA). After quality assessment, partial 16S and 18S rRNA genes were amplified via polymerase chain reaction (PCR). For eukaryotes, the 18S amplicon PCR forward primer = 5′-TCG TCG GCA GCG TCA GAT GTG TAT AAG AGA CAG CCA GCA SCY GCG GTA ATT CC-3′ and the reverse primer = 5′-GTC TCG TGG GCT CGG AGA TGT GTA TAA GAG ACA GAC TTT CGT TCT TGA TYR A-3′ were used. For bacteria, the 16S amplicon forward primer 5′-TCG TCG GCA GCG TCA GAT GTG TAT AAG AGA CAG CCT ACG GGN GGC WGC AG-3′ and the reverse primer = 5′-GTC TCG TGG GCT CGG AGA TGT GTA TAA GAG ACA GGA CTA CHV GGG TAT CTA while (E-G) are representatives of the soil sampling sites. Each sample is labelled "SP" (Sasso Pisano), followed by numbers. The picture inserted in (A) represents the yellowish-orange SP2 sampling site, formed by the surface deposition of the sulphur vapours probably from a solfatara-type vent. Hot spring sites (A,D) were either of clear or whitish liquid discharge. Hot springs (B,C) were splashing muddy greyish/blackish water. The fumarole rocky wall was continuously wrapped in hot steam (H). The greenish biofilm forms a tight layer on the crumbly rock of the fumarole (SP5) and was compared with a cold and pH-neutral water stream (SP4) (I).

Environmental DNA Extraction, Amplicon Generation and Sequencing
The Power soil DNA isolation kit (Qiagen, Hilden, Germany) was used to extract the whole genomic DNA of the stored samples. The samples were subjected to three extractions (triplicates) each, according to the manufacturer's protocol. The extracted gDNA was then concentrated with the Eppendorf Concentrator Plus (Eppendorf, Hamburg, Germany) at 45 or 60 • C for 35 min to obtain the optimal gDNA concentration. Qualitative and quantitative determination of DNA was performed by gel electrophoresis and a Nanodrop instrument (NanoDrop 2000 c Spectrophotometer, ThermoFisher Scientific, Waltham, MA, USA). After quality assessment, partial 16S and 18S rRNA genes were amplified via polymerase chain reaction (PCR). For eukaryotes, the 18S amplicon PCR forward primer = 5 -TCG TCG GCA  GCG TCA GAT GTG TAT AAG AGA CAG CCA GCA SCY GCG GTA ATT CC-3 and the  reverse primer = 5 -GTC TCG TGG GCT CGG AGA TGT GTA TAA GAG ACA GAC TTT  CGT TCT TGA TYR A-3 were used. For bacteria, the 16S amplicon forward primer 5 -TCG  TCG GCA GCG TCA GAT GTG TAT AAG AGA CAG CCT ACG GGN GGC WGC AG-3  and the reverse primer = 5 -GTC TCG TGG GCT CGG AGA TGT GTA TAA GAG  CTA CHV GGG TAT CTA ATCC-3 were applied [12,13]. These primers were additionally linked to the overhang adapter sequences to make the amplicon suitable for Illumina MiSeq sequencing. Both partial 16S and 18S rRNA genes were amplified via a modified PCR Master mixture: Double-distilled nuclease-free water (32 µL) was mixed with 10 µL Phusion GC Buffer-5×, 10 µM forward and reverse primers (1.0 µL each), 5% DMSO (2.5 µL), 50 mM MgCl 2 (0.15 µL), 10 mM dNTPs (1.0 µL), 25 ng template DNA (2.0 µL) and 0.5 µL of 2 U/µL Phusion HF DNA polymerase (ThermoFisher Scientific) to obtain a final volume of 50 µL reaction mixture. A thermal cycler (Biometra, Göttingen, Germany) was used to anneal the 18S and 16S rRNA gene-specific primers with the extracted gDNA at 56 and 60 • C, respectively. PCR was performed as described [13,14]. The amplification products were qualitatively and quantitatively controlled on a 0.8% agarose gel with 1× TAE buffer (ThermoFisher Scientific) and a NanoDrop spectrophotometer (ThermoFisher Scientific). PCR products were further processed by using the GeneRead Size Selection Kit (Qiagen) to wash off primers and PCR reagents from the resulting amplicons. Indexing of OCR products was executed with the Nextera XT DNA library prep kit (Illumina, San Diego, CA, USA); subsequently, paired-end sequencing at the Göttingen Genomics Laboratory using the 2 × 300 bp Paired-End mode with an Illumina MiSeq sequencer (Illumina) was performed. The obtained Fastq sequences are accessible at the NCBI database under the Sasso Pisano microbiome project number PRJNA725822.

Processing of 16S rRNA Gene-Based Amplicon Sequencing Data
The MetaAmp automated pipeline, an online resource for metagenomic analysis, was utilised to analyse 16S rRNA gene-based amplicon sequencing data (http://ebg.ucalgary. ca/metaamp/, accessed on 18 November 2020) [15]. Initially, the USEARCH software merged the uncompressed and demultiplexed sequence data and removed low-quality reads [16]. The subsequent step was the removal of non-matching and misaligned read pairs of small lengths (less than 350 bp). Moreover, the Mothur software package was utilised for the identification and trimming of forward and reverse primers [17]. Any reads that lacked forward and reverse primers or containing primer mismatches were discarded. In the next step, the reads were dereplicated using UPARSE software version 7.1 [18]. This software also discarded chimaeras and singletons. The resulting, high-quality reads were assembled into taxonomic units (OTUs) based on the 0.97 OTU clustering threshold (97% identity). The Mothur software package assigned taxonomic status to the OTUs with reference to the SILVA training dataset (http://www.mothur.org/wiki/Taxonomy_outline, accessed on 18 November 2020). Lastly, the aforementioned software generated relative abundance data, rarefaction curves, as well as alpha and beta diversity indexes. The samples were rarefied into subsamples. The lack of similarity within different samples was computed through the Bray-Curtis index and principal coordinate analysis (PCoA) was performed. The Permutational Multivariate Analysis of Variance (PERMANOVA) test allowed for hypothesis testing.

Analysis of 18S rRNA Gene-Based Amplicon Sequencing Data
With respect to analysis of 18S rRNA gene-based data, paired-end reads were merged using PEAR (Paired-End reAd mergeR) [19] and resulting sequences of small lengths (less than 400 bp) were removed. Furthermore, the BBMap tool was used to remove all sequences that did not match the primer sequences at the 5 and 3 ends [20]. Subsequently, chimeric sequences were removed with the VSEARCH tool [21]. Data cleaned in this way were converted from the FASTQ file format to the FASTA format and mapped with QIIME (Quantitative Insights into Microbial Ecology) against the reference genes from the SILVA 132 18S database [22] with a 97% match [23].

Transmission Electron Microscopy (TEM)
For TEM sample preparation, the samples SP4 and SP5 were placed in a fixing solution (0.2%, v/v, glutardialdehyde, 0.3%, w/v, formaldehyde). Subsequently, they were washed with a 50 mM K 2 HPO 4 buffer solution and parts of the biofilm were embedded in agar (Agar-Agar, Kobe I, Carl Roth, Karlsruhe, Germany). Sliced small agar blocks of samples were dehydrated with ethanol in increasing concentration, followed by infiltration with synthetic resin (LR White Resin, London Resin Company, Berkshire, UK), and then, embedded in gelatine capsules, which were polymerised at 55 • C for 24 h.
To obtain ultra-thin sections, the polymerised samples were first pre-milled with a sample trimming device (TM 60, Reichert-Jung, Wetzlar, Germany), and sliced with the ultramicrotome (Ultracut E, Reichert-Jung) using a diamond knife (DuPont, Wilmington, DE, USA). The quality of the ultrathin sections was assessed based on their interference pattern and "fished" out of the water using coated grids (Plano, Wetzlar, Germany) as specimen support. For staining of ultrathin sections, uranyl acetate and replacement stain (UAR-EMS Uranyl Acetate Replacement Stain, Electron Microscopy Sciences, Hatfield, UK) were applied at different dilutions (undiluted, 1:3, 1:5, v/v). The grid with the section side facing downwards was placed on a drop of 20 µL of the respective staining agent and was incubated for ten minutes. The ultrathin sections were examined by transmission electron microscopy (JEM-1011 Electron Microscopes, JEOL, Akishima, Tokyo, Japan) at a voltage of 80 kV and electron micrographs were recorded.

Prokaryotic Community Functional Profiling
The MetaAmp-generated taxonomic profile data were utilised to generate a holistic overview of functional profiles of various samples. In this context, the Piphilin online server (http://secondgenome.com/Piphillin, accessed on 5 October 2020) [24] was used. Piphilin allows for the direct search of representative OTU sequences against a database composed of 16S rRNA gene sequences using USEARCH version 8.0.1623 [16]. In the next step, gene copy numbers of every inferred genome are summed to generate KO abundances (the KEGG reference database was used) [25]. The online available server MicrobiomeAnalyst [26] was utilised for the completion of statistical and meta-analysis of gene abundance data. The R package DESeq2 was used to compute a differential abundance analysis [27]. To summarise the findings, important features of hot sampling sites were considered in regard to soil samples that exhibited over a 1-fold difference.

Phylogenetic Analysis
The phylogenetic analysis was focused on 16S rRNA gene sequences, including those of different Ktedonobacteria type strains. MUSCLE, implemented in MEGA-X software version 7.0, was utilised to align these sequences and also Ktedonobacteria class-related OTU sequences [28]. Lastly, as part of the phylogenetic analyses, evolutionary distances were assessed. The Kimura 2-parameter model was used for this. The construction of phylogenetic trees was performed utilising the maximum likelihood method [29]. Bootstrap values were calculated based on 1000 replications.

Diversity of Ktedonobacteria
The 30 most abundant OTU sequences potentially affiliated with Ktedonobacteria were aligned with 16S rRNA gene sequences of isolated Ktedonobacteria strains and metagenomic sequences resulting from NCBI blast. In the phylogenetic tree, the OTUs are sandwiched between mesophilic (Dictyobacter) and thermophilic (Thermogemmatispora) groups of the Ktedonobacteria, indicating the presence of a large pool of uncultured mesophilic to thermophilic Ktedonobacteria strains. Most strains could be classified as members of Ktedonobacteraceae; however, OTUs related to JG30-KF-AS9 and B10-SB3A seem to be outliers and are more closely related to thermophilic Thermogemmatisporaceae (Figure 4). The detected OTUs indicate that the natural hydrothermal field of Sasso Pisano offers a rich reservoir of novel uncultured Ktedonobacteria, which should be targeted with respect to cultivation, genome sequencing, and further exploration. The highest abundance of Ktedonobacteraceae was detected in the SP1 biofilm, which indicates a natural enrichment of these bacteria at a hot spring vent ( Figure S2). More diverse genus members belonging to Ktedonobacteria were identified in another hot spring sample (SP310; whitish discharge).

Diversity of Ktedonobacteria
The 30 most abundant OTU sequences potentially affiliated with Ktedonobacteria were aligned with 16S rRNA gene sequences of isolated Ktedonobacteria strains and metagenomic sequences resulting from NCBI blast. In the phylogenetic tree, the OTUs are sandwiched between mesophilic (Dictyobacter) and thermophilic (Thermogemmatispora) groups of the Ktedonobacteria, indicating the presence of a large pool of uncultured mesophilic to thermophilic Ktedonobacteria strains. Most strains could be classified as members of Ktedonobacteraceae; however, OTUs related to JG30-KF-AS9 and B10-SB3A seem to be outliers and are more closely related to thermophilic Thermogemmatisporaceae ( Figure  4). The detected OTUs indicate that the natural hydrothermal field of Sasso Pisano offers a rich reservoir of novel uncultured Ktedonobacteria, which should be targeted with respect to cultivation, genome sequencing, and further exploration. The highest abundance of Ktedonobacteraceae was detected in the SP1 biofilm, which indicates a natural enrichment of these bacteria at a hot spring vent ( Figure S2). More diverse genus members belonging to Ktedonobacteria were identified in another hot spring sample (SP310; whitish discharge).

Alpha and Beta Diversity
For alpha and beta diversity analysis, sequence datasets were rarefied to the lowest detected read size per sample. The Shannon index indicated that some of the hot spring samples comprise a lower diversity than soil group samples ( Figure 5). The soil group potentially exhibited a higher alpha diversity in some cases, as the environmental conditions are comparably less harsh and facilitate a broad range of microorganisms to grow. In contrast, due to the acidic pH, higher temperature, and constant washout at the hot spring spots, only acidic thermophilic microbes are supposed to grow. However, the t-test result does not show a significant difference between the two groups, indicating that the soil microbial communities may provide an active microbial influx to the hot springs and vice versa ( Figure 5A).
tected read size per sample. The Shannon index indicated that some of the hot spring samples comprise a lower diversity than soil group samples ( Figure 5). The soil group potentially exhibited a higher alpha diversity in some cases, as the environmental conditions are comparably less harsh and facilitate a broad range of microorganisms to grow. In contrast, due to the acidic pH, higher temperature, and constant washout at the hot spring spots, only acidic thermophilic microbes are supposed to grow. However, the t-test result does not show a significant difference between the two groups, indicating that the soil microbial communities may provide an active microbial influx to the hot springs and vice versa ( Figure 5A).
To visualize and explore the complex metagenomic data, the multidimensional scaling method-Principal Coordinates Analysis (PCoA)-was applied to observe the similarities at the OTU level. The PCoA of Sasso Pisano samples showed that the microbial communities of hot springs and soil samples were clustering separately, with a considerable variance among the two groups ( Figure 5B). According to the Permutational Multivariate Analysis of Variance (PERMANOVA), the beta diversity among both groups also suggested the inhabitant microbial communities are distinct, indicating the stringent effect of hot spring constraints in selection and enrichment of a unique microbial consortium as compared to the soil microflora. The tree diagram, calculated based on a Bray-Curtis index, also indicated that both groups are clearly distinct ( Figure 5C). To visualize and explore the complex metagenomic data, the multidimensional scaling method-Principal Coordinates Analysis (PCoA)-was applied to observe the similarities at the OTU level. The PCoA of Sasso Pisano samples showed that the microbial communities of hot springs and soil samples were clustering separately, with a considerable variance among the two groups ( Figure 5B). According to the Permutational Multivariate Analysis of Variance (PERMANOVA), the beta diversity among both groups also suggested the inhabitant microbial communities are distinct, indicating the stringent effect of hot spring constraints in selection and enrichment of a unique microbial consortium as compared to the soil microflora. The tree diagram, calculated based on a Bray-Curtis index, also indicated that both groups are clearly distinct ( Figure 5C).

Functional Profile
The abundance of functional genes summarised by KO identifiers (also called K numbers) was estimated from the OTU data. Based upon this estimation, the predicted function of genes was used to evaluate a metabolic profile of the microbial communities in different sites. The differential abundance analysis of KO numbers indicated that the hot spring microbiota was metabolically rich in terms of heavy metal resistance, methane, sulphur and nitrogen metabolism as well as aromatic compounds' degradation pathways as compared to the soil microbial consortium. Numerous transporters involved in the extrusion of Mn, Zn, Cu, W, Co, Ni and Mo were present in the hot spring microbial consortium ( Table 1). The functional profile also showed that the different KO numbers associated with methane metabolism are differentially present in hot spring and soil samples ( Figure 6). This indicates that the hot spring microbiome utilises different enzymes as compared to the soil microbiome to metabolise environmental methane. The soil microbiome shows preferences for aerobic methane metabolism such as methane oxidation to formaldehyde and formate dehydrogenation to CO 2 , while the hot spring microbiome seems to favour anaerobic methane metabolism through reverse-methanogenesis by means of methyl-coenzyme M reductase (Mcr-AOM) [30,31] (Figure S3). Moreover, with respect to carbon fixation, mainly formaldehyde seems to be fixed via the ribulose-monophosphate pathway (RuMP) [32]. More KO numbers in the hot spring microbial communities were also observed to be differentially abundant, involved in sulphur and nitrogen metabolism (data not shown). In the aromatic compound degradation pathways, genes and enzymes involved in the aminobenzoate, benzoate, chlorocyclohexane, chlorobenzene, chloroalkane, chloroalkene, dioxin, fluorobenzoate, nitrotoluene, naphthalene, styrene, xylene and toluene degradation were also differentially abundant in the hot spring samples ( Figure S4).

Eukaryotic Diversity at a Fumarole
According to rarefaction curves, 300-450 OTUs were identified in the neutral pH and mesophilic stream conditions; the extreme fumarole site comprised less than 50 OTUs ( Figure S5). A total of 57 OTUs affiliated with known taxonomic groups were identified in stream sample (SP4) data, while the fumarole rock (SP5) data comprised a total of 21 groups, which shows a considerably higher alpha diversity in the water stream biome than in the fumarole biome. It can be concluded that the extreme conditions of the fumarole limit microbial alpha diversity, while microbial communities could easily populate the neutral pH water stream.
Organisms known to typically colonise freshwater habitats, including amoebae (11.2%) which can be divided into the groups Dactylopodida (1.1%), Filamoeba (7%) and Ischnamoeba sp. (2.3%), were detected. The Incerta Sedis represents a kind of placeholder for unspecified species in the genome database. The paraphyletic group of green algae is represented by Chlorophyta (0.75%) and Charophyta (4%). Among the multicellular organisms, Chaetonotida (10%), belonging to the phylum Gastrotricha ("hairybacks"), was detected. Furthermore, Adineta vaga, belonging to the rotifers, was found with 2.1% and the lower fungi were represented with a total of 5.7% relative abundance. The group of SAR (Stramenopiles, Alveolata, Rhizaria) showed 19.7% and LEMD267 is listed in SILVA as a nonspecific taxonomic group that refers to lobose amoebae (22.5%). However, this cannot be further determined with the available data. Excluding the unspecified eukaryotes (19%), 62.02% of the identified species OTUs (Amoeba 22.5%, Archaeplastida 0.82%, Metazoa 14.3%, SAR 19.7%, low mushrooms 4.7%) can be assigned to a biofilm in a freshwater area with a neutral pH and a lower temperature around 15 • C. The other major abundant taxa were Intramacronucleata (Ciliates) (16%) and Chaetonotida (10%). The OTU assigned to Rhodophyceae with 0.07% and Echinamoeba thermarum with 4.2% can be potentially explained by the association of the stream channel to the thermo-acidic sampling site, since both OTUs possess a thermophilic character.
The apparently qualitative identical species composition of the sample duplicates (SP4.1 and SP4.2) differs especially in terms of the fraction of nematode-and mite-related OTUs (Figure 7). The unspecified Cyanidiaceae and the species Galdieria sulphuraria belonging to the family Cyanidiales of phototrophic eukaryotic algae represent the most abundant taxa of the fumarole biofilm (65.3%). Unspecified Collembola (springtails) and Folsomia candida contributed to Collembola sp. (16.8%). Collembola sp. (Arthropoda: Hexapoda) represent more highly developed organisms, as they are among the typical bottom dwellers and have been described for other extreme sites, in particularly hot and particularly cold areas [33]. Gregarina caledia (Apicomplexa) contributed 14% of the biofilm consortium (Figure 7).

Structural and Morphological Description of a Eukaryotic Biofilm
The electron micrographs of a fumarole biofilm (SP5) preparation allowed inspection of individual cells, which are separated from each other by a cell wall of approx. 0.2 µm thickness. Within these cells, organelles, chloroplasts, in particular, are visible ( Figure 8). Inside chloroplasts, membrane stacks of the thylakoids are prominent features. Densely packed cells ( Figure 8A) are separated from each other by an extracellular matrix formed between the organisms. Remineralisation processes have caused dark precipitates to settle as small particles in the extracellular matrix. Larger particles are separated from the cells by the matrix (Figure 8A). The marked particles imitate the shape of the matrix.
Based on the number of spore cells within the mother cell, potentially either the genus Galdieria or Cyanidium could be identified. More than four spores in the cell represent a distinct structural feature of Galdieria (Figure 8B), while a sporangium with tetraspores could be classified as Cyanidium ( Figure 8C). Based on the species description in Ciniglia et al. (2004) [34], the organism ( Figure 8D,E) can be morphologically determined even more precisely as Galdieria sulphuraria under the electron microscope. The cell has a thick cell wall and contains at least seven visible chloroplast shapes in the cytosol, surrounding a central vacuole, visible as a membrane-bounded empty lumen. Ultrathin sections also reveal the coexistence of the bacterial colony outside the extracellular matrix of the Cyanidiaceae biofilm ( Figure 8F).
Microorganisms 2021, 9, x FOR PEER REVIEW 13 of 20 Figure 7. Eukaryotic community composition of a neutral pH (water stream biofilm SP4) and extreme (fumarole biofilm SP5) biome. Taxonomic groups are listed which show at least 0.80% of the total OTU abundance (except for the Rhodophyceae (0.07%) from the stream water sample to show the contrast to the fumarole and Chlorophyta to distinguish them from Charophyta). In the stream water sample, 21 species groups are listed which represent 98.02% abundance of the total OTUs, and six groups are shown in the fumarole samples, which represent 97.90% of the total OTUs (Unk, unknown).

Structural and Morphological Description of a Eukaryotic Biofilm
The electron micrographs of a fumarole biofilm (SP5) preparation allowed inspection of individual cells, which are separated from each other by a cell wall of approx. 0.2 μm thickness. Within these cells, organelles, chloroplasts, in particular, are visible ( Figure  8). Inside chloroplasts, membrane stacks of the thylakoids are prominent features. Eukaryotic community composition of a neutral pH (water stream biofilm SP4) and extreme (fumarole biofilm SP5) biome. Taxonomic groups are listed which show at least 0.80% of the total OTU abundance (except for the Rhodophyceae (0.07%) from the stream water sample to show the contrast to the fumarole and Chlorophyta to distinguish them from Charophyta). In the stream water sample, 21 species groups are listed which represent 98.02% abundance of the total OTUs, and six groups are shown in the fumarole samples, which represent 97.90% of the total OTUs (Unk, unknown). raspores could be classified as Cyanidium ( Figure 8C). Based on the species description in Ciniglia et al. (2004) [34], the organism ( Figure 8D,E) can be morphologically determined even more precisely as Galdieria sulphuraria under the electron microscope. The cell has a thick cell wall and contains at least seven visible chloroplast shapes in the cytosol, surrounding a central vacuole, visible as a membrane-bounded empty lumen. Ultrathin sections also reveal the coexistence of the bacterial colony outside the extracellular matrix of the Cyanidiaceae biofilm ( Figure 8F).

Discussion
Ktedonobacteria, a deeply branched bacterial class comprising of mesophilic and thermophilic representatives, is characterised by its ubiquitous prevalence in terrestrial environments, complex life cycle, and in some cases, large genome size [35]. The class is divided into two orders, Ktedonobacterales and Thermogemmatisporales. The latter order includes Thermosporotrichaceae as well as Thermogemmatisporaceae strains, whereas Ktedonobacterales comprise Dictyobacteraceae, Ktedonobacteraceae and Ktedonosporobacter-

Discussion
Ktedonobacteria, a deeply branched bacterial class comprising of mesophilic and thermophilic representatives, is characterised by its ubiquitous prevalence in terrestrial environments, complex life cycle, and in some cases, large genome size [35]. The class is divided into two orders, Ktedonobacterales and Thermogemmatisporales. The latter order includes Thermosporotrichaceae as well as Thermogemmatisporaceae strains, whereas Ktedonobacterales comprise Dictyobacteraceae, Ktedonobacteraceae and Ktedonosporobacteraceae strains. Within these two orders, 20 proposed mesophilic and thermophilic species have been assigned to the genera Dictyobacter, Tengunoibacter, Ktedonobacter, Ktedonosporobacter, Thermosporothrix and Thermogemmatispora [36][37][38][39]. Ktedonobacteria isolates and related environmental DNA are derived from non-extreme sources [38,39] and extreme environments, such as an acid vapour-formed spring [40], naturally occurring CO 2 gas vents [41], a lava cave in a volcanic trench [42], volcanic fumaroles [43], steaming geothermal soil [44], and a mineral precipitating cave environment [45]. This suggests that Ktedonobacteria members appear to prevail in oligotrophic and extreme environments, implying these strains may have evolved versatile metabolic pathways to cope with extreme conditions. For instance, the potential to oxidise carbon monoxide (CO) has been reported for Ktedonobacteria members [46].
In this study, we gained insights into the abundance and diversity of Ktedonobacteria in the geologically diverse and environmentally extreme hydrothermal field environments of Sasso Pisano. Microbial diversity varied greatly with the type of sampling site, which included hot springs, a fumarole, and nearby soil as well as water samples. A high abundance and diversity of Ktedonobacteria at the hot springs and in some soil samples indicated that the Sasso Pisano hydrothermal region offers a natural reservoir of Ktedonobacteria members, which became naturally enriched under stringent environmental constraints including heated surface soil, extremely low pH values, low nutritional content of nitrogen, carbon and phosphorus and high content of heavy metals as well as trace elements, especially around fumaroles [5,9,10]. The functional profile indicated that the hot spring microbiome includes a significantly higher abundance of organisms harbouring genes involved in methane, sulphur, nitrogen and aromatic compounds' metabolism, since the acidic hot springs and fumaroles emit a mixture of various gases such as ammonia, methane, carbon dioxide, hydrogen, hydrogen sulphide, hydrocarbons and aromatic compounds [2,5,47]. Moreover, transporters conferring resistance against toxic transition metals were also detected in the hot spring microbiome, which implies that Ktedonobacteria members inhabiting this site could have evolved high metabolic plasticity to cope with acidification, nutrient depletion, and heavy metal resistance. In our previous study of biofilm samples from Marsberg copper mine, Germany, the capability to degrade toxic aromatic compounds as well as resistance against transition metals was detected in a metagenome-assembled genome affiliated with Ktedonobacteria [48]. The bacterial biofilm on steam vents from Sapichu volcanoes was also mainly colonised by related similar taxa of Chloroflexi (Ktedonobacteria), Acidobacteria, and Cyanobacteria. The metabolic potential analysis predicted similar cellular metabolic pathways related to methanogenesis, sulphur respiration, nitrogen fixation, and heavy metal transport except for photosynthesis by Cyanobacteria [43]. The observed higher abundance of Cyanobacteria in low-temperature soil samples (but not in hot springs samples) was consistent with colonisation patterns of Cyanobacteria in lower temperature vents, as reported by Wall et al. [49], while the higher temperature vents were abundant in Cloroflexi [49,50].
Two important metabolic pathways in oligotrophic geothermal sites are attributed to trace gas (CO and H 2 ) utilisation by the type I carbon monoxide dehydrogenase (cox genes) and the [NiFe]-hydrogenase (hyp genes) as an energy source for cell growth and persistence under nutrient-limiting conditions [51]. Mounting genomic evidence suggests that three phyla, Chloroflexi, Actinobacteria and Acidobacteria, use CO and H 2 as substrates [52]. A variety of thermophilic bacteria belonging to the classes Actinobacteria, Deinococci, Ktedonobacteria, Thermomicrobia and Clostridia contain cox operons and may be capable of aerobic CO oxidation. Hot springs from different regions host different cox encoding communities [53]. The Ktedonobacteria genome also includes the reductive TCA cycle along with multiple copies of cox operons, conferring CO oxidation potential to this taxa, allowing it to predominate the microbial community [46] under the influence of CO-rich gas vents, hydrothermal springs, and soil environments [41,54,55]. Other high abundant taxa have evolved diverse strategies, for instance, most Anaerolineaceae species metabolise various organic carbon sources under anaerobic conditions through fermentative metabolism [56]. Representatives of the thermophilic Aquificae grow in hot springs via oxidation of dissolved ferrous iron or iron-containing minerals and conduct nitrogen fixation even at 70 • C [57,58].
Deinococci are resistant to several stresses due to their highly efficient DNA damage repair ability and detoxification of several toxic compounds through hydrolytic activity [59,60]. Some Oxyphotobacteria (photosynthetic Cyanobacteria) are metabolically diverse primary producers and pioneer the colonisation where light is available including ecosystems with low light, low levels of O 2 and/or sulfidic conditions, because of their ability to perform, besides oxygenic photosynthesis, anoxygenic photosynthesis [61,62].
The eukaryotic alpha diversity at the fumarole was significantly lower than in the river water sample, whereas the alpha diversity of prokaryotes did not vary significantly in the hot springs and soil samples. This indicates that thermophilic eukaryotic organisms are less diverse than thermophilic prokaryotes. Cyanidiales dominated the biofilm with an OTU of 65.3%, as they represent the only phototrophic organisms growing under these conditions and have been extensively studied since the 1980s [63]. The Cyanidiales form a monophyletic group within the red algae, which inhabits the acidic hot springs at different sites. Cyanidiales were identified under the electron microscope based on the chloroplasts and reproduction pattern, as the genera Cyanidium and Galdieria. Cyanidium sp. propagate with tetraspores, whereas Galdieria is larger and forms more autospores and vacuoles than Cyanidium [64]. Yoon et al. (2006) described the structural composition of endolithic and interlithic populations of Cyanidiales at Pisciarelli in the Phlegraean fields near Naples, in Sasso Pisano (distinct from the site sampled in this study) and Monte Rotondo [65]. Overall, Galdieria dominated the fumaroles of Sasso Pisano, which was expected due to the mixotrophy, the resulting adaptation to fluctuating environmental factors, and availability of carbon to be exploited. Cyanidium caldarium was also identified.
Current descriptions of the organisms around geothermally active areas around the world usually only illuminate the photoautotrophic community of the order Cyanidiales without considering the heterotrophic eukaryotes. The investigation of the entire diversity is interesting as the extreme conditions can serve as models for living conditions in the early evolutionary stages of soil ecosystems and can provide further information on the development of biocoenoses in the history of the earth. On the other hand, they represent a reduced biocoenosis that can serve as a model for more complex ecosystems. Collembola sp. (Hexapoda) represent the largest and most complex species at this site. Collembola are among the typical bottom dwellers and have been described for other extreme sites, in particularly hot and cold areas [34]. They are known as consumers of bacteria, algae, and fungi [66][67][68], actively feeding and thriving on these biomass sources. Gregarina sp. are known to endoparasitise Hexapoda among invertebrates in numerous colonies [69]. Based on the detected taxa, a simple food web could be assumed in which Cyanidium and Galdieria, belonging to unicellular red algae Cyanidiales, represent the primary producers and build biomass which may be consumed by Collembola sp. as consumers, and Gregarina sp. endoparasitise Collembola sp. Symbiotic bacteria may facilitate the biofilm growth of Cyanidiales. Collembola sp. could alter the microbial communities, either directly (through selectively feeding) or indirectly (through the dissemination of microbial propagules). Ultimately, a more detailed description of the species for this site would be required to confirm the described food web as a next step. This can then serve as a model to describe the influences and interactions of the organisms involved on each other and relate these to more complex ecosystems.

Conclusions
The composition and morphology of microbiota around hot fluid springs, steam vent and soil samples from the geothermally active hydrothermal fields around the Sasso Pisano site (Pisa, Tuscany region) are presented. The prokaryotic community analysis indicated that Ktedonobacteria at the Sasso Pisano site ranged from thermophilic to mesophilic, and phylogenetically as well as metabolically diverse representatives were detected. Metabolic functional profiling of the thermophilic prokaryotic microbiota predicted a higher capability to utilise carbon (methane and aromatic compounds), sulphur and nitrogen compounds, and heavy metal resistance-conferring genes were also significantly more abundant in the hot spring microbiome as compared to the mesophilic soil microbial consortium. The diversity of eukaryotic microorganisms within biofilms at extreme conditions, with very high temperature (55 • C) and a very low pH value of 2, was evaluated. The detection of red algae Cyanidiales, Arthropoda (Collembola sp.) and Apicomplexa (Gregarina sp.) proposed a simplified food web at the thermophilic extremely nutrient-deprived acidic environment of a fumarole.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/microorganisms9071402/s1, Table S1: Complete details of the sampling sites, Figure S1: Rarefaction curve, Figure S2: Number of sampled reads of the Ktedonobacteria members. Figure S3: Differentially abundant genes involved in the methane metabolism. Figure S4: Relative abundance of differentially abundant genes involved in the aromatic compounds metabolism. Figure S5. Eukaryotic rarefaction curve. Funding: This research was funded by a DAAD Doctoral research grant granted to S.A. and the APC was funded by the Open Access Publication Fund of the University of Göttingen.

Data Availability Statement:
The amplicon sequencing data presented here is accessible at the NCBI database under the Sasso Pisano microbiome project number PRJNA725822.